Long-term Effects of a Middle School– and High School–Based Human
Immunodeficiency Virus Sexual Risk Prevention Intervention
Author Information
<http://archpedi.ama-assn.org/issues/v155n10/rfull/#aainfo> David M.
Siegel, MD, MPH; Marilyn J. Aten, PhD, RN; Maisha Enaharo, MPH
Objective To determine the longer-term effect (mean plusmnSD, 41.2
plusmn15.3 weeks; range, 14.1-80.5 weeks) of a middle school (MS)– and high
school (HS)–based human immunodeficiency virus and sexuality intervention
(Rochester AIDS Prevention Project for Youth [RAPP]) on knowledge,
self-efficacy, behavior intention, and behaviors.
Design Quasi-experimental design with 3 intervention groups and 1 control
group.
Setting Urban, predominantly ethnic, minority MS and HS health classes.
Participants Middle school and HS students (N = 4001) enrolled in health
classes in 10 schools. Fifty percent were African American; 16%, Hispanic;
20%, white; and 14%, other. Less than 10% of the students refused
participation.
Interventions There were 4 study conditions: (1) control, usual health
education curriculum taught by a classroom teacher; (2) RAPP adult health
educator, intervention curriculum implemented by highly trained health
educators; (3) RAPP peer educator, intervention implemented by extensively
trained HS students; and (4) a comparison of the RAPP intervention
curriculum taught by regular health teachers, implemented with MS students
only.
Main Outcome Measure A confidential questionnaire was administered to all
study subjects before and at long-term follow-up after the intervention,
containing scales to measure knowledge, self-efficacy, behavior intention,
and behaviors, including onset of sexual intercourse experience and
engagement in risky sexual behaviors.
Results Rates of baseline sexual activity in the sample were comparable to
those found in other urban school-based surveys. Long-term knowledge (MS
females, P<.001; and MS males, P<.01) and sexual self-efficacy (MS females,
P<.05; and HS females, P<.01) scores were higher among the intervention
groups (male and female are used in this study to describe those aged 9½-23
years). Intention to remain safe regarding sexual behavior was also greater
among intervention groups in MS but not HS. However, subjects who were
already sexually active at pretest were less likely to show a positive
intervention effect. An intervention effect for the onset of intercourse and
risky sexual behavior was found most significantly among MS females.
Conclusions A positive long-term effect from the RAPP intervention was
observed, particularly for youth who were involved in less risk (eg, not yet
sexually active) at study enrollment. Thus, we propose that the most
appropriate time for intervention implementation is earlier in adolescence,
before the onset of risky behaviors.
Arch Pediatr Adolesc Med. 2001;155:1117-1126
POA00437
DESPITE MANY efforts directed at prevention, adolescents continue to
represent a significant proportion of Americans who are diagnosed as having
sexually transmitted diseases. 1
<http://archpedi.ama-assn.org/issues/v155n10/rfull/#r1> , 2
<http://archpedi.ama-assn.org/issues/v155n10/rfull/#r2> In addition,
unintended pregnancy among US teenagers persists and carries medical,
emotional, and social costs. 3
<http://archpedi.ama-assn.org/issues/v155n10/rfull/#r3> , 4
<http://archpedi.ama-assn.org/issues/v155n10/rfull/#r4> Although only a few
of those in the United States with the acquired immunodeficiency syndrome
(AIDS) are adolescents, 5
<http://archpedi.ama-assn.org/issues/v155n10/rfull/#r5> given the long
incubation period for the human immunodeficiency virus (HIV), it is
estimated that one fifth of those with AIDS were infected as teenagers. 6
<http://archpedi.ama-assn.org/issues/v155n10/rfull/#r6> Since neither a
cure nor an effective vaccination has yet been developed, primary prevention
remains the most powerful strategy for curbing this epidemic. Since
behaviors determine the likelihood of HIV infection, prevention programs
must focus on maintaining safer behaviors (such as abstinence from sexual
intercourse or consistent use of condoms and other barriers) 7
<http://archpedi.ama-assn.org/issues/v155n10/rfull/#r7> , 8
<http://archpedi.ama-assn.org/issues/v155n10/rfull/#r8> and interventions
need to include many participants to effectively reduce HIV risk in an
adolescent population.
In an effort to capture large groups of teenagers for developing effective
HIV, sexually transmitted disease, and pregnancy prevention interventions,
research 9-12 <http://archpedi.ama-assn.org/issues/v155n10/rfull/#r9> has
been based in urban schools, where reported rates of having already
experienced sexual intercourse are relatively high. Because schools are
structured to gather students as a "captive audience," and some form of HIV
and family life education is already part of the curriculum, it is an ideal
setting for implementing and testing HIV and sexual risk prevention
interventions. In an earlier publication, 12
<http://archpedi.ama-assn.org/issues/v155n10/rfull/#r12> the short-term
results of the Rochester AIDS Prevention Project for Youth (RAPP) were
presented. The intervention was successful in increasing not only knowledge
but also intentions to behave in sexually safer ways. The RAPP intervention,
taught by peer educators (Peer Eds) and, in middle school (MS), by regular
health teachers (RHTs), demonstrated a positive effect relative to the
control and the health educator (H Ed)–taught groups. Meaningful behavior
change outcomes from the RAPP intervention can only be described after
sufficient follow-up. Little published work in this area reports on
follow-up beyond 3 to 6 months. 10
<http://archpedi.ama-assn.org/issues/v155n10/rfull/#r10> The present report
describes the longer-term (mean plusmnSD follow-up, 41.2 plusmn15.3 weeks)
effect of RAPP on knowledge, self-efficacy, behavior intentions, and, most
important, behaviors related to sexual intercourse and risk for HIV
infection.
