USEFULNESS OF A SURVEY ON UNDERAGE DRINKING IN A RURAL AMERICAN INDIAN COMMUNITY HEALTH CLINIC

Abstract

This study examined the usefulness of a survey on underage drinking in a rural American Indian community health clinic. One hundred ninety-seven youth (90 male, 107 female; age range 8–20 years) were recruited from clinic waiting rooms and through community outreach. The study revealed that the usefulness of the survey was twofold: Survey results could be used by clinic staff to screen for underage drinking and associated problems in youth served by the clinic, and the process of organizing, evaluating, and implementing the survey results accomplished several important goals of community-based participatory research.

INTRODUCTION

In the general U.S. population, underage drinking is associated with significant morbidity and mortality in childhood and adolescence (Hingson, 2009; Hingson, Edwards, Heeren & Rosenbloom, 2009; National Institute on Alcohol Abuse and Alcoholism [NIAAA], 2005). American Indian/ Alaska Native (AI/AN) adolescents are at higher risk than other U.S. ethnic minority adolescents for underage drinking and its associated morbidity and mortality (Bachman et al., 1991; Beals et al., 1997; Beauvais, Jumper-Thurman, & Burnside, 2008; Beauvais, Jumper-Thurman, Helm, Plested, & Burnside, 2004; Blum, Harmon, Harris, Bergeinsen, & Resnick, 1992; NIAAA, 2009; Miller, Beauvais, Burnside, & Jumper-Thurman, 2008; Substance Abuse and Mental Health Services Administration [SAMHSA], 2008; Indian Health Service, 2009; Wallace et al., 2003). Furthermore, adolescents living in rural areas may be at higher risk for underage drinking than suburban or urban youth (SAMHSA, 2004). Thus, AI/AN youth living in rural areas may be at higher risk for underage drinking and its associated problems than other U.S. youth.

In addition, in the general U.S. population, earlier ages of first use of and intoxication with alcohol have been associated with higher rates of problems in adulthood, including alcohol use disorders and alcohol-related risk behaviors (Grant & Dawson, 1997; Grant, Stinson, & Harford, 2001; Hingson, 2009; Hingson & Zha, 2009; Hingson, Heeren, & Winter, 2006a, 2006b; Hingson, Heeren, & Edwards, 2008). One study of an AI community sample from the same population from which the current study participants were drawn (Ehlers, Slutske, Gilder, Lau, & Wilhelmsen, 2006) showed that earlier ages of first alcohol intoxication were associated with higher rates of alcohol use disorders in adulthood. These findings raise the important question of whether earlier ages of first alcohol intoxication are associated with higher rates alcohol-related problems in youth themselves (prior to reaching adulthood), especially in AI and other high-risk youth. If earlier ages of first intoxication are related to higher risk for alcohol-related problems, then delaying age of first intoxication holds promise for decreasing alcohol-related harm in youth and rates of alcohol use disorders in adulthood.

Given these circumstances, finding ways to assess and intervene against underage drinking in rural AI settings is an important goal. Although primary care clinic screening for youth with psychiatric problems has shortcomings as currently practiced (Kelleher & Gardner, 2009; Ozer et al., 2009), such clinics continue to offer an important venue for information collection and screening, particularly in rural and high-risk minority communities where they may be the only source of medical treatment and health information. Thus, screening in the rural primary care setting is important because this setting offers a point of contact with youth where underage drinking and its associated risk behaviors and problems can be identified and appropriate intervention initiated.

Additionally, screening of adolescent health status in a primary care setting has the potential to build the human and infrastructure capacity of the clinic to promote positive health outcomes effectively in the community at large. Because it generates statistics on specific health issues and risk behaviors in the community served by the clinic, screening establishes a quantitative measure of a health problem that can be used by the community and its health care institutions to assess the importance of the problem, disseminate health information by education and outreach, evaluate alternative interventions against the problem, and measure the efficacy of those interventions. In addition, screening undertaken as part of a research process can promote an entity within the clinic dedicated to ongoing research on the health needs of the community and the translation of findings on those needs to interventions.

