National Spending on Mental Health and Substance Abuse Treatment by Age of Clients, 1997 Henrick J. Harwood, Tami L. Mark, PhD, David R. McKusick, PhD, Rosanna M. Coffey, PhD, Edward C. King, MA, James S. Genuardi, MA Tami Mark (Tami.Mark@thomson.com) and Rosanna Coffey are with Thomson Medstat. Henrick J. Harwood is with The Lewin Group, David R. McKusick, Edward C. King, and James S. Genuardi are with the Actuarial Research Corporation. Introduction Age is one of the most revealing lenses through which one can examine spending on mental health and substance abuse (MH/SA) treatment. There are several reasons why. The epidemiology of MH disorders and SA is st rongly related to age and presents distinct challenges for treatment of various age gr oups. Furthermore, public policy on health financing is often linked to th e age of the beneficiaries. For example, the elderly and young (along with the severely disabled) have been the primary recipients of public financing of health services : about 90% of Medicare populat ion is age 65 and older, and over 50% of the Medicaid population is younger than 18 years. This study is one of the first to analyze the age distribution of national spending on MH/SA services and is the first to look at the full age spectrum of MH/SA clients. The study builds on the comprehensive MH/S A spending estimates developed under Substance Abuse Mental Health Servi ces Administration (SAMHSA) Spending Estimates Project, which calculated spendi ng on MH treatment at $73.4 billion and on SA treatment at $11.9 billion for 1997. 1 1 That study found that between 1987 and 1997, MH/SA spending had increased 3.7% annually, versus 5% for all personal health care and public health spending. * MH spending grew slightly fast er, at 4%, compared to 3.4% for SA spending. Public payers reimbursed a disproportionate share of MH/SA treatment costs compared to costs for treatment of other health problems; 9.9% of public payer health spending was for MH/SA, compared to 6% of private spending. Only one other published study was found that examined MH/SA expenditures by age. Ringel and Sturm 2 estimated treatment expenditures for ch ildren aged 122617 for the year 1998. They used many of the same data sources used in the SAMHSA Spending Estimates Project, as well as additional survey data. Use of the Medical Expenditure Panel Survey (MEPS), which was not available when the SAMHSA es timates were developed, allowed them to allocate spending not only by ag e and type of service but also by type of payer. They put total MH/SA spending for children aged 122617 at $11.68 billion in 1998. Adolescents (aged 1222617) accounted for 60% of the total a nd had the highest expenditures per child, children aged 622611 accounted for 34% of the total, while children aged 12265 accounted for 6% of the total. They also learned that in 1998, private insuran ce covered the largest proportion of spending on these youth (46%), while Medicaid (a t 24%) and State and *The national MH/SA spending includes some expenditures that are not counted in the National Health Accounts (such as custodianship of group homes), but when this article compares MH/SA spending and NHA spending it uses MH/SA NHA-equivalent estimates, which exclude those other non-health services. National Spending on Mental Health and Substance Abuse Treatment by Age of Clients, 1997 1 local payers (at 21%) covered comparably sized shares of expenditu res for children 022617 years of age. Prevalence of MH/SA by Age The prevalence of MH/SA disorders differs across major age groups: children/adolescents, adults, and elders. Wh ile past-year prevalen ce ofMHdisorders is about 20% for each of the broad age groups, sp ecific disorders have different prevalence at different ages. 3 Among children and adolescents aged 922617 years, about1%to4%experience a conduct disorder, while this disorder is virtually undiagnosed in adults and older adults. At the other end of the age spectrum, severe cognitive impairment is experienced by 7% of adults aged 55 and older, while very low rates are seen in children and adults.3 Anxiety di sorders appear among youth, adults, and the elderly at 13%, 16%, and 11%, respectively, as do mood disorders (such as depression and bipolar disorder) at 6%, 7%, and 4%, respectively. The abuse of and dependence on alcohol and illi cit drugs is also strongly linked to age. Peak prevalence is typically seen in young adul ts and rapidly declines as adults age. Grant et al, using the 1992 National Longitudinal Alcohol Epidemiologic Survey, found that the annual rate of alcohol dependence or abuse was about 16% for age 1822629 years. In contrast, less than 1% of those 65 a nd older had alcohol dependence or abuse. 