Accuracy of the European Society of Cardiology 0/1-, 0/2-, and 0/3-hour Algorithms for Diagnosing Acute Myocardial Infarction

Diagnostics and Likelihood Ratios, Explained

Positive Findings (Patient Has This)

Finding Increased Disease Probability (Positive Likelihood Ratio)
0/2-h algorithm for ruling out AMI 21
0/1-h 14
0/3-h 13

Negative Findings (Patient Doesn't Have This)

Finding Decreased Disease Probability (Negative Likelihood Ratio)
0/1-h algorithm for ruling out AMI 0.01
0/2-h 0.02
0/3-h 0.09

Narrative

Chest pain is a common presentation to the emergency department (ED), representing over 5 million annual ED visits in the United States.1
However, only 10%–20% of ED patients with chest pain are ultimately diagnosed with acute myocardial infarction (AMI).2 Evaluation using accelerated diagnostic protocols with high- sensitivity cardiac troponin (hs-cTn) tests can provide more rapid detection of AMI and earlier discharge for patients in whom AMI has been excluded when compared to standard troponin assays.3, 4 Recent guidelines utilize hs-cTn in evaluation for AMI,5 while prior guidelines incorporated standard troponin assays at the initial time of presentation and at 3 h.6, 7, 8 The most recent European Society of Cardiology (ESC) guidelines use specific hs-cTn T or I thresholds at 0 h and 1 or 2 h and absolute changes to determine who may be ruled in or ruled out for AMI as well as those who require observation.5

The systematic review and meta-analysis discussed here included prospective observational cohort studies, implementation studies, and randomized controlled trials (RCTs) evaluating ESC hs-cTn protocols. Studies included adult patients in the ED or chest pain unit with suspected non-ST elevation myocardial infarction or acute coronary syndrome.9 The authors only included studies evaluating the 0/1-, 0/2-, and 0/3-h ESC protocols utilizing the Elecsys hs-cTnT (Roche), Architect hs-cTnI (Abbott), and Centaur/Atellica hs-cTnI (Siemens) assays based on 2015 ESC guideline thresholds.7 Diagnosis of AMI was determined based on the Third or Fourth Global Task Force Universal Definition of Myocardial Infarction.10, 11 The primary outcome was diagnostic accuracy for AMI using the 0/1-, 0/2-, and 0/3-h protocols.

The systematic review identified 32 publications (n = 30,066 patients; 4246 cases of AMI) conducted in Asia, Australasia, Europe, and North America that met the inclusion criteria.9 Among these 32 papers, they identified 33 total cohorts, of which 20 were unique subgroups. Sixteen cohorts evaluated the 0/1-h algorithm, seven cohorts the 0/2-h algorithm, and 10 cohorts the 0/3-h algorithm. Eighteen subgroups were evaluated using an observational study type, while one was evaluated with an RCT and one was quasi-experimental. The authors obtained primary data from the principal investigators for 16 subgroups and aggregate data from published articles for four subgroups. The prevalence of AMI ranged from 4% to 37% in the included studies.

The 0/1-h algorithm demonstrated high diagnostic accuracy to rule in (specificity = 94.0%, 95% confidence interval [CI] = 90.7%–96.2%; positive likelihood ratio [LR+] = 14, 95% CI = 9–20) and to rule out AMI (sensitivity = 99%, 95% CI = 98.5%–99.5%; negative likelihood ratio [LR−] = 0.01, 95% CI = 0.01–0.02), with 17% of patients ruled in and 54% of patients ruled out. The 0/2-h algorithm demonstrated similar test characteristics, with high accuracy for ruling in (specificity = 96%, 95% CI = 92.9%–97.9%; LR+ = 21, 95% CI= 13–35) and ruling out AMI (sensitivity = 98.6%, 95% CI = 97.2%–99.3%; LR− = 0.02, 95% CI = 0.01–0.04), with 15% of patients ruled in and 61% of patients ruled out. The 0/3-h algorithm had similar ability to rule in AMI (specificity = 93%, 95% CI = 86.9%–96.6%; LR+ = 13, 95% CI = 6.7–24), but lower ability to rule out AMI (sensitivity = 93.7%, 95% CI = 87.4%–97.0%; LR− = 0.09, 95% CI = 0.05–0.15), with 19% of patients ruled in and 66% of patients ruled out. The proportion of patients who remained undifferentiated or in the observational zone was 29% for the 0/1-h algorithm, 24% for the 0/2-h algorithm, and 15% for the 0/3-h algorithm. Stratification by assay demonstrated similar sensitivities and specificities.

Caveats

This study has several important limitations. While all three ESC algorithms had high specificities, there was significant heterogeneity across the included studies. This heterogeneity may be in part due to variation in AMI prevalence. Populations with a higher prevalence of AMI may also have a higher prevalence of non-AMI conditions associated with troponin elevations. The heterogeneity could also be due to differences in pinclusion and exclusion criteria used by different studies, with studies including only patients with chest pain having higher specificities compared to studies which included patients presenting with other symptoms. There were significant differences in implementation of the 0/3-h protocols among the included studies as well as potential miscalibration of the Elecsys hs-cTnT lots used globally from 2010 to 2012, further contributing to the heterogeneity. The lower sensitivity of the 0/3-h protocol and wide CIs could be due to the fact that some studies used this algorithm without clinical criteria (GRACE score <140 and pain-free) while others studies used it combined with clinical criteria (sensitivity without and with clinical criteria = 90% [95% CI = 82.9%–94.6%] vs. 98% [95% CI = 88.6%–99.8%], respectively). In fact, protocol performance may be suboptimal in some populations such as those at high-risk for AMI clinically.12 The performance of algorithms for diagnosis of AMI using sex-specific 99th percentile thresholds are still unclear as many studies did not adjudicate based on this factor. Several studies used samples collected in a different timing manner than that recommended in ESC guidelines, such as using a 0–2/3-h blood draw instead of a 0/2- or 0/3-h draw,13 and the systematic review authors included studies analyzing troponin samples collected over 30 min outside of the stipulated time by the ESC algorithms. The authors were unable to obtain individual patient-level data and could not account for patients falsely ruled in or out by the ESC algorithms. While most of these studies used the third universal definition of AMI, the adjudication criteria in each of these studies likely means that the difference between third and fourth universal definitions are unlikely to substantially bias the results. Finally, patient management was not dictated by the algorithms and, therefore, outcomes may have been influenced by troponin levels and clinical management.

The original manuscript was published in Academic Emergency Medicine as part of the partnership between TheNNT.com and AEM.

Author

Brit Long, MD; Michael Gottlieb, MD
Supervising Editors: Shahriar Zehtabchi, MD

Published/Updated

February 24, 2022

What are Likelihood Ratios?

LR, pretest probability and posttest (or posterior) probability are daunting terms that describe simple concepts that we all intuitively understand.

Let's start with pretest probability: that's just a fancy term for my initial impression, before we perform whatever test it is that we're going to use.

For example, a patient with prior stents comes in sweating and clutching his chest in agony, I have a pretty high suspicion that he's having an MI – let's say, 60%. That is my pretest probability.

He immediately gets an ECG (known here as the "test") showing an obvious STEMI.

Now, I know there are some STEMI mimics, so I'm not quite 100%, but based on my experience I'm 99.5% sure that he's having an MI right now. This is my posttest probability - the new impression I have that the patient has the disease after we did our test.

And likelihood ration? That's just the name for the statistical tool that converted the pretest probability to the posttest probability - it's just a mathematical description of the strength of that test.

Using an online calculator, that means the LR+ that got me from 60% to 99.5% is 145, which is about as high an LR you can get (and the actual LR for an emergency physician who thinks an ECG shows an obvious STEMI).

(Thank you to Seth Trueger, MD for this explanation!)

References:

Ankle-Brachial Index for Diagnosis of Arterial Injury in Penetrating Extremity Trauma

Diagnostics and Likelihood Ratios, Explained

Positive Findings (Patient Has This)

Finding Decreased Disease Probability (Negative Likelihood Ratio)
Ankle-brachial index: Sensitivity = 49% (95% CI = 39% to 60%) Not reported

Negative Findings (Patient Doesn't Have This)

Finding Decreased Disease Probability (Negative Likelihood Ratio)
Ankle-brachial index: Specificity = not reported (data not pooled due to significant heterogeneity) 0.59 (95% CI = 0.49 to 0.72)
Ankle-brachial index without any hard or soft signs (proximity to major artery was not studied) 0.01 (95% CI = 0.0 to 0.1)

Narrative

Penetrating extremity trauma (PET) is a common cause of arterial injury in the United States.1 PET can result in amputation, wound infection, venous thromboembolism, need for surgical interventions such as fasciotomy, and death.2 The diagnosis of arterial injury in patients with PET has been the subject of debate with differing recommendations by professional society and major trauma guidelines (Table 1).3, 4

The systematic review discussed here assesses the accuracy of the ankle–brachial index (ABI) in the diagnosis of arterial injury in patients with PET.5 The systematic review identified five prospective studies of 1,040 adult patients with upper or lower PET who underwent ABI. The prevalence of arterial injury was 14.3%. Due to considerable heterogeneity (I2 >75%), the authors chose not to calculate a pooled positive likelihood ratio (LR+) for ABI.6 The pooled negative likelihood ratio (LR–) for ABI was 0.59. Although this review concluded that ABI cannot independently exclude arterial injury, the systematic review suggested that such injuries can be excluded in patients with no hard or soft signs and a normal ABI (≥0.9).

Caveats

There are some limitations of the systematic review and metaanalysis. While all studies were prospective and the majority at low risk of bias, a considerable degree of heterogeneity existed among studies with regard to physical examination findings for arterial injury. The authors attributed this heterogeneity to the varying definition of hard and soft signs of arterial injury used across the included trials that originated from discrepancies in major trauma guidelines (Table 1).

Table 1. Definition of hard and soft signs in trauma guidelines
Image from Gyazo


Another limitation of the analysis is that not all patients received the reference standard diagnostic study. In almost all of the included trials, low-risk patients were observed for 24 h and did not undergo any of the predetermined reference tests: CT angiography, catheter angiography, or surgical exploration. The probability of clinically significant arterial injury (requiring intervention) in this population is likely to be low but cannot be accurately reported. A significant portion of these “low-risk” patients were then discharged and were lost to follow-up. This, therefore, creates risk of partial and differential verification bias.

Due to considerable heterogeneity, the systematic review did not report the posttest probability of arterial injury in patients with no hard signs of vascular injury. However, in patients with no hard or soft signs of vascular injury, a systematic review concluded that arterial injury can be ruled out if ABI is normal (≥0.9). This conclusion was drawn from two trials that included a group of PET patients without any hard or soft signs of arterial injury that underwent ABI testing. The review reports a pooled LR– of 0.01 for negative ABI in absence of soft or hard signs of vascular injury. The authors used the weighted prevalence of arterial injury in this subgroup of patients (16.3%) as an estimate of the pre-test probability and applied LR– of 0.01, arriving at a posttest arterial injury probability of 0% (95% CI = 0% to 1%). Based on this calculation, they suggested that this subset of patients may not need further testing. One limitation of this recommendation is its applicability in patients with PETs when the trajectory of injury is in proximity to a major artery. In contrast to the guidelines,3, 4 these two studies did not classify trajectory of injury in proximity to a major artery as a soft sign. In patients with such injuries, therefore, the review recommends that clinicians use their clinical judgment to decide whether to discharge, observe, or obtain further testing.

In summary, the existing evidence suggests that a normal ABI in the absence of soft or hard signs may rule out the presence of arterial injury in patients suffering from penetrating extremity injury. The data are insufficient to draw any conclusions about the use of ABI in any other penetrating extremity injury scenario.

The original manuscript was published in Academic Emergency Medicine as part of the partnership between TheNNT.com and AEM.

Author

Roshanak Benabbas, MD; Ian S. deSouza, MD
Supervising Editor: Kabir Yadav, MD

Published/Updated

April 30, 2021

What are Likelihood Ratios?

LR, pretest probability and posttest (or posterior) probability are daunting terms that describe simple concepts that we all intuitively understand.

Let's start with pretest probability: that's just a fancy term for my initial impression, before we perform whatever test it is that we're going to use.

For example, a patient with prior stents comes in sweating and clutching his chest in agony, I have a pretty high suspicion that he's having an MI – let's say, 60%. That is my pretest probability.

He immediately gets an ECG (known here as the "test") showing an obvious STEMI.

Now, I know there are some STEMI mimics, so I'm not quite 100%, but based on my experience I'm 99.5% sure that he's having an MI right now. This is my posttest probability - the new impression I have that the patient has the disease after we did our test.

And likelihood ration? That's just the name for the statistical tool that converted the pretest probability to the posttest probability - it's just a mathematical description of the strength of that test.

Using an online calculator, that means the LR+ that got me from 60% to 99.5% is 145, which is about as high an LR you can get (and the actual LR for an emergency physician who thinks an ECG shows an obvious STEMI).

(Thank you to Seth Trueger, MD for this explanation!)

References:

Diagnostic Accuracy of the History, Physical Examination, and Laboratory Testing for Giant Cell Arteritis

Diagnostics and Likelihood Ratios, Explained

Positive Findings (Patient Has This)

Finding Increased Disease Probability (Positive Likelihood Ratio)
Limb claudication 6.0
Jaw claudication 4.9
Temporal artery thickening 4.7
Temporal artery loss of pulse 3.3
Platelet count >400 × 103/μL 3.8
Temporal tenderness 3.1
ESR >100 mm/h 3.1

Negative Findings (Patient Doesn't Have This)

Finding Decreased Disease Probability (Negative Likelihood Ratio)
ESR <40 mm/h 0.18
C reactive protein <2.5 mg/dL 0.38
Age <70 years 0.48

Narrative

Giant cell arteritis (GCA), an inflammation of temporal arteries, can lead to multiple complications including vision loss if not diagnosed promptly.1 Unfortunately, diagnoses can often be delayed, particularly in those without classic features.2 Once the diagnosis is suspected, patients are typically started on high-dose glucocorticoids and referred for urgent temporal artery biopsy (TAB) or advanced imaging (e.g., computed tomography, magnetic resonance imaging). Unfortunately, high-dose glucocorticoids have significant side effects. Moreover, TAB is not sensitive enough to exclude the disease.3, 4 Therefore, it is important to incorporate the pre-test probability of GCA by using the history, examination, and laboratory findings in the clinical decision-making.

The systematic review discussed here included studies that evaluated historical features, physical examinations findings, and laboratory testing to predict the risk of GCA.5 The systematic review included studies that enrolled a consecutive sample of patients suspected of having GCA. Temporal artery biopsy, imaging, or clinical diagnosis were used as the reference standard for GCA. Studies should have included at least 5 patients with GCA and at least 5 patients that did not have GCA. Risk of bias was assessed with the quality assessment of diagnostic accuracy studies (QUADAS-2) tool.

The systematic review identified 68 studies (n = 14,037 patients), of whom 4,277 (30.5%) were classified as having GCA. Temporal artery biopsy was utilized as the reference standard in 38 out of 68 studies, while clinical diagnosis was the reference standard for the remainder.

The following findings were associated with a higher pre-test probability of GCA: limb claudication (positive likelihood ratio [LR+] 6.01; 95% confidence interval [CI] 1.38-26.16), jaw claudication (LR+ 4.90; 95% CI 3.74-6.41), temporal artery thickening (LR+ 4.70; 95% CI 2.65- 8.33), temporal artery loss of pulse (LR+ 3.25; 95% CI 2.49-4.23), platelet count of greater than 400 × 103/μL (LR+ 3.75; 95% CI 2.12-6.64), temporal tenderness (LR+ 3.14; 95% CI 1.14- 8.65), and erythrocyte sedimentation rate (ESR) >100 mm/h (LR+ 3.11; 95% CI 1.43-6.78). Findings associated with absence of GCA included an ESR <40 mm/h (negative likelihood ratio [LR-] 0.18; 95% CI 0.08-0.44), a C-reactive protein level <2.5mg/dL (LR- 0.38; 95% CI 0.25- 0.59), and age less than 70 years (LR- 0.48; 95% CI 0.27-0.86).

Caveats

The included studies had clinical heterogeneity with regard to patient populations and pre-test probabilities. Most studies were conducted in academic medical centers and none were conducted in the Emergency Department (ED), which may reduce generalizability to the ED or non-academic centers. Additionally, studies differed with regard to the gold standard, which could include TAB, imaging, or clinical diagnosis per the inclusion criteria. Those receiving TAB may have been subject to selection bias because a higher index of suspicion is generally needed to order that test due to the invasive nature of it. Moreover, TAB has imperfect sensitivity, so it is possible that some patients may have been misclassified if the TAB was a false negative (particularly if the portion of the temporal artery that was biopsied was not yet involved or patients were already on glucocorticoids for an extended period of time). Alternatively, those diagnosed clinically were at risk of verification bias because the index test contributed to the diagnosis. Some of the symptoms such as jaw claudication were not clearly defined in the included studies, which may artificially increase the LR+ of this feature. Finally, the systematic review included retrospective studies, which are subject to significant limitations. However, a sensitivity analysis assessing only prospective studies demonstrated comparable likelihood ratios.

In summary, the existing evidence indicates that several findings can significantly increase the pre-test probability of GCA. The strongest predictor was limb claudication, followed by jaw claudication, temporal artery thickness or loss of pulse, platelet count >400 × 103/μL, temporal tenderness, and ESR >100 mm/h. However, the absence of these findings does not sufficiently reduce the post-test probability to exclude further work up or testing.

Author

Brit Long, MD; Michael Gottlieb, MD, RDMS
Supervising Editor: Shahriar Zehtabchi, MD

Published/Updated

February 3, 2021

What are Likelihood Ratios?

LR, pretest probability and posttest (or posterior) probability are daunting terms that describe simple concepts that we all intuitively understand.

Let's start with pretest probability: that's just a fancy term for my initial impression, before we perform whatever test it is that we're going to use.

For example, a patient with prior stents comes in sweating and clutching his chest in agony, I have a pretty high suspicion that he's having an MI – let's say, 60%. That is my pretest probability.

He immediately gets an ECG (known here as the "test") showing an obvious STEMI.

Now, I know there are some STEMI mimics, so I'm not quite 100%, but based on my experience I'm 99.5% sure that he's having an MI right now. This is my posttest probability - the new impression I have that the patient has the disease after we did our test.

And likelihood ration? That's just the name for the statistical tool that converted the pretest probability to the posttest probability - it's just a mathematical description of the strength of that test.

Using an online calculator, that means the LR+ that got me from 60% to 99.5% is 145, which is about as high an LR you can get (and the actual LR for an emergency physician who thinks an ECG shows an obvious STEMI).

(Thank you to Seth Trueger, MD for this explanation!)

References:

Accuracy of Physical Examination and Imaging Findings for the Diagnosis of Elevated Intracranial Pressure

Diagnostics and Likelihood Ratios, Explained

Positive Findings (Patient Has This)

Finding Increased Disease Probability (Positive Likelihood Ratio)
Pupillary dilation 2.0
Compression or absence of the basal cisterns on CT 2.20
Midline shift >10 mm on CT 1.92

Negative Findings (Patient Doesn't Have This)

Finding Decreased Disease Probability (Negative Likelihood Ratio)
Pupillary dilation 0.89
Compression or absence of the basal cisterns on CT 0.23
Midline shift >10 mm on CT 0.89

Narrative

Elevated intracranial pressure (ICP) is a common but severe complication of several medical and traumatic conditions.1, 2 Prolonged increases in ICP are associated with poor patient outcomes in a variety of conditions, including traumatic brain injury (TBI), spontaneous subarachnoid or intracerebral hemorrhage, space-occupying lesion, meningitis, cerebral infarct, and cerebral edema from hepatic encephalopathy.3, 4, 5 Invasive ICP monitoring is the reference standard for cases suspected of increased ICP but is not universally available and has several potential complications, including intracranial infection and hemorrhage.5, 6 Consequently, clinicians must often rely on noninvasive methods for assessing increased ICP.

The systematic review and meta-analysis discussed here included all retrospective, prospective observational, and randomized controlled trials of adult patients (age ≥ 16 years) in the Emergency Department or Intensive Care Unit that assessed physical examination findings, brain computed tomography (CT), ocular nerve sheath diameter on ultrasound, or transcranial Doppler indices.7 The reference standard consisted of either an ICP reading > 20 mm Hg on invasive ICP monitoring or craniectomy with an operative diagnosis of elevated ICP. The primary outcome was the diagnostic accuracy of the aforementioned tests for diagnosing increased ICP.

The systematic review identified 40 studies (n = 5123 patients) which met the inclusion criteria. Twenty-four studies were prospective cohort studies, 15 were retrospective, and 1 was a randomized controlled trial. Twenty studies included patients with TBI, 3 subarachnoid hemorrhage, 2 intracerebral hemorrhage, 2 hepatic failure, 1 ischemic stroke, and 12 mixed populations of primary brain injury. Only 3 physical examination findings (pupillary dilatation, motor posturing, and altered mental status) had adequate relevant studies to allow for a metaanalysis. Table 1 demonstrates the findings associated with elevated ICP.

Table 1. Findings suggestive of elevated intracranial pressure (adapted from Fernando et al.).7
Image from Gyazo

Abbreviations: CI – confidence interval, LR – likelihood ratio, GCS – Glasgow Coma Scale, CT – computed tomography, mm – millimeters.

Measuring ocular nerve sheath diameter with ultrasound appeared to have a high discriminatory power with the pooled area under the receiver operating characteristic (AUROC) curve of 0.94 (95% CI 0.91 to 0.96). The sensitivity and specificity of the test were dependent on the cut-off used for increased optic nerve sheath in each study, precluding the possibility of meta-analysis to calculate pooled sensitivity and specificity. Authors calculated AUROC values for transcranial doppler pulsatility index (TCD-PI) to detect ICP > 20 mm Hg based on 3 studies, finding a pooled AUROC value of 0.85 (95% CI 0.78 to 0.91) with use of combined TCD arterial blood pressure methods.

Caveats

This study has several important limitations. Most of the included studies were relatively small, with only 13 studies enrolling more than 100 people. Additionally, the prevalence varied significantly between studies which may lead to spectrum bias, and evidence quality for findings suggestive of elevated ICP was predominantly low or moderate, with only midline shift > 10 mm on CT associated with high evidence quality. Nearly one-third of the studies were retrospective in nature. There was also significant heterogeneity with regard to the etiology and severity of the associated injuries, with no definition of the search strategy pertaining to the specific etiology of elevated ICP. Clinical signs were evaluated independently, which is not typical of clinical practice, and it is unclear how the diagnostic accuracy would change when using combinations of findings. Moreover, the inclusion of studies with either invasive monitoring or intraoperative diagnosis may lead to misclassification of the target condition in the latter case. Advances in imaging quality may limit the external validity of older studies with regard to current imaging. Furthermore, while 10 studies (1035 patients) evaluated ocular nerve sheath diameter, the variations in ocular nerve sheath diameter thresholds precluded the ability to perform meta-analyses on this group. Finally, there was limited data on transcranial Doppler with differences in both the parameters assessed between studies.

Based on the existing evidence, most findings are insufficiently sensitive or specific for the diagnosis of increased ICP. Therefore, physical examination findings, CT, ocular nerve sheath diameter, and transcranial Doppler do not appear to be reliable in identifying or excluding increased ICP.

The original manuscript was published in Academic Emergency Medicine as part of the partnership between TheNNT.com and AEM.

Author

Brit Long, MD; Alex Koyfman, MD; Michael Gottlieb, MD, RDMS
Supervising Editor: Shahriar Zehtabchi, MD

Published/Updated

February 14, 2020

What are Likelihood Ratios?

LR, pretest probability and posttest (or posterior) probability are daunting terms that describe simple concepts that we all intuitively understand.

Let's start with pretest probability: that's just a fancy term for my initial impression, before we perform whatever test it is that we're going to use.

For example, a patient with prior stents comes in sweating and clutching his chest in agony, I have a pretty high suspicion that he's having an MI – let's say, 60%. That is my pretest probability.

He immediately gets an ECG (known here as the "test") showing an obvious STEMI.

Now, I know there are some STEMI mimics, so I'm not quite 100%, but based on my experience I'm 99.5% sure that he's having an MI right now. This is my posttest probability - the new impression I have that the patient has the disease after we did our test.

And likelihood ration? That's just the name for the statistical tool that converted the pretest probability to the posttest probability - it's just a mathematical description of the strength of that test.

Using an online calculator, that means the LR+ that got me from 60% to 99.5% is 145, which is about as high an LR you can get (and the actual LR for an emergency physician who thinks an ECG shows an obvious STEMI).

(Thank you to Seth Trueger, MD for this explanation!)

References:

Lung Ultrasound for Diagnosis of Pneumonia in Children

Diagnostics and Likelihood Ratios, Explained

Positive Findings (Patient Has This)

Finding Increased Disease Probability (Positive Likelihood Ratio)
Pneumonia diagnosis 13.4
Sensitivity: 94%
Specificity: 94%

Negative Findings (Patient Doesn't Have This)

Finding Decreased Disease Probability (Negative Likelihood Ratio)
Pneumonia diagnosis 0.07

Narrative

The diagnosis of pneumonia remains a diagnostic challenge without an acceptable gold standard.1 Chest X-ray has poor test performance and only moderate inter-observer reliability,2 while computed tomography involves radiation exposure levels too high for routine use. A growing body of literature shows lung ultrasound (LUS) to be a potentially useful tool for the diagnosis of pneumonia in both adults3 and children,4 with the added benefit of no exposure to ionizing radiation.

Orso and colleagues performed a systematic review and meta-analysis on this topic.5 Their literature search identified 17 studies evaluating diagnostic accuracy of ultrasound for pneumonia in children. The meta-analysis was applied only to LUS for pneumonia and excluded studies of other diagnoses. All but two of the identified studies were prospective. Studies were evaluated for risk of bias by the QUADAS 2 tool.5 The systematic review reports the diagnostic accuracy results as sensitivity, specificity, and area under the receiver operating characteristic (AUC) curve with respective interquartile range (IQR) between 25% and 75% quartiles.

Overall LUS for pneumonia had a sensitivity of 94% (IQR, 89% to 97%) and specificity 94% (IQR, 86% to 98%). These correspond to a Likelihood ratio for a positive test (LR+) of 13.4 and a negative test (LR-) of 0.07. The area under the receiving operating curve was calculated to be 0.98 (IQR, 0.94 to 0.99), indicating a strong diagnostic accuracy.5

The studies had good agreement on sonographic definitions for pneumonia which included consolidation (hypoechoic subpleural area) and may have also included interstitial pattern, focal B-lines, and sonographic air bronchograms. One study did not report their definition. Those who reported scanning technique tended to follow imaging protocols similar to that described by Copetti, 2008.6

The authors identified retrospective study design of two included studies as the only high risk of bias. They assigned “uncertain risk” of bias to reference standard, convenience sampling, and timing of LUS.

Caveats

The most significant limitation is the lack of uniform reference standard. This is a challenge in pneumonia research as there is no accepted gold standard. The authors did an analysis evaluating the different reference standards, finding highest performance of LUS when compared to clinical diagnosis and adjudication. The concordance between Chest X-ray (CXR) and LUS was weaker, possibly because LUS likely outperforms CXR for identifying consolidation.6, 7 The comparison to CT is difficult to evaluate in this study given small number of subject who had CT. This metaanalysis found the use of CXR as reference standard to be the primary source of attributable heterogeneity amongst the reviewed studies.

The studies were variable in terms of setting and who was performing ultrasounds. In the majority, the sonographer was a pediatrician. Other scans were performed by emergency physicians, radiologists, and pediatric subspecialists. Similarly, the levels of ultrasound training and experience among sonographers were variable, ranging from 1 hour to 25 years. The most recent meta-analysis on this topic by Tsou et al. found significant difference between novice and experienced sonographers.8 While Tsou et al. included more studies due to their emphasis on sonographer training and not limiting to English language, they found similar pooled test characteristics for performance of LUS for pneumonia.

