National Expenditures for Mental Health Services and Substance Abuse Treatment 1991–2001

Chapter 1: Background and Methods
Organization of This Report
Chapter 1 summarizes the methods of estimation and limitations of the estimates contained in this volume. Chapter 2 summarizes the findings for total MHSA spending. In the subsequent chapters, mental health and substance use disorders are examined separately because expenditure patterns for these disorders differ in some important ways. Chapter 3 examines MH services spending for the latest year estimated, 2001, and compares this to all health care spending. Chapter 4 reviews the trends in MH and all health expenditures since 1991. Chapter 5 focuses on substance abuse (SA) treatment and explores the major providers and sources of support for substance use disorders in comparison to all health in 2001. Chapter 6 presents information on trends in SA spending from 1991 to 2001. Chapter 7 draws conclusions from the results of the spending estimates. Appendix A contains tables of estimates that serve as the foundation for the graphs displayed in this report. Appendix A displays estimates for 1991 and 2001, as well as average annual growth rates for 1991 - 2001, 1991 - 1996, and 1996 - 2001.
Rationale for the Estimates
The mission of the Substance Abuse and Mental Health Services Administration (SAMHSA), part of the U.S. Department of Health and Human Services, is to focus attention, programs, and funding on improving the lives of people with or at risk of mental and substance use disorders. The SAMHSA vision - "a life in the community for everyone" - focuses sharply on building resiliency and facilitating recovery for clients. The SAMHSA strategy to improve accountability, capacity, and effectiveness ensures that its resources are being used effectively and efficiently through State and community programs to serve all clients. SAMHSA´s programs have been aligned with a core set of priorities - access to services, retention in treatment, social support, abstinence from drug use and alcohol abuse, employment/return to school, criminal justice involvement, and stabilized family and living conditions. To build better systems, SAMHSA tracks national trends, establishes measurement and reporting systems, and develops and promotes standards to monitor and guide efforts to improve delivery of services to its clients.

The estimates in this report track national spending on treatment for mental and substance use disorders. This information aids SAMHSA, as well as policy makers, providers, and consumers, to understand what the nation spends on mental health services and substance abuse treatment, who funds that treatment, who delivers that treatment, and how the system has changed over time.
Purpose and Scope of Estimates
The estimates provide ongoing information of national spending on health care services related to the diagnosis and treatment of mental and substance use disorders. They also provide a view of MHSA treatment spending over time and compared with spending on all health care. This report describes estimates for 1991 through 2001.

These estimates focus on expenditures for MHSA treatment, not on the burden of MHSA illnesses. Burden of illness studies include costs not directly related to treatment, such as the impact of mental illness on productivity, societal costs linked to drug-related crimes, or housing and other accommodation subsidies to clients with MHSA disorders. The scope of the report also does not include the physical consequences of MHSA disorders. For example, physical consequences of MHSA problems include cirrhosis, trauma, and HIV and other infectious diseases. The report also does not include expenditures for the diagnosis and treatment of related disorders that are normally, or historically, covered by general medical insurance, such as dementias and tobacco addiction. Finally, the expenditures reported do not include those allocated to prevent substance abuse.

The reason for these exclusions is that the estimates include expenses in MHSA insurance coverage and MHSA public program funding. For example, treatment for cirrhosis of the liver would not be covered as substance abuse treatment under a third-party insurance policy (such as a managed behavioral health plan), nor would it be treated under a publicly funded substance abuse treatment program. For the most part, it would be covered under medical insurance or under general Medicaid or Medicare. HIV infection, which may result from injecting drugs, would be treated as a medical problem by a physician, not by a MHSA specialist, and expenses would be reimbursed under medical insurance.
The estimates integrate a wealth of national data sources from various government agencies and private organizations. Data are analyzed using both actuarial and statistical techniques. A number of complex issues must be addressed when combining the data to produce comprehensive estimates, such as assuring consistency across data sources, avoiding duplicate accounting, and adjusting for incomplete observations, among others.
Expert Advice. The methods for the estimation of national MHSA expenditures drew extensively upon suggestions from reviewers. The advisors included experts in mental health, substance use and abuse, expenditure estimation, actuarial science, and health economics. Experts on State programs (including the National Association of State Alcohol/Drug Abuse Directors (NASADAD) and the National Association of State Mental Health Program Directors (NASMHPD)) also reviewed the report and provided advice. Government experts on the SAMHSA specialty sector survey data shared information and insights on the imputation methods in those surveys. Appendix B lists members of the advisory panel.
Overview of Methods. The approach taken to estimate national MHSA spending was designed to be consistent with the National Health Accounts (NHA). The NHA constitutes the framework for which the estimates of spending for all health care are constructed by the Centers for Medicare and Medicaid Services (CMS). The framework can be considered as a two-dimensional matrix; along one dimension are health care providers or products that constitute the U.S. health care industry; along the other dimension are sources of funds used to purchase this health care.

