Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
9 Benefits and Costs of Prevention O n an intuitive level, preventing mental, emotional, and behav- ioral (MEB) disorders among young people is one of the soundest investments a society could make. The benefits include higher productivity, lower treatment costs, less suffering and premature mortality, and more cohesive familiesâand, of course, happier, better adjusted, more successful young people. Given the evidence that feasible actions can be taken to achieve these benefits, the case for action is compelling. Emerging evidence that some of these interventions are also cost-effective makes the case even stronger. In an analysis conducted for the committee, Eisenberg and Neighbors estimate that the annual costs of MEB disorders among young people totaled roughly $247 billion in 2007 (see Box 9-1). Demonstrating the effectiveness of interventions is necessary to establish a scientific basis for prevention approaches aimed at avoiding these costs. As outlined in this report, there is reason for optimism about the ability to successfully intervene in the lives of young people and prevent many negative out- comes. However, decisions about how to invest limited public resources must consider the cost of delivering the service and demonstrate that the benefits that can be expected from an interventionâboth those that can be readily valued in dollars (e.g., increased productivity, decreased treatment costs) and those that cannot (e.g., alleviation of pain and suffering of both â This chapter is based in part on a paper written for the committee by Daniel Eisenberg and Kamilah Neighbors in the Department of Health Management and Policy, School of Public Health, University of Michigan. 241
242 PREVENTING MENTAL, EMOTIONAL, AND BEHAVIORAL DISORDERS BOX 9-1 Methodology for Cost Estimates 1. Mental Health Service Costs a. Multiply Ringel and Sturmâs 1998 estimate of $11.7 billion by (73.7 + 29.45)/73.7 to expand age group to include ages 18-24 (they only included ages 0-17). b. Multiply by 2 to account for fact that their estimates do not account for full range of settings, as suggested by Costello, Copeland, and colleagues (2007). c. Inflate to 2007 dollars (multiply by 1.28), based on the Bureau of Labor Statis- ticsâ consumer price index (see http://www.bls.gov/cpi/). d. Multiply by population growth between 1998 and 2007 for people under age 25 (1.07) = $45 billion. 2. Health, Productivity, and Crime Costs a. Mental disorders: Multiply share of mental health and substance abuseâ related DALYs incurred by 0-24 age group (0.355), times National Institute of Mental Health (2002) estimate ($102 billion for 1995â$185 billion less the portion of total costs attributable to health care since counted in part 1), times inflation adjustment from 1995 to 2007 dollars (1.37), times population growth between 1995 and 2007 for people under 25 years old (1.07) = $54 billion. b. Drug abuse: Multiply share of mental health and substance abuseârelated DALYs incurred by 0-24 age group (0.355), times Office of National Drug Control Policy (2004) estimate ($165.1 billion for 2002â$180.9 billion less the $15.8 billion in health care costs since counted in part 1), times inflation adjustment from 2002 to 2007 dollars (1.15), times population growth between 2002 and 2007 for people under 25 years old (1.05) = $71 billion. c. Alcohol abuse: Multiply share of mental health and substance abuseârelated DALYs incurred by 0-24 age group (0.355), times Harwood (2000) estimate ($158 billion for 1998â$185 billion less the portion of total costs attributable to health care since counted in part 1), times inflation adjustment from 1998 to 2007 dollars (1.27), times population growth between 1998 and 2007 for people under 25 years old (1.07) = $77 billion. Total = $247 billion, which, divided by 104 million people ages 0-24, equals about $2,380 per young person. individuals and their families)âoutweigh the costs that would be incurred in a real-world environment. As one example of the complexity of measur- ing costs, a serious mental disorder in a parent or a child has obvious and measurable financial costs associated with treatment and lost productivity. However, the disorder also often profoundly affects the overall function- ing of the family in psychosocial ways that are devastatingly costly to the
BENEFITS AND COSTS OF PREVENTION 243 family but not readily susceptible to quantification, much less valuation in dollars and cents. This chapter opens with a brief tutorial on cost-benefit and cost- e Â ffectiveness analysis, as well as an explanation of what the terms âcost-beneficialâ and âcost-effectiveâ mean. The chapter then synthe- sizes existing knowledge on the benefits that could be achieved (namely, avoided costs) if prevention were widely implemented on a national scale. Next the available research on the benefits and costs of individual preven- tion programs or types of intervention is summarized. The chapter then discusses limitations of the available research and concludes by offering recommendations for future research in these areas. It is important at the outset to acknowledge the basic purpose and limitations of economic analysis in the context of prevention and preven- tion research. Economic analysis may be valuable at the beginning of prevention research by quantifying the costs associated with the disorder or problem being targeted for prevention, or at least those costs that lend themselves to quantification. This provides a sense of the potential value of prevention of the problem. Evaluating the cost-effectiveness of an intervention at the end of the prevention research cycle helps determine whether funding the intervention is a wise use of societal resources and hence desirable for dissemination. However, economic analysis has limi- tations as a decision-making aid. In particular, as mentioned above, even the best analyses are challenged to capture all of the psychological and emotional costs associated with MEB disorders in a manner that would be deemed universally acceptable. As a consequence, estimates of the cost of these disorders may misestimate the true social costs, possibly consider- ably. Equally importantly, economic analysis addresses efficiency but not equity. In some cases, a society or an organization may prefer investing in a less cost-effective program if it is more likely to reach disadvantaged populations. Also, particularly in the context of prevention, economic analysis may rely on a number of unproven assumptions (see Current Knowledge Regarding Intervention Benefit and Costs, below). The cost-effectiveness of an intervention also often depends on the perspective of the decision maker. In many cases, an intervention is cost- e Â ffective from the perspective of society as a whole, but not from the n Â arrower perspective of a single organization considering whether to fund the intervention. For example, consider the case of an investment in pre- vention of MEB disorders by a health care provider. That provider incurs the costs of the intervention and derives some of the benefits, in the form of reduced future costs of care. However, major social benefits of the inter- vention may be realized in other sectors of society, including the education s Â ector (e.g., when students are less disruptive in class) and the criminal justice sector (e.g., when recipients of the intervention are less likely to get
