Results from the National Longitudinal Study of Adolescent Health

Home , Longitudinal study

ARTICLE Predicting Adolescents’ Longitudinal Risk for Sexually Transmitted Infection Results From the National Longitudinal Study of Adolescent Health

Carol A. Ford, MD; Brian Wells Pence, MPH; William C. Miller, MD, PhD; Michael D. Resnick, PhD; Linda H. Bearinger, MS, PhD; Sandy Pettingell, PhD; Myron Cohen, MD

Background: Influencing adolescents’ sexual behav- Main Outcome Measure: Positive test result for C tra- iors has the potential to influence trajectories of risk for chomatis, N gonorrhoeae,orT vaginalis. sexually transmitted infections (STIs) among young adults. Results: Controlling for biological sex, age, race/ ethnicity, family structure, and maternal education, ado- Objective: To determine whether family, school, and lescents who perceived that their parents more strongly individual factors associated with increased duration of disapproved of their having sex during adolescence were virginity also protect against STIs in young adulthood. less likely to have STIs 6 years later (adjusted odds ratio, 0.89; 95% confidence interval, 0.81-0.99). Those with Design: Prospective cohort study. Wave I of the Na- a higher grade point average during adolescence were also tional Longitudinal Study of Adolescent Health oc- less likely to acquire STIs (adjusted odds ratio, 0.84; 95% confidence interval, 0.71-0.99). Stratified analyses con- curred in 1995 when participants were in grades 7 through firmed these findings among female, but not male, ado- 12. Six years later, all wave I participants who could be lescents. Feelings of connection to family or school, re- located were invited to participate in wave III and pro- ported importance of religion, attending a parochial vide a urine specimen for STI testing. school, and pledges of virginity during adolescence did not predict STI status 6 years later. Setting: In-home interviews in the continental United States, Alaska, and Hawaii. Conclusions: Perceived parental disapproval of sexual intercourse and higher grades in school during adolescence Participants: Population-based sample. Of 18 924 par- have protective influences on the trajectory of risk for ac- ticipants in the nationally representative weighted wave quiring STIs, primarily among female adolescents. Most fac- I sample, 14 322 (75.7%) were located and participated tors associated with increased duration of virginity in ado- in wave III. Test results for Chlamydia trachomatis, Neis- lescence do not influence the trajectory of STI risk. seria gonorrhoeae, and Trichomonas vaginalis were available for 11 594 (81.0%) of wave III participants. Arch Pediatr Adolesc Med. 2005;159:657-664

EXUALLY TRANSMITTED INFEC- acquiring STIs. Encouraging increased du- tions (STIs) cause consider- ration of virginity is one strategy to re- Author Affiliations: Adolescent able personal and societal duce STI prevalence because the imma- Medicine Program (Dr Ford); burden in the United States.1 ture cervix is biologically predisposed to Departments of Medicine Strategies to reduce this bur- becoming infected if exposed,9-11 and those (Drs Ford, Miller, and Cohen), den must target adolescents and young who delay initiation of sexual inter- Pediatrics (Dr Ford), and S2 3 Microbiology and Immunology adults. A recent population-based study course (sexual debut) until they are older (Dr Cohen), School of found that 4.7% of young women and 3.7% often engage in less risky sexual behav- 12-15 Medicine; and Department of of young men in the United States be- iors. Epidemiology (Mr Pence and tween the ages of 18 and 26 years are in- Wave I of the National Longitudinal Dr Miller), School of Public fected with Chlamydia trachomatis, and Study of Adolescent Health (Add Health) Health, The University of North that the prevalence of infection is substan- contributed to our understanding of the Carolina at Chapel Hill; tially higher among certain racial and eth- characteristics of families and schools as- Division of General Pediatrics nic minority groups. sociated with delaying sexual debut in a and Adolescent Medicine, Biological factors, sexual networks, large nationally representative adoles- Medical School (Drs Resnick partner characteristics, health care, and so- cent sample.16 By using a conceptual para- and Bearinger), and Center for 17 Adolescent Nursing, School of cial context contribute to the risk of digm from the resiliency literature and 1,4-8 Nursing (Drs Bearinger and STIs. Sexual behaviors are also key de- focusing on factors that were potentially Pettingell), University of terminants of risk, and virginal adoles- modifiable, cross-sectional analyses re- Minnesota, Minneapolis. cents and young adults are at low risk of vealed that adolescents who felt con-