4 Although drinking before age 21 is illegal, the 2000 National Household Survey on Drug Abuse showed that some children begin in preteen years, with about 2.4% of 12-year- olds consuming at least some alcohol in a month. This figure rises to 30% for 1622617- year-olds.5 Illicit drug use in the past month was about 3% for ages 1222613, rising to 20% for 1822620-year-olds, and declini ng to no more than 1% for those older than 65 years. According to the survey, approximately 4.1 million people required illicit drug abuse treatment in 2000, and 10 million needed treatme nt for alcohol use disorders (ie, were heavy users of alcohol). 5 Because the prevalence of MH/SA disorders diffe r across the life cycle, one might expect the spending on MH and SA treatment to va ry across broad age groups. SA disorders are highly age related, with low rates for young children, rapidly increasing prevalence from preteens to adults, the highest rates for young adults, and very low prevalence for older adults. Thus, for SA treatment expenditures, one might expect higher spending for adolescents and adults than for young childre n and the elderly. However, for all MH disorders combined, the prevalence patterns are relatively even across age groups. Thus, for MH treatment expenditures, one might exp ect fairly comparable levels across broad age groups of youth, adults, and the elderly. Spending, however, may differ for reasons other than overall prevalence rates. For exam ple, the types of MH/SA services provided and the associated costs may differ by age and the rates of identifica tion of disorders may also differ by age group. Methods The estimates presented here disaggregate the 1997 SAMHSA national estimates of spending on MH/SA treatment into 3 age groups (022617, 1822664, and 65 and older) for various categories of service relevant for MH/SA treatment.1, 6 National Spending on Mental Health and Substance Abuse Treatment by Age of Clients, 1997 2 Included diagnoses Like the SAMHSA national spending estimat es, the estimates presented here include most of the MH disorders contained in the Diagnostic and Statisti cal Manual, Fourth Edition (DSM-IV) . Spending for several 223groups224 of di agnoses and services was excluded because policymakers and financial systems tend to treat them like other health care conditions (i.e., mental retardation, developm ental disabilities, a nd severe cognitive impairment like Alzheimer222s and senile deme ntia) rather than as a distinct MH/SA disorder. The estimates presented here also do not include medical services obtained as a consequence of MH/SA (for example, accidents/ trauma, liver disease, or HIV infection) or MH/SA services delivered secondary to ca re for other health pr oblems. While treating these conditions costs $57 billion annually, these conditions were excluded from these analyses because they are generally covered by general medical insura nce, either private or public, unlike much of MH/SA treatment. Als o, costs to society such as social services for those who are mentally ill or criminal just ice costs associated with SA are excluded. Only primary MH/SA diagnoses were consid ered. Thus, for example, expenditures on persons with a primary diagnosis of cancer and a secondary diagnosis of depression would not be allocated to MH. Although this ma y result in overly conservative estimates, there is no way to accurately determine what proportion of expenditures for treatments with MH/SA secondary diagnoses should be allocated to the MH/SA condition. Overview of approach The MH/SA treatment spending figures were developed from the National Health Accounts (NHA) and other data sets using 2 different methods for general health care providers and MH/SA specialty providers. 7 The process for each type of provider is described below. The Appendix lists the data sources used to calculate each service. General sector provider estimation For general health care providers, such as primary care physicians, the method was to determine the proportion of the NHA (by type of provider, diagnosis, and age) spent on MH/SA treatment. Large, nationally representa tive data sets were used to identify the proportion of utilization (eg, visits to a physicia n, hospital days) of treatment of primary MH/SA diagnoses and to identify price diffe rences between MH/SA services and other disorders. In the data sets, each encounter r ecord (eg, a hospital discharge record) had an associated ICD-9-CM diagnosis and those pertaining to MH/S A disorders were used to identify MH/SA treatment. Age-specific utiliz ation and age-specific price factors within each diagnosis class were calculated to develop an estimate of the total spending on a particular type of general health care service for each age group. To provide an example: assume that the NHA indicates that $100 billion was spent on general hospital treatment in 1997. Assume that 10% of hospital days are for persons aged 1822664 with a MH diagnosis as measured using the National Hospital Discharge Survey (NHDS). Assume that the average pri ce per day for this group is 10% lower than the average price for persons with general health conditions as determined using the