244 PREVENTING MENTAL, EMOTIONAL, AND BEHAVIORAL DISORDERS into trouble with the law). It is quite plausible that the health care orga- nization may not perceive the intervention as worthwhile from its narrow perspective, whereas from a social perspective the intervention is highly cost-Âeffective (see Chapter 11 for a discussion of implementation issues). Addressing the disjunction between those who bear the costs of an interven- tion and those who experience its benefits may require coordinated plan- ning of interventions and, if possible, aligning of incentives across service systems. COST-BENEFIT AND COST-EFFECTIVENESS ANALYSIS Cost-benefit analysis (CBA) and cost-effectiveness analysis (CEA) are two methods of economic analysis used to assess whether an intervention is desirable from an economic perspective; put simply, they evaluate whether the benefits derived from the intervention are worth the cost invested in the intervention. The principal distinction between the two techniques lies in the measurement of desired outcomes. In CBA, all such outcomes are valued in monetary units (dollars), permitting a direct comparison of the benefits produced by the intervention with its costs. When benefits exceed costs, the intervention is said to be cost-beneficial. When benefits fall short of costsâand assuming that one is comfortable that all important posi- tive outcomes have been captured in monetary termsâthe conclusion is that the intervention is not worth undertaking. CBA is the ideal form of analysis given that it allows a comparison of desired outcomes (benefits) and Âundesired outcomes (costs) in the same metric. This permits a precise conclusion about the desirability of the intervention. Is the intervention âworth itâ? CEA, in contrast, is used when one or more major desired outcomes cannot be readily measured in monetary terms but a major outcome, mea- sureable in another metric, is common to the interventions being compared. A notable example in the health care literature pertains to interventions that avoid preventable premature deaths (or preventable illness or disability). Historically, the principal outcome in published studies was measured in terms of life-years saved. Now, most commonly, outcomes are measured as quality-adjusted life years (QALYs). Analysts typically employ CEA when they think that the desired outcome does not lend itself readily to monetization. Thus, breast or prostate cancer screening and treatment avoid premature deaths, but as they do so primarily for people beyond their working years, many analysts are uncomfortable attributing a dollar value to the beneficiariesâ extra years. It is possible to do so, using a mea- sure of willingness-to-pay (Gafni, 1997). Since the desired outcomes and the Â undesired outcomes (costs) are measured in different metrics in CEA (life years and dollars, respectively), the bottom line of a CEA is a ratio,
BENEFITS AND COSTS OF PREVENTION 245 in this case cost per QALY. An intervention is deemed cost-effective if it produces the desired outcome at a reasonable price, typically the lowest cost to realize a QALY among competing interventions. Thus, if an analyst is comparing three different interventions, all other things being equal, the cost-effective intervention is the one for which the cost per QALY is the least. (This simplification ignores additional concernsâthe other things not being equalâsuch as who benefits from the extra life years.) Often, analysts will label cost-effective an intervention not compared directly with alterna- tive investments. In such instances, typically they are comparing their find- ings to a standard in the literature. As a rule of thumb, ratios in the range of $50,000 to $100,000 or lower per life year lost are generally considered cost-effective (Ubel, Hirth, et al., 2003). In theory, a well-designed CBA and CEA of the same intervention should yield the identical conclusion about the desirability of the interven- tion (Bleichodt and Quiggin, 1999). An intervention will be cost-Âeffectiveâ that is, cost less per unit of benefit than alternative interventionsâif its benefits exceed its costs and do so with a net benefit that is greater than that of the alternative interventions. In practice, however, because Â researchers often focus on somewhat different outcomes depending on the method being used, one cannot assume that CBA and CEA will yield identical con- clusions about intervention desirability. Furthermore, all of these analyses rest on assumptions related to the quantification and valuation of important outcomes, assumptions that can drive the conclusions reached. Indeed, stan- dard practice in CBA and CEA should include use of sensitivity analysis, a family of methods to evaluate whether bottom-line conclusions are sensitive to assumptions made in the analysis (Gold, Russell, et al., 1996). The health care literature is dominated by CEAs; that is, one finds rela- tively few CBAs (Hammitt, 2002). The principal reason is the inability, or reluctance, of analysts or policy makers to place dollar values on important health outcomes. As we describe below, however, the prevention field seems to be an exception: the majority of studies to date have employed CBA. ECONOMIC NEED FOR PREVENTION Prevention, by definition, is undertaken to avoid harmful outcomes; the potential benefits of prevention are therefore equivalent to the net harms, or costs, of those outcomes. MEB disorders among young people account â Ubel, Hirth, and colleagues (2003) assert that the cost-effectiveness threshold should be raised to $200,000 or more per QALY. â For example, researchers using CBA may ignore improvements in health-related quality of life, because other benefits, such as reduced crime and increased employment, are easier to quantify in dollar terms.
