Prevalence of DSM-IV Disorders and Attendant Help-Seeking in 2 American Indian Reservation Populations FREE

Based on national studies,^1- 2 the Surgeon General’s report entitled Mental Health: Culture, Race, and Ethnicity³ concluded that African American and Hispanic individuals may not be at differential risk for mental disorder once demographic differences are taken into account. Regardless, significant mental health disparities likely exist for other ethnic minority populations, American Indians among them.⁴ Although American Indians have not been represented in sufficient numbers in these national efforts to permit independent estimates, other research specific to this population has revealed high rates of disorder, especially related to alcohol use and trauma.^5- 10 Unfortunately, the latter studies used case ascertainment, sampling methods, and measures of demographic correlates and help-seeking that preclude clear comparisons with national and international projections.

Clinically informed ethnographic work argues against the injudicious extension of current psychiatric epidemiologic methods to American Indian communities. Anthropologists have found that mental health problems may be described or understood in some American Indian communities in ways that diverge from mainstream conceptualizations. For instance, among the Hopi, the symptoms constituting major depression are involved in at least 5 illness categories¹¹; further, idioms such as “loneliness” are common markers of depression in other tribes.¹² Such examples argue for the careful consideration of culture both in the assessment of disorder and in the interpretation of subsequent findings. Perhaps less obvious, but no less important, are the cultural differences found within the larger American Indian population, which includes more than 300 federally recognized tribes, with an additional 200 Alaska Native groups.¹³ Consequently, rates of mental disorders have been thought to vary considerably among tribes³; however, little evidence has been available to support this conclusion.

Seminal work in the Epidemiologic Catchment Area Study demonstrated that relatively few individuals who have psychiatric disorders seek help in the specialty mental health care sector.¹⁴ Thus, describing help-seeking patterns has become integral to studies of this nature, which require special care in the unique services ecology in American Indian communities where traditional healers often play a critical role.¹⁵ The Indian Health Service (IHS) is the primary source of biomedical services in many reservation communities. Yet, this agency is dramatically underfunded,¹⁵ particularly with respect to mental health services,¹⁶ leaving its contribution to addressing local need unclear. In American Indian communities, alcohol and drug treatment has proceeded quite independent of treatment for physical and mental health problems, and, historically, communication between these systems has been rare.¹⁵ As a result, the comorbidity among these types of disorders has only recently been recognized, with only preliminary attempts at more comprehensive interventions.¹⁷ Thus, a fuller appreciation both of the comorbidity of mental health and substance use problems and of associated help-seeking has important implications for current debates regarding federal health policy toward American Indians.¹⁸

The American Indian Service Utilization, Psychiatric Epidemiology, Risk and Protective Factors Project (AI-SUPERPFP) was designed to address these gaps in the literature. In this article, we report on the prevalence of common DSM-IV disorders in 2 culturally distinct, reservation-based communities using a culturally adapted version of the Composite International Diagnostic Interview (CIDI). We also describe the comorbidity of mental and substance disorders, reporting the demographic correlates and help-seeking patterns associated with these disorders. This work offers unprecedented insight into the patterns of psychiatric disorders and service use in this special population and challenges assumptions about uniformity across American Indian tribes.

The primary objective of the AI-SUPERPFP was to estimate the prevalence of psychiatric disorders and accompanying service use within 2 American Indian reservation populations. The population of inference was 15- to 54-year-old enrolled members of 2 closely related Northern Plains tribes and a Southwestern tribe living on or within 20 miles of their respective reservations at the time of sampling (1997). (To protect the confidentiality of the participant communities, we use the general descriptors of Northern Plains and Southwest rather than specific tribal names.)¹⁹ The communities in question belong to different linguistic families, have different histories of migration, subscribe to different principles for reckoning kinship and residence, and have historically pursued different forms of subsistence. Yet both tribes have many experiences common to many American Indian groups. They share similar histories of colonization, including dramatic military resistance, externally imposed forms of governance, forced dietary changes, mandatory boarding school education, and active missionary movements. Although based on quite different epistemological perspectives, traditional systems of healing are active in both tribes. Unemployment is widespread, but within each tribe there is considerable variability in acculturation, education, and income. Thus, selection of these 2 tribes provided an opportunity to simultaneously examine both similarities and differences across tribes within a population that is relatively small yet extremely diverse.

