The Future of Research on Alcohol Health Disparities: A Health Services Research Perspective

Dear Editor:

In the previous issue of the Journal of Studies on Alcohol and Drugs, Zemore and colleagues (2018) describe a timely set of research priorities for studying and alleviating alcohol-related health disparities across race and ethnicity in the United States. While doing so, they raise the important point that we must study how to address alcohol-related health disparities from the perspective of many service settings— not just health care.

Indeed, limiting alcohol-related quality improvement efforts to health care settings could increase population-level health disparities because certain racial/ethnic groups are underrepresented in health care in various ways (Mulia et al., 2014; Zemore et al., 2018). The paucity of research on alcohol interventions in community and faith-based settings at best represents a missed opportunity. Innovative, culturally informed alcohol interventions need to be developed and implemented beyond the health care setting. Robust examples of such interventions exist and must be prioritized for future study (Allen et al., 2014; Ornelas et al., 2015).

Notwithstanding these issues, research to reduce alcohol-related health disparities in health care remains crucial because most persons do attend health care visits. According to the Behavioral Risk Factor Surveillance System, in 2016, roughly four in five people in all racial and ethnic groups in the United States attended a health checkup within the past 2 years: 79.1% of American Indian or Alaskan Natives, 86.1% of Asians, 89.4% of Black or African Americans, 83.9% of Native Hawaiian or Pacific Islanders, 83.8% of Whites, 81.3% of those of other race, 77.9% of those who stated no preferred race, and 85.4% of persons who reported Hispanic or Latino ethnicity (Centers for Disease Control and Prevention, 2018). Individuals who are not seen for routine care may obtain acute care where alcohol-related interventions may be provided (e.g., emergency departments, hospital units; Blow et al., 2017; Busch et al., 2017). Research is needed to ensure that alcohol-related care in these settings is delivered equitably and in culturally informed ways. Furthermore, research has demonstrated that new clinical roles and electronic tools can be implemented in health care to identify and address factors that affect health outside of health care settings (Gold et al., 2017; Hsu et al., 2016). Strategic plans to address alcohol-related health disparities on the population level must heavily emphasize a “both/and” approach—improving care provided to health disparities populations in health care systems and developing and implementing care in other venues.

To complement Zemore and colleagues’ (2018) plan of attack, below we note how investigators can address important research priorities to understand and alleviate alcohol-related disparities in the context of health care settings and data. Investigating disparities in access to and use of alcohol treatment and support services, comparing the efficacy of interventions, and monitoring disparities are a few areas noted by Zemore and colleagues (2018) that can benefit from this type of health services research.

Health care data can facilitate the study of health disparities research questions that are often deemed infeasible because of the rare occurrence of key health and health care outcomes (e.g., alcohol-related liver disease and death, receipt of pharmacotherapy and specialty care). Health data offer large sample sizes, enabling comparison of rare outcomes across smaller racial/ethnic groups and the ability to test effect modification by factors such as gender, when indicated (Glass et al., 2017b). When primary data collection is required, researchers can identify participants (Sterling et al., 2015) and oversample smaller racial/ethnic groups with health care data. Moreover, health care data facilitate longitudinal research. Exposure–outcome relationships can be observed prospectively across data-collection periods that are shorter or longer than what is common in existing survey data or that are prohibitively expensive in primary data collection. Fortunately, data from large health care systems, or combined data from multiple health systems sharing a common data model, provide a means to address such needs. For instance, the Department of Veterans Affairs has approximately 5 million enrollees and has proven to be ripe for research on disparities in alcohol-related care (e.g., Williams et al., 2017a). The Health Care Systems Research Network comprises 19 health care systems that share a common data model, serving more than 13 million patients within their integrated delivery systems (Health Care Systems Research Network, 2017). These data can be geocoded and linked to aggregated census data (Nelson et al., 2017) or linked via identifiers to other individual-level external sources (e.g., to ascertain suicides, overdoses; Bohnert et al., 2013; Simon et al., 2016).

Importantly, analyses of health care data can help influence clinicians and health system leaders to change care processes for the better to reduce health disparities. Electronic health record data are ideal for addressing quality improvement-related research questions because they are the same data that clinicians use to make day-to-day decisions. Consequently, data for research are routinely collected, longitudinal, and meaningful to stakeholders. Process measures (e.g., which interventions received when and where) can be used to monitor disparities to inform how care can be improved to reduce disparities. The National Institutes of Health is actively pursuing this type of health disparities research (e.g., PAR-18-286). With electronic health record data, outcomes of interventions can be compared retrospectively or prospectively, allowing researchers to monitor practice changes that have occurred in the past and to manipulate practice changes to see what happens in the future (Richesson et al., 2017; Williams et al., 2017b). Interrupted time series analysis of health system data is a powerful method to evaluate historical changes, and stepped-wedge and cluster-randomized designs can rigorously evaluate the effect of planned changes across many clinics in a larger health care system (Bobb et al., 2017; Thakral et al., 2017). Health care data offer the most efficient and accurate resource to inform development of conceptual frameworks and interventions to reduce disparities within health care.

Thus, although health care access remains a clear challenge with disproportionate access barriers in health disparities populations, health care systems have played and will continue to play a key role in disparities research. This role may increase as health systems take on new forms of addiction-related care. For instance, there have been calls to integrate more addiction care into primary care given the lack of efficacy of referral to treatment and the relatively weak effectiveness of alcohol screening and brief intervention (Bradley & Kivlahan, 2014; Glass et al., 2017a; McCambridge & Saitz, 2017). Equity in the provision of these services can be evaluated and promoted through health services research (Rieckmann et al., 2017). Partnered research via learning health system principles can accelerate the speed at which health disparities research findings can be put into practice (Greene et al., 2012).