(REPRINTED) ARCH PEDIATR ADOLESC MED/ VOL 159, JULY 2005 WWW.ARCHPEDIATRICS.COM 657

©2005 American Medical Association. All rights reserved. Downloaded From: https://jamanetwork.com/ on 09/26/2021 nected to family, lived in families with good mental health, parental disapproval of contraception (Cronbach ␣=.91) were and perceived that their parents more strongly disap- created from items scored on Likert-type scales.16,22 Family sui- proved of them having sex or using contraception were cide attempts and/or completions were measured by the fol- more likely to delay sexual debut than comparison lowing question: “Have any family members tried to kill them- groups.16 School factors associated with delay in sexual selves or succeeded in killing themselves over the past 12 debut included feelings of being connected to school and months?” attending a parochial school. Individual factors associ- Wave I School Variables. A previously described measure for ated with later sexual debut included adolescents’ feel- feelings of being connected to school was created using an 8-item ings about importance of religion, making a pledge of vir- scale (Cronbach ␣=.71).16,22 Measures for mean daily atten- ginity, perceived likelihood of early death, number of dance for the participants’ school and school type (coded as hours worked per week, perceived physical maturity com- parochial or not) were obtained from wave I school adminis- pared with peers, and grade point average (GPA). trator questionnaires. Wave III of Add Health provides an opportunity to determine whether factors that influence age of sexual de- Wave I Individual Variables. The importance of religion was mea- but also influence longitudinal trajectories of risk for sured by the following questions: How important is religion to STIs—using biological tests. In this study, we test the hy- you? (There were 4 response categories, ranging from very to not pothesis that family, school, and individual factors as- at all.) How often do you pray? (There were 5 response categories, ranging from at least once a day to never.) What is your re- sociated with increased duration of virginity during ado- 16 ligion? Responses to the first 2 items were recoded from 0 to 1 so lescence at wave I longitudinally decrease risk of that lower values represented lower importance of religion, and infection with C trachomatis, Neisseria gonorrhoeae, or the mean of the 2 items was calculated as an index of religious Trichomonas vaginalis at wave III. importance. Respondents who reported not having a religious affiliation and who did not answer these questions were coded as METHODS 0. Respondents who did report a religious affiliation had to answer at least 1 of the first 2 questions to receive a score. The index was then standardized to have a mean of 0 and an SD of 1. Add Health is a prospective cohort study following up almost Respondents who reported feelings of attraction toward a per- 18 20 000 adolescents into adulthood. The original Add Health son of the same sex or a romantic relationship with a person sample was identified from 1994 to 1995, using a stratified of the same sex were coded as having same-sex attractions or school-based sample of middle and high school students in relationships. The perceived risk of an early death was mea- the United States. Wave I interviews were conducted in 1995. sured by asking participants to indicate the chance that they Wave III occurred from 2001 to 2002, and included private will live until the age of 35 years. The number of hours en- interviews of original participants (N=14 322) and testing for gaged in paid work in a typical nonsummer week was coded C trachomatis and N gonorrhoeae by ligase chain reaction and as at least 20 per week or less. Self-report of physical appear- T vaginalis by polymerase chain reaction. The original sam- ance was used to measure whether participants believed they pling design, and all Add Health procedures, have been de- looked older or younger than most others their age. Grade point 3,16,19-21 scribed in detail elsewhere. For this study, we test average was calculated from self-report of grades in the most whether the wave I variables associated with age of sexual de- recent grading period in English/language arts, mathematics, 16 but predict STI status at wave III. The University of North history/social studies, and science. Virginity pledge was mea- Carolina at Chapel Hill Institutional Review Board approved sured by the following item: Have you taken a public or writ- all study procedures. ten pledge to remain a virgin until marriage?

MEASURES ANALYSES

Outcome Variable Descriptive frequencies were derived from unweighted data. All other analyses used data weighted to reflect a nationally repre- The outcome variable was a positive test result for C trachoma- sentative sample using survey estimation commands in com- tis, N gonorrhoeae,orT vaginalis on wave III urine specimens. puter software (Stata)23 that account for the stratification, clus- tering, and unequal probabilities of selection in the study design.20 Independent Variables We examined bivariate relationships between sociodemo- graphic characteristics, wave I independent variables (family, Demographic Variables. Age was categorized as 18 to 20, 21, school, and individual), and having a positive STI test result at 22, 23, or 24 years and older. All participants who reported that wave III. We conducted multivariate analyses to determine the they were of Hispanic or Latino origin were coded as Latino. ability of family, school, and individual variables to predict STI All other participants were coded as white, African American status, while controlling for sex, age, race/ethnicity, family struc- or black, Native American, or Asian American. Family structure, and level of maternal education. Because many factors in- ture during adolescence was measured by self-report at wave fluencing sexual behaviors and risk of STIs may differ by sex, I, and coded as living in a 2-parent home (biological or non- we assessed separate multivariate models for female and male biological) or not. Highest level of maternal education was mea- adolescents. sured by the mother’s report if she was interviewed at wave I In bivariate analyses, the range of each scaled variable was or by the adolescent’s report at wave I, and coded as less than divided into thirds. In multivariate regression analyses, these high school, high school graduate, or college or beyond. variables and age, mean school attendance, and GPA were mod- eled as continuous variables after confirmation that the as- Wave I Family Variables. Previously described measures of fam- sumption of linearity was met. Multivariate models included ily connectedness (Cronbach ␣=.88), perceived parental dis- only observations with no missing values for any included vari- approval of adolescent sex (Cronbach ␣=.90), and perceived able. For each multivariate model, correlations between inde-

(REPRINTED) ARCH PEDIATR ADOLESC MED/ VOL 159, JULY 2005 WWW.ARCHPEDIATRICS.COM 658

©2005 American Medical Association. All rights reserved. Downloaded From: https://jamanetwork.com/ on 09/26/2021 Table 1. Wave III Sample Characteristics and Proportions With a Positive Test Result for an STI*

