The Propensity Score

Propensity Score Matching in Nonrandomized Studies: A Concept Simply Explained Using Antidepressant Treatment During Pregnancy as an Example

“…As a result, the effect of treatment on outcome can be modeled with reduced bias. 6 There are other methods of matching, too. Each method has its strengths and limitations.…”

“…As a first step, the investigators use a large number of potential explanatory variables in a logistic regression procedure to mathematically model the relationship between these variables and the grouping variable. 5,6 For example, Oberlander et al 3 used over a dozen explanatory variables to mathematically model women who did or did not receive an SSRI to treat depression during pregnancy. These explanatory variables included age, income, number of prenatal visits, number of physician…”

“…Lastly, whereas propensity score matching cannot establish cause-effect relationships, as can RCTs, it can provide useful guidance for situations in which RCTs cannot be conducted. 6 …”

Reflections From a Statistical Editor: Elements of Great Manuscripts

“…For example, in observational studies that utilize propensity score matching to balance observed confounders between study groups, some methodologists recommend using analytic approaches that account for the paired nature of the data resulting from a propensity score matching procedure. 7 Great manuscripts utilize analytic approaches that match the study designs, maximizing the reason and benefits for such designs.…”

National comparison of oncologic quality indicators between open and robotic-assisted radical cystectomy

“…Propensity score adjustment was then performed using the covariate adjustment method [7]. The initial regression model for propensity adjustment for the treatment variable (surgical approach: RARC or ORC) was constructed for patient age, sex, race, Charlson/Deyo score, urban/rural and geographic location, insurance status, income, composite American Joint Committee on Cancer stage, hospital type, chemotherapy status, and year to best adjust each cohort.…”