The Centers for Medicare and Medicaid Services has a long history, as well as substantial expertise, in estimating national spending. The estimates of MHSA spending for non-MHSA specialty facilities were carved out of estimates of total national health services and supplies expenditures developed by CMS. Separate estimates were developed from SAMHSA data for specialty MHSA facilities. Duplicate expenditures between the two sectors were removed. Then, sector estimates were summed to obtain total national spending for mental health (MH), alcohol abuse (AA), illicit drug abuse (DA) and for total MHSA in the U.S. from 1991 through 2001. Finally, MHSA dollars were compared to all personal health care and government public health expenditures, which are referred to as national health care expenditures or all health expenditures. Table 1.1 summarizes the methods for estimating MHSA expenditures for the MHSA specialty facilities and other providers.
Strengths of Approach. The major benefit of this approach is that it levels the playing field for an analysis of the comparison between MHSA and all health care spending. When the same
Table 1.1: Overview of Methods for Estimating MHSA Expenditures
Methods Speciality Institutions Other Providers
Data Sources: Facility Surveys
(Facility-level reporting)
Encounter Data
(administrative claims and encounter-focused surveys)
Critical Data Elements:
Total Revenue
By: Facility
Modality of care (inpatient, outpatient, etc.)
Components of spending:
Service use
Payment rates
Each by:
Provider type
Basic Calculations:
Eliminate diagnoses out of scope (e.g., dementias, MR/DD) Eliminate speciality providers
Split multi-modality revenue by modality based on single modality providers' revenue Multiply "components of spending" together by diagnosis (mental, alcohol, illicit drug, all health disorders) and payer to estimate MHSA share of total health care expenditures by payer
Estimate total revenue by:
Provider type
Multiply national health care expenditures (excluding specialty MHSA providers) by "MHSA share"
Special Calculations:
Imputations for missing revenue = f (modality, ownership, reagion of country, number of client days) by facility  
Survey non-response adjustments Survey non-response adjustments
Extrapolations for missing years of data  
Projections for missing end years of data: CMS five-factor model with producer price indices  
Smooth expenditure estimates across all years Smooth expenditure estimates across all years
Results for 1991-2001: MHSA speciality expenditures by provider type, payer, and type of care MHSA other provider expenditures by provider type, payer, and type of care
methods and same underlying numbers are used for both calculations, the numbers can be made consistent for meaningful comparisons. This implies that MHSA and all health care spending can be followed over time as public programs and the health care system change. Furthermore, spending by clinical problem - mental illness, alcohol, and illicit drug abuse - can be studied to understand the patterns of public and private spending on these problems.
Basic Calculations. The specialty MHSA facility expenditure estimates were drawn from total revenues reported in the specialty surveys by facility and by payment source. Three major steps for the basic calculations were followed. First, spending on mental disorders that were beyond the scope of these estimates (dementia, tobacco addiction, mental retardation, and mental developmental delays) was subtracted from total revenues by facility. Second, revenues for providers who delivered multiple modes of care (inpatient, outpatient, and residential treatment) were re-estimated by modality using the average revenue per client and characteristics of single modality providers. Third, total revenues were configured by types of provider (for example, multi-service mental health organizations or specialty substance abuse centers), and by payer and diagnosis.

The estimates for other providers, in contrast, were dependent on and calibrated against the NHA totals. This was done for two reasons. First, specialty sector facility data sources included a census of facilities, while other data usually were based on samples. Second, the final results needed to be consistent with and comparable to the NHA estimates.

To develop MHSA expenditures for the other providers consistent with the methods of the NHA, the 2001 release of NHA health care expenditures was used. The NHA reports health care expenditures for all diagnoses only. Because the NHA encompasses both specialty institutions and general health care services, specialty institution MHSA providers had to be eliminated from the NHA estimates. This avoided double-counting the specialty service expenditures, which were estimated separately as noted above.

To distinguish MHSA from all-disease general health care expenditures, spending rates were estimated by type of diagnosis. Only the principal or primary diagnosis was used to identify spending on mental health (MH), alcohol abuse/dependence (AA), or drug abuse/dependence (DA), and all health treatments. Spending proportions for MH, AA, and DA were calculated by multiplying utilization by average prices (accounting for discounts and cost sharing) for each diagnostic group and dividing by the sum of all groups. These proportions were applied to the appropriate national health dollars from the NHA to estimate the MH, AA, and DA national dollars. Substance abuse (SA) expenditures were summed from AA and DA estimates. These estimations were made within type of payer and provider as described next.

