ACKNOWLEDGING PATIENT HETEROGENEITY IN DECISION ANALYSES: APPLICATION TO COLORECTAL CANCER SCREENING

Purpose: Different sources of patient heterogeneity (e.g., comorbidities related or unrelated to disease development or intervention effectiveness) may contribute to differential cost-effectiveness profiles when evaluating new healthcare interventions. Using the stratified analysis framework to examine the value of subgroup policies, we demonstrate the importance of capturing patient heterogeneity in colorectal cancer (CRC) screening. 

Methods: We developed a Markov model to capture the potential impacts of patient heterogeneity, represented by comorbidities related and unrelated to CRC development and screening performance, on the cost-effectiveness of CRC screening involving once-only sigmoidoscopy compared to no screening. We simulated cohorts of Norwegian men and women aged 60 years-old until death or 100 years with one of six comorbidity subgroups that differentially influenced the risks of developing CRC, dying from CRC, dying from background mortality or screening-related adverse events. Chronic obstructive pulmonary disease (COPD), dementia, and chronic renal failure reflected patients with “unrelated” comorbidities, while diabetes mellitus, obesity and smoking reflected patients with “related” comorbidities. Input parameters reflected Norwegian life tables adjusted for overall colorectal cancer mortality, and national epidemiologic and economic data. Screening effectiveness, aimed at reducing the future incidence of cancer, was based on a randomized controlled trial in Norway. In addition we reviewed the literature to derive comorbidity-specific risks of colorectal cancer incidence, adverse events, overall- and cancer specific mortality and utility values. For each comorbidity subgroup, we calculated the discounted (4%) incremental cost-effectiveness ratio (ICER), defined as the cost per quality-adjusted life year (QALY) gained, and the resulting net monetary benefit (NMB) gained from stratification, using a societal perspective.

Results: Unrelated comorbidities generally led to less attractive cost-effectiveness ratios (i.e., increased the ICER), while related comorbidity improved the cost-effectiveness profile of screening for CRC. The subgroup-specific ICERs ranged from €17,131 per QALY gained for obesity to €52,150 per QALY gained for dementia, compared to €25,071 per QALY gained for the average general population. Consequently, there was a positive NMB gained by stratification for a range of willingness-to-pay thresholds, indicating the potential value of considering patient subgroups in the analysis.

 

Conclusions: Decision makers should consider patient heterogeneity when evaluating population-based screening programs and other healthcare interventions, as decisions in favor of providing or rejecting treatment for the entire population can result in inefficient use of resources.