Appendix C
Mental and Substance-Use Health Services for Veterans: Experience with Performance Evaluation in the Department of Veterans Affairs

Robert Rosenheck, MD

Director, VA Northeast Program Evaluation Center (NEPEC), West Haven, CT

Professor of Psychiatry, Public Health and at the Child Study Center Yale Medical School, New Haven, CT


August 2004


Prepared for the Institute of Medicine Committee on Crossing the Quality Chasm: Adaptation to Mental Health and Addictive Disorders


Acknowledgement: Paul Errera MD, Thomas Horvath MD, Laurent Lehmann MD, Mark Shelhorse MD, Mary Jansen PhD, Gay Koerber MA, William Van Stone MD, Robert Gresen PhD and Anthony Campinell PhD and the staff of the Strategic Healthcare Group for Mental Health in VA Central Office have provided invaluable support over many years. The staff of NEPEC is responsible for most of the work reported here (but not for the errors, which are my own), specifically project directors Mayur Desai PhD, Rani Desai PhD, Alan Fontana PhD, Greg Greenberg PhD, Wesley Kasprow PhD, Douglas Leslie PhD, Alvin Mares PhD, James McGuire PhD, Michale Neale PhD, Sandra Resnick PhD. Thanks also to Michael Sernyak MD. Special analyses of the 2001 Survey of Veterans and the Schizophrenia PORT survey for this report were completed by Greg Greenberg and Rani Desai.

Summary

As the largest integrated health and social welfare agency in the United States, the Department of Veterans Affairs is a unique and potentially informative setting in which to examine the challenges of mental health and substance use treatment services quality and performance management.



The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement



Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.

OCR for page 423
Improving the Quality of Health Care for Mental and Substance-Use Conditions Appendix C Mental and Substance-Use Health Services for Veterans: Experience with Performance Evaluation in the Department of Veterans Affairs Robert Rosenheck, MD Director, VA Northeast Program Evaluation Center (NEPEC), West Haven, CT Professor of Psychiatry, Public Health and at the Child Study Center Yale Medical School, New Haven, CT August 2004 Prepared for the Institute of Medicine Committee on Crossing the Quality Chasm: Adaptation to Mental Health and Addictive Disorders Acknowledgement: Paul Errera MD, Thomas Horvath MD, Laurent Lehmann MD, Mark Shelhorse MD, Mary Jansen PhD, Gay Koerber MA, William Van Stone MD, Robert Gresen PhD and Anthony Campinell PhD and the staff of the Strategic Healthcare Group for Mental Health in VA Central Office have provided invaluable support over many years. The staff of NEPEC is responsible for most of the work reported here (but not for the errors, which are my own), specifically project directors Mayur Desai PhD, Rani Desai PhD, Alan Fontana PhD, Greg Greenberg PhD, Wesley Kasprow PhD, Douglas Leslie PhD, Alvin Mares PhD, James McGuire PhD, Michale Neale PhD, Sandra Resnick PhD. Thanks also to Michael Sernyak MD. Special analyses of the 2001 Survey of Veterans and the Schizophrenia PORT survey for this report were completed by Greg Greenberg and Rani Desai. Summary As the largest integrated health and social welfare agency in the United States, the Department of Veterans Affairs is a unique and potentially informative setting in which to examine the challenges of mental health and substance use treatment services quality and performance management.

OCR for page 423
Improving the Quality of Health Care for Mental and Substance-Use Conditions U.S. Veterans Of 25 million U.S. veterans, 21% used Veterans Affairs (VA) services in the past year, and 2.7% used VA mental health or substance abuse (MH/SA) services. Although all veterans are now eligible for VA services, those most likely to use VA services receive VA income benefits are older, poorer, and less likely to have health insurance. Treatment of MH/SA in VA In 2003, 1.2 million veterans received a MH/SA diagnosis in VA, about 25% of all VA users. While they were a diagnostically mixed, Global Assessment of Functioning scores averaged 53, suggesting poor functioning, and 19% were dually diagnosed. VA is a cabinet-level agency with many important stakeholders. Concern about war-related Post Traumatic Stress Disorder (PTSD) and homelessness among veterans have given mental health issues greater prominence in the VA community in recent years. In 1995, a major reform was initiated which closed most MH/SA inpatient beds, nearly doubled outpatients treated, and emphasized accountability and performance measurement. Linkage of VA with the Department of Defense (DoD) and Other Mental Health, Medical, and Social Service Systems There has been great interest recently in smoothing the transition from DoD to VA, although the integration of information systems has yet to take place. Most VA patients get all of their MH/SA and medical services from VA. Although there has been concern that with extensive recent bed closures, VA patients would be forced to seek care in other health systems and might experience an increased risk of incarceration or suicide, empirical studies conducted thus far have not shown a significant increase in these problems. Development of MH/SA Quality Measurement and Quality Management in VA During the past 20 years there have been two notable phases in the development of VA MH/SA services. The first was initiated by the leader of mental health programs in VA central office from 1985–1994 and involved expansion of specialized mental health programs such as Assertive Community Treatment, homeless

