To Advance Health Equity, We Must Bridge Gaps in Health Data and Measurement

The COVID-19 pandemic has heightened public awareness of racial inequities. The available data are undisputable and paint a stark reality: the disproportionately higher rate of hospitalizations and deaths from COVID-19 among people of color. As of November 10, 2020, the age-adjusted mortality from COVID-19 among Black Americans was three times higher than that of white Americans. If mortality rates had been the same across races, about 21,200 Black, 10,000 Latino, 1,000 Indigenous, and 70 Pacific Islander Americans would still be alive.

The pandemic has also underscored well-recognized and persistent data challenges that pose significant barriers to both assessing and addressing inequities in health and health care quality. Perhaps it’s time for a reminder that quality of care, as defined by the Institute of Medicine in its seminal 2001 report, Crossing the Quality Chasm, includes six domains: safety, effectiveness, patient-centeredness, timeliness, efficiency, and—not least—equity.

What we choose to measure (or not to measure) has significant implications on the story we can tell about the health of communities. This moment forces us to ask several crucial questions: Are current quality measures up to the task? Do they provide an accurate, non-biased picture of the effect of our health care system on people’s lives? Do they allow us to assess the six domains of quality? Especially, do they enable us to fully evaluate equity and identify where and why inequities exist in health and health care? 

Over the past 30 years, the U.S. health care system has created an impressive “quality measurement enterprise” with standardized methods to achieve scientific accountability as well as the development of over 2,000 measures and reporting programs. But despite decades of attention, quality remains uneven. Widespread variation is well known, although not well understood. The COVID-19 pandemic and current social justice movement beckon us to apply an equity lens to how we measure quality, and to focus on and address the many challenges that persist. Among them: 

  1. Gaps in data collection, especially the use of surrogate data for race and ethnicity
  2. Risk-adjustment complexities, especially when information about race and ethnicity is lacking or not accounted for appropriately
  3. Continued reliance on process over outcome measures
  4. Limited use of standardized measures of patient self-reporting
  5. Inadequacies in quality measures that have not kept pace with and are no longer relevant to innovations in care delivery  

Missing Data Are a Barrier to Assessing and Achieving Health Equity 

Even in the midst of a national and global pandemic, a time of crisis when the need for reliable and timely information is imperative, we’ve failed to gather the right types of data. The result is an incomplete story, especially when it comes to assessing the depth of health inequities that are pervasive in the U.S., at both community and individual levels. 

Between January and the end of April 2020, population-level data about Sars-Cov19 infections were collected and reported to the Centers for Disease Control and Prevention (CDC) from 1,802,416 people. However, data on race and ethnicity were only available for 866,693 people (48.1 percent). Data on emergency department (ED) visits did not include any information about race.

By May 2020, the CDC had instituted new data processes, working with states to provide more comprehensive information on race and ethnicity for reported COVID-positive cases. Since then, the proportion of reported cases that include race and ethnicity data has increased and provides a better picture of COVID trends and the disparate impact at the population level.

More detailed information about person-level inequities within a larger group is essential to our clinical understanding of individual risk factors and to develop the scientific evidence about people’s responses to treatments and their outcomes. Gaps in race and ethnicity data exist here too. In 2015, one third of commercial plans, half of Medicaid plans, and nearly three quarters of Medicare plans reported incomplete or partially complete claims documentation of race and ethnicity. These gaps are troubling—claims remain the predominant source of health care data.

At the peak of the pandemic, claims-based utilization data showed a significant drop in ED visits. The number of patients presenting with strokes, heart attacks, or other urgent health problems had plummeted. Were patients avoiding the hospital for fear of being infected? What might be the implications of forgoing needed care? Our understanding of this phenomenon is still evolving. Issues of equity are likely related to these observations. Race and ethnicity data will be essential to get answers and to address the contributing factors accordingly.

In addition to administrative claims, electronic medical records (EMR) are a promising source of rich and detailed information about individual patients and can help bolster efforts to advance equity.  

An analysis from an integrated health system in northern California provided person-level data on the outcomes of 16,957 patients who tested positive for the coronavirus. A patient’s age, gender, self-reported race and ethnicity, housing status, insurance type, and important clinical risk factors were accessed via the EMR. Non-Hispanic African American patients had 2.7 times the chance of hospitalization (after adjustment for age, sex, comorbidities, and income) compared to non-Hispanic white patients, suggesting an independent effect of systemic racism as well as the toll of lifelong toxic stress.    
And even in this analysis, data were incomplete: Race and ethnicity data were available for only 40 percent of the 16,957 people who tested positive for the coronavirus. That health system was subsequently successful at increasing race and ethnicity information to 67 percent, within one month.

Beyond population and person-level data, data intersectionality—the ability to link the information gathered from different databases to one individual—is an essential tool to fully assess the variation and impact of various factors on health outcomes. Linking databases is extremely challenging, yet some states are making great progress in this regard. 

Bridging the Health Data and Measurement Gaps to Advance Health Equity 

How do we respond to the gaps in our quality of care databases highlighted by the COVID-19 pandemic? And how can we do so on local, regional, and national levels?    

The good news is that standards for collection of race and ethnic data have been defined by the Office of Management and Budget and by the Office of Minority Health at the U.S. Department of Health and Human Services. Section 4302 of the Affordable Care Act requires the standardization, collection, analysis, and reporting of race, ethnicity, and language proficiency data in all national surveys and when health care services are provided. New York State also collects race and ethnicity data through numerous agencies.

Standards and regulations are necessary first steps, but they are not enough. Creating the health data system that will move us closer to health equity requires a systemic approach. The first step is to set measurable equity goals at all levels of the health system. These goals would encompass national and state policies, more equitable distribution of health care resources locally, and the provision of care to individuals. We also need an accountability mechanism to assess progress, and various incentive levers to motivate action toward goals.  

This will not be easy work. We need to prioritize data integrity, the building block to measuring quality and to identifying the existence and magnitude of inequities. Collecting data that can be trusted is a fundamental first step. At the same time, structural racism has shaped our institutions over hundreds of years, and health care is beginning to reckon with its roots. When it comes to data, we can’t ignore the mistrust and fear about how race and ethnicity information will be used. There may be good reasons for data gaps—people preferring not to report their race, for instance.  

We must learn from communities and organizations that have been working on the forefront of health equity and have successfully deployed an equity strategy, including engagement with the C-suite, staff, and people they serve. To advance a vision of health equity, transparent, reliable, and publicly reported information must inform community engagement and strategy development. Only trusted data can engage stakeholders around the evidence of health inequities and guide corrective and collective actions. 

The tragic toll of COVID-19 is proof that, when it comes to information we can trust, the stakes could not be higher. Such information is critical to rectifying the racial injustice built into all U.S. institutions, including our health system. To move forward, we must persevere to bridge the information gaps.

Anne-Marie J. Audet, MD, MSc, is United Hospital Fund's senior medical officer and oversees the organization's work on patient-centered care. 

United Hospital Fund has a long history of bringing together diverse perspectives to address critical challenges in health care in New York. In the current crisis, it’s more important than ever to hear from all parts of the health care system. Today’s commentary from UHF's senior medical officer, Anne-Marie J. Audet, MD, MSc, looks at how the pandemic has underscored long-standing data issues that pose significant barriers to both measuring and addressing inequities in health and health care quality. – UHF President Tony Shih

Dec. 16, 2020