HEALTH EDUCATION RESEARCH
Vol.32 no.3 2017 Pages 233–243 Advance Access published 3 May 2017
Randomized controlled trial of abstinence and safer sex intervention for adolescents in Singapore: 6-month follow-up Mee Lian Wong1*, Junice Y. S. Ng1, Roy K. W. Chan1,2, Martin T. W. Chio2, Raymond B. T. Lim1 and David Koh1,3 1
Saw Swee Hock School of Public Health, National University of Singapore, Singapore, Singapore, 2Department of STI Control, National Skin Centre, Singapore and 3PAPRSB Institute of Health Sciences, Universiti Brunei Darussalam, Gadong, Brunei *Correspondence to: Mee Lian Wong. E-mail: [email protected]
Received on June 1, 2016; editorial decision on February 18, 2017; accepted on April 25, 2017
Abstract We assessed the efficacy of an individual-based behavioral intervention on sexually transmitted infections’ (STI) risk-reduction behaviors in Singapore. A randomized controlled trial of a behavioral intervention compared to usual care was conducted on sexually active heterosexual adolescents aged 16–19 years attending the only public STI clinic. The intervention included two on-site skills-based sessions targeting individual, relational and environmental influences on sexual behaviors, followed by online support. Participants were assessed at baseline and 6-month follow-up. Primary outcomes were self-reported abstinence, number of partners and consistent condom use for vaginal sex. We recruited 337 adolescents to the intervention and 351 to usual care (controls). Fifty-nine percent of intervention participants and 53% of controls completed follow-up. Young men [adjusted risk ratio (RR) 2.03; 95% CI, 1.25–3.30], but not young women, in the intervention were more likely than controls to report secondary abstinence. More non-abstinent young women in the intervention than controls kept to one partner (adjusted RR, 1.25; 95% CI, 1.04– 1.50) compared to no differences in young men. There was no intervention effect on consistent condom use in both genders. Skill-based intervention can promote abstinence in young men
and keeping to one partner in young women in a clinic setting.
Introduction Adolescents are experiencing the most rapid increase in sexually transmitted infections (STIs) worldwide . Safer sexual practices have been shown to reduce STIs among them . Intervening early during adolescence has also been found to influence their sexual behaviors positively during adulthood . Many interventions for adolescents in Western countries [3–14] found mixed results ranging from an impact on condom use and partner reduction, [3, 6, 7, 9, 12] and STIs [3, 7, 9] to little or no impact [5, 8, 11, 13]. More than 50% of the adolescents in the world live in Asia . Recent surveys in Thailand , Taiwan, China and Vietnam  reported an increase in sexual initiation and unprotected intercourse among them than adults. Presently, South and Southeast Asia report the largest number of new STIs in the world. We need to develop and evaluate interventions to reduce adolescents’ risky sexual behaviors in this region. Data are lacking on such trials in Asia probably because of the cultural sensitivity of the topic, practical issues and infeasibility in accessing sexually active adolescents in the region. The few published trials that were
ß The Author 2017. Published by Oxford University Press. All rights reserved. For permissions, please email: [email protected]
M. L. Wong et al. conducted, mainly in schools in the Philippines  and Thailand [19, 20], found no impact, except for a reduction in the frequency of intercourse in the latter . The study in the Philippines  was a cluster randomized controlled trial (RCT) conducted on 845 high school students to evaluate a comprehensive theory-based AIDS prevention program aimed at providing them with accurate information about AIDS, fostering positive attitudes towards HIV-infected persons and developing their skills in sexual risk reduction. The 12-lesson intervention program, developed by public high school teachers together with local health experts, social scientists and health educators was successful in increasing AIDS-related knowledge. However, there was no impact on intended preventive behaviors such as intention to use condoms, plans to have only one sex partner and plans to refuse sex under the influence of alcohol. The authors attributed the ineffectiveness of the intervention in changing intended preventive behaviors to the lack of comfort and skills of teachers in delivering the lessons, insufficient student participation and lack of attention on skills training to resist peer pressure. This study did not assess change in sexual behaviors such as abstinence and condom use. In Thailand, Thato et al.  developed and assessed the impact of a comprehensive culturallysensitive theory-based sex education program on a convenience sample of 261 adolescent school children aged 13–18 years in the intervention and comparison group respectively. The intervention led to lower levels of self-reported sexual intercourse and delay in sexual initiation but there was no impact on condom use. This study might be limited by the small sample size because less than 11% of the intervention participants were sexually active, In addition, this group of high school adolescents might not represent the priority at-risk group given that only a small proportion of them were sexually active. Only one trial was conducted in a clinic setting where the target group was perinatally infected HIV-infected adolescents receiving HIV care at two hospitals in Thailand . The intervention focused on increasing their HIV knowledge, developing their coping skills, reducing their sexual risk behaviors and helping them to achieve their life goals. 234