OCR for page 423
Improving the Quality of Health Care for Mental and Substance-Use Conditions outreach, and transitional employment. The second was initiated in 1995 by the undersecretary for Health and brought changes in mental health service delivery as part of a major system-wide shift from operating as a hospital-based system of care to a community or population-based system of care. In both phases quality and performance measurement were crucial tools in guiding organizational change. A third phase, characterized by system-wide focus on building MH/SA quality is evolving in response to the recent report of the President’s New Freedom Commission on Mental Health. Quality of VA MH/SA Care The assessment of quality requires comparison of providers with standards or benchmarks, with risk adjustment for factors that may confound these comparisons. Evaluation of the quality of MH/SA care at VA facilities has been based on comparisons with: (1) VA system average performance, (2) VA performance in prior years, (3) the performance of other systems of care, and (4) comparison of care received by minorities with the majority population. Methods of quality measurement and benchmarking in VA are demonstrated for six aims highlighted in a previous Institute of Medicine (IOM) report: safety, effectiveness, person-centeredness, timeliness, efficiency, and equitability. Front-Line Experience Performance management in health care is sometimes experienced ambivalently by front line managers and clinicians. While they often feel empowered by access to data and find it allows them to improve the care they provide, there is also concern that measures are imperfect; that they do not take account of differences across facilities in case mix and in available community resources; that measures can be manipulated or “gamed,” resulting in unfair comparisons; and that managerial pressure to improve performance sometimes creates an atmosphere of personal criticism more than joint problem solving. Conclusion The complexity and uncertainty of the health care enterprise must be managed through comprehensive quality monitoring systems used by creative and committed leaders in competent organizations. VA has embraced this challenge.

OCR for page 423
Improving the Quality of Health Care for Mental and Substance-Use Conditions INTRODUCTION: THE DEPARTMENT OF VETERANS AFFAIRS IN AMERICAN MENTAL HEALTH CARE To bridge a chasm, one needs, at a minimum, a clear view of the terrain surrounding it. One can imagine building a bridge from one side of a chasm to the other, from both sides toward the middle, or even from a scaffolding erected on the floor in the center of the chasm out to both sides. But it is virtually impossible to imagine bridging a chasm if one were blocked from either viewing or accessing even one of its banks. The situation faced by those who would seek to bridge the many quality chasms in mental health care in the United States is in many respects like that of an engineer bridging a complex system of chasms with access to only one of its banks. People with serious mental illness often have needs for diverse services including psychiatric care, substance abuse care, primary and specialty medical care, and numerous social services including income supports, employment, education, and housing assistance, as well as help negotiating with the criminal justice system. And yet each of these needs is addressed by a different set of agencies at different levels of government. The advantage of this decentralized approach is that it increases local control, responsiveness, and flexibility (Peterson, 1995; Smith and Lipsky, 1993). The disadvantage is that agencies tend to compete for sources of funding, carefully guard their independence, and are often wary of sharing information on individual clients, let alone releasing systematic data on their overall operation. Mental health system engineers in America thus often find themselves trying to bridge system chasms while only being able to obtain information on the small patch of ground under their own feet. The U.S. Department of Veterans Affairs (VA) is a notable exception to this pattern. In most areas of U.S. social or health care policy, programs are operated at state or local levels by private or nonprofit providers, and the role of the federal government is limited. In contrast, the national government takes direct responsibility for providing comprehensive, lifelong, medical and social services to Veterans of the Armed Forces. National defense is the least contested area of federal dominance in the American system of government; as a result, the federal government has been given responsibility for the health and social welfare of military personnel, both on active duty and, for an increasingly large segment of veterans, after their period of military service is over. The VA thus represents the unusual case in which one agency accepts responsibility, at the national level, for providing comprehensive long-term care for a well-defined segment of the population. Mental health care provided by the VA may thus offer a uniquely informative, if atypical, opportunity to examine mental health performance monitoring and management in the American context. Taking a broad view, this presentation will: (1) describe the veteran population of the United States, compare the mental health needs (includ-

OCR for page 423
Improving the Quality of Health Care for Mental and Substance-Use Conditions ing substance use treatment needs) of veterans and those of other Americans, and compare veterans who use VA services with both nonveterans and other veterans who do not. It will then (2) present an overview of the VA health care system, the Veterans Health Administration (VHA)—and specifically its delivery of mental health and substance abuse services, paying special attention to the most prominent needs of the treated population, basic organizational structures, and notable changes in the delivery of those services in the past decade. (3) Next, data on the linkages between VA and non-VA mental health and social welfare systems will be presented to allow evaluation of the level of self-containment of VA mental health care and the nature of its linkages with other systems. The next section (4) describes the organizational processes through which quality management has been developed in VA mental health and substance abuse care in recent years. Having presented the context of mental health performance management, in the next section (5) we present evidence concerning the safety, effectiveness, person-centeredness, timeliness, efficiency and equitability of VA mental health and substance abuse care as it has changed in recent years and as it compares to other health care systems. Finally, (6) we touch on an area that has received virtually no systematic attention, the sometimes ambivalent reactions of front-line health system managers and clinicians to the implementation of performance management systems. I. AMERICA’S VETERANS: MENTAL HEALTH AND SUBSTANCE ABUSE STATUS AND USE OF VA SERVICES America’s Veterans In 2001, the national Survey of Veterans (SOV) conducted detailed interviews with a nationally representative sample of 20,000 veterans identified through VA administrative records and random digit dialing (USVA, 2004). Population estimates derived from the survey were based on an overall estimate of 25,196,036 living veterans in 2001, which included 12.4% of all U.S. adults and 24.5% of men 18 years or older (U.S. Census Bureau, 2001) (see column 1 of Table 1). The two most distinctive characteristics of the veteran population is that it is overwhelmingly male (94%) and that its age distribution is shaped by defense manpower needs and particularly wartime recruitment, rather than by the natural rate of population growth. Veterans are older than other Americans, first, because eligibility for military service begins at 18. In addition, and perhaps more important, World War II, the Korean conflict, and the Vietnam conflict, spanning 34 years from 1941 to 1975, were fought by far larger forces than have served in the 28 years since 1975. Altogether 30 million troops served during the