Characteristic No. of Participants† Weighted %‡ % Positive for an STI§ Unadjusted OR (95% CI) Biological sex Male 6759 50.8 5.3 Referent Female 7563 49.2 7.2 1.39 (1.15-1.69) Age, y 18-20 3295 28.6 5.9 Referent 21 2281 16.5 7.2 1.23 (0.86-1.77) 22 2731 16.6 6.8 1.17 (0.83-1.65) 23 2789 15.4 6.0 1.00 (0.67-1.51) 24-28 3226 22.8 5.7 0.96 (0.62-1.47) Race/ethnicity White 7741 67.6 3.2 Referent African American 3042 16.0 18.6 6.99 (5.38-9.09) Latino 2340 11.9 7.8 2.58 (1.90-3.51) Native American 136 0.8 12.3 4.29 (2.22-8.30) Asian/Pacific Islander 1026 3.7 4.0 1.29 (0.85-1.95) Family structure in adolescence 2 Parents at home 10 213 72.2 5.0 0.51 (0.41-0.62) All others 4109 27.8 9.5 Referent Mother’s education ϽHS graduate 2371 16.2 10.2 2.24 (1.76-2.85) HS graduate 4262 33.4 6.4 1.35 (1.08-1.70) ՆCollege 7338 50.4 4.8 Referent

Abbreviations: CI, confidence interval; HS, high school; OR, odds ratio; STI, sexually transmitted infection. *Overall, 6.2% of the participants had a positive test result for an STI (positive for Chlamydia trachomatis, Neisseria gonorrhoeae,orTrichomonas vaginalis). †This number may not sum to the total number of participants (N=14 322) because of missing data. ‡Reflects the representative proportion in the target US population. Percentages may not total 100 because of rounding. §Analyses using weighted data and taking into account the stratified sampling design to provide national estimates of proportions who test positive within each group.

pendent variables were calculated to ensure the absence of WAVE III STI STATUS strong colinearity. The 2 most strongly correlated variables were perceived parental disapproval of sex and contraception (␳=0.57 in the full model). All other correlations were Overall, 6.2% (95% confidence interval, 5.3%-7.2%) of below 0.40. Results did not change substantively when per- wave III respondents had a positive test result for C tra- ceived parental disapproval of contraception was excluded chomatis, N gonorrhoeae,orT vaginalis. from final models. The prevalence of a positive test result for C trachomatis, N gonorrhoeae,orT vaginalis was slightly higher RESULTS among young adult females than males later (Table 1). The prevalence was highest among African Americans and Native Americans, intermediate among Latinos, and low- STUDY POPULATION est among Asian/Pacific Islanders and whites. The prevalence was higher among young adults with mothers who Of the 18 924 Add Health participants in the nationally did not graduate from high school and those with moth- representative weighted wave I sample, 14 322 (75.7%) ers who had only completed high school when com- were located and agreed to participate in wave III. Of these pared with those with mothers who had some college edu- participants, 1130 (7.9%) refused to provide a urine speci- cation or a college degree. Young adults who lived in a men and 226 (1.6%) were unable to provide a specimen 2-parent home during adolescence had a lower preva- at the interview. Furthermore, 421 (2.9%) of the speci- lence of positive test results than those who did not live mens could not be processed due to shipping or labora- in 2-parent homes. The prevalence of positive test re- tory problems and 951 (6.6%) did not have results for sults did not significantly vary by age. all 3 STI tests.21 In sum, test results were available for C trachomatis, N gonorrhoeae, and T vaginalis for 11 594 (81.0%) of the wave III participants. WAVE I FACTORS THAT LONGITUDINALLY Including participants who did and did not undergo PREDICT STI STATUS IN urine testing, just more than half (52.8%) of the study BIVARIATE ANALYSES sample was female). Most participants (54.2%) were white, with substantial representation of African Americans In bivariate analyses, adolescents’ perceptions of their par- (21.3%), Latinos (16.4%), Asian Americans (7.2%), and ents’ approval of sex predicted their likelihood of hav- Native Americans (1.0%) (percentages do not total 100 ing STIs 6 years later (Table 2). Of the participants who because of rounding). Participants’ mean age was 22.0 perceived that their parents strongly disapproved of sex years (SD, 1.8 years). during adolescence, only 5.5% later tested positive com-

(REPRINTED) ARCH PEDIATR ADOLESC MED/ VOL 159, JULY 2005 WWW.ARCHPEDIATRICS.COM 659

©2005 American Medical Association. All rights reserved. Downloaded From: https://jamanetwork.com/ on 09/26/2021 Table 2. Ability of Wave I Family, School, and Individual Factors to Longitudinally Predict Infection With Chlamydia trachomatis, Neisseria gonorrhoeae,orTrichomonas vaginalis