A small number of 32 participants out of the 197 participants who did not attend any of the 4 intervention sessions were categorized as non-intervention (control) participants. The intervention led to an increase in HIV knowledge and attitude scores at six months post-intervention but there was no impact on abstinence and condom use. This study might be affected by selection bias because participants were not randomized to the control or intervention group. Additionally, as only 5% of the 197 participants were sexually active at baseline, the study might not have the statistical power to detect a change, if any, in sexual risk reduction. We need to arrest the rise in risky sexual behaviors among adolescents in Asia urgently by implementing interventions beyond schools to STI clinic settings in order to reach the core group of sexually active adolescents who are at high risk for acquiring and transmitting STIs. STI interventions for adolescents in clinic settings in the West [6, 7, 9] may not be transferable to adolescents in Asia. Most of these interventions focused on promoting condom use probably because of the ineffectiveness of promoting abstinence as the only  or main strategy [23, 24]. Promoting condom use alone without emphasizing abstinence may not be accepted in some Asian countries because of the concern that it may encourage premarital sex which is often viewed as immoral . We assessed the efficacy of an individual-based behavioral intervention to promote abstinence and safer sexual behaviors such as consistent condom use and partner reduction among sexually active heterosexual adolescents attending the only public STI clinic in Singapore. We hypothesized that the intervention would increase abstinence and condom use, and reduce the number of partners compared to the usual STI prevention program.
Materials and methods We conducted a RCT with 1:1 allocation ratio to compare participants receiving the new intervention with the usual care program (controls) at 6-month follow-up. Study participants were never married
Abstinence, safer sex intervention for adolescents sexually active heterosexual adolescents attending the only public Department of STI Control (DSC) clinic for screening or treatment of STIs in Singapore. Singapore has a small population of about 3.8 million citizens, with the majority (74.2%) being Chinese, 13.3% Malays and 9.1% Indians . About 80% of adolescents with reportable STIs attend this clinic. Eligibility criteria included being Singaporean citizen or permanent resident, aged 16–19, first-time attendees and being sexually active in the last six months. The assessors used a screening questionnaire to determine the adolescents’ eligibility for the study. The study was explained and adolescents were only assigned the study condition after they consented to participate in the study by signing the consent form (age of consent in Singapore: 16 years or older). As between 1 and 5 (median 1; interquartile range: 1–2) eligible adolescents attended the clinic daily, diffusion of intervention effects might occur due to them congregating and communicating with each other while waiting at the clinic. To reduce contamination bias, we randomized days  rather than individuals so that participants attending the clinic on the same day would be allocated to the same study group. The randomization sequence, using computerized non-deterministic algorithm, was prepared prior to enrolment of study participants. The study was approved by the Singapore National Healthcare Group Domain Specific Review Board: DSRB: E/09/073.
Intervention The intervention, consisting of 3 sessions, aimed to develop adolescents’ confidence and skills in negotiating and practicing abstinence and condom usage. We applied Green’s PRECEDE PROCEED framework , Bandura’s self-efficacy theory , findings from local studies on premarital intercourse  and condom use  to identify the most important modifiable factors to change adolescents’ risky sexual behaviors. We also obtained inputs from social workers, psychologists, youth workers and religious leaders in the development of the
intervention and continuous quality improvement principles were incorporated into the intervention to identify teething problems to continually improve the program. The counselors were social workers, nurses and preventive medicine doctors who were not staff of the STI clinic. They were trained on behavioral strategies and motivational interviewing by the principal investigator (PI: MLW). To maintain fidelity to the intervention, counselors received a training manual with a clearly defined protocol for each session. The PI also held regular meetings with the counsellors throughout the intervention period to discuss problems encountered. The chief counselor monitored the work of the counsellors and monthly progress reports were reviewed to ensure protocol adherence. The first session, held at the time of enrolment, focused on getting participants to self-reflect on the influences on their risk behaviors, and educating them on STIs/HIV using interactive flipcharts and videos depicting real life stories . They were also encouraged to discuss their views on love, sex and self-respect. The second session, held after two weeks, adopted motivational interviewing to help participants set goals on sexual abstinence or being faithful to one partner and consistent condom usage. They were trained in skills on negotiating for sexual abstinence and condom usage using role play; and on condom application using a model aid. Practical tips from their peers on how to abstain by using distraction strategies such as engaging in sports or fun activities, giving excuses and avoiding triggers to sex such as drinking were also shared with them. Tips on how to assert in a firm but nice way to say no to sex were also shared. The third session was an online chat or telephone session held 2 months later to provide further counseling and emotional support in boy girl relationships or other concerns raised by the adolescents. When smartphones became popular after 2010, we used the texting app ‘WhatsApp’ or SMS to counsel them depending on their preference. Outreach counseling was conducted for those (20%) who declined to return for follow-up at the clinic due to inconvenience or stigma, or were not able to attend the sessions during clinic opening hours. 235
M. L. Wong et al. The controls received the routine usual care from the health counselors at DSC. In the first session, all adolescents were counseled on STIs/HIV prevention. Adolescents diagnosed with STIs were treated and given a second counseling session two weeks later. Those who tested negative were notified of the test results through SMS and were not required to attend further counseling.
