Outcomes of three different models for sex education and citizenship programs concerning knowledge , attitudes , and behavior of Brazilian adolescents

Three different school-based sex education and citizenship programs in public schools in Rio de Janeiro, Belo Horizonte, and Salvador, Brazil, were evaluated in a cross-sectional study comparing knowledge, attitudes, and practices in sexuality, citizenship, and gender issues among adolescents participating in the programs’ activities as compared to adolescents enrolled in schools without such programs (controls). Results showed that Salvador’s program achieved good results, with significant changes in knowledge on sexuality and reproductive physiology, attitudes regarding citizenship, and current use of modern contraceptives; Rio de Janeiro’s program succeeded in improving students’ knowledge of reproductive physiology and attitudes towards sexuality; Belo Horizonte’s participants showed greater knowledge of reproductive physiology and STI/HIV prevention but had less positive attitudes towards gender issues, while reporting greater sexual activity. The main difference between Salvador’s program and the others was the focus on creative and cultural activities; Belo Horizonte’s main difference was its lack of interaction with health services and professionals. However, after the evaluation Belo Horizonte reframed its educational strategies and launched a scaling-up process in a joint effort with the health and school systems. Sex Education; Adolescent; Evaluation Introduction An important decrease in the fertility rate has occurred in the world over the past three decades. In parallel, a proportional increase has been observed in the 10-to-24-year age bracket, to the extent that this group now constitutes nearly one-third of the world’s population. In the case of Brazil, this increase in the proportion of adolescents and young adults is the consequence of the relatively high fertility rates of the 1960s, complemented by a sharp decrease in fertility over the past two decades. The size of this age group and its vulnerability in sexual and reproductive health are good reasons for considering it a high priority. According to a 1999 study by the Centro Brasileiro de Análise e Planejamento 1, 47.0% of male adolescents (15-19 years) reported sexual initiation before age 15, as compared to only 35.0% in 1994. Although prevalence of contraceptive use among the group of sexually active 15-19-year-olds has increased, reaching 58.8% and 67.4% among females and males respectively, according to the 1996 Brazilian Demographic and Health Survey 2, the consistency and effectiveness of use appear to be poor, since this is the only age group still showing no sign of a decline in fertility. Moreover, according to data from the National STD and AIDS Program, the 25-39-year age group has been the most heavily affected by the AIDS epidemic


Introduction
An important decrease in the fertility rate has occurred in the world over the past three decades.In parallel, a proportional increase has been observed in the 10-to-24-year age bracket, to the extent that this group now constitutes nearly one-third of the world's population.In the case of Brazil, this increase in the proportion of adolescents and young adults is the consequence of the relatively high fertility rates of the 1960s, complemented by a sharp decrease in fertility over the past two decades.The size of this age group and its vulnerability in sexual and reproductive health are good reasons for considering it a high priority.
According to a 1999 study by the Centro Brasileiro de Análise e Planejamento 1 , 47.0% of male adolescents (15-19 years) reported sexual initiation before age 15, as compared to only 35.0% in 1994.Although prevalence of contraceptive use among the group of sexually active 15-19-year-olds has increased, reaching 58.8% and 67.4% among females and males respectively, according to the 1996 Brazilian Demographic and Health Survey 2 , the consistency and effectiveness of use appear to be poor, since this is the only age group still showing no sign of a decline in fertility.Moreover, according to data from the National STD and AIDS Program, the 25-39-year age group has been the most heavily affected by the AIDS epidemic Cad.Saúde Pública, Rio de Janeiro, 21(2):589-597, mar-abr, 2005 in the country, representing 59.6% of total cases recorded between 1980 and 2000 3 .Considering the natural history of AIDS, it is estimated that the maximum frequency of these infections occurred in young people (15 to 24 years).
These problems were the base from which the Brazilian Government launched its National Health Program for Adolescents (PROSAD) in 1988, with the purpose of promoting the physical, sexual, and psychological health of adolescents as well as supporting their social inclusion.A decade later, in 1998, the Ministry of Education approved the National School Curriculum Guidelines, which established that information on sexual and reproductive health should be a crosscutting theme in all primary and secondary school curricula.
Parallel to this process, beginning in the latter 1980s several nongovernmental organizations (NGOs) and some public institutions launched initiatives focusing on sex education for adolescents.The results of only a few of these initiatives have been systematically evaluated, but most reports have assumed that they were good, effective, or useful because they were well-received by the target population.
Since 1988 the Odebrecht Foundation has supported one of the first and most important private initiatives in adolescent sex education in Brazil.From 1994 to 1997, the Foundation supported three sex education programs in three Brazilian cities, implemented by local NGOs in partnership with the public school and health systems.
The purpose of the current study was to evaluate the outcome of these programs three years after their inception with regard to knowledge, attitudes, and sexual practices of participating adolescents.In addition, some citizenship and gender issues were also studied.