OR (95% CI) No. of % Positive Wave I Characteristic Participants* Weighted %† for an STI‡ Unadjusted Adjusted§ Family Variables Feelings of connectedness to family Low 212 1.4 6.2 1.02 (0.47-2.23) 1.25 (0.78-2.00) Moderate 2689 17.9 7.1 1.18 (0.93-1.51) 1.11 (0.89-1.39) High 11 114 80.8 6.1 Referent Referent Family suicide attempt or completion No 13 562 95.6 6.2 Referent Referent Yes 620 4.4 7.1 1.15 (0.74-1.80) 0.98 (0.59-1.63) Perceived parental disapproval of sex Low 431 3.2 8.9 1.67 (1.07-2.60) 1.42 (1.05-1.93) Moderate 3221 23.1 8.0 1.49 (1.20-1.84) 1.22 (1.03-1.46) High 9739 73.8 5.5 Referent Referent Perceived parental disapproval of contraception Low 4570 33.7 7.7 1.49 (1.13-1.97) 1.10 (0.82-1.47) Moderate 3099 21.9 5.9 1.13 (0.84-1.52) 1.05 (0.90-1.22) High 5661 44.4 5.3 Referent Referent School Variables Feelings of connectedness to school Low 501 3.8 7.4 1.37 (0.83-2.24) 0.96 (0.63-1.46) Moderate 6941 49.6 6.6 1.19 (0.96-1.48) 0.98 (0.81-1.19) High 6281 46.6 5.6 Referent Referent Schools’ mean daily attendance, % 74-79 4558 37.9 4.8 Referent Referent 80-84 6202 45.7 5.6 1.18 (0.80-1.73) 1.13 (0.99-1.28) 85-89 2680 10.1 10.4 2.29 (1.47-3.56) 1.27 (0.97-1.65) 90-94 483 4.3 16.7 3.98 (2.39-6.61) 1.42 (0.96-2.11) 95-100 211 1.9 8.1 1.74 (0.97-3.09) 1.60 (0.95-2.71) Parochial school No 13 405 94.8 6.4 Referent Referent Yes 729 5.2 3.9 0.59 (0.41-0.86) 0.96 (0.64-1.46) Individual Variables Religious importance Low 2482 18.6 5.8 0.87 (0.65-1.18) 1.09 (0.84-1.42) Moderate 2433 18.7 6.0 0.91 (0.68-1.22) 1.04 (0.92-1.17) High 9165 62.7 6.6 Referent Referent Same-sex attraction or relationship No 13 170 92.4 6.2 Referent Referent Yes 1033 7.6 6.7 1.10 (0.79-1.52) 1.10 (0.74-1.65) Perceived likelihood of living to the age of 35 y Low 492 3.4 4.9 0.84 (0.47-1.49) 0.90 (0.63-1.29) Moderate 1499 10.2 10.6 1.92 (1.49-2.49) 0.94 (0.77-1.15) High 12 249 86.4 5.8 Referent Referent Paid work for Ն20 h/wk No 11 579 80.9 6.7 Referent Referent Yes 2615 19.1 4.5 0.66 (0.50-0.88) 0.78 (0.56-1.08) Perceives that one looks older than most peers No 12 316 86.4 6.5 Referent Referent Yes 1815 13.6 5.3 0.80 (0.59-1.09) 0.85 (0.60-1.21)

(continued)

pared with 8.0% and 8.9% of adolescents who perceived also predicted STI status 6 years later. Among adoles- that their parents had moderate and low disapproval of cents with a GPA between 3.25 and 4.0, 4.2% later sex, respectively. Findings related to perceived parental tested positive for STIs. In comparison, 7.2% of adoles- disapproval of contraception were similar. cents with a GPA between 2.25 and 3.0 and 7.8% of In general, adolescents who attended schools charac- adolescents with a GPA between 1.0 and 2.0 acquired terized as having higher mean daily attendance were STIs. Adolescents who worked 20 hours per week or more likely to acquire STIs than comparison groups. more were at lower risk of STIs 6 years later than those Adolescents who attended parochial schools were less who did not, and adolescents who perceived looking likely to acquire STIs than those who attended nonpa- younger than most peers were at higher risk of STIs rochial schools. School performance during adolescence than those who did not.

(REPRINTED) ARCH PEDIATR ADOLESC MED/ VOL 159, JULY 2005 WWW.ARCHPEDIATRICS.COM 660

OR (95% CI) No. of % Positive Wave I Characteristic Participants* Weighted %† for an STI‡ Unadjusted Adjusted§ Perceives that one looks younger than most peers No 12 699 90.3 6.1 Referent Referent Yes 1432 9.7 8.4 1.42 (1.06-1.90) 1.04 (0.76-1.42) Virginity pledge No 12 245 87.2 6.4 Referent Referent Yes 1919 12.8 5.1 0.79 (0.56-1.11) 0.85 (0.60-1.19) Grade point average 1.0-2.0 2932 20.4 7.8 1.93 (1.45-2.57) 1.41 (1.03-1.93) 2.25-3.0 5936 42.1 7.2 1.77 (1.38-2.27) 1.18 (1.01-1.38) 3.25-4.0 4906 37.6 4.2 Referent Referent

Abbreviations: See Table 1. *This number may not sum to the total number of participants (N=14 322) because of missing data. †Reflects the representative proportion in the target US population. Percentages may not total 100 because of rounding. ‡Analyses using weighted data and taking into account the stratified sampling design to provide national estimates of proportions who test positive within each group. §Adjusted for biological sex, age, race/ethnicity, family structure, mother’s education, and all other variables in the table. Family and school connectedness, perceived parental disapproval of sex and contraception, religious importance, perceived likelihood of early death, age, mean school attendance, and grade point average were entered into the multivariate model as continuous variables after preliminary analyses showed no significant violation of the assumption of linearity in the log odds for any of these variables; however, for ease of presentation in the table, categorical variables were constructed from these continuous variables to report adjusted ORs comparing each category with the reference category.