responding honestly because their responses would be used for program improvement. Frequency-based questions were used to assess sensitive topics. Instead of asking a leading question ‘Did you abstain from sex?’ we asked ‘How many times did you have vaginal sex?’
Consistent condom use for vaginal intercourse was used to compute sample size. Based on an estimated 1.7 fold relative increase in the proportion of consistent condom use in the intervention (34%) to the control group (20%) after 6 months post-intervention, a two-sided confidence level of 95% and power of 80%, a sample of 157 adolescents per group was required. As we estimated an attrition of 50% at 6-month follow-up, based on follow-up data in routine care, we computed the final sample size to be 314 per group.
Primary outcomes included consistent condom use, secondary abstinence and number of sexual partners in the last 6 months. Secondary abstinence for vaginal sex was defined as ever had sex before the intervention but not in the past 6 months post-intervention. It was assessed by the question ‘How many times did you have vaginal sex in the last 6 months?’ Those who responded ‘zero’ was defined as secondary abstinence. Consistent condom use was assessed by asking, ‘How often did you use a condom for vaginal sex in the last 6 months?’ Participants who responded ‘Yes, always’ were defined as consistent condom users, whereas those who responded ‘Yes, sometimes’, ‘No, not at all’ and ‘Can’t remember’ were defined as inconsistent condom users. The secondary outcome was incident acute STIs defined by laboratory tests for gonorrhea, chlamydia, trichomoniasis, syphilis and herpes simplex infection (Type II).
Data collection An independent group of trained assessors, who were blinded to the group assignment, provided both study groups with a printed self-administered questionnaire at baseline and 6-month follow-up. To avoid contamination bias at follow-up, the intervention and study groups were given different time appointments. To reduce social desirability bias, participants were assured that their responses were confidential and that code numbers rather than personal identifiers would be used on the questionnaires. We also acknowledged their difficulties in practicing safer sex and stressed the importance of 236
Statistical analyses To statistically determine the effect of the intervention, we compared change in outcomes between the control and intervention groups, adjusting for baseline differences. As preliminary analysis detected gender differences in outcomes, we also conducted gender-stratified analysis. To assess the effect size of the intervention, we used Poisson regression model to compute the crude and adjusted risk ratios (RR) of the intervention to the control group. Chi-square tests were used to compare categorical variables and Mann–Whitney U test for continuous variables. We used both complete case analysis, in which only participants with outcome data at 6-month follow-up were analyzed, and intention-to-treat analysis with imputed data for missing outcomes due to attrition to assess robustness of findings. For the intent-to-treat approach, we conducted sensitivity analyses using imputations to assess the potential effect of missing data. These include (i) worst-case scenario with all those who dropped out as having negative outcomes and (ii) appropriate multiple
Abstinence, safer sex intervention for adolescents imputation with ethnicity, residence, educational level, age, smoking, drinking alcohol and age of sex initiation imputed as covariates. Analyses were performed using Stata statistical software, version 14 .
