Evaluating options for a colorectal cancer screening programme in Ireland Sharp L, Tilson L, Whyte S, Ó Céilleachair A*, Walsh C, Usher C, Tappenden P, Chilcott J, Staines A, Barry M & Comber H. Population-based cancer research in Ireland, Davenport Hotel September 4th 2009

Background Over 2,000 new cases of colorectal cancer (CRC) are diagnosed in the Republic of Ireland each year. 2nd most common cancer for both genders Over 900 deaths per annum from CRC On most key indicators Irish people fare worse than their European contemporaries Higher incidence rates Lower survival rates Higher mortality amongst men As population ages incidence is projected to increase

Opportunity for Screening If caught early, CRC is very treatable Survival much higher in Stage I-II disease Screening in place for many European countries Numerous modalities exist for early detection of CRC: Guaic-based occult blood tests (gFOBT) Immunochemical-based stool tests (FIT) Flexible sigmoidoscopy

Cost-effectiveness analysis Comparing the cost-effectiveness of two policies, A and B: ICER = cost A – cost B/effect A – effect B Effects may be in life-years gained (LYG) or quality-adjusted life years gained (QALYs) The lower the ICER the “more” cost-effective A compared to B €45,000 per QALY is an informal threshold of “cost-effectiveness” in an Irish setting

Evaluating Screening Options Health technology assessment commissioned by HIQA Evaluate using cost-effectiveness analysis competing alternative strategies for CRC screening in Ireland Versus “No Screening” and also incrementally against each other Estimate the likely resource burden of screening for a range of key services and also health outcomes over a ten year time horizon after the introduction of screening.

Methods Core screening scenarios agreed with HIQA Expert Advisory Group: biennial FIT at ages 55-74 biennial gFOBT at ages 55-74, with reflex FIT FSIG once only at age 60 Supplementary scenarios also considered Diagnostic investigations for postive screen test: colonoscopy or CT colonography Surveillance for those with adenoma(s) ≥1cm removed: following current consensus recommendations (Atkins & Saunders, 2002)

Model Markov model adapted from an existing model developed by collaborators in ScHARR Natural history model of CRC Hypothetical cohort of 55 year-olds tracked over their lifetime used for cost-effectiveness Screening scenarios were then superimposed on this model Outcome measures: Cost per QALY and cost per Life Year Gained (LYG) Alternatives compared to “No Screening” and each other Costs and outcomes discounted @ 4% Healthcare payer perspective

Data Model parameters Natural history data Data on the performance of tests Cost data Other data such as uptake Data sourced from extensive literature reviews, information from existing screening programmes and expert opinion Sensitivity analysis One/multi way Probabilistic sensitivity analysis

Performance and Uptake Parameter Base-case Range FIT sensitivity adenomas cancers gFOBT sensitivity FSIG sensitivity Test uptake/compliance FIT uptake gFOBT uptake FSIG uptake 21% 71% 11% 36% 65% (low-risk) 74% (int/high-risk) 90% 53% 39% 19% - 22% 67% - 75% 10% - 12% 31% - 42% 60% - 70% 68% - 78% 85% - 95% 32% - 59% 24% - 67% COL compliance 86% 81% - 90%

Costs Parameter Base-case Range Screening tests FIT kit FIT processing and analysis gFOBT kit gFOBT processing and analysis FSIG Lifetime costs of managing symptomatic CRC stage I stage II stage III stage IV €3.75 €11.60 €1.70 €7.81 €150 €23,688 €37,180 €48,835 €36,602 €3.00 - €4.50 €9.28 - €13.92 €1.36 - €2.04 €6.25 - €9.37 €120 - €180 €18,950 – €28,425 €29,744 - €44,616 €39,068 - €58,602 €29,281 - €43,922

Incremental Cost Effectiveness vs. “No Screening” Scenario Cost of screening & CRC management Incremental cost per person1 Expected QALYs per person Incremental QALY per person1 ICER -Incremental cost per QALY gained No screening € 1074 - 10.96 gFOBT at 55-74 years € 1107 € 33.63 10.97 0.0076 € 4,4282 FIT at 55-74 years € 1114 € 40.17 10.98 0.0237 € 1,696 FSIG once at 60 years € 1077 € 3.43 0.0058 € 589 Costs and outcomes discounted at 4% 1 Each incremental value compares value for that strategy to common baseline of no screening 2 gFOBT considered dominated by a combination of FIT and FSIG

Cost-Effectiveness Plane

CE Plane: Extended Dominance

% reduction in CRC incidence rate2 % reduction in CRC mortality rate2 Health Outcomes Scenario % reduction in CRC incidence rate2 % reduction in CRC mortality rate2 gFOBT at 55-74 years 1.0% 11.8% FIT at 55-74 years 14.7% 36.0% FSIG once at 60 years 4.9% 7.5%

Health Outcomes Higher proportion of screen-detected with FIT (30% of all cancers, vs 14% with gFOBT and 3% with FSIG) Under all scenarios, screen-detected cancers have more favourable stage distribution than those detected symptomatically/clinically Sensitivity analysis found analysis to be robust. Findings did not change when using LYG as outcome measure

FSIG v FIT

But… Lifetime1 Rates per 100,00 Complications3 Scenario FSIG2 Colonoscopy2 Polypectomy2 Major bleeding4 Bowel perforation Death from perf FIT - 34,632 9,486 132 57 3.00 gFOBT 3,386 1,215 12 5 0.26 FSIG 40,177 2,543 2,487 22 0.25 FIT=faecal immunochemical test; FSIG= flexible sigmoidoscopy; gFOBT=guaiac-based faecal occult blood test 1 Over the entire lifetime of the cohort, therefore for gFOBT and FIT includes 10 screening rounds 2 Related to screening, diagnosis or surveillance 3 Complications associated with diagnostic and surveillance colonoscopy and, where relevant, FSIG 4 Major abdominal bleeding, requiring admission or intervention

Conclusions Compared to “No Screening” all of the options considered could be termed highly cost-effective. Biennial FIT 55-74 optimal strategy as it provides greater health gains at an acceptable ICER Not insignificant resource considerations and complications need to be borne in mind

Thank You