246 PREVENTING MENTAL, EMOTIONAL, AND BEHAVIORAL DISORDERS Mental, Emotional, and Behavioral Disorders Health of Individual Health of Others â¢ Mortality â¢ Family â¢ Health-related quality â¢ Victims of crimes of life â¢ Peers Economic Resources Informal Care Services Juvenile Justice Productivity Education Health Care Child Welfare FIGURE 9-1â Costs of mental, emotional, and behavioral disorders among young people. Fig9-1.eps SOURCE: Adapted from Eisenberg and Neighbors (2007). for considerable costs to the health care, child welfare, education, juvenile justice, and criminal justice systems, as well as enormous additional costs in terms of the suffering of individuals, families, and others affected (see Figure 9-1). The most direct and probably most significant economic cost is increased morbidity and decreased health-related quality of life of the individual experiencing a MEB disorder. The individualâs health problems, in turn, may lead to adverse conse- quences for other members of society, such as family members, victims of crime, and peers. Health problems typically also lead to additional costs, in the form of reduced productivity and earnings (Kessler, Heeringa, et al., 2008) and increased use of a range of social services. And, of course, MEB disorders place enormous stress on young people themselves and interfere with healthy development. Morbidity and Quality of Life MEB disorders among young people are associated with substantially increased morbidity and reduced health-related quality of life. These health
BENEFITS AND COSTS OF PREVENTION 247 problems are associated with psychological suffering (U.S. Public Health Service, 1999a) as well as increased risks of physical illnesses (Vreeland, 2007). These health consequences represent an enormous burden during childhood (Glied and Cuellar, 2003) and are also correlated with sig- nificantly increased risks to health and reduced productivity in adulthood (Kessler, Berglund, et al., 2005; Kessler, Ormel, et al., 2003). A young personâs mental disorder or substance abuse may also lead to negative health consequences for other members of society. For example, mental disorders lead to lost productivity and functioning not only for the children, but also for the parents and caregivers of the children (Tolan and Dodge, 2005). Untreated mental illness may also have intergenerational effects. Having a depressed mother, or having two parents with poor mental health, is associated with mental, behavioral, and emotional problems in children (Kahn, Brandt, and Whitaker, 2004; see also Chapter 7). Substance abuse, and to a lesser extent other MEB disorders, are also associated with more frequent risky behavior (such as driving under the influence) (Harwood, 2000), which often have substantial health repercus- sions for others. In addition, an individualâs health condition may affect his or her peers; in particular, substance abuse (Gaviria and Raphael, 2001) and suicidal behavior (Gould, Jamieson, and Romer, 2003) are thought to spread among peers via a contagion effect. To quantify the total health burdens posed by various illnesses and dis- orders, researchers with the Global Burden of Disease project of the World Health Organization (WHO) and the World Bank calculated disability- adjusted life years (DALYs) lost due to each health condition. This measure accounts for both morbidity (mainly measured by functional impairments) and mortality. For the United States, depression and alcohol use and abuse were among the top five sources of premature death and disability (Michaud, McKenna, et al., 2006). According to the most recent estimates by age group for the United States, in 1996 mental disorders and substance abuse accounted for 30 percent of DALYs lost by people under age 25 (calculation by Eisenberg and Neighbors based on Supplementary Material, Additional File 4 to Michaud, McKenna, et al., 2006). This represents by far the highest burden of any disease category for this broad age range. By more specific age intervals, the proportions were 3 percent for ages 0-4, 18 percent for ages 5-14, and 48 percent for ages 15-24. Given evidence that people with mental disorders are at greater risk for both communicable â This percentage was calculated by including all conditions in the Global Burden of Disease projectâs neuropsychiatric category except epilepsy and multiple sclerosis, which are not typi- cally considered mental disorders. Updated estimates for the United States, for the year 2005, will be available within the next few years, according to Catherine Michaud, the first author of the report used to generate the estimates here.
248 PREVENTING MENTAL, EMOTIONAL, AND BEHAVIORAL DISORDERS and noncommunicable diseases and that their disorders contribute to both intentional and unintentional injuries, the percentage may be even higher (Prince, Patel, et al., 2007). Economic Resource Costs Health problems associated with MEB disorders decrease productivity and significantly increase the utilization of services, thus reducing economic resources available to society for other purposes. Productivity During childhood and adolescence, when most people do not partici- pate in the labor market, the direct impacts of mental disorders and sub- stance use on economic productivity are small but real. Young people with MEB disorders may diminish the productivity of others closely involved in their lives, particularly family members. For example, the stress and unpredictability of having a child with a serious MEB disorder can interfere with parentsâ work lives (Busch and Barry, 2007), or a disruptive child in a classroom can interfere with other studentsâ learning. There may also be significant costs to the work or educational productivity of siblings (Fletcher and Wolfe, 2008). The indirect and long-term consequences are also likely to be large. These conditions interfere with young peopleâs ability to invest in their own human capital via education. Many studies show that poor mental health and substance use among young people are negatively related to participa- tion and performance in school (Diego, Field, and Sanders, 2003; Glied and Pine, 2002), as well as high school completion (Vander Stoep, Weiss, et al., 2003), important determinants of productivity in adulthood. These factors can increase risk for such behavioral problems as delinquent and antisocial behavior (Yoshikawa, 1994). Also, to the extent that MEB disorders in childhood carry over into adulthood, there will be further reductions in economic productivity. A large number of studies, many of which focus on depression, document that adults with mental illness and substance abuse disorders are less likely to be employed, and those who are employed work fewer hours and receive lower wages (see Ettner, Frank, and Kessler, 1997; Kessler, Heeringa, et al., 2008). Similarly, as adults, employees with mental â When aggregating the costs of mental disorders and substance abuse, it is important to keep in mind that productivity costs may already be reflected, at least to some extent, in measures of health burden, such as DALYs. Thus, one might be double-counting by claiming, for example, that a case of depression accounts for a certain number of DALYs in addition to productivity costs. This caveat, however, does not take away from the fact that in general productivity costs are large and important to consider in their own right.