Other limitations include, all published studies used convenience sampling, many had relatively small sample sizes, and none were randomized controlled trials. These limitations contribute to a risk of bias that the systematic review by Orso et. al. may have underestimated. It is important for us to emphasize that imaging findings (i.e. consolidation) may be different than clinical pathology (i.e. pneumonia) and may have variable clinical relevance. Are small (‘sub-centimeter’) lesions clinically significant? Could they suggest viral or bacterial etiology, or be anatomical variants? Are they prognostic? Do they indicate need for antibiotics? These and other questions continue to challenge this field of research.

Globally, it seems likely that LUS performed by experienced sonographers is a diagnostic test that may help identify lung findings among children being evaluated for pneumonia. The authors of the systematic review discussed here5 astutely cautioned against wholeheartedly taking the overly favorable results. Notably, prominent research limitations (e.g. heterogeneity of studies) raise concerns of reliability, and potential biases (e.g. lack of reference standard) challenge the optimistic results.

The original manuscript was published in Academic Emergency Medicine as part of the partnership between TheNNT.com and AEM.

Author

Lilly A. Bellman, MD; Yiju T Liu, MD
Supervising Editor: Shahriar Zehtabchi, MD

Published/Updated

February 3, 2020

What are Likelihood Ratios?

LR, pretest probability and posttest (or posterior) probability are daunting terms that describe simple concepts that we all intuitively understand.

Let's start with pretest probability: that's just a fancy term for my initial impression, before we perform whatever test it is that we're going to use.

For example, a patient with prior stents comes in sweating and clutching his chest in agony, I have a pretty high suspicion that he's having an MI – let's say, 60%. That is my pretest probability.

He immediately gets an ECG (known here as the "test") showing an obvious STEMI.

Now, I know there are some STEMI mimics, so I'm not quite 100%, but based on my experience I'm 99.5% sure that he's having an MI right now. This is my posttest probability - the new impression I have that the patient has the disease after we did our test.

And likelihood ration? That's just the name for the statistical tool that converted the pretest probability to the posttest probability - it's just a mathematical description of the strength of that test.

Using an online calculator, that means the LR+ that got me from 60% to 99.5% is 145, which is about as high an LR you can get (and the actual LR for an emergency physician who thinks an ECG shows an obvious STEMI).

(Thank you to Seth Trueger, MD for this explanation!)

References:

Accuracy of Point-of-Care Ultrasound for Diagnosing Soft Tissue Abscess

Diagnostics and Likelihood Ratios, Explained

Positive Findings (Patient Has This)

N/A

Negative Findings (Patient Doesn't Have This)

N/A

Narrative

Over 3 million cases of skin and soft tissue infections (SSTI) including cellulitis and abscesses are managed in U.S. emergency departments (EDs) each year.1, 2, 3, 4 Overlap in presentations of cellulitis and abscess, which require different therapeutic approaches, has prompted increasing research into point-of-care ultrasound (POCUS) to help differentiate the two.1, 5, 6

The systematic review summarized here included prospective cohort studies evaluating POCUS for diagnosis of abscess in ED patients.7 The authors of the systematic review included patients with clinical evidence of SSTI. Reference standards varied, typically including draining purulent discharge, computed tomography scan, or clinical follow-up. There were no restrictions with regard to POCUS machine, transducer, protocol, or clinician background. The primary outcome was diagnostic accuracy for abscess in the ED.

The authors identified 8 relevant studies (n = 747 patients), with 3 conducted in adult ED and 5 in pediatric ED. Calculation of the point estimates for the diagnostic accuracy of POCUS found a sensitivity of 95.5% (95% confidence interval [CI] 88.9-98.3) and specificity of 80.3% (95% CI 56.4-92.7).8

Caveats

There are important limitations to the validity of these data. First, patients with cellulitis but initially negative POCUS for abscess may develop abscess later, confounding the reported results. Second, the included studies incorporated various gold standards for abscess diagnosis due to absence of a definitive criterion. Perhaps more importantly, this review included few studies, all with convenience samples, routine contamination between clinicians and sonographers (for both diagnosis and management decisions), and shifting reference standards.

These methodologic challenges tend to inflate sensitivity and specificity estimates, a concern highlighted by findings from both the largest study in the systematic review and a larger study published after the review.9, 10 The largest study included in the analysis, comprising 25% of the review’s sample size, found that POCUS did not add to the diagnostic post-test probability (and may have lowered both sensitivity and specificity) when clinicians felt confident of the diagnosis before ultrasound (i.e. when pre-test probability was high). However, when the pre-test probability was low or moderate, ultrasound was found to be helpful in increasing the post-test probability.9 Similarly, a large recent study reported that when clinicians felt certain (>90% of cases) of the diagnosis, ultrasound was unhelpful, while in most uncertain cases it improved accuracy.10

Based on this evidence, the accuracy numbers reported in the systematic review do not appear reliably valid for typical or common POCUS use in SSTI. We believe that the diagnostic accuracy of POCUS is dependent on the pre-test probability of abscess. POCUS does appear, however, to be potentially helpful in identifying abscess in ED patients in cases of diagnostic uncertainty. Therefore, we have assigned a color recommendation of Yellow (Unclear if benefits), though we recognize that POCUS is helpful in cases with clinical uncertainty after clinical examination.

The original manuscript was published in Academic Emergency Medicine as part of the partnership between TheNNT.com and AEM.

Author

Brit Long, MD; Alex Koyfman, MD; Michael Gottlieb, MD, RDMS
Supervising Editor: Shahriar Zehtabchi, MD

Published/Updated

November 20, 2019

What are Likelihood Ratios?

LR, pretest probability and posttest (or posterior) probability are daunting terms that describe simple concepts that we all intuitively understand.

Let's start with pretest probability: that's just a fancy term for my initial impression, before we perform whatever test it is that we're going to use.

For example, a patient with prior stents comes in sweating and clutching his chest in agony, I have a pretty high suspicion that he's having an MI – let's say, 60%. That is my pretest probability.

He immediately gets an ECG (known here as the "test") showing an obvious STEMI.

Now, I know there are some STEMI mimics, so I'm not quite 100%, but based on my experience I'm 99.5% sure that he's having an MI right now. This is my posttest probability - the new impression I have that the patient has the disease after we did our test.

And likelihood ration? That's just the name for the statistical tool that converted the pretest probability to the posttest probability - it's just a mathematical description of the strength of that test.

Using an online calculator, that means the LR+ that got me from 60% to 99.5% is 145, which is about as high an LR you can get (and the actual LR for an emergency physician who thinks an ECG shows an obvious STEMI).

(Thank you to Seth Trueger, MD for this explanation!)

References:

Markers of Fluid Responsiveness in Hemodynamically Unstable Patients

Diagnostics and Likelihood Ratios, Explained

Positive Findings (Patient Has This)

Finding Increased Disease Probability (Positive Likelihood Ratio)
Increase in cardiac output or related parameters following passive leg raising 11.0 [95% CI, 7.6 - 17]
Sensitivity: 88%
Specificity: 92%
Respiratory variation in vena cava diameter measured by ultrasound in mechanically ventilated patients (dispensability index >15%)* 5.3 [95% CI, 3.5 - 8.1]
Sensitivity: 85%
Specificity: 84%
Pulse pressure variation in mechanically ventilated patients, Vt <7.0mL/kg: 7.9 [95% CI, 4.1 - 16]
Sensitivity: 72%
Specificity: 91%

Negative Findings (Patient Doesn't Have This)

Finding Decreased Disease Probability (Negative Likelihood Ratio)
No increase in cardiac output with passive leg raising 0.13 [95% CI, 0.07-0.22]
Sensitivity: 88%
Specificity: 92%
Respiratory variation in vena cava diameter measured by ultrasound in mechanically ventilated patients (dispensability index <15%)* 0.27 [95% CI, 0.08 - 0.87]
Sensitivity: 77%
Specificity: 85%
*Subgroup of intubated patients without spontaneous respiratory efforts.

Narrative

Hemodynamically unstable patients are often given intravenous (IV) crystalloids and colloids to increase cardiac output and improve tissue perfusion.1 If successful, this is called fluid responsiveness. However, IV fluid therapy may cause pulmonary or peripheral edema, abdominal compartment syndrome, or impair oxygen diffusion.2 The clinician must decide, based on static and dynamic tests, whether vasopressors or inotropes should be used instead.

Static tests measure central venous pressure, often a good approximation of right atrial pressure, and pulmonary capillary wedge pressure.3 In contrast, dynamic measurements analyze changes in cardiac output or related parameters in response to bedside maneuvers that transiently change preload. The purpose of the systematic review summarized here is to provide summary estimates of the accuracy of the various symptoms, signs, and measurements used to predict fluid responsiveness in patients with refractory hypotension, signs of organ hypoperfusion, or both.

The systematic review discussed here, included 50 studies (n=2260) from intensive care unit settings, using a variety of invasive and noninvasive reference standards.4 Fluid responsiveness was defined as an increase in cardiac output of at least 15% in 39 of 50 trials and at least 10% in the other 11 trials. Mean prevalence of fluid responsiveness was 50% (95% CI, 42% - 56%).4 In all studies indices were measured before assessment of fluid responsiveness.

In pooled estimates, an increase in cardiac output with passive leg raising had the strongest likelihood ratios for an indicator of fluid responsiveness, as did the presence of respiratory variation in vena cava diameter measured with ultrasound and pulse pressure variation in the subset of mechanically ventilated patients. Lack of change in cardiac output with passive leg raise was the strongest indicator of the absence of fluid responsiveness.

Adverse Effects
No direct adverse effects were reported for the tests summarized here.

Caveats

This meta-analysis has several limitations, mainly that generalizability is uncertain. Individuals with arrhythmias, cardiogenic pulmonary edema, oxygen deficits, valvular disease, and ventricular failure were excluded.4 Some were also excluded if it was deemed dangerous or unethical to withhold fluids, and in some cases patients had already been extensively fluid resuscitated and on vasopressors prior to enrollment. Cardiac output measurements via TTE in these studies were performed by individuals with specific echocardiography training, hence applicability to physicians without equivalent skill is low.

Definitions of fluid responsiveness varied (from 10 to 15%) and most studies did not use thermodilution via pulmonary artery catheter, the reference standard for cardiac output.5, 6 Some trials also used poorly validated methods to determine fluid responsiveness, and virtually all studies enrolled a small convenience sample.4 The review authors deemed the overall quality of evidence to be high. However, there is tremendous variation in reference standards, shifting definitions of fluid responsiveness, and small convenience samples of often dissimilar subjects. This brings us to question the generalizability of the studies and validity of pooling data.

A meta-analysis published in 2016 found similar numbers to this review for the passive leg raise test, noting that mode of ventilation, type of fluid, starting position for the test, and measurement technique did not affect diagnostic performance.7 Conversely, a more recent, more rigorous review of dynamic tests found fewer acceptable studies, and weaker estimates of accuracy, particularly for passive leg raise. The authors also found pooling of data to be inappropriate based on high heterogeneity, low study quality, and different reference standards.8 Finally, a recent review of respiratory variation in vena cava diameter also found this test’s utility to be limited, especially when used in spontaneously ventilated patients, and negative results could not be used to rule out fluid responsiveness.9

Notably, no patient-centered outcomes were examined or reported in this systematic review. The tests assessed were designed to predict fluid responsiveness, a characteristic never rigorously evaluated as a guide for improving resuscitation, despite widespread use. Ultimately, therefore, the tests are physician-centered, and while we have reason to hope they are patient-centered as well, they may or may not lead to improved patient outcomes even if accurate.

In conclusion, we find serious flaws in the evidence base. It is promising that results across these studies—the best data presently available—were generally consistent, but more rigorous, larger investigations are badly needed, both to determine the accuracy of these tests and, of greater importance, to determine the utility of fluid responsiveness as a resuscitation guide. In terms of findings, change in cardiac output following passive leg raising was the most accurate predictor of fluid responsiveness in critically ill subjects who had already undergone initial resuscitation. The accuracy of change in IVC diameter with respiration to predict fluid responsiveness was modest in mechanically ventilated patients, and no test summarized here could rule out fluid responsiveness with certainty.

Image from Gyazo


The original manuscript was published in Journal of Evidence-Based Healthcare as part of the partnership between TheNNT.com and the journal.

Author

Jia Jian Li; James Hassel, MD: Joel Gernsheimer, MD
Supervising Editors: Kabir Yadav, MD; Allan Wolfson, MD; Shahriar Zehtabchi, MD

Published/Updated

November 1, 2019

What are Likelihood Ratios?

LR, pretest probability and posttest (or posterior) probability are daunting terms that describe simple concepts that we all intuitively understand.

Let's start with pretest probability: that's just a fancy term for my initial impression, before we perform whatever test it is that we're going to use.

For example, a patient with prior stents comes in sweating and clutching his chest in agony, I have a pretty high suspicion that he's having an MI – let's say, 60%. That is my pretest probability.

He immediately gets an ECG (known here as the "test") showing an obvious STEMI.

Now, I know there are some STEMI mimics, so I'm not quite 100%, but based on my experience I'm 99.5% sure that he's having an MI right now. This is my posttest probability - the new impression I have that the patient has the disease after we did our test.

And likelihood ration? That's just the name for the statistical tool that converted the pretest probability to the posttest probability - it's just a mathematical description of the strength of that test.

Using an online calculator, that means the LR+ that got me from 60% to 99.5% is 145, which is about as high an LR you can get (and the actual LR for an emergency physician who thinks an ECG shows an obvious STEMI).

(Thank you to Seth Trueger, MD for this explanation!)

References:

Use of the Clinical Examination in the Diagnosis of Cardiac Syncope

Diagnostics and Likelihood Ratios, Explained

Positive Findings (Patient Has This)

Finding Increased Disease Probability (Positive Likelihood Ratio)
Patient History
Age at first syncopal episode older than 35 years 3.3
History of atrial fibrillation or flutter 7.3
Known severe structural heart disease 3.3 to 4.8
Dyspnea prior syncope 3.5
Chest pain/angina prior to syncope 3.4 to 3.8
Cyanotic during syncope 6.2
Diagnostic Tests
High-sensitivity cardiac troponin T >42 pg/mL 5.1
High-sensitivity cardiac troponin I >31.3 pg/mL 5.4
NT-proBNP ≥210.5 pg/ml 47
NT-proBNP >1966 pg/ml 5.8
BNP >302 pg/mL 6.3
Multivariable Evaluation
Heart disease, abnormal ECG, or both 2.3
EGSYS Score 3 or more 2.8 to 3.3
Vasovagal Score less than -2 1.7 to 8.6

Negative Findings (Patient Doesn't Have This)

Finding Decreased Disease Probability (Negative Likelihood Ratio)
Patient History
Age at first syncopal episode 35 years or younger 0.13
Diagnostic Tests
Normal cardiac troponin T or I 0.15 to 0.39
Normal BNP level 0.16 to 0.21
Multivariable Evaluation
Absence of heart disease, abnormal ECG or both 0.20
EGSYS Score less than 3 0.12 to 0.17
Vasovagal score -2 or more 0.10 to 0.84

Source

Albassam OT, Redelmeier RJ, Shadowitz S, Husain AM, Simel D, Etchells EE. Did This Patient Have Cardiac Syncope? JAMA. 2019. doi:10.1001/jama.2019.8001

Study Population: 4317 patients from 11 studies who presented to either the emergency department or primary care, or were referred to specialty clinics for evaluation

Narrative

Syncope or transient loss of consciousness is a common problem seen in the emergency department (ED), accounting for 1% to 1.5% of ED visits annually.1 Cardiac syncope caused by cerebral hypoperfusion secondary to cardiopulmonary events such as arrhythmia or structural heart disease, accounts for 5% to 21% of syncope events.2 Cardiac syncope is associated with an increased risk of premature death and cardiac events.3, 4 It is therefore important emergency providers differentiate cardiac syncope from other causes.

This systematic review by Albassam et al, evaluated patient characteristics, physical exam findings, and diagnostic tests to identify cardiac causes of syncope.5 The authors searched the MEDLINE, Embase, CINAHL and Cochrane databases, selecting 11 studies that met inclusion and exclusion criteria. Each study included at least 10 subjects aged 12 years or older, for a total of 4317 patients. Studies were assigned levels of evidence developed for the Rational Clinical Exam series.6

Several historical factors were associated with an increased likelihood of cardiac syncope including: age at first syncopal spell 35 years or older, history of atrial fibrillation or flutter, known severe structural heart disease, dyspnea or chest pain prior to syncope, and witnessed cyanosis during syncope. An elevated cardiac troponin T or I, and an elevated B-type natriuretic peptide (BNP) both modestly increased the probability of cardiac syncope.

Factors that decreased the probability included age less than 35 years at first spell, normal cardiac troponin T or I, and normal BNP. Combinations of findings such as Evaluation of Guidelines in Syncope Study (EGSYS) score ≥3, Vasovagal score <-2, and abnormal electrocardiogram, heart disease, or both were more useful when absent than when present.

Caveats

The results of this review should be interpreted cautiously. Included studies generally defined cardiac syncope based on cardiologist judgment. Five of 11 included specialty referral populations or inpatients, leading to the potential for spectrum bias. Applying these results to a general ED population might lead to additional testing or interventions in patients who have lower risk of cardiac syncope than the studied population. Misclassification may have further skewed the results as patients with unexplained syncope were excluded from a number of studies. Many of the clinical findings resulted from single studies. Studies often included a wide age range of patients despite the incidence of syncope, related ED visits, and serious outcomes increasing sharply after the age of 60.7

Cardiac biomarkers such as troponin or BNP testing appear to be an attractive diagnostic option however they did not rule in or out cardiac syncope. Moreover, these levels were likely used to diagnose cardiac syncope leading to incorporation bias. ACEP clinical policy, appreciating these limitations, suggests a risk-stratification approach focusing on patient history and physical examination to avoid unnecessary testing and hospital admissions.1

In summary, the accurate diagnosis of cardiac syncope is helpful in determining an appropriate plan of care. While no single variable can independently diagnose or exclude cardiac syncope, several clinical findings may be used cohesively to help guide healthcare providers.

The original manuscript was published in Academic Emergency Medicine as part of the partnership between TheNNT.com and AEM.

Author

Kathryn Wiesendanger, BSc; Daniel K. Nishijima, MD, MAS
Supervising Editor: Shahriar Zehtabchi, MD

Published/Updated

October 16, 2019

What are Likelihood Ratios?

LR, pretest probability and posttest (or posterior) probability are daunting terms that describe simple concepts that we all intuitively understand.

Let's start with pretest probability: that's just a fancy term for my initial impression, before we perform whatever test it is that we're going to use.

For example, a patient with prior stents comes in sweating and clutching his chest in agony, I have a pretty high suspicion that he's having an MI – let's say, 60%. That is my pretest probability.

He immediately gets an ECG (known here as the "test") showing an obvious STEMI.

Now, I know there are some STEMI mimics, so I'm not quite 100%, but based on my experience I'm 99.5% sure that he's having an MI right now. This is my posttest probability - the new impression I have that the patient has the disease after we did our test.

And likelihood ration? That's just the name for the statistical tool that converted the pretest probability to the posttest probability - it's just a mathematical description of the strength of that test.

Using an online calculator, that means the LR+ that got me from 60% to 99.5% is 145, which is about as high an LR you can get (and the actual LR for an emergency physician who thinks an ECG shows an obvious STEMI).

(Thank you to Seth Trueger, MD for this explanation!)

References:

Factors Predicting Difficult Endotracheal Intubation

Diagnostics and Likelihood Ratios, Explained

Positive Findings (Patient Has This)

Finding Increased Disease Probability (Positive Likelihood Ratio)
Factors Affecting ETT Placement
History: History of difficult intubation 16-19
Signs: Upper lip bite test grade 3 14 (95% CI, 8.9-22)
Signs: Shorter hyomental distance 6.4 (95% CI, 4.1- 10)
Signs: Retrognathia 6 (95% CI, 3.1-11)
Signs: Combination of findings on Wilson score 9.1 (95% CI, 5.1-16)
Signs: Impaired neck mobility 4.2 (95% CI, 1.9-9.5)
Signs: Modified Mallampati score ≥3 4.1 (95% CI, 3.0-5.6)

Negative Findings (Patient Doesn't Have This)

Finding Decreased Disease Probability (Negative Likelihood Ratio)
Factors Affecting ETT Placement
History: Absence of a history of difficult intubation 0.72-0.82
Signs: Upper lip bite test grade 3 0.42
Signs: Shorter hyomental distance 0.84
Signs: Retrognathia 0.85
Signs: Combination of findings on Wilson score 0.60
Signs: Impaired neck mobility 0.77
Signs: Modified Mallampati score ≥3 0.52

Narrative

Endotracheal intubation is a common procedure in emergency medicine, and recognizing a potentially difficult intubation is imperative in planning for the procedure. While the “can’t intubate, can’t ventilate” scenario is rare, it is catastrophic if the airway operator is not prepared.1, 2, 3 Thus, predicting factors associated with difficult endotracheal intubation is important for emergency clinicians, with consideration of airway adjuncts such as video laryngoscopy, supraglottic airway devices, and cricothyrotomy.4 Some of the factors associated with intubation failure (or difficult intubation) include a history of prior difficult intubation, limited upper lip bite test (the patient bites the upper lip with his/her lower incisors), retrognathia, short thyromental and hyomental distance, decreased cervical spinal motion, higher modified Mallampati classification (defined by visibility of oropharyngeal structures with maximal mouth opening and tongue protrusion), and composite scores such as the Wilson Score (incorporating weight, mobility of the cervical spine and jaw, retrognathia, and incisor appearance).4, 5, 6, 7

The systematic review discussed here included studies evaluating risk factors (based on medical history or physical examination) or clinical tests that could predict difficult intubation (outcome) in adults (> 18 years) undergoing endotracheal intubation with direct laryngoscopy.8 Authors assessed the quality of the included trials using the Rational Clinical Examination (RCE) series quality checklist.9

The authors of the meta-analysis identified 62 relevant studies (n = 33,559 patients), which were all performed in the operating room (OR). The overall prevalence of difficult intubation was 10% (95% confidence interval [CI], 8.2% - 12%) which was most commonly defined by Cormack-Lehane grade 3 or 4.10 Cormack-Lehane grade 3 is defined as only the epiglottis visualized and grade 4 by neither glottis nor epiglottis seen on direct laryngoscopy.10 Other definitions included combination of Cormack-Lehane grade with additional requirements such as number of intubation attempts, time, use of bougie in 6 studies; percentage of glottic opening in 1 study; Intubation Difficulty Scale score > 5 in 3 studies; or minimum intubation time or number of attempts in 5 studies. History of prior difficult intubation was associated with an increased likelihood of difficult intubation (positive likelihood ratio [LR+]: 16 - 19).

Clinical examination findings including upper lip bite test class 3, defined as inability to bite any part of the upper lip with lower incisors, was a strong predictor of difficult intubation (LR+: 14; 95% CI, 8.9 - 22). Other findings, such as retrognathia (LR+: 6.0; 95% CI, 3.1 - 11), hyomental distance <3 - <5.5 cm (LR+: 6.4; 95% CI, 4.1 - 10), impaired neck mobility (LR+: 4.2; 95% CI, 1.9 - 9.5), impaired mouth opening (LR+: 3.6; 95% CI, 2.1 - 6.1), and the modified Mallampati score > 3 (LR+: 4.1; 95% CI, 3.0 - 5.6) also predicted difficult intubation. The Wilson score was also a strong predictor of difficult intubation (LR+: 9.1; 95% CI, 5.1 - 16). However, no clinical factor or composite score was useful in excluding difficult intubation. Sensitivity analyses did not change interpretation of results.8

Caveats

The trials included in the systematic review (rated as high-quality) identified certain findings are associated with an increased risk of difficult intubation. However, none of the findings were sufficient to exclude this. There was some variability in the reference standard used among studies to define a difficult airway, though the majority of studies incorporated the Cormack-Lehane classification system.10 In addition, studies that used the time of intubation or number of intubation attempts to define a difficult airway might have been influenced by the individual clinician’s ability or experience in intubation. Several predictors such as impaired cervical motion and retrognathia are subjective and vulnerable to inter-observer variability.

Authors of the systematic review limited their analysis to studies with independent assessments of predictors and outcomes in order to reduce bias. This led to exclusion of studies conducted in emergency settings. Therefore, all studies included in the systematic review were performed in the OR setting, limiting the applicability to the emergency department (ED) setting. Endotracheal intubation in the OR setting is more commonly associated with a nonemergent need for endotracheal intubation. While ED patients may differ with regard to mental and hemodynamic status, presence of gastric contents or vomiting, and ability to cooperate well with the assessments, knowledge of factors associated with difficult intubation and adequate preparation are still essential. Finally, this analysis evaluated only direct laryngoscopy. Therefore, the results of this review may not reflect current airway technology incorporating video laryngoscopy, extraglottic airway devices, and other advanced techniques.

In summary, the existing evidence indicates that several findings predict a difficult endotracheal intubation, but their absence cannot reliably exclude this scenario. The most accurate assessment was the upper lip bite test, followed by shorter hyomental distance, retrognathia, impaired neck mobility, modified Mallampati score > 3, and the Wilson score. Future studies should incorporate new airway technology such as video laryngoscopy and include emergency situations.

The original manuscript was published in Academic Emergency Medicine as part of the partnership between TheNNT.com and AEM.

Author

Brit Long, MD; Alex Koyfman, MD; Michael Gottlieb, MD, RDMS
Supervising Editor: Shahriar Zehtabchi, MD

Published/Updated

July 1, 2019

What are Likelihood Ratios?

LR, pretest probability and posttest (or posterior) probability are daunting terms that describe simple concepts that we all intuitively understand.

Let's start with pretest probability: that's just a fancy term for my initial impression, before we perform whatever test it is that we're going to use.

For example, a patient with prior stents comes in sweating and clutching his chest in agony, I have a pretty high suspicion that he's having an MI – let's say, 60%. That is my pretest probability.

He immediately gets an ECG (known here as the "test") showing an obvious STEMI.

Now, I know there are some STEMI mimics, so I'm not quite 100%, but based on my experience I'm 99.5% sure that he's having an MI right now. This is my posttest probability - the new impression I have that the patient has the disease after we did our test.

And likelihood ration? That's just the name for the statistical tool that converted the pretest probability to the posttest probability - it's just a mathematical description of the strength of that test.

Using an online calculator, that means the LR+ that got me from 60% to 99.5% is 145, which is about as high an LR you can get (and the actual LR for an emergency physician who thinks an ECG shows an obvious STEMI).

(Thank you to Seth Trueger, MD for this explanation!)

References:

Diagnostic Accuracy of Ultrasound for Confirmation of Endotracheal Tube Placement

Diagnostics and Likelihood Ratios, Explained

Positive Findings (Patient Has This)

Finding Increased Disease Probability (Positive Likelihood Ratio)
POCUS for ETT Placement
Confirmation of ETT Placement 34.4 (95% CI 12.7-93.1)

Negative Findings (Patient Doesn't Have This)

Finding Decreased Disease Probability (Negative Likelihood Ratio)
POCUS for ETT Placement
Confirmation of ETT Placement 0.01 (95% CI 0.01-0.02)

Narrative

Endotracheal intubation is a common intervention in the emergency department (ED) and prehospital setting. Direct visualization of endotracheal tube (ETT) placement through the vocal cords is limited at times, and esophageal intubation can be dangerous if not recognized.1 Therefore, additional methods (e.g., lung auscultation, esophageal detector devices, capnography) are necessary for confirmation of tube placement. However, these methods are not always reliable.2, 3, 4 Point-of-care ultrasonography (POCUS) has increasingly been used as a potential confirmatory tool for ETT confirmation. The 2015 Advanced Cardiac Life Support guidelines state that POCUS may be a useful adjunct for ETT confirmation.5

The meta-analysis discussed here included prospective observational or randomized controlled trials evaluating transtracheal POCUS for ETT placement confirmation in patients older than 18 years.6 All studies included a confirmatory test for comparison (e.g. end-tidal capnography, colorimetric capnography, direct visualization). The primary outcome was diagnostic accuracy of transtracheal POCUS for ETT confirmation, with subgroup analyses including location, provider specialty, provider experience, transducer type, and POCUS technique.

