Comparing treatments via the propensity score: Stratification or modeling?

Jessica A. Myers, Thomas Louis

Research output: Contribution to journalArticle

Abstract

In observational studies of treatments or interventions, propensity score (PS) adjustment is often useful for controlling bias in estimation of treatment effects. Regression on PS is used most often and can be highly efficient, but it can lead to biased results when model assumptions are violated. The validity of stratification on PS depends on fewer model assumptions, but this approach is less efficient than regression adjustment when the regression assumptions hold. To investigate these issues, we compare stratification and regression adjustments in a Monte Carlo simulation study. We consider two stratification approaches: equal frequency strata and an approach that attempts to choose strata that minimize the mean squared error (MSE) of the treatment effect estimate. The regression approach that we consider is a generalized additive model (GAM) that estimates treatment effect controlling for a potentially nonlinear association between PS and outcome. We find that under a wide range of plausible data generating distributions the GAM approach outperforms stratification in treatment effect estimation with respect to bias, variance, and thereby MSE. We illustrate each approach in an analysis of insurance plan choice and its relation to satisfaction with asthma care.

Original languageEnglish (US)
Pages (from-to)29-43
Number of pages15
JournalHealth Services and Outcomes Research Methodology
Volume12
Issue number1
DOIs
StatePublished - Mar 2012

Fingerprint

Propensity Score
Social Adjustment
Insurance
Observational Studies
Asthma

Keywords

  • Causal inference
  • Generalized additive model
  • Nonlinear modeling
  • Observational study
  • Optimal stratification
  • Propensity score

ASJC Scopus subject areas

  • Health Policy
  • Public Health, Environmental and Occupational Health

Cite this

Comparing treatments via the propensity score : Stratification or modeling? / Myers, Jessica A.; Louis, Thomas.

In: Health Services and Outcomes Research Methodology, Vol. 12, No. 1, 03.2012, p. 29-43.

Research output: Contribution to journalArticle

@article{de97ae3de0e14394b518abd7207d1adb,
title = "Comparing treatments via the propensity score: Stratification or modeling?",
abstract = "In observational studies of treatments or interventions, propensity score (PS) adjustment is often useful for controlling bias in estimation of treatment effects. Regression on PS is used most often and can be highly efficient, but it can lead to biased results when model assumptions are violated. The validity of stratification on PS depends on fewer model assumptions, but this approach is less efficient than regression adjustment when the regression assumptions hold. To investigate these issues, we compare stratification and regression adjustments in a Monte Carlo simulation study. We consider two stratification approaches: equal frequency strata and an approach that attempts to choose strata that minimize the mean squared error (MSE) of the treatment effect estimate. The regression approach that we consider is a generalized additive model (GAM) that estimates treatment effect controlling for a potentially nonlinear association between PS and outcome. We find that under a wide range of plausible data generating distributions the GAM approach outperforms stratification in treatment effect estimation with respect to bias, variance, and thereby MSE. We illustrate each approach in an analysis of insurance plan choice and its relation to satisfaction with asthma care.",
keywords = "Causal inference, Generalized additive model, Nonlinear modeling, Observational study, Optimal stratification, Propensity score",
author = "Myers, {Jessica A.} and Thomas Louis",
year = "2012",
month = "3",
doi = "10.1007/s10742-012-0080-3",
language = "English (US)",
volume = "12",
pages = "29--43",
journal = "Health Services and Outcomes Research Methodology",
issn = "1387-3741",
publisher = "Springer Netherlands",
number = "1",

}

TY - JOUR

T1 - Comparing treatments via the propensity score

T2 - Stratification or modeling?

AU - Myers, Jessica A.

AU - Louis, Thomas

PY - 2012/3

Y1 - 2012/3

N2 - In observational studies of treatments or interventions, propensity score (PS) adjustment is often useful for controlling bias in estimation of treatment effects. Regression on PS is used most often and can be highly efficient, but it can lead to biased results when model assumptions are violated. The validity of stratification on PS depends on fewer model assumptions, but this approach is less efficient than regression adjustment when the regression assumptions hold. To investigate these issues, we compare stratification and regression adjustments in a Monte Carlo simulation study. We consider two stratification approaches: equal frequency strata and an approach that attempts to choose strata that minimize the mean squared error (MSE) of the treatment effect estimate. The regression approach that we consider is a generalized additive model (GAM) that estimates treatment effect controlling for a potentially nonlinear association between PS and outcome. We find that under a wide range of plausible data generating distributions the GAM approach outperforms stratification in treatment effect estimation with respect to bias, variance, and thereby MSE. We illustrate each approach in an analysis of insurance plan choice and its relation to satisfaction with asthma care.

AB - In observational studies of treatments or interventions, propensity score (PS) adjustment is often useful for controlling bias in estimation of treatment effects. Regression on PS is used most often and can be highly efficient, but it can lead to biased results when model assumptions are violated. The validity of stratification on PS depends on fewer model assumptions, but this approach is less efficient than regression adjustment when the regression assumptions hold. To investigate these issues, we compare stratification and regression adjustments in a Monte Carlo simulation study. We consider two stratification approaches: equal frequency strata and an approach that attempts to choose strata that minimize the mean squared error (MSE) of the treatment effect estimate. The regression approach that we consider is a generalized additive model (GAM) that estimates treatment effect controlling for a potentially nonlinear association between PS and outcome. We find that under a wide range of plausible data generating distributions the GAM approach outperforms stratification in treatment effect estimation with respect to bias, variance, and thereby MSE. We illustrate each approach in an analysis of insurance plan choice and its relation to satisfaction with asthma care.

KW - Causal inference

KW - Generalized additive model

KW - Nonlinear modeling

KW - Observational study

KW - Optimal stratification

KW - Propensity score

UR - http://www.scopus.com/inward/record.url?scp=84859099785&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84859099785&partnerID=8YFLogxK

U2 - 10.1007/s10742-012-0080-3

DO - 10.1007/s10742-012-0080-3

M3 - Article

C2 - 25419169

AN - SCOPUS:84859099785

VL - 12

SP - 29

EP - 43

JO - Health Services and Outcomes Research Methodology

JF - Health Services and Outcomes Research Methodology

SN - 1387-3741

IS - 1

ER -