In this sense, primary care clinic screening of adolescent health issues can serve as a vehicle for implementing many goals of community-based participatory research (CBPR) for health. To paraphrase Israel and colleagues (2003), CBPR is an approach to health research which equitably involves community members, health care institutions, and researchers in a project to understand health problems in the unique context of the community and to use that knowledge to improve the health of community members. CBPR emphasizes the importance of local factors in health problems and in health solutions; facilitates collaborative, equitable partnerships in the research process; promotes co-learning and capacity building among all partners; aims to translate research to health intervention/prevention in the community; builds the capacity of the community to engage in further health care research and translation through a cyclical, iterative process in which the partners pursue new research questions and interventions based on their evolving understanding of the health dynamics in the community; disseminates findings to and solicits feedback from all partners; and often represents a long-term commitment by researchers, health care institutions, and the community (Israel et al., 2003).

We report here the results of a survey on underage drinking and associated problems undertaken at a rural AI community health clinic. As noted above, in a previous study (Ehlers et al., 2006) involving retrospective reporting by adults from the same community, earlier ages of first alcohol intoxication were found to be associated with higher rates of alcohol use disorders in adulthood. When feedback from this study was shared with the community, the board of the community health clinic requested assessment of current underage drinking and proposal of a strategy to reduce underage drinking in the reservation population served by the clinic. This paper reports the results of the first phase of the project: assessment of underage drinking in youth served by the clinic.

The goals of this study were to determine whether a survey on underage drinking and associated problems in a rural AI community health clinic is feasible; can yield useful information for further individual and family intervention; can inform further research on individual, family, and community-wide intervention and prevention efforts; and can effectively accomplish many of the goals of CBPR.

METHOD

This study of the usefulness of a survey for underage drinking was conducted as part of a larger project to build the capacity of an AI community health center to assess and implement a program to reduce underage drinking on contiguous Southern California tribal reservations. The capacity-building project is an ongoing collaboration among the AI community health center, The Scripps Research Institute (TSRI), and the Pacific Institute for Research and Evaluation (PIRE). The Institutional Review Board (IRB) representing the nine Southern California tribes included in the study requires that the research team not reveal the name of the health clinic or the names or exact locations of the reservations. For the purposes of this report, the clinic will be called the Southern California Tribal Health Center (SCTHC). The funding for this study came from NIAAA as part of a grant to assess and reduce underage drinking in rural communities. The protocol for the study was approved by the IRB of PIRE, TSRI, and the IRB representing the nine tribes served by the clinic.

Measure

The research team developed a survey to assess underage drinking in the population of youth served by the clinic, as a preliminary step toward the design and implementation of culturally appropriate individual intervention and community-wide prevention research to study the efficacy of measures to reduce underage drinking. The survey was constructed as a collective effort by scientific advisory personnel at NIAAA and their Rural Underage Drinking Initiative grantees, including the University of Pittsburgh, the SCTHC, TSRI, and PIRE. The Principal Investigator (Internal Medicine, Pediatrics) AI physician. To of the project at the SCTHC is a board certified enhance cultural appropriateness of the survey, the research team solicited input on the design and implementation from a local advisory board made up of reservation elders, community volunteers, clinic administrators, and tribal law enforcement personnel. Several young adult AI research assistants at the SCTHC, including the initial study coordinator (JAL), started working on the project at the beginning of survey construction. They provided valuable input into crafting questions concerning substance use and alcohol problems, so that the items were culturally appropriate and tailored to known specifics about youth drinking on the reservations. The survey included demographics (10 questions); ages of first alcohol drink, first alcohol intoxication, and first drinking ≥ 5 drinks (for boys) or ≥ 4 drinks (for girls) in a 2-hour period (3 questions); past 30-day, past year, and lifetime drinking quantities and frequencies (7 questions); other substance use (8 questions); psychiatric symptoms (4 questions); past year alcohol-related problems and risk behaviors (13 questions); and alcohol availability (13 questions). The full survey is included as Appendix A to this report. All research team members completed the National Institutes of Health human subjects protection training program and a course in research and survey administration organized by TSRI and PIRE study personnel.