The authors also assessed time to confirmation as a secondary outcome. The authors of the meta-analysis identified 17 studies (n = 1,595 patients) that met their inclusion criteria. Twelve studies were performed in the ED, and five studies were conducted in the operating room. Overall, POCUS was 98.7% sensitive (95% confidence interval [CI] = 97.8% to 99.2%) and 97.1% specific (95% CI = 92.4% to 99.0%), with a positive likelihood ratio (LR+) of 34.4 (95% CI = 12.7 to 93.1) and a negative likelihood ratio (LR–) of 0.01 (95% CI = 0.01 to 0.02). Area under the receiver operating characteristic curve demonstrated a high degree of accuracy (area under the curve = 0.994; 95% CI = 0.982 to 0.998). The mean time to confirmation was 13.0 seconds (95% CI = 12.0 to 14.0 seconds).6 Subgroup analyses demonstrated no statistically significant difference with respect to enrollment location, provider, training, transducer type, or technique.

Caveats

Although POCUS is a valuable tool for confirmation of ETT placement, it is dependent on the individual provider’s ability to obtain and interpret appropriate images. Thus, it is important that providers receive adequate training before this technique is utilized routinely. Included studies demonstrated significant variation in POCUS training protocols. However, there was no significant difference in accuracy with respect to the training protocol. Of note, a prior study has demonstrated that the learning curve for transtracheal POCUS is relatively rapid.7 POCUS operator experience, specialty, and level of training varied in the included trials, but again, no significant difference was identified on subgroup analysis. Eight studies utilized dynamic technique (using POCUS to guide the intubation), nine used static technique (verifying tube placement after intubation), and one used both. However, there was no statistically significant difference in the diagnostic accuracy between techniques. This was supported by another recent study directly comparing the two techniques, which also demonstrated similar accuracy.8 In addition, there is also significant variability in the POCUS visual findings for guiding or verifying intubation. These findings include the number of air–mucosa interfaces, “snowstorm” flutter in the trachea, and a change in shape of the cricothyroid area, known as the “bullet” sign. Most of the included studies utilized the presence of a single air–mucosa interface to confirm trachea placement and two air–mucosa interfaces to identify esophageal intubation, known as the “double tract” sign. Three studies used the “snowstorm” sign to confirm trachea intubation, which refers to the brief flutter within the trachea as the tube is passed. Further studies are needed to determine the accuracy of these individual findings.

Transducer type varied between studies, but there was no statistically significant difference in accuracy present on subgroup analysis. A recent study comparing transducer types did not demonstrate a difference in accuracy but noted that providers preferred the linear transducer.9 Of note, most studies did not describe the size of the ETT used, although recent literature has suggested that accuracy remains consistent regardless of ETT size.10

The overall statistical heterogeneity was low. The overall risk of bias was also low, but the risk of bias for index testing, patient selection, and flow and timing was unclear, a limitation most likely due to inclusion of observational studies in the meta-analysis. This issue together with presence of publication bias warrants caution in interpreting the findings.

Further studies are needed to directly evaluate whether static versus dynamic technique, linear versus curvilinear transducer, and the selected POCUS finding impact diagnostic accuracy of POCUS for ETT confirmation.

Based on the existing evidence, POCUS appears to be highly sensitive and specific for guiding and verifying ETT placement. POCUS is easily available, rapid, noninvasive, and does not depend on ventilation for confirmation. Therefore, we have assigned a color recommendation of green (benefit > harm) to this technique.

The original manuscript was published in Academic Emergency Medicine as part of the partnership between TheNNT.com and AEM.

Author

Brit Long, MD; Alex Koyfman, MD; Michael Gottlieb, MD, RDMS
Supervising Editor: Shahriar Zehtabchi, MD

Published/Updated

June 14, 2019

What are Likelihood Ratios?

LR, pretest probability and posttest (or posterior) probability are daunting terms that describe simple concepts that we all intuitively understand.

Let's start with pretest probability: that's just a fancy term for my initial impression, before we perform whatever test it is that we're going to use.

For example, a patient with prior stents comes in sweating and clutching his chest in agony, I have a pretty high suspicion that he's having an MI – let's say, 60%. That is my pretest probability.

He immediately gets an ECG (known here as the "test") showing an obvious STEMI.

Now, I know there are some STEMI mimics, so I'm not quite 100%, but based on my experience I'm 99.5% sure that he's having an MI right now. This is my posttest probability - the new impression I have that the patient has the disease after we did our test.

And likelihood ration? That's just the name for the statistical tool that converted the pretest probability to the posttest probability - it's just a mathematical description of the strength of that test.

Using an online calculator, that means the LR+ that got me from 60% to 99.5% is 145, which is about as high an LR you can get (and the actual LR for an emergency physician who thinks an ECG shows an obvious STEMI).

(Thank you to Seth Trueger, MD for this explanation!)

References:

Retinal Pathology in Patients with Acute Onset Flashes and Floaters

Diagnostics and Likelihood Ratios, Explained

Positive Findings (Patient Has This)

Finding Increased Disease Probability (Positive Likelihood Ratio)
POCUS for Retinal Pathology
Symptoms: Visual reduction 5 (95% CI 3.1-8.1)
Symptoms: >10 new floaters 8-36
Signs: Vitreous hemorrhage 10. (95% CI 5.1-20)
Signs: Vitreous pigment 44. (95% CI 2.3-852)

Negative Findings (Patient Doesn't Have This)

Finding Decreased Disease Probability (Negative Likelihood Ratio)
POCUS for Retinal Pathology
Signs: Vitreous pigment 0.23

Source

Hollands H, Johnson D, Brox AC, Almeida D, Simel DL, Sharma S. Acute-onset floaters and flashes: is this patient at risk for retinal detachment? JAMA. 2009;302:2243-9.

Study Population: 2496 patients referred to ophthalmologists for evaluation of acute onset floaters and flashes of suspected ophthalmologic origin and/or diagnosed posterior vitreous detachment

Narrative

Posterior vitreous detachment (PVD) or separation of the posterior vitreous from the retina, is usually due to degeneration and age. PVD may present with floaters and/or flashes or may be asymptomatic for years.1 The incidence of PVD increases with age, from 25% in those 50-59 years to 87% in those 80-89 years.2 This condition sets the stage for further ophthalmologic pathologies, including a retinal tear.3, 4 Thus, previously diagnosed PVD is an important component of the history in a patient with an ophthalmologic complaint. A feared complication of PVD and retinal tear is retinal detachment. Retinal detachment occurs in 33-46% of patients with retinal tear and affects 0.8-1.8 per 10,000 persons per year,3, 4with a population prevalence of 0.3%.5, 6 Rapid diagnosis and treatment of retinal detachment can reduce visual loss and may even restore vision.7

This review included studies evaluating elements of the history and physical examination in patients with acute floaters and flashes of suspected ophthalmologic origin and suspected or diagnosed PVD.8 The systematic review and meta-analysis assessed the symptoms or clinical findings that could predict retinal tear in patients with acute onset floaters and/or flashes and a diagnosis of PVD established by the ophthalmologist.8

The authors of the meta-analysis identified 12 relevant studies (n = 2496 patients). All 12 were performed in ophthalmology clinics, with patients referred from primary care, optometry, or general ophthalmology. All patients had floaters and/or flashes of suspected ophthalmologic origin and suspected or diagnosed PVD. The prevalence of retinal tear was 14%. Only one study evaluated subjective vision reduction, finding that this increased likelihood of retinal tear (LR+: 5.0, 95% CI 3.1-8.1).1 Normal visual acuity decreased the likelihood of retinal tear (LR+: 0.60, 95% CI 0.5-0.7). Twelve studies evaluated slit lamp examination findings, showing vitreous hemorrhage and vitreous pigment increased the likelihood of retinal tear (LR+: 10, 95% CI 5.1-20 and 44, 95% CI 2.3-852, respectively).

Caveats

While certain features on history and examination increase the likelihood of retinal tear in the presence of flashes and floaters and/or PVD, several findings depended upon a thorough fundoscopic and slit lamp examination. Thus, ED provider experience and equipment for fundoscopic and slit lamp examination play an important role in evaluation for retinal tear.9 Importantly, slit lamp examination does not evaluate the posterior chamber of the eye. In order to appropriately visualize the posterior chamber, fundoscopy, typically with full eye dilation, is needed. While point-of-care ocular ultrasound can be diagnostic of several ophthalmologic conditions, this modality was not evaluated in this meta-analysis.10 The prevalence of retinal tear in patients presenting to the ED with visual symptoms such as vision loss and flashes/floaters is likely lower than found in this meta-analysis, suggesting spectrum bias, and in fact, most patients enrolled in the trials were diagnosed with PVD first or were suspected to have PVD, making this a highly select group likely to be different from patients with these complaints in the emergency department (ED). Regardless, this information might help ED physicians suspect retinal tear early and refer the patients to an ophthalmologist as soon as possible, as retinal tear leads to retinal detachment in 33-46% of patients.3, 4

Lastly, many of the likelihood ratios generated by the meta-analysis had wide confidence intervals, with lower limits approaching 1.0, most likely resulting from the small sample size. Larger studies are needed to generate better precision.

Based on the existing data, new onset floaters and/or flashes, subjective vision loss, and vitreous pigment or hemorrhage on ophthalmologic examination are associated with increased risk of retinal tear among patients with suspected or known PVD. In most studies included in the meta-analysis, PVD was often established by the ophthalmologist during initial encounter, suggesting further data are needed to determine whether these findings remain consistent among patients in acute care settings such as ED. Nonetheless, in the absence of other relevant data, this review suggests that findings outlined here may be useful in the evaluation of acute floaters/flashes. Newer data also suggest point-of-care ultrasound may be a useful indicator of retinal detachment.10 Future studies should determine how ultrasound may be used in combination with the above findings to risk stratify patients with new vision changes in non-ophthalmology settings.

The original manuscript was published in Academic Emergency Medicine as part of the partnership between TheNNT.com and AEM.

Author

Brit Long, MD; Alex Koyfman, MD; Michael Gottlieb, MD, RDMS
Supervising Editor: Shahriar Zehtabchi, MD

Published/Updated

What are Likelihood Ratios?

LR, pretest probability and posttest (or posterior) probability are daunting terms that describe simple concepts that we all intuitively understand.

Let's start with pretest probability: that's just a fancy term for my initial impression, before we perform whatever test it is that we're going to use.

For example, a patient with prior stents comes in sweating and clutching his chest in agony, I have a pretty high suspicion that he's having an MI – let's say, 60%. That is my pretest probability.

He immediately gets an ECG (known here as the "test") showing an obvious STEMI.

Now, I know there are some STEMI mimics, so I'm not quite 100%, but based on my experience I'm 99.5% sure that he's having an MI right now. This is my posttest probability - the new impression I have that the patient has the disease after we did our test.

And likelihood ration? That's just the name for the statistical tool that converted the pretest probability to the posttest probability - it's just a mathematical description of the strength of that test.

Using an online calculator, that means the LR+ that got me from 60% to 99.5% is 145, which is about as high an LR you can get (and the actual LR for an emergency physician who thinks an ECG shows an obvious STEMI).

(Thank you to Seth Trueger, MD for this explanation!)

References:

Diagnostic Accuracy of Ultrasound for the Evaluation of Small Bowel Obstruction

Diagnostics and Likelihood Ratios, Explained

Positive Findings (Patient Has This)

Finding Increased Disease Probability (Positive Likelihood Ratio)
POCUS for Small Bowel Obstruction
Small bowel obstruction diagnosis 27 (95% CI 8-98)

Negative Findings (Patient Doesn't Have This)

Finding Decreased Disease Probability (Negative Likelihood Ratio)
POCUS for Small Bowel Obstruction
Small bowel obstruction diagnosis 0.08 (95% CI 0.06-0.11)

Narrative

Small bowel obstruction (SBO) comprises 2% of patients presenting to the emergency department (ED) with abdominal pain, with over 300,000 hospitalizations per year.1, 2 SBO occurs when bowel contents cannot pass through the small bowel, due to either mechanical or functional etiology. If not appropriately diagnosed, SBO can result in intestinal ischemia, necrosis, and perforation.3 Traditional means of diagnosis such as plain film radiography suffer from poor sensitivity and specificity.4 Other imaging modalities include computed tomography (CT) or magnetic resonance imaging (MRI). Ultrasound has demonstrated promise in evaluation for SBO, as it is rapid, repeatable, inexpensive, does not expose the patient to radiation, and can be performed at the bedside.5, 6

The meta-analysis discussed here included trials that evaluated the accuracy of ultrasound for diagnosing SBO.7 The gold standard confirmatory test was determined by individual study definition and included CT, enteroclysis, diagnosis at surgery or discharge, or clinical follow up. The primary outcome was diagnostic accuracy of ultrasound for SBO, with subgroup analysis based on specific clinical setting (ED vs non-ED setting). The authors of the meta-analysis also conducted a sensitivity analysis that categorized inconclusive ultrasound examinations as false negatives.

The meta-analysis included 11 prospective, observational studies (n = 1178 patients). These studies enrolled patients suspected of acute SBO, with most studies using a convenience sample. Five studies were conducted in the ED, and 6 were conducted in other settings. ED providers performed the ultrasound in 3 studies. The mean age of the patients included in the analysis was 50 years, and 74% of patients were male. Overall, ultrasound was 92% sensitive (95% confidence interval [CI], 89% to 95%) and 97% specific (95% CI, 88% to 99%), with a positive likelihood ratio (LR+) of 27 (95% CI, 8 to 98) and a negative likelihood ratio (LR-) of 0.08 (95% CI, 0.06 to 0.11).

Sixteen examinations were non-diagnostic. Assuming all of these examinations were false negatives (worst case scenario), a sensitivity analysis demonstrated a slightly lower sensitivity of 91% (95% CI 86% to 94%) and +LR of 25 (95% CI 7 to 84), though specificity and -LR remained similar.

Caveats

While ultrasound can be a valuable tool in diagnosis of SBO, it depends on the individual provider’s ability to obtain and interpret appropriate images. This meta-analysis included studies with differences in sonographer experience and diagnostic criteria, which limit the applicability of the data. However, previous literature has suggested good diagnostic accuracy for identifying SBO after a 10-minute training session and 5 practice ultrasound examinations.5 ED providers were the primary sonographers in only 3 studies, limiting the generalizability of the findings to ED. Most studies enrolled patients based on a convenience sample, which could be a source of selection bias. Several studies did not explicitly state whether providers were blinded to the CT results, which may have led to potential bias. Inclusion of studies from different clinical settings can also contribute to spectrum bias; however, test characteristics for ultrasound performed by ED providers did not differ significantly from other providers based on the authors’ subgroup analysis. A recent prospective observational study published in 2019 suggests that the diagnostic accuracy of ultrasound for SBO may be lower, with a sensitivity of 88% and specificity of 54% (95% CI 45% to 63%).8 Of note, this study utilized providers with variable ultrasound experience and was performed at centers which did not regularly use ultrasound for SBO.8 Therefore, further studies evaluating the diagnostic accuracy among ED providers are needed.

Included studies in the meta-analysis were at low to moderate risk of bias, and limited data were present concerning pediatric patients.7 Moderate heterogeneity was present regarding patient population and outcome assessment, and studies demonstrated mild heterogeneity for sensitivity and moderate heterogeneity for specificity.

Although ultrasound offers several advantages compared to CT including the ability to rapidly diagnose SBO and assess progression of SBO over time, CT may still be necessary to determine the underlying etiology, which can influence the decision for operative management.

The existing evidence indicates that ultrasound is sensitive and specific in diagnosis of SBO. While the exact test characteristics for diagnosing SBO in an ED setting performed by ED physicians are not clear, this modality is rapid, repeatable, inexpensive, easily available, noninvasive, and not associated with radiation. Therefore, we have assigned a color recommendation of Green (Benefit > Harm) to the use of ultrasound for evaluation of SBO.

Author

Brit Long, MD; Alex Koyfman, MD; Michael Gottlieb, MD, RDMS
Supervising Editor: Shahriar Zehtabchi, MD

Published/Updated

May 3, 2019

What are Likelihood Ratios?

LR, pretest probability and posttest (or posterior) probability are daunting terms that describe simple concepts that we all intuitively understand.

Let's start with pretest probability: that's just a fancy term for my initial impression, before we perform whatever test it is that we're going to use.

For example, a patient with prior stents comes in sweating and clutching his chest in agony, I have a pretty high suspicion that he's having an MI – let's say, 60%. That is my pretest probability.

He immediately gets an ECG (known here as the "test") showing an obvious STEMI.

Now, I know there are some STEMI mimics, so I'm not quite 100%, but based on my experience I'm 99.5% sure that he's having an MI right now. This is my posttest probability - the new impression I have that the patient has the disease after we did our test.

And likelihood ration? That's just the name for the statistical tool that converted the pretest probability to the posttest probability - it's just a mathematical description of the strength of that test.

Using an online calculator, that means the LR+ that got me from 60% to 99.5% is 145, which is about as high an LR you can get (and the actual LR for an emergency physician who thinks an ECG shows an obvious STEMI).

(Thank you to Seth Trueger, MD for this explanation!)

References:

Diagnostic Accuracy of Point-of-Care Ultrasound for Retinal Detachment

Diagnostics and Likelihood Ratios, Explained

Positive Findings (Patient Has This)

Finding Increased Disease Probability (Positive Likelihood Ratio)
POCUS for Retinal Detachment
Retinal detachment diagnosis 25 (95% CI 8-78)

Negative Findings (Patient Doesn't Have This)

Finding Decreased Disease Probability (Negative Likelihood Ratio)
POCUS for Retinal Detachment
Retinal detachment diagnosis 0.06 (95% CI 0.01-0.25)

Source

Gottlieb M, Holladay D, Peksa GD. Point-of-Care Ocular Ultrasound for the Diagnosis of Retinal Detachment: A Systematic Review and Meta-Analysis. Acad Emerg Med. 2019 Jan 13. doi: 10.1111/acem.13682

Study Population: 11 studies comprising 844 patients suspected of having retinal detachment, with 5 studies conducted in the emergency department

Narrative

Ocular complaints are common in acute care settings. Although most patients will ultimately be diagnosed with a benign condition, 10-26% of people presenting to medical providers with flashes and floaters, for instance, will be diagnosed with retinal detachment.1, 2 Rapid diagnosis and treatment of retinal detachment are crucial to prevent irreversible vision loss.3 Unfortunately, the patient’s history has been shown to be poorly predictive, and physical examination performed by a non-ophthalmologic acute-care provider has demonstrated poor positive and negative predictive value.4 The gold standard for diagnosis of retinal detachment is largely indirect fundoscopy by an ophthalmologist. Point of care ultrasound (POCUS) may provide a fast and accurate assist in diagnosis, particularly when specialist consultation for fundoscopy may be delayed or inaccessible.

The meta-analysis discussed here evaluated the test performance of POCUS for the evaluation of suspected retinal detachment.5 Prospective (randomized-controlled or not) studies of POCUS in patients suspected of retinal detachment were included with a confirmatory test (ophthalmologic examination, surgical findings, computed tomography, magnetic resonance imaging, or clinical follow up). The primary outcome was diagnostic accuracy of POCUS for retinal detachment.

The authors of the meta-analysis identified 11 observational studies (n = 844 patients) that met their inclusion criteria.5 Consistent with other population estimates,1, 2 21% of patients were ultimately diagnosed with retinal detachment. Overall, POCUS was 94% sensitive (95% confidence interval [CI] 78% to 99%) and 96% specific (95% CI 89% to 99%), with a positive likelihood ratio (LR+) of 25 (95% CI 8 to 78) and a negative likelihood ratio (LR-) of 0.06 (95% CI 0.01 to 0.25). The area under the receiver operating characteristic curve demonstrated high accuracy (0.99; 95% CI 0.97 to 0.99).

Caveats

The performance of POCUS is dependent on provider experience, and the included studies involved a range of clinician experience and training, possibly accounting for some heterogeneity. Of note, most studies had consistently high sensitivity and specificity regardless of operator specialty, with the exception of one study which included a medical student sonographer and demonstrated reduced sensitivity compared to the other included publications. Training protocols varied between studies, ranging from 30 minutes to 2 hours in length, so it is unclear what the ideal training protocol should entail. The authors of the meta-analysis were unable to evaluate accuracy based on clinician experience, as most studies included sonographers of varying experience levels.

Of 11 studies included, only 5 were conducted in Emergency Department (ED) patients. The accuracy of POCUS was high in these patients [93.9% (95% CI 78.7% to 98.5%) sensitive and 92.4% (95% CI 85.6% to 96.1%) specific], but was significantly lower in other populations [74.1% (95% CI 61.0% to 84.7%) sensitive and 85.3% (95% CI 75.3% to 92.4%) specific]. This may be due to differences in the patient populations (e.g., acuity) or location sonography. Importantly, test characteristics were comparable when truly point of care ultrasound was performed by emergency medicine physicians (6 studies) as when performed by radiologists (4 studies). Further studies are recommended to identify which subgroup characteristics are associated with decreased diagnostic accuracy in these various cohorts.

Studies were at low risk of bias overall, though reference standards were an issue. The multiple different reference standards used across studies would normally be a concerning source of potential bias and variation for pooled accuracy data, but the consistently high sensitivities and specificities across studies with different reference standards are reassuring. The confidence intervals for both pooled sensitivity and specificity were wide. This indicates that high quality studies with larger sample sizes are needed. Future research should assess the influence of sonographer experience, as well as the ideal training protocols and number of examinations to obtain proficiency in this modality. It would also be valuable to determine the effect of high-frequency linear transducers, color Doppler, and contrast-enhanced ultrasound on diagnostic accuracy, as color or focused Doppler ultrasound may be more accurate than B-mode.6

Based on the existing evidence, POCUS appears to be a rapid, accessible, noninvasive tool that is highly sensitive and specific for diagnosis of retinal detachment. In fact, the American College of Emergency Physicians lists ocular examination with POCUS as a core emergency ultrasound application.7 Therefore, we have assigned a color recommendation of Green (Benefit > Harm) for use of POCUS for diagnosis of retinal detachment. POCUS should be considered a powerful tool to expedite definitive care for patients when direct ophthalmologic consultation may be otherwise delayed or limited.

Author

Brit Long, MD; Michael Gottlieb, MD
Supervising Editors: Gary Green, MD; Joshua Quaas, MD

Published/Updated

April 15, 2019

What are Likelihood Ratios?

LR, pretest probability and posttest (or posterior) probability are daunting terms that describe simple concepts that we all intuitively understand.

Let's start with pretest probability: that's just a fancy term for my initial impression, before we perform whatever test it is that we're going to use.

For example, a patient with prior stents comes in sweating and clutching his chest in agony, I have a pretty high suspicion that he's having an MI – let's say, 60%. That is my pretest probability.

He immediately gets an ECG (known here as the "test") showing an obvious STEMI.

Now, I know there are some STEMI mimics, so I'm not quite 100%, but based on my experience I'm 99.5% sure that he's having an MI right now. This is my posttest probability - the new impression I have that the patient has the disease after we did our test.

And likelihood ration? That's just the name for the statistical tool that converted the pretest probability to the posttest probability - it's just a mathematical description of the strength of that test.

Using an online calculator, that means the LR+ that got me from 60% to 99.5% is 145, which is about as high an LR you can get (and the actual LR for an emergency physician who thinks an ECG shows an obvious STEMI).

(Thank you to Seth Trueger, MD for this explanation!)

References:

Point-of-Care Ultrasound for the Diagnosis of Thoracoabdominal Injuries After Blunt Trauma

Diagnostics and Likelihood Ratios, Explained

Positive Findings (Patient Has This)

Finding Increased Disease Probability (Positive Likelihood Ratio)
POCUS for Thoracoabdominal Injury
Significant thoracic injuries or abdominal free fluid 18.5 (95% CI 10.8-40.5)

Negative Findings (Patient Doesn't Have This)

Finding Decreased Disease Probability (Negative Likelihood Ratio)
POCUS for Thoracoabdominal Injury
Significant thoracic injuries or abdominal free fluid 0.27 (95% CI 0.19-0.37)

Narrative

Trauma is a major cause of morbidity and mortality, representing one of the top ten causes of both death and disability-adjusted life years by the World Health Organization.1, 2 Point-of-care ultrasound (POCUS) is commonly performed during or after the primary survey to identify whether significant thoracic injuries or abdominal free fluid are present, particularly when patients are unstable or cannot receive a computed tomogram (CT).3 However, it is important to determine the accuracy of this modality to ensure proper application in trauma patients.

The Cochrane Review discussed here4 included retrospective and prospective studies assessing the diagnostic accuracy of POCUS for thoracoabdominal injuries in patients with blunt trauma (defined as any non-penetrating force). The reference standard included CT scan, magnetic resonance imaging (MRI), laparotomy or laparoscopy, thoracotomy, or autopsy. The primary outcome was the diagnosis of any thoracoabdominal injury, which was defined as: free fluid in the thoracic or abdominal cavity, retroperitoneum, pericardium, or mediastinum; organ injury (e.g. splenic, other solid organ, hollow viscera, or other organ laceration); a vascular lesion (e.g. dissection of rupture of aorta or other vessels); and other injuries (e.g. pneumothorax). Subgroup analyses were performed for pediatric patients versus adult patients, as well as abdominal versus thoracic injury.4 Approximately, half of the trials were conducted in the United States, and half of the study subjects were enrolled in level 1 trauma centers.

The authors of the Cochrane review identified 2872 records, of which 34 studies (n = 8635 patients) met inclusion criteria.4 Overall, POCUS was 74% sensitive (95% confidence interval [CI], 65% to 81%) and 96% specific (95% CI, 94% to 98%) with a positive likelihood ratio (LR+) of 18.5 (95% CI, 10.8 to 40.5) and a negative likelihood ratio (LR-) of 0.27 (95% CI, 0.19 to 0.37). Among pediatric patients, POCUS was 63% sensitive (95% CI, 46% to 77%) and 91% specific (95% CI, 81% to 96%). In adults alone, POCUS was 78% sensitive (95% CI, 69% to 84%) and 97% specific (95% CI, 96% to 99%). POCUS was 68% sensitive (95% CI, 59% to 75%) and 95% specific (95% CI, 92% to 97%) for diagnosing abdominal injuries specifically. For thoracic injuries, POCUS was 96% sensitive (95% CI, 88% to 99%) and 99% specific (95% CI, 97% to 100%). Assuming an overall baseline pretest probability of thoracoabdominal injury of approximately 28% (based on the median prevalence of such injuries in all of the included studies), POCUS would hypothetically miss injuries in 7.3% of patients and falsely suggest the presence of injuries in 2.9% of the patients. Assuming a pretest probability of thoracoabdominal injury of approximately 31% in pediatric patients (based on the median prevalence of such injuries in the included studies), the miss rate (false negative) would increase to 11.8%, and the false positive rate would be 6.2%.4

Caveats

The overall quality of the trials included in this meta-analysis was unclear due to limited reporting of the selection of participants and choice of diagnostic testing used for the reference standard. The most important limitation of the original review was the inclusion of all organ injuries (rather than free fluid) in the outcome assessment. The American College of Emergency Physicians guidelines states that the primary indication of the FAST examination is to “identify pathologic collections of free fluid or air released from injured organs or structures.”3 Consequently, assessing for the presence of any organ injury is beyond the scope of the FAST examination. Additionally, it is unclear how many injuries were significant enough to require an intervention. In this case, while a solid organ injury may have been missed with the FAST exam, it may not have been clinically significant. Therefore, it may have been preferable to focus on the identification of intraperitoneal free fluid and clinically significant organ injuries.