OCR for page 423
Improving the Quality of Health Care for Mental and Substance-Use Conditions three major war eras from 1941 to 1975 while only 5.7 million living veterans entered military service since 1975 (U.S. Census Bureaus, 2001). Thus 46% of veterans are 60 years old or more as compared to only 16.6% of the general population and 14.5% of men. A detailed characterization of the U.S. veteran population based on the 2001 SOV is presented in the first column of Table 1. Veterans do not differ markedly from the rest of the U.S. population in racial composition, education, or employment although there are somewhat fewer blacks and Hispanics and educational levels are somewhat higher among veterans, most likely due to the availability of specific veterans’ educational benefits. Mental Health and Substance Abuse Disorders Among Veterans and Nonveterans in the General Population Self-report data from the SOV show that 6.6% of veterans report having received services for a MH/SA problem in the past year, but these data do not allow comparison with the MH/SA status of the general population. The most useful study for comparing MH/SA problems of veteran and nonveteran men is a secondary analysis of data from the Epidemiological Catchment Area (ECA), which surveyed the mental health status of 18,572 Americans, including 10,954 men, and oversampled older Americans, in five locations in 1980 (Norquist et al., 1990). ECA data reveal no differences in lifetime prevalence of mental health disorder among veterans of World War II or either the Korean or Vietnam conflict eras and age-matched nonveteran men. In contrast, veterans of the post-Vietnam era (the initial period of the All Volunteer Force [AVF]) show a greater prevalence of lifetime mental disorder (54.6% of veterans vs. 40.9% of nonveterans [p <.0001]). Data on 6-month prevalence of mental disorder show a similar pattern, although World War II era veterans had a significantly lower overall prevalence of mental disorder than nonveterans (11.8% vs. 17.7%, p <.01). Examination of specific lifetime disorders shows that World War II era veterans had lower prevalences of any nonsubstance abuse disorder than nonveterans (12.2% vs. 18.5%, p <.01); Vietnam era veterans had lower prevalence of schizophrenic disorders (0.8% vs. 2.2%, p <.05) and affective disorders (4.4% vs. 8.3%, p <.01); and post-Vietnam veterans had higher lifetime prevalence of substance abuse disorders (47.4% vs. 30.6%, p <.01, including both alcohol and drug disorders) and antisocial personality disorder (14.9% vs. 5.8%) but lower prevalence of schizophrenic disorder than nonveterans (0.3% vs. 1.5%, p <.01). Findings of greater rates of mental illness and especially substance abuse among veterans of the AVF are consistent with several studies showing greater substance use among military personnel in the immediate post-

OCR for page 423
Improving the Quality of Health Care for Mental and Substance-Use Conditions Vietnam era (Rosenheck et al., 1996a) and with studies that have demonstrated a three to four times greater risk of homelessness among post-Vietnam veterans as compared to nonveterans, as well as among female veterans (who have always served on a voluntary basis) (Gamache et al., 2001, 2003). The increased risk of homelessness is completely absent or not statistically significant among veterans of earlier eras in which the draft guaranteed a more representative military force. Thus a major issue in the psychiatric epidemiology of U.S. veterans appears to be the shift from the draft to the AVF. There have been far fewer veterans since the end of the Vietnam conflict, but they appear to have a greater risk of MH/SA problems, not because of the hazards of military service, but because of self-selection processes among those who volunteer. Findings from the ECA are also consistent with the results of the National Vietnam Veterans Readjustment Study (NVVRS) (Kulka et al., 1990), a major epidemiologic study of representative samples of Vietnam era veterans and a matched sample of nonveterans. While the NVVRS found higher rates of posttraumatic stress disorder (PTSD) among veterans exposed to high levels of combat than among other veterans who were not and civilians; rates of other mental disorders did not differ between veteran and nonveteran populations. A recent analysis of data from 12,480 male respondents aged 25–60 in the National Household Drug Abuse Surveys from 1994, 1997, and 1998 showed veterans reported greater rates of near-daily alcohol use in the past year (22.9 vs. 19.2%, p <.001) but lower rates of illicit drug use (10.0% vs. 12.9%, p <.001) (Tessler et al., in press). Similarly, an epidemiologic study that compared homeless veteran and nonveteran men in Los Angeles found that veterans were less likely to have nonsubstance abuse mental health disorders (47.5% vs. 65.2%, p <.01) but more likely to meet criteria for alcohol abuse or dependence (72.3% vs. 59.8%, p <.05) (Rosenheck and Koegel, 1993). Both these studies involve representative samples of the veteran population, not those involved in treatment. MH/SA Status Among Veteran and Nonveteran Users of Mental Health Services Three studies have compared veteran and nonveteran men who were using MH/SA services (Desai et al., in press-b; Rosenheck et al., 2000a; Tessler et al., 2002). These studies generally have found veterans to be older, less likely to be minorities, better educated, and with higher incomes, and analyses were adjusted for these differences in comparisons of mental health status. The Schizophrenia Patient Outcomes Research Team (PORT) study of representative samples of patients treated for schizophrenia in Ohio and