Tribal rolls formed the sampling universe; these records list all individuals meeting minimal requirements for recognition as tribal members. A critical point for the AI-SUPERPFP was that tribal enrollment coincided with eligibility for IHS services, the major service provider in rural American Indian communities and a focus of our services research.

Stratified random sampling procedures were used, with tribe, age, and sex as strata. Records were selected randomly for inclusion into replicates, which were then released as needed to reach the goal of approximately 1500 interviews per tribe. Altogether, 46.6% and 39.2% of those listed in the Southwest and Northern Plains tribal rolls were found to be living on or near their reservations. Of those located and found eligible, 73.7% in the Southwest tribe (n = 1446) and 76.8% in the Northern Plains tribes (n = 1638) agreed to participate, with response rates lowest for men and younger tribal members. Sample weights, used in all analyses presented here, accounted for differential selection probabilities across all strata and for nonresponse biases. The AI-SUPERPFP methods are described in greater detail elsewhere²⁰; our Web site, http://www.uchsc.edu/ai/ncaianmhr/presentresearch/superprj.htm, provides copies of the interview and the training manual.

Tribal and institutional review board approvals were obtained prior to data collection. All adult participants provided informed consent; parental or guardian consent was obtained before requesting minor assent. Interviews were computer-assisted and -administered by tribal members intensively trained in research and interviewing methods. Extensive quality control procedures verified that location, recruitment, and interview procedures were conducted in a standardized, reliable manner. These included relocating 10% of those deemed ineligible because they lived away from the reservations, verifying 10% of the refusals, and reviewing more than 10% of the audiotaped interviews to ensure that questions were read verbatim with appropriate tempo and that the interviewer established suitable rapport with the participant.

Concerns about confidentiality were aggressively addressed. Each staff member signed a confidentiality statement that included the specific provision that any breach would be grounds for immediate dismissal. Each interviewer carried a copy of this policy to show to participants. Furthermore, field supervisors (also tribal members) ensured that interviewers were not assigned to potential participants who were family members or otherwise well known to them.

Psychiatric disorders were measured with the University of Michigan version of the CIDI (UM-CIDI),² adapted for use in American Indian communities in the context of a previous project.⁷ As described in more detail elsewhere,²⁰ focus group reviews by community members and biomedical and traditional service providers led to several changes. Psychoses and mania were excluded because of concerns regarding the cultural validity of the measures (for instance, asking about hallucinations in cultures where the seeking of visions is nurtured must be more nuanced than is possible in a structured lay interview). Simple and social phobias and agoraphobia were excluded because of concerns about respondent burden. Other changes included provision of definitions of seldom-used words (eg, period) or the deconstruction of complex questions (eg, “Has there been a period of 2 weeks or more when you thought a lot about death—either your own, someone else’s, or death in general?” was deconstructed into 3 questions). The UM-CIDI provided for diagnoses based on DSM-III-R criteria, but the AI-SUPERPFP Team completed a careful analysis of the changes in diagnostic criteria made in DSM-IV and added items for the assessment of DSM-IV disorders (the World Health Organization’s CIDI 2.0²¹ was not generally available when this work commenced). Thus, the AI-SUPERPFP CIDI yielded both DSM-III-R and DSM-IV diagnoses of the following disorders: major depressive episode (MDE), dysthymic disorder, generalized anxiety disorder, panic disorder, posttraumatic stress disorder (PTSD), alcohol abuse, alcohol dependence, drug abuse, and drug dependence.