PARTICIPANTS AND METHODS
SAMPLE
The subjects (N = 4001) ( Table 1
<http://archpedi.ama-assn.org/issues/v155n10/fig_tab/poa00437_t1.html> )
were drawn from 10 urban schools in a medium-sized northeastern city with a
population of approximately 250 000. The criteria for study inclusion were
that students were (1) enrolled in required health education classes and (2)
fluent in either English or Spanish. The ethnicity of the sample was
diverse: 50% were African American; 16%, Hispanic; 20%, white, non-Hispanic;
and 14%, other (including Asians, Native Americans, and those who indicated
that they were biracial). Although the student population of the school
district is generally of low socioeconomic status (70% of the families have
incomes less than the federal poverty level), some differences might have
confounded findings. For confidentiality reasons, and because younger
teenagers often do not know about family income, employment, or education,
we used a socioeconomic status proxy based on median house value, rent, and
family income and on educational level of the adult population within each
census tract. The proxy (socioeconomic area [SEA]) consisted of
subject-reported ZIP code and street address, and the mean SEA was 5.2 (SD,
2.7), slightly lower for MS than high school (HS) students.
PROCEDURE
Intervention
Students were recruited (as classroom cohorts) within their regular school
health education classes to participate in RAPP, a quasi-experimental
classroom-based intervention designed to increase knowledge and skills aimed
at safe behavior regarding sexuality and HIV/AIDS. Passive parental consent
for student participation was obtained. The parents of all students
scheduled to take health in the upcoming school year are routinely sent a
letter from the district director of health and physical education informing
them that family life education, including sexuality, will be taught and
they can request their son or daughter not participate in that unit. During
the study, a description of the RAPP program was a part of this letter, and
parents were given the opportunity to inquire further about RAPP and/or
refuse participation. Questions were directed to the study principal
investigator (D.M.S.), who met with parents individually to address their
concerns; few (<10 families) withdrew their son or daughter. The opportunity
to refuse study participation was verbally offered in the classroom after
the study description and before the first session. In addition, students
were told that they could withdraw from the study at any time, and on the
study instrument, instructions indicated that a student could choose not to
complete the questionnaire.
Classes were assigned within semesters to 1 of 3 conditions: (1) control,
the usual health education curriculum taught by the RHT; (2) RAPP adult H
Ed, the RAPP intervention implemented by a male-female ethnically diverse
pair of highly trained adult educators; or (3) RAPP Peer Ed, volunteer HS
students who completed approximately 50 hours of preparation by RAPP staff
and taught the RAPP curriculum as pairs of educators. Health education in MS
was taught in seventh grade only, while in HS, students had the option to
take health in 10th, 11th, or 12th grade; most students chose 10th or 11th
grade. The semester assignment of classes to the intervention condition was
based on feasibility issues and the availability of Peer Eds. However, by
study conclusion, all health classes in each of the participating MSs and
HSs had been assigned to each of the study groups (experimental and
control). At no time did intervention and control conditions take place in
the same school during a given semester. These design features enhanced
generalizability by ensuring that the study groups were spread across all
different schools, while avoiding contamination between intervention and
control classes. Within the year following the main study, regular MS
teachers only were trained to implement the curriculum. This fourth study
condition tested the transfer of the content and process of teaching to
regular school personnel.
The RAPP intervention (H Ed, Peer Ed, or RHT) consisted of 10 (HS) or 12
(MS) consecutive health class sessions (usually 2 or 3 sessions per week)
delivered for 2 to 7 weeks. The intervention was integrated into the regular
school health education schedule to avoid disruption within schools and to
build an intervention that might generalize to other schools in the future.
The content was based on literature 9
<http://archpedi.ama-assn.org/issues/v155n10/rfull/#r9> concerning
school-based interventions, the expertise of the RAPP H Eds, and principles
from the Theory of Reasoned Action 13
<http://archpedi.ama-assn.org/issues/v155n10/rfull/#r13> , 14
<http://archpedi.ama-assn.org/issues/v155n10/rfull/#r14> and normal
adolescent development. The intervention is most similar to the fourth
generation of "abstinence plus" interventions, which promote sexual
abstinence but also include safer sex messages. 11
<http://archpedi.ama-assn.org/issues/v155n10/rfull/#r11> Early sessions
emphasized self-esteem and decision-making strategies, while later classes
progressed through in-depth discussion and skill-based activities concerning
sexuality, sexually transmitted diseases, pregnancy, and, finally, HIV/AIDS.
This last topic received particular emphasis, and all sessions included
small- and large-group activities such as games, role plays, and take-home
exercises often requiring parental input. Priority was placed on maximum
engagement of the students in a highly interactive and dynamic learning
experience in both intervention conditions.
