Diagnostic accuracy measures the likelihood that a test will correctly classify an individual, calculated as the sum of true positives and true negatives divided by the total number of individuals tested.

This measure reflects a weighted average of sensitivity and specificity, where sensitivity is adjusted by the prevalence of the condition in the study population, and specificity is adjusted by the complement of the prevalence. However, diagnostic accuracy rarely aligns closely with both sensitivity and specificity, except in cases where these values are nearly equal or when the disease prevalence is around 50%. While diagnostic accuracy might seem like a straightforward indicator of test validity, its reliance on prevalence limits its effectiveness compared to a more nuanced evaluation of individual test characteristics and performance.

The diagnostic accuracy can be calculated by using the following formulae:

Diagnostic Accuracy = (TP + TN) / Total

For example, if the results from a diagnostic test are Total = TP + TN + FP + FN = 80 + 70 + 10 + 20 = 180,

then, the diagnostic accuracy = (TP + TN) / Total = (80 + 70) / 180 ≈0.83

This means that the diagnostic test correctly classifies the individual as either having or not having the disease 83% of the time.

However, the diagnostic accuracy changes with disease prevalence. Similarly, predictive values are influenced by prevalence: as prevalence decreases, the positive predictive value declines while the negative predictive value increases. Whereas, positive and negative likelihood ratios and the area under the ROC curve are derived solely from sensitivity and specificity and remain unaffected by disease prevalence.

Reference:

1. Šimundić, A.M., 2009. Measures of diagnostic accuracy: basic definitions. EJIFCC, 19(4), p.203.
2. McGee, S., 2012. Evidence-based physical diagnosis. Elsevier Health Sciences.
3. Fletcher, R.H., Fletcher, S.W. and Fletcher, G.S., 2012. Clinical epidemiology: the essentials. Lippincott Williams & Wilkins.