WAVE I FACTORS THAT ity. This study focuses on determining whether modifi- LONGITUDINALLY PREDICT STI STATUS able family, school, and individual factors associated with IN MULTIVARIATE ANALYSES increased duration of virginity during adolescence also protect against having STIs 6 years later. Adjusting for biological sex, age, race/ethnicity, family Adolescents who perceive that their parents more structure, and maternal education, adolescents who per- strongly disapprove of their having sexual intercourse as ceived that their parents more strongly disapproved of an adolescent are less likely to be infected with C tra- their having sex during adolescence were less likely to chomatis, N gonorrhoeae,orT vaginalis 6 years later. To have C trachomatis, N gonorrhoeae,orT vaginalis 6 years our knowledge, this is the first report of a longitudinal later (adjusted odds ratio [95% confidence interval], 0.89 link between perceptions of parental opinions about sex [0.81-0.99]) (Table 2). Adolescents with a higher GPA during adolescence and biologically measured STIs. This during adolescence were less likely to have these STIs 6 finding is consistent with previous research showing that years later (adjusted odds ratio [95% confidence inter- the more disapproving adolescents perceive their moth- val], 0.84 [0.71-0.99]). Other wave I family, school, and ers to be toward their engagement in sexual inter- individual variables were not associated with wave III STI course, the less likely adolescents are to become preg- status in multivariate analyses. nant24 and to self-report STI diagnoses.25 From a practical perspective, if parents who disap- ABILITY OF FACTORS TO LONGITUDINALLY prove of their adolescent children engaging in sexual in- PREDICT STI STATUS AMONG FEMALE tercourse convey this attitude effectively, then the long- VS MALE ADOLESCENTS term risk of STIs may be reduced. Unfortunately, effective conveyance of parental attitudes about sexual inter- In stratified multivariate analyses, female adolescents course seems to be challenging. Adolescents often have who perceived that their parents were more disapprov- inaccurate perceptions of their parents’ attitudes about ing of their having sex during adolescence had a lower their engaging in sexual intercourse, and the tendency risk of STIs 6 years later (Table 3). Female adolescents is for adolescents to underestimate parental disap- with a higher GPA had a reduced likelihood of later proval.24,26,27 Furthermore, the correspondence between STIs. Among male adolescents, family, school, and indi- parent and child reports of whether they have discussed vidual factors during adolescence did not predict STI sexual topics is often low, and many parents underesti- status 6 years later. mate their adolescent children’s engagement in sexual activity.26,28,29 Taken together, this suggests a need to iden- COMMENT tify strategies to increase the congruence between adolescents’ perceptions of parental attitudes about sexual We confirm that STIs are common among young adults intercourse during adolescence and actual parental atti- in the United States, and that prevalence varies substan- tudes. Strategies will need to take into account the com- tially by nonmodifiable factors such as race and ethnic- plex process of conveying parental attitudes to chil-

