Presentation (CTI)

OBJECTIVES: Health economic models are increasingly used to inform healthcare decision-making; however, cost-effectiveness analyses (CEAs) are implemented using diverse modeling platforms that differ in transparency and usability. This study compared cost-effectiveness results and implementation characteristics of identical short-term CEAs independently implemented using TreeAge Pro and R with a shared Excel-based input dataset.
METHODS: A short-term decision-tree CEA was conducted over a 40-week time horizon from a U.S. healthcare payer perspective using data derived from the SURPASS-2 randomized controlled trial. The same decision problem, model structure, clinical inputs, and cost parameters were defined in a single Excel-based dataset and independently implemented in two platforms: TreeAge Pro (graphical, proprietary software) and R (open-source, script-based implementation using the heemod package). Effectiveness outcomes included change in hemoglobin A1c and body weight; costs included drug acquisition and treatment-related adverse events. Incremental costs, incremental effects, and incremental cost-effectiveness ratios (ICERs) were calculated in each platform. Deterministic sensitivity analyses varied key parameters. Implementation characteristics, including transparency, flexibility, and usability, were qualitatively assessed.
RESULTS: Both implementations produced directionally consistent CE results. ICERs for the primary outcome were highly consistent across platforms (TreeAge Pro: approximately $5,700 per 1% reduction in hemoglobin A1c; R: approximately $5,600), with relative differences below 2%. In both models, the intervention of interest was more effective and more costly than the comparator. Sensitivity analyses identified identical key drivers of CE across platforms, and conclusions regarding relative value were unchanged. The R-based implementation enabled greater transparency and flexibility for scenario testing, whereas TreeAge Pro facilitated rapid model visualization.
CONCLUSIONS: Independent implementation of an identical short-term CEA in TreeAge Pro and R yielded consistent CE conclusions, demonstrating that results were driven by assumptions and inputs rather than software choice. Open-source modeling platforms may enhance transparency and reproducibility without materially altering conclusions, supporting their complementary use alongside proprietary tools in HTA practice.