Data Collection
Students completed a confidential survey at baseline (preintervention),
immediately after the intervention, and at long-term follow-up. The survey
instrument, available in English and Spanish, was read to students during a
40-minute health class by the project H Eds. The study was reviewed and
approved by the administration of the local school district and the
university institutional research review board. Students were assured that
their answers were confidential to the research staff, and that they could
participate in the health classes without completing the research
instrument. More than 90% of the students completed the survey
preintervention. Those who declined participation were provided alternate
school activities. Subjects were tracked over time by using (1) a school
district–assigned identification (ID) number and (2) an RAPP study ID
number. Specifically, each health class teacher maintained a roster of
student names and associated school ID numbers. A separate study team list
associated the school ID numbers with unique RAPP-assigned subject numbers.
Surveys maintained by the study team were only labeled with the subject
number, while the school maintained the separate roster that contained the
student name and district ID number. This dual-list procedure confirmed
that, despite student mobility, duplicate subject enrollment did not occur
and student confidentiality was carefully preserved.
Study Instrument
The survey questionnaire, pilot tested on 450 students preceding the main
study, measured constructs determined to be important in assessing the
impact of the RAPP curriculum and has been described in detail elsewhere. 12
<http://archpedi.ama-assn.org/issues/v155n10/rfull/#r12> Those variables
reported herein include demographics, knowledge, self-efficacy regarding
sexual matters, behavior intention within the next year, history of risk
behaviors, history of sexual experiences, and self-reports of behavior. The
variable items and their psychometric properties are summarized in Table 2
<http://archpedi.ama-assn.org/issues/v155n10/fig_tab/poa00437_t2.html> .
VARIABLES MEASURED
In addition to demographics (age, sex, ethnicity, and the SEA proxy for
socioeconomic status), scales (summarized in Table 2
<http://archpedi.ama-assn.org/issues/v155n10/fig_tab/poa00437_t2.html> )
were developed during a pilot phase, because of the lack of reliable and
valid scales reported in the literature. Existing literature was reviewed to
include variables represented in the Theory of Reasoned Action.
Reliabilities were computed separately for MS and HS students and ranged
from 0.74 to 0.80 (n = 3800 students) for alpha(internal consistency) and
from 0.66 to 0.84 for test-retest reliability (n = 450 students). Factor
analyses for each construct revealed 1-factor solutions for all constructs
except "involvement in sexual risk behaviors," which supported 3 factors
("somesex" represented the initiation of sexual intercourse, and "risksex"
represented some risk involvement). Variance accounted for by the factors
ranged from 10% to 36%.
Knowledge
Students responded to the 26 items with yes if they believed the statement
to be true; no, if false; and not sure (a choice scored as incorrect and
included to minimize guessing and inflation of correct response scores). To
avoid a ceiling effect, individual items were included only if they had less
than 80% correct responses during the pilot phase.
Sex Self-efficacy
This scale, developed from similar work by Misovich et al, 15
<http://archpedi.ama-assn.org/issues/v155n10/rfull/#r15> asked how hard or
easy it would be to carry out each of 8 behaviors in relation to sexuality.
Behavior Intention
An index of intention to behave in safe ways 12
<http://archpedi.ama-assn.org/issues/v155n10/rfull/#r12> included 9 items
to indicate student agreement or disagreement on 7-point scales with
selected statements concerning sexual behavior and substance use.
Life Risk History
Items from the Youth Risk Behavior Survey 16
<http://archpedi.ama-assn.org/issues/v155n10/rfull/#r16> were used to tap
15 questions concerning school- and community-related risks. Furthermore, we
asked a panel of 25 experts in adolescent health (clinicians and behavioral
scientists) to rank the level of risk: 0 indicates no or minimal risk (eg,
missed school without permission); 1, some risk (eg, tried marijuana); and
2, substantial risk (eg, used marijuana regularly).
History of Sexual Intercourse
Preintervention, students were asked about their history of sexual
intercourse as part of 7 different items addressing onset, frequency, and
multiple partner experience. The degree to which students were consistent
across all 7 items in which there was an opportunity to answer "I have never
had sex" was examined to be confident regarding the validity of responses.
Particularly for younger students, it was important that subjects understood
the concept of sexual intercourse before initiating the intervention.
Students were categorized as ever having had sexual intercourse (score of 1)
or never having had sexual intercourse (score of 0).
Involvement in Sexual Risk Behaviors
To assess involvement in sexual behavior, students were queried as to their
initiation and engagement in 13 selected behaviors. Factor analyses resulted
in 3 dimensions: (1) somesex or the initiation and onset of sexual
intercourse experience, which included 5 items (ever carry condoms, ever had
sexual intercourse, communication with a partner about sex, have had sex 1-5
times within the past 3 months, and planning ahead to have sex),
representing 33% of the variance; (2) risksex or the engagement in some
risky behaviors, which included 5 items (tried to get pregnant or get a
partner pregnant, actual pregnancy involvement, having had sex when the
teenager really did not want to, having sex while using alcohol or other
drugs, and having had sex >5 times in the past 3 months), representing 10%
of the variance; and (3) engagement in high-risk sexual behaviors (history
of sex with an intravenous drug user, anal sex, oral sex, and a history of
HIV testing), representing 8% of the variance. The third dimension was not
meaningful to further analyses as there were few students who engaged in
these most serious behaviors. The first 2 factors were used as dependent
variables in analyses of whether and how much engagement in actual sexual
behavior was reported at longer-term follow-up for all intervention groups.