The public sector payer categories are: Medicare, Medicaid, State and local government sources other than Medicaid, and Federal sources other than Medicare and Medicaid (e.g., Veterans Affairs, Department of Defense, and Federal Block Grants). Medicaid expenditures are combined Federal and State and local funds. The private sources are: private insurance, out-of-pocket expenditures, and other private sources (e.g., philanthropy).

The provider categories are: specialty hospitals, general hospital specialty units, non-specialty care in general hospitals, psychiatrists, non-psychiatrist physicians, other non-physician professionals, multi-service mental health organizations, free-standing nursing homes, specialty substance abuse centers, home health, and retail prescription drugs. Although the definition has differed across SAMHSA surveys and across time, multi-service mental health organizations generally include any facility that provides a variety of MH services and that is not hospital-based. Similarly, specialty substance abuse centers are generally clinics and residential treatment centers that specialize in chemical dependency.

Expenditures by provider and payer were further divided into inpatient, outpatient, and residential care. In some cases, providers offered all three types of care. For example, hospital expenditures could comprise inpatient, outpatient, or residential services. Pharmaceutical (which includes retail pharmacy only) and home health expenditures were classified as outpatient expenditures. Nursing home expenditures were classified as residential expenditures.
Special Calculations. Several complex methodological adjustments were made to develop national spending estimates from multiple and disparate data sets. Methods were devised to allocate spending by diagnosis for facility-level data where disease classifications differed across surveys. Specifically, when co-occurring alcohol and drug abuse was adopted as a survey classification for clients, those joint diagnoses were apportioned according to spending on single-diagnosis care. Missing total revenues from MH and SA facility surveys were imputed based on numbers of clients and facility characteristics. Estimates from data sources with small samples and high variance in estimates from year-to-year were smoothed. Estimates based on incomplete survey response rates were adjusted. Missing years of survey data were extrapolated and projected to 2001 when necessary. The costs of health insurance administration for MHSA coverage were estimated based on percentages from the NHA. Finally, an NHA-equivalent estimate was computed by eliminating a small proportion of expenditures for social services in order to compare MHSA estimates to total national spending.
Data. Table 1.2 lists the data sources used to develop the estimates, how they were used, and the years of data that contributed to the estimates. For specialty institutional providers, SAMHSA generally conducts censuses of facilities that treat mental or substance use disorders, through the Survey of Mental Health Organizations (SMHO, formally called the Inventory of Mental Health Organizations (IMHO)) and the National Survey of Substance Abuse Treatment Services (NSSATS, formally called the Uniform Facilities Data Set (UFDS)), respectively. Facility administrators answer these surveys and report data at the aggregate facility level (for example, total number of Medicaid clients or total revenues for clients treated for alcohol abuse).