BENEFITS AND COSTS OF PREVENTION 249 health or substance abuse disorders can reduce the productivity of other workers, particularly if the job affects the work of others (e.g., assembly line work). Utilization of Services As one would expect, mental disorders and substance abuse are strongly associated with increased utilization of mental health and substance abuse services. Ringel and Sturm (2001) estimated the annual national costs of mental health treatment for children under age 18, as of 1998, at $11.68 billion, or $172 per child. They found that expenditures were $293 per child for ages 12-17, $163 per child for ages 6-11, and $35 per child for ages 0-5. Adjusted to current dollars using the consumer price index, the annual national costs in 2007 would be $14.8 billion. We are not aware of analogous estimates for substance abuse treatment of young people, although estimates are available for adults for alcohol abuse Â (Harwood, 2000) and drug abuse treatment (Office of National Drug Control Policy, 2004). In the past 15 to 20 years, the mix of mental health services for young people has shifted from inpatient to outpatient settings (Ringel and Sturm, 2001), as in the adult population (Wang, Demler, et al., 2006). Also, as in the adult population, the relative treatment mix for childrenâs mental health has shifted from specialty settings to primary care (Wang, Demler, et al., 2006) and from therapy and counseling to medication (Glied and Cuellar, 2003) (although this latter shift was interrupted in 2003 by the Food and Drug Administrationâs warnings about the use of antidepressant medications for children) (Libby, Brent, et al., 2007). These changes are also not fully reflected in the estimates cited. Young people with MEB disorders have higher utilization of mental health services across a range of social service systems, not just health care. Costello, Copeland, and colleagues (2007) considered data from a range of settings and demonstrated that mental health service costs in health care set- tings represent only a modest fraction of the total costs incurred by children with mental disorders for these services. Using a sample of adolescents ages 13-16 in western North Carolina, they estimated that mental health service costs for adolescents with mental disorders equated to $894 per adolescent in the local population, with more than one-quarter (27 percent) of the total costs incurred in the school and juvenile justice systems. The overall estimate is over three times that in the Ringel and Sturm (2001) study, â This number is based on converting the total costs per 100,000 population in Table 2 in Costello, Copeland, and colleagues (2007) to total costs per person. The percentage attribut- able to the school and the juvenile justice systems is based on dividing the sum of these costs ($10.9 million and $13.2 million, respectively) by the total costs ($89.4 million).
250 PREVENTING MENTAL, EMOTIONAL, AND BEHAVIORAL DISORDERS which focused mainly on mental health service costs in health care settings. The findings of Costello, Copeland, and colleagues (2007) are consistent with other empirical studies showing that MEB disorders are associated with increased use of services in nonmedical settings, such as foster care ( Â Harman, Childs, and Kelleher, 2000), special education (Bussing, Zima, et al., 1998), and juvenile justice (Teplin, Abram, et al., 2002). Youth with MEB disorders who become involved with the juvenile justice system also often incur costs related to law enforcement and court expenses, detention, placement and incarceration, and other forms of treat- ment that are publicly provided (National Center on Addiction and Sub- stance Abuse, 2004). In addition, violent crimes can result in victim costs, such as medical care, treatment through public programs, and property damages to victims. The costs associated with all juvenile (under age 18) arrests in 2004 were estimated at about $14.4 billion (National Center on Addiction and Substance Abuse, 2004), and the costs of medical care, treatment through public programs, and property damages to victims of juvenile violence were estimated at about $95 million (Miller, Sheppard, et al., 2001). Although not all of these crimes were committed by young people with MEB disÂorders, overall costs of these disorders would be higher if the cost of relevant juvenile crimes were included with service use estimates. In addition, these health problems lead to significantly increased use of informal (unpaid) care by family members and others. For example, family members with a child with mental health care needs are more likely than family members whose children do not have these needs to reduce their working hours or stop working to care for their child (Busch and Barry, 2007). Using data from the Fast Track project, Foster, Jones, and colleagues (2005) estimated that each youth with conduct disorder incurs public costs of more than $70,000 over a seven-year period, with costs incurred by the juvenile justice, education, and general health care systems in addition to the mental health system. Similarly, a study in the United Kingdom (Scott, Knapp, et al., 2001) documented societal costs from childhood conduct disorder that extended into adulthood. Children who had diagnosed con- duct disorder at age 10 incurred public service costs by age 28 that were 10 times higher than those considered to have no problems and 3.5 times higher than those with conduct problems but not diagnosed with conduct disorder. This suggests that preventive interventions aimed at addressing behavioral problems before they reach the threshold for a diagnosis could yield significant savings.
BENEFITS AND COSTS OF PREVENTION 251 Estimates of Total Costs Comprehensive âcost of illnessâ studies quantify and aggregate, in monetary terms, the various costs associated with particular illnesses or disorders. Although there are many recent studies of this type in European countries, the most recent estimates in the United States correspond to 1995 for mental disorders (National Institute of Mental Health, 2000), 2002 for drug abuse (Office of National Drug Control Policy, 2004), and 1998 for alcohol abuse (Harwood, 2000). Aggregating service costs and health and productivity costs for individuals age 18 and older, the annual economic costs of mental disorders were estimated at $185 billion in 1995 (National Institute of Mental Health, 1999), the annual economic costs of drug abuse in 2002 were estimated at $180.9 billion (Office of National Drug Control Policy, 2004), and the annual economic costs of alcohol abused in 1998 were estimated at $185 billion (Harwood, 2000). These reports do not permit an estimate of costs specific to people from birth to age 24. However, in an analysis for the committee, Eisenberg and Neighbors used data in these reports to make a rough approximation, for the year 2007, by making the following two assumptions: (1) the full cost of services for this age group per person is twice as high as the mental health care costs per person estimated by Ringel and Sturm (2001) and (2) the population share of health, productivity, and crime-related costs for people ages 0-24 is 35.5 percent (a calculation based on Supplementary Material, Additional File 4 to Michaud, McKenna, et al., 2006). Under these assumptions, Eisenberg and Neighbors estimated that the total annual economic costs are roughly $247 billion as of 2007 (in 2007 dollars),10 or about $2,380 per person under age 25. This per-Âperson total includes about $500 in health service costs and $1,900 in health, productivity, and crime-related costs. Several caveats pertain to this estimate. Perhaps most notably, one would not be able to prevent all of these costs, no matter how much one â The authors measured health and productivity costs by estimating the lost or diminished income due to morbidity and mortality. This is typically called a human capital approach to valuing health. Estimates from the human capital approach tend to be lower than estimates from willingness-to-pay approaches and are typically considered lower bound estimates (Hirth, Chernew, et al., 2000). Note that they also accounted for costs to other members of society, such as informal care and crime. â Although the reports are not specific about the age groups included, one can infer that they apply to those age 18 and over based on the data sources used. â This is a conservative assumption in two respects. First, it is lower than the adjustment factor of 3-4 estimated by Costello, Copeland, and colleagues (2007). Second, treatment costs are rising over time; for example, Mark, Coffey, and colleagues (2005) found that mental health and substance abuse treatment costs for the full population increased from $60 billion in 1991 to $104 billion in 2001. 10âNote also that the estimate of total costs accounted for population growth.