OCR for page 423
Improving the Quality of Health Care for Mental and Substance-Use Conditions Georgia included an over-sampling of VA patients to allow adequate power for comparison of VA–non-VA male service users (N = 466 VA patients and 279 non-VA male patients) (Rosenheck et al., 2000a). After adjusting for age and race differences, no significant differences were found on measures of psychosis, depression, or substance abuse. A comparison of 1,252 veteran and 3,236 nonveteran men treated at 18 sites in the ACCESS demonstration of service system integration for homeless people with severe mental illness also found no differences between veterans and nonveterans on psychiatric or drug problems, although veterans had somewhat more severe alcohol problems as measured by the Addiction Severity Index (Tessler et al., 2002). Finally, the Connecticut Outcome Study compared 196 VA patients and 337 non-VA patients treated at nearby state-operated Community Mental Health Centers and also found no significant differences in measures of psychiatric symptoms or substance abuse, after adjustment for age, race and income (Desai et al., in press). Use of VA Services Data from the 2001 SOV show that 20.5% of veterans reported using any VA services (i.e., not specifically MH/SA services) in the past year and 34.4% in their lifetimes. These figures are substantially higher than those recorded in a similar national survey conducted in 1987. In that survey, only 5.8% reported VA service use in the past year and 21.2% lifetime (Rosenheck and Massari, 1993). These substantial changes reflect at least three factors. First, eligibility for VA services was vastly expanded in 1996 from those who receive VA compensation or pension benefits or have low incomes, estimated to have represented only 9.4 million veterans (Kizer, 1999), to the entire population of 25 million veterans. In addition, major changes in the configuration of VA facilities have made services far more accessible. In the 1990s major reductions in inpatient beds allowed expansion of outpatient care and the establishment of over 500 accessible community-based outpatient clinics (GAO, 2001). In 2003, the average veteran lived 12.2 miles from the nearest VA facility as compared to 32.0 miles in 1994 (Greenberg and Rosenheck, 2003; Rosenheck and Cicchetti, 1995). In addition, the increasing numbers of uninsured Americans and the growing cost and importance of prescription drugs have also contributed to the growing demand for virtually free VA services. In spite of the changes in eligibility, veteran characteristics that are associated with use of VA services have changed little since 1987. Columns 2 and 3 in Table 1 show the proportions of veterans in each subgroup who used VA services in the past year and in their lifetimes. Both recent and lifetime VA service use is associated with greater age, minority status, low

OCR for page 423
Improving the Quality of Health Care for Mental and Substance-Use Conditions education and income, lack of private insurance, poor health, mental health service use, and receipt of VA compensation. Table 2 presents a logistic regression analysis of factors that predict recent and lifetime VA service use to illustrate both the independent and the relative magnitude of the effects of each factor. The fourth and seventh columns rank the absolute value of the magnitude of these effects (both positive and negative) and show the strongest correlates of VA service use to be receipt of VA compensation, low income, lack of private insurance, poor health, age less than 30, African American race, Prisoner of War experience, and mental health service use and related disability. These factors are virtually the same as those identified in the 1987 SOV (Rosenheck and Massari, 1993). Use of VA Mental Health Services SOV data further reveal that 6.6% of all veterans used mental health services in the previous year, and 2.7% used VA mental health services (41% of those who used any mental health services) (Columns 4 and 5 of Table 1). Among mental health service users, too, those who used VA services are older, more likely to be minority group members, had less education and lower incomes, lacked private insurance, had poorer health, and received VA compensation (see also Table 3 for logistic regression analysis and ranking). It is notable that veterans who sought services for PTSD were especially likely to have used VA mental health services, replicating a finding from a previous analysis of the NVVRS data (Rosenheck and Fontana, 1995) and showing that, contrary to what was once popular belief, veterans with PTSD related to their military service do not avoid using VA mental health services. The PORT survey of the treatment of schizophrenia in Ohio and Georgia allows further comparison, with the group of severely mentally ill veterans, of those who used VA services (N = 350) and those who used non-VA services (N = 170) (reanalysis based on data in Rosenheck et al., 2000a). Stepwise logistic regression showed veterans who used VA services to be 2.7 times more likely to be receiving VA compensation, 3.0 times more likely to be living in a supervised residence, 37% less likely to be black, and to have used fewer emergency services and have had less severe symptoms. A study that focused on administrative data from state mental hospitals in eight states between 1984 and 1989 (Desai and Rosenheck, 2000) found that from 7 to 27% of men in these non-VA facilities were veterans as compared to 29–34% in the general male population, suggesting that veterans are less likely to use non-VA service than other men. In comparison with other state hospital patients, veterans were older, more like to have alcoholism and bipolar disorder, and perhaps of greatest interest: (a) lived

OCR for page 423
Improving the Quality of Health Care for Mental and Substance-Use Conditions further away from VA hospitals than nonveterans; (b) were from states with lower per capita expenditure on VA mental health care; and (c) were from states with higher per capita expenditure on state hospitals. Thus, in addition to personal characteristics, both residential remoteness from a VA facility and scant supply of VA services in relation to non-VA services increased veterans’ use of non-VA services. A study of the proportion of veterans in each U.S. county who use VA MH/SA services similarly found distance from veterans’ residences to the nearest VA facility to be the strongest predictor of VA MH/SA services use, along with the relative local supply of VA and non-VA services (Rosenheck and Stolar, 1998). In fact, VA service use among veterans service connected for psychoses was specifically reduced in association with a high supply of state and county mental hospital resources, while VA use among nonpsychotic veterans was negatively associated with the supply of non-Federal general hospital resources. The impact of the supply and proximity of VA services has also been demonstrated in a sample of homeless mental health services users (Gamache et al., 2000). II. TREATMENT OF MENTAL HEALTH AND SUBSTANCE ABUSE IN THE VA: PATIENTS, ADMINISTRATION, RELATIONSHIPS WITH OTHER FEDERAL AGENCIES, STAKEHOLDERS, AND CHANGES SINCE 1995 The VA is a cabinet level federal department that includes two major subdivisions that provide services to people with mental illness: (1) The VHA, which delivered health care services to approximately 5 million veterans in fiscal year (FY) 2003 at 162 medical centers and more than 850 facility and community based clinics; and (2) the Veterans Benefits Administration (VBA), which provided income benefits to over 2.5 million veterans in FY 2003 in addition to rehabilitation and educational support and housing loan guarantees (U.S. Department of Veterans Affairs, 2003). Of the 2.5 million veterans who received compensation from VBA in 2003, 481,000 received compensation for mental illness, and of these 47% used VA mental health services (Greenberg and Rosenheck, 2004a). VA Patients Diagnosed with Mental Health and Substance Abuse Disorders Administrative workload data from the VHA show that in FY 2003, 1,218,327 veterans (about 25% of all those who received VA health services) received a mental health or substance abuse (MH/SA) diagnosis (ICD-9 codes 290.00-312.99) during an inpatient, nursing home, residential, or outpatient encounter (Table 4).

