Interpreting a Screening Program's Falling Death Rate
A county launches a colorectal cancer screening program. Over five years, screening uptake rises from 40% to 62%, and colorectal cancer mortality in the county falls from 18 to 14 per 100,000. The program director claims the screening caused the decline and wants to expand it countywide with a large budget. A critic notes that statewide mortality fell similarly during the same period, that the county's population is aging, and that treatment improvements occurred concurrently. The health department must decide whether to recommend expansion and how to evaluate the program.
1. What are the key threats to the claim that screening caused the mortality decline, and how should the evaluation be designed?
Read the worked answer
Threats to validity:
- Secular trends: Statewide mortality fell similarly, suggesting broad factors (treatment, awareness, other screening) rather than the county program.
- Confounding by age: An aging population could change both screening uptake and mortality patterns.
- Lead-time and length bias: Screening detects cancers earlier; survival improves without necessarily reducing mortality.
- Ecologic fallacy: County-level associations do not prove individual-level causation.
- Concurrent treatment improvements: Better chemotherapy and surgery could drive the decline.
Evaluation design: Use a difference-in-differences comparison: compare the change in county mortality with the change in comparable counties or the state, adjusting for age and other factors. Track intermediate outcomes (screening uptake, stage at diagnosis, adenoma detection). Ideally, use individual-level linkage of screening and outcomes. Pre-specify outcomes and allow enough follow-up (mortality benefits lag).
2. Given the evidence, should the health department recommend countywide expansion now? What conditions would justify expansion?
Read the worked answer
Recommendation: Do not recommend full expansion on the current evidence alone. The observed decline is consistent with the program but not attributable to it; the statewide parallel decline and aging population are serious alternative explanations.
Conditions that would justify expansion:
- Stronger evidence of effect - difference-in-differences showing the county's decline exceeds the comparison, with adjustment for age and treatment trends.
- Process success - sustained uptake gains, especially in underserved groups, and improved stage at diagnosis.
- Cost-effectiveness - screening is generally cost-effective, but expansion should target high-yield populations.
- Capacity - colonoscopy and follow-up capacity to handle increased demand without delays.
Interim step: Continue and possibly expand to underserved areas with strong monitoring, rather than a blanket countywide scale-up. Frame the decision as evidence-informed, not evidence-proof.