Purpose
Investigational testing can aid clinicians in the etiological diagnosis of uveitis. However, commonly performed batteries of tests can be costly and non-specific. The purpose of this study was to evaluate commonly ordered studies for the work-up of a patient presenting to the ophthalmologist with uveitis and determine the diagnostic value of these tests in an incremental cost perspective.
Methods
This analysis was conducted by performing a current literature search of both infectious and inflammatory causes of uveitis for epidemiological data, including disease prevalence in the United States, laboratory testing for each disease entity ordered at the University of Virginia Department of Ophthalmology, sensitivity and specificity data for each available test, positive and negative predictive values of studies, pretest probability for disease, and Medicare and Medicaid costs and provider reimbursements for the state of Virginia. A revised Bayes’ theorem statistical analysis utilizing TreeAge software was performed to determine cost effectiveness ratios for each test, further grouped by diagnosis. Etiologies were ranked using cost-effectiveness units (CEU), with lower values indicating a more cost-effective test. Diagnoses involving more than one applicable test were ranked by an average cost-effectiveness.
Results
A total of 16 diagnoses were considered for this analysis. The average cost effectiveness for each diagnosis was determined and the diagnoses were ranked by most cost-effective to exclude. This revealed that completing the evaluation in a step-wise manner resulted in the most effective use of resources. For Medicare patients, Rheumatoid arthritis was the most effective first diagnostic evaluation, with an average cost effectiveness of 10.4 CEU. Subsequently, this was followed by syphilis (14.5 CEU), bartonella infection (15.6 CEU), granulomatosis with polyangiitis (Wegener’s Granulomatosis) (18.1 CEU), and Polyarteritis Nodosa (19.0CEU).
Conclusions
Due to the broad differential of causes of uveitis, diagnostic testing by the ophthalmologist can be useful in determining an etiology. However, full batteries of studies are often non-specific and costly to perform. Conducting diagnostic testing using a pre-determined algorithm can aid the ophthalmologist in arriving at a diagnosis within resource constraints.