Estimation of conditional treatment effect and prediction of binary outcomes using the joint use of propensity and prognostic scores
Jonggyu Baek et al.
Abstract
Marginal treatment effect (MTE), including the average treatment effect on treated (ATT) and the average treatment effect (ATE), informs policymakers of a treatment's effect at the population level. Relative to MTE, conditional treatment effect (CTE) and patient outcome prediction are more informative at the individual level. In clinical settings with shared decision-making, patients and clinicians alike may want to know the effect of a treatment and its resulting health outcome in the context of the patient's specific health status and comorbidities. To estimate CTE on the individual level, we have proposed a method that employs a generalized additive model (GAM) with a tensor smoother of estimated propensity and prognostic scores. Using this method to control for confounders and to utilize patient's conditions, we posit that outcome prediction in the setting of observational studies will be improved. In order to apply this proposed method to practice, we examined the association between 30-day statin discontinuation and 1- and 2-year mortality among newly admitted nursing home residents in the US, 2011–2016.
Evidence weight
Balanced mode · F 0.40 / M 0.15 / V 0.05 / R 0.40
| F · citation impact | 0.50 × 0.4 = 0.20 |
| M · momentum | 0.50 × 0.15 = 0.07 |
| V · venue signal | 0.50 × 0.05 = 0.03 |
| R · text relevance † | 0.50 × 0.4 = 0.20 |
† Text relevance is estimated at 0.50 on the detail page — for your query’s actual relevance score, open this paper from a search result.