Participants

All participants were registered patients at the SCTHC. The SCTHC serves the AI/AN population of San Diego County, which is estimated to be 24,437 individuals (SCTHC, personal communication, January 24, 2013). In 2011, the SCTHC received 4325 unduplicated visits, of which 1301 were from individuals in the age range of 5–19 years. Thus, 30% of unduplicated SCTHC visits were in the approximate age range of the participants in this study (SCTHC, personal communication, January 24, 2013). Participants were recruited from clinic waiting rooms, at community-wide health fairs, at powwows, and at after-school programs. Children ages 8–12 years verbally assented with written parental consent. Adolescents ages 13–17 years gave written assent with written parental consent. Adults ages 18–20 years gave written informed consent. Prior to providing informed assent and consent, youth and their parents were informed of the nature of the study and were given contact information for the study coordinator in case they had any questions later. Prospective participants were told that, after completion of the survey, they would be given a $15 gift card to Walmart, Target, or Barnes & Noble. Prospective participants were also told that, if at any point they felt uncomfortable answering any question, they could skip that question, and that they could also discontinue their participation in the survey entirely. In either case, the participant would still receive the gift card. The survey was administered by an AI/AN young adult research team member at the clinic in a private room, or at an outside venue using a folding table and chairs in a private area. All survey data were recorded anonymously and subsequently were managed using a coded number unique to each individual. To further safeguard confidentiality, no list linking the participant’s name or other unique information with his/her coded number was kept.

Analyses

This report has two aims. The first is to report quantitative information on frequencies of selected alcohol and substance use problems, as well as logistic regression analysis results regarding the association of earlier age of first intoxication with alcohol and substance use problems. The second aim is to report qualitative information on the usefulness of the survey as a vehicle for promoting CBPR, including building the capacity of the SCTHC to intervene to reduce underage drinking in the population that it serves.

To accomplish the first aim, two sets of analyses were undertaken. The first set of analyses determined demographics, age of onset, and frequencies of each use and problem variable. Because there are often significant differences in alcohol-related measures between males and females, frequencies of use and problem variables are reported separately for boys and girls. To determine if there were significant gender differences in use or problem variables, frequencies of dichotomous variables in boys and girls were compared using Fisher exact tests. Continuous variables were compared using the Mann-Whitney test. In these analyses, alpha was set at 0.05, and p values < 0.05 were considered significant. Gender comparisons for dichotomous demographic, alcohol use, and alcohol problem variables where the number of responses in any cell was < 5 are not reported in the tables.

Because alcohol use and problems often increase with age during childhood and adolescence, and given that age of initiation is confounded with current age, age at interview was included as a covariate when assessing the relationship of age of first alcohol problems. Therefore, the second set of analyses used multivariate logistic regression to assess the association of each use and problem variable as the outcome variable with the independent variables age and age of first alcohol intoxication, using the Backward Stepwise approach of Wald (SPSS, Scientific Software International). Independent variables with the highest p value (> 0.10) were removed in each step. Odds ratios and their p values were calculated for each variable in each final logistic model. In preliminary analysis of the entire sample, several outcome variables showed significant gender interaction. Therefore, the entire sample was divided on the basis of gender, and these subsamples were analyzed separately. In these analyses, alpha was set at 0.05, and p values < 0.05 were considered significant.

The second aim of this study is to report qualitative information on the usefulness of the survey as a vehicle for promoting CBPR and for building the capacity of the SCTHC to intervene to reduce underage drinking in the population it serves. We report on several dimensions of the process of survey design, implementation, and data analysis that illustrate ways in which the survey facilitated clinic and community participation, partnership between outside (PIRE, TSRI) and community (SCTHC) personnel, training and mentoring of young adult AI research personnel, dissemination of research findings, and use of research to build capacity of the clinic to address a pressing community health need.