MarketScan database. Then spending for MH care in general hospitals by persons aged 1822664 would be $9 billion, or 9% of the total NHA spending. National Spending on Mental Health and Substance Abuse Treatment by Age of Clients, 1997 3 Data sources for utilization factors incl uded the National Hospital Discharge Survey (NHDS), the National Ambulatory Medical Care Survey (NAMCS), the National Hospital Ambulatory Medical Care Survey (NHAMCS), the National Nursing Home Survey (NNHS), and the National Home and Hospice Care Survey (NHHCS). To reduce yearly sampling error on surveys with small sa mples, utilization data were averaged over 3 consecutive years closest to 1997. It shoul d be noted that the NHAMCS data for 1996 showed significantly different results from other years because of changes in the coding of payer data; therefore, the 1996 NHAMCS wa s excluded from the estimation process. The source for price fact ors was the MarketScan256 database, an annual collection of private health insurance claims from large em ployers. Price differentials for a particular age group were computed from the ratio of the average approved amount per claim for the age group to the average approved amount for claimants of all ages. Occasionally, when price data for an age-diagnosis-s ervice group were una vailable, or when calculations for an age group were out of line (30% above or below the price for all ages combined), the price ratio was replaced with a value of 1.0 so that the prices for all ages combined would be used for that age group. The only service that was not identified usi ng a primary MH/SA diagnosis as indicated on the encounter records was pharmaceuticals. The databases used to measure utilization of pharmaceuticals227 the NAMCS and NHAMCS 227provided nationally representative data on prescriptions written during physician office visits, hospital outpatient visits, and emergency room visits. The type of medication prescribed is indicated by its name and 4- digitNDCclassification rather than by the indicated diagnosis. Six classifications of psychopharmacologic drugs were used in thes e surveys to identif y MH/SA medications: (1) sedatives and hypnotics, (2 ) antianxiety, (3) antipsycho tic, (4) antidepressant, (5) central nervous system (CNS) stimulant a nd anorexiant, and (6) miscellaneous CNS drugs. Specialty sector provider estimation For specialty providers, information on MH/S A spending was estimated directly from surveys of the universe of sp ecialty providers. These include d the inventory of Mental Health Organizations (IMHO), a census of MH service facilities in the United States, and the Uniform Facility Data Set (UFDS), a census of US substance abuse facilities. In these data sets, aggregate information on the number of patients by class of diagnosis (MH, SA, other) was used to allocate total re venues by facility by diagnosis. Note that 223other224 made up approximately 2.5% of patie nts. These surveys asked providers about the age distribution of clients served. Data on the distribution were th en used to allocate facility revenues by age group. F acilities with no client-age di stribution were assigned the average distribution of comparable facilities (e.g., based on size, service type, modality, geographic region, and ownership status). Data were adjusted to account for non- response on utilization and revenue items th rough imputation. Variables in the regression imputation model included service type (inpa tient, residential, outpatient), ownership, region of the country, client da y, and type of modality (si ngle or multiple modalities). The most recent spending estimates that could be developed for the specialty facilities (1996 from UFDS and 1994 from IMHO) were trended forward to 1997. Projections of spending on each service used the Centers fo r Medicare and Medica id (CMS) projection National Spending on Mental Health and Substance Abuse Treatment by Age of Clients, 1997 4 method that breaks overall spending growth into components227prices, utilization, and demographic changes. They also used the histor ic changes in age dist ributions by service. Comparison to all health care spending The age estimates for MH/SA spending in 1997 were compared to age estimates of national spending on all health care. These health care figures by age were derived from the NHA. For all health spending, whenever the surveys mentioned previously were insufficient for deriving service unit expend itures (such as the average expense for a hospital stay) by age and paye r, the 1997 MEPS was used. In comparing the resulting national all-heal th expenditures to MH /SA expenditures, one other detail must be acknowledged. First, because the total MH/SA spending included some costs (such as custodianship of group homes) that are not on heal th services per se, and therefore would not be counted in the NHA, a second estimate of MH/SA spending227the MH/SA NHA-equivalent expe nditures227was calculated