JOURNAL OF ADOLESCENT HEALTH 19X$13:582-588
An Evaluation of a School-Based AIDS/HIV Education Program for High School Students CAROLYN SEYMORE ASHWORTH, M.D., CHERYL NEWMAN,
M.D.,
ROBERT H. DuRANT, Ph.D.,
AND GREG GAILLARD, A,&
TIN+effect of a l-hr school-based AIDS/HIV education program on the knowledge and attitudes of high school students was evaluated with a modified version of the Centers for Disease Control Health Risk Survey. One urban and one suburban school each were randomly assigned to an educational inte~ention b = 5351 or a control group (?I = 659). All students received a posttest 2 weeks after the intervention. Knowledge was based on responses to 12 true-false questions (pretest OL= .76, posttest ix = 0.811. Principal components analysis was used to develop three attitude scales and risk-taking behavior was assessed by self-report. Data were analyzed with KNskall-~allis analysis of variance (ANOVA) and multivarlate ANOVA. The groups did not differ in knowledge level at pretest. At posttest the education group had significantly (p s 0.0061 higher knowledge even after controlling for the effects of previous AIDS education fp s 0.0191, gender fp I 0.0071, and Hispanic ~nic~~ (p 5 0.0481.After the education program, students were less worried about exposure to the AIDS virus, but were more worried fp I 0.0481 about AIDS acquisition during their adult life. Although single school-based AIDS/HIV education programs may increase knowledge, more extensive education may be neededto change the behavior and attitudes of older high school students.
From the ~rtn~~lf ofPedjafri~ K.S.E.), A??~b~~iaforyCnrc Cmfer, University ofAl&mm at ~jr~rfillg}rat~l, Bimiughartr, AMMFII~Z nrtd the Lkpartrnentof Pediatrics,Sectbt ofGtweral Pediatrics and Adolescent Medicine(R.H.D., G.G.), and t/w Department of lnferrml Medicine, Section ofZufectious Diseases W-J.), Medical College of Georgia,
Augusta,Georgia. Addressco~~~oF!de~?ce to: Robert H. D;i~Fft,
Ph.D., CE 122, Muff ~~le~of Georgia, A~gusf~*GA 30912. fltfortpiwfs a~iiabte. 1 Mauu~~~f accepted March 2 7, 1992. 582 1054-139x/92/$5.00
KEY
WORDS:
AIDS/I-NV education School intervention Students’ AIDS knowledge
Addescents are at higher risk of exposure to the acquired immunodeficiency syndrome (AIDS) virus than is suggested by the relatively low prevalence rate in this age group. Almost one-quarter of the cases now reported have occurred in young adults ages 20-29 years. because the time between infection and the appearance of clinical symptoms is approximately 4-7 years, it is assumed that many of these individuals were exposed to the virus during their adolescence (1,2). Ex~~men~tion with behaviors that might expose youth to the AIDS virus frequently occurs as adolescents attempt to establish their own identity (3,4). Furthermore, adolescents’ limited ability to understand and anticipate the long-term consequences of such behaviors increases their risk for human immunode~~ency virus (HIV) infection (5). Prevention in this age group has been approached through educational programs about behaviors that increase adolescents’ exposure to the virus (6). Accurate information about the severity of the disease and specific methods of protection have influenced the risk-taking behavior of homosexual and bisexual males in several San Francisco studies (5,7,8). Early descriptive studies of adolescents’ knowledge about AIDS found that they had many misconceptions about HIV transmission and behavioral risks factors (9-11). More recent data suggest that adolescents remain misinformed about AIDS/HIV in ways which
Publishedby ElsevierScience
D
Society for Adolescent Medicine, 1992 Publishing Co., Inc., 655 Avenue of the Americas, New Yolk, NY 10010
November 1992
SCHOOL-BASED AIDS/HIV EDUCATION PROGRAM