(REPRINTED) ARCH PEDIATR ADOLESC MED/ VOL 159, JULY 2005 WWW.ARCHPEDIATRICS.COM 661

©2005 American Medical Association. All rights reserved. Downloaded From: https://jamanetwork.com/ on 09/26/2021 parents, with adolescent boys being more likely to be en- Table 3. Ability of Family, School, and Individual Factors couraged to be independent and adolescent girls more to Longitudinally Predict Acquisition of an STI, Stratified likely to be encouraged to be obedient. Borawski et al33 by Biological Sex* also found that adolescent girls report less negotiated unsupervised time than adolescent boys. The link between Female Male Wave I Characteristic Adolescents Adolescents perceived parental disapproval of sex and lower longitudinal risk of STIs among adolescent girls may be, at least Family Variables Feeling connected to family 0.94 (0.80-1.11) 0.93 (0.74-1.16) in part, the result of internalization of more conven- Family suicide attempt or 0.78 (0.44-1.40) 1.37 (0.57-3.34) tional standards or a stronger desire among adolescent completion girls to behave in ways that will not result in parental dis- Perceives that parent 0.84 (0.74-0.94) 0.94 (0.80-1.11) approval. disapproves of sex We also found sex differences in the protective rela- Perceives that parent 0.97 (0.81-1.16) 0.96 (0.78-1.18) disapproves of contraception tionship between higher GPA and lower longitudinal STI risk. Previous research30 has shown that adolescents with School Variables higher scores on a standardized measure for vocabulary Feeling connected to school 1.03 (0.90-1.18) 1.00 (0.81-1.24) Schools’ mean daily attendance 1.09 (0.92-1.29) 1.18 (0.97-1.42) are less likely to engage in vaginal intercourse and other Parochial school 1.13 (0.71-1.81) 0.80 (0.49-1.32) noncoital sexual behaviors than those with lower scores, and that this relationship is stronger among adolescent Individual Variables Religious importance 0.93 (0.79-1.08) 1.00 (0.85-1.18) girls than adolescent boys. Our findings are generally con- Same-sex attraction or 1.00 (0.60-1.67) 1.17 (0.63-2.17) sistent with this observation. However, GPA is influ- relationship enced by many factors other than verbal ability, and it is Perceived likelihood of living to 1.02 (0.88-1.17) 1.06 (0.92-1.23) possible that higher grades, especially for adolescent girls, the age of 35 y may be another manifestation of internalization of con- Paid work for Ն20 h/wk 0.67 (0.44-1.02) 0.86 (0.53-1.39) Perceives that one looks older 0.83 (0.54-1.29) 0.86 (0.49-1.50) ventional standards and a desire to avoid parental dis- than most peers approval—which in the realm of sexual development may Perceives that one looks younger 1.39 (0.96-2.01) 0.74 (0.43-1.29) be linked to less risky sexual behaviors and partner than most peers choices. Further research will clearly be needed to bet- Virginity pledge 0.83 (0.56-1.22) 0.87 (0.47-1.59) ter understand predictors of longitudinal risk of STIs for Grade point average 0.82 (0.66-1.00) 0.85 (0.66-1.09) adolescent girls and boys, and how and why these predictors may differ by sex. Abbreviation: See Table 1. *Data are given as adjusted odds ratio (95% confidence interval). The Overall, our negative findings are perhaps the most odds ratios were adjusted for age, race, ethnicity, family structure, mother’s interesting. Family, school, and individual factors linked education, and all other variables in the table. Acquisition of an STI was to delayed sexual debut among adolescents16 in large part based on a positive test result for Chlamydia trachomatis, Neisseria gonorrhoeae,orTrichomonas vaginalis. did not predict STI status 6 years later. Perhaps the cor- relation between early sexual debut and risky sexual behaviors (and, therefore, risk of STIs)12-15 weakens as adolescents approach adulthood. Furthermore, it may be dren, which likely involves nonverbal and verbal unrealistic to expect broad protective models based on communication. the adolescent resiliency literature to longitudinally pre- We found that adolescents with higher grades in school dict complex health outcomes. Future research will be are less likely to have acquired STIs 6 years later than needed to better understand our negative findings. those with lower grades. This finding is consistent with There were several limitations of this study. First, we previous research showing that adolescents with higher did not investigate the process by which family, school, grades are less likely to participate in high-risk behav- and individual factors influence risk of later STIs, in- iors across many domains,16 and that adolescents with cluding the potential role of sexual behaviors. Second, higher intelligence are less likely to engage in risk- the original Add Health sample was school based, so drop- associated sexual behaviors.30 We cannot exclude bidi- outs were excluded. However, Add Health has followed rectional relationships between academic achievement up all originally enrolled participants, including drop- and long-term risk of STIs. Others31 have reported bidi- outs, and previous research suggests that potential bias rectional influences between academic achievement and because of missing out-of-school youth in Add Health is age of sexual debut, with variation in these relation- small.34 Third, 24.3% of the original wave I sample could ships among adolescent boys and girls. not be located at wave III, and results of all 3 STI tests Our results show that the longitudinal risk of STIs is were not available for 19.0% of wave III respondents. How- influenced by perceptions of parental disapproval of sex ever, careful assessment has shown that risk of wave III for adolescent girls, but not adolescent boys. The link be- nonresponse bias seems minimal,35 and comparisons of tween parental disapproval of sex and lower likelihood wave III participants who did and did not provide urine of STIs in late adolescence likely works through influ- specimens for STI testing have shown few differences.36 ences on sexual behaviors (eg, fewer partners, fewer in- Fourth, the performances of STI tests on urine speci- cidents of sexual activity, or more consistent condom use) mens are not perfect,21 which influences the accuracy of or some aspect of partner choice (eg, monogamous re- our main outcome variable. We are reassured by previ- lationship, marriage, or “safer” partners). There is lit- ous analyses quantifying the effect of nonresponse and erature32 documenting sex-differentiated socialization by test performance on prevalence estimates that show the