CLASS CLIMATE
To test for any differences across various learning settings, the existing
health education class environment was observed and scored by the adult RAPP
educators for all participating teachers/classrooms. Working independently,
each member of a pair of RAPP H Eds rated the physical environment and the
classroom health teacher's facilitation of the RAPP curriculum. The 18 items
were summed to form an overall "class climate" score (range, 0-36). Rater
agreement was high (r>0.80), and the 2 scores were averaged.
DATA ANALYSES
Since the study design was a quasi experiment, it was important to consider
whether there were preintervention differences in study variables that might
confound findings in relation to long-term outcomes. Therefore, demographics
(age, sex, SEA, and ethnicity), the proportions of females and males (female
and male are used in this study to describe those aged 9½-23 years) who
reported sexual intercourse experience, and the study variables of interest
(knowledge, sexual self-efficacy, general life risk, sex safe behavior
intention, and the self-reports of behaviors concerning initiation of sexual
behavior and involvement in risky sexual behavior) were compared within the
MS and HS groups and by sex. Previous work 17
<http://archpedi.ama-assn.org/issues/v155n10/rfull/#r17> , 18
<http://archpedi.ama-assn.org/issues/v155n10/rfull/#r18> has documented
differences in sexual behavior by age and sex (eg, earlier sexual debut in
males vs females). Also, we compared the observation rating for the class
climate score among the MS and HS groups. Then, the major study analyses
concerning long-term outcomes were tested by repeated-measures analyses of
variance.
The study was a 3-factor design (4 by 2 by 4); the factors were ethnicity
(African American, Hispanic, non-Hispanic white, and other), history of
sexual intercourse experience (yes or no), and intervention group (control,
H Ed–taught, Peer Ed–taught, and, in MS, RHT-taught students). The procedure
for analyses was that demographics, the general life risk score, the mean
score for length of time from intervention, the class climate score, and the
relevant pretest score for each dependent variable were entered as
covariates. Then, each of the factors was entered with the intervention
factor considered last. The strategy was to delineate the long-term
intervention effect beyond other related variables. The interaction between
sexual activity status and intervention group was also examined to determine
a given intervention's differential effectiveness based on whether subjects
were sexually active. Since the sample was large and statistical
significance is easily reached with large sample sizes, effect sizes (eta
values) are reported for the intervention, as is the total variance
accounted for by the model (R2).
RESULTS
The mean plusmnSD duration of long-term follow-up in this study was 41.2
plusmn15.3 weeks (range, 14.1-80.5 weeks); overall, two thirds of the
subjects were present at study conclusion (72% in MS and 55% in HS). Among
12th graders, 73% did not complete long-term follow-up because of graduation
or drop out, while the attrition for 10th and 11th grade was 55% and 43%,
respectively. The 4 intervention groups among the MSs (control, H Ed, Peer
Ed, and RHT) and the 3 groups among the HSs (control, H Ed, and Peer Ed)
were compared to determine whether there were significant differences in
demographic variables ( Table 1
<http://archpedi.ama-assn.org/issues/v155n10/fig_tab/poa00437_t1.html> ) and
for the study variables of interest. While there were significant
differences for age, SEA, ethnicity, and the proportions of students who
reported a history of sexual intercourse experience preintervention, the
magnitude of the differences was, for the most part, small. The most
important differences were that, for MS students, the proportion of females
who reported having experienced intercourse preintervention ranged from
18.4% for RHT-taught students to 34.2% for H Ed–taught students. In
contrast, many more males among the MS sample reported sexual experience
preintervention, ranging from 53.2% of RHT-taught students to 66.2% of Peer
Ed–taught students. Overall, in MS, about 30% of females and 63% of males
reported sexual experience at baseline. Among the HS students, the range of
sexual experience among females was from 60.6% of Peer Ed–taught students to
73.2% of the controls, while the range among males was from 69.3% of Peer
Ed–taught students to 82.3% of controls. Demographics and the self-reported
history of sexual intercourse experience were entered as covariates in all
analyses.
Pretest scores for the major study variables were also used as covariates in
analyses of long-term intervention effects since there were slight
preintervention differences across study groups. It was also important to
control for each subject's pretest score before assessing the intervention
effect. Most differences, however, while statistically significant, did not
indicate meaningful distinctions. For example, MS sexual self-efficacy
scores were higher (safer) among H Ed–taught students (37.5), Peer Ed–taught
students (37.5), and RHT-taught students (37.9) compared with control
students (35.7) (P<.001). Classroom climate scores represented the greatest
preintervention differences between control and intervention groups, with
the largest discrepancy between MS control (22.5) and RHT (29.9) groups
(P<.001) on a scale that ranged from 12 to 31.