may increase their exposure to HIV (12,13). Schoolbased general sex education programs have modest effects on risk-taking behaviors reported after extensive skills-based interventions (14,15). Several intervention programs designed to educate adolescents specifically about AIDS/HIV infection have had a positive effect on knowledge, but did not seem to affect attitudes or behavioral intent (16-19). Other short-term interventions have documented modest changes in both HIV knowledge and attitudes (2023). The discrepancy between knowledge gain and attitude change about AIDS/HIV infection may be due to methodologic problems with the program evaluation strategies. Interventional studies often included a wide range of ages and infrequently included control groups for comparison (9-11, 16,17,19,23,24). Because development of the cognitive ability to process information and associate it with personal risk increases with age, the inclusion of younger adolescents may dilute the efficacy of the interventions in altering risk-taking behaviors. Determining the true effect of a specific intervention is also difficult without an adequate control group because exposure to external factors and other extraneous variables often affects the outcome measured (25). The purpose of this study was to determine the level of knowledge about AIDS and HIV infection among high school students in Augusta, Georgia, and to evaluate the effect of the current AIDS education program on knowledge about AIDS prevention and transmission, attitudes concerning HIV infection, and perceived risk of exposure to and acquiring AIDS.
random numbers table, one suburban and one urban school were each randomly assigned to either an AIDS/HIV education intervention group or to a control group receiving no intervention. The students in each urban and suburban school, respectively, had similar socioeconomic and ethnic profiles. Approximately 10 weeks later the two high schools in the intervention group received a standardized educational program about AIDS and HIV infection. The 1-hr education program was developed and taught by Red Cross certified educators. The program consisted of an age-appropriate video [A Letter to Brian (26) or Don’t Forget Sherry (27)] about the transmission and prevention of AIDS and was followed by a discussion session led by two trained AIDS educators. All content areas and components measured on the questionnaires were addressed during the intervention. Two weeks after completion of the intervention, a posttest questionnaire was administered. At the pretest, a list of students in each class completing the questionnaire was made and kept by the investigators in a separate file from
Methods Evaluation Design The state of Georgia mandates AIDS education in grades K-12. The timing, content, and format of the program are left to the individual school system. To comply with state law, the Richmond County Board of Education enlisted the help of the Augusta Chapter of the American Red Cross and the Medical College of Georgia to provide and evaluate the educational programs. Eleventh and 12th grade students (n = 1194) attending two urban and two suburban public high schools were pretested with an anonymous, standardized questionnaire in February 1990. Using a
583
the anonymous questionnaires. Using this list, only students completing the initial pretest questionnaire completed the posttest questionnaire. Because the
questionnaire was anonymous, pre- and p&test questionnaires of the students were not paired. Questionnaire A modified version of the Centers for Disease Control (CDC) Health Risk Survey was used to evaluate the program (12). The CDC questionnaire was designed for epidemiologic surveillance of adolescents’ knowledge, attitudes, and behaviors and has not been evaluated for test-retest reliability; however, several scales from the survey have been found to have acceptable internal reliabilities (28). Certain questions were omitted because of the school board’s concern that the more explicit questions regarding personal sexual activity and condom use would be construed as promoting sexual intercourse. Additional questions examining beliefs and attitudes about AIDS were included in the questionnaire (29). These questions evaluated the perceived risk of acquiring AIDS as well as the level of worry about the possibility of infection. Except for the knowledge questions, all other questions had either Likert-type or ranked-order response formats. The questionnaire was pretested on adolescents of similar age and race who were seen at the medical university’s Adolescent Clinic. This study was
584
ASHWORTHET AL.