The interventions
The programs evaluated were: the EDUCARTE Program in the city of Rio de Janeiro, Education: an Exercise in Citizenship in Salvador, Bahia, and Sexuality and Affectivity Education in Belo Horizonte, Minas Gerais.Although the three programs were developed independently, they had several commonalities, including the same objective of supporting the social and individual development of adolescents enrolled in public schools.
The three programs shared similar guiding principles, concepts, and methods for working with sexuality, gender, and citizenship as crosscutting themes.All three focused on an expanded concept of sexuality, changing the common comprehension of sexuality as intercourse and adding affectivity to this new dimension.They also focused on the social construction of gender roles, attempting to deconstruct traditional societal roles attributed to men and women.The concept of citizenship was anchored in the principles of human rights.Moreover, adolescents were actively stimulated to take part in the teaching-learning process, and a positive adult-adolescent relationship was promoted in both their schools and families.Finally, schools were encouraged to interact with their own local communities.
Table 1 summarizes the three programs' main characteristics: duration; number of training hours; number of schools and health centers participating; number of schoolteachers, health professionals, and adolescent peer educators trained; and number of adolescents directly involved in each program.

Evaluation
The programs were evaluated through a joint effort by the Odebrecht Foundation (a nonprofit foundation created and maintained by the Odebrecht Construction and Engineering Company, with headquarters in Salvador, Bahia) and the Maternal and Child Research Center of Campinas (Cemicamp), using a cross-sectional KAP (knowledge, attitudes, and practice) study.The knowledge, attitudes, and practices of adolescents who participated in the sex education programs were compared with those of adolescents enrolled in schools that did not have such programs in their curricula, separately for each municipality.Although the evaluation was performed by a research team not involved in the interventions, almost all phases of the process were participatory, with stakeholders taking part in the elaboration of the research instrument, field work, and discussion of results.

Data collection
Data were collected in late 1997 in all cities using the same self-administered, multiple-choice questionnaire.All students from grades 5 to 8 in classes indicated by the schools as having participated in the program in either Rio de Janeiro or Belo Horizonte were invited to complete the questionnaire.In the schools used for comparison (controls), all students from grades 5 to 8 were invited to respond to an identical questionnaire.Enrollment of schools to participate in the sex education programs was determined by local program managers, so it was not random.The program in Salvador had been developed in all the city schools, so there were no true comparison schools.Therefore, in Salvador the control schools were defined as those that begun implementing the program in 1997, while target schools were those that had initiated activities before 1997.Schools in each city were selected to participate in the study by simple random sampling (separately) from a list of program and comparison schools.
Only students ages 10 through 19 from target schools who declared having participated in sex education activities at school were considered participating adolescents.Adolescents in the same age bracket from control schools and who declared never having participated in any sex education activity comprised the control group.
The data collection instrument was developed by the research team from Cemicamp in collaboration with the local program managers in each city.The pre-tested, self-administered, multiple-choice KAP questionnaire contained 50 questions designed to provide answers to the program evaluation's specific objectives.