252 PREVENTING MENTAL, EMOTIONAL, AND BEHAVIORAL DISORDERS invested in prevention. Not all MEB disorders are preventable, given cur- rent knowledge, and some may never be preventable. On one hand, from this perspective, the estimate of $247 billion overstates the potential value of prevention. On the other hand, this estimate includes only costs avoided from preventing disorders that would meet full clinical criteria and does not include costs that would be avoided from reducing problem behaviors and symptoms in the range in which symptoms are not severe enough to meet diagnostic criteria. These costs are generally not included in cost-of-illness studies, but they may be very large. From this perspective, the estimate of $247 billion understates the aggregate costs of MEB disorders among young people. As well, the estimate does not fully capture the quality of life of the children and their families. Quantifying the costs of MEB disorders among young people is useful as a way to approximate the potential value of prevention and to compare the burden of these disorders11 among young people with other disease burdens, but very few studies have addressed this topic. In general, as Hu (2006) describes, methodologies in cost-of-illness studies vary and often depend on several assumptions that require further study. In the context of MEB disorders among young people, one important next step for this research literature is to conduct a comprehensive cost-of-illness study for the United States that builds on previous studies, such as Harwood, Ameen, and colleagues (2000) and Ringel and Sturm (2001), and the estimates cre- ated for this report by Eisenberg and Neighbors (2007) and accounts for the substantial use of services outside medical settings shown by Costello, Copeland, and colleagues (2007). After the initial work is completed to refine the methodology and identify data sources, periodic updates will be much easier to produce. In addition, further research is needed to improve the ability to project lifetime consequences of mental disorders in child- hood. In particular, researchers face the challenge of disentangling con- founding factors from true causal relationships in observed relationships between mental disorders in childhood and later outcomes. Miller (in Biglan, Brennan, et al., 2004) provides a much higher esti- mate of $435.4 billion in 1998 ($557.3 in 2007 dollars) for the costs of problem behaviors among youth, defined as underage drinking, heroin or cocaine abuse, high-risk sex, youth violence, youth smoking, high school dropout, and youth suicide acts. More than half was attributable to suffer- ing and quality of life, with the balance consisting of work losses, medical spending, and other resource costs. Averaged across all youth, this would be an average cost of $12,300 per youth ages 12-20 ($15,744 in 2007 dollars). 11â The discussion that follows refers specifically to emotional and behavioral disorders rather than problems, as it is referring to costs associated with actual disorders.
BENEFITS AND COSTS OF PREVENTION 253 COST-EFFECTIVENESS OF PREVENTIVE INTERVENTIONS Although the potential benefits from preventing MEB disorders are clearly large, and there is a substantial and growing body of evidence documenting the positive outcomes of prevention interventions, relatively few evaluations have been conducted to assess the cost-effectiveness of the interventions. The evaluations that are available tend to be those associ- ated with the interventions with the longest follow-up and include some of the most successful programs. Similarly, cost-effectiveness evaluations tend to be limited to such areas as early childhood development, youth development, and prevention of violence, depression, and substance abuse, in which there has been more research overall. In addition, most of the favorable cost-effectiveness results apply to interventions for higher risk populations, although a small number of universal prevention programs have also been shown to be cost-effective. Aos, Lieb, and colleagues (2004) reviewed the economic analyses of a large number of relevant interventions. The authors conducted a compre- hensive and detailed review and analysis for the Washington State govern- ment of prevention and early intervention programs designed to (1) reduce crime; (2) lower substance abuse; (3) improve educational outcomes, such as test scores and graduation rates; (4) decrease teen pregnancy; (5) reduce teen suicide attempts; (6) lower child abuse or neglect; and (7) reduce domestic violence. In addition to the discussion below based in part on their analysis, we refer the reader to this study as a resource for additional empirical results as well as a detailed discussion of methodological issues. Early Childhood Interventions Perhaps the most heavily researched preventive programs are early childhood interventions for children from birth to age 5. Some of these programs are primarily home-based, whereas others are primarily cen- ter-based. In a meta-analysis of over 25 studies of home visitation pro- grams (by nurses or other trained professionals), Aos, Lieb, and colleagues (2004) concluded that the average benefits per child were about $11,000 and costs were about $5,000.12 The benefit-cost ratio has been shown to be higher for certain programs; for example, in an economic evaluation of the Nurse-Family Partnership Program, Karoly, Kilburn, and Cannon 12â All dollar values in this section are in 2002 or 2003 dollars. In addition to average Âeffects for this group of programs, Aos, Lieb, and colleagues also estimated the benefits of the Nurse- Family Partnership at $26,298 and the costs at $9,118 and the benefits of the HIPPY (Home Instruction Program for Preschool Youngsters) at $3,313 and the costs at $1,837. They esti- mated that benefits exceeded costs for the Comprehensive Child Development Program and the Infant Health and Development Program.
