Analytical approach:
A lifetime decision-analytic model, based on all NBCSP data for 2008 (681,915 people), simulated the costs and benefits of the screening programme. The authors stated that the perspective was that of Australia's national health service.
Effectiveness data:
The effectiveness of screening was defined as detecting disease at an earlier stage; identifying and monitoring adenomas with a high-risk of developing colorectal cancer or identifying and treating cancer at an earlier stage. The distribution of disease stage at detection was from the NBCSP for both screened and unscreened groups. Data on five-year survival from time of diagnosis of colorectal cancer were from BioGrid Australia (a large database). After five years, survival was assumed to revert to the normal age-specific life expectancy from Australian government statistics. Other key parameters for the model included the underlying incidence of colorectal cancer, which was assumed to be the same for screening and no screening; and the participation rates for screening and diagnostic tests. All parameters were age specific, and most of the data were from the NBCSP.
Monetary benefit and utility valuations:
Not relevant.
Measure of benefit:
The primary measure of benefit was life-years saved in relation to colorectal cancer. The benefits were discounted at 3% annually.
Cost data:
Costs were assigned for each part of the screening programme: initial screening and invitation, diagnostic tests, surveillance after diagnosis, and treatment for colorectal cancer. Screening costs included programme administration, FOBT kits, pathology testing and a national colorectal cancer register. These were from the NBCSP Monitoring Report and were assumed to be zero for no screening. Diagnostic costs included general practitioner visits and colonoscopies undertaken. Surveillance costs were the age-specific costs of colonoscopy. Treatment costs, for both groups, were from a published model, based on stage of cancer at presentation. Throughout the model, except for treatment, participation rates influenced the costs. All costs were reported in 2008 Australian dollars (AUD), and they were discounted at 3% annually.
Analysis of uncertainty:
Several scenarios were run. Some simulated the impact of screening older people, with different participation and incidence rates, and with proportional increases in costs based on the additional people screened. Others tested alternative FOBT participation rates (25%, 50%, 100%), alternative five-year survival data, using recommended surveillance frequencies rather than observed frequencies, increasing the proportion of colorectal cancer prevented by resection or surveillance of adenomas, and using a zero discount rate.