Report in English with a Dutch summary (KCE reports 45A)

KCE reports vol.45 Screening for Colorectal Cancer 111
CRC screening using hemoccult was approximately 8470 (£5685) for males
and 7380 (£4951) for females. Longer-term simulations, which relied on
modelling, estimated these costs at approximately 3050 (£2047) per QALY
gained for males and 2040 (£1371) per QALY gained for females. Screening in
women was more cost-effective than screening of men mainly due to the longer
life expectancy.
Gyrd-Hansen analysed alternative programs depending on combinations of
different target populations, i.e. inviting different age groups, and screening
intervals. Fifty-four of the 60 programs were found to be inefficient as being
subject to extended dominance. The estimated incremental costs per life-year
gained of the identified programs lying on the efficiency frontier ranged from
2210 to 5525 (17.000 42.500DKK). The six most efficient programs
evaluated included biennial screening of 65-74 year olds, of 60-74 year-olds, and
of 55-74 year olds; screening 55-74 year olds every 1,5 years; and annual
screening of 55-74 year-olds and 50-74 year-olds. The highest incremental cost
occurred when expanding the program from screening the 55 74 year olds
every year to include also the 50 54 year olds in the program.
Helm estimated costs per life year saved to be on average approximately
16.195 ($20.500) for screening based on the Minnesota protocol, 2150
($2700) for screening based on the Funen-1 protocol, and 1975 ($2500) for
screening based on the Nottingham protocol. The high estimate for the
Minnesota-based result was probably explained by the smaller survival benefit
associated with the trial s healthy volunteer recruitment and the practice of
FOBT rehydration, which increased the number of false positive results and
generated substantial numbers of unnecessary endoscopic investigations.
The Canadian study 498 estimated the incremental cost per life year saved due to
FOBT screening, compared with no screening, at 8335 (CAD11.907). The
most recent study of Whynes and colleagues (2004) was in line with previous
long-term simulations (see above). A screening program based on the
Nottingham trial protocol was estimated to have an incremental costeffectiveness
ratio of 2360 (£1584) (Whynes, 2004).
The Australian study 502 estimated net cost per DALY at 7200 (AUD12.000)
for a target population of individuals aged 55-69 years. This cost per DALY was
lower for older age groups, i.e. 3180 (AUD5300) and 3980 (AUD6600) for
respectively 70-74 and 75+ year olds, and higher for younger persons, i.e.
14.400 (AUD24.000) and 30.000 (AUD50.000) for respectively 50-54 and
45-49 year old persons.
Finally, the French study of Lejeune estimated the incremental cost per life year
gained at 4705 when calculated over a 10-year period and 3357 when this
period was extended to 20 years.
Sensitivity analysis
Robustness of results is checked through sensitivity analysis. In several studies,
cost items, discount rates, screening intervals, age, compliance, sensitivity and
specificity of the screening tests, survival, and complications were varied to see
how results could be influenced. All but one (Stone) only performed one waysensitivity
analysis.
Helm only performed sensitivity analysis on costs derived from the 10 th and 90 th
percentile of charges. The ICER was found to vary within about 50% of the base
values. Costs per life year saved ranged from approximately 9000 to 25.675
(baseline 16.195) for screening based on the Minnesota protocol, 1250 to