Challenges in Assessing the Process-Outcome Link in Practice
Published in: Journal of General Internal Medicine, v. 30, no. 3, Mar. 2015, p. 359-364
Posted on RAND.org on January 19, 2015
The expanded use of clinical process-of-care measures to assess the quality of health care in the context of public reporting and pay-for-performance applications has led to a desire to demonstrate the value of such efforts in terms of improved patient outcomes. The inability to observe associations between improved delivery of clinical processes and improved clinical outcomes in practice has raised concerns about the value of holding providers accountable for delivery of clinical processes of care. Analyses that attempt to investigate this relationship are fraught with many challenges, including selection of an appropriate outcome, the proximity of the outcome to the receipt of the clinical process, limited power to detect an effect, small expected effect sizes in practice, potential bias due to unmeasured confounding factors, and difficulties due to changes in measure specification over time. To avoid potentially misleading conclusions about an observed or lack of observed association between a clinical process of care and an outcome in the context of observational studies, individuals conducting and interpreting such studies should carefully consider, evaluate, and acknowledge these types of challenges.