(REPRINTED) ARCH PEDIATR ADOLESC MED/ VOL 159, JULY 2005 WWW.ARCHPEDIATRICS.COM 662

©2005 American Medical Association. All rights reserved. Downloaded From: https://jamanetwork.com/ on 09/26/2021 risk of this potential bias seems small.3 Fifth, if wave III Entwisle, PhD, for their assistance in the original Add positive STI test results represent infections that have per- Health design. sisted since wave I, we would overestimate the influence of our predictive model. However, it is unlikely that REFERENCES these STIs would persist for 6 years because of the pro- 37 pensity of infections to resolve spontaneously or be ef- 1. Institute of Medicine Committee on Prevention and Control of Sexually Trans- 38 fectively treated with antibiotics, which may be given mitted Diseases. The Hidden Epidemic: Confronting Sexually Transmitted Diseases. for other reasons. Of more concern, our STI outcomes Washington, DC: National Academy Press; 1997. most likely reflect recent sexual behaviors; no serologic 2. Chesson HW, Blandford JM, Gift TL, Tao G, Irwin KL. The estimated direct medi- studies were available to test for antibody responses to cal cost of sexually transmitted diseases among American youth, 2000. Per- spect Sex Reprod Health. 2004;36:11-19. STIs, which would more closely approximate cumula- 3. Miller WC, Ford CA, Morris M, et al. Prevalence of chlamydial and gonococcal tive risk of infection. Sixth, the reliability and validity of infection among young adults in the United States. JAMA. 2004;291:2229- most independent variables have not been evaluated and 2236. measures included few items (eg, perceived parental dis- 4. Ellen J, Aral S, Madger L. Do differences in sexual behaviors account for the racial/ ethnic differences in adolescents’ self-reported history of a sexually transmitted approval of sex). Seventh, our independent measures may disease? Sex Transm Dis. 1998;25:125-129. not have captured important dimensions of family and 5. Ford K, Sohn W, Lepkowski J. Characteristics of adolescents’ sexual partners school context during adolescence that may influence later and their association with use of condoms and other contraceptive methods. Fam risk of STIs (eg, familial patterns of health care use), lim- Plann Perspect. 2001;33:100-105. iting the predictive ability of our model. 6. Ford K, Sohn W, Lepkowski J. American adolescents: sexual mixing patterns, bridge partners, and concurrency. Sex Transm Dis. 2002;29:13-19. In summary, our results show that adolescents’ aca- 7. Aral S. Understanding racial-ethnic and societal differentials in STI. Sex Transm demic success and perceptions of parental disapproval Infect. 2002;78:2-4. of sexual intercourse diminish the risk of acquiring STIs 8. Adimora AA, Schoenbach VJ, Martinson FE, et al. Social context of sexual rela- in young adulthood, particularly for females. Further tionships among rural African Americans. Sex Transm Dis. 2001;28:69-76. 9. Kahn JA, Rosenthal SL, Succop PA, Ho GY, Burk RD. Mediators of the associa- research will be needed to understand how these factors tion between age of first sexual intercourse and subsequent human papilloma- diminish STI risk, and to identify other strategies to virus infection. Pediatrics [serial online]. 2002;109:E5. Available at: http:// reduce STI prevalence among adolescents and young pediatrics.aappublications.org.libproxy.lib.unc.edu/cgi/content/full/109/1/e5. adults. Parents should be encouraged to convey clear Accessed April 29, 2005. messages that they do not approve of their children 10. Moss GB, Clementson D, D’Costa L, et al. Association of cervical ectopy with heterosexual transmission of human immunodeficiency virus: results of a study having sexual intercourse during adolescence. Finally, of couples in Nairobi, Kenya. J Infect Dis. 1991;164:588-591. prevention and intervention programs will need to 11. Harrison HR, Costin M, Meder JB, et al. Cervical Chlamydia trachomatis infec- account for the complex phenomena that contribute to tion in university women: relationship to history, contraception, ectopy, and health problems faced by adolescents and young cervicitis. Am J Obstet Gynecol. 1985;153:244-251. 12. Durbin M, DiClemente R, Siegel D, Krasnovsky F, Lazarus N, Camacho T. Fac- adults, including STIs and the human immunodefi- tors associated with multiple sex partners among junior high school students. ciency virus. J Adolesc Health. 1993;14:202-207. 13. Coker A, Richter D, Valois RF, McKeown R, Garrison C, Vincent M. Correlates Accepted for Publication: February 14, 2005. and consequences of early initiation of sexual intercourse. J Sch Health. 1994; Correspondence: Carol A. Ford, MD, Adolescent Medi- 64:372-377. 14. Santelli J, Brener N, Lowry R, Bhatt A, Zabin L. Multiple sexual partners among cine Program, Campus Box 7220, The University of North U.S. adolescents and young adults. Fam Plann Perspect. 1998;30:271-275. Carolina, Chapel Hill, NC 27599-7220 ([email protected] 15. O’Donnell BL, O’Donnell CR, Stueve A. Early sexual initiation and subsequent sex- .edu). related risks among urban minority youth: the reach for health study. Fam Plann Funding/Support: This study was supported in part by Perspect. 2001;33:268-275. 16. Resnick MD, Bearman PS, Blum RW, et al. Protecting adolescents from harm: grant UO131496 from the National Institute of Allergy findings from the National Longitudinal Study of Adolescent Health. JAMA. 1997; and Infectious Diseases, Bethesda, Md (STD Coopera- 278:823-832. tive Research Center at The University of North Caro- 17. Resnick MD. Protective factors, resiliency and healthy youth development. Ado- lina at Chapel Hill); grant HD38210 from the National lesc Med. 2000;11:157-164. Institutes of Health, Bethesda; and The Robert Wood 18. The National Longitudinal Study of Adolescent Health. Carolina Population Cen- ter, The University of North Carolina at Chapel Hill Web site. Available at: http: Johnson Foundation Generalist Physician Faculty Schol- //www.cpc.unc.edu/addhealth. Accessed March 3, 2004. ars Program. Mr Pence is a Howard Hughes Medical In- 19. Bearman PS, Jones J, Udry JR. The National Longitudinal Study of Adolescent stitute Predoctoral Fellow. Health: research design: 1997. Carolina Population Center, The University of North Additional Information: This research uses data from Add Carolina at Chapel Hill Web site. Available at: http://www.cpc.unc.edu/projects /addhealth/design.html. Accessed March 3, 2004. Health, a program project designed by J. Richard Udry, 20. Chantala K, Tabor J. Strategies to perform a designed-based analysis using the PhD, Peter S. Bearman, PhD, and Kathleen Mullan Har- Add Health data: 1999. Carolina Population Center, The University of North Caro- ris, PhD, and funded by grant P01-HD31921 from the lina at Chapel Hill Web site. Available at: http://www.cpc.unc.edu/projects National Institute of Child Health & Human Develop- /addhealth/files/weight1.pdf. Accessed September 25, 2004. ment, Bethesda, with cooperative funding from 17 other 21. Cohen MS, Florey FA, Ford CA, et al. Biomarkers in wave III of the Add Health study: 2003. Carolina Population Center, The University of North Carolina at Chapel agencies. Persons interested in obtaining data files from Hill Web site. Available at: http://www.cpc.unc.edu/projects/addhealth/files Add Health should contact Add Health, Carolina Popu- /biomark.pdf. Accessed September 1, 2004. lation Center, 123 W Franklin St, Chapel Hill, NC 27516- 22. Sieving RE, Beuhring T, Resnick MD, et al. Development of adolescent self- 2524 (http://www.cpc.unc.edu/addhealth). report measures from the National Longitudinal Study of Adolescent Health. J Adolesc Health. 2001;28:73-81. Acknowledgment: We thank Carolyn Tucker Halpern, 23. Stata Corp. Intercooled Stata 7.0. College Station, Tex: Stata Corp; 2001. PhD, for her contributions to our interpretation of sex 24. Dittus PJ, Jaccard J. Adolescents’ perceptions of maternal disapproval of sex: differences; and Ronald R. Rindfuss, PhD, and Barbara relationship to sexual outcomes. J Adolesc Health. 2000;26:268-278.

