In recent years managed health care principles developed in the private sector have increasingly been applied in public-sector mental health and addiction service systems. At the state level, this change in health care management has consistently reduced costs, but it has had mixed results for access and quality of care for severely disabled individuals (1,2).
In 1995 the Department of Veterans Affairs (VA) also began to transform its vast health care system along the lines of managed care (3). The health care system was decentralized into 22 regional service networks, each charged with providing an integrated continuum of care to its population of veterans (4). Furthermore, service networks were given a mandate and the autonomy to shift the focus of services from hospital-based inpatient care to outpatient and primary care.
The study described here used program data from before and after the shift to outpatient care to assess the effects of these new policies on VA's national substance abuse treatment system, which is the nation's largest.
The Program Evaluation and Resource Center in Palo Alto, California, has been tracking VA's substance abuse treatment system since 1990 (5). As part of this work, the center mailed surveys to all VA substance abuse treatment program directors in October of 1990, 1991, 1994, and 1997. Programs were identified at each wave by telephone surveys, using several criteria. The program had to have at least two full-time-equivalent staff, to identify substance abuse treatment as its primary purpose, and to provide services beyond detoxification.
Program directors completed surveys that covered the program's setting, structure, services, treatment orientation, staffing, and patients. Directors who did not initially complete the surveys were recontacted by center staff until a response rate of 100 percent was attained.
F1 presents data on VA substance abuse treatment programs for survey years 1990 to 1997. The number of programs increased steadily from 1990 to 1994 due to a $100 million substance abuse programming enhancement initiative developed and implemented by the U.S. Congress, the Department of Veterans Affairs, and the Office of National Drug Control Policy (5). From 1990 to 1994 the total number of programs grew from 273 to 389, an increase of 42.5 percent. Outpatient programs had the greatest growth rate, 80.6 percent, an increase of 98 to 177 programs. However, the system's growth was halted in 1995 by decentralization and also by pressure to reduce the size of government after the 1994 congressional elections (5). From 1994 to 1997, the number of programs declined by 22 percent, from 389 to 304.
When managed care principles have been introduced in other public mental health and substance abuse systems, the most notable effect has been a decline in inpatient services (1,2,3,4,5,6). The 63 percent decrease in VA inpatient programs from 1994 to 1997 is the most remarkable change attributable to decentralization and the mandate to convert from hospital-based to outpatient-based services. Most of the inpatient programs that survived were focused on treating substance abuse patients who had comorbid psychiatric conditions.
Due to the closure of inpatient programs, inpatient beds decreased from 5,920 in 1994 to 2,960 in 1997, a 50 percent decrease. The loss of inpatient capacity was partly offset by an increase in the number of residential programs, from 32 to 57 (78 percent), and by a concomitant increase in residential beds, from 1,190 to 1,750 (47.1 percent).
The average length of stay for both inpatient and residential programs decreased between 1994 and 1997. The average length of stay in inpatient programs decreased from 24 to 20 days, and in residential programs it decreased from 110 to 65 days. The average number of annual admissions to inpatient programs decreased 25 percent from 1994 to 1997, whereas the average number of admissions to residential programs increased 77 percent over the same period.
Despite the growth in residential programs, the proportion of residential programs that had patients on waiting lists for treatment on the last day of the fiscal year increased from 58 percent in 1994 to 74 percent in 1997. The proportion of inpatient programs with patients waiting for care stayed constant and high at 68 percent.
F1 shows a steady increase in the number of outpatient programs from 1990 to 1994. Although the growth in the number of outpatient programs virtually stopped after 1994, the intensity of outpatient services continued to expand. In 1994 only 11 percent of VA outpatient programs provided four or more hours of care per treatment day to the average patient—for example, in day hospitals and intensive outpatient programs. By 1997 a total of 48 percent of VA outpatient programs met this standard. It appears that while networks exercised their mandate to reduce inpatient services, they frequently converted inpatient programs to day hospitals rather than eliminating them outright. Thus from 1994 to 1997, the number of outpatient programs classified as intensive increased 347 percent, from 19 to 85 programs.
Consistent with the growth in programs, systemwide staffing grew from 4,147 full-time-equivalent (FTE) employees in 1991 to 4,733 in 1994. By 1997 the number had declined to 3,326 FTE employees. The proportion of staff working in outpatient programs increased from 23 percent of all FTE employees in 1994 to 41 percent of all FTE employees in 1997. The only notable shift in type of staff was an increase in the proportion of nursing staff from 1994 to 1997—especially in intensive outpatient programs, with a 10 percent increase, and in residential programs, with a 16 percent increase.