approved by both the County Board of Education and the university’s Human Assurance Committee. AIDS/HIV knowledge level was based on the students’ reT>onses to 12 questions with a yes-nounsure response format. The internal reliability of the knowledge scale was assessed with Cronbach’s o (30). Reliability for the pretest and posttest scales was 0.76 and 0.81, respectively. These are acceptable levels of internal reliability for a scale limited to 12 dichotomous items and are slightly higher than reliability reported by DiClemente for a 16item knowledge scale from the CDC survey instrument (28). Principal components analysis, with varimax rotation, using a parallel analysis technique to determine the number of components to retain, was used to identify specific scales from nine attitude questions (31). The magnitude of the eigenvalue used to determine the number of components to retain was based on a comparison with eigenvalues from 15 randomly generated data sets of the same sample sizes and number of scale items as the original data set. This approach identifies groups of questions that measure similar dimensions and, when combined together, produce scales with high internal consistency. Three distinct components were identified: (1) tolerance toward AIDS/HIV infected students (two items), (2) self-efficacy concerning AIDS/ HIV knowledge, testing, and prevention (three items), and (3) peer and family communication about AIDS/HIV (two items). These same three components were found on the posttest data. The means, standard deviations, and the total possible range of the scales developed by principal components analysis, as well as the knowledge, behavior, and other attitude scales are recorded in Table 1.
JOURNAL OF ADOLESCENT HEALTH Vol. 13, No. 7
Table 1. Pretest Attitudes and Knowledge of AIDS/HIV of Hieh School Students Range of Scale (Low to High)
x
SD
3.67
1.38
l-5
1.57
1.04
1-5
3.55
1.24
l-5
2.15
1.46
l-5
1.69
0.83
1-3
1.97
0.79
l-3
1.37
0.67
l-3
1.62
0.73
1-4
0.08 9.01
0.88 2.30
o-4 o-12
6.21
1.49
4-8
IO.68
1.56
6-12
6.40
I.62
4-8
Statistical Analysis
Would you feel uncomfortable asking a sexual partner if they were at risk for AIDS/HIV infection? If you got AIDS/HIV infection would you worry about dying? Should AIDS/HIV infected patients be isolated from the rest of society? Have you changed your behavior to avoid getting AIDS/HIV infection? How worried are you that you have been exposed to AIDS/HIV infection? How worried are you that you might get AIDS/HIV infection some time in your life? What do you think the chance is that you have the AIDS!HIV infection virus? What do you think the chance is of you getting AIDS/HIV infection in your adult life? Intravenous drug use risktaking behavior AIDS/HIV Knowledge Level Tolerance toward AIDS/HIV infected students Self-efficacy concerning AIDS/ HIV knowledge, testing and prevention Peer and family communication about AIDS/HIV
Because the variables measured on ordinal scales had either normal or uniform distributions and because of the large sample size, both interval and ordinal or ranked scales are summarized as mean 2 standard deviation for simplicity of interpretation. Differencesbetween the intervention and control group in each of the demographic variables were assessed with X2 (30). Pretest differences in the scales that are described in Table 1 were first assessed with Kruskall-Wallis analysis of variance (ANOVA). TO control for the effects of extraneous factors, the pretest data were then tested with multivariate analysis of covariance, with intervention group, previous AIDS/HIV education, gender, racial
group, and Hispanic ethnic background as the main effects and age as the covariate (32). The ANOVA model that was used assessed the effect of the intervention group &er controlling for the other main and covariate effects (ANCOVA) in the model. Racial group ai~d Hispanic ethnic background were two separate questions on the CDC questionnaire. Because Hispanic students recorded white, black, and other as their race, these were treated as two separate variables in the analysis. Identical multivariate ANOVAs and ANCOVAs were used to assess between group differences of the posttest variables.
586
ASHWORTH ET AL.
significantly (p s 0.048) higher worry that they might get AIDS than the nonintervention group after controlling for the effect of age (p 5 0.028). No other independent variables were significant in the ANCOVA model. The educational program had no effect on any of the other attitude scales listed in Table 2. Behavior Scales The educational intervention did not have a significant effect on either self reported i.v. drug use or behavior change since learning about AIDS.
