High participation rates are not necessary for cost-effective colorectal cancer screening

Abstract

Background: In many countries high participation is an explicit target in screening programmes. The desire for high participation often appears to drive screening policy, although it is increasingly recognized that encouraging high participation may impinge upon the rights of an individual to make an informed choice. One argument offered in support of high participation is that it improves the cost-effectiveness of screening. This is questionable on theoretical grounds, and empirically there are conflicting results. Two recent cost-effectiveness models of faecal occult blood test (FOBT) screening for colorectal cancer (CRC) showed that cost-effectiveness was improved, another showed that cost-effectiveness was worsened and a fourth indicated that cost-effectiveness was unaffected by increasing the participation rate.

Methods: We assessed the extent to which different levels and patterns of participation affect cost-effectiveness, using decision modelling of three CRC screening with FOBT scenarios. We estimate the incremental cost-effectiveness (value for money) ratios for each scenario.

Results: The way in which participation is modelled, particularly assumptions made about the subsequent screening behaviour of non-participants ("if" and "when" a non-participant attends for subsequent screening), affects the cost-effectiveness estimates for FOBT screening programmes. 100% participation in all screening rounds gives a cost per life year saved (LYS) of USD 9705. Cost-effectiveness is worst when people who do not take part in one screening round (initial or subsequent) never take part in any future rounds of screening. Under this scenario, a participation rate of 20% in second and subsequent rounds gives a cost per LYS of USD 29,500. Under more realistic assumptions, for example the attendance of even a small proportion of non-participants in subsequent rounds, cost-effectiveness is more favourable and similar to that achieved for full participation: the scenario with a random participation rate of 20% in second and subsequent rounds for both participants and non-participants has a cost per LYS of USD 11,270.

Conclusions: Contrary to a commonly held view, high participation in screening programmes is not necessary to achieve cost-effectiveness. Setting high target participation rates in screening programmes does not guarantee cost-effectiveness and may in certain circumstances reduce the cost-effectiveness.