for accurate comparison between MH/SA and all health care spending. Thus, in this study of spending by age of client, figures are presented for both 223total MH/SA expenditures224 and 223MH/SA NHA-equivalent expenditures.224 Only 4% of total expenditu res relate to non-health expenditures by specialty MH/SA pr oviders in these estimates. Variation from prior nationa l SAMHSA spending estimates There were some differences between thes e age-group estimates a nd the prior SAMHSA estimates by Coffey et al.1 In these estimates of MH/SA expenditures by age group, insurance administration expenses were excl uded because of the lack of such data by beneficiary age. Also, in this current study of spending by ag e, some aspects of spending on pharmaceuticals for MH treatment that had been inadvertently excluded from the original study were included227specifically, spending on methylphenidate (Ritalin). Thus, the national estimates in this article are the revised, slightly higher, estimates of MH/SA spending. Finally, unlike the prior estimate s of MH/SA for the whole US population, these estimates by age group do not divide spending by payer type because of a lack of data on the age distribution of clients by payer. Results Table 1 presents estimates of national spending for treatment of MH/SA by age of clients treated in 1997. For completeness, 223total MH /SA expenditures224 are also shown, but only 223MH/SA NHAequivalent expenditures224 are comp ared with all health care spending in Tables 1 and 2. The majority of MH/SA spending, about 72%, supported the treatment of adults between the ages of 18 and 64 in 1997, yet only 51% of all health care spendi ng related to adults aged 1822664. Children and youth younger than 18 comprised about 13% of MH/SA spending and about 11% of all health care spending. For adults 65 years of age and older, the reverse was true. Only 15% of MH/SA spending went toward treatment of MH/SA disorders of the elderly, while 38% of a ll health care spending was on elders. Table 1 also reveals the MH/SA spending estimates by diagnosis for MH, SA, and both combined. For each of these diagnoses, the per centage spent across the age groups also is National Spending on Mental Health and Substance Abuse Treatment by Age of Clients, 1997 5 displayed. Table 2 examines MH/SA spending as a percentage of a ll health care spending by age group. Table 1. Estimated 223Revised224 Spendi ng on Mental Health and Substance Abuse (MH/SA) Treatment by Age and Diagnosis Compared to All Health Care Spending, 1997 MH/SA NHA Equivalent spending (millions) MH NHA Equivalent spending (millions) SA NHA Equivalent spending (millions) All health care spending (millions) Client age group Total MH/SA spending (millions) Amount % Amount % Amount % Amount % Total $82,437 $79,433 100.0 $68,465 100.0 $10,968 100.0 $1,057,494 100.0 Younger than 18 $10,975 $10,646 13. 4 $10,042 14.7 $604 5. 5 $116,150 11.0 1822664 years $59,072 $56,7 96 71.5 $46,949 68.6 $9,84 6 89.8 $537,609 50.8 65 and older $12,390 $11,991 15.1 $11,474 16.8 $518 4. 7 $403,735 38.2 Source: SAMHSA Spending Estimates Project. These 223revi sed224 dollar estimates differ slightly from prior estimates in Coffey et al1 because pharmaceutical ex penses for MH treatment were revised. Estimates exclude administrative insurance expenses. The $84, 243 million above is 0.007% higher than the $81,849 million estimate previously given for total MH/SA spending, and the same increase occurred for MH/SA NHA-equivalent spending, which was $78,845 million in the earlier report. There are material differences in the age patt erns of spending for MH and SA treatment. Youth and older adults accounted for larger proportions of spending on MH services (15% and 17%, respectively) than of spendi ng on SA services (6% and 5%, respectively) in 1997. Adults aged 1822664 accounted for 69% of spending on MH, and 90% of spending on SA services. The lower rate of spendi ng on SA services among clients younger than 18 and older than 64 probably reflects th e lower treatment prevalence ofSAdisorders among these 2 age groups compared to adults aged 1822664. As Table 2 shows, overall MH/SA spending cons tituted about 8% of all health spending in 1997. For children and adolescents, MH/S A spending represented more, 9% of all health care spending. For adults aged 1822664, MH/SA spending represented even more, 11% of all health car e spending, while for adults ages 65 and older, MH/SA spending was a much smaller proporti on of all of their health care spending, only 3%. The age patterns were most pronounced for SA treatment expenses, with over 3 times the share of all health spending occurring fo r adults (2% of all health care spending was for SA treatment) as compared with youth (0.5%) and elders (0.1%). In contrast, MH spending constituted a similar proporti on (9% and 11%) of all health spending for youth and adults younger than 65, respectively, although it represented a smaller