(REPRINTED) ARCH PEDIATR ADOLESC MED/ VOL 159, JULY 2005 WWW.ARCHPEDIATRICS.COM 663

©2005 American Medical Association. All rights reserved. Downloaded From: https://jamanetwork.com/ on 09/26/2021 25. Crosby R, Leichliter JS, Brackbill R. Longitudinal prediction of sexually trans- 33. Borawski EA, Ievers-Landis CE, Lovegreen LD, Trapel ES. Parental monitoring, mitted diseases among adolescents: results from a national survey. AmJPrev negotiated unsupervised time, and parental trust: the role of perceived parent- Med. 2000;18:312-317. ing practices in adolescent risk behaviors. J Adolesc Health. 2003;33:60-70. 26. Jaccard J, Dittus PJ, Gordon VV. Parent-adolescent congruency in reports of ado- 34. Udry JR, Chantala K. Missing school dropouts in surveys does not bias risk lescent sexual behavior and in communication about sexual behavior. Child Dev. estimates. Soc Sci Res. 2003;32:294-311. 1998;69:247-261. 35. Chantala K, Kalsbeek WD, Andraca E. Non-response in wave III of the Add Health 27. Sieving R, McNeely C, Blum R. Maternal expectations, mother-child connected- Study: 2004. Carolina Population Center, The University of North Carolina at Chapel ness, and adolescent sexual debut. Arch Pediatr Adolesc Med. 2000;154:809- Hill Web site. Available at: http://www.cpc.unc.edu/projects/addhealth/files 816. /W3nonres.pdf. Accessed September 1, 2004. 28. Newcomer SF, Udry JR. Parent-child communication and adolescent sexual 36. Ford CA, Jaccard J, Millstein SG, Miller WC. Who refuses to provide a urine speci- behavior. Fam Plann Perspect. 1985;17:169-174. men for Chlamydia trachomatis and Neisseria gonorrhoeae testing [abstract]? 29. Young T, Zimmerman R. Clueless: parental knowledge of risk behaviors of middle J Adolesc Health. 2004;34:144. school students. Arch Pediatr Adolesc Med. 1998;152:1137-1139. 30. Halpern CT, Joyner K, Udry J, Suchindran C. Smart teens don’t have sex (or kiss 37. Parks KS, Dixon PB, Richey CM, Hook EW. Spontaneous clearance of Chla- much either). J Adolesc Health. 2000;26:213-225. mydia trachomatis infection in untreated patients. Sex Transm Dis. 1997;24: 31. Schvaneveldt P, Miller B, Berry E, Lee T. Academic goals, achievement, and age 229-235. at first sexual intercourse: longitudinal, bidirectional influences. Adolescence. 2001; 38. Bachmann LH, Stephens J, Richey CM, Hook EW. Measured versus self- 36:767-787. reported compliance with doxycycline therapy for Chlamydia-associated syn- 32. Block JH. Differential premises arising from differential socialization of the sexes: dromes: high therapeutic success rates despite poor compliance. Sex Transm some conjectures. Child Dev. 1983;54:1335-1354. Dis. 1999;26:272-278.

Announcement

Topic Collections. Archives offers collections of articles in specific topic areas to make it easier for physi- cians to find the most recent publications in a field. These are available by subspecialty, study type, disease, or prob- lem. In addition, you can sign up to receive a Collection E-Mail Alert when new articles on specific topics are pub- lished. Go to http://archpedi.ama-assn.org/collections to see these collections of articles.

(REPRINTED) ARCH PEDIATR ADOLESC MED/ VOL 159, JULY 2005 WWW.ARCHPEDIATRICS.COM 664