RESULTS

Of the 197 youth who participated in the study, 90 were male (age range 8–20 years, median age 13.5 years) and 107 were female (age range 8–20 years, median age 15.0 years). Recruitment of sample youth from each venue was as follows: 65% percent were recruited from the clinic itself (via fliers, clinic staff, and direct solicitation in waiting rooms), 19% from after-school programs, 14% at community health fairs, and 2% at powwows. Recruitment at the last three venues sometimes coincided with presentations on underage drinking given by the research assistants as part of a “giving back” program included in the survey project. Overall, there were 12 recruitment refusals (6%): 7 at the clinic, 2 from community health fairs, and 3 at powwows. In the sample, 98% of boys and 97% of girls self-identified as AI/AN. Ninety-eight percent of boys and 99% of girls reported speaking English in their homes. Boys were not significantly different from girls in any demographic variable. These results are displayed in Table 1. Not every survey question was answered by each participant, leading to some instances of missing data. In this report, we report positive responses for each variable and the percentage of the total responses (positive and negative) to the survey question for that variable.

Table 1.

Demographic and Alcohol Age of Onset Variables Comparing Boys and Girls (N = 197)

Variable	Boys			Girls			Mann-Whitney	p Value
	na	Range, Median (Years)	Mean Rank	n	Range, Median (Years)	Mean Rank
Age	90	8–20, 13.5	91.7	107	8–20, 15.0	105.1	4162.00	0.10
Age first drink	33	8–16, 13.0	44.8	58	8–18, 14.0	46.7	916.00	0.73
Age first intoxication	24	12–17, 14.5	32.6	42	8–18, 14.0	34.0	483.00	0.78
Age first drinking ≥5 (male)/≥4 (female) drinks in a 2-hour period	23	12–17, 15.0	32.8	43	8–18, 14.0	33.9	478.00	0.82

	nb	%	n	%	Fisher p Value Exact Test
Currently in school	77	86	90	84	1.00

^an indicates the number of boys or girls responding to the question assessing each continuous variable. For example, 33 boys and 58 girls reported their age of first whole drink (vs. never having had a whole drink).

^bn indicates the number of boys or girls responding positively to the question asking whether they were currently in school

Frequencies of selected substance use, mood, and alcohol-related problem variables and their comparisons in boys and girls can be seen in Tables 2 and 3. Thirty-seven percent of boys and 54% of girls had drunk one or more standard drinks in their lifetimes. Twenty-seven percent of boys and 39% of girls reported having been intoxicated with alcohol one or more times in their lifetimes. Of boys who had drunk alcohol in their lifetimes, 31% had drunk in the past month and 17% had drunk at least weekly in the past year. Of boys who had drunk in the past month, 50% reported drinking ≥ 5 drinks on occasions when they did drink, and 38% reported drinking ≥ 5 drinks at least once a week in the past month. Of girls who had drunk in their lifetimes, 47% had drunk in the past month and 23% had drunk at least weekly in the past year. Of girls who had drunk in the past month, 47% reported drinking ≥ 5 drinks on occasions when they did drink, and 16% reported drinking ≥ 4 drinks at least once a week in the past month. In the entire sample, 13% of boys and 28% of girls reported feeling so sad or hopeless for at least 2 weeks in the past year that they gave up some usual activities (“impairment”). In the entire sample, 9% of boys and 18% of girls had seriously considered suicide in the past year. Frequencies of substance use variables and serious consideration of suicide were not significantly different in boys as compared to girls. However, depressive episode with impairment was significantly more common in girls as compared to boys.

Table 2.