proportion for those 65 and older (only 3%). The fact that children ar e relatively healthy a nd have relatively low general health service use and th at elders are much higher users of general health services helps to explain the larger proportion of MH/SA among children than among the elders. In addition, it should be remembered that th e proportion of SA spending on youth reflects the low prevalence of SA among very young chil dren and the much higher level of SA among adolescents. Breakouts by smaller age gr oups were not possible in this study. National Spending on Mental Health and Substance Abuse Treatment by Age of Clients, 1997 6 Table 2. Estimated 223Revised224 Spendi ng on Mental Health and Substance Abuse (MH/SA) Treatment, by Age and Diagnosis, 1997 MH/SA NHA-Equivalent spending (millions) Percentage of all health spending Client age group MH/SA MH SA All health spending MH/SA MH SA Total $79,433 $68,465 $10,968 $1, 057,494 7.5% 6.5% 1.0% Younger than 18 years $10,646 $10,0 42 $604 $116,150 9.2% 8.6% 0.5% 1822664 years $56,796 $46,949 $9,846 $537,609 10.6% 8.7% 1.8% 65 and older $11,991 $11,474 $51 8 $403,735 3.0% 2.8% 0.1% Source: SAMHSA Spending Estimates Project. These 223revi sed224 dollar estimates differ slightly from prior estimates in Coffey et al.1 because pharmaceutical expenses for MH treatment were revised. Estimates exclude spending on insurance administration. T he $79,433 million above is 0.007% higher than the $78,845 million estimate previously given for total MH/SA spending, and the same increase occurred for MH/SA NHA-equivalent spending, which wa s $78,845 million in the earlier report. Table 3 examines MH/SA spending estimates by type of service, using the 223total MH/SA spending224 estimates (rather than MH/SA NHA- equivalent estimates). Overall, 74% of MH/SA spending was with specialty MH/SA providers. This figure would have been even higher if psychotropic medi cations could have been associated with the specialty of the professional prescribing the medication (for example, psyc hiatrist or non- psychiatrist physicians) or were wholly counted in the sp ecialty sector. Clearl y, specialty providers deliver a preponderance of MH /SA care, although general se rvice providers are seeing and treating an appreciable number of MH/SA patients. Across the 3 age groups some distinct differe nces emerged in where patients received their MH/SA care. About 85% of spendi ng for youth was with specialty MH/SA providers, compared to 76% for younger and middle-aged adults and 51% for older adults. There were 2 major differences in pa tterns of care between youth and other adults. First, youth had a much lowe r share of spending from pr escription pharmaceuticals (6%) than did adults aged 1822664 (14%). Second, children and adolescents were more likely than the other age groups to have gotten ca re from organized MH/S A providers such as MH clinics and residential treatment cente rs in contrast to independent MH/SA practitioners such as psychiatrists, psyc hologists, counselors, and social workers. Twenty-five percent of total MH/SA expenditures on children were for care in multi- service MH organizations as compared to 14% and 8% for adults and older adults. One of the most significant differences across the age groups in the service received was that 33% of MH/SA spending for older adults went to nursing home care, compared to essentially nothing for the other age groups. Prior analyses found th at total nursing home MH/SA expenditures were unchanged in nominal dollars between 1987 and 1997, and declined as a share of total MH/SA expenditures from about 10% to 6%. During this time period, Federal law required nursi ng home residents to be scr eened for mental illness and placed in appropriate settings. However, it appears that mental illness was still the primary or first-listed diagnosis for a bout 6% of nursing home expenditures. Overall, 13% of MH/SA spending was in either specialty psychiatric or SA hospitals (the payments for the 3 age groups were 14%, 13% , and 10% respectivel y). The pattern of variation was reversed for spending in speci alty MH/SA treatment units affiliated with general hospitals (16% overall and 19%, 16% , and 15% for the 3 age groups). Less than National Spending on Mental Health and Substance Abuse Treatment by Age of Clients, 1997 7 3% of MH/SA spending was related to services being delivered in hospital units that did not specialize in the treatment of MH/SA. Table 3. Mental Health and Substance Abuse (MH/SA) Expenditures by Age Group and Provider, 1997 Percentage of MH/SA spending on each age group by service type Type of provider Total MH/SA spending (millions) by service type MH/SA clients, all ages Younger than 18 years 18-64 years 65 and older General service providers (tot al) $21,863 26.5 15.1 24.0 48.6 Non-specialty hospital care 206 $2,331 2.8 2.3 2.9 2.7 Non-psychiatrist physicians $4,718 5.7 6.4 5.5 6.3 Freestanding home health $428 0.5 0.2 0.4 1.6 Freestanding nursing home $4,722 5.7 0.0 1.1 33.0 Retail prescription drugs 207 $9,664 11.7 6.2 14.2 5.0 Specialty providers (total) $60,574 73.5 84. 