Results Between November 2009 and October 2014, we screened 1068 adolescent patients, of which 688 (75.8%) out of 908 eligible adolescents participated in the study (Fig. 1). The participants and non-participants did not differ significantly in age (P ¼ 0.687), gender (P ¼ 0.103) and ethnicity (P ¼ 0.849). We recruited 337 adolescents to the intervention and 351 to the control group. Both
groups were comparable for baseline demographic characteristics, sexual behaviors and STI prevalence rates. However, significant gender differences were observed for consistent condom use in both study groups, with more young men than young women reporting consistent condom use at baseline (P < 0.001, Table I). More than half (59%) of the participants (93 young men, 105 young women; total: 198) in the intervention group and 53% of participants (89 young men, 98 young women; total: 187) in the control group completed the 6-month follow-up (Fig. 1) with no significant difference in attrition rates between groups (P ¼ 0.352). Of those lost to follow-up, 49.2% of intervention participants and 51.6% of controls were not contactable due to
Fig. 1. Progress of Study Participants through the Trial at the STI clinic in Singapore, November 2009–May 2015. *Only participants who have untreated acute STIs (that is, chlamydia, gonorrhea and syphilis) were required to return for 2-week follow-up for treatment and counseling.
238 75 (50.0) 75 (50.)
35 (23.6) 113 (76.4) 3 (2–6)
36 (23.8) 115 (76.2)
54 (35.5) 98 (64.5)
18 (18–19) 35 (23.0) 117 (77.0)
C (n ¼ 152)
71 (39.0) 111 (61.0)
17 (9.2) 168 (90.8) 4 (2–7)
54 (28.9) 133 (71.1)
66 (35.3) 121 (64.7)
18 (17–19) 50 (26.7) 137 (73.3)
I (n ¼ 187)
70 (36.8) 120 (63.2)
24 (12.4) 169 (87.6) 3 (2–6)
60 (30.2) 139 (69.8)
82 (41.2) 117 (58.8)
18 (17–19) 52 (26.1) 147 (73.9)
C (n ¼ 199)
75 (38.3) 121 (61.7)
32 (16.6) 161 (83.4) 3.5 (2–6)
52 (26.3) 146 (73.7)
54 (27.3) 144 (72.7)
18 (18–19) 46 (23.2) 152 (76.8)
I (n ¼ 198)
71 (38.4) 114 (61.6)
37 (20.4) 144 (79.6) 3 (2–6)
46 (24.7) 140 (75.3)
60 (32.1) 127 (67.9)
18 (17–19) 49 (26.2) 138 (73.8)
C (n ¼ 187)
38 (40.9) 55 (59.1)
24 (26.7) 66 (73.3) 3 (2–6.5)
27 (29.0) 66 (71.0)
27 (29.0) 66 (71.0)
18 (18–19) 20 (21.5) 73 (78.5)
I (n ¼ 93)
41 (46.6) 47 (53.4)
23 (26.1) 65 (73.9) 3 (2–5.5)
18 (20.5) 70 (79.5)
27 (24.7) 66 (75.3)
18 (17–19) 24 (27.0) 65 (73.0)
C (n ¼ 89)
37 (35.9) 66 (64.1)
8 (7.8) 95 (92.2) 4 (2–6.5)
25 (23.8) 80 (76.2)
27 (25.7) 78 (74.3)
18 (17.5–19) 26 (24.8) 79 (75.2)
I (n ¼ 105)
30 (30.9) 67 (69.1)
14 (15.1) 79 (84.9) 3 (2–6)
28 (28.6) 70 (71.4)
38 (38.8) 60 (61.1)
18 (17–19) 25 (25.5) 73 (74.5)
C (n ¼ 98)
Values are numbers (percentages) unless stated otherwise. More than 80% Singaporeans live in HDB flats. The different room types are used as a proxy indicator of socioeconomic status. I ¼ Intervention group, C ¼ Control group receiving usual care. a For young women followed-up at 6-month, there were marginally significant differences (P ¼ 0.051) between intervention and control groups. b HDB ¼ Housing Development Board apartments (public housing flats).
68 (45.3) 82 (54.7)
145 (42.6) 195 (57.4)
44 (29.3) 106 (70.7)
96 (27.4) 254 (72.6)
31 (22.3) 108 (77.7) 3 (1.75–6)
47 (31.3) 103 (68.7)
136 (38.7) 215 (61.3)
sex 59 (17.3) 282 (82.7) 3 (2–6)
19 (18–19) 32 (21.3) 118 (78.7)
18 (18–19) 87 (24.8) 264 (75.2)
Age (median IQR) 18 (18–19) 16–17 82 (24.3) 18–19 255 (75.7) Ethnicitya Malay 113 (33.5) Non-Malay 224 (66.5) Residenceb HDB 1/2/3-room/rental 98 (29.1) HDB 4/5-room/Private 239 (70.9) residence/Mansion Sexual behaviors First sex age (Median 16 (15–17) (IQR) Lifetime consistent condom use for vaginal Yes 48 (14.8) No 276 (85.2) Lifetime number of sexual 4 (2–7) partners (Median (IQR) STI status at baseline Positive 139 (41.9) Negative 193 (58.1)
I (n ¼ 150)
C (n ¼ 351)
I (n ¼ 337)
Participants followed-up at 6 months
Recruited participants at baseline
Table I. Characteristics of intervention and control groups at baseline and at 6-month follow-up
