Improving teenagers' knowledge of emergency contraception: cluster randomised controlled trial of a teacher led intervention

Recruitment

Head teachers of all mixed sex, state secondary schools in the four local education authorities in the county of Avon were sent a letter in May 1999 inviting their school to participate in the study. We asked for reasons for declining to take part.

Consent for the study was elicited from the head teacher, as the “gatekeeper” of the cluster; the director of education in each of the four education authorities; parents, in the form of a letter from the head teacher, with an opt out slip to return should they not want their child to take part in the study; and pupils, at the time of delivery of the questionnaires. The relevant ethics committees approved the study.

Randomisation

To ensure that the control and intervention groups were balanced with respect to factors likely to influence the outcome, we used a minimisation strategy in randomising schools to the groups.12 These factors were the percentage of pupils entitled to free school meals, size of year group, whether sex education was taught by a tutor or specialised team of teachers, and whether sex education was taught mainly in year 9 or in year 10. For size of year group and the percentage of pupils entitled to free school meals, the respective binary variable indicated whether the school was above or below the median. Allocation of schools to the groups, which occurred after all schools were recruited to the study, was determined by a sequential minimisation protocol. The order in which schools were minimised was determined by computer generated random numbers. Teachers and pupils were not blind to allocation.

Intervention

The aims of the in-service training of the teachers were to improve their knowledge about emergency contraception, to examine myths about the method (box), and to develop skills for use in a lesson for year 10 pupils on emergency contraception. The two hour training session was given by one of the authors (AG, who is trained in family planning) to teachers involved in delivering sex education at each school. Costs of extra staff to cover teachers taken out of lessons for the training were reimbursed. The proposed lesson to the pupils was demonstrated to the teachers during the training session. The lesson was taken from a resource produced for schools by Brook Advisory Centres.13 This was the only resource available at the time that included materials for the delivery of a lesson on emergency contraception. We chose year 10 for the intervention because a survey we did of schools in Avon indicated that most schools covered contraception by the end of this year.

The content and delivery of the in-service training of teachers and the delivery by teachers of the lesson to pupils were piloted in one school in the area in the summer term of 1999. The pilot included two “active learning” methods (the involvement of pupils in experiential activities): a scenario activity and a role play, which have been shown in a review of the effectiveness of sex education to be the most effective teaching methods.14 The scenario chosen for the pilot involved a male lead role, because research has shown that boys tend to feel excluded from sex education lessons.15 However, our observation of the lesson, and focus group discussions with pupils who took part in the pilot, showed that the pupils saw the male lead as unrealistic, and so we changed the scenario to that of a young woman having unprotected sex while drunk on a Friday night (box). Furthermore, the role play was removed, as pupils found it difficult to actively engage with the activity (as shown by previous research16), and was replaced by a quiz on emergency contraception from the same resource.

Evaluation of the intervention

The outcome data comprised the pupils' answers to a confidential, self completed questionnaire administered at baseline and to a similar follow up questionnaire six months later. The questionnaires included questions used in a previous study by one of the authors (AG) and in a concurrent study in Scotland. 8 16 In schools in the intervention group, the teachers were trained and the pupils were given the lesson within two weeks of the baseline questionnaire. Schools in the control group received the intervention after the follow up questionnaire was completed. All the schools continued to deliver their usual sex education programme. The study was completed in the academic year 1999-2000.

Sample size and study outcomes

The primary outcome was the proportion of pupils able to identify correctly the 72 hour limit for use of hormonal emergency contraception. To detect a difference of 18% between the intervention and control groups with 80% power and a level of significance of P<0.05 (two tailed test), 11 schools were required in each group. This assumed an intention to treat analysis, an absentee rate of 15%, and an intra-cluster correlation coefficient of 0.05, and that at baseline 26% of the girls would know the correct 72 hour time limit for use of hormonal emergency contraception (data derived from previous research8). Secondary outcomes were knowledge of the correct time limit (five days) for the use of the intrauterine device as emergency contraception, the proportion of pupils who weren't virgins, the proportion of pupils who had used emergency contraception, and the pupils' intention to use emergency contraception in the future.

Analysis of data

A regression analysis, weighted for schools rather than individuals, was used to analyse primary and secondary outcomes separately for boys and girls. The analysis was adjusted for baseline score and the four factors used in the minimisation strategy.17

We did a sensitivity analysis of the primary outcome to ascertain the influence of absent pupils on the findings. The number of pupils needed to be taught for one more pupil to know the correct time limit was calculated as the reciprocal of the difference between the groups in the proportion of pupils knowing the time limit, obtained from the adjusted analysis and rounded to the nearest whole number. We used Stata version 6.0 (StataCorp, College Station, TX, 1999) for the analyses.

Participants

Of the 82 secondary schools in the four local education authorities in Avon, 49 were eligible to take part in the study, and 24 were recruited. Table 1 gives the reasons for exclusion, and table 2 gives the reasons schools gave for declining to take part in the study. A total of 1974 boys and 1820 girls were eligible to take part in the study.

Nine pupils in the intervention group and 24 pupils in the control group were withdrawn from the study by their parents. Pupils were able to leave their questionnaires blank; 18 boys and two girls did so. A researcher who was blind to allocation to group assessed whether questionnaires were spoilt. Forty four boys and no girls had spoilt questionnaires; these were removed from further analysis. No schools dropped out of the study. Numbers of pupils who completed the study are given in the figure, which shows the flow of pupils and schools though the study.