For other providers, various data sources were used. These included administrative claims data and surveys that collect encounter-level or patient-level data. In some cases, these surveys often sample a first stage of providers and then a second stage of encounters between providers and patients. With characteristics on each encounter or patient, expenditures for specific diagnoses such as mental health, substance abuse, or all health care can be calculated.
Table 1.2: Data Sources for the MHSA Spending Estimates
Data Source Use in Spending Estimates Years Used
National Health Accounts (NHA)
  • National health care expenditures by provider and payer.
1986 - 1997, 1998, 1999, 2000, 2001
National Hospital Discharge Survey (NHDS)
  • Proportion of general hospital inpatient days devoted to MHSA diagnoses.
1986 - 1997 , 1998, 1999, 2000, 2001
National Hospital Ambulatory Medical Care Survey (NHAMCS)
  • Proportion of general hospital outpatient visits devoted to MHSA diagnoses.
  • Proportion of emergency room visits devoted to MHSA diagnoses.
  • Proportion of MHSA drug mentions during visits to general hospital outpatient departments and emergency rooms for MHSA.
1992 - 1997, 1998, 1999, 2000, 2001
National Ambulatory Medical Care Survey (NAMCS)
  • Proportion of physician office visits devoted to MHSA.
1985, 1990 - 1997, 1998, 1999, 2000, 2001 - office visits;
  • Proportion of MHSA drug mentions during physician office visits.
1985, 1992 - 1997, 1998, 1999, 2000. 2001 - drugs
National Nursing Home Survey (NNHS)
  • Proportion of nursing home residents with MHSA diagnoses.
1985, 1995, 1997, 1999
National Home and Hospice Care Survey (NHHCS)
  • Proportion of home health users with MHSA diagnoses.
1994, 1996, 1998, 2000
  • Payment ratios for MHSA and other disorders.
  • Proportion of physician bills for MHSA by inpatient, outpatient, and emergency room care.
  • Proportion of other provider bills (e.g., psychia-trists and home health agencies) for MHSA.
  • Average copayment amounts.
1995, 1996, 1997, 1998, 1999
IMS Health Inc. data
  • To verify NAMCS, NHAMCS, and MEPS prescription drug estimates.
1994 - 1997
Medicaid drug rebate data
  • To corroborate estimates from MEPS and MarketScan® for the ratio of MHSA prescriptions to non-MHSA drugs.
1994 and later
Healthcare Cost and Utilization Project, Nationwide Inpatient Sample (HCUP-NIS)
  • Charge differential between MHSA services and other health care services.
1988 - 2001
National Medical Expenditure Survey (NMES)
  • Distribution of payments among multiple payers for services.
Medical Expenditure Panel Survey (MEPS)
  • Distribution of payments among multiple payers for services.
  • Basic data on spending for psychologists and counselors.
  • Size, frequency, and cost of refills of prescription drugs by class of drug.
1996 - 2000
Economic Census, Health Care and Social Assistance Sector
  • Data on number of establishments and receipts for establishments based on the North American Industrial Classification System (NAICS) that now identifies several specialty MHSA providers: offices of physicians, mental health specialists, offices of mental health practitioners (except physicians), outpatient mental health and substance abuse centers, psychiatric and substance abuse hospitals, and residential mental health and substance abuse facilities.
CMS Medicare and Medicaid Statistics (in published reports and special tabulations)
  • Inpatient services provided by physicians by diagnosis group for Medicare patients.
  • Relative Medicare payments for physician services in offices, hospital outpatient departments, and emergency rooms.
  • Distribution of hospital-based nursing home, home health, and personal care agency payments out of total community hospital payments.
Alcohol and Drug Services Study (ADSS)
  • Expenditures in substance abuse specialty organizations.
Inventory/Survey of Mental Healthcare Organizations (IMHO/SMHO)
  • Expenditures in mental health specialty organizations.
1986, 1988, 1990, 1992, 1994, 1998
National Survey of Substance Abuse Treatment Services (NSSATS) / Uniform Facility Data Set (UFDS)
  • Expenditures in substance abuse specialty organizations.
1987, 1990, 1991, 1993, 1995, 1996, 1998, 2000
Changes from Prior Estimation Methodology
Current estimates reflect improvements resulting from suggestions made during the substantial review process for prior estimates and from the use of new data sources not available when prior estimates were developed. The changes result in more accurate estimates. Because the improvements are complex and involve various aspects of the estimation process, net dollar impact of a particular change was not determined.

For the current estimates, new data became available for almost all provider data sources. Two completely new data sources also were introduced. The Medical Expenditure Panel Survey (MEPS), collected by the Agency for Healthcare Research and Quality (AHRQ), was used to estimate aspects of "other providers" such as payment amounts and payer categories. MEPS is a nationally representative household survey that collects information on MHSA care by asking about the reason for medical care utilization and then assigning diagnosis codes to those reported reasons.

Another new database was the SAMHSA Alcohol and Drug Services Survey (ADSS). The ADSS was a one-time survey of the universe of substance abuse providers in the United States. It had certain advantages over the NSSATS/UFDS. In particular, it had a more complete universe than earlier UFDS. It also generated more accurate data on revenue through the use of a telephone survey. Thus, the ADSS results were used to adjust the data from NSSATS/UFDS.

Some surveys also changed their structure. The Survey of Mental Health Organization (SMHO), conducted by SAMHSA, replaced the prior census of specialty mental health facilities known as the Inventory of Mental Health Organization (IMHO). In contrast to the IMHO, revenue data from specialty MH organizations in the SMHO were not collected from all providers but rather from a representative sample.
The estimates in this report were prepared using standard estimating techniques and the best available survey information. They represent the only MHSA estimates comparable to total health care spending in the U.S. As in any effort of this type, multiple data sources were used to piece together and cross check information that ultimately formed the basis for these estimates. Each data source comes with its own set of strengths and weaknesses.

Adjustments were made through estimation techniques to compensate for potential identified problems that weaknesses may cause. Among the data-related problems addressed were unavailability of recent information, item-specific non-response or undisclosed information on surveys (i.e., missing information in specific fields), surveys that overlap providers, and inconsistency in survey questions from year to year - each of which will influence the accuracy of the estimates. For example, SAMHSA stopped collecting revenue data for specialty substance abuse facilities after 1998. Therefore, for estimates after 1998, revenues for specialty substance abuse facilities were imputed based on actual client counts. In addition, substantial survey and item non-response occurred in the substance abuse specialty facility data prior to 2000; therefore, estimates were adjusted using the 1996 ADSS survey.

Back to previous webpage

File Date: 9/24/2008