OCR for page 423
Improving the Quality of Health Care for Mental and Substance-Use Conditions The most frequent MH/SA diagnoses (Table 4, numbered column 2) were dysthymia (41%), PTSD (20%), anxiety disorder (20%), and major depressive disorder (20%). Altogether 22% received a substance abuse diagnosis (17% alcohol abuse/dependence and 11% drug abuse/dependence) and 18% were dually diagnosed. Altogether, 38% received VA compensation for medical or psychiatric problems, almost four times that in the general population. Among veterans who received a MH/SA diagnosis, 930,098 (76%) received a primary diagnosis for MH/SA (Table 4, numbered column 3), meaning that the MH/SA diagnosis was the primary focus of at least one contact during the year; and 705,209 of these received treatment in a MH/SA specialty program (numbered column 4) (76% of those who received a primary MH/SA diagnosis and 57% of those who received any primary or secondary MH/SA diagnosis). An additional 89,372 veterans received services in a specialty mental health program but did not receive a primary mental health diagnosis (not shown on table), for a total of 794,581 or 17% of all VA patients who received mental health services in a specialty clinic setting (Greenberg and Rosenheck, 2004a). The most frequent MH/SA diagnoses among veterans treated in specialty clinics (Table 4, numbered column 5) were also dysthymia (43%), PTSD (31%), anxiety disorder (22%), and major depressive disorder (24%), with 26% receiving a substance abuse diagnosis (20% alcohol abuse/dependence and 16% drug abuse/dependence), and 18% were dually diagnosed. Global Assessment of Functioning (GAF) scores based on a single item rating scale ranging from 0 to 100, which is a standard part of the psychiatric diagnosis, average 41.8 (s.d. = 13.1) among inpatients at the time of discharge and 53.3 (s.d. = 11.3) among outpatients (Greenberg and Rosenheck, 2004a). A GAF score of 50 is often used as a cutoff for severe mental illness. Thus although fewer than 15% of VA patients with MH/SA diagnoses have the most severe illnesses such as schizophrenia or bipolar disorder, there is considerable functional impairment among these patients. It is also noteworthy that although only 76% of veterans with a primary MH/SA diagnosis receive care in specialty clinics, 95% or more of those with the most serious illnesses (schizophrenia, major depressive disorder, bipolar disorder, or PTSD) receive care in MH/SA specialty clinics, and 45% receive VA compensation. Annual surveys conducted from FY 1995 (Rosenheck et al., 1996c) to FY 2000 (Seibyl et al., 2001) showed that almost 30% of VA psychiatric inpatients, and almost 50% of those in inpatient substance abuse programs, had been homeless at the time of admission. Over 100,000 MH/SA outpatients are identified as homeless each year, about 12% of the total, which is most likely a substantial undercount, since coding for homelessness is not uniform in the outpatient files.

OCR for page 423
Improving the Quality of Health Care for Mental and Substance-Use Conditions   Odds Ratio 95% Confidence Interval Rank Order Military Experience Active Duty 0–2 years 1.00 Reference 3–5 years 1.27 1.26-1.28 26 5 years or more 1.36 1.34-1.38 22 Served in a Combat or War Zone 1.01 1.00-1.02 38 Exposed to Dead or Wounded 1.19 1.18-1.20 33 Prisoner of War 1.34 1.29-1.39 23 Health Status Self-Reported Health Status Fair-Poor 1.85 1.82-1.87 14 Good 1.95 1.93-1.98 11 Excellent-Very Good 1.00 Reference Any Health Service Use For Alcohol or Drugs 1.02 1.00-1.03 36 PTSD 1.50 1.48-1.51 18 Mental Health 1.25 1.24-1.26 30 Service Connected Not Service Connected 1.00 Reference <50% 7.08 6.99-7.16 2 >50% 16.56 16.28-16.85 1 Mental health interferes with work or activities at all 0.64 0.64-0.65 17 Proxy Answered Questions 1.93 1.88-1.99 12 *All odds ratios were significant at p <.0001 except for an employment status of unemployed or other, which was not significant. (Combat p value is .503 and use of mental health services for drugs and alcohol p value is .0325)