9 76.0 51.4 Specialty hospitals 247 $10,746 13.0 14. 3 13.4 10.0 Specialty units of general and VA hospitals $13,371 16.2 18. 7 16.0 15.0 Psychiatrists $7,39 6 9.0 7.6 9.9 5.8 Other MH/SA professionals ** $10,147 12.3 9.8 13.1 11.0 Residential treatment centers (formerly for children only) $2,807 3.4 6.8 3.1 1.7 Multi-service MH organizations 206206 $12,135 14.7 25.1 14.3 7.6 Specialty SA centers 207207 $3,974 4.8 2.5 6.2 0.3 Total expenditures $82, 437 100.0 100.0 100.0 100.0 Source: SAMHSA Spending Estimates Project. These 223revi sed224 dollar estimates differ slightly from prior estimates in Coffey et al.1 because pharmaceutical expenses for MH treatment were revised. Estimates exclude insurance administration expenses. The $82,4 37 million above is 0.7% higher than the $81,849 million estimate previously given for total MH/SA spending, and the same increase occurred for MH/SA NHA-equivalent spending, which was $78,845 million in the earlier report. 206223Nonspecialty hospital care224 includes care provided in general hospitals outside of specialized psychiatric and SA units. 207223Retail prescription drugs224 refers to drugs obtained through retail (pharmacy or mail order) distribution. Inpatient drug treatment and facilities that dispense drugs through public programs, such as methadone clinics, are not included in this category, but rather are included with the specific facility expenditures. 247223Specialty hospitals224 include psychi atric and SA hospitals, which specia lize in MH or SA treatment, and all of their hospital-based services. **223Other MH/SA professionals224 includes psychologists, counselors, social workers, and nurse practitioners. 206206 223Multiservice MH organizations224 includes a variety of providers such as community MH centers, residential treatment facilities for the mentally ill, and partial care facilities. 207207223Specialty SA centers224 includes freestanding SA centers and units of other facilities. Thus, for example, it includes methadone maintenance clinics, other facilities that primarily serve persons with SA problems, and units of public health clinics, charitable organi zations, correctional facilities, and other entities. Some of these organizations have SA as their primary mission and others treat SA as a secondary function. National Spending on Mental Health and Substance Abuse Treatment by Age of Clients, 1997 8 Less than 6% of MH/SA spending was for services provided by non-MH/SA-specialty physicians, with each age group spending about the same percentage for these providers. Implications for Behavior al Health Services This analysis provides new data on nationa l spending on MH/SA treatment by 3 client age groups for 1997: youth (022617 years), adults (1822664 years), and olde r adults (65 years and older). In the next round of estimates , spending by age groups over a 10-year period will be derived. Also, methodological improvements will be undertaken to allow estimation of source of payment for MH/SA se rvices by age. With these advances, the effect of program changes on services to sp ecific age groups, such as the State Child Health Insurance Program, could be evaluated. In addition, this analys is could not refine the age estimates for children, for example, examining estimates for young children and adolescents separately. Future analyses may be able to discern trends among subgroups of youth. There is one other study that has produced MH/SA spending estimates for a specific age group.2 That study estimated that MH/SA spending on youth aged 122617 was $11.68 billion in 1998. This figure is very close to the $10.98 billion estimate for the year 1997 for youth aged 022617 presented in this analysis. Th e difference is comparab le to the rate of medical care price inflation between the 2 years. On one hand, the similarity is not surprising given that many of the same data se ts were used in both studies; on the other hand, the similarity is unexpected given the complexity of the estimation process, the potential for different methods , and the newer data sources used in the later study by Ringel and Sturm. The similar findings serv e as a partial corr oboration of the study methods. The data presented in this article have a number of important potential uses and implications. They provide a baseline agains t which future estimates can be compared. For example, the pace of MH/SA expenditure growth during the 1980s was of concern to private employers, particularly expe nditures on children and adolescents. 