Selected Substance Use and Mood Variables Comparing Boys and Girls

Variable	Condition (Boys; Girls)a	Boys n (%)	Girls n (%)	Fisher Exact Test
Drank in past month	Ever drank (35; 62)	11 (31%)	29 (47%)	0.20
Drank ≥ 5 drinks when did drink in past month	Drank in past month (12; 30)b	6 (50%)	14 (47%)	1.00
Drank ≥ 5 (male)/ ≥ 4 (female) drinks in a 2-hour period at least once/week in past month	Drank in past month (13; 31)b	5 (38%)	5 (16%)	0.13
Drank at least weekly in past year	Ever drank (35; 62)	6 (17%)	14 (23%)	0.61
Typically drank ≥ 5 drinks when drank in past year	Drank in past year (25; 47)	13 (52%)	22 (47%)	1.00
Use cannabis ≥ 40 times in lifetime	Ever used cannabis (30; 43)	11 (37%)	19 (44%)	0.63
Used methamphetamine in past year	Entire sample (90; 107)	5 (6%)	6 (6%)	1.00
Used OTC meds to get high in past year	Entire sample (90; 107)	36 (4%)	54 (50%)	0.15
Used prescription pills to get high in past year	Entire sample (90; 107)	32 (36%)	48 (45%)	0.19
Sad or hopeless with impariment for ≥ 2 weeks in past year	Entire sample (90; 107)	12 (13%)	30 (28%)	0.01
Seriously considered suicide in past year	Entire sample (90; 107)	8 (9%)	19 (18%)	0.10

^a(Boys; Girls) indicates the number of boys or girls responding positively to each condition. For example, in the first row, 35 boys and 62 girls responded positively to the question asking whether they had ever had one or more full drinks of alcohol. Then, 11 of the 35 boys (31%) and 29 of the 62 girls (47%) indicated they had drunk in the past month.

^bNumbers of those who drank in past month are discrepant because different numbers of boys and girls answered each question.

Table 3.

Selected Alcohol-related Problem Variables Comparing Boys (n = 90) and Girls (n = 107)

Variable	Boys n (%)	Girls n (%)	Fisher Exact Test
Drove a vehicle within 1 hour after drinking	6 (7%)	9 (8%)	0.79
Rode with driver who drove vehicle within 1 hour after drinking	27 (30%)	50 (47%)	0.02
Drunk while in school	12 (13%)	22 (21%)	0.19
Alcohol blackout	8 (9%)	24 (22%)	0.01
Missed school or class because of drinking	11 (12%)	27 (25%)	0.03
Passed out (lost consciousness) while drinking	11 (12%)	26 (24%)	0.04

When frequencies of alcohol-related problems in the past year were examined in the entire sample, 7% of boys and 8% of girls had driven a motor vehicle within one hour of drinking. Thirty percent of boys and 47% of girls had ridden in a motor vehicle with a driver who had been drinking within one hour of driving. Thirteen percent of boys and 21% of girls reported having been drunk in school, 9% of boys and 22% of girls reported having experienced an alcohol blackout, and 2% of boys and 8% of girls reported having been seriously injured due to drinking. Riding in a vehicle whose driver had been drinking within the hour (p = 0.02) and experiencing an alcohol blackout (p = 0.01) were significantly more common in girls as compared to boys.

When logistic regression models were used to examine for significant age of first intoxication taking into account age as covariate, earlier age of first intoxication (OR, p value) was associated with several substance use and alcohol-related problems in boys and girls (Table 4). In boys, earlier age of first intoxication was associated with drinking in the past month (1.99, 0.046) and experiencing an alcohol blackout in the past year (2.15, 0.04). In girls, earlier age of first intoxication was associated with smoking ≥ 1/2 pack of cigarettes per day in the past month (1.60, 0.04), using cannabis ≥s 40 times in one’s lifetime (1.70, 0.02), using methamphetamine in the past year (1.97, 0.02), being drunk in school in the past year (1.89, 0.01), and suffering a serious injury due to drinking in the past year (1.60, 0.03).

Table 4.

Final Logistic Regression Models for Factors Significantly Associated with Selected Substance Use and Alcohol-related Problem Variables for Boys (n = 90) and Girls (n = 107)^a