OCR for page 423
Improving the Quality of Health Care for Mental and Substance-Use Conditions TABLE 4 Veterans Treated for Primary or Secondary Mental Health (MH) Diagnosis in the VHA in FY 2003, by Specialty Clinic Treatment Diagnosis (1) Total with any MH Dx (2) Percent of total by Dx/SC Primary MH Diagnosis* (8) Secondary MH Dx only (9) Percent with secondary Dx only (3) Total with Primary MH Dx/SC (4) Treated in Specialty MH Clinic (5) Percent by Dx/SC (6) Treated only in non-MH clinic (7) Percent Treated in Specialty Clinic Total 1,218,327 100% 930,098 709,432 100% 220,666 76% 288,229 24% Dysthymia 495,519 41% 360,511 302,070 43% 58,441 84% 135,008 27% PTSD 242,850 20% 227,448 216,970 31% 10,478 95% 15,402 6% Anxiety disorder 238,191 20% 183,062 155,108 22% 27,954 85% 55,129 23% Major depressive disorder 182,927 15% 177,500 173,010 24% 4,490 97% 5,427 3% Schizophrenia 100,054 8% 95,733 91,568 13% 4,165 96% 4,321 4% Adjustment disorder 79,103 6% 73,439 67,674 10% 5,765 92% 5,664 7% Bipolar disorder 76,702 6% 73,202 70,383 10% 2,819 96% 3,500 5% Other psychosis 61,985 5% 53,764 39,502 6% 14,262 73% 8,221 13% Dementia 60,269 5% 38,871 23,047 3% 15,824 59% 21,398 36% Personality disorder 36,620 3% 35,450 34,427 5% 1,023 97% 1,170 3% Other psychiatric disorder 274,209 23% 194,649 138,118 19% 56,531 71% 79,560 29% Substance abuse 266,358 22% 214,251 188,252 27% 25,999 88% 52,107 20% Alcohol 209,930 17% 166,005 144,854 20% 21,151 87% 43,925 21% Drug 132,412 11% 121,743 114,994 16% 6,749 94% 10,669 8% Dually diagnosed (MH and SA) 221,288 18% 211,021 187,808 26% 23,213 89% 10,267 5% Service connected (1–100%) 462,069 38% 384,133 320,831 45% 63,302 84% 77,936 17% <=50% 238,690 20% 190,236 150,669 21% 39,567 79% 48,454 20% >50%<100% 120,818 10% 102,600 88,391 12% 14,209 86% 18,218 15% 100% 102,561 8% 91,297 81,771 12% 9,526 90% 11,264 11% Total diagnoses/veteran 1.8     2.2   1.0   1.4   *The primary diagnosis is the principal condition treated during the episode of care. NOTE: MH = mental health, Dx = diagnosis, SA = substance abuse, SC = service connected, PTSD = posttraumatic stress disorder

OCR for page 423
Improving the Quality of Health Care for Mental and Substance-Use Conditions TABLE 5 Workload of Specialized VA Mental Health Programs (FY 2003) Program Outpatient/Community Patients Treated Visits/Pt/Yr. Total Visits Outpatient Substance Abusea 115,954 25.8 2,991,613 Mental Health Intensive Case Management (ACT) 3,961 59.3 234,887 Day Hospital 6,218 27.4 170,373 Day Treatment Center 14,389 39.7 571,243 PTSD Clinical Teams 72,849 8.9 645,895 Community Residential Care 9,436 8.4 79,262 Health Care for Homeless Veteransb 61,123 4.1 252,642 Psychosocial Rehabilitationc 38,302 33.5 1,283,117 Total 283,930 17.4 4,945,917 Inpatient/Residential Episodes of Care Days/Ep. Total Days General Psychiatry Inpatient 87,002 12.8 1,113,626 Psychosocial Residential Rehabilitation and Treatmentd 12,771 40.1 512,117 Inpatient Substance Abuse 5,763 7.4 42,646 Specialized PTSD Programs 4,302 38.8 166,908 Homeless Veterans Grant and Per Diem Program 10,982 92.6 1,016,933 Health Care for Homeless Veterans Residential Contracts 5,430 52.8 286,599 Domiciliary Care for Homeless Veterans 5,156 110.3 568,707 Total 126,250 24.9 3,138,829 aIncludes methadone maintenance and substance day treatment bIncludes supported housing, homeless outreach, and case management cIncludes Compensated Work Therapy, Incentive Therapy, Vocational Assistance dIncludes residential programs for PTSD, substance abuse, general psychiatry, and vocational rehabilitation

OCR for page 423
Improving the Quality of Health Care for Mental and Substance-Use Conditions Appendix Performance Measures Used by the Northeast Program Evaluation Center in the Evaluation and Monitoring of VA Mental Health Programs (Data from FY 2002) Health Care for Homeless Veterans (HCHV) Program and Domiciliary Care for Homeless Veterans (DCHV) Programs Compensated Work Therapy (CWT) Program and Compensated Work Therapy /Transitional Residence (CWT/TR) Program PTSD Performance Monitors and Outcome Measures Mental Health Intensive Case Management Performance Measures from the National Mental Health Program Performance Monitoring System Adherence to PORT Pharmacotherapy Guidelines for Patients with Schizophrenia Outcomes on the Global Assessment of Functioning (GAF) Scale I. HEALTH CARE FOR HOMELESS VETERANS (HCHV) PROGRAM AND DOMICILIARY CARE FOR HOMELESS VETERANS (DCHV) PROGRAM Program Structure Unique veterans served/stops per clinician (HCHV mean = 147.5 veterans; 619 visits) Percent of allocated staff slots that are filled (HCHV mean = 96.3%) Literally homeless veterans seen per clinician (HCHV mean = 90.3) Per diem cost (HCHV mean = $37.67) Annual turnover rate1 (DCHV mean = 3.3) Process Measures Patient Characteristics Percent strictly homeless (living outdoors/shelter) (HCHV mean = 67.8%; DCHV mean = 34.8%) Percent with no time homeless (HCHV = 8.1%; DCHV mean = 4.1%) Percent with a psychiatric disorder, substance abuse problem, or medical illness (HCHV mean = 82%; DCHV mean = 99.8%) 1   Annual turnover rate is determined by dividing the total number of discharges in the DCHV Program by the number of DCHV operating beds. Average length of stay and occupancy rates will influence a site’s value for annual turnover rate.

