Quality measures and sociodemographic risk factors: to adjust or not to adjust.

In 2006, the National Quality Forum (NQF) established a policy against adjusting quality measures for sociodemographic risk factors, ie, socioeconomic status and other social risk factors.1 This policy was based on concern that statistical adjustment could mask poor care provided to socially disadvantaged patients and create lower standards of care. However, given the potential influence of sociodemographic risk on health and health care, this policy of not including risk adjustment, adopted by the Centers for Medicare & Medicaid Services (CMS) and others, potentially results in unfair comparisons among clinicians, hospitals, and other health care organizations.2 Two health care trends have likely spurred reexamination of this policy. The first reflects a trend toward measuring patient outcomes (eg, hospital readmissions) rather than care processes (eg, smoking cessation counseling).2,3 Process measures largely under hospital or clinician control, such as drug administration during hospitalization, are little influenced by social determinants of health (ie, patient life circumstances, constraints, and community resources). In contrast, outcomes such as patient hospital readmission ormortality aremore strongly influenced by “social risk.” Ignoring these effects can produce misleading conclusions about comparative performance.4,5 Specifically, clinicians and hospitals serving disadvantaged patients may appear worse on quality measures than they really are, and those serving more affluent patients may appear better than they really are. The second trend ties payment to performance for hospitals, clinicians, and other entities, eg, accountable care organizations (ACOs).3,6 This trend creates the potential for greater financial penalties for hospitals, clinicians, and ACOs caring for socially disadvantaged patient populations. The 2006 policy could have the paradoxical effect of exacerbating health care disparities by depriving safety-net hospitals and clinicians of resources needed to provide quality care, creating the public appearance that hospitals and clinicians serving disadvantaged patients are low quality, and as a consequence, generating perverse incentives to avoid serving disadvantaged patients or communities.7 In response to these concerns about fairness, the NQFundercontract fromCMSconvenedanexpertpanel to reconsider its policy. The panel considered the questions of whether, when, and how to adjust measures used for “accountability applications” (public reporting of quality and pay-for-performance) for sociodemographic risk factors. The panel sought to balance possible unintended consequences of adjustment, includingperceivedmaskingofdisparitiesandthepotential for creation of lower standards of quality for socially disadvantaged patients, against the unintended consequences of not adjusting for socioeconomic status.