Outcome Variable	Independent Variable	Boys			Girls
		Odds Ratio	95% CI	p Value	Odds Ratio	95% CI	p Value
Drank in past month	Age first intoxication	1.99	1.01–3.91	0.046
	Current age
Smoked ≥ half pack cigarettes/day in past month	Age first intoxication
	Current age				1.60	1.01–2.54	0.04
Used cannabis ≥10 times in past month	Age first intoxication
	Current age				1.74	1.03–2.94	0.04
Use cannabis ≥40 times in lifetime	Age first intoxication				1.70	1.09–2.66	0.02
	Current age				1.70	1.06–2.72	0.03
Used methamphetamine in past year	Age first intoxication				1.97	1.14–3.42	0.02
	Current age				2.26	1.05–4.85	0.04
Rode with driver who drove vehicle within 1 hour after drinking in past year	Age first intoxication
	Current age				1.65	1.00–2.69	0.049
Been drunk in school in past year	Age first intoxication				1.89	1.18–3.05	0.01
	Current age
Blackout while drinking in past year	Age first intoxication	2.15	1.04–4.42	0.04
	Current age
Worried he/she drank too much/often in past year	Age first intoxication				1.60	1.07–2.40	0.02
	Current age
Serious injury or injuries due to drinking in past year	Age first intoxication				1.60	1.04–2.46	0.03
	Current age

¹Multivariate logistic regression was used to assess the association of each substance use and alcohol- related problem variable (as the outcome variable) with independent variables entered into the initial model in each regression: age of first alcohol intoxication and current age. Final logistic models were determined using the Backward Stepwise Approach of Wald. Odds ratios and p values are given for all independent variables that appeared in each final model. A blank cell indicates that the independent variable did not appear or was not significant in the final model.

From a qualitative standpoint, the successful implementation of this survey demonstrated several important findings from the survey project. First, a detailed and informative survey of underage drinking, including quantitative information on age of onset of alcohol use variables, substance use, psychiatric problems, and alcohol-related problems, is feasible and acceptable to youth enrolled at a rural AI community health clinic.

Second, young adult AI research assistants are able to implement such a survey effectively. Examples of the research assistants’ effective implementation of the survey include: (1) giving advice on culturally appropriate terminology and wording for survey questions; (2) generating a picture of the most common locally consumed alcoholic beverages as an aid in estimating standard drinks with the participant during survey administration; (3) developing and implementing culturally sensitive strategies for recruitment, both from within the clinic and from the larger community; (4) purchasing and operating an optical scanner capable of generating scannable survey forms, which were then scanned to transfer data to an Excel file for statistical analysis; (5) attending four training sessions with a statistician at TSRI to learn principles of statistical analysis using SPSS; (6) giving presentations on underage drinking to adult community members and clinicians at the SCTHC, and to middle and high school students in after-school programs; (7) creating and implementing a game-oriented education program on underage drinking for elementary school students; (8) reporting back to the SCTHC board on the implementation and results of the survey; and (9) giving presentations at four national meetings on underage drinking and AI/AN health.

Third, such a survey can yield valuable information for screening and treatment interventions in the clinic. These data can also inform further research on individual and family intervention, as well as community-wide prevention research studies. The information in the survey allows for improved screening for underage drinking and its associated problems at the SCTHC, by providing a few questions that are easy to administer yet informative of risk. Information gained from these questions can be used to advise youth to obtain further evaluation for substance use and emotional problems. Using information obtained in the survey, the SCTHC, with TSRI and PIRE, also submitted a successful application for ongoing funding to study specific individual and family intervention and community-wide prevention strategies in the population served by the SCTHC. In this case, a continuing research program, which studied the efficacy of (1) individual and family Motivational Interviewing to reduce underage drinking and (2) a community-wide reward and reminder program to reduce teenage drinking and driving, was based directly on the findings and on the research capacity that had been built during the survey research effort.

Fourth, the project demonstrates that a survey research project is an effective vehicle for accomplishing many of the qualitative goals of CBPR (see Israel et al., 2003) beyond the quantitative information on underage drinking and associated problems obtained in the survey itself. Examples of the qualitative accomplishments of the survey include:

1. Building on the strengths of the SCTHC (in this case, the administrative, clinical, and educational expertise of the SCTHC, including its student program, whose participants made an invaluable contribution to the project);

2. Building on the strengths in the community (in this case, the advice and support of an informed local advisory board, as well as a generation of elders and parents—some of whom were themselves in recovery—who were committed to building programs to reduce underage drinking on the reservations served by the SCTHC);

3. Facilitating a collaborative research partnership between the SCTHC and outside research entities (in this case, TSRI and PIRE);