Table 3
<http://archpedi.ama-assn.org/issues/v155n10/fig_tab/poa00437_t3.html> ,
Table 4
<http://archpedi.ama-assn.org/issues/v155n10/fig_tab/poa00437_t4.html> ,
Table 5
<http://archpedi.ama-assn.org/issues/v155n10/fig_tab/poa00437_t5.html> , and
Table 6
<http://archpedi.ama-assn.org/issues/v155n10/fig_tab/poa00437_t6.html>
present the analyses of variance models conducted for each of the dependent
variables. The format of these tables is the same as that used in an earlier
article. 12 <http://archpedi.ama-assn.org/issues/v155n10/rfull/#r12> Using
Table 3
<http://archpedi.ama-assn.org/issues/v155n10/fig_tab/poa00437_t3.html> as
an example, the data presentation is as follows. Subjects were stratified by
school level and sex, with variables sequentially entered into the analysis
of variance, shown in the table as rows. In Table 3
<http://archpedi.ama-assn.org/issues/v155n10/fig_tab/poa00437_t3.html>
(predicting knowledge at posttest), for MS females, the collective
contribution of the covariates to the model yielded an F value of 14.7, with
a significance of P<.001. Taking the covariates separately, age was a
meaningful predictor, with an F value of 11.4, as was SEA, with an F value
of 24.9.
Also in Table 3
<http://archpedi.ama-assn.org/issues/v155n10/fig_tab/poa00437_t3.html> , for
MS females, the combined contribution of all main effects (to the posttest
knowledge score) was significant, with an F value of 12.2, while ethnicity
alone accounted for an F value of 10.1. In the lower half of the table, the
mean score from the study instrument for the dependent variable in question
is provided. In Table 3
<http://archpedi.ama-assn.org/issues/v155n10/fig_tab/poa00437_t3.html> ,
this would be the mean knowledge scale score. For example, among the MS
females, the F value for the contribution of intervention group membership
to knowledge score at posttest was 20.3, and the mean knowledge scale score
for the control group was 12.9 compared with 15.6, 15.9, and 15.5 for the H
Ed, Peer Ed, and RHT groups, respectively.
For long-term knowledge ( Table 3
<http://archpedi.ama-assn.org/issues/v155n10/fig_tab/poa00437_t3.html> ),
covariates were important contributors to outcomes. Increasing age, a higher
SEA, a better class climate score, time from intervention (except HS males),
and the pretest knowledge score (except HS males) were significant. The
long-term knowledge means were consistently greater for the intervention
groups compared with the controls, and were significant for MS females and
males. There were some ethnic differences; white non-Hispanic students
generally had higher knowledge scores, followed by African Americans and
Hispanics. The means for self-efficacy ( Table 4
<http://archpedi.ama-assn.org/issues/v155n10/fig_tab/poa00437_t4.html> )
were higher for each of the intervention groups compared with the controls,
reaching significance for the MS and HS females. Ethnic differences were
noted in that Hispanic youth generally reported less self-efficacy than did
other groups. Covariate significance was almost entirely accounted for by
the pretest self-efficacy score and was highly significant (F values ranged
from 80.0 to 175.0). There were no differences for either knowledge or
self-efficacy in relation to whether there was a history of sexual
intercourse. Long-term etavalues ranged from 0.17 to 0.29 for knowledge and
from 0.11 to 0.15 for self-efficacy. The proportions of variance explained
by the models (R2) ranged from 0.11 to 0.27 for knowledge and from 0.12 to
0.35 for self-efficacy, with more variance explained in the HS models.
Intention to remain safe in regard to sexual behavior was the third variable
considered for short-term outcomes, 12
<http://archpedi.ama-assn.org/issues/v155n10/rfull/#r12> and tested again
for longer-term outcome in the present analyses ( Table 5
<http://archpedi.ama-assn.org/issues/v155n10/fig_tab/poa00437_t5.html> ).
Among MS students, the means were lower (representing less intention to
remain safe, including remaining abstinent) for the controls than the
intervention groups. There was no intervention effect for HS students,
likely reflecting the high prevalence of sexual experience that preceded the
intervention. In all groups, the pretest score for intention was again the
most significant and powerful covariate (F values ranged from 98.2 to
223.7). Other covariates, especially the general life risk score (F values
ranged from 9.9 to 37.6), were also significant. Unlike knowledge and
self-efficacy, the means for intention to be safe were lower for those
students who reported a history of sexual experience preintervention,
especially for the MS males (P<.01). The etavalues ranged from 0.11 to 0.26,
and the R2 for the model ranged from 0.26 to 0.54; greater variance was
explained among HS students.
Finally, we examined the long-term intervention effect for the index scores
described (in the "Variables Measured" subsection of the "Participants and
Methods" section), which represented initiation of sexual activity (somesex
[ Table 6
<http://archpedi.ama-assn.org/issues/v155n10/fig_tab/poa00437_t6.html> ])
and engagement in more risky sexual behavior (risksex). For somesex, with
the exception of HS females, the means were in the expected direction for an
intervention effect, ie, the higher means (representing less involvement in
sexual exploration) were found for the intervention groups compared with the
controls. Significance, however, was reached only among the MS females,
likely because this group reported the least (30%) history of sexual
experience preintervention. There was also a significant interaction effect
for sex history by intervention for Peer Ed–taught males at the HS level.
