Practice Perfect - A PRESENT Podiatry eZine

Jarrod Shapiro, DPM
Practice Perfect Editor
Assistant Professor,
Dept. of Podiatric Medicine,
Surgery & Biomechanics
College of Podiatric Medicine
Western University of
Health Sciences,
St, Pomona, CA

How Accurate Is Your Clinical Exam? Part 1

One of the great benefits of being in academic medicine is the constant demand to sharpen your clinical skills. Imagine working in your office or clinic and having students following along, seeing patients with you. At various points, your students will ask you to answer questions that you’ve taken for granted for years. For example, how accurate is palpating pulses in predicting peripheral arterial disease? Or, what is the reliability of the talar tilt and anterior drawer tests for diagnosis of lateral ankle ligament ruptures? How about Mulder’s click in diagnosing intermetatarsal space neuromas? Sometimes I’ll have a response, but more often than not my students ask very sharp questions that I can’t answer. As a result, I proceed to the medical literature to find the answer. For those of you already expert at this use of the medical literature, feel free to stop reading. For the rest of us, I hope this week’s and next week’s quick review helps you utilize the written evidence and keep you questioning.

Tonight's Premier Lecture is
Indications Risks and Outcomes of Lower Extremity Revascularization in Patients with PVD
by Allan Hamdan, MD

Determining clinical reliability: the terminology

Students will ask you to answer questions that you’ve taken for granted for years.

The first step in any process is understanding the pertinent terminology. Here are the major terms utilized in the medical literature when discussing reliability and diagnostic accuracy of clinical tests.

2x2 Contingency Table: This table is used when comparing a clinical test with a reference standard (the best way to determine the accuracy of a test). For example, the best way to determine if the probe to bone test accurately predicts osteomyelitis would be to compare it to bone biopsy and microbiological analysis. Take a look at the table below. From this table we are able to generate the rest of the important terms.

2x2 Contingency Table.	Reference Standard Positive	Reference Standard Negative
Clinical Test Positive	True-positive result (a)	False-positive result(b)
Clinical Test Negative	False-negative result (c)	True-negative result (d)

Overall Accuracy: This is a calculation that gives us a picture of the test’s general accuracy. It is found by dividing the correct responses (true positives and negatives) by the total number of patients. It has the following formula: (a+d)/(a+b+c+d) x 100%.

Positive Predictive Value (PPV): This value estimates the likelihood that a patient who tests positive actually has the disease. It has the following formula: a/(a+b) x 100%. The higher the PPV the more likely the patient has the disease.

Each of us unconsciously creates a pretest probability during multiple points of a patient encounter.

Negative Predictive Value (NPV): This value estimates the likelihood that a patient without the disease will test negative(i.e. the number of true negatives divided by the other negatives). It has the following formula: d/c+d x 100%

Sensitivity: This is the proportion of patients who actually have the disorder and have a positive test. It has the following formula: a/(a+c) x 100%.

Specificity: The proportion of patients without the disorder who have a negative test. It has the following formula: d/(b+d) x 100%.

Pretest Probability: The likelihood that a patient has the disorder before the actual examination or test. Each of us unconsciously creates a pretest probability during multiple points of a patient encounter. A patient complaining of inability to walk and lateral ankle pain after a fall may – in your mind - have a high pretest probability of an ankle fracture.

Posttest Probability: This is the likelihood that a patient has the disorder after the particular test is performed. For example, after examining your ankle pain patient, you find he is, i fact, is unable to walk and has focal pain to touch at the lateral malleolus. An ankle fracture diagnosis now has a high posttest probability. New pretest and posttest probabilities now exist around the radiographs you order next.

Likelihood Ratio (LR): This term is used directly with pre- and post-test probabilities. LR uses a test’s sensitivity and specificity to demonstrate a shift in the pretest probability. LRs are either positive (favoring the existence of the disorder) or negative (favoring the absence of a disorder).

Positive LR formula: LR = Sensitivity/(1-Specificity)
Negative LR formula: LR = (1-Sensitivity)/Specificity

Look at the table below for interpretation of likelihood ratios.

+ LR	- LR	Interpretation
> 10	< 0.1	Conclusive shift in probability
5-10	< 0.1 - 0.2	Moderate shift in probability
2 - 5	< 0.2 - 0.5	Small, sometimes important shift in probability
1 - 2	< 0.5 - 1.0	Rarely important shift in probability

*Jaeschke, et al. JAMA, Feb 1994; 271(5): 389-391.

Fagan Normogram: For those of you who like a very organized and mathematical approach to things, you’ll enjoy the Fagan normogram. This is a previously created chart that allows one to graph your pretest probability (either determined by you or found in the literature) compared with the published likelihood ratio to find a posttest probability. Consider this example. You want to know how smart one of your podiatric colleagues is, so you look at his “PRESENT Sign” (an extra gleam in the eye as a result of greater knowledge). You consider your colleague to have an 80% pretest probability of being well educated and found that the PRESENT sign has a likelihood ratio of 10 in predicting level of knowledge. Plotting these numbers of the normogram below demonstrates a posttest probability of 97% (i.e. your colleague with a positive PRESENT sign has a 97% chance of being well educated). Most of us don’t have the time to run through this exercise while in clinic, but it provides an important method when considering examination techniques and laboratory and imaging methods.

The terms discussed above provide us with the basic methods to analyze and discuss the medical literature in regards to the accuracy and reliability of various diagnostic testing methods. Next week we’ll apply these terms to several podiatric physical examination techniques.

Keep writing in with your thoughts and comments. Better yet, post them in our eTalk forum.
Best wishes.

###

Jarrod Shapiro, DPM
PRESENT Practice Perfect Editor
[email protected]