Throughout the 1990s VA substance abuse patients have experienced a steady increase in the severity of their clinical and social problems. For example, from 1990 to 1997 the proportion of patients who were married or in a marriage-like relationship declined from 48 percent to 32 percent in inpatient programs, from 39 percent to 23 percent in residential programs, and from 55 percent to 36 percent in outpatient programs.
Similarly, from 1991 to 1997 the proportion of patients with a comorbid psychiatric diagnosis increased among patients in inpatient programs from 30 percent to 47 percent, in residential programs from 26 percent to 33 percent, and in outpatient programs from 32 percent to 42 percent.
This increase in social isolation and psychiatric problems appears to have been steady throughout the various policy trends in services. Several factors outside of treatment systems themselves seem to be contributing to increases in the problems of patients in public-sector substance abuse treatment programs, raising the issue of whether a rapidly changing service system will be able to respond to the needs of an increasingly troubled patient population.
VA is sometimes perceived as an overly centralized organization that implements changes at a glacial pace (7). The data presented here belie that stereotype. The changes in the substance abuse treatment system since 1990 are nothing short of dramatic. To the extent that the mandate to emphasize outpatient care and the decentralization of the system were intended to increase the ability of VA facilities to alter the services they offered, these measures clearly were successful. However, the more important questions are how access, outcomes, and costs of care were affected by the changes.
In terms of access, the proportion of programs with waiting lists indicates that despite decreased lengths of stay, more patients desire care at inpatient and residential programs than VA has space for at any given time. Waiting lists are a particular point of concern in VA, because many patients live a significant distance from a facility and thus find it difficult to keep regular outpatient appointments. Furthermore, a substantial proportion of VA substance abuse patients lack stable housing.
To preserve access to care, VA should continue to expand residential services such as community residential facilities (8), domiciliaries, and residential rehabilitation programs. These settings can be used to provide care and also to house homeless patients who are currently receiving treatment in outpatient programs such as day hospitals. Because residential beds are cheaper than inpatient beds on a per diem basis, VA should be able to expand residential beds within current budget parameters.
What will be the effect of the shift from inpatient to outpatient care on treatment outcomes of veterans with substance use disorders? Patients who are clinically eligible for care in either setting appear to have similar outcomes in outpatient and inpatient programs (9), but we currently do not know whether more severely impaired patients, such as those with dual diagnoses and socially isolated patients, benefit equally from, for example, an intensive outpatient program and an inpatient program. In this sense, VA is conducting a national experiment, and outcome data will be required to assess its ultimate impact. Fortunately, because the Program Evaluation and Resource Center has archived data on how large national samples of VA substance abuse patients have fared in inpatient programs, we will have a baseline to judge whether the new system produces comparable, better, or worse outcomes for patients in general and for more severely impaired patients in particular.
Because of the decline in the number of substance abuse programs and staff and the sharp reduction in acute inpatient beds, direct costs for specialized substance abuse services clearly were reduced. However, there is some evidence that closing specialty substance abuse programs may result in substance-dependent patients' receiving more care in medical-surgical and general psychiatry units (10). Therefore, the center is currently evaluating whether decreasing specialty substance abuse services truly reduces costs or merely shifts them to other VA and non-VA programs.
We emphasize that VA is a public system with a public mission: to care for the nation's veterans, including those with severe illnesses and few social and economic resources. Thus, although VA may adopt some principles of managed care and mirror the private-for-profit sector in some respects, its ultimate success must be measured by veterans' access to and outcomes of care. In the coming years, the analysis of national evaluation data being gathered by the Program Evaluation and Resource Center and other centers will be critical to assessing whether VA is living up to its mission of providing accessible, integrated, and cost-effective services to veterans with substance use disorders.
Preparation of this paper was supported by the VA Mental Health Strategic Health Group and Health Services Research and Development Service. Opinions expressed in this paper do not necessarily reflect official VA policy positions.
The authors are affiliated with the Program Evaluation and Resource Center at the VA Palo Alto Health Care System in Menlo Park, California, and the Mental Health Strategic Health Group at VA headquarters in Washington, D.C. Send correspondence to Dr. Humphreys at the VA Palo Alto Health Care System, Menlo Park Division, 795 Willow Road (152-MPD), Menlo Park, California 94025 (e-mail, email@example.com). Richard J. Frances, M.D., is editor of this column.