M. L. Wong et al.
Abstinence, safer sex intervention for adolescents change in phone numbers. With the exception of less young Malay women in the intervention than the control group (P ¼ 0.051), both groups were comparable for socio-demographic characteristics at follow-up. Participants’ attendance for the first two sessions was high (100%; 80%); 67% completed all three sessions. At 6-month follow-up, the intervention participants were 1.7 times more likely than the controls to report secondary abstinence (adjusted RR; 1.70, 95% CI 1.18–2.45) at follow-up. (Table II) On stratifying by gender, the effect on secondary abstinence was found in young men (adjusted RR, 2.03; 95% CI 1.25–3.30) but not in young women (21.9 versus 15.3%, P ¼ 0.234). There was also a significant dose–response relationship, with an increase in abstinence rates among young men receiving all three sessions compared to those receiving attending one or two sessions (chi square trend, P < 0.001, not in Table). Among those young men who abstained, more than half (58%) used to engage in sex with casual partners (36%) or sex workers (22%). Among the young women however, almost all (92%) of them reported sex with boyfriends and only 8% with casual partners at follow-up. The intervention participants were also more likely than the controls to report keeping to one partner (adjusted RR, 1.21; 95% CI 1.02–1.43) at follow-up. Upon stratification by gender, there was no intervention effect on the young men with similar proportions in both groups reporting having one partner (53.5 versus 51.5%, P ¼ 0.664). For the young women however, a significant effect was found, with the majority (82.9%) in the intervention group reporting having only one partner compared to 68.2% in the controls (adjusted RR, 1.25; 95% CI 1.04–1.50). There was no intervention effect on consistent condom use for vaginal sex overall (adjusted RR, 1.37; 95% CI 0.96–1.96, P ¼ 0.086) and when stratified by gender. The overall incident STI rate was lower in the intervention group than the control group, but it did not achieve statistical significance (intervention group: n ¼ 18, 9.9%, controls: n ¼ 24, 13.4%, P ¼ 0.297, RR: 0.71, 95% CI 0.41–1.28). Stratification by gender also showed lower but not statistically significant rates in the intervention
group for young men (8 versus 10.7%, P ¼ 0.549) and young women (11.6 versus 16.8%, P ¼ 0.299), respectively. In sensitivity analyses (Table III), the statistically significant effect of the intervention on abstinence for young men was maintained with the effect size being similar on using worst-case imputation and being slightly lower with multiple imputation. For young women, the significant effect size on keeping to one partner also remained similar on using worstcase imputation and multiple imputation.
Discussion Following the delivery of an abstinence and safer sex intervention to sexually experienced adolescent clinic attendees, young men in the intervention group were twice as likely than the usual care group to report secondary abstinence. For young women, the intervention group was almost 1.3 times more likely than the controls to keep to one sex partner only. The observed intervention effect is unlikely to be wholly due to Hawthorne effect because the usual care group also received the same attention from the caring clinic counselors and assessors. The effectiveness of the intervention is unlikely to be wholly due to the effect of participating in a study  because of the differential gender effect on abstinence and partner reduction. If it is due to study effect, the outcomes should go in the same direction for both genders. Other findings that support the efficacy of the intervention are the magnitude of the effect and the dose response relationship, with increasing abstinence rates among those attending more sessions. The application of a theoretical framework and findings from local needs assessment to develop the behavioral intervention could have contributed to its success. The new intervention adopted motivational interviewing and skills training to negotiate and practice abstinence and safer sex, and provided reinforcement sessions to both STI positive and negative patients, unlike the usual care program which provided one follow-up session to the STI positives only. 239
