Background:
Decision analysis models are often used in economic evaluations of sexually transmitted disease (STD) screening programs. Because decision analysis models rarely incorporate transmission effects beyond an immediate group of partners, these models can understate the true benefits of screening. Recently, more emphasis has been placed on the importance of using dynamic transmission models to evaluate screening programs.

Objective:
To examine the use of simple linear adjustments to decision analysis models of STD screening to incorporate population-level transmission impacts.

Method:
We examined the cost-effectiveness of a hypothetical screening program. We assumed that X was the number of adverse outcomes averted estimated by a decision analysis model, compared to yX, the true number of adverse outcomes averted (y > 1). We assumed y was unknown, but that an estimate z could be obtained based on transmission models previously applied in similar populations. We examined how the estimated cost-effectiveness ratio estimated by the decision analysis would change as z ranged from 1 to 2y.

Result:
As z increased from 1 to y, the estimated cost-effectiveness ratio approached the true cost-effectiveness ratio. When z exceeded y, the cost per adverse outcome averted was underestimated, but in many cases was closer to the actual cost per outcome averted than the unadjusted estimate.

Conclusion:
If general, rule-of-thumb estimates of the ratio of the number of adverse outcomes prevented to the number predicted by static models were developed, the cost-effectiveness estimates generated by static models could be adjusted to incorporate the effects of STD transmission.

Implications:
Although decision analysis models are easier to create and use than dynamic transmission models, dynamic models can provide more accurate estimates of the cost-effectiveness of STD prevention programs. These findings highlight the potential to improve the accuracy of decision analysis models without sacrificing their relative simplicity.