Discussion Early cross-sectional surveys of adolescents described deficiencies in knowledge and intolerant attitudes about AIDS that varied with the cultural, ethnic, and geographic profile of the adolescents (912,16,17). Although these studies described gaps in knowledge and problematic attitudes, they could not offer solutions because of their study design. Intervention programs to educate adolescent students were generally short, one-time exposures that were not specifically focused on adolescent behaviors. Hingson et al. (16) compared results from a survey of 16- to 19-year olds in 1986 with information from a similar telephone survey in 1988 (12). Although there was an increase in knowledge among adolescents, self-reported behavior did not change. The nonmatched pretest-posttest design of the study limited the conclusions that could be drawn. Also a spedfic educational program was not evaluated; instead, the generalized approach of the mass media and individual learning experiences were assumed to be the intervention. The ability of such nonspecific and nondirected efforts to change adolescents’ subsequent behavior is limited. Using a pretest-posttest noncontrol design, Miller and Downer (23) found a 13% increase in AIDS/HIV knowledge, and an increase in tolerance toward AIDS patients following a 50-min education program. The failure to include a control group limits the conclusions that can be drawn from this study. Brown et al. (17) presented pilot data of the impact of a brief AIDS-education program on knowledge and attitudes about AIDS among seventh and tenth grade public school students, using a quasiexperimental, pretest-posttest design. Students reported more knowledge, greater tolerance of AIDS patients,
JOURNAL OF ADOLESCENT HEALTH Vol. 13, No. 7
and more hesitancy toward high-risk behaviors after the program, but these changes were modest, as in our. students. In that sample, the change in knowledge was independent of a cl nge in tolerance, attitudes regarding high-risk behaviors, or coping strategies. A control group was not included, making it impossible to determine whether the changes from pretest to posttest were due to external factors such as media exposure, a testing effect, or other factors. In a follow-up study Brown et al. (20) used a nonrandomized pretest-posttest control group design to evaluate a 5-hr AIDS-education program among seventh through 12th grade students. The education program had a significant effect on knowledge, tolerant attitudes, and intention to avoid sexual intercourse as a means of AIDS prevention. Using a nonrandomized five by two repeatedmeasures factorial design, Rhodes and Walitski (33) tested the effect of four commercial videotape presentations on college students. Knowledge and attitudes about AIDS were measured prior to, immediately after, and 4-6 weeks following presentation of the videotapes. There was a significant, immediate increase in AIDS knowledge attributable to all four videotapes (p C 0.001). These gains persisted until follow-up, with erosion found only with one videotape. The sole attitude change was an increase in the perceived effectiveness of AIDS prevention methods (p 5 0.001); however, classes were not randomly assigned to the intervention and control groups, possibly biasing the results. Their results are of questionable relevance to younger high school students due to differences in cognitive maturation. Huszti et al. (18) evaluated tenth grade students in Oklahoma. Individual classes within two schools were assigned to one of two intervention groups or a control group. The interventions consisted of video educational programs and question-and-answer sessions. Although a nonintervention group was included, it is not clear how the classes of subjects were assigned “randomly” to a particular group. The possibility of contamination between friends in different treatment groups within the same school was introduced. Similar to our study, Huszti’s students demonstrated a significant increase in the level of AIDS/HIV knowledge in the intervention groups. Although knowledge dropped after the immediate posttest, it was still maintained at a higher level than at the initial examination. Unlike our study, they found that the intervention program had a significant effect on the 26-item AIDS attitude score. Huszti
November
1992