Combined Secondary abstinence Only one partner in the past 6 monthsd Consistent condom use for vaginal sexd Young men Secondary abstinence Only one partner in the past 6 monthsd Consistent condom use for vaginal sexd Young women Secondary abstinence Only one partner in the past 6 monthsd Consistent condom use for vaginal sexd
NA 6 (7.3)
NA 10 (19.2)
NA 16 (11.9)
NA 15 (19.0)
NA 17 (25.0)
NA 32 (21.8)
23 (21.9) 68 (82.9)
38 (40.9) 28 (53.5)
61 (30.8) 96 (71.6)
15 (15.3) 54 (68.2)
19 (21.3) 35 (51.5)
34 (18.2) 89 (60.5)
C n (%)
1.43 (0.79–2.58) 1.21 (1.01–1.45)e
1.52 (0.84–2.76) 1.25 (1.04–1.50)e
2.03 (1.25–3.30)e 1.08 (0.76–1.53)
1.91 (1.2–3.06)e 1.05 (0.74–1.47) 1.05 (0.62–1.79)
1.70 (1.18–2.45)e 1.21 (1.02–1.43)e
Adjustedb RR (95% CI)
1.69 (1.17–2.45)e 1.18 (1.00–1.40)
Unadjusted RR (95% CI)
I n (%)
I n (%)
C n (%)
Behaviors at 6-month follow-up
1.24 (0.74–2.07) 1.22 (1.02–1.46)e
1.79 (1.17–2.75)e 1.07 (0.76–1.51)
1.50 (1.06–2.10)e 1.19 (1.00–1.41)e
ITT (Imputed)c RR (95% CI)
Except for abstinence, model for young men adjusted for baseline STI status, respective baseline outcome, and for young women it was adjusted for ethnicity, respective baseline outcome. For abstinence, model for young men was adjusted for baseline STI status and for young women, ethnicity. I ¼ Intervention, C ¼ control group receiving usual care, NA¼ Not applicable. a Only includes those followed up to 6 months. b Complete case analysis; Poisson models with goodness-of-fit X2 > 0.9. c Multiple imputation, intention-to-treat analysis with demographics (ethnicity, age, education level, type of residence), frequency of smoking and alcohol drink, and age of first sex). d Numbers (percentages) at baseline differ from Table 1 as these are only reflective of those who did not abstain from sex at 6-month follow-up. e P < 0.05.
Numbers at 6-month follow-up
Table II. Change in outcomes from baseline to 6-month follow-up in the intervention and control groups
M. L. Wong et al.
Abstinence, safer sex intervention for adolescents Table III. Sensitivity analyses using intention-to-treat analyses with imputed data for those who defaulted 6-month follow-up as compared with complete case analysis Young men Reported secondary abstinence at 6 months SET
RR (95% CI)
Worst outcome scenario Demographics (Ethnicity, Residence, Education level, Age), Alcohol, Smoking, First sex age Complete case analysis
2.03 (1.23–3.35) 1.79 (1.17–2.75)
Young women Reported keeping to one partner at 6 months SET
RR (95% CI)
Worst outcome scenario Demographics (Ethnicity, Residence, Education level, Age), Alcohol, Smoking, First sex age Complete case analysis
1.41 (1.06–1.89) 1.22 (1.02–1.46)
Sets 1 and 2 are the results of the intention-to-treat analyses (ITT). For set 1, all the missing outcomes in both groups were assigned poor outcomes. For Set 2, the covariates obtained from baseline were used to impute the outcome variables.
The strong intervention effect on secondary abstinence in young men but not in young women could be because more young men had stopped engaging in sex with sex workers and casual partners. Young men, being the ones who often initiated sex , would have better control in curbing this behavior with the strategies learnt from the intervention. In contrast, almost all the young women reported sex with boyfriends. It would be more difficult for them to refuse sex with their boyfriends because they perceived sex as a way to convey love  and trust. The ineffectiveness of the intervention to increase condom use might be due to reduced pleasure for the young men. This was reported as the main reason for not using condoms by 34% of the young men in the intervention group. One study limitation is the low follow-up rate of around 60%. Similar low follow-up rates have been reported in trials on one-to-one behavioral STI intervention  and health promotion  for adolescents in clinic settings in United States. The low
retention rate is unlikely to be due to treatment assignment because the dropout rates and reasons for dropout were similar in both groups. Missing data from loss to follow-up might affect our interpretation on the intervention effects. However, sensitivity analyses yielded similar results. The assessment of incident laboratory-confirmed STIs is a methodological strength but the statistical power of the effect estimate was limited by the small sample size. Our 5-year study was powered to detect effect estimates in behavioral indicators, our primary outcomes, but not incident STIs because of the small number of adolescent patients attending the clinic. Future studies, using STIs particularly as a primary outcome indicator, should be carried out on larger samples in clinic settings in other Asian countries. Our study in a STI clinic-based sample limits the generalizability of the findings to the general adolescent population. However, our trial meets an urgent need because it is targeted at high-risk adolescents, of whom 80% with notifiable STIs attend the public STI clinic. 241
