Using machine learning to examine heterogeneous treatment effects (MA)
Background:
Large observational data are increasingly available in health, economic and social sciences, where researchers are interested in causal questions rather than prediction. Advances in machine learning (ML) and artificial intelligence offer large potential benefits to patients. ML models can learn patterns of health trajectories of vast numbers of patients. These techniques can help physicians to anticipate future events drawing from information well beyond the physician’s practice experience. Thus, machine learning is particularly promising for the search and detection of complex relationships in high-dimensional data such as in electronic health records. One important application of ML in practice is the estimation of heterogeneous treatment effects in healthcare. The overall average effect, as commonly estimated in randomized experiments alone is often of limited value and physicians and researchers want to know when treatments do and do not work. Heterogeneous treatment effects can for instance occur in a drug treatment which has varying efficacy depending on individual characteristics. More general, the estimation of treatment effect heterogeneity plays an essential role in selecting the most effective treatment from a number of available treatments, determining subpopulations for which a treatment is effective or harmful, designing individualized optimal treatment regimes, and generalizing causal effect estimates obtained from an experimental sample to a target population. The appeal of machine learning is that it manages to uncover generalizable patterns and discovers relationships not specified in advance.
Objective:
The aim of the thesis is an evidence-based assessment of benefits, risks, and central challenges of using machine learning models in health economics. In particular, the focus should be directed to the application of ML around the subject heterogeneous treatment effects.
Literature:
Athey, S. (2019). “The Impact of Machine Learning on Economics”, in Agrawal, A., Gans, J. & Goldfarb, A. (eds.). The Economics of Artificial Intelligence: An Agenda. Chicago: University of Chicago Press. doi:10.7208/9780226613475
Caron, A., Baio, G., & Manolopoulou, I. (2022). “Estimating individual treatment effects using non-parametric regression models: A review”, Journal of the Royal Statistical Society (A): Statistics in Society, vol. 185, no. 3, pp. 1115–1149. doi: 10.1111/rssa.12824.
Padula, W. V., Kreif, N., Vanness, D. J., Adamson, B., Rueda, J.D., Felizzi, F., Jonsson, P., IJzerman, M.J., Butte, A., & Crown W. (2022). “Machine Learning Methods in Health Economics and Outcomes Research—The PALISADE Checklist: A Good Practices Report of an ISPOR Task Force”, Value in Health, vol. 25, no. 7, pp. 1063-1080. doi: 10.1016/j.jval.2022.03.022.
Note: Writing in English language desired.