Furthermore, a number of different gold standards were used, which can lead to differential verification bias. There was also significant heterogeneity with regard to both the patient and study characteristics. Training protocols also varied and were not explicitly defined in most studies. Additionally, provider experience and specialty, as well as the POCUS machine utilized differed significantly between studies. These factors may have further contributed to the heterogeneity between the trials. Moreover, most studies assessed abdominal injury and only four studies assessed thoracic injury. Finally, there was limited reporting of multiple methodological parameters, including how studies accounted for inconclusive results.4

While the trials included in the meta-analysis do not report the harms associated with using POCUS for identifying thoracoabdominal injuries, false positive findings might subject the patients to unnecessary diagnostic or therapeutic procedures. False negative findings might falsely reassure the providers and cause them to miss injuries. The level of training and the experience of the operators might reduce the risk of such harms. Future trials are needed to assess the risk of such harms and identify the methods by which they could be reduced.

Based on the existing evidence, POCUS appears to be highly sensitive and specific for identifying significant thoracic injury. Additionally, POCUS appears to be highly specific but insufficiently sensitive to exclude abdominal injury. This suggests that a positive POCUS is strongly suggestive of an abdominal injury and should prompt subsequent targeted intervention, particularly in unstable patients. However, a negative POCUS examination does not exclude significant injury and should be followed by advanced imaging (e.g. CT), especially in pediatric patients. Keeping the limitation of this diagnostic modality in mind, this non-invasive and rapid test could provide useful information to the clinician and guide the proper diagnosis and treatment. Therefore, we have assigned a color recommendation of Green (Benefit > Harm) to the use of POCUS for identifying thoracoabdominal injuries after blunt trauma.

The original manuscript was published in Academic Emergency Medicine as part of the partnership between TheNNT.com and AEM.

Author

Michael Gottlieb, MD; Alex Koyfman, MD; Brit Long, MD

Published/Updated

March 15, 2019

What are Likelihood Ratios?

LR, pretest probability and posttest (or posterior) probability are daunting terms that describe simple concepts that we all intuitively understand.

Let's start with pretest probability: that's just a fancy term for my initial impression, before we perform whatever test it is that we're going to use.

For example, a patient with prior stents comes in sweating and clutching his chest in agony, I have a pretty high suspicion that he's having an MI – let's say, 60%. That is my pretest probability.

He immediately gets an ECG (known here as the "test") showing an obvious STEMI.

Now, I know there are some STEMI mimics, so I'm not quite 100%, but based on my experience I'm 99.5% sure that he's having an MI right now. This is my posttest probability - the new impression I have that the patient has the disease after we did our test.

And likelihood ration? That's just the name for the statistical tool that converted the pretest probability to the posttest probability - it's just a mathematical description of the strength of that test.

Using an online calculator, that means the LR+ that got me from 60% to 99.5% is 145, which is about as high an LR you can get (and the actual LR for an emergency physician who thinks an ECG shows an obvious STEMI).

(Thank you to Seth Trueger, MD for this explanation!)

References:

Dyspnea Due to Acute Heart Failure Syndrome

Diagnostics and Likelihood Ratios, Explained

Positive Findings (Patient Has This)

Finding Increased Disease Probability (Positive Likelihood Ratio)
Chest radiography
Kerley B lines 6.5 (95% CI 2.6-16.2)
Interstitial edema 6.4 (95% CI 3.4-12.2)
Cephalization 5.6 (95% CI 2.9-10.4)
Alveolar edema 5.3 (95% CI 3.3-8.5)
Lung ultrasonography
Positive B-line scan 7.4 (95% CI 4.2-12.8)
Bedside echocardiography
Restrictive mitral pattern 8.3 (95% CI 4.0-16.9)

Negative Findings (Patient Doesn't Have This)

Finding Decreased Disease Probability (Negative Likelihood Ratio)
Chest radiography
Kerley B lines 0.88 (95% CI 0.69-1.13)
Interstitial edema 0.73 (95% CI 0.68-0.78)
Cephalization 0.53 (95% CI 0.39-0.72)
Alveolar edema 0.95 (95% CI 0.94-0.97)
Lung ultrasonography
Positive B-line scan 0.16 (95% CI 0.5-0.51)
Bedside echocardiography
Restrictive mitral pattern 0.21 (95% CI 0.12-0.36)

Narrative

Dyspnea is a common acute symptom in patients presenting to the emergency department and who are ultimately diagnosed with acute heart failure syndrome (AHFS).1 However, in patients with undifferentiated dyspnea, an accurate diagnosis of AHFS may be difficult with the standard initial evaluation that includes patient history, physical examination, electrocardiography (ECG), chest radiography, and natriuretic peptide testing. This systematic review and meta-analysis comprehensively evaluated the diagnostic accuracy of the clinical assessment and index tests that physicians may use to distinguish AHFS from other clinical conditions in patients presenting to the emergency department with dyspnea.2

This review included 57 mostly prospective or cross-sectional studies, 52 unique cohorts, and a total of 17,893 patients.2 There was no single historical variable, symptom, or physical examination finding that could significantly reduce the likelihood of AHFS. An S3 gallop marginally increased the likelihood of AHFS (positive likelihood ratio [LR+] = 4.0; 95% confidence interval [CI], 2.7 to 5.9) but was an insensitive finding. None of the abnormal ECG findings substantially increased or decreased the probability of AHFS. The presence of radiographic findings that represented edema moderately increased the likelihood of AHFS (LR+ = 4.8 to 6.5; Table 1), but a negative chest radiograph was unhelpful.

Image from Gyazo

Serum B-type natriuretic peptide (BNP) testing increased the probability of AHFS (LR+ > 5.0) but only at markedly elevated concentrations (more than 800 pg per mL [800 ng per L]; Table 2); both BNP and serum N-terminal proB-type natriuretic peptide (NT-proBNP) testing were useful at excluding AHFS at very low concentrations (less than 100 pg per mL [100 ng per L]). Bedside lung ultrasonography of three or more B-line artifacts in two bilateral lung zones had the greatest discriminatory value among index tests (LR+ = 7.4; 95% CI, 4.2 to 12.8; negative likelihood ratio [LR–] = 0.16; 95% CI, 0.05 to 0.51). Bedside echocardiography of visually estimated reduced ejection fraction was somewhat helpful (LR+ = 4.1; 95% CI, 2.4 to 7.2), and the finding of restrictive mitral pattern was highly predictive of AHFS (LR+ = 8.3; 95% CI, 4.0 to 16.9).

Image from Gyazo

Caveats

Patients presenting to the emergency department with dyspnea are different than patients presenting to the primary care clinic with more subacute, less severe dyspnea. Although this possible spectrum bias may potentially influence study validity, the review analyzed all emergency department patients who presented with dyspnea and did not discriminate between time courses of presenting illness.2 Therefore, index test operative characteristics should maintain applicability, but considering the likely lower AHFS prevalence (and lower pretest probability) in the clinic population, relatively more tests may be required to confirm AHFS in the primary care clinic setting.

The estimation of each index test’s diagnostic accuracy was limited by the lack of a true, objective criterion or “gold” standard against which the test was measured. The criterion standard for AHFS was typically a subjective assessment by two or more physicians with data from the clinical record. The review did not evaluate the diagnostic accuracy of historical elements, symptoms, or examination findings in combination, also known as clinical gestalt.2 Clinical gestalt likely outperforms these diagnostic elements in isolation and plays an important role in determining the pretest probability of AHFS.

The studies that were included in the review varied in quality.2 Differences in inclusion and exclusion criteria among the included studies put them at varying degrees of risk for spectrum bias, and the consideration of pooled results should factor in the moderate to high statistical heterogeneity among included studies. With regards to the natriuretic peptide analyses, this review did not consider age as a variable that may alter BNP and NT-proBNP values. Age-stratified NT-proBNP cut points have since been demonstrated to be moderately useful in the diagnosis of AHFS.3

The accuracy of lung ultrasonography will be provider dependent, but emergency physician sonographers in most included studies were workshop-trained only. Furthermore, trainees may rapidly achieve procedural competency,4 and lung ultrasonography has been shown to have fair predictive value for pulmonary edema from AHFS when performed by novices and experts.5 Lastly, the performance of lung ultrasonography is likely time-sensitive because B-line artifacts appear to decrease coincident with treatment for lung edema caused by AHFS6 or other etiology.7

Conclusion: The individual components of the clinical history and physical examination, ECG, and chest radiography are not useful independently for confirming or excluding the diagnosis of AHFS in patients presenting to the emergency department. BNP concentrations above 800 pg per mL are helpful for establishing the diagnosis. BNP and NT-proBNP results are also valuable in ruling out AHFS when concentrations are below the threshold of 100 pg per mL. Lung ultrasonography, although provider-dependent, may be learned quickly and appears to have the best combination of test characteristics to confirm or exclude the diagnosis of AHFS.

The original manuscript was published in Medicine by the Numbers, American Family Physician as part of the partnership between TheNNT.com and AFP.

Author

Ian S. deSouza, MD; Jennifer L. Martindale, MD, MSc

Published/Updated

March 5, 2019

What are Likelihood Ratios?

LR, pretest probability and posttest (or posterior) probability are daunting terms that describe simple concepts that we all intuitively understand.

Let's start with pretest probability: that's just a fancy term for my initial impression, before we perform whatever test it is that we're going to use.

For example, a patient with prior stents comes in sweating and clutching his chest in agony, I have a pretty high suspicion that he's having an MI – let's say, 60%. That is my pretest probability.

He immediately gets an ECG (known here as the "test") showing an obvious STEMI.

Now, I know there are some STEMI mimics, so I'm not quite 100%, but based on my experience I'm 99.5% sure that he's having an MI right now. This is my posttest probability - the new impression I have that the patient has the disease after we did our test.

And likelihood ration? That's just the name for the statistical tool that converted the pretest probability to the posttest probability - it's just a mathematical description of the strength of that test.

Using an online calculator, that means the LR+ that got me from 60% to 99.5% is 145, which is about as high an LR you can get (and the actual LR for an emergency physician who thinks an ECG shows an obvious STEMI).

(Thank you to Seth Trueger, MD for this explanation!)

References:

Rule Out Ectopic Pregnancy

Diagnostics and Likelihood Ratios, Explained

Positive Findings (Patient Has This)

Finding Decreased Disease Probability (Negative Likelihood Ratio)
Visualized IUP 0.08
Note: accuracy of ultrasound is operator-dependent. Reported LRs may not be reproducible by an inexperienced sonographer.

Negative Findings (Patient Doesn't Have This)

Source

Narrative

This was a meta-analysis including 10 studies (n=2057) examining the operating characteristics of bedside emergency provider (EP) performed ultrasound to rule out ectopic pregnancy. All studies used the visualization of an IUP (intrauterine pregnancy) on ultrasound as the rule out criteria. The ultrasound examinations performed by EPs included transabdominal, transvaginal or both. Reference standards included formal radiology US, gynecology US, over-read of EP performed US by radiology, or clinical record review. They did not report specificity and positive LR's because of significant heterogeneity in these results when data was pooled. This study supports the utility of EP performed ultrasound as a strong rule out test for ectopic pregnancy.

Caveats

Note: accuracy of ultrasound is operator-dependent. Reported LRs may not be reproducible by an inexperienced sonographer.

Published in collaboration with The POCUS Atlas

Author

Garrett Ghent, MD

Published/Updated

September 13, 2018

What are Likelihood Ratios?

LR, pretest probability and posttest (or posterior) probability are daunting terms that describe simple concepts that we all intuitively understand.

Let's start with pretest probability: that's just a fancy term for my initial impression, before we perform whatever test it is that we're going to use.

For example, a patient with prior stents comes in sweating and clutching his chest in agony, I have a pretty high suspicion that he's having an MI – let's say, 60%. That is my pretest probability.

He immediately gets an ECG (known here as the "test") showing an obvious STEMI.

Now, I know there are some STEMI mimics, so I'm not quite 100%, but based on my experience I'm 99.5% sure that he's having an MI right now. This is my posttest probability - the new impression I have that the patient has the disease after we did our test.

And likelihood ration? That's just the name for the statistical tool that converted the pretest probability to the posttest probability - it's just a mathematical description of the strength of that test.

Using an online calculator, that means the LR+ that got me from 60% to 99.5% is 145, which is about as high an LR you can get (and the actual LR for an emergency physician who thinks an ECG shows an obvious STEMI).

(Thank you to Seth Trueger, MD for this explanation!)

POCUS Atlas: Abdominal Aortic Aneurysm (AAA)

Diagnostics and Likelihood Ratios, Explained

Positive Findings (Patient Has This)

Finding Increased Disease Probability (Positive Likelihood Ratio)
Abdominal aorta >3 cm 10.8
Note: accuracy of ultrasound is operator-dependent. Reported LRs may not be reproducible by an inexperienced sonographer.

Negative Findings (Patient Doesn't Have This)

Finding Decreased Disease Probability (Negative Likelihood Ratio)
Abdominal aorta >3 cm 0.025
Note: accuracy of ultrasound is operator-dependent. Reported LRs may not be reproducible by an inexperienced sonographer.

Narrative

This was a systematic review evaluating the operating characteristics of emergency department (ED) performed ultrasonography for abdominal aortic aneurysm (AAA). A total of 7 studies (n = 655) were included in the analysis, all of which were prospective studies which enrolled adult patients with symptoms/signs suggestive of AAA. The reference standard was varied among studies including CT, MRI, aortography, radiology performed ultrasound, exploratory laparotomy, or autopsy results. AAA was defined as > 3 cm dilation of the aorta. Individual sensitivity and specificity for AAA detection among studies were as follows: sensitivity 97.5-100%, specificity 94.1-100%, LR+ 10.8 - infinite, and LR- 0.00-0.025.

Caveats

Note: accuracy of ultrasound is operator-dependent. Reported LRs may not be reproducible by an inexperienced sonographer.

Published in collaboration with The POCUS Atlas

Author

Matthew Riscinti, MD; John F Kilpatrick, MD

Published/Updated

What are Likelihood Ratios?

LR, pretest probability and posttest (or posterior) probability are daunting terms that describe simple concepts that we all intuitively understand.

Let's start with pretest probability: that's just a fancy term for my initial impression, before we perform whatever test it is that we're going to use.

For example, a patient with prior stents comes in sweating and clutching his chest in agony, I have a pretty high suspicion that he's having an MI – let's say, 60%. That is my pretest probability.

He immediately gets an ECG (known here as the "test") showing an obvious STEMI.

Now, I know there are some STEMI mimics, so I'm not quite 100%, but based on my experience I'm 99.5% sure that he's having an MI right now. This is my posttest probability - the new impression I have that the patient has the disease after we did our test.

And likelihood ration? That's just the name for the statistical tool that converted the pretest probability to the posttest probability - it's just a mathematical description of the strength of that test.

Using an online calculator, that means the LR+ that got me from 60% to 99.5% is 145, which is about as high an LR you can get (and the actual LR for an emergency physician who thinks an ECG shows an obvious STEMI).

(Thank you to Seth Trueger, MD for this explanation!)

POCUS Atlas: Small Bowel Obstruction

Diagnostics and Likelihood Ratios, Explained

Positive Findings (Patient Has This)

Finding Increased Disease Probability (Positive Likelihood Ratio)
Dilated small bowel (diameter ≥25 mm) with peristaltic activity 27.18
Note: accuracy of ultrasound is operator-dependent. Reported LRs may not be reproducible by an inexperienced sonographer.

Negative Findings (Patient Doesn't Have This)

Finding Decreased Disease Probability (Negative Likelihood Ratio)
Dilated small bowel (diameter ≥25 mm) with peristaltic activity 0.08
Note: accuracy of ultrasound is operator-dependent. Reported LRs may not be reproducible by an inexperienced sonographer.

Narrative

This was a meta-analysis including 11 studies (n=1178) to evaluate the test characteristics of US in diagnosis of small bowel obstruction (SBO). There was mild to moderate heterogeneity in diagnostic criteria, study location, sonographer experience, and reference standard. Specifically, most studies used 2.5 cm as the cutoff to diagnose SBO while one study used the cutoff of 3.0 cm, and several other studies only noted the presence of “dilated bowel loops” as a diagnostic criteria. Of the 11 studies included, only 3 were emergency department studies. Reference standards included surgery, clinical diagnosis, CT, or other advanced imaging. While there were multiple components to the index test and varied reference standard, this does appear to be the best and biggest review on this topic. These operating characteristics suggest ultrasound to be a valuable tool in the diagnosis of SBO, however further studies are needed specifically with regards to the emergency department setting. *Other diagnostic criteria included visualizing collapsed distal loops of bowel with decreased peristalsis

Caveats

Note: accuracy of ultrasound is operator-dependent. Reported LRs may not be reproducible by an inexperienced sonographer.

Published in collaboration with The POCUS Atlas

Author

Roshanak Benabbas, MD

Published/Updated

What are Likelihood Ratios?

LR, pretest probability and posttest (or posterior) probability are daunting terms that describe simple concepts that we all intuitively understand.

Let's start with pretest probability: that's just a fancy term for my initial impression, before we perform whatever test it is that we're going to use.

For example, a patient with prior stents comes in sweating and clutching his chest in agony, I have a pretty high suspicion that he's having an MI – let's say, 60%. That is my pretest probability.

He immediately gets an ECG (known here as the "test") showing an obvious STEMI.

Now, I know there are some STEMI mimics, so I'm not quite 100%, but based on my experience I'm 99.5% sure that he's having an MI right now. This is my posttest probability - the new impression I have that the patient has the disease after we did our test.

And likelihood ration? That's just the name for the statistical tool that converted the pretest probability to the posttest probability - it's just a mathematical description of the strength of that test.

Using an online calculator, that means the LR+ that got me from 60% to 99.5% is 145, which is about as high an LR you can get (and the actual LR for an emergency physician who thinks an ECG shows an obvious STEMI).

(Thank you to Seth Trueger, MD for this explanation!)

POCUS Atlas: Pediatric Appendicitis

Diagnostics and Likelihood Ratios, Explained

Positive Findings (Patient Has This)

Finding Increased Disease Probability (Positive Likelihood Ratio)
Non-compressible appendix >6 mm 9.24

Negative Findings (Patient Doesn't Have This)

Finding Decreased Disease Probability (Negative Likelihood Ratio)
Non-compressible appendix >6 mm 0.17
Note: while the authors did find a negative LR of 0.17, they concluded that a positive ED‐POCUS is diagnostic but a negative ED‐POCUS is not enough to rule out AA. As always, use clinical judgment in interpreting these results in patients.

Narrative

This systematic review and meta-analysis of 4 studies (n=461) evaluated the accuracy of emergency department POCUS performed by EM or PEM physicians for diagnosis of acute appendicitis in children. The main limitation of the study was the high prevalence of equivocal results. However, authors did a sensitivity analysis with and without equivocal cases and results were similar. A mathematical model was used to compare POCUS to CT scan and MRI. The authors concluded that a positive POCUS exam is diagnostic, obviating the need for CT, however if POCUS is equivocal or negative, further imaging with CT or MRI is necessary.

Caveats

Note: accuracy of ultrasound is operator-dependent. Reported LRs may not be reproducible by an inexperienced sonographer.

Published in collaboration with The POCUS Atlas

Author

Roshanak Benabbas, MD

Published/Updated

What are Likelihood Ratios?

LR, pretest probability and posttest (or posterior) probability are daunting terms that describe simple concepts that we all intuitively understand.

Let's start with pretest probability: that's just a fancy term for my initial impression, before we perform whatever test it is that we're going to use.

For example, a patient with prior stents comes in sweating and clutching his chest in agony, I have a pretty high suspicion that he's having an MI – let's say, 60%. That is my pretest probability.

He immediately gets an ECG (known here as the "test") showing an obvious STEMI.

Now, I know there are some STEMI mimics, so I'm not quite 100%, but based on my experience I'm 99.5% sure that he's having an MI right now. This is my posttest probability - the new impression I have that the patient has the disease after we did our test.

And likelihood ration? That's just the name for the statistical tool that converted the pretest probability to the posttest probability - it's just a mathematical description of the strength of that test.

Using an online calculator, that means the LR+ that got me from 60% to 99.5% is 145, which is about as high an LR you can get (and the actual LR for an emergency physician who thinks an ECG shows an obvious STEMI).

(Thank you to Seth Trueger, MD for this explanation!)

POCUS Atlas: Cholecystitis

Diagnostics and Likelihood Ratios, Explained

Positive Findings (Patient Has This)

Finding Increased Disease Probability (Positive Likelihood Ratio)
Cholelithiasis, wall thickening, pericholecystic fluid, sonographic Murphy's 15.6
Note: accuracy of ultrasound is operator-dependent. Reported LRs may not be reproducible by an inexperienced sonographer. LR+ is a range but simplified here for the purposes of the interactive tool. See below for details.

Negative Findings (Patient Doesn't Have This)

Finding Decreased Disease Probability (Negative Likelihood Ratio)
Cholelithiasis, wall thickening, pericholecystic fluid, sonographic Murphy's 0.21
Note: accuracy of ultrasound is operator-dependent. Reported LRs may not be reproducible by an inexperienced sonographer. LR- is a range but simplified here for the purposes of the interactive tool. See below for details.

Narrative

This was a systematic review including 4 prospective studies evaluating the operating characteristics of bedside ultrasound for acute cholecystitis (AC) in adult patients seen in the emergency department with a clinical suspicion for AC or right upper quadrant pain. Sample size of the studies varied from 30 to 193 subjects. Reference standard was surgical pathology. The experience of the sonographers varied between the studies and in one study no documentation of sonographer experience was noted. There was significant heterogeneity across the included studies precluding the authors ability to pool the results hence a range is noted in the operating characteristics table.

Caveats

Note: accuracy of ultrasound is operator-dependent. Reported LRs may not be reproducible by an inexperienced sonographer.

Published in collaboration with The POCUS Atlas

Author

Mike Macias, MD

Published/Updated

What are Likelihood Ratios?

LR, pretest probability and posttest (or posterior) probability are daunting terms that describe simple concepts that we all intuitively understand.

Let's start with pretest probability: that's just a fancy term for my initial impression, before we perform whatever test it is that we're going to use.

For example, a patient with prior stents comes in sweating and clutching his chest in agony, I have a pretty high suspicion that he's having an MI – let's say, 60%. That is my pretest probability.

He immediately gets an ECG (known here as the "test") showing an obvious STEMI.

Now, I know there are some STEMI mimics, so I'm not quite 100%, but based on my experience I'm 99.5% sure that he's having an MI right now. This is my posttest probability - the new impression I have that the patient has the disease after we did our test.

And likelihood ration? That's just the name for the statistical tool that converted the pretest probability to the posttest probability - it's just a mathematical description of the strength of that test.

Using an online calculator, that means the LR+ that got me from 60% to 99.5% is 145, which is about as high an LR you can get (and the actual LR for an emergency physician who thinks an ECG shows an obvious STEMI).

(Thank you to Seth Trueger, MD for this explanation!)

POCUS Atlas: Cholelithiasis

Diagnostics and Likelihood Ratios, Explained

Positive Findings (Patient Has This)

Finding Increased Disease Probability (Positive Likelihood Ratio)
Presence of gallstone 7.48
Note: accuracy of ultrasound is operator-dependent. Reported LRs may not be reproducible by an inexperienced sonographer.

Negative Findings (Patient Doesn't Have This)

Finding Decreased Disease Probability (Negative Likelihood Ratio)
Presence of gallstone 0.16
Note: accuracy of ultrasound is operator-dependent. Reported LRs may not be reproducible by an inexperienced sonographer.

Narrative

This was a meta-analysis including 8 studies (n=710) evaluating the operating characteristics of emergency ultrasound (EUS) for identifying cholelithiasis in adult patients presenting to the emergency department with symptoms suggestive of biliary colic (RUQ pain, epigastric pain, or right flank pain). Reference standards included radiology performed ultrasound, CT, MRI, and/or surgical pathology. There was quite a variation in the technical ability between operators which may distort the pooled sensitivity and specificity.

Caveats

Note: accuracy of ultrasound is operator-dependent. Reported LRs may not be reproducible by an inexperienced sonographer.

Published in collaboration with The POCUS Atlas

Author

Roshanak Benabbas, MD

Published/Updated

What are Likelihood Ratios?

LR, pretest probability and posttest (or posterior) probability are daunting terms that describe simple concepts that we all intuitively understand.

Let's start with pretest probability: that's just a fancy term for my initial impression, before we perform whatever test it is that we're going to use.

For example, a patient with prior stents comes in sweating and clutching his chest in agony, I have a pretty high suspicion that he's having an MI – let's say, 60%. That is my pretest probability.

He immediately gets an ECG (known here as the "test") showing an obvious STEMI.

Now, I know there are some STEMI mimics, so I'm not quite 100%, but based on my experience I'm 99.5% sure that he's having an MI right now. This is my posttest probability - the new impression I have that the patient has the disease after we did our test.

And likelihood ration? That's just the name for the statistical tool that converted the pretest probability to the posttest probability - it's just a mathematical description of the strength of that test.

Using an online calculator, that means the LR+ that got me from 60% to 99.5% is 145, which is about as high an LR you can get (and the actual LR for an emergency physician who thinks an ECG shows an obvious STEMI).

(Thank you to Seth Trueger, MD for this explanation!)

POCUS Atlas: DVT

Diagnostics and Likelihood Ratios, Explained

Positive Findings (Patient Has This)

Finding Increased Disease Probability (Positive Likelihood Ratio)
Positive proximal CUS, complete CUS, or color flow duplex US 30.03
Note: accuracy of ultrasound is operator-dependent. Reported LRs may not be reproducible by an inexperienced sonographer.

Negative Findings (Patient Doesn't Have This)

Finding Decreased Disease Probability (Negative Likelihood Ratio)
Negative proximal CUS, complete CUS, or color flow duplex US 0.04
Note: accuracy of ultrasound is operator-dependent. Reported LRs may not be reproducible by an inexperienced sonographer.

Narrative

This was a meta-analysis including 16 studies (n=2379) evaluating the operating characteristics of emergency physician (EP) performed ultrasound for the diagnosis of DVT. The providers used color-flow duplex ultrasound in two studies, proximal venous ultrasound in 13 studies (not looking at the calf), and whole-leg venous ultrasound in one study. Reference standard was radiology department ultrasound, vascular lab, or angiography. Considering only high quality studies which met QUADAS-2 Criteria (11 out of 16 initially selected studies), the sensitivity and specificity improved to 97.6% and 96.8% respectively.

Caveats

Note: accuracy of ultrasound is operator-dependent. Reported LRs may not be reproducible by an inexperienced sonographer.

Published in collaboration with The POCUS Atlas

Author

John F Kilpatrick, MD

Published/Updated

What are Likelihood Ratios?

LR, pretest probability and posttest (or posterior) probability are daunting terms that describe simple concepts that we all intuitively understand.

Let's start with pretest probability: that's just a fancy term for my initial impression, before we perform whatever test it is that we're going to use.

For example, a patient with prior stents comes in sweating and clutching his chest in agony, I have a pretty high suspicion that he's having an MI – let's say, 60%. That is my pretest probability.

He immediately gets an ECG (known here as the "test") showing an obvious STEMI.

Now, I know there are some STEMI mimics, so I'm not quite 100%, but based on my experience I'm 99.5% sure that he's having an MI right now. This is my posttest probability - the new impression I have that the patient has the disease after we did our test.

And likelihood ration? That's just the name for the statistical tool that converted the pretest probability to the posttest probability - it's just a mathematical description of the strength of that test.

Using an online calculator, that means the LR+ that got me from 60% to 99.5% is 145, which is about as high an LR you can get (and the actual LR for an emergency physician who thinks an ECG shows an obvious STEMI).

(Thank you to Seth Trueger, MD for this explanation!)

POCUS Atlas: Severe Systolic Dysfunction by EPSS

Diagnostics and Likelihood Ratios, Explained

Positive Findings (Patient Has This)

Finding Increased Disease Probability (Positive Likelihood Ratio)
End point septal separation (EPSS) >7 mm 2.07
Note: accuracy of ultrasound is operator-dependent. Reported LRs may not be reproducible by an inexperienced sonographer.

Negative Findings (Patient Doesn't Have This)

Finding Decreased Disease Probability (Negative Likelihood Ratio)
End point septal separation (EPSS) >7 mm 0.0
Note: accuracy of ultrasound is operator-dependent. Reported LRs may not be reproducible by an inexperienced sonographer.