Sample size
Sample size was calculated from the estimated proportion (prevalence) of 50.4% of 15 to 19 years-old who declared having had sexual intercourse in a survey performed in the cities of Rio de Janeiro, Curitiba, and Recife 4 .Alpha was set at 0.05 and beta at 0.10.The absolute difference between the programs' and the control groups' proportions was set at 8.0%, resulting in 838 students per group in each municipality 5 .
The final sample size did not reach the desired number, although it was never less than 40 cases per group for questions regarding sexual practices.In this case, sample precision was estimated at 32.0% (alpha was set at 0.05 and beta at 0.20) based on data from the 1996 Brazilian Demographic and Health Survey 2 , in which 20.0% of adolescents reported previous sexual intercourse.

Study variables
The main knowledge outcomes referred to sexuality, reproductive physiology, contraception, and sexually transmitted infections (STI), including HIV/AIDS.Attitudes referred to sexuality, citizenship, and gender, including whether the respondents felt at ease talking about sexuality.Finally, there were questions on current sexual activity, condom use, and modern contraceptive methods utilization.
For variables related to adolescents' knowledge, indicators were constructed from scores derived from the number of correct answers to a battery of questions for each specific variable.The indicator of gender attitudes was defined by a group of 16 activities related to the individual, domestic, and public dimensions of gender.Students were asked to state whether they considered each of them as being associated with males, females, or both genders.The expected attitude was that students would associate the highest possible number of activities with both genders.The indicator of attitudes toward sexuality involved agreement or disagreement with 13 statements concerning myths and taboos frequently associated with this topic.The indicator of attitudes toward citizenship was based on 18 daily situations involving personal interrelationships which students were asked to rank as never, sometimes, or always acceptable to them.The scores had different ranges and were dichotomized using median values of sample data.
The questionnaire also allowed the research team to collect demographic information on age, sex, skin color, grade in school, employment, and number of extra-curricular activities taken.These variables were included in the analysis as statistical controls.

Data analysis
Data analysis was performed separately for each city.Initially, bivariate statistical analyses investigated possible differences between groups in students' demographic characteristics and sexual practices.These analyses were performed separately for programs and controls, using Yates chi-square test 6 .
Multiple logistic regression analyses examined the relationship between groups (programs and controls) and each of the following dependent variables: knowledge about sexuality, reproductive physiology, contraception, STI/AIDS, and HIV transmission; attitudes on sexuality, gender, citizenship, whether the respondent felt at ease talking about sexuality, and current sexual activity, current condom use, and current use of a modern contraceptive method.Six other control variables: sex (male: 1/female: 0); age (years); skin color (white: 1/non-white: 0); school grade (5 th and 6 th : 0/7 th and 8 th : 1); employment (yes: 1/no: 0); number of extra-curricular activities (≤ 1: 0/> 1: 1) were entered into the models.Stepwise selection was used to adjust the regression models for all variables except "group", which was entered into all models.Results of logistic regression analyses are presented only for "group" as adjusted odds ratio with the respective 95% confidence interval 7 .
All data entry and statistical analyses used SPSS 6.0.

Ethical considerations
Before participating in the study, students were asked to sign a free informed consent form ensuring adherence to the ethical principles of social science research, under the respective Resolution by the Brazilian Health Council, Ministry of Health 8 .Students 17 years or under were required to provide written consent from a parent or legal guardian.The research program and consent form were approved by the Institutional Review Board/Ethics Committee of the State University of Campinas.

Sample description
More females than males were participating in the program in Belo Horizonte.In Salvador the program adolescents were significantly younger than the controls, while no age differences were found between the two groups in Rio de Janeiro or Belo Horizonte.In addition, more controls than program adolescents described themselves white-skinned in Belo Horizonte.Self-declared skin color distribution showed no statistical differences between participants and controls in Salvador and Rio de Janeiro (Table 2).