Analyses of the MDE diagnoses indicated that among these samples, requiring that symptoms co-occurred and were not due to illness, medications, or substance use dramatically deflated the prevalence rates and decreased their diagnostic validity. Further inquiry suggested that the cognitive demands of assessing the multiple time frames in the context of a list of symptoms proved especially foreign and difficult for study participants in this population.²² Thus, the MDE rates reported here were derived using simplified diagnostic criteria based only on the endorsement of at least 5 of the 9 MDE symptoms. A clinical significance criterion for MDE, dysthymic disorder, generalized anxiety disorder, and PTSD was implemented as described elsewhere²³ and required that participants report either “a lot” of distress or impairment due to the symptoms of disorder or “some” distress and “some” impairment. Aggregations included any depressive disorder (MDE or dysthymic disorder), any anxiety disorder (those with generalized anxiety disorder, panic disorder, or PTSD), any depressive and/or anxiety disorder, any substance disorder (alcohol or drug abuse or dependence), and any disorder. Because we were particularly interested in the co-occurrence of depressive and/or anxiety and substance disorders in the context of help-seeking, we also categorized individuals by whether they exhibited depressive and/or anxiety disorders only, substance disorders only, or comorbid depressive and/or anxiety and substance disorders.

We examined the relationship of disorder to tribe and sex group (Southwest women as referent group), age (15-24 years, 25-34 years, and 35-44 years compared with ≥45 years), formal educational attainment (high school or general equivalency diploma and some postsecondary education compared with less than high school), employment status (student and unemployed compared with working), and marital status (separated/widowed/divorced and never married compared with married/cohabitating).

Questions about help-seeking were included in each diagnostic module and were asked of all individuals who endorsed at least some symptoms of the disorder. These questions were patterned after the UM-CIDI² and adapted to reflect the service ecologic environment of American Indian reservation communities, including questions about traditional healers (eg, medicine men and spiritual or religious leaders), expanded examples of specialty care (from both mental health and substance abuse professionals), and other medical professionals. Based on focus group suggestions, the wording in these questions was altered from the questions on the UM-CIDI about “seeing” to “talking to” the service providers. Furthermore, focus groups pointed out that speaking of the use of traditional healing “services” made little sense; rather one approaches a healer to ask for help and the healer, “patient,” and his or her family come together to seek resolution. Thus, this construct is labeled “help-seeking.”

Variable construction was completed using SPSS²⁴ and SAS²⁵; all inferential analyses were conducted using Stata’s svy procedures²⁶ with sample and nonresponse weights.²⁷ In the prevalence tables, we present estimates, by tribe, for the total population and also separately for men and women. Given the number of comparisons made, 99% confidence intervals are reported. Multinomial logistic regression, in which the differences between multiple mutually exclusive groups are compared simultaneously,²⁸ was used to examine the relationships of demographic correlates to diagnostic status. To calculate estimates reflecting the differential patterns across sex and tribe groups, a 4-category sex/tribe variable was used in the regression models (3 dummy indicators for Southwest men, Northern Plains women, and Northern Plains men, with Southwest women as the referent group). We report 95% confidence intervals for multinomial logistic regression parameters.

Table 1 presents a demographic description of the AI-SUPERPFP samples. Substantially more women than men were interviewed in the Southwest, likely reflecting differential migration patterns where men were more likely to pursue employment in off-reservation urban areas. Those in the Northern Plains were more likely to live in poverty than were those in the Southwest, although rates in both tribes were high. The Southwest participants were more likely than their Northern Plains counterparts to be married and less likely to be divorced.

Focusing first on tribal differences, as shown in Table 2, the overall lifetime prevalence of any AI-SUPERPFP DSM-IV disorder did not vary significantly by tribe. Similarly, both depressive disorders and anxiety disorders were found at comparable levels in the Southwest and Northern Plains tribes. Substance disorders, especially alcohol dependence, were significantly less prevalent in the Southwest tribe than in the Northern Plains tribes.

Sex differences in the prevalence of PTSD were dramatic; women’s rates were almost twice men’s rates in both tribal groups. The Southwest women showed the most distinctive patterns of disorder. The rates of substance use disorders for Southwest women were about 35% of those for the Southwest and Northern Plains men and 50% of those of Northern Plains women. These differences in prevalence of substance use disorders accounted for the finding that Southwest women were less likely to qualify for any DSM-IV disorder than were men of either tribe.