Narrative

This was a prospective observational study (n=80) comparing emergency medicine (EM) physician performed mitral valve EPSS to formal TTE LVEF estimation. A convenience sample of unselected hospitalized patients undergoing comprehensive TTE for any indication was used. While EPSS > 7 mm was noted to be 100% sensitive for predicting severe systolic dysfunction (EF < 30%), a second cutoff of 8 mm was used for assessing any systolic dysfunction. The sensitivity and specificity of an EPSS > 8 mm for any systolic dysfunction were 83.3% (95% CI, 62.6-95.2) and 50.0% (95% CI, 29.2-70.9), respectively. The corresponding positive LR was 1.67, and the negative LR was 0.33.

Caveats

Note: accuracy of ultrasound is operator-dependent. Reported LRs may not be reproducible by an inexperienced sonographer.

Published in collaboration with The POCUS Atlas

Author

Matthew Riscinti, MD

Published/Updated

What are Likelihood Ratios?

LR, pretest probability and posttest (or posterior) probability are daunting terms that describe simple concepts that we all intuitively understand.

Let's start with pretest probability: that's just a fancy term for my initial impression, before we perform whatever test it is that we're going to use.

For example, a patient with prior stents comes in sweating and clutching his chest in agony, I have a pretty high suspicion that he's having an MI – let's say, 60%. That is my pretest probability.

He immediately gets an ECG (known here as the "test") showing an obvious STEMI.

Now, I know there are some STEMI mimics, so I'm not quite 100%, but based on my experience I'm 99.5% sure that he's having an MI right now. This is my posttest probability - the new impression I have that the patient has the disease after we did our test.

And likelihood ration? That's just the name for the statistical tool that converted the pretest probability to the posttest probability - it's just a mathematical description of the strength of that test.

Using an online calculator, that means the LR+ that got me from 60% to 99.5% is 145, which is about as high an LR you can get (and the actual LR for an emergency physician who thinks an ECG shows an obvious STEMI).

(Thank you to Seth Trueger, MD for this explanation!)

POCUS Atlas: Acute Heart Failure

Diagnostics and Likelihood Ratios, Explained

Positive Findings (Patient Has This)

Finding Increased Disease Probability (Positive Likelihood Ratio)
Visually "reduced" ejection fraction 4.1
≥3 B-lines in two bilateral lung zones 7.40
Note: accuracy of ultrasound is operator-dependent. Reported LRs may not be reproducible by an inexperienced sonographer.

Negative Findings (Patient Doesn't Have This)

Finding Decreased Disease Probability (Negative Likelihood Ratio)
Visually "reduced" ejection fraction 0.24
≥3 B-lines in two bilateral lung zones 0.16
Note: accuracy of ultrasound is operator-dependent. Reported LRs may not be reproducible by an inexperienced sonographer.

Narrative

This was a large systematic review and meta-analysis (57 studies, n = 17,893) of the operating characteristics for diagnostic elements available to the emergency physician for diagnosing acute heart failure (AHF) including the history and physical, ECG, chest radiography, BNP/NT-proBNP (NPs), bedside echocardiography, lung ultrasound, and bioimpedance. They concluded that bedside lung US and echocardiography appear to the most useful tests for affirming the presence of AHF while NPs are valuable in excluding the diagnosis. Reduced ejection fraction was determined to have the highest +LR compared to other elements of the exam. However, the studied included in the final pooling appear to have only used "visual estimation" of reduced EF. With regards to lung US, a positive finding was defined in every study by the presence of at least three B lines in two bilateral lung zones.

Caveats

Note: accuracy of ultrasound is operator-dependent. Reported LRs may not be reproducible by an inexperienced sonographer.

Published in collaboration with The POCUS Atlas

Author

Matthew Riscinti, MD

Published/Updated

What are Likelihood Ratios?

LR, pretest probability and posttest (or posterior) probability are daunting terms that describe simple concepts that we all intuitively understand.

Let's start with pretest probability: that's just a fancy term for my initial impression, before we perform whatever test it is that we're going to use.

For example, a patient with prior stents comes in sweating and clutching his chest in agony, I have a pretty high suspicion that he's having an MI – let's say, 60%. That is my pretest probability.

He immediately gets an ECG (known here as the "test") showing an obvious STEMI.

Now, I know there are some STEMI mimics, so I'm not quite 100%, but based on my experience I'm 99.5% sure that he's having an MI right now. This is my posttest probability - the new impression I have that the patient has the disease after we did our test.

And likelihood ration? That's just the name for the statistical tool that converted the pretest probability to the posttest probability - it's just a mathematical description of the strength of that test.

Using an online calculator, that means the LR+ that got me from 60% to 99.5% is 145, which is about as high an LR you can get (and the actual LR for an emergency physician who thinks an ECG shows an obvious STEMI).

(Thank you to Seth Trueger, MD for this explanation!)

POCUS Atlas: Pulmonary Embolism

Diagnostics and Likelihood Ratios, Explained

Positive Findings (Patient Has This)

Finding Increased Disease Probability (Positive Likelihood Ratio)
Goal-directed RV dysfunction Echo score ≥1 90
Note: accuracy of ultrasound is operator-dependent. Reported LRs may not be reproducible by an inexperienced sonographer.

Negative Findings (Patient Doesn't Have This)

Finding Decreased Disease Probability (Negative Likelihood Ratio)
Goal-directed RV dysfunction Echo score ≥1 0.0
Note: accuracy of ultrasound is operator-dependent. Reported LRs may not be reproducible by an inexperienced sonographer.

Narrative

This was a prospective observational study (n=116) of consecutive normotensive patients with confirmed pulmonary embolism, assessing the diagnostic accuracy of biomarkers, CT, and goal-directed echocardiography for right ventricular dysfunction. Emergency physicians, blinded to clot burden and biomarkers, performed qualitative goal-directed echocardiography for right ventricular (RV) dysfunction*: RV enlargement (RV diameter greater than or equal to that of the left ventricle), severe RV systolic dysfunction (RV free wall hypokinesis or TAPSE < 1.0 cm), and/or interventricular septum flattening or bowing into the left ventricle. If any one of these were present, right ventricular dysfunction was diagnosed. Goal-directed echocardiography results were compared to comprehensive echocardiography as the gold-standard.

Caveats

Note: accuracy of ultrasound is operator-dependent. Reported LRs may not be reproducible by an inexperienced sonographer.

Published in collaboration with The POCUS Atlas

Author

Matthew Riscinti, MD

Published/Updated

What are Likelihood Ratios?

LR, pretest probability and posttest (or posterior) probability are daunting terms that describe simple concepts that we all intuitively understand.

Let's start with pretest probability: that's just a fancy term for my initial impression, before we perform whatever test it is that we're going to use.

For example, a patient with prior stents comes in sweating and clutching his chest in agony, I have a pretty high suspicion that he's having an MI – let's say, 60%. That is my pretest probability.

He immediately gets an ECG (known here as the "test") showing an obvious STEMI.

Now, I know there are some STEMI mimics, so I'm not quite 100%, but based on my experience I'm 99.5% sure that he's having an MI right now. This is my posttest probability - the new impression I have that the patient has the disease after we did our test.

And likelihood ration? That's just the name for the statistical tool that converted the pretest probability to the posttest probability - it's just a mathematical description of the strength of that test.

Using an online calculator, that means the LR+ that got me from 60% to 99.5% is 145, which is about as high an LR you can get (and the actual LR for an emergency physician who thinks an ECG shows an obvious STEMI).

(Thank you to Seth Trueger, MD for this explanation!)

POCUS Atlas: B-Lines for Acute Cardiogenic Pulmonary Edema

Diagnostics and Likelihood Ratios, Explained

Positive Findings (Patient Has This)

Finding Increased Disease Probability (Positive Likelihood Ratio)
≥3 B-lines in two bilateral lung zones 12.38
Note: accuracy of ultrasound is operator-dependent. Reported LRs may not be reproducible by an inexperienced sonographer.

Negative Findings (Patient Doesn't Have This)

Finding Decreased Disease Probability (Negative Likelihood Ratio)
≥3 B-lines in two bilateral lung zones 0.06
Note: accuracy of ultrasound is operator-dependent. Reported LRs may not be reproducible by an inexperienced sonographer.

Narrative

This was a systematic review including 7 prospective case control or cohort studies (n=1075) evaluating the sensitivity and specificity of B-lines in diagnosing acute cardiogenic pulmonary edema (ACPE). The included studies recruited patients presenting to the hospital with acute dyspnea, or where there was a clinical suspicion of congestive heart failure. The setting was either the emergency department (ED) , ICU, or inpatient wards. Ultrasound examinations were performed by any non-radiologist physician. *Various lung ultrasound protocols were used, including the Volpicelli method, the Lichtenstein protocol, and the Comet Score. All involved using B-lines to make the diagnosis of ACPE. The varied protocols used for diagnosis may explain the increased sensitivity noted in this study compared to other meta-analysis. Gold standard was heterogeneous amongst studies with 'final diagnosis from clinical follow-up' being an acceptable reference standard.

Caveats

Note: accuracy of ultrasound is operator-dependent. Reported LRs may not be reproducible by an inexperienced sonographer.

Published in collaboration with The POCUS Atlas

Author

Kyle Kelson, MD

Published/Updated

What are Likelihood Ratios?

LR, pretest probability and posttest (or posterior) probability are daunting terms that describe simple concepts that we all intuitively understand.

Let's start with pretest probability: that's just a fancy term for my initial impression, before we perform whatever test it is that we're going to use.

For example, a patient with prior stents comes in sweating and clutching his chest in agony, I have a pretty high suspicion that he's having an MI – let's say, 60%. That is my pretest probability.

He immediately gets an ECG (known here as the "test") showing an obvious STEMI.

Now, I know there are some STEMI mimics, so I'm not quite 100%, but based on my experience I'm 99.5% sure that he's having an MI right now. This is my posttest probability - the new impression I have that the patient has the disease after we did our test.

And likelihood ration? That's just the name for the statistical tool that converted the pretest probability to the posttest probability - it's just a mathematical description of the strength of that test.

Using an online calculator, that means the LR+ that got me from 60% to 99.5% is 145, which is about as high an LR you can get (and the actual LR for an emergency physician who thinks an ECG shows an obvious STEMI).

(Thank you to Seth Trueger, MD for this explanation!)

POCUS Atlas: Pneumothorax

Diagnostics and Likelihood Ratios, Explained

Positive Findings (Patient Has This)

Finding Increased Disease Probability (Positive Likelihood Ratio)
No lung sliding and lack of comet tails 50.5
Note: accuracy of ultrasound is operator-dependent. Reported LRs may not be reproducible by an inexperienced sonographer.

Negative Findings (Patient Doesn't Have This)

Finding Decreased Disease Probability (Negative Likelihood Ratio)
No lung sliding and lack of comet tails 0.09
Note: accuracy of ultrasound is operator-dependent. Reported LRs may not be reproducible by an inexperienced sonographer.

Narrative

This was a systematic review of 8 prospective studies (n=1048) of adult patients. Included manuscripts evaluated for traumatic or iatrogenic pneumothorax. No studies that screened for spontaneous pneumothorax were included. Examiners were surgeons, radiologists, or emergency providers. Reference standard was pneumothorax found on CT or a rush of air upon tube thoracostomy. All studies but one used the ultrasonographic signs of lung sliding and comet tail to rule out pneumothorax. Although the exact technique used to perform the ultrasound examination is not reported with enough detail in some studies, most agree on requiring the examination of more than one intercostal space in both the midclavicular line and laterally and inferiorly at the anterior or midaxillary lines. Lastly, this data does not evaluate whether the pneumothoraces identified were clinically significant.

Caveats

Note: accuracy of ultrasound is operator-dependent. Reported LRs may not be reproducible by an inexperienced sonographer.

Published in collaboration with The POCUS Atlas

Author

Kyle Kelson, MD

Published/Updated

What are Likelihood Ratios?

LR, pretest probability and posttest (or posterior) probability are daunting terms that describe simple concepts that we all intuitively understand.

Let's start with pretest probability: that's just a fancy term for my initial impression, before we perform whatever test it is that we're going to use.

For example, a patient with prior stents comes in sweating and clutching his chest in agony, I have a pretty high suspicion that he's having an MI – let's say, 60%. That is my pretest probability.

He immediately gets an ECG (known here as the "test") showing an obvious STEMI.

Now, I know there are some STEMI mimics, so I'm not quite 100%, but based on my experience I'm 99.5% sure that he's having an MI right now. This is my posttest probability - the new impression I have that the patient has the disease after we did our test.

And likelihood ration? That's just the name for the statistical tool that converted the pretest probability to the posttest probability - it's just a mathematical description of the strength of that test.

Using an online calculator, that means the LR+ that got me from 60% to 99.5% is 145, which is about as high an LR you can get (and the actual LR for an emergency physician who thinks an ECG shows an obvious STEMI).

(Thank you to Seth Trueger, MD for this explanation!)

POCUS Atlas: Pleural Effusion

Diagnostics and Likelihood Ratios, Explained

Positive Findings (Patient Has This)

Finding Increased Disease Probability (Positive Likelihood Ratio)
Anechoic fluid above diaphragm 47
Note: accuracy of ultrasound is operator-dependent. Reported LRs may not be reproducible by an inexperienced sonographer.

Negative Findings (Patient Doesn't Have This)

Finding Decreased Disease Probability (Negative Likelihood Ratio)
Anechoic fluid above diaphragm 0.06
Note: accuracy of ultrasound is operator-dependent. Reported LRs may not be reproducible by an inexperienced sonographer.

Narrative

This was a meta-analysis including 12 retrospective and prospective studies (n=1554) of adults and pediatric patients. Ultrasound was used to diagnose pleural effusion, with the reference standard either CT, surgery, or a more formal “high quality ultrasound in conjunction with expert end diagnosis.” Ultrasound examinations were performed by a variety of operators including emergency physicians, intensivists, radiologists, and nurses. Exact criteria for diagnosis of a pleural effusion by ultrasound was not defined.

Caveats

Note: accuracy of ultrasound is operator-dependent. Reported LRs may not be reproducible by an inexperienced sonographer.

Published in collaboration with The POCUS Atlas

Author

Kyle Kelson, MD; Matthew Riscinti

Published/Updated

What are Likelihood Ratios?

LR, pretest probability and posttest (or posterior) probability are daunting terms that describe simple concepts that we all intuitively understand.

Let's start with pretest probability: that's just a fancy term for my initial impression, before we perform whatever test it is that we're going to use.

For example, a patient with prior stents comes in sweating and clutching his chest in agony, I have a pretty high suspicion that he's having an MI – let's say, 60%. That is my pretest probability.

He immediately gets an ECG (known here as the "test") showing an obvious STEMI.

Now, I know there are some STEMI mimics, so I'm not quite 100%, but based on my experience I'm 99.5% sure that he's having an MI right now. This is my posttest probability - the new impression I have that the patient has the disease after we did our test.

And likelihood ration? That's just the name for the statistical tool that converted the pretest probability to the posttest probability - it's just a mathematical description of the strength of that test.

Using an online calculator, that means the LR+ that got me from 60% to 99.5% is 145, which is about as high an LR you can get (and the actual LR for an emergency physician who thinks an ECG shows an obvious STEMI).

(Thank you to Seth Trueger, MD for this explanation!)

POCUS Atlas: Pneumonia (Adults and Peds)

Diagnostics and Likelihood Ratios, Explained

Positive Findings (Patient Has This)

Finding Increased Disease Probability (Positive Likelihood Ratio)
Hepatization, shred sign, air bronchograms 12.14
Note: accuracy of ultrasound is operator-dependent. Reported LRs may not be reproducible by an inexperienced sonographer.

Negative Findings (Patient Doesn't Have This)

Finding Decreased Disease Probability (Negative Likelihood Ratio)
Hepatization, shred sign, air bronchograms 0.16
Note: accuracy of ultrasound is operator-dependent. Reported LRs may not be reproducible by an inexperienced sonographer.

Narrative

The was a systematic review including 20 prospective adult and pediatric studies (n=2513) with varied settings including the emergency department, inpatient wards, or ICU. A positive finding on ultrasound was identified as an alveolar and interstitial pattern or consolidation, although this is not further expanded upon. Gold standard was either CT, chest radiography, or “clinical diagnosis” depending on the study. One large caveat of this study is that it has a very large degree of heterogeneity, with ultrasound examinations performed by emergency physicians, intensivists, and radiologists of varying levels of expertise, on patients ranging from ambulatory to critically ill.

Caveats

Note: accuracy of ultrasound is operator-dependent. Reported LRs may not be reproducible by an inexperienced sonographer.

Published in collaboration with The POCUS Atlas

Author

Kyle Kelson, MD

Published/Updated

What are Likelihood Ratios?

LR, pretest probability and posttest (or posterior) probability are daunting terms that describe simple concepts that we all intuitively understand.

Let's start with pretest probability: that's just a fancy term for my initial impression, before we perform whatever test it is that we're going to use.

For example, a patient with prior stents comes in sweating and clutching his chest in agony, I have a pretty high suspicion that he's having an MI – let's say, 60%. That is my pretest probability.

He immediately gets an ECG (known here as the "test") showing an obvious STEMI.

Now, I know there are some STEMI mimics, so I'm not quite 100%, but based on my experience I'm 99.5% sure that he's having an MI right now. This is my posttest probability - the new impression I have that the patient has the disease after we did our test.

And likelihood ration? That's just the name for the statistical tool that converted the pretest probability to the posttest probability - it's just a mathematical description of the strength of that test.

Using an online calculator, that means the LR+ that got me from 60% to 99.5% is 145, which is about as high an LR you can get (and the actual LR for an emergency physician who thinks an ECG shows an obvious STEMI).

(Thank you to Seth Trueger, MD for this explanation!)

POCUS Atlas: Pneumonia (Adults)

Diagnostics and Likelihood Ratios, Explained

Positive Findings (Patient Has This)

Finding Increased Disease Probability (Positive Likelihood Ratio)
Subpleural consolidation and/or focal B-lines 15.3

Negative Findings (Patient Doesn't Have This)

Finding Decreased Disease Probability (Negative Likelihood Ratio)
Subpleural consolidation and/or focal B-lines 0.09
Note: while the meta-analysis found a negative LR of 0.09, the authors noted that there are some concerns regarding the lack of standardized methodology in the studies included (reference standards, sonographer experience, and ultrasound patterns considered significant for CAP). As always, apply clinical judgment when interpreting these results in patients.

Narrative

This was a systematic review including 17 prospective studies (n=5108) evaluating the operating characteristics of lung ultrasound for pneumonia in adult patients seen in the emergency department with a clinical suspicion for this diagnosis. Reference standard was either chest radiograph, chest CT, or final clinical diagnosis. Included studies varied with regards to which and how many lung fields were evaluated. The operators performing ultrasound examinations were exclusively emergency physicians or radiologists.*Subpleural consolidation and/or focal B-lines were the diagnostic criteria in the majority of manuscripts included, however in 4 studies, no clear positive findings were specified.

Caveats

Note: accuracy of ultrasound is operator-dependent. Reported LRs may not be reproducible by an inexperienced sonographer.

Published in collaboration with The POCUS Atlas

Author

Kyle Kelson, MD

Published/Updated

What are Likelihood Ratios?

LR, pretest probability and posttest (or posterior) probability are daunting terms that describe simple concepts that we all intuitively understand.

Let's start with pretest probability: that's just a fancy term for my initial impression, before we perform whatever test it is that we're going to use.

For example, a patient with prior stents comes in sweating and clutching his chest in agony, I have a pretty high suspicion that he's having an MI – let's say, 60%. That is my pretest probability.

He immediately gets an ECG (known here as the "test") showing an obvious STEMI.

Now, I know there are some STEMI mimics, so I'm not quite 100%, but based on my experience I'm 99.5% sure that he's having an MI right now. This is my posttest probability - the new impression I have that the patient has the disease after we did our test.

And likelihood ration? That's just the name for the statistical tool that converted the pretest probability to the posttest probability - it's just a mathematical description of the strength of that test.

Using an online calculator, that means the LR+ that got me from 60% to 99.5% is 145, which is about as high an LR you can get (and the actual LR for an emergency physician who thinks an ECG shows an obvious STEMI).

(Thank you to Seth Trueger, MD for this explanation!)

POCUS Atlas: Nephrolithiasis

Diagnostics and Likelihood Ratios, Explained

Positive Findings (Patient Has This)

Finding Increased Disease Probability (Positive Likelihood Ratio)
Any hydronephrosis 2.85
Moderate to severe hydronephrosis 0.39
Note: accuracy of ultrasound is operator-dependent. Reported LRs may not be reproducible by an inexperienced sonographer.

Negative Findings (Patient Doesn't Have This)

Finding Decreased Disease Probability (Negative Likelihood Ratio)
Any hydronephrosis 0.39
Moderate to severe hydronephrosis 0.76
Note: accuracy of ultrasound is operator-dependent. Reported LRs may not be reproducible by an inexperienced sonographer.

Narrative

This was a meta-analysis including 5 studies (n=1773) evaluating the accuracy of POCUS to diagnose nephrolithiasis in adult patients presenting to the emergency department with symptoms suggestive of renal colic (flank pain, dysuria, abdominal pain radiating to groin). Reference standards included CT, direct stone visualization, or surgical findings. Specificity improved significantly (~94.4%) for moderate to severe hydronephrosis (i.e. exclusion of mild hydronephrosis). This systematic review has some flaws (most importantly not having a single uniform gold standard) but appears to be the best available evidence.

Caveats

Note: accuracy of ultrasound is operator-dependent. Reported LRs may not be reproducible by an inexperienced sonographer.

Published in collaboration with The POCUS Atlas

Author

Roshanak Benabbas, MD

Published/Updated

What are Likelihood Ratios?

LR, pretest probability and posttest (or posterior) probability are daunting terms that describe simple concepts that we all intuitively understand.

Let's start with pretest probability: that's just a fancy term for my initial impression, before we perform whatever test it is that we're going to use.

For example, a patient with prior stents comes in sweating and clutching his chest in agony, I have a pretty high suspicion that he's having an MI – let's say, 60%. That is my pretest probability.

He immediately gets an ECG (known here as the "test") showing an obvious STEMI.

Now, I know there are some STEMI mimics, so I'm not quite 100%, but based on my experience I'm 99.5% sure that he's having an MI right now. This is my posttest probability - the new impression I have that the patient has the disease after we did our test.

And likelihood ration? That's just the name for the statistical tool that converted the pretest probability to the posttest probability - it's just a mathematical description of the strength of that test.

Using an online calculator, that means the LR+ that got me from 60% to 99.5% is 145, which is about as high an LR you can get (and the actual LR for an emergency physician who thinks an ECG shows an obvious STEMI).

(Thank you to Seth Trueger, MD for this explanation!)

POCUS Atlas: Abscess

Diagnostics and Likelihood Ratios, Explained

Positive Findings (Patient Has This)

Finding Increased Disease Probability (Positive Likelihood Ratio)
Direct visualization 5.71
Note: accuracy of ultrasound is operator-dependent. Reported LRs may not be reproducible by an inexperienced sonographer.

Negative Findings (Patient Doesn't Have This)

Finding Decreased Disease Probability (Negative Likelihood Ratio)
Direct visualization 0.04
Note: accuracy of ultrasound is operator-dependent. Reported LRs may not be reproducible by an inexperienced sonographer.

Narrative

This was a systematic review including 6 prospective observational studies (n=800) comparing physical exam and POCUS to diagnose abscess in both children and adults in the emergency department (ED). Inclusion criteria for all but one study was the presence of a skin & soft tissue infection (SSTI) while once study specifically included patients where abscess was suspected with plan to incise and drain. The reference standard for abscess was pus drainage on initial I&D or at follow up. In many studies, facial and genital/rectal area SSTIs were excluded therefore these results should be interpreted with caution in evaluation of SSTIs in these areas.

Caveats

Note: accuracy of ultrasound is operator-dependent. Reported LRs may not be reproducible by an inexperienced sonographer.

Published in collaboration with The POCUS Atlas

Author

Matthew Riscinti, MD

Published/Updated

What are Likelihood Ratios?

LR, pretest probability and posttest (or posterior) probability are daunting terms that describe simple concepts that we all intuitively understand.

Let's start with pretest probability: that's just a fancy term for my initial impression, before we perform whatever test it is that we're going to use.

For example, a patient with prior stents comes in sweating and clutching his chest in agony, I have a pretty high suspicion that he's having an MI – let's say, 60%. That is my pretest probability.

He immediately gets an ECG (known here as the "test") showing an obvious STEMI.

Now, I know there are some STEMI mimics, so I'm not quite 100%, but based on my experience I'm 99.5% sure that he's having an MI right now. This is my posttest probability - the new impression I have that the patient has the disease after we did our test.

And likelihood ration? That's just the name for the statistical tool that converted the pretest probability to the posttest probability - it's just a mathematical description of the strength of that test.

Using an online calculator, that means the LR+ that got me from 60% to 99.5% is 145, which is about as high an LR you can get (and the actual LR for an emergency physician who thinks an ECG shows an obvious STEMI).

(Thank you to Seth Trueger, MD for this explanation!)

POCUS Atlas: Shoulder Dislocation and Reduction

Diagnostics and Likelihood Ratios, Explained

Positive Findings (Patient Has This)

Finding Increased Disease Probability (Positive Likelihood Ratio)
Abnormal humeral head location 100000
Note: accuracy of ultrasound is operator-dependent. Reported LRs may not be reproducible by an inexperienced sonographer. LR+ is infinity; simplified to 100,000 for hte purposes of this interactive tool.

Negative Findings (Patient Doesn't Have This)

Finding Decreased Disease Probability (Negative Likelihood Ratio)
Abnormal humeral head location 0.0
Note: accuracy of ultrasound is operator-dependent. Reported LRs may not be reproducible by an inexperienced sonographer.

Narrative

This was a prospective observational study (n= 73) evaluating the diagnostic accuracy of ultrasound for both identification of shoulder dislocation and confirmation of successful reduction. Patients with suspected shoulder dislocation were enrolled, with 69/73 having dislocation confirmed by shoulder radiographs. Patients with multi trauma or decreased level of consciousness were excluded. Both the anterior and lateral approach were used to evaluate for dislocation. Examinations were performed by either an experienced ultrasound trained emergency physician or a senior emergency medicine resident. An ultrasound exam was performed prior to initial shoulder radiographs and after reduction attempt. Ultrasound was found to be 100% accurate with respect to both shoulder dislocation identification and reduction confirmation. *This is a limited study as the sample size is relatively small and this ultrasound examination is dependent on operator experience therefore these results should be interpreted with caution.

Caveats

Note: accuracy of ultrasound is operator-dependent. Reported LRs may not be reproducible by an inexperienced sonographer.

Published in collaboration with The POCUS Atlas

Author

Mike Macias, MD

Published/Updated

What are Likelihood Ratios?

LR, pretest probability and posttest (or posterior) probability are daunting terms that describe simple concepts that we all intuitively understand.

Let's start with pretest probability: that's just a fancy term for my initial impression, before we perform whatever test it is that we're going to use.

For example, a patient with prior stents comes in sweating and clutching his chest in agony, I have a pretty high suspicion that he's having an MI – let's say, 60%. That is my pretest probability.

He immediately gets an ECG (known here as the "test") showing an obvious STEMI.

Now, I know there are some STEMI mimics, so I'm not quite 100%, but based on my experience I'm 99.5% sure that he's having an MI right now. This is my posttest probability - the new impression I have that the patient has the disease after we did our test.

And likelihood ration? That's just the name for the statistical tool that converted the pretest probability to the posttest probability - it's just a mathematical description of the strength of that test.

Using an online calculator, that means the LR+ that got me from 60% to 99.5% is 145, which is about as high an LR you can get (and the actual LR for an emergency physician who thinks an ECG shows an obvious STEMI).

(Thank you to Seth Trueger, MD for this explanation!)