8 Such concerns, however, were based on data from one employe r or convenience samples. In the future, one should be able to use the estimates pr esented here to examine trends in MH/SA expenditure growth by age group and to addres s directly the adequ acy of spending and cost containment concerns for these age groups. Expenditure data by age is also important fo r planning and evaluation of public programs. One implication of this study for public payers, such as state Medicaid programs, is that policymakers need to pay attention to MH/SA service planning and evaluation specifically aimed at children since more th an 1 in 11 dollars of all MH/SA expenditures are likely to be spent on services to childre n. The share-of-spending estimates can also be used as benchmarks for state-level estimate s. A Medicaid program, for example, that spent less than 5% of its me dical expenditures on MH/SA ma y have created barriers to access of MH/SA services. Expenditure distributions by provider also suggest that planning and evaluation of MH/SA for diffe rent age groups should focus on different services. For the elderly, for example, efforts might focus on care in nursing homes, while for children efforts might focu s on multi-service MH organizations. National Spending on Mental Health and Substance Abuse Treatment by Age of Clients, 1997 9 The results of this study of 1997 data show that significant diffe rences existed across major age groups in spending on MH/SA. Adults received a mix of services across the specialty and non-specialty sectors. One question raised by these expenditure distributions is whether the mix is appropria te. For example, 14% of expenditures were for prescription medications, is this level of medication use appropr iate? Is it supplanting or augmenting behavioral treatment? Data from this study indicat ed that although MH/SA servi ce expenditures were highest among adults aged 1822664, expenditures on children were also significant227exceeding $10 billion. Indeed, a number of MH/SA research, tr eatment, and financing issues are unique to children and adolescents. Given that 9% of medical care expend itures on youth went to MH/SA treatment, efforts must be made to ensure that psychiatric and SA treatment research focuses on youth and does not merely attempt to transfer results from adults to the youth population. For example, the most recent comprehensive epidemiologic survey measuring the prevalence of MH disord ers227the National Comorbidity Survey227was limited to individuals aged 1522654. Similarly, clin ical trials of medi cations often exclude children despite the fact th at a growing aspect of MH/SA treatment on youth involves medications. Institutional hospital services were a larger share of expenditures on youth younger than 18 than of adults aged 1822664. More than 35% of expenditure s on children222s MH/SA treatment were for care in hospitals. For children, expenditures on residential services were second only to those for the elderly. Aggressive managed care has limited access to and ultimately the supply of inpatient and residential treatment opportunities over the past 10 years.6, 9 This analysis illustrates that these policies may have had more of an impact on youth than on other age groups since they ar e heavy users of inpa tient and residential services. Whether the decline in institutiona l use was beneficial and whether current levels of institutional care us e by youth are appropriate clearly need further elucidation.3 In addition, whether inpatient services can be provided in community settings has important implications for quality of life as well as cost-effectiveness. As President Bush222s New Freedom Commission on Mental Health recently documented, there are numerous barriers and gaps that prevent chil dren and adults from receiving effective community treatment ( http://www.mentalhealthcomm ission.gov/reports/reports.htm ). One limitation of this analysis was that because it is consistent with the approach taken in the national health accounts it may exclude spending on school-based MH/SA services where children may receive a significant shar e of their MH/SA care. Future studies need to document expenditures on MH/SA treatment in schools and its relationship to services provided in other settings. Among senior adults, MH/SA spending was a less significant portion of total health care spending but an important dimension nonetheless. MH/SA spending constituted only about 3% of all health care spending for those 65 and older. This lower share of spending on MH/SA treatment in senior adults as compared to younger gr oups certainly is relate d to the substantial expenditures on physical health that occur in th e last decades of life for most adults and to the fact that dementia is excluded from the spending estimates shown here because it is National Spending on Mental Health and Substance Abuse Treatment by Age of Clients, 1997 10 a disorder typically covered under medical in surance rather than the MH/SA portion of the benefits. However, the finding that 33% of this older-adult MH/SA spending was for nursing home care (excluding dementia patien ts) merits the attention of financing specialists and providers, especially