Alcohol abuse and dependence were the most common substance use problems. The exception to this was among Southwest women, for whom alcohol dependence was relatively rare and not significantly different from rates of drug abuse or dependence.

The 12-month prevalence rates are presented in Table 3. The distribution of 12-month disorders was similar to the distribution of lifetime prevalence (albeit rates were lower, reflecting the truncated time span). Combined across sex, the 12-month prevalence of the AI-SUPERPFP DSM-IV disorders did not differ by tribe, and levels of depressive disorders or anxiety disorders did not differ between the Northern Plains and Southwest tribes. As with lifetime rates, substance disorders were significantly less prevalent in the Southwest tribe than in the Northern Plains tribes.

Examination of prevalence rates by sex shows that substance problems were again the most common disorders for men in both tribes and significantly more prevalent than either anxiety or depressive disorders. For Southwest women, anxiety disorders were most common and were more prevalent than either substance use disorders or depressive disorders. For Northern Plains women, rates of anxiety and substance use disorders were comparable, and both were more prevalent than depressive disorders.

Respondents with depressive and/or anxiety disorders were at increased risk for substance disorders and vice versa (odds ratio, 2.96 [99% confidence interval, 2.32-3.78]). The Southwest men and women with 1 type of disorder were about 3 times as likely to also have a disorder of a second type (odds ratio [99% confidence interval], respectively, 3.00 [1.68-5.37] and 3.51 [2.03-6.07]); Northern Plains men and women with 1 type of disorder were around 5 times as likely to have a second type (odds ratio [99% confidence interval], respectively, 4.70 [2.57-8.60] and 5.11 [3.19-8.17]).

Multinomial logistic methods were used to examine the demographic correlates of DSM-IV disorders in these 2 American Indian populations. The first 3 data columns of Table 4 compare the disorder groups (depressive and/or anxiety disorder(s) only, substance disorder only, and comorbid depressive and/or anxiety and substance disorders) with the group with no disorder. The second set of 3 columns compares the disorder groups with one another.

Controlling for other correlates, Southwest women were at greater risk than any other group for depressive and/or anxiety disorders but at lower risk for substance disorders, either alone or comorbid with depressive and/or anxiety disorders.

Participants aged 15 to 24 years at the time of interview were at greater risk for substance disorders compared with no disorder, depressive and/or anxiety disorders, or comorbid disorders than those older than 44 years.

Those with more than a high school education were at greater risk for depressive and/or anxiety or comorbid disorders compared with no disorder than were those with less formal education.

These demographic correlates were not associated with psychiatric disorder.

Table 5 includes estimates of lifetime help-seeking by type of provider (mental health professional, medical professional, traditional healer, and any help-seeking) for depressive and/or anxiety disorders only, substance disorders only, and for comorbid depressive and/or anxiety and substance disorders. Help-seeking from traditional sources was more common in the Southwest tribe than in the Northern Plains tribes for substance and comorbid disorders. Comparing across disorder types, participants with comorbid disorders sought help more than did those with depressive and/or anxiety or substance disorders alone, although these differences were statistically significant only among Northern Plains men. Most striking is the degree to which help for psychiatric disorders was sought from traditional healers. In fact, Southwest men with substance disorders and Southwest women with depressive and/or anxiety disorders were more likely (P<.05) to consult traditional healers than medical professionals about such problems.

