A recent systematic review of economic evaluations suggests that screening for genital chlamydia infection is “cost effective.”¹ We are concerned about how the authors reached this conclusion since the reviewers did not take into account the fact that Chlamydia trachomatis is infectious. The methodological problems arising from this fundamental flaw raise questions about the validity of the conclusion.

The correct model to use in the evaluation of an infectious disease must be capable of encompassing all its effects, including the potential for transmission. Bernoulli first reported such transmission dynamic models in the 18th century.² The wide misuse of static, as opposed to transmission dynamic, models has been noted in the economics literature on vaccination programmes,³ but the message has been slow to transcend to the economics literature on sexually transmitted infections, with a few notable exceptions.⁴ In the case of screening for genital chlamydia, someone who is successfully treated might be re-infected; the benefits of treatment in preventing long term sequelae will be lost, and the person could continue to infect others. If they are successfully treated without re-infection, however, they will not transmit infection. Since the two possibilities have opposing effects on the number of cases, the direction of change in the cost effectiveness ratio is uncertain; it could overestimate or underestimate the true cost effectiveness. Economic evaluations that do not incorporate these effects are, therefore, very unlikely to model the outcomes of a chlamydia screening programme accurately.

Although the use of objective criteria to assess the quality of identified papers was praised in a recent STI editorial,⁵ the checklist used by Honey et al¹ is outdated and was not applied appropriately for an infectious disease. This led the authors to include papers whose results might be unreliable. The use of more recent and widely used guidelines, which ask questions about the choice of model type and the justification for the key parameters on which the model is based,⁶ may have drawn attention to the problems of static models. Furthermore, the review included studies that used “cost per case detected,” which is an inadequate outcome for screening programmes because it does not take into account resource implications associated with the course of action taken by individuals after case detection.

We have recently concluded our own systematic review of economic analyses of screening programmes for genital chlamydia infection, as part of the ongoing Chlamydia Screening Studies project (ClaSS). While the majority of studies we identified had used an incorrect modelling approach, we did identify a full economic evaluation that had used a dynamic model to evaluate chlamydia screening. This was identified by Honey et al. but excluded because they thought that it did not fulfil their inclusion criteria.¹

We propose that all future economic evaluations of chlamydia screening should use a dynamic modelling approach. A consensus panel to develop guidelines for the conduct of economic evaluations of interventions for sexually transmitted infections could take this recommendation into account.⁶