POCUS Atlas: Free Fluid on FAST for Ectopic Pregnancy Requiring the OR

Diagnostics and Likelihood Ratios, Explained

Positive Findings (Patient Has This)

Finding Increased Disease Probability (Positive Likelihood Ratio)
Free fluid in Morison's pouch 112
Free fluid in pelvis 9.5
Note: accuracy of ultrasound is operator-dependent. Reported LRs may not be reproducible by an inexperienced sonographer.

Negative Findings (Patient Doesn't Have This)

Finding Decreased Disease Probability (Negative Likelihood Ratio)
Free fluid in Morison's pouch 0.5
Free fluid in pelvis 0.47
Note: accuracy of ultrasound is operator-dependent. Reported LRs may not be reproducible by an inexperienced sonographer.

Narrative

This was a prospective observational study (n=242) of patients presenting to the emergency department with suspected ectopic pregnancy, evaluating if free fluid in the peritoneal cavity identified by bedside ultrasound was predictive of the need for operative intervention. The included patients were suspected to have an ectopic pregnancy based on positive pregnancy test results, in their first trimester, with either abdominal pain or vaginal bleeding. All examinations were performed by emergency providers (EPs) using a transabdominal approach, to determine if free fluid was present in the hepatorenal space (Morison’s Pouch) and/or pelvis. Reference standard was chart review by one of four study investigators who were blinded to the EP performed US results. There was one patient who had free fluid in Morison’s Pouch but had a confirmed IUP (suspected to be a ruptured corpus luteum cyst) and no ruptured ectopic pregnancy. This study supports that free fluid present in Morison’s pouch in patients with suspected ectopic pregnancy strongly predicts the need for operative intervention.

Caveats

Note: accuracy of ultrasound is operator-dependent. Reported LRs may not be reproducible by an inexperienced sonographer.

Published in collaboration with The POCUS Atlas

Author

Garrett Ghent, MD

Published/Updated

What are Likelihood Ratios?

LR, pretest probability and posttest (or posterior) probability are daunting terms that describe simple concepts that we all intuitively understand.

Let's start with pretest probability: that's just a fancy term for my initial impression, before we perform whatever test it is that we're going to use.

For example, a patient with prior stents comes in sweating and clutching his chest in agony, I have a pretty high suspicion that he's having an MI – let's say, 60%. That is my pretest probability.

He immediately gets an ECG (known here as the "test") showing an obvious STEMI.

Now, I know there are some STEMI mimics, so I'm not quite 100%, but based on my experience I'm 99.5% sure that he's having an MI right now. This is my posttest probability - the new impression I have that the patient has the disease after we did our test.

And likelihood ration? That's just the name for the statistical tool that converted the pretest probability to the posttest probability - it's just a mathematical description of the strength of that test.

Using an online calculator, that means the LR+ that got me from 60% to 99.5% is 145, which is about as high an LR you can get (and the actual LR for an emergency physician who thinks an ECG shows an obvious STEMI).

(Thank you to Seth Trueger, MD for this explanation!)

POCUS Atlas: Central Line Confirmation

Diagnostics and Likelihood Ratios, Explained

Positive Findings (Patient Has This)

Finding Increased Disease Probability (Positive Likelihood Ratio)
Saline bubbles in the right heart 100000
Note: accuracy of ultrasound is operator-dependent. Reported LRs may not be reproducible by an inexperienced sonographer. LR+ is infinity; set as 100,000 for the purposes of this interactive tool.

Negative Findings (Patient Doesn't Have This)

Finding Decreased Disease Probability (Negative Likelihood Ratio)
Saline bubbles in the right heart 0.13
Note: accuracy of ultrasound is operator-dependent. Reported LRs may not be reproducible by an inexperienced sonographer.

Narrative

This was a prospective study (n=78) in critically ill patients in the emergency department (ED) and intensive care unit (ICU) at a single academic center, evaluating the use of POCUS to identify correct placement of supra-diaphragmatic central venous catheter (CVC) placement. All CVC placements and POCUS exams were performed by resident trainees. Correct positioning of the CVC was considered if turbulent flow was visualized in the right atrium on sub-xiphoid, parasternal, or apical cardiac ultrasound after injecting 5 cc of sterile, non-agitated, normal saline through the CVC. Reference standard was confirmatory chest radiography.

Caveats

Note: accuracy of ultrasound is operator-dependent. Reported LRs may not be reproducible by an inexperienced sonographer.

Published in collaboration with The POCUS Atlas

Author

Matthew Riscinti, MD; John F Kilpatrick, MD

Published/Updated

What are Likelihood Ratios?

LR, pretest probability and posttest (or posterior) probability are daunting terms that describe simple concepts that we all intuitively understand.

Let's start with pretest probability: that's just a fancy term for my initial impression, before we perform whatever test it is that we're going to use.

For example, a patient with prior stents comes in sweating and clutching his chest in agony, I have a pretty high suspicion that he's having an MI – let's say, 60%. That is my pretest probability.

He immediately gets an ECG (known here as the "test") showing an obvious STEMI.

Now, I know there are some STEMI mimics, so I'm not quite 100%, but based on my experience I'm 99.5% sure that he's having an MI right now. This is my posttest probability - the new impression I have that the patient has the disease after we did our test.

And likelihood ration? That's just the name for the statistical tool that converted the pretest probability to the posttest probability - it's just a mathematical description of the strength of that test.

Using an online calculator, that means the LR+ that got me from 60% to 99.5% is 145, which is about as high an LR you can get (and the actual LR for an emergency physician who thinks an ECG shows an obvious STEMI).

(Thank you to Seth Trueger, MD for this explanation!)

POCUS Atlas: Positive FAST for Detecting Intra-abdominal Injury

Diagnostics and Likelihood Ratios, Explained

Positive Findings (Patient Has This)

Finding Increased Disease Probability (Positive Likelihood Ratio)
Positive free fluid 30
Note: accuracy of ultrasound is operator-dependent. Reported LRs may not be reproducible by an inexperienced sonographer.

Negative Findings (Patient Doesn't Have This)

Finding Decreased Disease Probability (Negative Likelihood Ratio)
Positive free fluid 0.26
Note: accuracy of ultrasound is operator-dependent. Reported LRs may not be reproducible by an inexperienced sonographer.

Narrative

This was a meta-analysis including 22 studies (n=12089) in adult patients presenting with blunt trauma, evaluating the accuracy of the FAST examination for the detection of intra-abdominal injury. All included studies had at least 1 reference standard including abdominal computed tomography, diagnostic peritoneal lavage, laparotomy, autopsy, and/or clinical course. 5 of the studies excluded patients with hemodynamic instability. When these studies were looked at alone, the positive LR increased to 82.

Caveats

Note: accuracy of ultrasound is operator-dependent. Reported LRs may not be reproducible by an inexperienced sonographer.

Published in collaboration with The POCUS Atlas

Author

Matthew Riscinti, MD

Published/Updated

What are Likelihood Ratios?

LR, pretest probability and posttest (or posterior) probability are daunting terms that describe simple concepts that we all intuitively understand.

Let's start with pretest probability: that's just a fancy term for my initial impression, before we perform whatever test it is that we're going to use.

For example, a patient with prior stents comes in sweating and clutching his chest in agony, I have a pretty high suspicion that he's having an MI – let's say, 60%. That is my pretest probability.

He immediately gets an ECG (known here as the "test") showing an obvious STEMI.

Now, I know there are some STEMI mimics, so I'm not quite 100%, but based on my experience I'm 99.5% sure that he's having an MI right now. This is my posttest probability - the new impression I have that the patient has the disease after we did our test.

And likelihood ration? That's just the name for the statistical tool that converted the pretest probability to the posttest probability - it's just a mathematical description of the strength of that test.

Using an online calculator, that means the LR+ that got me from 60% to 99.5% is 145, which is about as high an LR you can get (and the actual LR for an emergency physician who thinks an ECG shows an obvious STEMI).

(Thank you to Seth Trueger, MD for this explanation!)

POCUS Atlas: Increased Intracranial Pressure (Optic Nerve Sheath Diameter >5mm)

Diagnostics and Likelihood Ratios, Explained

Positive Findings (Patient Has This)

Finding Increased Disease Probability (Positive Likelihood Ratio)
Optic nerve sheath diameter >5 mm 12.4
Note: accuracy of ultrasound is operator-dependent. Reported LRs may not be reproducible by an inexperienced sonographer.

Negative Findings (Patient Doesn't Have This)

Finding Decreased Disease Probability (Negative Likelihood Ratio)
Optic nerve sheath diameter >5 mm 0.05
Note: accuracy of ultrasound is operator-dependent. Reported LRs may not be reproducible by an inexperienced sonographer.

Narrative

This was a systematic review, which including 12 studies (n = 478) using ultrasound measurement of optic nerve diameter (cut point of 5 mm for adult studies, 4.5 mm for age 1–17 years, and 4 mm for age <1 year) to evaluate for increased intracranial pressure (ICP). There was moderate to high heterogeneity among these studies given multiple patient populations. This resulted in wide confidence intervals: sensitivity of 95.6% (95% CI, 87.7%–98.5%), specificity of 92.3% (95% CI, 77.9%–98.4%), positive likelihood ratio of 12.5 (95% CI, 4.2–37.5), and a negative likelihood ratio of 0.05 (95% CI 0.016–0.14). It is also important to mention that the gold standard in this review was CT, which is not as accurate as invasive ICP monitoring. Overall their conclusions were that ocular sonography had a very low LR- (0.05) making it a good test for ruling out raised ICP in a low-risk group, and a high LR+ (12.4) making it a good test for ruling in raised ICP in a high-risk group.

Caveats

Note: accuracy of ultrasound is operator-dependent. Reported LRs may not be reproducible by an inexperienced sonographer.

Published in collaboration with The POCUS Atlas

Author

Matthew Riscinti, MD

Published/Updated

What are Likelihood Ratios?

LR, pretest probability and posttest (or posterior) probability are daunting terms that describe simple concepts that we all intuitively understand.

Let's start with pretest probability: that's just a fancy term for my initial impression, before we perform whatever test it is that we're going to use.

For example, a patient with prior stents comes in sweating and clutching his chest in agony, I have a pretty high suspicion that he's having an MI – let's say, 60%. That is my pretest probability.

He immediately gets an ECG (known here as the "test") showing an obvious STEMI.

Now, I know there are some STEMI mimics, so I'm not quite 100%, but based on my experience I'm 99.5% sure that he's having an MI right now. This is my posttest probability - the new impression I have that the patient has the disease after we did our test.

And likelihood ration? That's just the name for the statistical tool that converted the pretest probability to the posttest probability - it's just a mathematical description of the strength of that test.

Using an online calculator, that means the LR+ that got me from 60% to 99.5% is 145, which is about as high an LR you can get (and the actual LR for an emergency physician who thinks an ECG shows an obvious STEMI).

(Thank you to Seth Trueger, MD for this explanation!)

POCUS Atlas: Retinal Detachment

Diagnostics and Likelihood Ratios, Explained

Positive Findings (Patient Has This)

Finding Increased Disease Probability (Positive Likelihood Ratio)
Positive proximal CUS, complete CUS, or color flow duplex US 12.1
Note: accuracy of ultrasound is operator-dependent. Reported LRs may not be reproducible by an inexperienced sonographer. LR+ reported as a range and listed here as upper limit for the purposes of this interactive tool (see below for detail).

Negative Findings (Patient Doesn't Have This)

Finding Decreased Disease Probability (Negative Likelihood Ratio)
Negative proximal CUS, complete CUS, or color flow duplex US 0.03
Note: accuracy of ultrasound is operator-dependent. Reported LRs may not be reproducible by an inexperienced sonographer. LR- reported as a range and listed here as upper limit for the purposes of this interactive tool (see below for detail).

Narrative

This was a small systematic review (n = 201) that evaluated the utility of emergency department (ED) performed ocular ultrasound. A total of 3 studies were included in the final analysis and overall the data was low risk for bias. Operating characteristics were not pooled. Among the 3 studies included, there was not a clear definition for a positive test result. In one of the included studies, no positive test definition was provided. These results should be interpreted with caution given the small number of patients included and the operator dependent nature of this POCUS exam. A more recent study with larger variability in operator experience suggests the all-comer sensitivity may be much lower.

Caveats

Note: accuracy of ultrasound is operator-dependent. Reported LRs may not be reproducible by an inexperienced sonographer.

Published in collaboration with The POCUS Atlas

Author

Matthew Riscinti, MD

Published/Updated

What are Likelihood Ratios?

LR, pretest probability and posttest (or posterior) probability are daunting terms that describe simple concepts that we all intuitively understand.

Let's start with pretest probability: that's just a fancy term for my initial impression, before we perform whatever test it is that we're going to use.

For example, a patient with prior stents comes in sweating and clutching his chest in agony, I have a pretty high suspicion that he's having an MI – let's say, 60%. That is my pretest probability.

He immediately gets an ECG (known here as the "test") showing an obvious STEMI.

Now, I know there are some STEMI mimics, so I'm not quite 100%, but based on my experience I'm 99.5% sure that he's having an MI right now. This is my posttest probability - the new impression I have that the patient has the disease after we did our test.

And likelihood ration? That's just the name for the statistical tool that converted the pretest probability to the posttest probability - it's just a mathematical description of the strength of that test.

Using an online calculator, that means the LR+ that got me from 60% to 99.5% is 145, which is about as high an LR you can get (and the actual LR for an emergency physician who thinks an ECG shows an obvious STEMI).

(Thank you to Seth Trueger, MD for this explanation!)

Protected: POCUS Atlas: DVT

Diagnostics and Likelihood Ratios, Explained

Positive Findings (Patient Has This)

Finding Increased Disease Probability (Positive Likelihood Ratio)
Positive proximal CUS, complete CUS, or color flow duplex US 30.03
Note: accuracy of ultrasound is operator-dependent. Reported LRs may not be reproducible by an inexperienced sonographer.

Negative Findings (Patient Doesn't Have This)

Finding Decreased Disease Probability (Negative Likelihood Ratio)
Negative proximal CUS, complete CUS, or color flow duplex US 0.04
Note: accuracy of ultrasound is operator-dependent. Reported LRs may not be reproducible by an inexperienced sonographer.

Narrative

This was a meta-analysis including 16 studies (n=2379) evaluating the operating characteristics of emergency physician (EP) performed ultrasound for the diagnosis of DVT. The providers used color-flow duplex ultrasound in two studies, proximal venous ultrasound in 13 studies (not looking at the calf), and whole-leg venous ultrasound in one study. Reference standard was radiology department ultrasound, vascular lab, or angiography. Considering only high quality studies which met QUADAS-2 Criteria (11 out of 16 initially selected studies), the sensitivity and specificity improved to 97.6% and 96.8% respectively.

Caveats

Note: accuracy of ultrasound is operator-dependent. Reported LRs may not be reproducible by an inexperienced sonographer.

Published in collaboration with The POCUS Atlas

Author

John F Kilpatrick, MD

Published/Updated

June 18, 2018

What are Likelihood Ratios?

LR, pretest probability and posttest (or posterior) probability are daunting terms that describe simple concepts that we all intuitively understand.

Let's start with pretest probability: that's just a fancy term for my initial impression, before we perform whatever test it is that we're going to use.

For example, a patient with prior stents comes in sweating and clutching his chest in agony, I have a pretty high suspicion that he's having an MI – let's say, 60%. That is my pretest probability.

He immediately gets an ECG (known here as the "test") showing an obvious STEMI.

Now, I know there are some STEMI mimics, so I'm not quite 100%, but based on my experience I'm 99.5% sure that he's having an MI right now. This is my posttest probability - the new impression I have that the patient has the disease after we did our test.

And likelihood ration? That's just the name for the statistical tool that converted the pretest probability to the posttest probability - it's just a mathematical description of the strength of that test.

Using an online calculator, that means the LR+ that got me from 60% to 99.5% is 145, which is about as high an LR you can get (and the actual LR for an emergency physician who thinks an ECG shows an obvious STEMI).

(Thank you to Seth Trueger, MD for this explanation!)

Acute Coronary Syndrome

Diagnostics and Likelihood Ratios, Explained

Positive Findings (Patient Has This)

Cardiac Risk Increased Disease Probability (Positive Likelihood Ratio)
Abnormal prior stress test 3.1x (2.0-4.7)
Peripheral arterial disease 2.7x (1.5-4.8)
Prior history of CAD 2.0x (1.4-2.6)
Prior MI 1.6x (1.4-1.7)
Diabetes 1.4x (1.3-1.6)
CVA 1.4x (1.1-1.8)
Male gender 1.3x (1.2-1.3)
Hyperlipidemia 1.3x (1.1-1.5)
Hypertension 1.2x (1.1-1.3)
Any tobacco use 1.1x (0.9-1.3)
Family history of CAD 1.0x (0.9-1.2)
Obesity 1.0x (0.9-1.2)
History of CABG 0.97x (0.5-2.1)
Chest pain characteristics Increased Disease Probability (Positive Likelihood Ratio)
Radiation to both arms 2.6x (1.8-3.7)
Pain similar to prior ischemia 2.2x (2.0-2.6)
Change in pattern over prior 24 hours 2.0x (1.6-2.5)
"Typical" chest pain 1.9x (0.94-2.9)
Pain worse with exertion 1.5x-1.8x
Radiation to neck or jaw 1.5x (1.3-1.8)
Recent episode of similar pain 1.3x (1.1-1.4)
Radiation to left arm 1.3x (1.2-1.4)
Radiation to right arm 1.3x (0.78-2.1)
Pain with diaphoresis 1.3x-1.4x
Pain with dyspnea 1.2x (1.1-1.3)
Abrupt onset 1.1x (1.0-1.2)
Any improvement with nitroglycerin 1.1x (0.93-1.3)
"Typical" radiation 1.0x-5.7x
Burning pain 1.0x-1.4x
Associated nausea/vomiting 0.92x-1.1x
Associated palpitations 0.71x (0.37-1.3)
Associated syncope 0.55x (0.39-0.76)
Pleuritic pain 0.35x-0.61x
Physical finding Increased Disease Probability (Positive Likelihood Ratio)
Hypotension (SBP<100) 3.9x (0.98-15)
Lung rales 2.0x (1.0-4.0)
Tachypnea 1.9x (0.99-3.5)
Tachycardia (heart rate>120) 1.3x (0.42-3.94)
Pain reproduced on palpation 0.28x (0.14-0.54)
ECG findings Increased Disease Probability (Positive Likelihood Ratio)
ST depression 5.3x (2.1-8.6)
Ischemic ECG 3.6x (1.6-5.7)
T wave inversion 1.8x (1.3-2.7)
Clinical Decision Tools Summary Likelihood Ratio
High
HEART score 7-10 13x (7.0-24)
TIMI score 5-7 6.8x (5.2-8.9)
Intermediate
HEART score 5-6 2.4x (1.6-3.6)
TIMI score 3-4 2.4x (2.1-2.7)
HFA/CSANZ rule (high risk) 2.8x (2.6-3.0)
Indeterminate
HEART score 4 0.79x (0.53-1.2)
TIMI score 2 0.94x (0.85-1.0)
Low
HEART score 0-3 0.20x (0.13-0.30)
TIMI score 0-1 0.31x (0.23-0.43)
HFA/CSANZ rule (low to intermediate) 0.24x (0.19-0.31)

Abbreviations: HEART, History, Electrocardiogram, Age, Risk Factors, Troponin; HFA/CSANZ, The Heart Foundation of Australia and Cardiac Society of Australia and New Zealand; TIMI, Thrombolysis in Myocardial Infarction.

Negative Findings (Patient Doesn't Have This)

Cardiac Risk Decreased Disease Probability (Negative Likelihood Ratio)
Male gender 0.70x (0.64-0.77)
Prior history of CAD 0.75x (0.56-0.93)
Hypertension 0.78x (0.72-0.85)
Hyperlipidemia 0.85x (0.77-0.93)
Prior MI 0.88x (0.81-0.93)
Diabetes 0.90x (0.86-0.94)
Abnormal prior stress test 0.92x (0.88-0.96)
Any tobacco use 0.96x (0.85-1.1)
Peripheral arterial disease 0.96x (0.94-0.98)
CVA 0.97x (0.94-0.99
Obesity 0.99x (0.88-1.1)
Family history of CAD 0.99x (0.91-1.1)
History of CABG 1.00x (0.92-1.1)
Chest pain characteristics Decreased Disease Probability (Negative Likelihood Ratio)
"Typical" chest pain 0.52x (0.35-0.69)
Pain worse with exertion 0.66x-0.83x
Pain similar to prior ischemia 0.67x (0.60-0.74)
Abrupt onset 0.75x (0.61-0.91)
"Typical" radiation 0.78x-0.9x
Recent episode of similar pain 0.80x (0.71-0.90)
Change in pattern over prior 24 hours 0.84x (0.79-0.90)
Radiation to left arm 0.88x (0.81-0.96)
Pain with dyspnea 0.89x (0.82-0.96)
Any improvement with nitroglycerin 0.90x (0.85-0.96)
Radiation to neck or jaw 0.91x (0.87-0.95)
Pain with diaphoresis 0.91x-0.93x
Radiation to both arms 0.93x (0.89-0.9)
Burning pain 0.97x-1.0x
Associated nausea/vomiting 0.98x-1.0x
Radiation to right arm 0.99x (0.96-1.0)
Associated palpitations 1.0x (0.98-1.1)
Associated syncope 1.1x (1.1-1.1)
Pleuritic pain 1.1x-1.2x
Physical finding Decreased Disease Probability (Negative Likelihood Ratio)
Tachypnea 0.95x (0.89-1.0)
Lung rales 0.95x (0.90-1.0)
Hypotension (SBP <100) 0.98x (0.95-1.0)
Tachycardia (HR >120) 0.99x (0.96-1.0)
Pain reproduced on palpation 1.2x (1.0-1.2)
ECG findings Decreased Disease Probability (Negative Likelihood Ratio)
Ischemic ECG 0.74x (0.68-0.81)
ST depression 0.79x (0.71-0.87)
T wave inversion 0.89x (0.86-0.93)

Narrative

This rational clinical examination confirms the findings of previous studies indicating that traditional coronary artery risk factors (e.g. diabetes, hypertension, etc.) are not useful predictors of acute coronary syndrome. This systematic review also shows that isolated signs and symptoms are not helpful in identifying the underlying ischemic etiology for chest pain. The review states that the most rational approach would be to use TIMI or HEART tools in combination with the institution's background prevalence of ACS to calculate the initial pre- stress test probability of ACS. Unfortunately, the risk scores alone may not be adequate to lower the probability sufficiently to achieve a desired miss rate lower than 1%.

The authors of this systematic review employed the following definition of ACS as their reference standard: either final hospital discharge diagnosis of ACS [either as determined by the treating physician or by systematic central adjudication by reviewers using a pre specified definition of ACS] or clinical cardiac events [encompassing at least cardiovascular death, myocardial infarction, and revascularization] through 14 days to 6 weeks after presentation). A review of the included trials revealed that some of the studies considered events such as “revascularization” and even “conservative management of coronary artery disease” as indicator of ACS diagnosis at discharge. We acknowledge that this method of determining ACS outcome is far from ideal. Therefore we expect that the results of future high quality trials with a proper gold standard for ACS (death or MI) change the likelihood ratios dramatically.

Caveats

Author

Shahriar Zehtabchi, MD

Published/Updated

February 1, 2017

What are Likelihood Ratios?

LR, pretest probability and posttest (or posterior) probability are daunting terms that describe simple concepts that we all intuitively understand.

Let's start with pretest probability: that's just a fancy term for my initial impression, before we perform whatever test it is that we're going to use.

For example, a patient with prior stents comes in sweating and clutching his chest in agony, I have a pretty high suspicion that he's having an MI – let's say, 60%. That is my pretest probability.

He immediately gets an ECG (known here as the "test") showing an obvious STEMI.

Now, I know there are some STEMI mimics, so I'm not quite 100%, but based on my experience I'm 99.5% sure that he's having an MI right now. This is my posttest probability - the new impression I have that the patient has the disease after we did our test.

And likelihood ration? That's just the name for the statistical tool that converted the pretest probability to the posttest probability - it's just a mathematical description of the strength of that test.

Using an online calculator, that means the LR+ that got me from 60% to 99.5% is 145, which is about as high an LR you can get (and the actual LR for an emergency physician who thinks an ECG shows an obvious STEMI).

(Thank you to Seth Trueger, MD for this explanation!)

Dyspnea Due to Heart Failure (With Chronic Respiratory Disease)

Diagnostics and Likelihood Ratios, Explained

Positive Findings (Patient Has This)

Risk FactorsIncreased Disease Probability (Positive Likelihood Ratio)
History of Atrial fibrillation4.1× (2.5-6.6)
History of Coronary artery bypass grafting2.8× (1.3-5.8)
History of Myocardial infarction2.2× (1.4-3.5)
History of Diabetes mellitus2.0× (1.3-3.2)
History of Coronary artery disease2.0× (1.5-2.6)
History of Angina1.7× (1.0-2.5)
History of Hypertension1.2× (0.95-1.5)
SymptomsIncreased Disease Probability (Positive Likelihood Ratio)
Orthopnea1.3× (1.1-1.5)
Fatigue1.1× (0.96-1.3)
Nocturnal cough0.93× (0.73-1.2)
SignsIncreased Disease Probability (Positive Likelihood Ratio)
3rd heart sound (ventricullar filling gallop)57.0× (7.6-425)
JVD4.3× (2.8-6.5)
Lower extremity edema2.7× (2.2-3.5)
Rales2.6× (2.1-3.3)
Hepatic congestion2.4× (1.2-4.7)
Enlarged heart1.6× (0.43-6.2)
Wheezing0.85× (0.65-1.1)
OtherIncreased Disease Probability (Positive Likelihood Ratio)
Clinician's Gestalt9.9× (5.3-18)
CXR FindingsIncreased Disease Probability (Positive Likelihood Ratio)
Edema11.0× (5.8-22.0)
Cardiolmegaly7.1× (4.5-11.0)
Pleural effusion(s)4.6× (2.6-8.0)
Pneumonia1.0× (0.46-2.3)
Hyperventilation0.53× (0.25-1.1)
Normal0.11× (0.04-0.28)
EKG FindingsIncreased Disease Probability (Positive Likelihood Ratio)
Atrial fibrillation6.0× (3.4-10.0)
Ischemic ST-T waves4.6× (2.4-8.7)
Q waves3.1× (1.8-5.5)
BNP LevelsIncreased Disease Probability (Positive Likelihood Ratio)
BNP ≥100 pg/ml4.1 (3.3-5.0)

Negative Findings (Patient Doesn't Have This)

Risk FactorsDecreased Disease Probability (Negative Likelihood Ratio)
Coronary artery disease0.67× (0.54-0.84)
Atrial fibrillation0.74× (0.63-0.85)
Myocardial infarction0.84× (0.74-0.96)
Hypertension0.84× (0.65-1.1)
Diabetes mellitus0.85× (0.74-0.97)
Angina0.90× (0.80-1.0)
Coronary artery bypass grafting0.92× (0.84-0.99)
SymptomsDecreased Disease Probability (Negative Likelihood Ratio)
Orthopnea0.68× (0.48-0.95)
Fatigue0.79× (0.54-1.2)
Nocturnal cough1.1× (0.85-1.4)
SignsDecreased Disease Probability (Negative Likelihood Ratio)
Rales0.39× (0.28-0.55)
Lower extremity edema0.41× (0.30-0.57)
JVD0.65× (0.54-0.78)
3rd heart sound (ventricullar filling gallop)0.83× (0.75-0.91)
Hepatic congestion0.91× (0.84-1.0)
Enlarged heart0.99× (0.95-1.0)
Wheezing1.2× (0.94-1.4)
OtherDecreased Disease Probability (Negative Likelihood Ratio)
Clinician's Gestalt0.65× (0.55-0.77)
CXR FindingsDecreased Disease Probability (Negative Likelihood Ratio)
Cardiolmegaly0.54× (0.44-0.67)
Edema0.68× (0.58-0.79)
Pleural effusion(s)0.78× (0.69-0.89)
Pneumonia1.0× (0.93-1.1)
Hyperventilation1.1× (1.0-1.2)
Normal1.7×(1.5-1.8)
EKG FindingsDecreased Disease Probability (Negative Likelihood Ratio)
Atrial fibrillation0.73× (0.63-0.84)
Ischemic ST-T waves0.83× (0.74-0.93)
Q waves0.84× (0.75-0.94)
BNP LevelsDecreased Disease Probability (Negative Likelihood Ratio)
BNP ≥100 pg/ml0.09× (0.04-0.19)

Narrative

A low BNP might be helpful in ruling out CHF if the pre-test probability is equivocal. However, if pretest probability for CHF is low and BNP is above threshhold, BNP would not be helpful in ruling in CHF. In other words, very low BNP values in patients with low probability of heart failure and very high BNP values in patients with high probability of the disease can confirm clinical suspicions.