4. Promoting co-learning by both the SCTHC staff and the outside investigators (in this case, research assistants at the SCTHC learned research techniques and investigators from TSRI and PIRE learned culturally appropriate assessment of underage drinking and associated problems);

5. Integrating and balancing research and clinically actionable information (in this case, by providing information on the scope of the clinical problem, and screening questions that could easily be used for rapid assessment of underage drinking in all departments of the SCTHC and could serve as a basis for referral for more extensive evaluation and treatment, as appropriate);

6. Achieving relevance to the local community. In this case, elders and parents in the community were well aware of the association of underage drinking with serious, potentially lethal problems, including motor vehicle crashes. From a research standpoint, in this community earlier ages been found to be associated with high rates of alcohol use disorders of first in adulthood (Ehlers et al., 2006), but little was known about their association with problems in children and teens themselves. The survey was able to confirm an association between underage drinking and serious alcohol-related problems in youth and to provide new quantitative information on the frequencies of alcohol use, other substance use, mood problems, and alcohol-related risk behaviors as well as their association with earlier ages of intoxication;

7. Promoting a cyclical, iterative process which uses research to build capacity, determine relevant clinical information, and propose culturally appropriate and effective interventions on the basis of that information, leading to further research on the efficacy of those interventions, beginning the cycle anew. In this case, the survey project provided the data and helped build the capacity follow-on efforts to research the effectiveness of individual and family intervention and community-wide prevention strategies to reduce underage drinking in youth served by the SCTHC;

8. Disseminating critical health-related information to the community. In this case, the survey project confirmed existing information and added new information on the clinical significance of underage drinking, and provided a mechanism to disseminate that information to youth and parents in the reservation communities served by the SCTHC, using presentations and the clinic newsletter; and

9. Establishing a long-term commitment to research in the service of health promotion in the participating communities. In this case, under the aegis of funding from the National Institute on Minority Health and Health Disparities and NIAAA, an 8-year research program on the efficacy of clinical and community-wide interventions to reduce underage drinking, with its attendant promotion of a young generation of AI researchers at the SCTHC, is currently underway.

DISCUSSION

This report demonstrates that a brief survey on youth substance use, mood problems, and alcohol-related problems can be administered successfully in a rural AI community health clinic, and can be useful in both quantitative and qualitative terms as a vehicle for improving health in the community served by the clinic.

From a quantitative standpoint, survey findings can inform clinic personnel and community leaders on the specific kinds and extent of health problems experienced by youth. Survey findings can be used for rapid, targeted screening for those problems in behavioral health, medical, and dental departments; provide for referral to appropriate clinic services; and serve as a basis for clinic outreach, education, and prevention programs in the community. For example, the survey reported here found that five (6%) boys and five (5%) girls reported bingeing (defined here as ≥ 5 drinks per drinking day for boys and ≥ 4 per drinking day for girls) at least weekly in the past month. Weekly bingeing in adults has been associated with higher risk of alcohol use disorder symptoms in both those with predominantly European ancestry (Saha, Stinson, & Grant, 2007) and an AI sample (Gilder et al., 2011). In the 1994, 1995, and 1998 National Household Survey on Drug Abuse surveys, teenagers who drank ≥ 5 drinks per day on one or more days in the past month were more likely than those who did not to have used other illicit substances, driven under the influence of alcohol or drugs, experienced school problems, engaged in fights, and had legal problems (Greenblatt, 2000). Thus, a single question about weekly bingeing identifies a high-risk subgroup in the clinic population that can be quickly referred for intervention and treatment. Similarly, use or heavy use of other substances, depressive episode or suicidal ideation, and driving a motor vehicle or riding with a driver under the influence of alcohol are all serious, potentially lethal—but easily screened-for—conditions.

A second important quantitative finding in the survey reported here is that, compared to boys, girls are as much or more at risk, depending on the variable, for alcohol and drug use and for alcohol-related problems. For example, 47% of girls as compared to 30% of boys (p = 0.02) reported riding in a motor vehicle in the past year with a driver who had been drinking within the hour. These data suggest that a determined community campaign to educate youth on ways to avoid riding with a driver who has been drinking might be useful in reducing harm in youth. This example again illustrates how a community-based survey can lead to important information on risk the clinic serves, which can easily be used in brief screening behaviors specific in any department.

