Cost-Effectiveness Analysis from a Randomized Controlled Trial of Tailored Exercise Prescription for Women with Breast Cancer with 8-Year Follow-Up

Abstract

Studies show conflicting results on whether exercise interventions to improve outcomes for women with breast cancer are cost-effective. We modelled the long-term cost-effectiveness of the Exercise for Health intervention compared with usual care. A lifetime Markov cohort model for women with early breast cancer was constructed taking a societal perspective. Data were obtained from trial, epidemiological, quality of life, and healthcare cost reports. Outcomes were calculated from 5000 Monte Carlo simulations, and one-way and probabilistic sensitivity analyses. Over the cohort's remaining life, the incremental cost for the exercise versus usual care groups were $7409 and quality-adjusted life years (QALYs) gained were 0.35 resulting in an incremental cost per QALY ratio of AU$21,247 (95% Uncertainty Interval (UI): Dominant, AU$31,398). The likelihood that the exercise intervention was cost-effective at acceptable levels was 93.0%. The incremental cost per life year gained was AU$8894 (95% UI Dominant, AU$11,769) with a 99.4% probability of being cost effective. Findings were most sensitive to the probability of recurrence in the exercise and usual care groups, followed by the costs of out-of-pocket expenses and the model starting age. This exercise intervention for women after early-stage breast cancer is cost-effective and would be a sound investment of healthcare resources.

Keywords: breast cancer; cost-effectiveness analysis; cost-utility analysis; exercise.

Figure 1

Illustration of Markov model health states.

Figure 2

Probabilistic sensitivity analyses, incremental cost per QALY gain scatterplot. AU$ = Australian dollars, QALY = quality-adjusted life year.

Figure 3

Cost-effectiveness acceptability curves.

Figure 4

One-way sensitivity analyses; incremental cost per QALY gain.