254 PREVENTING MENTAL, EMOTIONAL, AND BEHAVIORAL DISORDERS (2005) found that the program cost about $7,000 per child and produced total benefits of about $41,000 per child for the higher risk sample and about $9,000 per child for the lower risk sample.13 In general, some of the main benefits of home visitation programs, converted into dollar estimates of their value, have been reduced child abuse, improved achievement test scores, and decreased likelihood of arrest later in life. The benefits from reduced child abuse are generally estimated on the basis of reductions in medical, child welfare, and other public service costs and crime costs, based on epidemiological evidence showing correlations between child abuse and these costs later in life. Improved achievement test scores are usually valued on the basis of how earnings relate to education. Finally, arrests are valued in terms of both the costs to the criminal justice systems and victims (particularly health costs for crimes involving injuries) and lost productivity while incarcerated (see also the technical appendix to Aos, Lieb, et al., 2004). Several different center-based early interventions also appear to have benefits that exceed their costs (see Targeting Early Childhood Devel- opment in Preschool in Chapter 6 for further discussion of these pro- grams). In a meta-analysis of over 50 studies of early childhood education programs for low-income 3- and 4-year-olds, Aos, Lieb, and colleagues (2004) found that, on average, benefits per child were $17,000 and costs were $7,000. In an economic analysis of the Abecedarian Early Childhood project, an intensive, multiyear intervention for children from birth to age 5, ÂBarnett and Masse (2006) found that per-child benefits were $158,000 and costs were $63,000; the primary benefits were related to cognitive abilities and education, which were valued in terms of estimated impact on future earnings. The intervention was also associated with a reduction in smoking, which was valued in terms of estimated reduction in pre- mature mortality (with a year of life then valued at $150,000, based on willingness-to-pay estimates in the literature). The Perry Preschool proj- ect, which included 1-2 years of intensive preschool, home visiting, and group meetings of parents, had estimated per-child benefits of $240,000 and costs of $15,000 (Belfield, Nores, et al., 2006); the primary benefits, some of which were observed well into adulthood, were reduced crime, positive academic outcomes, and reduced smoking. The Chicago Child- Parent Centers, a center-based preschool education for disadvantaged children, had estimated benefits per child of $75,000 and costs of $7,400 (Temple and Reynolds, 2007); the primary benefits were improved aca- demic outcomes and reduced crime. Temple and Reynolds (2007) compared the benefit-to-cost ratios of 13â Although the estimates provided by Aos, Lieb, and colleagues (2004) and Karoly, Kilburn, and ÂCannon (2005) differ, the difference between benefits and costs is substantial for both.
BENEFITS AND COSTS OF PREVENTION 255 the Perry Preschool project, the Carolina Abecedarian project, and the Chicago Child-Parent Centers to other types of interventions designed to benefit childrenâs development. They concluded that preschool education has a more favorable benefit-to-cost ratio than the Special Supplemental Nutrition Program for Women, Infants, and Children, the Nurse-Family Partnership, a class-size reduction initiative for grades K-3, and the Job Corps. There has been debate, however, regarding the benefits and costs of pre-K programs, including Head Start (Cook and Wong, 2007), the most heavily funded and widespread early childhood education program. Ludwig and Phillips (2007) attempt to resolve the debate by pointing out that Head Start costs about $9,000 per child, and would need to produce academic achievement gains only on the order of .1 to .2 standard deviations to con- fer equivalent benefits. They argue that the evaluation literature on Head Start strongly favors a benefit of this size or more, and that the program should be viewed as cost-beneficial. Heckman (1999, 2007) argues that investments in early childhood development, particularly for disadvantaged children, have greater payoff in terms of the development of skills needed for future success than do investments in any other period of life. A sys- tematic review of economic analyses of programs targeting mental health outcomes or accepted risk factors for mental illness by Zeichmeister, Kilian, and colleagues (2008) concluded that, among the few available studies, the most favorable results were for early childhood education programs. Youth Development Interventions Although comprehensive interventions in early childhood have prob- ably received more attention from scholars and policy makers, many com- prehensive interventions for older (school-age) children and adolescents also appear to be cost-effective. Aos, Lieb, and colleagues (2004) found that five of the six youth development programs reviewed,14 whose aims include improving parentâchild relationships and reducing problem behaviors, such as substance use and violence, are cost-beneficial, with benefit-cost ratios ranging from 3 to 28. These authors also found that several programs for juvenile offenders, with a range of goals mostly pertaining to improved behavior, are highly cost-effective, yielding net benefits per child well over $10,000 in many cases. 14â The programs determined to have benefits that exceed costs include the Seattle Social Development project, Guiding Good Choices, the Strengthening Families Program for Parents and Youth 10-14, the Child Development project, and the Good Behavior Game. CASASTART was determined to have costs exceeding benefits.
256 PREVENTING MENTAL, EMOTIONAL, AND BEHAVIORAL DISORDERS Interventions Targeted at Specific Mental, Emotional, and Behavioral Disorders or Substance Use There are currently few economic analyses of interventions that target the prevention of specific MEB disorders or substance use among young people, although there are a large number of studies that document effi- cacy, effectiveness, or both (see Chapters 6 and 7). The interventions in these economic analyses address depression, violence and conduct disorder, and substance use. Lynch, Hornbrook, and colleagues (2005) performed a cost-effectiveness analysis of a highly successful group cognitive-behavioral therapy intervention to prevent depression among adolescent children of depressed parents (see Box 7-4). They found that the intervention is very likely to be cost-effective, with an incremental cost of $610 per child and a cost-effectiveness ratio of $9,275 per QALY15 (95 percent CI, â$12,148 to $45,641). Foster, Jones, and Conduct Problems Research Group (2006) found that the Fast Track intervention, designed to reduce violence and conduct disorders among at-risk children, was about 70 percent likely to be cost-effective in preventing conduct disorder for the higher risk group, but it had less than a 0.01 probability of being cost-effective for the lower risk group, which represented the majority of the sample. Aos, Lieb, and colleagues (2004) found that 10 of the 12 substance use prevention pro- grams (including two programs that focus on smoking prevention) they analyzed were highly cost-effective, with benefitâcost ratios ranging from 3 to over 100.16 The estimated benefits per child were generally small (less than $1,000 in most cases), but the costs were even smaller (less than $200 in all but one program). Current Knowledge Regarding Intervention Benefits and Costs Overall, knowledge about the benefits and costs of specific interven- tions aimed at preventing MEB disorders is promising but still limited. Rela- tive to the number of efficacious or effective interventions (see Chapters 6 and 7), few investigators have conducted cost-effectiveness or cost-benefit analyses. There is also considerable uncertainty about many of the estimates in the available literature. For interventions that exhibit dramatically dif- 15â QALYs, like DALYs, are measures of health that account for both morbidity and mortality. As mentioned earlier, ratios under $50,000 per life year lost are generally considered cost- effective; this is regardless of whether life years are adjusted for quality of life (Ubel, Hirth, et al., 2003). 16â The programs determined to have benefits that exceed costs include the Adolescent Tran- sitions Program, Project Northland, Family Matters, Life Skills Training, Project STAR, the Minnesota Smoking Prevention Program, the Other Social Influence/Skills Building Substance Prevention Program, Project Toward No Tobacco Use, All Stars, and Project Alert. DARE and STARS for Families were ineffective, making the costs exceed the benefits by definition.