A third example of important quantitative information from the survey reported here is shown in Table 4, which portrays those substance use and alcohol-related problems that were significantly associated (OR, p value) with earlier age of first intoxication and/or with age in boys and girls separately. Other substance use, mood problems, and alcohol-related problems assessed in the survey were not found to be significantly related to either age of first intoxication or age in either boys or girls. As can be seen in Table 4, earlier age of first intoxication was associated with several substance use and past year alcohol problem variables, including experiencing a blackout while drinking (2.15, 0.04) in boys and being drunk in school (1.89, 0.01) and serious injury due to drinking (1.60, 0.03) in girls. Taken together with data on median age of first intoxication (Table 1), these findings suggest that concerted efforts to reduce and delay underage drinking might have an important impact on reducing harm from alcohol-related problems in clinic youth. Also of interest is the regression model of riding in a vehicle driven by a driver who had been drinking within the hour. This behavior is significantly associated with age in girls, but not boys, and not with earlier age of first intoxication in boys or girls. This finding indicates that as girls move through their teen and young adult years, they increasingly risk riding with drivers who have been drinking, perhaps because they are riding in vehicles driven by older boyfriends. If this hypothesis could be confirmed, targeting this behavior in teen and young adult girls might be particularly useful in reducing harm.

From a qualitative standpoint, this research project demonstrates that a survey is an effective vehicle for accomplishing many of the qualitative goals of CBPR (see Israel et al., 2003) beyond the quantitative information obtained in the survey itself. In this sense, the process of designing, implementing, analyzing data from, and translating findings of the survey, rather than the survey data itself, accomplishes the goals of CBPR. As noted in the Results, these accomplishments include building on the strengths of the existing SCTHC infrastructure and the reservation communities, facilitating collaborative partnerships between and mutual learning by AI and outside research personnel that benefited both groups, undertaking research in the service of actionable about clinical problems important to the community, and promoting ongoing research by the clinic on health care problems in the community. The process of undertaking the survey was perhaps most important in fostering young adult AI researcher training and promoting an ethos of ongoing health research to enhance health care by AI researchers.

The results of this study should be viewed in the light of several limitations. The survey was not administered to a random sample, but to a sample of convenience, which may not be representative of the youth population served by the SCTHC, other AI youth populations served by rural clinics, or rural AI youth in general. The survey uses cross-sectional and retrospective data, which may be subject to recall bias. A prospective study that assessed randomly selected youth participants at several time points as they moved from childhood to young adulthood would address these concerns. Our logistic models did not test several covariates that might have been significantly associated with the alcohol and other substance use, depression and suicidal ideation, and alcohol-related problem variables if they had been included in the models. Examples include earlier age of first use of substances other than alcohol, anxiety, history of personal trauma, family disruption, and socioeconomic status. Another limitation is that of multiple comparisons. In total, 28 logistic regression analyses were undertaken for both boys and girls. None of the significant findings in these analyses would survive Bonferroni correction. A larger sample size might have demonstrated significant differences in alcohol, substance use, mood, and alcohol-related problems or significant associations between age of first alcohol intoxication and those problems, which were not found in the current study. The major weakness of this manuscript is the very small numbers of participants for some items assessed in the survey. The small numbers render assessments of their association with gender, age, and earlier age of first intoxication unreliable from a scientific standpoint. A larger prospective study would remedy this limitation and provide a more firm foundation for community action.

Despite these limitations, we believe the findings of this study represent an important contribution in identifying the quantitative and qualitative ways a brief, easily administered survey of youth served by the SCTHC can be used to identify important substance use, mood, and alcohol-related problems in these youth, and to provide a population-specific and culturally appropriate platform for community-wide education and prevention efforts to reduce the morbidity and mortality associated with underage drinking.

Appendix A. Survey on Underage Drinking and Associated Problems