The nation's health care safety net plays a critical role in providing care to millions of Americans who have few options for accessing important health care services. Public hospitals are in a unique position within the health care safety net because they often operate systems of care, including inpatient care, primary and specialty outpatient care, and emergency departments. They therefore can offer patients with chronic conditions coordinated care throughout their life cycles.
In an effort to improve care for patients with chronic conditions, seven large and diverse public hospitals joined with the NPHHI to form the Consortium for Quality Improvement in Safety Net Hospitals and Health Systems. Choosing diabetes as their initial focus because of its prevalence among their patients, the consortium members aimed to learn about the quality of diabetes care in public hospitals, describe mechanisms for delivering diabetes management services, and identify lessons that might apply to other chronic conditions. When the study began, 41.4 percent of the patients with diabetes at these hospitals were uninsured and an additional 15.5 percent were covered by Medicaid.
Consortium hospital systems provided information on laboratory tests commonly given to patients with diabetes to monitor their care over time. These tests record HbA1C values and lipid levels (low-density lipoprotein, or LDL, and total cholesterol). Project staff then compared these values to those found in four other studies of diabetes patients: one study of patients who received care through the Department of Veterans Affairs (VA); another of patients enrolled in commercial managed care plans; and two nationally representative government surveys that provide averages for adults with diabetes in the United States.
Despite the vulnerability of public hospital patients—as evidenced by the high number of uninsured and publicly insured individuals—outcomes for consortium patients are comparable to or, in some cases, better than national averages on standard measures of diabetes management, although they fall short of diabetes-related outcomes for VA patients. At consortium hospitals and managed care plans, 65 and 67 percent of patients, respectively, had HbA1c values below 8.5 percent, indicating moderate control of diabetes. A higher percentage of VA patients had controlled diabetes. At consortium hospital systems, in national surveys, and in managed care populations, about one-fifth (21%, 18%, and 20%, respectively) of patients had HbA1C values of 9.5 percent or above, indicating very poor glycemic control. This figure was much lower for VA patients: only 8 percent of patients with diabetes in the VA study showed very poor glycemic control.
In addition, consortium patients had similar or better cholesterol levels than patients in the national samples.
These findings may help dispel the perception that patients at public hospital systems receive only episodic care that does not effectively manage long-term chronic conditions. The consortium outcomes are equal to and, in some cases, better than privately covered managed care plan enrollees and the general population. Clearly, public hospitals are doing something right for patients with diabetes. Nevertheless, outcomes in public hospitals, managed care plans, and the general population are far below outcomes for VA patients on key measures of quality. The VA is a large public system with rigorous protocols for quality improvement. Its success in managing patients with diabetes should be studied and, whenever possible, models should be spread to public hospitals and other providers.Outcomes by Race/Ethnicity and Coverage
There are important differences in health outcomes across subgroups of the NPHHI consortium hospital systems. After controlling for insurance, gender, age, number of outpatient visits, and hospital characteristics, we found significant variation across patient groups by race. For example, compared with white patients, black patients at consortium hospitals were significantly less likely to have well-controlled diabetes (OR=.46).
Insurance coverage also was a significant factor in glycemic control, independent of race, age, gender, hospital characteristics, and outpatient utilization. Medicare enrollees were more likely than Medicaid enrollees to have better diabetes control (OR=1.18), and uninsured patients were significantly less likely than those with any type of insurance coverage to have well-controlled diabetes (OR=.85). Even at safety net hospitals, with services designed to provide access to high-quality care for low-income and other vulnerable patients, insurance and race continue to influence health care utilization and ultimately health outcomes for patients with diabetes.Conclusions
The consortium project provides important lessons for safety net hospitals and other providers of diabetes services.
- Continuity of care and a range of services are essential. The consortium hospital systems employ various strategies such as education classes, special diabetes clinics, and language services to improve care for at least a subset of their diabetes patients.
- Information management systems are central to providing comprehensive diabetes care. Six of the consortium participants were able to provide substantial race/ethnicity data on their diabetic patients and half were able to provide detailed data on patients' preferred languages. Such data enable safety net hospitals to compare chronic care utilization and outcomes by racial and ethnic groups—an essential first step in addressing disparities.
- Insurance coverage influences patient outcomes even when there are programs to mitigate financial barriers to care. At consortium sites, uninsured individuals receive less care, including critically important outpatient care, than patients covered by Medicaid or Medicare.