Table 7.

 Strengths and weaknesses of the main research methods used to quantify overdiagnosis from cancer screening

Research method	Strengths	Weaknesses
Follow-up of randomized controlled trials	Best able to minimize biases Directly answers question of interest	Substantial time and resource requirements Limited external validity Not useful for monitoring over time
Modeling	Can project through areas of uncertainty Not limited by time constraints Can evaluate multiple screening situations Can be used for monitoring over time	Validity of results depends on assumptions (poor directness) Needs constant updating of model constraints to reflect changing nature of cancer diagnosis and treatment Small changes in assumptions and model can lead to large changes in estimates Difficult to critically appraise (a “black box”) Need long follow-up to determine overdiagnosis, yet uncertainty increases with time in models May give false sense of precision, insufficient attention to uncertainty
Pathological and imaging studies	Can be used for monitoring over time One of the simplest approaches	Validity of results depends on assumptions (poor directness) Not able to account for competing mortality Need to be sure all diagnosed cases are ascertained, and that data are collected in same way
Ecological and cohort studies	Directly answers question of interest Provides “real world” view of overdiagnosis Able to compare results from different settings Can be used for monitoring over time	Potential for confounding factors related to diagnosis, treatment, and health status between populations Moderate time requirements Needs investment in population registries, full and accurate ascertainment of all cases, and full and accurate collection of potential confounders