Caveats

Author

Khaled Hassan, MD and Shahriar Zehtabchi, MD

Published/Updated

April 8, 2013

What are Likelihood Ratios?

LR, pretest probability and posttest (or posterior) probability are daunting terms that describe simple concepts that we all intuitively understand.

Let's start with pretest probability: that's just a fancy term for my initial impression, before we perform whatever test it is that we're going to use.

For example, a patient with prior stents comes in sweating and clutching his chest in agony, I have a pretty high suspicion that he's having an MI – let's say, 60%. That is my pretest probability.

He immediately gets an ECG (known here as the "test") showing an obvious STEMI.

Now, I know there are some STEMI mimics, so I'm not quite 100%, but based on my experience I'm 99.5% sure that he's having an MI right now. This is my posttest probability - the new impression I have that the patient has the disease after we did our test.

And likelihood ration? That's just the name for the statistical tool that converted the pretest probability to the posttest probability - it's just a mathematical description of the strength of that test.

Using an online calculator, that means the LR+ that got me from 60% to 99.5% is 145, which is about as high an LR you can get (and the actual LR for an emergency physician who thinks an ECG shows an obvious STEMI).

(Thank you to Seth Trueger, MD for this explanation!)

Dyspnea Due to Heart Failure (Without Chronic Respiratory Disease)

Diagnostics and Likelihood Ratios, Explained

Positive Findings (Patient Has This)

Risk FactorsIncreased Disease Probability (Positive Likelihood Ratio)
History of Heart Failure5.8× (4.1-8.0)
History of Myocardial infarction3.1× (2.0-4.9)
History of Coronary artery disease1.8× (1.1-2.8)
History of Dyslipidemia1.7× (0.43-6.9)
History of Diabetes mellitus1.7× (1.0-2.7)
History of Hypertension1.4× (1.1-1.7)
History of Smoking0.84× (0.58-1.2)
History of COPD0.81× (0.60-1.1)
SymptomsIncreased Disease Probability (Positive Likelihood Ratio)
Paroxysmal nocturnal dyspnea2.6× (1.5-4.5)
Orthopnea2.2× (1.2-3.9)
Edema2.1× (0.92-5.0)
Dyspnea on exertion1.3× (1.2-1.4)
Fatigue and weight gain1.0× (0.74-1.4)
Cough0.93× (0.70-1.2)
SignsIncreased Disease Probability (Positive Likelihood Ratio)
3rd heart sound (ventricullar filling gallop)11× (4.9-25.0)
Abdominojugular reflux6.4× (0.81-51.0)
JVD5.1× (3.2-7.9)
Rales2.8× (1.9-4.1)
Any murmur2.6× (1.7-4.1)
Lower extremity edema2.3× (1.5-3.7)
Valsalva maneuver2.1× (1.0-4.2)
SBP <100 mmHg2.0× (0.60-6.6)
4th heart sound (atrial gallop)1.6× (0.47-5.5)
SBP >= 150 mmHg1.0× (0.69-1.6)
Wheezing0.52× (0.04-2.9)
OtherIncreased Disease Probability (Positive Likelihood Ratio)
Clinician's Gestalt4.4× (1.8-10.0)
CXR FindingsIncreased Disease Probability (Positive Likelihood Ratio)
Pulmonary venous congestion12.0× (6.8-21.0)
Interstitial edema12.0× (5.2-27.0)
Alveolar edema6.0× (2.2-16.0)
Cardiomegaly3.3× (2.4-4.7)
Pleural effusion(s)3.2× (2.4-4.3)
Any edema3.1× (0.60-16.0)
Pneumonia0.50× (0.29-0.87)
Hyperinflation0.38× (0.20-0.69)
EKG FindingsIncreased Disease Probability (Positive Likelihood Ratio)
Atrial fibrillation3.8× (1.7-8.8)
New T-wave changes3.0× (1.7-5.3)
Any abnormal finding2.2× (1.6-3.1)
ST elevation1.8× (0.80-4.0)
ST depression1.7× (0.97-2.9)
BNP LevelsIncreased Disease Probability (Positive Likelihood Ratio)
≥2504.6× (2.6-8.0)
≥2003.7× (2.6-5.4)
≥1503.1× (2.1-4.5)
Clinical judgement or BNP ≥100 pg/ml3.1× (2.8-3.5)
≥1002.7× (2.0-3.9)
≥501.7× (1.2-2.6)

Negative Findings (Patient Doesn't Have This)

Risk FactorsDecreased Disease Probability (Negative Likelihood Ratio)
History of Heart Failure0.45× (0.38-0.53)
History of Myocardial infarction0.69× (0.58-0.82)
History of Coronary artery disease0.68× (0.48-0.96)
History of Hypertension0.71× (0.55-0.93)
History of Dyslipidemia0.89× (0.69-1.1)
History of Diabetes mellitus0.86× (0.73-1.0)
History of Smoking1.4× (0.58-3.6)
History of COPD1.1× (0.95-1.4)
SymptomsDecreased Disease Probability (Negative Likelihood Ratio)
Dyspnea on exertion0.48× (0.35-0.67)
Edema0.64× (0.39-1.1)
Orthopnea0.65× (0.45-0.92)
Paroxysmal nocturnal dyspnea0.70× (0.54-0.91)
Fatigue and weight gain0.99× (0.85-1.1)
Cough1.0× (0.87-1.3)
SignsDecreased Disease Probability (Negative Likelihood Ratio)
Valsalva maneuver0.41× (0.17-1.0)
Rales0.51× (0.37-0.70)
Lower extremity edema0.64× (0.47-0.87)
JVD0.66× (0.57-0.77)
Abdominojugular reflux0.79× (0.62-1.0)
Any murmur0.81× (0.73-0.90)
3rd heart sound (ventricullar filling gallop)0.88× (0.83-0.94)
SBP <100 mmHg0.97× (0.91-1.0)
4th heart sound (atrial gallop)0.98× (0.93-1.0)
SBP ≥ 150 mmHg0.99× (0.84-1.2)
Ascities1.0× (0.99-1.1)
Wheezing1.3× (1.1-1.7)
OtherDecreased Disease Probability (Negative Likelihood Ratio)
Clinician's Gestalt0.45× (0.28-0.73)
CXR FindingsDecreased Disease Probability (Negative Likelihood Ratio)
Cardiomegaly0.33× (0.23-0.48)
Any edema0.38× (0.11-1.3)
Pulmonary venous congestion0.48× (0.28-0.83)
Interstitial edema0.68× (0.54-0.85)
Pleural effusion(s)0.81× (0.77-0.85)
Alveolar edema0.95× (0.93-0.97)
Pneumonia1.0× (1.0-1.1)
Hyperinflation1.1× (1.0-1.1)
EKG FindingsDecreased Disease Probability (Negative Likelihood Ratio)
Any abnormal finding0.64× (0.47-0.88)
Atrial fibrillation0.79× (0.65-0.96)
New T-wave changes0.83× 0.74-0.92)
ST elevation0.98× (0.94-1.0)
ST depression0.95× (0.90-1.0)
BNP LevelsDecreased Disease Probability (Negative Likelihood Ratio)
≥500.06× (0.03-0.12)
≥800.06× (0.03-0.13)
Clinical judgement or BNP ≥100 pg/ml0.09× (0.06-0.11)
≥1000.11× (0.07-0.16)
≥2000.11× (0.07-0.18)
≥2500.14× (0.06-0.33)
≥1500.15× (0.11-0.21)

Narrative

A low BNP might be helpful in ruling out CHF if the pre-test probability is equivocal. However, if pretest probability for CHF is low and BNP is above threshhold, BNP would not be helpful in ruling in CHF. In other words, very low BNP values in patients with low probability of heart failure and very high BNP values in patients with high probability of the disease can confirm clinical suspicions.

Caveats

Author

Khaled Hassan, MD and Shahriar Zehtabchi, MD

Published/Updated

What are Likelihood Ratios?

LR, pretest probability and posttest (or posterior) probability are daunting terms that describe simple concepts that we all intuitively understand.

Let's start with pretest probability: that's just a fancy term for my initial impression, before we perform whatever test it is that we're going to use.

For example, a patient with prior stents comes in sweating and clutching his chest in agony, I have a pretty high suspicion that he's having an MI – let's say, 60%. That is my pretest probability.

He immediately gets an ECG (known here as the "test") showing an obvious STEMI.

Now, I know there are some STEMI mimics, so I'm not quite 100%, but based on my experience I'm 99.5% sure that he's having an MI right now. This is my posttest probability - the new impression I have that the patient has the disease after we did our test.

And likelihood ration? That's just the name for the statistical tool that converted the pretest probability to the posttest probability - it's just a mathematical description of the strength of that test.

Using an online calculator, that means the LR+ that got me from 60% to 99.5% is 145, which is about as high an LR you can get (and the actual LR for an emergency physician who thinks an ECG shows an obvious STEMI).

(Thank you to Seth Trueger, MD for this explanation!)

Hemorrhagic Stroke

Diagnostics and Likelihood Ratios, Explained

Positive Findings (Patient Has This)

Risk FactorsIncreased Disease Probability (Positive Likelihood Ratio)
Age <=60y1.7× (1.4-1.9)
Alcohol consumption1.6× (1-2.5)
Male1.2× (1.1-1.3)
Hypertension1.1× (1.0-1.2)
Cigarette smoking0.79× (0.45-1.4)
Diabetes mellitus0.64× (0.43-0.95)
Prior stroke0.59× (0.17-2.0)
Hyperlipidemia0.48× (0.2-1.1)
Coronary artery disease0.44× (0.31-0.61)
Atrial fibrillation0.44× (0.25-0.78)
Peripheral artery disease0.41× (0.2-0.83)
Prior TIA0.34× (0.18-0.65)
SymptomsIncreased Disease Probability (Positive Likelihood Ratio)
Seizures w/ neurological deficit4.7× (1.6-14)
Vomiting3.0× (1.7-5.5)
Headache2.9× (1.7-4.8)
Loss of consciousness2.6× (1.6-4.2)
Acute onset deficit0.65 (0.52-0.81)
SignsIncreased Disease Probability (Positive Likelihood Ratio)
Kernig's or Brudzinski's or both8.2× (0.44-150)
LOC: coma6.2× (3.2-12)
Neck stiffness5.0× (1.9-12.8)
Diastolic BP >110 mmHg4.3× (1.4-14)
LOC: drowsy2.0× (1.0-3.9)
Plantar response: both extensor1.8× (0.99-3.4)
Plantar response: single extensor1× (0.87-1.2)
Hemiparesis0.96× (0.9-1.0)
Plantar response: both flexor0.45× (0.25-0.81)
LOC: alert0.35× (0.24-0.5)
Cervical bruit0.12× (0.03-0.47)
Labs/StudiesIncreased Disease Probability (Positive Likelihood Ratio)
Xanthochromia in CSF15× (7.7-29)
Atrial fibrillation on ECG0.19× (0.06-0.59)
OtherIncreased Disease Probability (Positive Likelihood Ratio)
Clinician's Gestalt6.2× (4.2-9.3)

Negative Findings (Patient Doesn't Have This)

Risk FactorsDecreased Disease Probability (Negative Likelihood Ratio)
Age <=60y0.71× (0.63-0.82)
Alcohol consumption0.75× (0.51-1.1)
Male0.85× (0.77-0.94)
Hypertension0.88× (0.77-1.01)
Diabetes mellitus1.1× (1.0-1.2)
Prior stroke1.1× (0.88-1.4)
Hyperlipidemia1.1× (1.1-1.1)
Coronary artery disease1.1× (1.0-1.3)
Atrial fibrillation1.1× (1.05-1.1)
Peripheral artery disease1.1× (1.0-1.1)
Prior TIA1.2× (1.1-1.3)
Cigarette smoking1.2× (0.79-1.8)
SymptomsDecreased Disease Probability (Negative Likelihood Ratio)
Vomiting0.73× (0.59-0.91)
Headache0.66× (0.56-0.77)
Loss of consciousness0.65× (0.52-0.82)
Seizures w/ neurological deficit0.93× (0.9-0.96)
Acute onset deficit1.7× (1.4-2.1)
SignsDecreased Disease Probability (Negative Likelihood Ratio)
Diastolic BP >110 mmHg0.59× (0.93-0.89)
Neck stiffness0.83× (0.75-0.92)
Kernig's or Brudzinski's or both0.87× (0.73-1.0)
Hemiparesis1.1× (1.0-1.2)
Cervical bruit1.1× (1.0-1.1)
Labs/StudiesDecreased Disease Probability (Negative Likelihood Ratio)
Xanthochromia in CSF0.31× (0.19-0.49)
Atrial fibrillation on ECG1.2× (1.0-1.5)
OtherDecreased Disease Probability (Negative Likelihood Ratio)
Clinician's Gestalt0.28× (0.20-0.39)

Narrative

Clinical exam findings help form an impression of hemorrhage vs ischemic stroke but head CT (when possible) is the best test to rapidly distinguish the stroke subtype.

Caveats


  • Authors' RCE study selection criteria excluded those that enrolled patients with subarachnoid hemorrhage (which will present with some similar findings of hemorrhagic stroke).

  • Diagnostic value of Xanthochromia was only reported by one study (Britton et al., Acta Med Scand 1983), which determined xanthochromia by both visual assessment and spectophotometry for comparison. LRs calculated by each technique were almost the same.

Author

Khaled Hassan, MD and Shahriar Zehtabchi, MD

Published/Updated

What are Likelihood Ratios?

LR, pretest probability and posttest (or posterior) probability are daunting terms that describe simple concepts that we all intuitively understand.

Let's start with pretest probability: that's just a fancy term for my initial impression, before we perform whatever test it is that we're going to use.

For example, a patient with prior stents comes in sweating and clutching his chest in agony, I have a pretty high suspicion that he's having an MI – let's say, 60%. That is my pretest probability.

He immediately gets an ECG (known here as the "test") showing an obvious STEMI.

Now, I know there are some STEMI mimics, so I'm not quite 100%, but based on my experience I'm 99.5% sure that he's having an MI right now. This is my posttest probability - the new impression I have that the patient has the disease after we did our test.

And likelihood ration? That's just the name for the statistical tool that converted the pretest probability to the posttest probability - it's just a mathematical description of the strength of that test.

Using an online calculator, that means the LR+ that got me from 60% to 99.5% is 145, which is about as high an LR you can get (and the actual LR for an emergency physician who thinks an ECG shows an obvious STEMI).

(Thank you to Seth Trueger, MD for this explanation!)

Spinal Stenosis in the Elderly

Diagnostics and Likelihood Ratios, Explained

Positive Findings (Patient Has This)

SymptomsIncreased Disease Probability (Positive Likelihood Ratio)
No pain when seated7.4× (1.9-30)
Burning sensation around the buttocks, Intermittent priapism associated with walking, or both7.2× (1.6-32)
Urinary disturbance6.9× (2.7-17)
Improvement when bending forward6.4× (4.1-9.9)
Bilateral buttock or leg6.3× (3.1-13)
Neurogenic claudication3.7× (2.9-4.8)
Numbness of perineal region3.7× (1.0-13)
Improve when seated3.3× (1.4-7.7)
Exacerbation when standing up2.3× (1.8-2.8)
Bilateral plantar numbness2.2× (1.4-3.2)
Treatment for symptoms needs to be repeated every year2.0× (1.5-2.8)
Orthopedic disease2.0× (1.2-3.5)
Pain below buttocks1.4× (1.0-1.8)
Exacerbated while standing up1.2× (1.1-1.3)
Wake up to urinate at night1.2× (1.1-1.3)
Thigh1.1× (1.0-1.2)
Gluteal0.88× (0.79-0.98)
Signs on Physical ExamIncreased Disease Probability (Positive Likelihood Ratio)
Wide-based gait13× (1.9-95)
Abnormal Romberg test result4.2× (1.4-13)
Vibration deficit2.8× (1.3-6.2)
Pinprick deficit2.5× (1.1-5.5)
Age >65 (vs ≤65)2.5× (1.4-4.2)
Weakness2.1× (1.0-4.4)
Absent Achilles reflex2.1 (1.0-4.4)
No pain with flexion1.4 (1.0-2.0)
Symptoms worsened with bending forward0.48 (0.34-0.66)

Negative Findings (Patient Doesn't Have This)

SymptomsDecreased Disease Probability (Negative Likelihood Ratio)
Neurogenic claudication0.23× (0.17-0.31)
Pain below buttocks0.34× (0.13-0.88)
Thigh0.36× (0.12-1.1)
Exacerbated while standing up0.38× (0.21-0.69)
Exacerbation when standing up0.46× (0.37-0.56)
Wake up to urinate at night0.50× (0.33-0.78)
Improvement when bending forward0.52× (0.46-0.60)
Bilateral buttock or leg0.54× (0.43-0.68)
No pain when seated0.57× (0.43-0.76)
Improve when seated0.58× (0.41-0.81)
Treatment for symptoms needs to be repeated every year0.75× (0.65-0.86)
Bilateral plantar numbness0.84× (0.76-0.92)
Urinary disturbance0.88× (0.83-0.93)
Orthopedic disease0.90× (0.83-0.98)
Burning sensation around the buttocks, Intermittent priapism associated with walking, or both0.95× (0.92-0.98)
Numbness of perineal region0.97× (0.94-1.0)
Gluteal3.3× (1.2-8.8)
Signs on Physical ExamDecreased Disease Probability (Negative Likelihood Ratio)
Pain below buttocks0.34× (0.13-0.88)
Age >65 (vs ≤65)0.34× (0.19-0.61)
Thigh0.36× (0.12-1.1)
Exacerbated while standing up0.38× (0.21-0.69)
Exacerbation when standing up0.46× (0.37-0.56)
No pain with flexion0.48× (0.24-0.96)
Wake up to urinate at night0.50× (0.33-0.78)
Improvement when bending forward0.52 (0.46-0.60)
Bilateral buttock or leg0.54 (0.43-0.68)
Vibration deficit0.57 (0.40-0.82)
No pain when seated0.57 (0.43-0.76)
Improve when seated0.58 (0.41-0.81)
Wide-based gait0.60 (0.46-0.78)
Pinprick deficit0.66 (0.48-0.91)
Abnormal Romberg test result0.67 (0.51-0.87)
Weakness0.69 (0.49-0.96)
Absent Achilles reflex0.69 (0.49-0.96)
Treatment for symptoms needs to be repeated every year0.75 (0.65-0.86)
Bilateral plantar numbness0.84 (0.76-0.92)
Urinary disturbance0.88 (0.83-0.93)
Orthopedic disease0.90 (0.83-0.98)
Burning sensation around the buttocks, Intermittent priapism associated with walking, or both0.95 (0.92-0.98)
Numbness of perineal region0.97 (0.94-1.0)
Symptoms induced by having patients bend forward1.3 (1.2-1.5)
Gluteal3.3 (1.2-8.8)

Narrative

Outcome was clinical syndrome of Lumbar Spinal Stenosis (LSS).

Caveats


  • A characteristic clinical presentation, including neurogenic claudication, radicular pain, or both; and

  • Radiographic or anatomic LSS

Author

Rodrigo Kong, MD and Shahriar Zehtabchi, MD

Published/Updated

February 25, 2013

What are Likelihood Ratios?

LR, pretest probability and posttest (or posterior) probability are daunting terms that describe simple concepts that we all intuitively understand.

Let's start with pretest probability: that's just a fancy term for my initial impression, before we perform whatever test it is that we're going to use.

For example, a patient with prior stents comes in sweating and clutching his chest in agony, I have a pretty high suspicion that he's having an MI – let's say, 60%. That is my pretest probability.

He immediately gets an ECG (known here as the "test") showing an obvious STEMI.

Now, I know there are some STEMI mimics, so I'm not quite 100%, but based on my experience I'm 99.5% sure that he's having an MI right now. This is my posttest probability - the new impression I have that the patient has the disease after we did our test.

And likelihood ration? That's just the name for the statistical tool that converted the pretest probability to the posttest probability - it's just a mathematical description of the strength of that test.

Using an online calculator, that means the LR+ that got me from 60% to 99.5% is 145, which is about as high an LR you can get (and the actual LR for an emergency physician who thinks an ECG shows an obvious STEMI).

(Thank you to Seth Trueger, MD for this explanation!)

Streptococal Pharyngitis

Diagnostics and Likelihood Ratios, Explained

Positive Findings (Patient Has This)

SymptomsIncreased Disease Probability (Positive Likelihood Ratio)
Strep exposure in the past 2 weeks1.9× (1.3-2.8)
Myalgias1.4× (1.1-1.7)
No cough1.1-1.7
History of sore throat1.0-1.1
Reported fever0.97-2.6
Headache0.81-2.6
Nausea0.76-3.1
Duration <3d0.72-3.5
Signs on Physical ExamIncreased Disease Probability (Positive Likelihood Ratio)
Tonsillar exudates3.4× (1.8-6.0)
Pharyngeal exudates2.1× (1.4-3.1)
Tonsillar or pharyngeal exudates1.8× (1.5-2.3)
Any Exudates1.5-2.6
Tonsillar swelling/enlargement1.4-3.1
Palatine petechiae1.4× (0.48-3.1)
Ant. Cervical lymph node tenderness1.2-1.9
Measured temp >37.8 C1.1-3.0
Male sex0.87× (0.72-1.05)
No coryza0.86-1.6
Measured temp >=38.3 °C0.68-3.9
Pharynx injected0.66-1.63
Ant. Cervical lymph node swollen/enlarged0.47-2.9
Rash0.06-35
Centor Criteria Points (Ignoring Age Modification)Increased Disease Probability (Positive Likelihood Ratio)
4 Points6.3×
3 Points2.1×
2 Points0.75
0 Points0.16
1 Points0.3

Negative Findings (Patient Doesn't Have This)

SymptomsDecreased Disease Probability (Negative Likelihood Ratio)
Duration <3d0.15-2.2
Reported fever0.32-1.0
No cough0.53-0.89
Headache0.55-1.1
History of sore throat0.55-1.2
Nausea0.91× (0.86-0.97)
Strep exposure previous 2 wk0.92× (0.86-0.99)
Myalgias0.93× (0.86-1.0)
Signs on Physical ExamDecreased Disease Probability (Negative Likelihood Ratio)
Pharynx injected0.18-6.42
Measured temp >37.8 C0.27-0.94
No coryza0.51-1.4
Measured temp >=38.3 C0.54-1.3
Ant. Cervical lymph node swollen/enlarged0.58-0.92
Ant. Cervical lymph node tenderness0.60× (0.49-0.71)
Tonsillar swelling/enlargement0.63× (0.56-0.72)
Any Exudates0.66-0.94
Tonsillar exudates0.72× (0.60-0.88)
Tonsillar or pharyngeal exudates0.74× (0.66-0.82)
Pharyngeal exudates0.90× (0.75-1.1)
Rash0.90-1.1
Palatine petechiae0.98× (0.92-1.1)
Male sex1.1× (0.93-1.2)

Narrative

The Bottom Line: Clinician judgment performs better when looking at the constellation of findings to determine diagnosis.

Caveats


  • Authors identify Centor scoring method as being validated with an ROC of 0.79 but this is for an adult population that is not where Strep throat incidence peaks (children 24-36% vs 5-24% adults).

  • There are 9 clinical prediction rules for strep throath. Not all references provide LRs or ROCs and most have limitations for their use).

Author

Khaled Hassan, MD and Shahriar Zehtabchi, MD

Published/Updated

What are Likelihood Ratios?

LR, pretest probability and posttest (or posterior) probability are daunting terms that describe simple concepts that we all intuitively understand.

Let's start with pretest probability: that's just a fancy term for my initial impression, before we perform whatever test it is that we're going to use.

For example, a patient with prior stents comes in sweating and clutching his chest in agony, I have a pretty high suspicion that he's having an MI – let's say, 60%. That is my pretest probability.

He immediately gets an ECG (known here as the "test") showing an obvious STEMI.

Now, I know there are some STEMI mimics, so I'm not quite 100%, but based on my experience I'm 99.5% sure that he's having an MI right now. This is my posttest probability - the new impression I have that the patient has the disease after we did our test.

And likelihood ration? That's just the name for the statistical tool that converted the pretest probability to the posttest probability - it's just a mathematical description of the strength of that test.

Using an online calculator, that means the LR+ that got me from 60% to 99.5% is 145, which is about as high an LR you can get (and the actual LR for an emergency physician who thinks an ECG shows an obvious STEMI).

(Thank you to Seth Trueger, MD for this explanation!)

Myasthenia Gravis

Diagnostics and Likelihood Ratios, Explained

Positive Findings (Patient Has This)

SymptomsIncreased Disease Probability (Positive Likelihood Ratio)
Food in mouth after swallowing13.0 (0.85-212.0)
Unintelligible speech after prolonged speaking4.5× (1.2-17.0)
Signs on Physical ExamIncreased Disease Probability (Positive Likelihood Ratio)
Sleep Test53.0× (3.4-832.0)
Ice Test24.0× (8.5-67.0)
Rest Test16.0 (0.98-261.0)
Anticholinesterase Test15.0× (7.5-31.0)
Quiver eye movements4.1 (0.22-75.0)

Negative Findings (Patient Doesn't Have This)

SymptomsDecreased Disease Probability (Negative Likelihood Ratio)
Unintelligible speech after prolonged speaking0.61 (0.46-0.8)
Food in mouth after swallowing0.70 (0.58-0.84)
Signs on Physical ExamDecreased Disease Probability (Negative Likelihood Ratio)
Sleep Test0.01× (0.00-0.16)
Anticholinesterase Test0.11× (0.06-0.21)
Ice Test0.16× (0.09-0.27)
Rest Test0.52 (0.29-0.95)
Quiver eye movements0.82 (0.57-1.2)
Peek Sign0.88 (0.76-1.0)

Narrative

Myasthenia gravis is the most common neuromuscular transmission disease, with prevalence rates as high as 20.4 per 100,000.1 The disorder is associated with severe morbidity and mortality, however is very amenable to treatment.

The review summarized here examines the accuracy and utility of history, physical examination, and bedside testing for diagnosing myasthenia. 15 studies met inclusion criteria for the review, including 12 prospective studies (n=896) evaluating the performance of individual elements of the history, physical exam, and bedside clinical tests.

There are several clinical tests that can be performed at the bedside to help in the evaluation for myasthenia gravis. The ice test involves placing ice over the more ptotic eye, while the rest test involves placing a small amount of cotton over a closed lid to rest the eye. Either is placed for 2 minutes, with a positive result being a 2 mm improvement in ptosis. The sleep test is performed by having the patient close their eyes in a dark room for 30 minutes, and then re-evaluating for improvement in ptosis. The peek test is performed by having the patient try to gently close their eyes for 30 seconds. A positive result is the so-called “peeking” of the sclera due to the patient’s inability to maintain eye closure for 30 seconds.

Unintelligible speech after a prolonged period (LR+ 4.5 ) and a positive sleep test (LR+ 53), appear to be useful findings that increase the likelihood of the diagnosis. The ice test (LR- 0.16) and sleep test (LR- 0.01) both appear to significantly decrease the likelihood of the diagnosis when negative. Acetylcholinesterase testing (typically with edrophonium) appears to be among the most useful bedside clinical tests with a LR+ of 15 and a LR- of 0.1.

