Performance measurement and equity

Abstract

To maximise benefits and minimise harm, equity must be built in from the start

Performance measurement is now a reality for clinicians around the world. It involves measuring and monitoring quality of care using standardised indicators. Shortcomings in the quality of care—the gap between what we know and what we do—are well documented.¹ So too are inequities in access, quality, and outcomes linked to gender, ethnic origin, and socioeconomic status.² Recognition of substandard and uneven quality of care has fuelled calls for providers to be more publicly accountable and for health systems to change.

Interest is growing in performance measurement as a way to drive improvements in health care. In this week's BMJ, McDonald and colleagues describe an ethnographic case study in which two English general practices changed their organisation to achieve high performance scores under the quality and outcomes framework.³ The quality and outcomes framework, and other high profile measurement and reporting efforts such as those in the US Veterans' Health Administration, have met with some early success.^{4 5} Adding to this enthusiasm is a recent study that attributes declining mortality from acute coronary syndromes and heart failure—two conditions in which performance measurement has been widely used—to increased use of evidence based treatments.⁶ However, optimism about potential benefits is tempered by growing concerns about potential harms.⁷

McDonald and colleagues were particularly concerned with adverse effects on practitioners' clinical autonomy and motivation. However, they found that incentives were mostly aligned with professional values about optimising quality of care. What the study does not tell us, though, is how these organisational changes were perceived by patients or what impact they had on patients from different communities in different practice settings.

Socially disadvantaged patients may stand to benefit most from structured efforts to measure and improve quality, as they often experience the largest quality gaps. Importantly, however, they may also be at greatest risk of harm.⁸ Equity is a major dimension of healthcare quality and a key attribute of high performing health systems,⁹ so initiatives to improve quality will be incomplete unless inequities are reduced as performance improves. Performance measurement and quality improvement alone will not result in more equitable systems of care.

Interventions to improve quality can impact on health inequities in three ways: they may narrow, maintain, or widen existing inequities, depending on their relative effectiveness in different groups of people and how they deal with the root causes of inequity. In a randomised controlled trial, a complex intervention designed to improve the quality of primary care for depression reduced disparities by improving health outcomes and unmet need significantly more among Latinos and African Americans than among whites.¹⁰ A longitudinal study examined the impact of performance measurement for patients with end stage renal disease insured by Medicare in the United States. It found that racial and gender disparities were reduced in relation to the adequacy of haemodialysis but were unchanged for the management of anaemia and nutritional status.¹¹ A retrospective analysis of performance data from Medicare managed care showed steady improvement over many years, with narrowing of disparities in process indicators. Control of glucose and cholesterol improved in both white patients and black patients.¹² However, racial disparities in outcome measures widened because the improvements were greater for white patients. This shows that it is more difficult to improve outcomes than processes of care for disadvantaged populations.

If we are to identify persistent disparities between populations that will otherwise be masked by overall gains in quality, we need performance measures that are stratified by sex, ethnic origin, or socioeconomic status. In Canada, the project for an Ontario women's health evidence based report card (POWER) is developing explicit methods for assessing equity as a routine part of performance measurement (www.powerstudy.ca).

In the US and the UK, practices that serve socioeconomically disadvantaged patients have shown poorer performance on commonly used quality indicators than have practices serving more advantaged patients.^{13 14} Reporting these measures—particularly when pay is linked to performance—can inadvertently penalise providers who care for those most in need, creating perverse incentives to exclude these patients. Risk adjustment models sometimes include socioeconomic status, but these can also mask real disparities in quality. An “equity blind” approach cannot account for the non-clinical factors that influence health outcomes, and it may stop us learning which components reduce disparities and which do not. Equity oriented performance measurement takes these factors into account, and it can make systems and providers publicly accountable for the communities they serve.

Indeed, performance measurement can be a blessing, not a curse, for efforts to reduce inequities in quality. With adequate data, we can routinely measure and monitor progress, learn what tools and interventions work, develop and test new interventions to eliminate disparities, and understand a dimension of quality that has thus far seemed intractable. Ultimately, equity in health outcomes will probably be achieved only if we target the barriers that stop the providers serving disadvantaged patients and communities from reaching their quality targets. To investigate and eliminate disparities, we need to stratify performance data by the patients' sex, ethnic origin, and other socioeconomic variables. This will allow us to build an evidence base for implementing change that will maximise benefits and minimise harms. Equity must become an integral component of performance measurement.