M. L. Wong et al. To our knowledge, this is the first RCT in Asia that is conducted on sexually active adolescents in a clinic setting. This is also one of the few trials [3, 14] that examine gender effects. As our study compared the new intervention with usual care rather than with a group with no STI prevention intervention, the outcome differences between the groups might be underestimated, thus giving a more conservative estimate of the intervention effect. However, the strength of comparing the new intervention with usual care is that we are testing a relevant public health question aimed at improving existing health services. This trial differs from the few trials [9, 14] on sexually active adolescents in clinic settings in one important aspect—the strong two-fold intervention effect on secondary abstinence among young men at 6-month follow-up. A smaller effect (unadjusted risk ratio: 1.27) of secondary abstinence was reported in St Lawrence’s study on sexually active African American adolescents in a health center albeit at 12-month follow-up . As gender-specific analysis was not reported in the analysis, it is unclear whether the effect differed by gender. A meta-analysis of HIV reduction trials on adolescents mainly in school or community settings in United States also found a smaller 1.2 effect size on abstinence . The strong effect on secondary abstinence in young men in our intervention might be explained by differences in intervention activities and socio-cultural characteristics and partner type of the adolescents. In addition to providing other safer sex alternatives, our intervention emphasized self-reflection and the sharing of practical tips on abstinence gathered from their peers. As about half of their partners were casual partners or sex workers, it is probably easier to abstain from sex with these partners than with girlfriends where there is an emotional attachment. The partner types were not reported in St Lawrence’s study.
Conclusion The intervention, compared to usual care, led to a significant effect on abstinence in young men and 242
keeping to one partner in young women. An intervention like this that integrates needs assessment with motivational interviewing and skills training to reduce sexual partners, as well as to negotiate and practice abstinence could help patient educators in clinic settings to increase public health impact on risky sexual behaviors. Resources should be made available for additional and enhanced counseling. The strong effect on secondary abstinence is promising. Future research should test this effect in other Asian countries and among Asian adolescents who study abroad or reside outside Asia.
Acknowledgements We thank the counselors and research assistants who have assisted in the conduct of this study.
Funding This work was supported by National Medical Research Council, Singapore (NMRC/1188/2008).
Conflict of interest statement None declared. References 1. Dehne KL, Riedner G. Sexually transmitted infections among adolescents: the need for adequate health services. Reprod Health Matters 2001; 9:170–83. 2. Jackson CA, Henderson M, Frank JW et al. An overview of prevention of multiple risk behaviour in adolescence and young adulthood. J Public Health 2012; 34: i31–40. 3. Bolu OO, Lindsey C, Kamb ML et al. Is HIV/sexually transmitted disease prevention counseling effective among vulnerable populations?: a subset analysis of data collected for a randomized, controlled trial evaluating counseling efficacy (Project RESPECT). Sex Transm Dis 2004; 31:469–74. 4. Chacko MR, Wiemann CM, Kozinetz CA et al. Efficacy of a motivational behavioral intervention to promote chlamydia and gonorrhea screening in young women: a randomized controlled trial. J Adolesc Health 2010; 46:152–61. 5. DeLamater J, Wagstaff DA, Havens KK. The impact of a culturally appropriate STD/AIDS education intervention on black male adolescents’ sexual and condom use behavior. Health Educ Behav 2000; 27:454–70.
