Unmeasured Confounding in Observational Studies of Management of Cerebellar Intracranial Hemorrhage

To the Editor In a propensity score –matched cohort of 578 patients from 4 observational cohort studies, Dr Kuramatsu and colleagues showed that evacuation of medium-sized intracerebellar hematomas (approximate volume, 20 cm3) was not associated with better functional outcome. Assessing treatment effectiveness in observational data is challenging because treatment decisions are based on patient characteristics that also are typically predictive of outcome, causing confounding by indication. Although the authors addressed this potential bias with propensity scores, we would like to emphasize the possibility of residual confound ing.

Source: JAMA - February 18, 2020 Category: General Medicine Source Type: research