Caveats

The data presented here are based on a small number of studies, each with a small number of subjects. The sleep test, for instance, generated the most extreme (i.e. theoretically useful) LR’s, but comes from a single study of 68 subjects, 42 of whom were found to have MG. The high prevalence suggests verification and referral bias in these studies, as the LR findings were generated from samples in which myasthenia was common—a characteristic that may be quite different in many settings where the tests will be used in common clinical practice. Furthermore, because several of these tests depend on the skill and experience of the examiner, and because the inter-observer reliability of the findings was not assessed in any study, the reliability of these findings remains unknown.

Finally, not every patient with weakness can be evaluated with sophisticated examinations like acetylcholinesterase testing or single fiber electromyography, or by an experienced neurologist. Therefore while this is the best published information available on diagnostic findings for myasthenia gravis, the generalizability of these data is difficult to predict.

Author

Rodrigo Kong, MD, and Maxwell Morrison

Published/Updated

February 16, 2013

What are Likelihood Ratios?

LR, pretest probability and posttest (or posterior) probability are daunting terms that describe simple concepts that we all intuitively understand.

Let's start with pretest probability: that's just a fancy term for my initial impression, before we perform whatever test it is that we're going to use.

For example, a patient with prior stents comes in sweating and clutching his chest in agony, I have a pretty high suspicion that he's having an MI – let's say, 60%. That is my pretest probability.

He immediately gets an ECG (known here as the "test") showing an obvious STEMI.

Now, I know there are some STEMI mimics, so I'm not quite 100%, but based on my experience I'm 99.5% sure that he's having an MI right now. This is my posttest probability - the new impression I have that the patient has the disease after we did our test.

And likelihood ration? That's just the name for the statistical tool that converted the pretest probability to the posttest probability - it's just a mathematical description of the strength of that test.

Using an online calculator, that means the LR+ that got me from 60% to 99.5% is 145, which is about as high an LR you can get (and the actual LR for an emergency physician who thinks an ECG shows an obvious STEMI).

(Thank you to Seth Trueger, MD for this explanation!)

Migraine

Diagnostics and Likelihood Ratios, Explained

Positive Findings (Patient Has This)

CriteriaProbability of Disease (Positive Likelihood Ratio)
≥ 4 POUNDing Criteria5.8×

Negative Findings (Patient Doesn't Have This)

None identified.

Narrative


  • Outcome: Migraine, defined by International Headache Society criteria (applied by a neurologist)

  • Level I Evidence Study: Independent, blinded comparisons of components of the clinical examination with a gold standard among 100 or more consecutive patients with headache.

Caveats

Criteria developed based on a single study with Level I Evidence

Author

Rodrigo Kong, MD and Shahriar Zehtabchi, MD

Published/Updated

February 15, 2013

What are Likelihood Ratios?

LR, pretest probability and posttest (or posterior) probability are daunting terms that describe simple concepts that we all intuitively understand.

Let's start with pretest probability: that's just a fancy term for my initial impression, before we perform whatever test it is that we're going to use.

For example, a patient with prior stents comes in sweating and clutching his chest in agony, I have a pretty high suspicion that he's having an MI – let's say, 60%. That is my pretest probability.

He immediately gets an ECG (known here as the "test") showing an obvious STEMI.

Now, I know there are some STEMI mimics, so I'm not quite 100%, but based on my experience I'm 99.5% sure that he's having an MI right now. This is my posttest probability - the new impression I have that the patient has the disease after we did our test.

And likelihood ration? That's just the name for the statistical tool that converted the pretest probability to the posttest probability - it's just a mathematical description of the strength of that test.

Using an online calculator, that means the LR+ that got me from 60% to 99.5% is 145, which is about as high an LR you can get (and the actual LR for an emergency physician who thinks an ECG shows an obvious STEMI).

(Thank you to Seth Trueger, MD for this explanation!)

Pertussis (Whooping Cough)

Diagnostics and Likelihood Ratios, Explained

Positive Findings (Patient Has This)

SymptomsIncreased Disease Probability (Positive Likelihood Ratio)
Posttussive Emesis1.8× (1.4-2.2)
Paroxysmal Cough1.1× (1.1-1.2)
SignsIncreased Disease Probability (Positive Likelihood Ratio)
Inspiratory Whoop1.9× (1.4-2.6)

Negative Findings (Patient Doesn't Have This)

Sign / SymptomDecreased Disease Probability (Negative Likelihood Ratio)
Paroxysmal Cough0.52× (0.27-1.0)
Posttussive emesis0.58× (0.44-0.77)
SignsDecreased Disease Probability (Negative Likelihood Ratio)
Inspiratory Whoop0.78× (0.66-0.93)

Narrative


  • Outcome: Pertussis

  • RCE Conclusion: In a non-outbreak setting, data to determine the diagnostic usefulness of symptoms classically associated with pertussis are limited and of relatively weak quality. The presence or absence of posttussive emesis or an inspiratory whoop modestly change the likelihood of pertussis; therefore, clinicians must use their overall clinical impression to decide about additional testing or empirical treatment.

Caveats


  • None of the above LR's would dramatically change pretest probability

  • All patients in the 3 included studies were enrolled in outpatient clinics

Author

Rodrigo Kong, MD and Shahriar Zehtabchi, MD

Published/Updated

What are Likelihood Ratios?

LR, pretest probability and posttest (or posterior) probability are daunting terms that describe simple concepts that we all intuitively understand.

Let's start with pretest probability: that's just a fancy term for my initial impression, before we perform whatever test it is that we're going to use.

For example, a patient with prior stents comes in sweating and clutching his chest in agony, I have a pretty high suspicion that he's having an MI – let's say, 60%. That is my pretest probability.

He immediately gets an ECG (known here as the "test") showing an obvious STEMI.

Now, I know there are some STEMI mimics, so I'm not quite 100%, but based on my experience I'm 99.5% sure that he's having an MI right now. This is my posttest probability - the new impression I have that the patient has the disease after we did our test.

And likelihood ration? That's just the name for the statistical tool that converted the pretest probability to the posttest probability - it's just a mathematical description of the strength of that test.

Using an online calculator, that means the LR+ that got me from 60% to 99.5% is 145, which is about as high an LR you can get (and the actual LR for an emergency physician who thinks an ECG shows an obvious STEMI).

(Thank you to Seth Trueger, MD for this explanation!)

Osteomyelitis in Diabetic Patients

Diagnostics and Likelihood Ratios, Explained

Positive Findings (Patient Has This)

Signs/FindingsIncreased Disease Probability (Positive Likelihood Ratio)
Bone Exposure9.2× (0.57-146)
Ulcer Area > 2cm²7.2× (1.1-4.9)
Positive probe-to-bone test6.4× (3.6-11)
Clinical Gestalt5.5× (1.8-17)
Ulcer inflammation (erythema, swelling, purulence)1.5× (0.51-4.7)
Lab and Study FindingsIncreased Disease Probability (Positive Likelihood Ratio)
ESR > 70 mm/h11× (1.6-79.0)
Abnormal findings [indicating osteomyelitis] on plain Radiograph2.3× (1.56-3.3)
Swab Culture1× (0.65-1.5)

Negative Findings (Patient Doesn't Have This)

Signs/FindingsIncreased Disease Probability (Negative Likelihood Ratio)
Positive probe-to-bone test0.39× (0.20-0.76)
Ulcer Area > 2cm²0.48× (0.31-0.76)
Clinical Gestalt0.54× (0.30-0.97)
Bone Exposure0.70× (0.53-0.92)
Ulcer inflammation (erythema, swelling, purulence)0.84× (0.56-1.3)
Lab and Study FindingsDecreased Disease Probability (Negative Likelihood Ratio)
ESR > 70 mm/h0.34× (0.06-1.9)
Abnormal findings [indicating osteomyelitis] on plain Radiograph0.63× (0.51-0.78)
Swab Culture1× (0.08-13)

Narrative


  • Outcome: osteomyelitis

  • Conclusion: "No single historical feature or physical examination reliably
    excludes osteomyelitis."

Caveats

Studies were retro- or prospective; mostly level III or IV evidence; most patients had foot ulcers, or infection, or suspected osteomyelits; all tests used gold standard (bone biopsy w histopathology diagnosis, bone culture or both) for comparison; In- or out-patient location was not clear in most cases; likely strong selection bias.

Not all criteria for evaluation of osteomyelitis on radiographic examination were reported in each study.

Author

Rodrigo Kong, MD and Shahriar Zehtabchi, MD

Published/Updated

What are Likelihood Ratios?

LR, pretest probability and posttest (or posterior) probability are daunting terms that describe simple concepts that we all intuitively understand.

Let's start with pretest probability: that's just a fancy term for my initial impression, before we perform whatever test it is that we're going to use.

For example, a patient with prior stents comes in sweating and clutching his chest in agony, I have a pretty high suspicion that he's having an MI – let's say, 60%. That is my pretest probability.

He immediately gets an ECG (known here as the "test") showing an obvious STEMI.

Now, I know there are some STEMI mimics, so I'm not quite 100%, but based on my experience I'm 99.5% sure that he's having an MI right now. This is my posttest probability - the new impression I have that the patient has the disease after we did our test.

And likelihood ration? That's just the name for the statistical tool that converted the pretest probability to the posttest probability - it's just a mathematical description of the strength of that test.

Using an online calculator, that means the LR+ that got me from 60% to 99.5% is 145, which is about as high an LR you can get (and the actual LR for an emergency physician who thinks an ECG shows an obvious STEMI).

(Thank you to Seth Trueger, MD for this explanation!)

Carpal Tunnel Syndrome

Diagnostics and Likelihood Ratios, Explained

Positive Findings (Patient Has This)

Finding (Sign/Symptom/Lab/Study)Number Needed to Diagnose
(Positive Likelihood Ratio)
Flick Sign  21.4×
Closed Fist Sign  7.3×
Hypalgesia 3.1×
Square Hand Sign  2.7×
Classical or Probable (Hand diagram, JAMA 2000)2.4×
Weak Thumb Abduction 1.8×
Thenar Atrophy 1.6×
Abnormal Vibration 1.6×
Abnormal Monofilament 1.5×
Bilateral Symptoms 1.4×
Tinnel Sign 1.4×
Phalen Sign 1.3×
Age > 40 years 1.3×
2 point discrimination 1.3×
Nocturnal Paresthesia 1.2×
Pressure Provocation Test 1.0×
Tourniquet Test 1.0×

Negative Findings (Patient Doesn't Have This)

Finding (Sign/Symptom/Lab/Study)Number Needed to Diagnose
(Negative Likelihood Ratio)
Flick Sign  0.1×
Closed Fist Sign  0.4×
Classical or Probable (Hand diagram, JAMA 2000) 0.5×
Age > 40 years 0.5×
Weak Thumb Abduction 0.5×
Square Hand Sign  0.5×
Bilateral Symptoms 0.7×
Nocturnal Paresthesia 0.7×
Abnormal Monofilament 0.7×
Hypalgesia 0.7×
Phalen Sign 0.7×
Abnormal Vibration 0.8×
Tinnel Sign 0.8×
Tourniquet Test 1.0×
Thenar Atrophy 1.0×
2 point discrimination 1.0×
Pressure Provocation Test 1.0×

Narrative

This review addresses the accuracy of the history and physical examination in diagnosing CTS, as confirmed by electrodiagnostic studies.
The data are derived from symptomatic patients presenting to an orthopedic surgeon, physical therapist, or an electrodiagnostic laboratory.
There are no data addressing the value of physical diagnosis in patients presenting to a primary care physician/ ED with symptoms suggestive of CTS.

Caveats

Author

Rodrigo Kong, MD and Shahriar Zehtabchi, MD

Published/Updated

January 9, 2013

What are Likelihood Ratios?

LR, pretest probability and posttest (or posterior) probability are daunting terms that describe simple concepts that we all intuitively understand.

Let's start with pretest probability: that's just a fancy term for my initial impression, before we perform whatever test it is that we're going to use.

For example, a patient with prior stents comes in sweating and clutching his chest in agony, I have a pretty high suspicion that he's having an MI – let's say, 60%. That is my pretest probability.

He immediately gets an ECG (known here as the "test") showing an obvious STEMI.

Now, I know there are some STEMI mimics, so I'm not quite 100%, but based on my experience I'm 99.5% sure that he's having an MI right now. This is my posttest probability - the new impression I have that the patient has the disease after we did our test.

And likelihood ration? That's just the name for the statistical tool that converted the pretest probability to the posttest probability - it's just a mathematical description of the strength of that test.

Using an online calculator, that means the LR+ that got me from 60% to 99.5% is 145, which is about as high an LR you can get (and the actual LR for an emergency physician who thinks an ECG shows an obvious STEMI).

(Thank you to Seth Trueger, MD for this explanation!)

Malaria in Returning Travelers

Diagnostics and Likelihood Ratios, Explained

Positive Findings (Patient Has This)

Finding (Sign/Symptom/Lab/Study)Number Needed to Diagnose
(Positive Likelihood Ratio)
Hyperbilirubinemia 7.3×
Splenomegaly 6.5×
Thrombocytopenia 5.6×
Fever 5.1×
Jaundice/icterus 4.5×
Pallor 2.8×
Hepatomegaly 2.4×
Vomiting 2.0×
Headache 1.8×
Chills/rigors 1.7×
Nausea 1.3×
Diarrhea 0.6×
Dyspnea 0.11×
Cough 0.04×

Negative Findings (Patient Doesn't Have This)

Finding (Sign/Symptom/Lab/Study)Number Needed to Diagnose
(Negative Likelihood Ratio)
Fever 0.12×
Thrombocytopenia 0.32×
Headache 0.40×
Chills/rigors 0.47×
Hyperbilirubinemia 0.65×
Splenomegaly 0.79×
Pallor 0.80×
Vomiting 0.86×
Nausea 0.88×
Jaundice/icterus 0.91×
Hepatomegaly 0.95×
Dyspnea 1.08×
Diarrhea 1.1×
Cough 1.3×

Narrative

  • Outcome: Malaria
  • There were 14 studies conducted in malaria endemic areas, with adults, children, or adults and children populations. In 7 of 9 studies that species was clearly reported, P. falciparum predominated.
  • Most of the Pediatric studies were conducted at sub Saharan Africa. Other areas included India and Thailand.
  • Adult studies were mostly conducted at India and Tanzania. Children and Adult combined studies were conducted in Uganda, Zimbabwe, and the Philippines.
  • There were 7 studied investigating travelers returning from endemic areas. In 5 of 7 studies P. Falciparum predominated.

Caveats

In returning travelers, the adequacy of malaria prophylaxis was not reported in a uniform fashion.

Author

Rodrigo Kong, MD

Published/Updated

March 15, 2012

What are Likelihood Ratios?

LR, pretest probability and posttest (or posterior) probability are daunting terms that describe simple concepts that we all intuitively understand.

Let's start with pretest probability: that's just a fancy term for my initial impression, before we perform whatever test it is that we're going to use.

For example, a patient with prior stents comes in sweating and clutching his chest in agony, I have a pretty high suspicion that he's having an MI – let's say, 60%. That is my pretest probability.

He immediately gets an ECG (known here as the "test") showing an obvious STEMI.

Now, I know there are some STEMI mimics, so I'm not quite 100%, but based on my experience I'm 99.5% sure that he's having an MI right now. This is my posttest probability - the new impression I have that the patient has the disease after we did our test.

And likelihood ration? That's just the name for the statistical tool that converted the pretest probability to the posttest probability - it's just a mathematical description of the strength of that test.

Using an online calculator, that means the LR+ that got me from 60% to 99.5% is 145, which is about as high an LR you can get (and the actual LR for an emergency physician who thinks an ECG shows an obvious STEMI).

(Thank you to Seth Trueger, MD for this explanation!)

Deep Venous Thrombosis (DVT)

Diagnostics and Likelihood Ratios, Explained

Positive Findings (Patient Has This)

Finding (Sign/Symptom/Lab/Study)Number Needed to Diagnose
(Positive Likelihood Ratio)
High Sensitivity D-Dimer with Low Pre-Test Probability2.4×
High Sensitivity D-Dimer with Mod Pre-Test Probability1.7×
High Sensitivity D-Dimer with High Pre-Test Probability1.5×

Negative Findings (Patient Doesn't Have This)

Finding (Sign/Symptom/Lab/Study)Number Needed to Diagnose
(Negative Likelihood Ratio)
High Sensitivity D-Dimer with Mod Pre-Test Probability0.05×
High Sensitivity D-Dimer with High Pre-Test Probability0.07×
High Sensitivity D-Dimer with Low Pre-Test Probability0.10×

Narrative

  • All LR's above are composite
  • Study Selection: Studies were selected for inclusion in the systematic review if they enrolled consecutive, unselected outpatients with suspected DVT and applied clinical prediction rules before D-dimer testing or diagnostic imaging.
  • Wells' criteria used as clinical prediction rule
  • Diagnostic accuracy for DVT improves when clinical probability is estimated before diagnostic tests.
  • When combined with a negative mod sensitivity D-dimer result, diagnostic imaging and anticoagulant therapy can be safely withheld for patients with a low clinical probability estimate since the negative LR (0.20; 95% CI, 0.12-0.31) is such that the probability after testing for DVT is less than 1%.
  • When combined with a negative high sensitivity D-dimer result, diagnostic imaging and anticoagulant therapy can be safely withheld in patients with a low (LR, 0.10; 95% CI, 0.03-0.37) or moderate clinical probability estimate (LR, 0.05; 95% CI, 0.01- 0.21) because they create a probability estimate after testing for DVT of less than 1%.

Caveats

Author

Rodrigo Kong, MD

Published/Updated

What are Likelihood Ratios?

LR, pretest probability and posttest (or posterior) probability are daunting terms that describe simple concepts that we all intuitively understand.

Let's start with pretest probability: that's just a fancy term for my initial impression, before we perform whatever test it is that we're going to use.

For example, a patient with prior stents comes in sweating and clutching his chest in agony, I have a pretty high suspicion that he's having an MI – let's say, 60%. That is my pretest probability.

He immediately gets an ECG (known here as the "test") showing an obvious STEMI.

Now, I know there are some STEMI mimics, so I'm not quite 100%, but based on my experience I'm 99.5% sure that he's having an MI right now. This is my posttest probability - the new impression I have that the patient has the disease after we did our test.

And likelihood ration? That's just the name for the statistical tool that converted the pretest probability to the posttest probability - it's just a mathematical description of the strength of that test.

Using an online calculator, that means the LR+ that got me from 60% to 99.5% is 145, which is about as high an LR you can get (and the actual LR for an emergency physician who thinks an ECG shows an obvious STEMI).

(Thank you to Seth Trueger, MD for this explanation!)

Hypovolemia

Diagnostics and Likelihood Ratios, Explained

Positive Findings (Patient Has This)

Finding (Sign/Symptom/Lab/Study)Number Needed to Diagnose
(Positive Likelihood Ratio)
Capillary refill time6.9×
Sunken eyes3.4×
Speech not clear or expressive3.1×
Dry axilla2.8×
Upper or Lower extremity weakness2.3×
Dry tongue 2.1×
Confusion2.1×
Dry mouth and nose m.m.2.0×
Tongue longitutinal furrows2.0×
Pulse change > 30bpm1.7×
Postural hypotension (SVP decr. > 20 mmHg)1.5×

Negative Findings (Patient Doesn't Have This)

Finding (Sign/Symptom/Lab/Study)Number Needed to Diagnose
(Negative Likelihood Ratio)
Dry mouth and nose m.m.0.3×
Tongue longitutinal furrows0.3×
Sunken eyes0.5×
Speech not clear or expressive0.5×
Dry tongue0.6×
Dry axilla0.6×
Confusion0.6×
Capillary refill time0.7×
Upper or Lower extremity weakness0.7×
Pulse change > 30bpm0.8×
Postural hypotension (SVP decr. > 20 mmHg)0.9×

Narrative

Caveats

  • Determination of some of the physicial findings such as axillary moisture and capillary refill are very operator dependent and have a poor inter-rater reliability (in case of 'dry axilla" the kappa is 0.5)

  • Vital signs or postural changes have poor sensitivity in identifying blood loss or hypovolemia.

Author

Rodrigo Kong, MD

Published/Updated

What are Likelihood Ratios?

LR, pretest probability and posttest (or posterior) probability are daunting terms that describe simple concepts that we all intuitively understand.

Let's start with pretest probability: that's just a fancy term for my initial impression, before we perform whatever test it is that we're going to use.

For example, a patient with prior stents comes in sweating and clutching his chest in agony, I have a pretty high suspicion that he's having an MI – let's say, 60%. That is my pretest probability.

He immediately gets an ECG (known here as the "test") showing an obvious STEMI.

Now, I know there are some STEMI mimics, so I'm not quite 100%, but based on my experience I'm 99.5% sure that he's having an MI right now. This is my posttest probability - the new impression I have that the patient has the disease after we did our test.

And likelihood ration? That's just the name for the statistical tool that converted the pretest probability to the posttest probability - it's just a mathematical description of the strength of that test.

Using an online calculator, that means the LR+ that got me from 60% to 99.5% is 145, which is about as high an LR you can get (and the actual LR for an emergency physician who thinks an ECG shows an obvious STEMI).

(Thank you to Seth Trueger, MD for this explanation!)

Aortic Dissection

Diagnostics and Likelihood Ratios, Explained

Positive Findings (Patient Has This)

SymptomsIncreased Disease Probability (Positive Likelihood Ratio)
Tearing/Ripping Pain10.8× (5.2-22.0)
Migrating pain7.6× (3.6-16.0)
Sudden Chest Pain2.6× (2.0-3.5)
Hx of Hypertension1.5× (0.8-3.0)
Signs on Physical ExamIncreased Disease Probability (Positive Likelihood Ratio)
Focal Neurologic Deficit33.0× (2.0-549.0)
Diastolic Heart Murmur4.9× (0.6-40.0)
Pulse Deficit2.7× (0.7-9.8)
Lab and Study FindingsIncreased Disease Probability (Positive Likelihood Ratio)
Enlarged Aorta or Wide Mediastinum3.4× (2.4-4.8)
LVH on Admission EKG3.2× (1.5-6.8)

Negative Findings (Patient Doesn't Have This)

SymptomsDecreased Disease Probability (Negative Likelihood Ratio)
Sudden Chest Pain0.3× (0.2-0.4)
History of Hypertension0.4× (0.3-0.6)
Tearing/Ripping Pain0.4× (0.3-0.5)
Migrating Pain0.6× (0.5-0.7)
Signs on Physical ExamDecreased Disease Probability (Negative Likelihood Ratio)
Pulse deficit0.63× (0.4-1.0)
Focal Neuro deficit0.87× (0.8-0.9)
Diastolic murmur1.1× (0.6-1.7)
Lab and Study FindingsDecreased Disease Probability (Negative Likelihood Ratio)
Enlarge aorta or wide mediastinum0.13× (0.02-1.00)*
LVH on admission EKG0.84× (0.7-0.9)

Narrative

  • Outcome: Thoracic Aortic Dissection
  • Multiple types of studies were reviewed in this RCE.
  • This spreadsheet reviews prospective studies only.
  • LR's are listed for these prospective studies.

Caveats

Other studies were omitted due to methodological limitations.

Author

Rodrigo Kong, MD

Published/Updated

February 22, 2012

What are Likelihood Ratios?

LR, pretest probability and posttest (or posterior) probability are daunting terms that describe simple concepts that we all intuitively understand.

Let's start with pretest probability: that's just a fancy term for my initial impression, before we perform whatever test it is that we're going to use.

For example, a patient with prior stents comes in sweating and clutching his chest in agony, I have a pretty high suspicion that he's having an MI – let's say, 60%. That is my pretest probability.

He immediately gets an ECG (known here as the "test") showing an obvious STEMI.

Now, I know there are some STEMI mimics, so I'm not quite 100%, but based on my experience I'm 99.5% sure that he's having an MI right now. This is my posttest probability - the new impression I have that the patient has the disease after we did our test.

And likelihood ration? That's just the name for the statistical tool that converted the pretest probability to the posttest probability - it's just a mathematical description of the strength of that test.

Using an online calculator, that means the LR+ that got me from 60% to 99.5% is 145, which is about as high an LR you can get (and the actual LR for an emergency physician who thinks an ECG shows an obvious STEMI).

(Thank you to Seth Trueger, MD for this explanation!)

Temporal Arteritis

Diagnostics and Likelihood Ratios, Explained

Positive Findings (Patient Has This)

Finding (Sign/Symptom/Lab/Study)Number Needed to Diagnose
(Positive Likelihood Ratio)
Beaded temporal artery 4.6×
Prominent or enlarged temporal artery 4.3×
Jaw claudication 4.2×
Diplopia 3.4×
Absent temporal artery pulse 2.7×
Tender temporal artery 2.6×
Any temporal artery abnormality 2.0×
ESR > 100 mm/h 1.9×
Scalp tenderness 1.6×
Optic atrophy or ischemic optic neuropathy 1.6×
Anemia 1.5×
Temporal headache 1.5×
Weight loss 1.3×
Any headache 1.2×
ESR > 50 mm/h 1.2×
Fatigue 1.2×
Fever 1.2×
Anorexia 1.2×
ESR abnormal 1.1×
White Race 1.1×
Any visual symptom 1.1×
Arthralgia 1.1×
Any funduscopic abnormality 1.1×
Polymyalgia rheumatica 0.97×
Myalgia 0.93×
Unilateral visual loss 0.85×
Male se× 0.83×
Vertigo 0.71×
Synovitis 0.41×

Negative Findings (Patient Doesn't Have This)

Finding (Sign/Symptom/Lab/Study)Number Needed to Diagnose
(Negative Likelihood Ratio)
ESR abnormal 0.2×
ESR > 50 mm/h 0.35×
Any temporal artery abnormality 0.53×
Prominent or enlarged temporal artery 0.67×
Any headache 0.7×
Absent temporal artery pulse 0.71×
Jaw claudication 0.72×
Anemia 0.79×
Optic atrophy or ischemic optic neuropathy 0.8×
ESR > 100 mm/h 0.8×
Temporal headache 0.82×
Tender temporal artery 0.82×
Anorexia 0.87×
Weight loss 0.89×
Fever 0.92×
Scalp tenderness 0.93×
Beaded temporal artery 0.93×
Fatigue 0.94×
Diplopia 0.95×
Any visual symptom 0.97×
Polymyalgia rheumatica 0.99×
Arthralgia 1.0×
Any funduscopic abnormality 1.0×
Myalgia 1.1×
Vertigo 1.1×
Synovitis 1.1×
Unilateral visual loss 1.2×

Narrative

  • Data derived from a highly selected (referral) population.
  • Prevalence: 39% overall in 21 studies (prevalence of TA in general Population:1%)
  • Outcome: Positive temporal artery biopsy
  • The authors' analysis suggests that there are only a few signs, symptoms, or tests that have significant predictive value for this preselected population.
  • The rest of the signs, symptoms or tests have little or no predictive value for this preselected population.

Caveats

It is unclear how these LR's are generalizable to the population at large.

Author

Rodrigo Kong, MD

Published/Updated

What are Likelihood Ratios?

LR, pretest probability and posttest (or posterior) probability are daunting terms that describe simple concepts that we all intuitively understand.

Let's start with pretest probability: that's just a fancy term for my initial impression, before we perform whatever test it is that we're going to use.

For example, a patient with prior stents comes in sweating and clutching his chest in agony, I have a pretty high suspicion that he's having an MI – let's say, 60%. That is my pretest probability.

He immediately gets an ECG (known here as the "test") showing an obvious STEMI.

Now, I know there are some STEMI mimics, so I'm not quite 100%, but based on my experience I'm 99.5% sure that he's having an MI right now. This is my posttest probability - the new impression I have that the patient has the disease after we did our test.

And likelihood ration? That's just the name for the statistical tool that converted the pretest probability to the posttest probability - it's just a mathematical description of the strength of that test.

Using an online calculator, that means the LR+ that got me from 60% to 99.5% is 145, which is about as high an LR you can get (and the actual LR for an emergency physician who thinks an ECG shows an obvious STEMI).

(Thank you to Seth Trueger, MD for this explanation!)