The pretest covariate for somesex was highly significant (F values ranged
from 207.4 to 320.9), and the history of sexual experience was also
significantly related to long-term sexual behavior (F values ranged from
30.3 to 45.4), with higher (safer) scores among those not yet sexually
active. Furthermore, the general life risk covariate was also significant
for all groups, suggesting that preintervention risk (whether sexual or
nonsexual) was significant in predicting later sexual behavior. In relation
to ethnicity, African American youth had lower mean scores for this
variable, indicating that they reported greater engagement in sexual
activity. The etavalues were small (0.06-0.17), and the R2 values for the
models ranged from 0.40 to 0.54.
Prediction of engagement in risky sex behavior (risksex) was more difficult;
significance for the intervention was not demonstrated, as most students did
not report these behaviors. However, the means were in the expected
direction for the intervention groups, with the controls having slightly
lower mean scores (representing less safety) than the intervention groups
among all students. There was a statistically significant interaction for
sex activity status by intervention for MS females, indicating that sexually
active (vs nonsexually active) females were more positively affected by the
intervention (P<.001). Analogous to the other analyses, increasing age,
greater life risk, a history of sexual experience preintervention, and
especially the pretest score for risky behavior (F values ranged from 44 to
223) were highly significant. The degree of variance explained by each of
the 4 models for risksex ranged from an R2 of 0.17 to an R2 of 0.43.
COMMENT
Numerous efforts to reduce behavioral risk (especially sexual risk) have
examined short-term change, while fewer studies follow up subjects beyond 6
months. Findings describing 12-month after the intervention follow-up in
school-based studies are unusual. 10
<http://archpedi.ama-assn.org/issues/v155n10/rfull/#r10> , 19
<http://archpedi.ama-assn.org/issues/v155n10/rfull/#r19> The etavalues for
interventions are most often larger for nonbehavioral outcomes (such as
knowledge) and frequently lessen over time. 19
<http://archpedi.ama-assn.org/issues/v155n10/rfull/#r19> , 20
<http://archpedi.ama-assn.org/issues/v155n10/rfull/#r20> The sample in this
report, by virtue of the available school population, included students at
serious sexual risk and those youngsters who were at potentially no risk for
early life entry into sexual risk. Longer-term behavior etavalues are
usually small in such studies, 19
<http://archpedi.ama-assn.org/issues/v155n10/rfull/#r19> , 21
<http://archpedi.ama-assn.org/issues/v155n10/rfull/#r21> and in many
instances have not been considered at all. Our long-term etavalues ranged
from a medium effect of 0.25 (knowledge) to smaller behavior etavalues of
0.14 (somesex) and 0.13 (risksex), findings that are consistent with other
adolescent sex risk reduction studies and with behavior studies aimed at
risk reduction in other domains.
There was a positive, sustained, long-term effect of the RAPP intervention
when compared with the control intervention in the areas of knowledge,
self-efficacy regarding sexual matters, behavior intention, and
self-reported behaviors. While statistical significance was not universally
reached, when compared with control across intervention groups, mean scores
were consistently in the desired safer direction. This was especially true
for the Peer Ed– and (at MS) the RHT-taught groups. Since students' pretest
variable score heavily influenced their score on that same variable at
follow-up, this observation speaks to the need for development and testing
of school-based sexual risk reduction interventions among younger students,
such as those in the late elementary grades. In a preliminary analysis of
our data (not presented herein), we had not initially included sexual
history as a covariate and found that there was more significant effect of
the intervention compared with the control. When sexual history was entered
into the analysis of variance before examining for the intervention effect,
the impact of the latter was diminished. This is consistent with the
association of a more positive outcome in the presence of "safer" pretest
scores, because the initiation of sexual intercourse is included as a
contributing item to the somesex and risksex indexes.
Interpretations of school-based intervention trials such as RAPP are
bolstered by the inclusion of many subjects who are representative of a
general adolescent population. The prospects of implementing programs found
to be successful are also brighter compared with community-based efforts
since school structures already exist in every community and are, in fact,
obligated to provide curriculum time related to family values and sexuality.
At the same time, there are limitations to be considered in evaluating RAPP.
The quasi-experimental design we used was necessary to carry out a large
school-based study that would not disrupt the usual class and grade
structure to such an extent that the project would have been rejected by the
participating institutions. This resulted, however, in lack of true subject
randomization. Our analyses for group differences in baseline
characteristics revealed statistical significance across several variables,
but the magnitude of scale score differences was not clinically meaningful,
and the relevant variables were entered in the analysis of variance to
further control for their potential effect on the dependent variable. In
addition, not all subjects who enrolled in the study were present at longest
follow-up. Such attrition, greater in HS and particularly among 12th
graders, presents potential bias in results. We examined relevant pretest
characteristics of those students who did and did not participate in late
follow-up, and there was a slight overrepresentation of higher-risk (related
to sexual behavior) students among study dropouts. Their inclusion at
follow-up might have diminished the observed intervention effect, but again
we expected and found that the most intervention-responsive students were
those who were yet to engage in risk behaviors.
Another reality of this type of work is that the most meaningful outcome,
behavior, is measured by self-report. With this potential source of error in
mind, we asked subjects about sexual experience in multiple ways (see the
"Variables Measured" subsection of the "Participants and Methods" section)
and assessed their behavior only based on consistent responses. Thus, we
believe our index of sexual activity is accurate. This and the other
measures for the study were pilot tested at MS and HS levels, and the
instrument was not finalized until reliability and validity were addressed.