These findings build on work within 2 literatures, that of other American Indian research and that composing the broader context of psychiatric epidemiologic methods. Comprehensive assessments of adults within American Indian communities have been rare. In an early effort, Shore et al⁸ found that 69% of a small tribal village in the Northwest had definite or probable psychiatric impairment (using DSM-I and DSM-II), compared with 57% from the Stirling County study using the same methods.²⁹ The most common disorder was alcoholism. A reinterview 19 years later in the same village using the DSM-III-R version of the Schedule for Affective Disorders and Schizophrenia-Lifetime Version³⁰ found elevated rates of lifetime alcohol abuse or dependence (57%) and lifetime MDE (21%) and somewhat elevated rates of lifetime PTSD (5%) when compared with those of the Epidemiologic Catchment Area Study.⁹ More recently, Robin et al¹⁰ completed a comprehensive assessment of a Southwestern tribal sample of 3 large family pedigrees, using the DSM-III-R version of the Schedule for Affective Disorders and Schizophrenia-Lifetime Version, and reported high rates of lifetime alcohol abuse or dependence (71%), drug abuse or dependence (37%), PTSD (22%), and affective disorders (27%). Kunitz et al⁶ reported high rates of alcohol dependence (70% for men) using a modified version of the DSM-III-R Diagnostic Interview Schedule³¹ with a sample derived from those using IHS within the past 10 years and matched by age, sex, and community to cases receiving treatment for alcohol problems. As with the AI-SUPERPFP, these studies found alcohol disorders to be most common; however, the AI-SUPERPFP rates reported here, especially for alcohol disorders, are substantially lower than those reported in these previous efforts. Methodological differences in case ascertainment and sampling likely account for these discrepancies.

Our findings also need to be considered in the context of other population-based psychiatric epidemiologic studies that have used similar methods. The Australian National Mental Health and Well-Being Survey³² used the World Health Organization’s CIDI 2.0 to derive DSM-IV 12-month prevalence estimates with a population-based sample of Australian residents; their rates of depressive disorders are comparable with those in the AI-SUPERPFP, their rates of PTSD lower than all our American Indian samples, and their rates of alcohol abuse and dependence lower than all but Southwest women. Few US prevalence studies to date have used DSM-IV. The National Comorbidity Survey Replication³³ promises to provide the best contextual data; however, at this point, only the MDE findings from this study are available. The National Comorbidity Survey Replication lifetime MDE rate of 16.2% (range, 15.1%-17.3%) is higher than that reported here for the AI-SUPERPFP, while the 12-month rate of 6.6% (range, 5.7%-7.3%) is not significantly different. The Epidemiologic Catchment Area¹ studies and the original National Comorbidity Survey² used earlier versions of DSM, but comparisons with the combined estimates from those studies³⁴ are consistent with the conclusions drawn from the extant DSM-IV studies—namely, the AI-SUPERPFP samples had elevated rates of alcohol disorders and PTSD. The AI-SUPERPFP MDE rates are either equivalent to or lower than rates in these other populations but certainly not higher, as we had originally anticipated.²²

Indeed, the differential patterns for depression and PTSD in these American Indian populations compared with others are striking, and previous ethnographic studies provide guidance in their interpretation. Anthropologists have documented that “depression” may be conceptualized differently in at least some American Indian groups.^11- 12 Although such findings do not preclude these populations from experiencing DSM-defined depression, American Indians responding to such questions in the CIDI may not endorse the symptoms in the patterns necessary for a diagnosis. Previous studies showing high rates of depressive disorders in American Indian communities have often relied on instruments administered by mental health care professionals, where additional probing and discussion of the meaning of questions are encouraged,^9,35 in contrast to the structured interactions required in lay interviews, where participants are asked to rely on their own understandings of the queries. Added to this, the impact and importance of trauma is clear in American Indian communities, which have been decimated historically by disease and conflict³⁶ and in which accidents and other traumas are everyday occurrences.³⁷ Such experiences may encourage participants to conceptualize or relate their distress explicitly in terms of reactions to trauma rather than depression. Future analyses of these data will investigate more closely the reversal of the depression and PTSD rates evident in our data and will examine the comorbidity between these disorders.

Tribe and sex proved the most important demographic correlates of disorder in the AI-SUPERPFP. The Southwest women, in particular, were at lower risk for substance disorders and at greater risk for depressive and/or anxiety disorders than were others. Previous research by our group has indicated that the Northern Plains tribes consistently appear to be at higher risk for the development of substance use and associated problems compared with the Southwest tribe.^38- 39 Here, focusing on disorders rather than use, we find the same pattern. Age and education also demonstrated relationships with disorder; specifically, younger tribal members were at greater risk for substance problems and those with more formal education were at greater risk for depressive and/or anxiety and comorbid disorders. Unlike other studies, however, demographic correlates such as employment and marital status were generally unrelated to these aggregated disorder groups.