Abstinence, safer sex intervention for adolescents 6. DiClemente RJ, Wingood GM, Harrington KF et al. Efficacy of an HIV prevention intervention for African American adolescent girls: a randomized controlled trial. JAMA 2004; 292:171–9. 7. DiClemente RJ, Wingood GM, Rose ES et al. Efficacy of sexually transmitted disease/human immunodeficiency virus sexual risk-reduction intervention for african american adolescent females seeking sexual health services: a randomized controlled trial. Arch Pediatric Adolesc Med 2009; 163:1112–21. 8. James NJ, Gillies PA, Bignell CJ. Evaluation of a randomized controlled trial of HIV and sexually transmitted disease prevention in a genitourinary medicine clinic setting. Aids 1998; 12:1235–42. 9. Jemmott JB , 3rd, Jemmott LS, Braverman PK et al. HIV/ STD risk reduction interventions for African American and Latino adolescent girls at an adolescent medicine clinic. A Randomized Controlled Trial. Arch Pediatr Adolesc Med 2005; 159:440–9. 10. Jemmott JB , 3rd, Jemmott LS, Fong GT. Abstinence and safer sex HIV risk-reduction interventions for African American adolescents: a randomized controlled trial. JAMA 1998; 279:1529–36. 11. Metzler CW, Biglan A, Noell J et al. A randomized controlled trial of a behavioral intervention to reduce high-risk sexual behavior among adolescents in STD clinics. Behav Therapy 2000; 31:27–54. 12. Milhausen RR, DiClemente RJ, Lang DL et al. Frequency of sex after an intervention to decrease sexual risk-taking among African-American adolescent girls: results of a randomized, controlled clinical trial. Sex Educ 2008; 8:47–57. 13. Shrier LA, Ancheta R, Goodman E et al. Randomized controlled trial of a safer sex intervention for high-risk adolescent girls. Arch Pediatric Adolesc Med 2001; 155:73–9. 14. St Lawrence JS, Brasfield TL, Jefferson KW et al. Cognitivebehavioral intervention to reduce African American adolescents’ risk for HIV infection. J Consult Clin Psychol 1995; 63:221–37. 15. UNICEF. Progress for Children: A Report Card on Adolescents. UNICEF, 2012. 16. Vuttanont U, Greenhalgh T, Griffin M et al. “Smart boys” and “sweet girls”–sex education needs in Thai teenagers: a mixed-method study. Lancet 2006; 368:2068–80. 17. Zabin LS, Emerson MR, Nan L et al. Levels of change in adolescent sexual behavior in three Asian cities. Stud Family Plan 2009; 40:1–12. 18. Aplasca MR, Siegel D, Mandel JS et al. Results of a model AIDS prevention program for high school students in the Philippines. Aids 1995; 9(Suppl. 1):S7–13. 19. Thato R, Jenkins R, Dusitsin N. Effects of the culturallysensitive comprehensive sex education programme among
Thai secondary school students. J Adv Nurs 2008; 62:457– 69. Thongkrajai E, Stoeckel J, Kievying M et al. AIDS prevention among adolescents: an intervention study in northeast Thailand Bangkok: The Population Council Quarterly Report, ICRW: Women and AIDS Program 1994. Chokephaibulkit K, Tarugsa J, Lolekha R et al. Outcomes of a comprehensive youth program for HIV-infected adolescents in Thailand. J Assoc Nurses AIDS Care 2015; 26:758–69. Underhill K, Montgomery P, Operario D. Sexual abstinence only programmes to prevent HIV infection in high income countries: systematic review. BMJ 2007. Johnson BT, Scott-Sheldon LA, Huedo-Medina TB et al. Interventions to reduce sexual risk for human immunodeficiency virus in adolescents: a meta-analysis of trials, 19852008. Arch Pediatr Adolesc Med 2011; 165:77–84. Mullen PD, Ramirez G, Strouse D et al. Meta-analysis of the effects of behavioral HIV prevention interventions on the sexual risk behavior of sexually experienced adolescents in controlled studies in the United States. J Acquir Immune Defic Syndr 2002; 30(Suppl. 1):S94–S105. MOH. Ministry of Heath: Health Facts Singapore 2013. Singapore, 2013. Katz MH. Evaluating Clinical and Public Health Interventions: A Practical Guide to Study Design and Statistics. Cambridge: Cambridge University Press, 2010. Green LW, Kreuter MW. Health Program Planning: An Educational and Ecological Approach. New York: McGraw-Hill, 2005. Glanz K, Rimer BK, Viswanath K. Health Behavior and Health Education: Theory, Research, and Practice. San Francisco: Jossey-Bass, 2008. Wong ML, Chan RK, Koh D et al. Premarital sexual intercourse among adolescents in an Asian country: multilevel ecological factors. Pediatrics 2009; 124:e44–52. Wong ML, Chan RK, Tan HH et al. Gender differences in partner influences and barriers to condom use among heterosexual adolescents attending a public sexually transmitted infection clinic in Singapore. J Pediatr 2013; 162:574–80. StataCorp. Stata 14 Base Reference Manual: College Station, TX: Stata Press., 2015. de Bruin M, McCambridge J, Prins JM. Reducing the risk of bias in health behaviour change trials: improving trial design, reporting or bias assessment criteria? A review and case study. Psychol Health 2015; 30:8–34. Walker Z, Townsend J, Oakley L et al. Health promotion for adolescents in primary care: randomised controlled trial. BMJ 2002; 325:524.