It is clear that the goal for school-based interventions of this type should
be the primary prevention of risky sexual behavior. Since this can only
occur in younger, precoital populations, we propose that late elementary
school students, before the transition to sexual activity, should be the
target group for the next phase of study.
Author/Article Information
From the Department of Pediatrics (Drs Siegel and Aten) and the School of
Nursing (Dr Aten), University of Rochester, and the Department of
Pediatrics, Rochester General Hospital (Dr Siegel and Ms Enaharo),
Rochester, NY.
Corresponding author and reprints: David M. Siegel, MD, MPH, Department of
Pediatrics, Rochester General Hospital, 1425 Portland Ave, Rochester, NY
14621 (e-mail: [log in to unmask]
<mailto:[log in to unmask]> ).
Accepted for publication April 10, 2001.
This study was supported by grant 49037 from the National Institute of
Mental Health, Rockville, Md.
We thank Barbara Thompson for her tireless preparation of the manuscript;
the staff of the Rochester AIDS Prevention Project for Youth, including the
health educators (Margaret Cain, BA; Raúl Corujo-Molina; Desiree Voorhies,
RN, MSEd; and Lennard Wedderburn, CSW) and the research assistant (Terri
Vaughn, CSW), for their dedication, commitment, and hard work on behalf of
the project; and the staff and students of the participating schools.
What This Study Adds
A significant proportion of young adults with AIDS became infected with HIV
through sexual contact during adolescence. A prior report showed that a
school-based intervention was successful in the short-term in improving
knowledge and changing behavior.
This study found that at an average of 10½ months after the intervention, MS
and HS students still demonstrated a positive effect of the program on
several important measures related to safe sex. Intensively trained HS Peer
Eds can function as effective implementers of an HIV sexual risk prevention
program.
REFERENCES
1. <http://archpedi.ama-assn.org/issues/v155n10/rfull/#rr1>
Cates W.
The epidemiology and control of sexually transmitted diseases in
adolescents.
In: Schydlower M, Shafer M, eds. Adolescent Medicine: State of the Art
Reviews. Philadelphia, Pa: Hanley & Belfus Inc; 1990:409-428.
2. <http://archpedi.ama-assn.org/issues/v155n10/rfull/#rr2>
Burstein GR, Gaydos CA, Diener-West M, Howell MR, Zenilman JM, Quinn TC.
Incident chlamydia trachomatis infections among inner-city adolescent
females.
JAMA.
1998;280:521-526.
ABSTRACT <http://jama.ama-assn.org/issues/v280n6/abs/joc80009.html> |
FULL TEXT <http://jama.ama-assn.org/issues/v280n6/rfull/joc80009.html> |
PDF <http://jama.ama-assn.org/issues/v280n6/rpdf/joc80009.pdf> | MEDLINE
<http://www.ncbi.nlm.nih.gov/htbin-post/Entrez/query?db=m&form=6&Dopt=r&uid=
9707141>
3. <http://archpedi.ama-assn.org/issues/v155n10/rfull/#rr3>
Epner JEG, ed.
Policy Compendium on Reproductive Health Issues Affecting Adolescents.
Chicago, Ill: American Medical Association; 1996.
4. <http://archpedi.ama-assn.org/issues/v155n10/rfull/#rr4>
Alan Guttmacher Institute.
Sex and America's Teenagers.
New York, NY: Alan Guttmacher Institute; 1994.
5. <http://archpedi.ama-assn.org/issues/v155n10/rfull/#rr5>
Centers for Disease Control and Prevention.
HIV/AIDS Surveillance Report.
Atlanta, Ga: Centers for Disease Control and Prevention; 1998.
6. <http://archpedi.ama-assn.org/issues/v155n10/rfull/#rr6>
Boyer CB, Kegeles SM.
AIDS risk and prevention among adolescents.
Soc Sci Med.
1991;33:11-23.
MEDLINE
<http://www.ncbi.nlm.nih.gov/htbin-post/Entrez/query?db=m&form=6&Dopt=r&uid=
1882237>
7. <http://archpedi.ama-assn.org/issues/v155n10/rfull/#rr7>
Rotheram-Borus MJ, Mahler KA, Rosario M.
AIDS prevention with adolescents.
AIDS Educ Prev.
1995;7:320-336.
MEDLINE
<http://www.ncbi.nlm.nih.gov/htbin-post/Entrez/query?db=m&form=6&Dopt=r&uid=
7577308>
8. <http://archpedi.ama-assn.org/issues/v155n10/rfull/#rr8>
Stanton B, Kim N, Galbraith J, Parrott M.
Design issues addressed in published evaluations of adolescent HIV-risk
reduction interventions: a review.
J Adolesc Health.
1996;18:387-396.
MEDLINE
<http://www.ncbi.nlm.nih.gov/htbin-post/Entrez/query?db=m&form=6&Dopt=r&uid=
8803730>
9. <http://archpedi.ama-assn.org/issues/v155n10/rfull/#rr9>
Kirby D, Short L, Collins J, et al.
School-based programs to reduce sexual risk behaviors: a review of
effectiveness.
Public Health Rep.
1994;109:339-360.