because public policies have attempted to move such care into community settings. Overall, the low MH/SA spending proportion for the elderly raises que stions since it is known that roughly 20% of the older adult population needs MH car e227very similar to the proportion of children and a dults with such requirements.3 Because older adults may present with different clinical symptoms th an do other adults, such as more somatic complaints and comorbid medical conditi ons, detection of MH/SA conditions by non- specialists may be particularly difficult and require extra efforts. In summary, analyses of the economic size of the MH/SA treatment system by age group can provide important information to policym akers and program planners charged with making decisions about resour ce allocation. Such informati on can also make clear to researchers and providers the importance of ag e considerations in the design of studies and interventions so that im portant segments of the populat ion in specific treatment settings are not ignored. Acknowledgments A contract from the Center for Substan ce Abuse Treatment (CSAT) and Center for Mental Health Services (CMHS) of the Substance Abuse and Mental Health Services Administration (SAMHSA) supported the prepar ation of this article (Contract No. 270- 96-0007). The content is solely the responsibil ity of the authors and does not necessarily reflect the official views of SAMHSA, or its components CSAT or CMHS, or the Department of Health and Human Services. References 1Coffey RM, Mark T, King E, et al. National Estimates of Expenditures for Mental Health and Substance Abuse Treatment, 1997 . Rockville, Md: Center for Substance Abuse Treatment and Center for Mental Health Services, Substance Abuse and Mental Hea lth Services Administration; July 2000. SAMHSA Publication No. SMA-002263499. 2Ringel JS, Sturm R. 223National Estimates of Mental Health Utilization and Expenditures for Children in 1998224. The Journal of Behavioral Health Services & Research . 2001;28(3):319226333. 3Surgeon General of the United States. Mental Health: A Report of the Surgeon General . Washington, DC: US Dept of Health and Human Servi ces, Public Health Service; 1999. 4Grant BF, Harford TC, Dawson DA, et al. 223Prevalence of DSM-IV Alcohol Abuse and Dependence: United States, 1992224. Alcohol Health Research World . 1994;18:243226248. 5Office of Applied Studies (OAS). 2000 National Household Survey on Drug Abuse . Rockville, Md: US Dept Health and Human Services, Substance Abuse and Mental Health Services Administration; 2001. 6Mark TL, Coffey RM, King E, et al. 223Spending on Mental Health and Substance Abuse Treatment, 1987- 1997224. Health Affairs . 2000;19(4):107-120. National Spending on Mental Health and Substance Abuse Treatment by Age of Clients, 1997 11 7Braden BR, Cowan CA, Lazenby HC, et al. 223National Health Expenditures224. Health Care Financing Review . 1998;20(1):83226126. 8Frank RG, Salkever DS, Sharfstein SS. 223A New Lo ok at Rising Mental Health Insurance Costs224. Health Affairs . 1991;10(2):116226123. 9Witkin MJ, Atay JE, Mandersheid RW, et al. 223Highlights of Organized Mental Health Services in 1994 and Major National and State Trends224. In: Mental H ealth United States, 1998. Rockville, Md: Substance Abuse and Mental Health Services Administration; 1999. USDHHS Publication Number (SMA) 99-3285. National Spending on Mental Health and Substance Abuse Treatment by Age of Clients, 1997 12 Appendix Data sources for estimati ng MH/SA providers/services * Provider/service Utilization Average payments Payment channels Nonspecialty care in general hospitals NHDS, NHAMCS HCUP, MarketScan256 , Medicare claims, Medicare statistics NMES, MarketScan256 Psychiatrists NAMCS, HCFA Medicare statistics MarketScan256 , Medicare statistics NMES, MarketScan256 Nonpsychiatrist physicians NAMCS, HCFA Medicare statistics MarketScan256 , Medicare statistics NMES, MarketScan256 Other nonphysician professionals MEPS MarketScan256 NMES, MarketScan256 Freestanding nursing homes NNHS Freestanding home health NHHCS MarketScan256 NMES, MarketScan256 Retail prescription drugs NAMCS, NHAMCS, IMS MarketScan256, Medicaid drug rebate data NMES, MarketScan256 Residential treatment centers IMHO Specialty hospitals IMHO/UFDS Specialty units of general hospitals and VA hospitals IMHO/UFDS Specialty substance abuse centers IMHO/UFDS Multiservice MH organizations IMHO/UFDS *NHDS indicates National Hospital Discharge Surv ey; NHAMCS, National Hospital Ambulatory Medical Care Survey; HCUP, Healthcare Cost and Utiliza tion Project; NMES, National Medical Expenditure Survey; NAMCS, National Ambulatory Medical Care Survey; NNHS, National Nursing Home Survey; NHHCS, National Home Health Care Survey; IMHO, Inventory of Mental Health Organizations; UFDS, Uniform Facility Data Survey; HCFA, Health Care Financing Administration. National Spending on Mental Health and Substance Abuse Treatment by Age of Clients, 1997