Flow of schools and pupils through the study

• Download figure
• Open in new tab
• Download powerpoint

Schools eligible to take part in the study were similar to all state secondary schools in Avon when compared on factors important to the primary outcome. Schools declining to take part in the study were less likely, compared with schools agreeing to take part, to teach sex education to younger year groups or to use a specialised team of teachers (table 3). Schools in the control group had larger year groups and more specialist teams teaching sex education than schools receiving the intervention (table 4).

Effect of the intervention on the primary outcome

Table 5 and Table 6 show the results of the intervention on the proportions of pupils knowing the correct time limit for hormonal emergency contraception. At six months' follow up the proportion of pupils in the intervention group who knew the correct time limit was significantly higher than the proportion in the control group (boys 15.9% higher (95% confidence interval 6.5% to 25.3%, P<0.01), girls 20.4% higher (10.4% to 30.4%, P<0.01)). To examine the effect of absent pupils on the results, we did a sensitivity analysis assuming that all absent pupils at both baseline and follow up in both groups did not know the correct time limit for hormonal emergency contraception. The results show that the intervention effect remained (boys 14.8% higher (6.6% to 23.0%, P<0.01), girls 19.8% higher (12.0% to 27.6%, P<0.01). The number of pupils needed to be taught for one more pupil to know the correct time limit was six for boys (1/0.159=6.29) and five for girls (1/0.204=4.90).

Secondary outcomes

Table 7 shows the effect of the intervention on the secondary outcomes. The proportion of pupils knowing the correct time limit for use of the intrauterine device as emergency contraception was significantly higher in the intervention group than in the control group (boys 4.2% higher (0.7% to 7.7%, P=0.02), girls 10.7% higher (0.4% to 21.0%, P=0.04). The proportion of pupils who were not virgins did not differ significantly between the groups, and of these pupils the proportion who said they had used emergency contraception did not differ significantly between the groups. There was no difference between the groups in the proportion of pupils intending to use emergency contraception in the future.

Design of the study

Reviews of interventions in sexual health education aimed at young people concluded that the design of such evaluations needed to be improved so that evidence of the effectiveness of different interventions could be generated. 14 18 We believe our evaluation to be rigorously designed, with a defined focus and clear objectives. The lesson was developed by an organisation that is well known for providing resources that are appropriate for particular age groups and that use active learning rather than didactic teaching methods. The in-service training and the lesson were piloted and amended to ensure their relevance for, respectively, teachers and year 10 pupils.

Kirby et al state that active learning activities help pupils personalise information.14 Teachers need to feel confident enough to deliver such lessons. This “one off” intervention was limited in its ability to develop such teaching skills. However, our observations of a number of lessons in the course of the study led us to believe that active rather than didactic methods were used. There were some differences between the intervention and control groups at baseline. This did not, however, result in bias, because there were no real differences between the results of the adjusted and unadjusted analyses.

Analyses were done at the level of schools rather than individuals for two reasons: the intervention was delivered to clusters of pupils (schools), and the pilot lesson showed that pupils felt more confident completing the questionnaires without individual identification codes.

Our intervention showed no evidence of changed sexual behaviour—whether increased sexual activity or greater use of emergency contraception. It was important for us to show this lack of effect, in view of the ongoing debate on the effects of sex education and the argument that promoting the use of contraception encourages sexual activity.19

The schools in the study were similar to state schools in England generally in terms of standard measures of deprivation and academic attainment, but sex education in our schools was more likely than in schools nationally to be team led, which may reduce the generalisability of the findings. However, this method of teaching may become more common, after a recent report from the Office for Standards in Teaching encouraged the use of teams.20 Even in the schools that already had this good practice, pupils' understanding of emergency contraception improved.

Usefulness of the intervention

Lack of awareness of correct time limits for use of emergency contraception is not the only factor preventing its use. It is likely that concerns about the safety of emergency contraception also act as a deterrent. Women seeking emergency contraception, as well as personal, social, and health education teachers, have been shown to have concerns of this nature. 11 21 Of 87 teachers in a survey undertaken in the study area, only 45 agreed with the statement “Using emergency contraception is safe.”11

There is now some evidence that hormonal emergency contraception is more effective the sooner it is used after unprotected sexual intercourse.5 However, it remains effective—and is licensed for use—up to 72 hours after sex. Our study was undertaken before emergency contraception was made available over the counter.

An interesting question is whether the intervention effect lasts: whether the outcome is sustained over a longer period than six months, and how frequently the in service training would need to be repeated in order to continue to be effective.

Limitations of the study

The main limitation is that we did not include an evaluation of health gains, as advocated by previous reviews. 14 18 None of the randomised controlled trials in a recent review of research into the primary prevention of pregnancy among adolescents showed any benefit in health based outcomes.22 Also, our sample size was too small to show a reduction in the rate of conceptions.

The change in the pupils' knowledge may not translate to a change in behaviour, as so many other factors undoubtedly play a role. However, educating teenagers on the time limits for use of emergency contraception after unprotected sex is more likely to have an impact on behaviour than, say, encouragement to use a condom.