The cost‐effectiveness of evidence‐based guidelines and practice for screening and prevention of tuberculosis

Introduction: The potential cost‐effectiveness of screening depends on the risk of tuberculosis (TB) in the population being screened and the rate at which the screening outcome (prevention) is achieved. Aims: To compare the cost‐effectiveness of contact screening for TB for: (1) contact screening as it actually occurred in Victoria in 1991 (Model 1); (2) the process which should have occurred had the 1991 contact screening guidelines been followed (Model 2); (3) a hypothetical evidence‐based model (Model 3). Methods: Three models were constructed according to the aims. The cost‐effectiveness of contact screening is presented as costs to government per unit outcome (in the form of cases prevented, cases found and contacts traced) for each model. Assumptions about disease behaviour were consistent between models. A sensitivity analysis was performed to examine the effect of the assumptions made in Model 3 about rates of referral and treatment of infected contacts, and about the efficacy of isoniazid (INH) in preventing TB. Results: The total cost of Model 1 was greater than that of the other Models. Model 1 is the least cost‐effective, costing $309 065 per case prevented, and Model 3 is the most cost‐effective, costing $32 210 per case prevented. The cost of Model 2 was $58 742 per case prevented. The incremental cost‐effectiveness of Model 3 compared to Model 2 is $107 per additional contact screened, and $3881 per additional case prevented. Case finding is not as cost‐effective as best‐practice case prevention, ranging from $231 799 per case found in Model 1 to $205 596 per case found in Model 2. The sensitivity analysis shows that the cost‐effectiveness of Model 3 decreases with lower referral rates, lower rates of preventive therapy, and lower efficacy of INH. However, even allowing for reduced programme parameters, Model 3 is most cost‐effective. Discussion: Costing policy options is an important component of programme delivery, but needs to be considered in the context of the product being purchased, e.g. the prevention of disease, or case finding. Case finding as a product of contact screening is expensive in all three models. Prevention of TB, on the other hand, can be cost‐effective, as shown in Model 3. It was least cost‐effective in Model 1, largely because prevention was not considered a priority, and few infected contacts actually received preventive therapy. Clear programme aims, adherence to guidelines and high rates of preventive therapy are essential in order to achieve cost‐effectiveness. Copyright © 2000 John Wiley & Sons, Ltd.