et al. did not use any scaling technique, such as principal components analysis, to delineate specific components of attitudes concerning AIDS, which may account for the differences between their findings and our study. Similar to our data, Huszti et al. did not find that the educational program had a sustained effect on the intent to change risk-taking behaviors. DiClemente et al. (22) used a similar design to evaluate a S-class period intervention in three middle and three high schools in San Francisco. Significant increases in knowledge were found among students in the intervention classes. Using a quasiexperimental pretest-posttest control group design, we documented that the intervention had a significant effect on knowledge of AIDS and HIV infection in 11th and 12th grade students. These results held true after controlling for previous AIDS education, gender, and Hispanic ethnicity. These results support previously cited findings (11,15-23). Although at pretest the intervention group reported greater worry about exposure to HIV than the control group, at posttest the differences were not significant. These pretest differences were not associated with any other factors and were assumed to be due to sampling error. The effect of the decrease in worry about exposure to the AIDS virus may be an indication that the educational effort returns concern about contracting the virus to a more realistic level. We did not find a significant difference in attitudes about AIDS patients or perceived risk of getting AIDS. Education efforts from the community at large may have affected both groups equally, or students may have failed to utilize knowledge to change their attitudes. The failure of this intervention to influence risk-taking behaviors should be interpreted with care; only one item concerning i.v. drug use was evaluated by the questionnaire because of modification required by the Human Assurance Committee and the Board of Education. The number of students admitting to i.v. drug use was small, thus limiting the usefulness of measuring their behavioral change. Moreover, the veracity of self-report could not be ascertained” Another limitation of this study was the inability to measure sexual risk-taking behavior owing to the School Board’s restrictions. This is a problem that has been encountered by many investigators. The majority of published studies demonstrate, as we did, a modest increase in high school students’ knowledge about AIDS after an initial educational program. Our study is one of the few quasiexperimental designs that included a control group. AS
SCHOOL-BASED
AIDS/HIV
EDUCATION
PROGRAM
587
such, our study suggests that brief educational programs have a limited impact on adolescents’ attitudes and knowledge about AIDS. Innovative programs that integrate information into daily situations and repetitive practice of the social skills involved may prove to have a more sustained effect on attitudes and knowledge (15). In order to effectively educate adolescents about the dangers of risktaking behaviors associated with AIDS!HIV infection, we must look beyond conventional approaches and tackle the problem with more interactive and innovative methods.
References 1.
Hcin K. Commentary on adolescent acquired immunodeficiency syndrome: The next wave of the human immunodeficiency virus epidemic? J Pediatr 1989;114:144-9.
2. Manoff SB, Gayle HD, Mays MA, Rogers MF. Acquired immunodeficiency syndrome in adolescents: Epidemiology, prevention, and public health issues. Pediatr Infect Dis 1989;8:309-14. 3. Johnson RL. Adolescent growth and development. In: Hoffman A, Greydanus, eds. Ado!escent Medicine. Norwalk, CT: Appleton and Lange, 1989:9-15. 4. Strasburger VC. Tne adolescent in contemporary Western society. In: Hoffman A, Greydanus D, eds. Adolescent Medicine. Norwalk, CT: Appleton and Lange, 1989;3-7. 5. Becker MH, JosephJG. AIDS and behavioral change to reduce risk: A review. Am J Public Health 1988;78394-410. 6. DuRant RH, Gaillard GL: Advances in health education and intervention programs for adolescents. Curr Opin Pediatr 1990;2:671-6. 7, Winkclstein W, Samuel M, Padian NS, et al. The San Francisco Men’s Health Study: III. Reduction in human immunodeficiency virus transmission among homosexual/bisexual men, 1982-86. Am J Public Health 1987;77:685-9. 8. Centers forDisease Control. Self-reported behavioral changes among homosexual and bisexual men-San Francisco. JAMA 1985;254:2537-8. 9. Helgerson SD, Petersen LR. Acquired immunodeficiency syndrome and secondary school students: Their knowledge is limited and they want to learn more. Pediatrics 1988;81: 350-5. 10. Goodman E, Cohall AT. Acquired immunodeficiency syndrome and adolescents: Knowledge, attitudes, beliefs, and behaviors in a New York City adolescent minority population. Pediatrics 1989;84:36-42. syn11. Strunin L, Hingson R. Acquired immunodeficiency drome and adolescents: Knowledge, beliefs, attitudes, and behaviors. Pediatrics 1987;79:825-8. 12. Anderson JE, Kann L, Haltzman D, et al. HIV/AIDS Knowledge and sexual behavior among high school students. Fam Plann Perspect 1990;22:252-5. 13. DuRant RH, Seymore C, Newman C, Gaillard G. High dmd students level of knowledge of AIDS/HIV and the perceived risk of currently having the AIDS infection. J School Health 1992;62:59-63. 14. Stout JW, Rivara FE. Schools and sex education: Does it work? Pediatrics 1989;83:375-9.
