Rational laboratory testing in uveitis: A Bayesian analysis

Uveitis encompasses a heterogeneous group of clinical entities with the common feature of intraocular inflammation. In addition to patient history and examination, a focused set of laboratory investigations is frequently necessary to establish a specific diagnosis. There is limited consensus among uveitis specialists regarding appropriate laboratory evaluation for many distinct patient presentations. The appropriateness of a laboratory test for a given case of uveitis will depend on patient-specific as well as epidemiologic factors. Bayesian analysis is a widely used framework for the interpretation of laboratory testing, but is seldom adhered to in clinical practice. Bayes' theorem states that the predictive value of a particular laboratory test depends on the sensitivity and specificity of that test, as well as the prevalence of disease in the population being tested. In this review we will summarize the performance of commonly-utilized laboratory tests for uveitis, as well as the prevalence of uveitic diagnoses in different geographic practice settings. We will propose a logical framework for effective laboratory testing in uveitic disease through rigorous application of Bayesian analysis. Finally, we will demonstrate that while many highly sensitive laboratory tests offer an effective means to rule out associated systemic disease, limited test specificity and low pre-test probability often preclude the diagnosis of systemic disease association with any high degree of certainty, even in the face of positive testing.