BENEFITS AND COSTS OF PREVENTION 257 ferent levels of benefits compared with costs, this uncertainty may be moot, but in other cases, it is important to consider carefully. Perhaps the most important source of uncertainty pertains to longer term outcomes. In many economic evaluations, longer term outcomes of participants in an intervention are not observed and instead must be pro- jected on the basis of other data. Many long-term benefits of early preven- tion programs cannot be measured until middle childhood and adolescence (e.g., juvenile crime). Longitudinal data used to make projections, such as correlations between the incidence of MEB disorders in childhood and in adulthood, do not necessarily represent accurate causal estimates, as Foster, Dodge, and Jones (2003) note. Another important source of uncertainty is a lack of statistical power. As Mrazek and Hall (1997) observe, many studies in this literature have modest sample sizes and are not sufficiently powered to look at key measures of effectiveness; typically, adequately powered estimates of cost-effectiveness require even larger samples than estimates of effectiveness per se (Ramsey, McIntosh, and Sullivan, 2001). A third, related source of uncertainty results from the outcomes measured: that is, whether interventions that appear to be cost-effective in reducing risk fac- tors closely connected to MEB disorders, but do not measure disorders as an outcome, can actually prevent the incidence of these disorders. Another source of uncertainty includes potential differences between cost-efficacy and cost-effectiveness. Evaluations of interventions conducted in research settings (efficacy studies) may get different results if conducted in real-world settings (effectiveness studies), raising potential questions about whether the cost-effectiveness (or more accurately, cost-efficacy) would be realized if the intervention were implemented in a nonresearch environment (see Foster, Dodge, and Jones, 2003, for a brief discussion of this). Similarly, the costs of interventions implemented in real-world settings may differ from the costs in a research setting. In addition, as discussed in more detail in Chapter 11, a major challenge in prevention research, particularly when dealing with whole communities, is that preventive interventions are likely to have differential impact on individuals in different contexts because (a) participants have different risk and protective factors that cause different responses to the intervention; (b) the level of participation in interventions varies; and (c) interventions are routinely delivered with varying levels of fidelity and adoption. These factors can reduce overall impact compared to that seen in efficacy trials; thus some analyses of behavioral or economic outcomes in community implementation studies may not find significant effects. There are challenges in measuring the cost of the time of children and other people involved in interventions. Those challenges can lead to poor estimates of costs, creating either an over- or underestimate. Often, however, analysts omit such time costs, introducing a clear bias toward
258 PREVENTING MENTAL, EMOTIONAL, AND BEHAVIORAL DISORDERS underestimating total costs. For example, some studies do not consider the opportunity cost incurred by teachers delivering an intervention who might otherwise be engaged in productive teaching activities (Aos, Lieb, et al., 2004). Finally, and importantly, other intangibles, most notably the suffer- ing of children and their families, are likely to be costly but extremely dif- ficult to quantify and assign a monetary value. The difficulty in measuring and valuing these costs restricts the potential of CBA and CEA to accurately evaluate the relative merits of preventive interventions for MEB disorders, which may lead to a substantial underestimation of the benefits of success- ful interventions. Research needs to be devoted to improving measurement methods that will permit assessment of the economic value associated with suffering related to these disorders. Another important caveat is that the quality of the underlying evidence used to project costs and benefits varies. Aos, Lieb, and colleagues (2004) account for this in their meta-analysis by assigning different weights to studies based on indicators of quality, but such a solution has unavoidable limitations, as the authors acknowledge. Many evaluations do not meet some of the important guidelines for quality of evidence, as stated by such organizations as the Food and Drug Administration (1998) and the Society for Prevention Research (Flay, Biglan, et al., 2005). For example, evaluators have not always published a specific plan of analysis before collecting data, which leaves open the possibility of selectively reporting positive results among many outcomes and analytical approaches. A final caveat for this literature is the reminder that, while some Âstudies employ CEA, most of the studies in the prevention field have employed CBA. In practice, CEA and CBA results are not strictly comparable. However, in this literature, because most of the studies yield strong conclusions (posi- tive in most cases), it is unlikely that the basic findings would be sensitive to the choice of method. As this literature evolves and more interventions with borderline cost-Âeffectiveness are evaluated, examining the sensitivity of conclusions to alternative assumptions will be important. CONCLUSIONS AND RECOMMENDATIONS The potential value of prevention of MEB disorders among young p Â eople is enormous. MEB disorders among young people result in sig- nificant costs to multiple service sectors. Such disorders threaten childrenâs future productivity and wellness and disrupt the lives of those around them. Conclusion: The economic, social, and personal costs of MEB disorders among young people are extraordinarily high.