These correlates do show some associations with individual AI-SUPERPFP disorders, however. For instance, Mitchell et al⁴⁰ reported a relationship between marital status and DSM-IV drug abuse or dependence, while Spicer et al⁴¹ reported some associations, varying by tribe, between education, employment, and marital status, with both DSM-III-R lifetime and 12-month alcohol dependence. Even given these findings, however, when compared with similar analyses of the original National Comorbidity Survey,² one is struck by the relative lack of relationship between disorder and poverty, in particular. In poor populations, such demographic characteristics are unlikely to serve as strongly and consistently as antecedents or consequences of disorder as they might in the general population.⁴² The AI-SUPERPFP is not the first to find such lack of associations. For instance, in the Great Smoky Mountain Study, Costello et al⁴³ found poverty to be related to disorder status for white but not American Indian youths. As suggested by others,^{10,36,44- 45} we anticipate that abuse, violence, and trauma will be more critical than poverty in understanding disorder in these communities. Indeed, a recent AI-SUPERPFP publication demonstrates the association between childhood abuse, both physical and sexual, and subsequent AI-SUPERPFP substance disorders.⁴⁶

Substantial comorbidity was apparent in these American Indian samples. Individuals with substance disorders were at significantly increased risk for depressive and/or anxiety disorders. Not surprisingly, individuals with comorbid disorders tended to seek services more often than did those with depressive and/or anxiety or substance disorders alone—and from all types of providers. Such results suggest that substantial help-seeking occurs in these communities in both specialty and primary care sectors as well as from traditional cultural sources. Inclusion of service providers such as mental health technicians and Community Health Representatives—paraprofessionals hired through IHS and tribes to provide many direct services—rendered these questions more inclusive and reflective of the services ecology in American Indian communities than is often typical. Moreover, seeking help from traditional cultural sources is equally common, if not more so, than from specialty or primary care providers, especially in the Southwest tribe.

We did not discuss participants’ assessments of the treatments received or their perceived efficacy and acceptability, although at least some literature suggests that substantial proportions of those in need receive short-term treatment with little positive effect.⁴⁷ Understanding the factors that affect help-seeking and the importance of paraprofessional providers is an important next step in our research agenda.

Recognizing the limitations of a prevalence study such as this is critical. The response rates, while less than optimal, are similar to those of comparable national efforts.^1- 3 Here, we focused on only 2 of the larger tribal populations in the country and, of these, only members living on or near their reservations. Thus, extrapolations to other tribal populations should be made carefully, if at all. In addition, the AI-SUPERPFP sample, mirroring the National Comorbidity Survey, included a limited age range.

The measurement of diagnoses cross-culturally is fraught with potential sources of error. Although we used a standard structured interview protocol, adaptations increased its cultural validity.²⁰ Still, research in other populations suggests that such retrospective self-reports may yield biased estimates of disorder; typically, these are thought to be underestimates.⁴⁸ Further, here the lay interviewers were themselves tribal members. Although instructed not to interview family or others known to them, use of community members may have further depressed the rates. If these biases were marked, they should have been more severe in the smaller Northern Plains communities; the findings of higher prevalence in this tribal group suggest, however, that such potential bias was outweighed by other factors.

In the final analysis, the AI-SUPERPFP samples, while well defined and justified, were limited in cultural representation, age range, and residence. The instrumentation and data collection methods, while comparable with national psychiatric epidemiology studies, were also limited. Thus, the AI-SUPERPFP should be considered a critical but preliminary step to better understanding the prevalence of psychiatric disorders and help-seeking among American Indian populations.

