Background:

Colorectal cancer screening is now recommended in the general population beginning at age 50 for those at average risk. The most common colorectal cancer screening test in use in the United States is the guaiac based fecal occult blood test (FOBT). Colorectal cancer screening is now covered by Medicare with a reimbursement level of $4.50 for the guaiac test. Immunochemical fecal occult blood tests (IFOBT) have tended to be more expensive and have not yet been widely used in the US. In order to inform coverage and payment decisions related to the use of these tests, this report estimates the cost effectiveness of an immunochemical test with test performance parameters that are equivalent to or better than those associated with the guaiac test. We also report the threshold payment level of the immunochemical test relative to the guaiac test, the level of payment for the immunochemical test that would result in cost-effectiveness equivalent to that of the comparative guaiac test.

Methods:

We use a micro-simulation model, MISCAN-COLON, developed and validated by Erasmus University to describe the natural history of the adenoma carcinoma sequence and the impact of screening on reducing colorectal cancer incidence and mortality. The cost effectiveness of life years gained relative to costs for screening are derived for screening tests with different test performance characteristics. We review the literature for guaiac and immunochemical tests to establish reasonable test performance levels of sensitivity and specificity for these tests. Although the efficacy of FOBT screening was established using the guaiac based Hemoccult II test, the guaiac based Hemoccult SENSA test has higher sensitivity but lower specificity than Hemoccult II and recently has increased in use. Consequently we consider both Hemoccult II and Hemoccult SENSA as base cases. We assume base case values for Hemoccult II of 40% sensitivity for colorectal cancer, 10% sensitivity for adenomas ≥1.0 cm, 5% sensitivity for adenomas < 1 cm, and 98% specificity (for not having cancers or adenomas). For base case values for Hemoccult SENSA we assume 70% sensitivity for colorectal cancer, 17% sensitivity for adenomas ≥ 1 cm, 9% sensitivity for adenomas < 1.0 cm and 92.5% specificity. We found less definitive evidence for establishing estimates of sensitivity and specificity for most of the immunochemical tests in a general population. Therefore we assumed a more favorable and a less favorable case for the immunochemical tests. We assumed that the immunochemical tests have sensitivities comparable to Hemoccult SENSA but with higher specificity (98% and 95%). Consequently we assumed that the immunochemical test had 70% sensitivity for colorectal cancer, 17% sensitivity for adenomas ≥ 1 cm, 9% sensitivity for adenomas < 1 cm, and 98% specificity for the more favorable case and 95% specificity for the less favorable case. We also assumed the same sensitivity parameters for the immunochemical tests but with 95% specificity. Furthermore in a sensitivity analysis, we assumed that the sensitivity of the immunochemical test increases 25%, 50%, 75%, and 100% over that of the Hemoccult II base case. In the base case we assumed extended intervals of surveillance for those with lower risk adenomas as specified in the most recent surveillance guidelines. We repeated these analyses, assuming a more intensive pattern of 3-year surveillance colonoscopy for all with adenomas detected. The cost and health effect outcomes were derived under the assumption that no colorectal cancer screening occurred prior to age 65 and all (100%) were compliant. In a sensitivity analysis, we assumed more realistic compliance rates and a higher compliance for immunochemical FOBT than for Hemoccult FOBT screening. We also determined the threshold analysis value for Hemoccult II and Hemoccult SENSA when an immunochemical test was the base case.

Results:

The cost effectiveness of the Hemoccult II FOBT ($1,071 per life year gained) is a very favorable level of cost-effectiveness in comparison to other cancer screening modalities. Immunochemical tests, even with costs per test of $28 per test, still have a cost effectiveness ratio of no more than $4,500 per life year saved. At a paymemt level of $28 for IFOBT and $4.50 for Hemoccult II, the incremental cost effectiveness ratio (ICER) for IFOBT is $11,000 per additional life-year saved assuming a specificity of 98% for IFOBT and $21,000 per additional life-year saved assuming a specificity of 95% for IFOBT. The threshold payment level of the IFOBT, with 98% specificity for most test parameters considered, was in the range of $7.00 to $13.00, which is only somewhat higher than the $4.50 of the base case Hemoccult II. However when the IFOBT has specificity of 95%, then the threshold values for most test parameters considered were less than zero dollars. Results for IFOBT are much more favorable if Hemoccult SENSA is assumed to be the base case and especially if IFOBT is assumed to operate at the more favorable specificity value of 98%. A threshold payment level of $28 for IFOBT is exceeded if either or both of the following conditions are met: a) IFOBT is assumed to have the lower specificity value of 95% but much better values of sensitivity for the detection of adenomas than Hemoccult SENSA, or b) IFOBT is assumed to have sensitivity values equal to Hemoccult SENSA but the higher specificity value of 98%. If we assume payment rates of $18 and $27 for IFOBT, then the corresponding threshold payment levels are $10 and $17 for Hemoccult II when IFOBT has 98% specificity and $5 and $14 for Hemoccult SENSA when assuming 95% specificity for IFOBT.

Conclusion:

Fecal occult blood tests, either guaiac based or immunochemical based, provide for a very cost effective intervention for reducing colorectal cancer incidence and mortality. If the immunochemical fecal occult blood test maintains the high specificity of Hemoccult II (98%) and increases sensitivity for colorectal cancer to 70% over that of Hemoccult II (40%), then a unit cost level of approximately $13.00 would provide a comparable cost-effectiveness to Hemoccult II at $4.50 per unit cost. If the specificity of the immunochemical fecal occult blood test is assumed to be 95% when the sensitivity for colorectal cancer increases to 70%, then the threshold payment level for IFOBT would actually be lower than the current $4.50. However, further threshold analysis using Hemoccult SENSA as the base case with a sensitivity of 70% for colorectal cancer and specificity of 92.5% indicates that the immunochemical test could achieve a threshold payment level in excess of $28 when the more favorable assumptions about IFOBT are made.

Evidence about the relative specificity and sensitivity of IFOBT in comparison to Hemoccult II and Hemoccult SENSA is sparse and highly uncertain. Therefore the scenarios under which the threshold payment level of $28 is exceeded for IFOBT, although potential possible, cannot be considered to be strongly evidence based. If payment level of $18 and $27 are assumed for IFOBT, corresponding threshold payment levels for Hemoccult II would be higher than current payment levels while this would be true for Hemoccult SENSA only if the lower specificity value of 95% is assumed for IFOBT.