Guertin JR, Conombo B, Langevin R, et al. A Systematic Review of Methods Used for Confounding Adjustment in Observational Economic Evaluations in Cardiology Conducted between 2013 and 2017. Medical Decision Making. 2020;40(5):582-595. doi:10.1177/0272989X20937257

Langevin, R. Policy Learning with Observational Data: The Case of Hepatitis C Treatment for HIV/HCV Co-Infected Patients (Job Market Paper)

Abstract: Decision-makers often must select a single option from a finite set of alternatives: physicians choose a medical treatment, investors select a risk level for their portfolio, and insurers set a premium for an insurance policy. To achieve better outcomes, policymakers sometimes provide policy rules or guidelines to direct such decisions. In this paper, I show how to design an optimal policy rule using observational data under relatively weak assumptions about the underlying structure of the heterogeneous sampled population. Consistent estimation of conditional average treatment effects is performed via a weighted K-means algorithm, assuming the outcome model is correctly specified within each group of homogeneous observations. Feasible and optimal policy rules are implemented through a standard decision tree under both perfect and imperfect adherence to treatment. This methodology is applied to the case of treatment choice for Hepatitis C (HepC) in patients co-infected with the human immunodeficiency virus (HIV) and the HepC virus (HCV), a setting where there is no uniform guideline regarding modern pharmaceutical treatment options. Estimation results show that one group of patients has an approximate 80% chance of spontaneously clearing HCV without any treatment. Estimation results also show that reallocating treatments among treated individuals could have reduced total treatment costs by between CAN$2.7 million and CAN$3.3 million, while slightly increasing total health benefits under either perfect or imperfect adherence to treatment compared to the status quo. Improved guideline recommendations for the management of HIV/HCV co-infected patients can be easily derived using the proposed methodology.

Langevin, R. Bias-Reduced Estimation of Finite Mixtures: An Application to Latent Group Panel Structures

Abstract: Finite mixtures are often used in econometric analyses to account for unobserved heterogeneity. This paper demonstrates that maximizing the likelihood of a finite mixture of parametric densities yields estimates that can be largely biased in finite samples under weak regularity conditions. The large bias is caused by the presence of outliers in component densities with unbounded or large support. The size of the bias is positively correlated with the degree of overlap between the densities within the mixture. In contrast, I show that maximizing the classification-mixture likelihood function equipped with a consistent classifier leads to less biased and more efficient estimates for all parameters in the mixture compared to those obtained from standard MLE procedures. Monte Carlo simulations confirm the presence of large finite-sample biases in the estimated parameters, with the size of the biases reducing as either the sample size or the overall distance between the component densities tends to infinity. Simulation results also show that the proposed estimation strategy generally outperforms standard MLE procedures in finite samples, both in terms of estimation bias and efficiency. In an application using latent group panel structures and health administrative data, estimation results show that the proposed strategy leads to a reduction in out-of-sample prediction error of around 17.6\% compared to the best results obtained from standard MLE procedures.

Langevin, R. Bending the Cost Curve in Health Care: The Role of Community-Based Organizations

Abstract: Healthcare costs have been rising worldwide in recent decades, hence putting pressure simultaneously on households, governments, insurers, and healthcare providers. This paper demonstrates that higher funding for community-based organizations (CBOs) is likely to yield substantial long-term savings for healthcare systems through improvements in social determinants of health (SDoH). Estimation of elasticities over time is performed via a first-differenced two-way fixed effects (FD-TWFE) estimator that accounts for the structural breaks caused by the COVID-19 pandemic. I show that this estimator is more efficient and less prone to biases coming from feedback effects compared to other fixed effects estimators when the error part is strongly persistent over time. I use regional-level health administrative data from Quebec, Canada, between 2009 and 2023 to estimate all savings generated by CBO public funding. Estimation results show that each additional dollar given to CBOs in the province yields average savings of around CAN$8.00 2 to 7 years later. Average savings tend to increase over the study period, namely those generated by CBOs funded by Quebec's Ministry of Health and Social Services (MSSS). The estimated savings also vary across groups of regions, with more densely populated regions featuring large and statistically significant savings over the entire study period. This heterogeneity could be leveraged to help less effective regions reduce the growth of healthcare expenditures over time.

Ammi, M., Langevin, R., Strumpf, E. C., & Arpin, E. (2024). Effets de la pandémie de COVID-19 sur la réallocation des dépenses de santé publique par fonction : estimation de court terme et analyse prédictive contrefactuelle (2024RP-11, Rapports de projets, CIRANO. In French.) https://doi.org/10.54932/LSLR2977

Ammi, M., Langevin, R., Strumpf, E. C., & Arpin, E. (2024). S’attaquer aux crises épidémiologiques : oui, mais à quel prix ? (2024PJ-07, Revue PERSPECTIVES, CIRANO. In French.) https://doi.org/10.54932/TUPX6305

Langevin, R., Couturier, E. (2023). Spirale salaires-inflation: mythe ou réalité? Institut de recherche et d’informations socioéconomiques (IRIS).

Devin N., Langevin R. (2022), Market Basket Measure Research: Additional Income Inequality Indicators using the Market Basket Measure, Income Research Paper Series, Statistics Canada, Catalogue no. 75F0002M

Langevin, R. (2021). Hausser l’impôt des plus riches: des bénéfices qui dépassent les inconvénients, Institut de recherche et d’informations socioéconomiques (IRIS).

Langevin, R. (2019). Les projets de loi sur la laïcité augmentent-ils le nombre de crimes haineux au Québec? Ligue des droits et libertés, section Québec.

Langevin, R., Guay, E. (2018). L’austérité a-t-elle contribué à la relance économique au Québec? Institut de recherche et d’informations socioéconomiques (IRIS).

Dufour, M., Langevin, R. (2016). Quels seraient les effets réels d’une hausse marquée du salaire minimum? Institut de recherche et d’informations socioéconomiques (IRIS).