BENEFITS AND COSTS OF PREVENTION 259 To date, there is some evidence that the benefits of some specific inter- ventions outweigh the costs. However, the scientific literature on the cost- effectiveness of prevention is still young, and it faces a number of conceptual and practical obstacles. Conclusion: The current body of research on costs, cost-effectiveness, and cost-benefits of preventive mental, emotional, and behavioral inter- ventions is very limited. Much of the strongest evidence to date is for interventions that improve protective factors or reduce risk factors demonstrated through research to be closely related to MEB disorders (see Chapter 4). For example, multiple economic evaluations of early childhood development programs have dem- onstrated benefits that exceed costs. It is also notable that among the limited number of interventions shown to be cost-effective, many were either targeted to higher risk children (e.g., the early childhood programs such as the Perry Preschool project) or were cost-effective only for a higher risk subgroup within the analysis (e.g., the Fast Track study). Aside from a small number of substance use prevention programs (see review by Aos, Lieb, et al., 2004), few universal interventions have been demonstrated to be cost-effective for preventing MEB disorders. Future research is needed to determine whether selective and indicated prevention programs are inherently more likely to be cost-effective in the context of MEB disorders, or if this finding is an artifact of the programs that happen to have been subjected to economic evaluations thus far. Conclusion: Of those few intervention evaluations that have included some economic analysis, most have presented cost-benefit findings and demonstrate that intervention benefits exceed costs, often by substan- tial amounts. However, few studies measure effects on diagnosable MEB disorders as an outcome, and most do not conduct sufficient longitudinal follow-up to fully capture potential long-term benefits. Also, considerable uncertainty remains about some of these estimates. Economic analyses are important for quantifying the potential value of prevention and assessing the actual value of existing interventions. Many scholars in the prevention field have called for more regular eco- nomic analyses (Flay, Biglan, et al., 2005; Spoth, Greenberg, and Turrisi,
260 PREVENTING MENTAL, EMOTIONAL, AND BEHAVIORAL DISORDERS 2008; Zeichmeister, Kilian, et al., 2008).17 Many preventive interventions have been shown to be highly effective but have not yet been evaluated for cost-effectiveness in real-world settings. Guidelines on how to conduct high-quality cost-effectiveness Âstudies are needed to help shape the develop- ment of this area of research as it continues to evolve. Recommendation 9-1: The National Institutes of Health, in consulta- tion with government agencies, private-sector organizations, and key researchers should develop outcome measures and guidelines for eco- nomic analyses of prevention and promotion interventions. The guide- lines should be widely disseminated to relevant government agencies and foundations and to prevention researchers. For interventions involving young people, long-term outcomes are often pivotal for determining cost-effectiveness, as significant benefits are likely to accrue into adulthood, yet current knowledge is remarkably weak in most contexts. Long-term follow-up data should be collected whenever possible. As electronic data systems become more integrated and acces- sible, one promising avenue is through administrative databases, which do not necessarily depend on expensive efforts to track down and interview participants.18 CEAs should also make clear the various sources of uncer- tainty. If the cost-effectiveness results are dramatically positive or nega- tive, wide intervals may not raise questions about the overall conclusion that an intervention is cost-effective, but publishing such information will make the assessment more transparent. Special attention should be given to addressing the fact that costs from an intervention in one sector may be evident in other sectors. While this has been done for early childhood, less attention has been focused on this issue in other developmental stages, such as adolescence. Economic analyses should also be comprehensive in their accounting of relevant costs and benefits. The work by Costello, Copeland, and colleagues (2007), for example, illustrates the importance of measuring costs across a range of service venues. Again, integration of electronic data systems may be a valuable tool for capturing these costs. To capture the benefit of reduc- tions in specific MEB disorders, interventions should measure diagnostic outcomes whenever possible. 17â This is an issue not only for prevention but also for treatment of mental disorders in chil- dren. A comprehensive review of economic evaluations of child and adolescent mental health interventions (most of which are treatment, not prevention) found only 14 had been published to date, although the authors speculated that two or three times that many would be in print within five years (Romeo, Byford, and Knapp, 2005). 18â Of course, researchers would need to overcome hurdles related to informed consent and privacy restrictions.
BENEFITS AND COSTS OF PREVENTION 261 Evaluations should begin to address the fact that multiple interventions over the span of childhood may have important dynamic complementarities (Heckman, 2007). For example, participation in an early childhood inter- vention such as Head Start may enhance a childâs ability to benefit from a later intervention to prevent substance use. Although it would be difficult to randomize children to different sequences of interventions over a long time span, empirical research to address these complementarities to the extent possible would be very informative. Similarly, understanding the causal links between aspects of poverty (e.g., food insecurity, disadvantaged neighborhoods, low-quality schools) and mental health should be improved. These links may reveal some of the most important mechanisms by which to prevent MEB disorders in cost- effective ways, but it is very difficult to establish incontrovertible causal relationships due to the many likely confounders in observational data.19 While there have been calls for increased economic analyses, the num- ber of projects that include calculation of costs and cost-effectiveness will increase only if guidelines on how to conduct these types of analyses are widely available and the additional costs recognized. Recommendation 9-2: Funders of intervention research should incorpo- rate guidelines and measures related to economic analysis in their pro- gram announcements and provide supplemental funding for projects that include economic analyses. Once available, supplemental funding should also be provided for projects with protocols that incorporate recommended outcome measures. Although one might argue that grant awards should be increased rather than providing supplemental funding to those that conduct eco- nomic analyses, there is a precedent for providing supplemental funding in other areas. For example, the National Institutes of Health (NIH) provides research supplements for projects involving underrepresented minori- ties and individuals to improve the diversity of the research workforce. Although these supplements are modest, NIH has reported that they are an effective means of encouraging institutions to recruit from currently underrepresented groups. Evaluations of the costs and cost-effectiveness of prevention inter- ventions will increase only if researchers include them in their protocols. S Â tudies designed to determine the effectiveness of interventions in a real- world setting should be clear not only on what the intervention costs, so that a community can judge the feasibility of funding the project, but also 19â Fordiscussions of links between poverty and mental health among children, see, for ex- ample, Ripple and Zigler (2003) and the Center on the Developing Child (2007).
262 PREVENTING MENTAL, EMOTIONAL, AND BEHAVIORAL DISORDERS the cost-effectiveness, or expected benefits, so that the community can determine the potential value of their investment. Recommendation 9-3: Researchers should include analysis of the costs and cost-effectiveness (and whenever possible cost-benefit) of inter- ventions in evaluations of effectiveness studies (in contrast to efficacy trials). Finally, cost-benefit and cost-effectiveness studies of mental health pro- motion interventionsâscarce in the literature to dateâwould be very useful in permitting a meaningful comparison of the relative desirability of preven- tion and promotion approaches. In concluding this discussion, it is important to note that the significant societal benefits of preventing mental, emotional, and behavioral problems among young people may warrant intervention even when there is no spe- cific cost-effectiveness data available, particularly if there is evidence that an effective intervention is available. Waiting for future cost-effectiveness analyses to become available, which might take years to develop, would put many young people at unnecessary risk.