588
JOURNAL OF ADOLESCENT HEALTH Vol. 13, NC. 7
ASHWORTH ET AL.
15. Kirby D, Barth RP, Leland N, Fetro D. Reducing the risk: Impact of a new curriculum on sexual risk-taking. Fam Plmn Perspect 1991;23%3-63. 16. Hingson R, Strunin L, Berlin B. Acquired immunodeficiency syndrome transmission: Changes in knowledge and behaviors among teenagers, Massachusetts statewide surveys, 198&1988. Pediatrics 1990;85:24-9. 17. Brown LK, Fritz GK, uarone V. The impact of AIDS education on junior and senior high school students. J Adolesc Health Care 1989;10:386-92. 18. Huszti HC, Clopton JR, Mason PJ. Acquired immunodeficiency syndrome educational program: Effects on adolescents’ knowledge and attitudes. Pediatrics 1989;84:986-94. 19. Johnson JA, Sellow JF, Campbell AE, et al. A program using medical students to teach high school students about AIDS. J Med Ed 1988;63:522-30. 20. Brown LK, Barone V, Fritz GK, et al. AIDS educations: The Rhode Island experience. lQQ1;18:195-206, 21. Schinke SP, Gordon AN, Weston RE. Self-instruction to prevent HIV infection among African-American and HispanicAmerican adolescents. J Cen Clin Psych 1990;58:432-6. 22. DiClemente RJ. Evaluation of school-based AIDS education curricula in San Francisco. j Sex Res 1989;26:188-98. 23. Miller L, Downer A. AIDS: What you and your friends need to know-a lesson plan for adolescents. J Sch Health 1988;58:137-41. 24. Slap GB, Plotkin SL, Khalid N, et al. A human immunodeBciency virus peer education program for adolescent females. J Adolesc Health lQQ1;12:434-42.
25. Reichart CS. The statistical analysis of data from nonequi-
26. 27. 28.
29.
30. 31.
32. 33.
valent group designs. In: Cook T, Campbell D, eds. QuasiExperimentation: Design and Analysis Issues for Field Settings. Boston, Houghton-Mifflin, 1979:147-205. American Red Cross. A Letter to Brian (videotape). American Red Cross Productions. American Red Cross. Don’t Forget Sherry (videotape). American Red Cross Productions. DiClemente RJ. Pmdictors of HIV-preventive sexual behavior in a high-risk adolescent population: The influence of perceived peer norms and sexual communications on incarcerated adolescents’ consistent use of condoms. J Adolesc Health lQQ1;12:385-90. Zimet GD, Anglin TM, Lazebnik R, et al. Adolescents’ knowledge and beliefs about AIDS: Did the government brochure help? Am J Dis Child 1989;148:518. SPSS Inc. SPSS X Users’s Guide, Edition 2. Chicago, IL; SPSS Inc, 1986857-60. Zwick WR, Velicer WF: Comparison of five rules for determining the number of components to retain. Psych Bull 1986;99:432-42. Winer BJ: Statistical principles in experimental design, 2nd ed. New York, McGraw-Hill, 1971. Rhodes F, Wolitski R: Effect of instructional videotapes on AIDS knowledge and attitudes. J Am Co11Health 1989;3:26671.