OCR for page 423
Improving the Quality of Health Care for Mental and Substance-Use Conditions Program Participation Mean days in residential treatment (HCHV mean = 59 days; DCHV mean = 104.1 days) Percent of successful residential treatment completions (HCHV = 52%; DCHV mean = 69%) Percent of disciplinary discharges from residential treatment (DCHV mean = 14%) Percent of premature program departures from residential treatment (DCHV mean = 12%) Percent contacted through outreach (HCHV = 78%) Appropriateness for residential treatment (HCHV = 89%) Outcome Measures Percent with clinical improvement in alcohol problems (HCHV = 73.1%; DCHV mean = 84% unadjusted) Percent with clinical improvement in non-substance abuse psychiatric problems(HCHV = 70.7%; DCHV mean = 83% unadjusted) Percent with clinical improvement with medical problems (HCHV mean = 66.3%, DCHV mean = 88.4%) Percent discharged to an apartment, room, or house (HCHV mean = 37.9%; DCHV mean = 57.4% unadjusted) Percent with no housing arrangements after discharge (Supported Housing Program = 21.8%; DCHV mean = 19.3% unadjusted) Percent discharged with arrangements for full- or part-time employment (HCHV = 49.3%; DCHV mean = 54.5% unadjusted) II. COMPENSATED WORK THERAPY (CWT) PROGRAM AND COMPENSATED WORK THERAPY/TRANSITIONAL RESIDENCE (CWT/TR) PROGRAM Process Measures Program participation Mean hours worked per week in CWT (CWT mean = 26.1 hr/wk; CWT/TR mean = 33.0) Mean days in residential treatment (CWT/TR mean = 192 days) Percent of successful completions (CWT mean = 51.4%; CWT/TR mean = 56.1%) Percent of disciplinary discharges (CWT mean = 15.4%; CWT/TR mean = 26.8%)

OCR for page 423
Improving the Quality of Health Care for Mental and Substance-Use Conditions Veteran satisfaction in the residential treatment environment2 (CWT/TR Community Oriented Program Evaluation Scale (COPES) index mean = 2.95) Veteran satisfaction in the therapeutic work environment3 (CWT/TR Ward Environment Scale (WES) mean = 6.22) Outcome Measures Mean work improvement score4 (CWT mean = 1.57) Percent with clinical improvement in alcohol problems (CWT mean = 63.6% unadjusted) Percent with clinical improvement in drug problems (CWT mean = 63% unadjusted) Percent with clinical improvement in nonsubstance abuse psychiatric problems(CWT mean = 47.3% unadjusted) Percent with clinical improvement with medical problems (CWT mean = 35.2% unadjusted) Percent competitively employed after discharge (CWT mean = 41.3%) Percent unemployed after discharge (CWT mean = 24.8%) Percent employment status unknown after discharge (CWT mean = 14.6%) Percent of veterans relocated and reinterviewed 3 months after discharge (CWT/TR mean = 55.8%) Mean Addiction Severity Index (ASI) index for alcohol problems 3 months after discharge (CWT/TR mean = 0.08 unadjusted) Mean ASI index for drug problems 3 months after discharge (CWT/TR mean = 0.03 unadjusted) Mean ASI index for psychiatric problems 3 months after discharge (CWT/TR mean = 0.18 unadjusted) Mean days competitively employed past month at 3 months after discharge (CWT/TR mean = 11 days unadjusted) Mean days housed past 90 days at 3 months after discharge (CWT/TR mean = 71.2 days unadjusted) III. PTSD PERFORMANCE MONITORS AND OUTCOME MEASURES Outpatient Programs (Specialized PTSD Outpatient Programs) Patient Characteristics War zone service (87%) DD 214 (Discharge Certification form) service validation (65%) 2   Range equals 0–4. 3   Range equals 0–9. 4   Range equals 0–2.

OCR for page 423
Improving the Quality of Health Care for Mental and Substance-Use Conditions PTSD clinical diagnosis (82%) Substance abuse diagnosis (41%) Prior psychiatric treatment (74%) Prior specialized PTSD treatment (22%) Workload Number of visits per filled Full Time Employee Equivalents (FTEE) (1,160) Number of veterans treated per filled FTEE (88) Costs Direct costs per visit ($73) Direct costs per capita ($958) Outpatient Care Measures (All VA PTSD treatment, specialized and non-specialized) Service Utilization and Continuity of Care Six Months Following Inpatient Index Stay Any outpatient stop in 6 months after discharge (DC) (92.0%) Any outpatient stop in 30 days after DC (72.0%) Days to first outpatient stop in 6 months after discharge (23.7) Number of stops in 6 months among those with any stops (18.6) Continuity: Bi-months (2 month intervals) with two stops (2.5) (Next measure applies only to those with a dual diagnosis (PTSD and SA) At least 1 Psychiatric and 1 SA outpatient (OP) stop in 6 months after DC (14.4%) Continuity of Care Among Outpatients with PTSD Diagnosis Number of outpatient stops (15.5) Number of days with outpatient stops (11.9) Continuity: bi-months with 2 stops (2.61) Continuity: months with any MH visit of six months (4.17) Dropout (6 months with no O/P visit) 13% Continuity of care index (.57) Modified Continuity Index (MCI) (.80) Inpatient/Residential Programs (Specialized PTSD programs) Patient Characteristics (Same as for Outpatient programs) Costs Direct costs per diem ($136) Direct costs per capita ($4,662)