MEDLINE
<http://www.ncbi.nlm.nih.gov/htbin-post/Entrez/query?db=m&form=6&Dopt=r&uid=
8190857>
10. <http://archpedi.ama-assn.org/issues/v155n10/rfull/#rr10>
Kim N, Stanton B, Li X, Dickerson K, Gailbraith J.
Effectiveness of the 40 adolescent AIDS-risk reduction interventions: a
quantitative review.
J Adolesc Health.
1997;20:204-215.
MEDLINE
<http://www.ncbi.nlm.nih.gov/htbin-post/Entrez/query?db=m&form=6&Dopt=r&uid=
9069021>
11. <http://archpedi.ama-assn.org/issues/v155n10/rfull/#rr11>
Thomas MH.
Abstinence-based programs for prevention of adolescent pregnancies.
J Adolesc Health.
2000;26:5-17.
MEDLINE
<http://www.ncbi.nlm.nih.gov/htbin-post/Entrez/query?db=m&form=6&Dopt=r&uid=
10638713>
12. <http://archpedi.ama-assn.org/issues/v155n10/rfull/#rr12>
Siegel DM, Aten MJ, Roghmann KJ, Enaharo M.
Early effects of a school-based human immunodeficiency virus infection and
sexual risk prevention intervention.
Arch Pediatr Adolesc Med.
1998;152:961-970.
ABSTRACT <http://archpedi.ama-assn.org/issues/v152n10/abs/poa7567.html> |
FULL TEXT <http://archpedi.ama-assn.org/issues/v152n10/rfull/poa7567.html>
| PDF <http://archpedi.ama-assn.org/issues/v152n10/rpdf/poa7567.pdf> |
MEDLINE
<http://www.ncbi.nlm.nih.gov/htbin-post/Entrez/query?db=m&form=6&Dopt=r&uid=
9790605>
13. <http://archpedi.ama-assn.org/issues/v155n10/rfull/#rr13>
Ajzen I, Fishbein M.
Understanding Attitudes and Predicting Social Behavior.
Englewood Cliffs, NJ: Prentice-Hall International Inc; 1980.
14. <http://archpedi.ama-assn.org/issues/v155n10/rfull/#rr14>
Fishbein M.
AIDS and behavior change: an analysis based on Theory of Reasoned Action.
Interamerican J Psychol.
1990;24:37-56.
15. <http://archpedi.ama-assn.org/issues/v155n10/rfull/#rr15>
Misovich SJ, Fisher WA, Fisher JD.
Understanding and promoting AIDS preventive behaviors: measures of AIDS risk
reduction information, motivation, behavioral skills, and behavior.
In: Davis CM, Yarbor WH, Bauserman R, Scheer G, Davis SL, eds. Sexuality
Related Measures: A Compendium. Thousand Oaks, Calif: Sage Publications;
1998.
16. <http://archpedi.ama-assn.org/issues/v155n10/rfull/#rr16>
Division of Adolescent and School Health, Center for Chronic Diseases
Prevention and Health Promotion, Centers for Disease Control and Prevention.
Youth Risk Behavior Survey.
Atlanta, Ga: Division of Adolescent and School Health, Center for Chronic
Diseases Prevention and Health Promotion, Centers for Disease Control and
Prevention; 1990.
17. <http://archpedi.ama-assn.org/issues/v155n10/rfull/#rr17>
Siegel DM, Aten MJ, Roghmann KJ.
Self-reported honesty among middle and high school students responding to a
sexual behavior questionnaire.
J Adolesc Health.
1998;23:20-28.
MEDLINE
<http://www.ncbi.nlm.nih.gov/htbin-post/Entrez/query?db=m&form=6&Dopt=r&uid=
9648019>
18. <http://archpedi.ama-assn.org/issues/v155n10/rfull/#rr18>
Paikoff RL.
Early heterosexual debate: situations of sexual possibility during the
transition to adolescence.
Am J Orthopsychiatry.
1995;65:389-401.
MEDLINE
<http://www.ncbi.nlm.nih.gov/htbin-post/Entrez/query?db=m&form=6&Dopt=r&uid=
7485424>
19. <http://archpedi.ama-assn.org/issues/v155n10/rfull/#rr19>
Kalichman SC, Carey MP, Johnson BT.
Prevention of sexually transmitted HIV infection: a meta-analytic review of
the behavioral outcome literature.
Ann Behav Med.
1996;18:6-15.
20. <http://archpedi.ama-assn.org/issues/v155n10/rfull/#rr20>
Walter HJ, Vaughan MS, Ragin DR, Cohall AT, Kasen S, Fullilove RE.
Prevalence and correlates of AIDS-risk behaviors among urban minority high
school students.
Prev Med.
1993;22:813-824.
MEDLINE
<http://www.ncbi.nlm.nih.gov/htbin-post/Entrez/query?db=m&form=6&Dopt=r&uid=
8115340>
21. <http://archpedi.ama-assn.org/issues/v155n10/rfull/#rr21>
Stanton B, Kim N, Galbraith J, Parrott M.
Design issues addressed in published evaluations of adolescent HIV risk
reduction interventions: a review.
J Adolesc Health.
1996;18:387-396.
Edward E. Rylander, M.D.
Diplomat American Board of Family Practice.
Diplomat American Board of Palliative Medicine.