Populations at Risk: A New Section for JGIM

Physicians and other health care providers want to assure the best possible health outcomes for their patients. However, we do not work in a vacuum and often find ourselves in a conundrum.^1–3 Health outcomes are influenced by the political, economic, and organizational environments in which we practice, teach, and do research, as well as by our own knowledge, skills, and attitudes, and the socioeconomic circumstances, health beliefs, and community characteristics of our patients. Disparities in the receipt of health care services by race/ethnicity, socioeconomic status, and gender have all been well described⁴ and the health risks of impaired access to care have been well documented.⁵ Federal efforts to eliminate racial and ethnic disparities in health are drawing attention to the need to move beyond descriptive studies and to identify effective interventions to address these challenges.⁶ In this issue of JGIM, we announce a new section, Populations at Risk, aimed at enhancing our understanding of these dynamics, and building the evidence base to improve the quality of care delivered to all patients.

In Populations at Risk we will publish original articles, brief reports, and review articles as well as case studies describing efforts—positive and negative—to improve care delivery for “populations at risk,” defined broadly. We include groups who are at risk for suboptimal health outcomes by virtue of their sociodemographic characteristics (gender, race, ethnicity, age, income, geographic area of residence, education, or sexual preference), health insurance status, functional status (disability), or diagnosis (HIV infection, mental illness, or substance abuse). JGIM has always published and will continue to publish many articles on these groups in all of its sections. Our objective here is to highlight articles that contribute to understanding and eliminating disparities in access to care, quality of care, and health outcomes among high-risk populations.

The two articles in this issue that launch Populations at Risk ask important questions and set the grounds for future work. Sarver and Baker found that Spanish-speaking patients who did not speak the same language as the emergency department physician who saw them were less likely to have a follow-up appointment recommended, whether or not an interpreter was used during the visit.⁷ Close to 90% of the interpreters used were “ad hoc” emergency room staff or family members, which raises questions about the quality of interpretation received. However, when a follow-up appointment was recommended, patients with language barriers were equally likely to complete it as those without language barriers.⁷ Prior analyses of this survey showed that language concordance and interpreter use influenced patients' perceived understanding of their condition⁸ and satisfaction with the patient–provider relationship.⁹

Of note, even when a follow-up appointment was recommended, lapses in communication of this information to the patient were evident, independent of language barriers. Overall, 20% of patients were not aware a follow-up appointment had been recommended. This did not vary by language concordance, as Spanish-speaking nurses conducted exit counseling. Knowledge of appointment did vary by type of appointment given, with those who were told to call for the next available appointment in primary care faring the worst. Forty percent of these patients were unaware of this recommendation. Sixty percent of patients completed recommended follow-up visits within 1 month. One wonders how many of the rest encountered difficulty in scheduling a follow-up appointment in the public hospital that was the site of the study.

In a survey of homeless women, an exceedingly vulnerable group, Wenzel and colleagues found that 1 in 8 reported having been raped in the last year.¹⁰ Half of these women were raped more than once. The women who reported rape fared worse on every physical and mental health measure used in this study.¹⁰ This is consistent with findings that women veterans who reported sexual abuse reported more physical symptoms and medical conditions.¹¹ These women are at risk for HIV infection, sexually transmitted disease, and post-traumatic stress disorder. In addition, 20% of women who reported rape stated that they desired treatment for substance abuse, but were unable to obtain it, highlighting the need to coordinate services.

Notwithstanding limitations, these articles suggest that there is much we can do at the encounter, practice, and organizational level that can potentially serve to improve care for high-risk populations. For example, when seeing a patient whose language we do not speak, we can be cognizant of how this barrier may impact the care delivered, and assess the quality of interpretation used. We can enhance our skills in the bilingual interview and examine any personal biases we may harbor to individuals whose backgrounds differ from our own. When seeing a homeless woman, it seems prudent to ask about a history of sexual assault, as recommended by Wenzel et al.¹⁰ We should also become knowledgeable about the range of services available in the community for these women. Nonetheless, these studies highlight how many questions remain unanswered and how much further research is needed to provide solutions. There is especially a need for longitudinal studies that assess outcomes over time and allow us to understand the interplay of multiple factors.

In the public hospital where I practiced for many years, we confronted these challenges daily. Residents of our catchment area came from 110 different countries and spoke over 50 different languages. Half of our patients were uninsured, and many had complex social needs that could not be addressed adequately in an office visit. Our Emergency Department was chronically overcrowded, and there were often long waits for primary care appointments for new patients. In response, we actively recruited a multilingual staff at all levels, provided formal interpreter training, and had an active language bank. We designated emergency follow-up slots in medical and nonmedical subspeciality clinics for use by the Emergency Department and assigned nurse practitioners and physician assistants for primary care follow-up, so that no patient who needed a follow-up appointment would leave the emergency room without one scheduled. We built formal relationships with community-based organizations that offered vital services that we could not provide for our patients. We even changed the color of the clinic card from green to gold so that patients wouldn't have to get anxious when asked for their “green card.” We did not have the capacity to formally evaluate the effectiveness of these interventions or to identify which elements could be generalizable to other settings. Growing interest in the need to eliminate health disparities and enhance cultural competence in practice by federal and private funders will likely lead to more opportunities for investigators to study these issues.

No matter how diligent and industrious we are at addressing these concerns at a local level, the ability to accomplish these objectives will be influenced by a number of additional factors. Federal, state, and local health policy, and market forces all influence practice experience. “Nonhealth” policies, such as education, housing, and welfare can all have a significant influence on health. Well-intentioned policies may have unintended consequences when implemented.¹² Policy results from political compromise and incremental changes, as well as from analysis of existing evidence. In addition, the interaction between national, state, and local policies may not be predicted easily. Policy inaction can also have adverse consequences, such as the current societal failure to provide access for ever-growing numbers of the uninsured. At a time when physicians are increasingly held accountable for the outcomes of the care they provide, it is critical to understand how these nonclinical factors influence health outcomes. Good research can contribute to our understanding of the interplay between the multiple factors that contribute to health disparities and guide the development of evidence-based interventions to address them.

Clearly, a wide range of subject areas would be of interest for publication in this new section, including the following examples:

Primary care physicians know all too well the human consequences of the failings of our health system. Readers of JGIM play a critical role in providing care for high-risk populations, as well as designing, implementing, and evaluating strategies to improve their access to care and health outcomes.¹³ We encourage you to submit to us your best work in these areas. Our Editor noted that it is not sufficient merely to publish these papers, but that we need to think about ways to maximize their impact. We enlist your help and welcome your suggestions in accomplishing this. We hope to stimulate dialogue, debate, and action.—Arlene S. Bierman, MD,Associate Editor.