OCR for page 423
Improving the Quality of Health Care for Mental and Substance-Use Conditions Outcomes (4 months after discharge)5 PTSD symptoms Short Mississippi Scale (−5.6%) NEPEC PTSD Scale (−6.7%) Substance abuse ASI Composite for Alcohol problems (−23.5%) ASI Composite for Drugs problems (−14.3%) Violence NVVRS Scale (−38.8%) Work Number of days worked for pay (+1.8%) Satisfaction with treatment Client Satisfaction Scale (Attkisson et al.) (15.6) All PTSD Inpatient Care (General and Specialized Programs) Bed days six months after DC (5.6) Number of admissions 6 months after DC (.45) Percentage readmitted within 14 days (5.0%) Percentage readmitted within 30 days (8.0%) Percentage readmitted within 180 days (30.0%) Days to readmission first year after discharge (74.8) IV. MENTAL HEALTH INTENSIVE CASE MANAGEMENT (MHICM) Program Structure Percent allocated FTEE that are filled (84%) Availability of appropriate medical support (89%) Availability of appropriate nursing support (96%) Unfilled FTEE lagged greater than 6 months (43% of teams) Caseload size (average = 15.4)(should be less than 15) Appropriateness of Admissions Hospitalized 30 days or more Hospitalized more than 30 days in the previous year (82%) Diagnosis of psychotic illness (78%) GAF < 50 at admission 5   Negative values on symptom measures represent improvement (i.e., declining symptoms)

OCR for page 423
Improving the Quality of Health Care for Mental and Substance-Use Conditions Treatment Process Clients terminated Intensity (greater than 1 hour per week)(66.7%) Services provided in the community 60% of the time (87%) Face-to-face contacts per week (1.45) Seen for rehabilitation (36.1%) Improvement in Therapeutic Alliance (10.2%) ACT Fidelity Score (4.0) Outcomes6 Change in inpatient days (6 months before entry to 6 months after) (−47 days; −73%) Brief Psychiatric Rating Scale (BPRS, Symptoms)(−3.92; −10%) Brief Symptom Inventory (BSI, Symptoms)(−.22; −11%) Global Assessment of Functioning (GAF)(−2.01; −5%) Instrumental Activities of Daily Living (IADL)(+.95, 2%) Lehman Quality of Life Question (+2.8, 11%) Housing Independence (+.43, 15%) Satisfaction with VA Mental Health Services (+1.54, 17.6%) Satisfaction with MHICM (+.62, 21%) V. PERFORMANCE MEASURES FROM THE NATIONAL MENTAL HEALTH PROGRAM PERFORMANCE MONITORING SYSTEM Population Coverage Proportion of All U.S. Veterans who received VA MH services (2.5%) Proportion of Veterans Service Connected for Mental Illness who received VA MH services (42.2%) Proportion of low-income Non Service Connected Veterans who received VA MH services (7.0%) Proportion of Female Veterans who received VA MH services (3.1%) Inpatient Care Measures Bed days six months after DC (6.47) Number of admissions 6 months after DC (.55) Percentage readmitted within 14 days (6.9%) Percentage readmitted within 30 days (11.5%) 6   Negative values on symptom measures represent improvement (i.e., declining symptoms).

OCR for page 423
Improving the Quality of Health Care for Mental and Substance-Use Conditions Percentage readmitted within 180 days (32.1%) Days to readmission first year after discharge (64.7%) Outpatient Care Measures Service Utilization and Continuity of Care Six Months Following Inpatient Index Stay Any outpatient stop in 6 months after DC (82.1%) Any outpatient stop in 30 days after DC (63.3%) Days to first outpatient stop in 6 months after discharge (24.9) Number of stops in 6 months among those with any stops (25.8) Continuity: Bi-months with two stops (2.16) Any medical outpatient stop in 6 months after DC (82.8%) Days to 1st medical OP stop in 6 months after DC (36.6) Number of OP medical stops in 6 months among those with any stops (10.6) (Next 4 measures apply to those with a dual diagnosis (Psyc. and SA) At least 1 Psyc. and 1 SA OP stop in 6 months after DC (21.2%) At least 3 Psyc. and 3 SA OP stop in 6 months after DC (17.3%) Continuity: bi-months with two stops (2.17) Number of Psyc. and SA visits among those with any stops (25.89) Continuity of Care among Outpatients with Psychotic Diagnoses Number of outpatient stops (16.8) Number of days with outpatient stops (12.5) Continuity: bi-months with 2 stops (2.56) Continuity: months with any MH visit of 6 months (3.98) Dropout (6 months with no OP visit) 15% Continuity of care index (.56) Modified MCI (.80) Inpatient Satisfaction Measures Coordination of Care (.72) Provision of Information (.66) Timeliness/Access to Care (.61) Courtesy (.66) Emotional Support (.63) Respect for Patient Preferences (.71) Family Involvement (.54) Physical Care (.62) Transition Home (.62) General Satisfaction (.53)

OCR for page 423
Improving the Quality of Health Care for Mental and Substance-Use Conditions VI. ADHERENCE TO PORT PHARMACOTHERAPY GUIDELINES FOR PATIENTS WITH SCHIZOPHRENIA Percentage of patients receiving oral antipsychotic medication (81%). Percentage of veterans dosed higher than recommended guidelines (12.3%) Percentage of veterans dosed lower than recommended guidelines (28.6%) Percentage of patients receiving polypharmacy (two antipsychotic medications)(8.1%) VII. OUTCOMES ON THE GLOBAL ASSESSMENT OF FUNCTIONING (GAF) SCALE Improvement after inpatient discharge Change from inpatient GAF to last outpatient GAF in first six months after discharge (4.9) Change from inpatient GAF to the last outpatient GAF of the fiscal year (6.4) Improvement after during outpatient treatment Change from first outpatient GAF to last outpatient GAF in the next 6 months (0.46) Change from first outpatient GAF to last outpatient GAF of the fiscal year (0.33) Change from first outpatient GAF in the second 6 months of the fiscal year to the last outpatient GAF of the fiscal year (0.38) Among newly admitted outpatient veterans (those with no outpatient mental health visits in the first 3 months of the fiscal year) the change from first outpatient GAF to last outpatient GAF of the fiscal year (1.9)