The "ROsafe" program was designed as an intervention with an educational as well as a service providing component, based on evidence and theory to increase STI testing behavior by senior vocational students. The results of this cluster randomized study suggest that the intervention that offered education combined with health services had positive effects on the uptake of STI testing among students with sexual experience. Compared to a control group, students that received the health education and were offered sexual health services in the schools were more often tested (OR 4.25) for STI than students who received only one of these components.
The prevalence rate of STI of students in the study group tested after the intervention was lower than expected. Only 1.4% of the sample tested positive for Chlamydia, while in a large national Chlamydia study in the Netherlands an overall prevalence of 2.0% was found, with 3.2% in highly urban areas (in which our study group lives) . Furthermore, in higher risk populations such as in visitors of the Rotterdam STI clinic, Chlamydia prevalence is around 10% .
The hypothesis that offering school based facilities can enhance the uptake of sexual health care is supported by other studies [25–29]. Characteristics of the services provided in this study are in line with those that are identified as facilitators for the use of school based sexual health services for adolescents . For example, the services were offered in a space that was confidential, easy accessible, free of charge, with the possibility to take a friend along. Our study is rather unique because, to our knowledge, no studies exist on school-based interventions to promote STI testing among adolescents that combine health education and sexual health services. However, school-based interventions aiming at testing (among other outcomes) comprising only educational components could not discover significant effects on STI testing [16, 17].
Female students were more often tested than male students. The difference in test-rate between female students and male students was large: 21.5% of sexually experienced women versus 5.4% of the men performed an STI test. Also at the school based sexual health service most of the clients were female. Perhaps school based health services are more accepted by women, as was also suggested in a review by Kirby who observed that older males may not be easily reached by school- based or school-linked clinics in the United States . European studies also suggest that male students have other preferences than female students with respect to school based health services and school based sexual health services. E.g. a Swedish study by Makenzius among male students revealed that a majority of them felt the need for counseling and advice about their sexuality, but that they needed more male-friendly youth health services, such as male staff, special hours for males and alternative methods of STI testing, such as Internet-service for Chlamydia testing . The preference for male staff by boys was also reported in a study among British secondary school students. The school drop-in clinics were most frequently visited by girls . However, also among the general STI clinic visitors (not school- based) below 25 years in the Netherlands, 65% is female [32, 33].
Students who were absent at baseline were excluded from the analysis. However, in the entire sample which also includes those without a baseline measurement (n = 1762), reported STI prevalence was 8.0%. This indicates that students that were tested for STI during the time of the intervention were more often absent at baseline. Possibly, these students were more at risk for STI than those who were present al baseline. Data on their sexual behavior is available from the 2 months preceding their follow up measurement, and shows that their sexual behavior was not riskier from students who were compliant at baseline. However, evidence exist that students who drop out or have less attachment to school have more sexual risk behavior due to a riskier lifestyle which might be related to absenteeism and drop out .
Several methodological limitations should be considered in interpreting this study. Although teachers were requested to administer questionnaires to absent students in one of the following lessons, this study was confronted with a high attrition. Only 57% of the baseline sample was compliant at one or more follow-up measurements, due to high levels of absenteeism from school among these students. It is unknown whether this absenteeism is in any way related to the study, because the participating schools could not provide figures on absenteeism at school or class level. However, the Ministry of Education calculates school drop-out rates on the base of registrations at the start and finish of the school year. The 2009 figures showed a 13% drop out at vocational schools in the Rotterdam-Rijnmond area in the school year 2007/2008 .
Another reason for loss to follow-up was transfer of students to another class or school. Also, not all teachers were cooperative in administering each follow up questionnaire to their students, due to workloads, illness, cancellations of lessons, exams, and periods of work placement when students were not in the school. The cooperation of teachers also depended on their degree of involvement with the subject. High attrition rates for intervention research or low compliance in a school screening program are not rare in this type of research in schools, especially when targeting high risk youth [16, 35, 36]. The outcome measure of STI testing could be constructed if a student was compliant for at least 1 follow-up measurement. Due to missing data at follow-up measurements, test rates can be underestimated. If a student reported no test at t1, but dropped out after t1, a possible test after t1 is not reported. Self reports were the principal mean of data collection. Where possible, client registrations of the health services were matched to the self reports. However, only 61% of self reported STI tests could be successfully matched. Apart from the low response rates at the follow-up questionnaires this was due to the fact that students could choose to be seen anonymously, and at the questionnaires only a few identifying questions were asked (not their name).
A second limitation, also caused by a high degree of absenteeism, was that only a minority of the students received the complete intervention: only 32% of students who filled out the questionnaire at t1 reported to have received the full intervention, However, 26% did not fill out the t1 questionnaire, and it is unknown whether they received any of the intervention components.