Several clinical implications emerge from these findings. As in studies of non–American Indian populations, many participants did not qualify for any DSM-defined disorder, with fewer than 25% qualifying in the 12-month period. Posttraumatic stress disorder may be especially common among at least some American Indian populations, while MDE may be uncommon or expressed idiomatically in ways not captured by the available epidemiologic tools. Thus, health care professionals working with American Indians should be especially prepared to address PTSD and also should be sensitive to alternate expressions of depression. The dearth of relationships between the classic demographic correlates and disorders warns against assuming that individuals in these communities who are poorer, unemployed, or unmarried are necessarily at heightened risk for psychiatric problems. Furthermore, while substance disorders—and especially alcohol disorders—were common among both tribal groups of men and Northern Plains women, our findings point out that depressive and/or anxiety disorders were also common and that individuals often experienced both. Such findings suggest that successful secondary and tertiary prevention efforts may hinge on screening help-seekers for comorbid disorders.⁴⁸ As tribes and IHS renegotiate service provision under current government policies, opportunities to better coordinate mental health and substance services should be pursued.

Correspondence: Janette Beals, PhD, American Indian and Alaska Native Programs, University of Colorado Health Sciences Center, MS F800, PO Box 6508, Aurora, CO 80045-0508 (jan.beals@uchsc.edu).

Submitted for Publication: December 9, 2003; final revision received October 15, 2004; accepted May 2, 2004.

Additional Authors/the AI-SUPERPFP Team: Cecelia K. Big Crow, Dedra Buchwald, MD, Buck Chambers, Michelle L. Christensen, PhD, Denise A. Dillard, PhD, Karen DuBray, Paula A. Espinoza, PhD, Candace M. Fleming, PhD, Ann Wilson Frederick, Diana Gurley, PhD, Lori L. Jervis, PhD, Shirlene M. Jim, Carol E. Kaufman, PhD, Ellen M. Keane, Suzell A. Klein, Denise Lee, Monica C. McNulty, Denise L. Middlebrook, PhD, Laurie A. Moore, Tilda D. Nez, Ilena M. Norton, MD, Carlette J. Randall, Angela Sam, James H. Shore, Sylvia G. Simpson, MD, and Lorette L. Yazzie.

Funding/Support: The study was supported by grants R01 MH48174 (Dr Manson) and P01 MH42473 (Dr Manson) from the National Institutes of Health (NIH), Bethesda, Md; manuscript preparation was supported by NIH grants R01 DA14817 (Dr Beals), R01 AA13420 (Dr Beals), and R01 AA13800 (Dr Novins).

Acknowledgments: The AI-SUPERPFP would not have been possible without the significant contributions of many people. The following interviewers, computer/data management, and administrative staff supplied energy and enthusiasm for an often difficult job: Anna E. Barón, PhD, Antonita Begay, Amelia T. Begay, Cathy A. E. Bell, Phyllis Brewer, Nelson Chee, Mary Cook, Helen J. Curley, Mary C. Davenport, Rhonda Wiegman Dick, Marvine D. Douville, Pearl Dull Knife, Geneva Emhoolah, Fay Flame, Roslyn Green, Billie K. Greene, Jack Herman, Tamara Holmes, Shelly Hubing, Cameron R. Joe, Louise F. Joe, Cheryl L. Martin, Jeff Miller, Robert H. Moran, Jr, Natalie K. Murphy, Melissa Nixon, Ralph L. Roanhorse, Margo Schwab, PhD, Jennifer Settlemire, Donna M. Shangreaux, Matilda J. Shorty, Selena S. S. Simmons, Wileen Smith, Tina Standing Soldier, Jennifer Truel, Lori Trullinger, Arnold Tsinajinnie, Jennifer M. Warren, Intriga Wounded Head, Theresa (Dawn) Wright, Jenny J. Yazzie, and Sheila A. Young. We would also like to acknowledge the contributions of the Methods Advisory Group: Margarita Alegria, PhD, Evelyn J. Bromet, PhD, Dedra Buchwald, MD, Peter Guarnaccia, PhD, Steven G. Heeringa, PhD, Ronald Kessler, PhD, R. Jay Turner, PhD, and William A. Vega, PhD. Finally, we thank the tribal members who so generously answered all the questions asked of them.