Knowledge on sexuality
Adolescents participating in the sex education program in Salvador had better knowledge on sexuality than controls.Participating adolescents from the three cities had significantly greater knowledge of reproductive physiology than controls, with the odds ratio varying from 1.60 to 1.79.In Belo Horizonte, participating adolescents also showed greater knowledge of STI/HIV/AIDS than controls (Table 3).

Attitudes on sexuality, citizenship, and gender issues
Adolescents participating in the programs in Rio de Janeiro and Salvador showed more positive attitudes toward sexuality-related issues than controls.Only participating adolescents from Salvador reported feeling more at ease talking about sexuality and showed the expected attitudes towards citizenship issues in greater proportion than controls.Participating adolescents from Belo Horizonte expressed the expected gender attitudes in a significantly lower proportion than controls (Table 4).

Self-reported sexual behavior by participating adolescents and controls
More participants than controls in Belo Horizonte reported sexual activity in the previous three months, both in the bivariate and multivariate analyses (60.4% and 36.4%,respectively) (OR = 2.49).Reported use of modern contraceptive methods was almost twice as frequent among Salvador's participating adolescents as compared to controls in the multivariate analysis, while in the bivariate analysis this proportion was one third more frequent.Both results were statistically significant (Tables 5 and 6).

Discussion
All three models of sex education and citizenship programs proved capable of improving adolescent knowledge on some sexuality-related issues, especially reproductive physiology.Salvador's program also appeared to have improved participants' attitudes towards sexuality and citizenship issues, besides contributing to the adoption of modern contraceptives by the more sexually active adolescents.Belo Horizonte's program, on the other hand, appeared to be associated with less positive attitudes towards gender equity, and more of its participants reported having engaged in sexual activity.Rio de Janeiro's model had no impact on the adoption of safer sex practices, but did improve the adolescents' attitudes towards sexuality.
There are a number of possible explanations for the observed differences between sites, not all necessarily related to the programs themselves, including the selection of classes in each of the intervention schools where the program took place.Since the programs were not part of an operational research study but had purely educational goals, the schools, classes, and students were not randomly selected to participate.It is thus possible that the adolescents who were considered to be at higher risk were selected to participate in the programs.As there was no baseline information available for the three programs, it was not possible to assess differences between control and intervention schools prior to the initiation of the programs.If adolescents with less knowledge and less positive attitudes or behaviors were included in the programs, any improvement resulting from the intervention may have placed them just at the level of control adolescents, who already had better knowledge, attitudes, or behavior beforehand.If so, any impact from the programs would be undetectable by the methodological design used in this evaluation.
In addition, the evaluation focused on program results rather than processes.Due to these limitations, some questions remain unanswered.Despite such limitations, there were strategic differences between the three models which could also explain greater changes in Salvador than in Rio de Janeiro or Belo Horizonte.For example, Belo Horizonte's program did not involve health professionals or health centers in its implementation strategy, while a partnership was established between health centers and schools in the other two cities.Another striking difference between Salvador's program and the other two was its strong emphasis on participants' involvement in creative activities such as theater skits and dramatization.It was not clear why Rio de Janeiro's program did not appear to have promoted the adoption of safe sex practices among its participants, since it implemented two important strategies for reaching this objective: a partnership with the health sector and training of over two thousand peer educators.
Other factors may also have influenced the different results in the three cities.Exposure to other sources of information and other factors that may have induced attitude and behavioral  provide valuable information for the continued improvement of educational programs aimed at reducing the risk of unwanted pregnancy and disease among adolescents.

Contributors
All the authors participated in drafting the article and in all stages of the research.

Table 2
Proportion of participants in the programs and control students according to certain characteristics, in each of the three cities.Rio de Janeiro, Belo Horizonte, and Salvador, Brazil.

Table 3
Adjusted odds ratio of having better knowledge of sexually-related issues: adolescents participating in the sex education programs compared with controls* in the three cities.Rio de Janeiro, Belo Horizonte, and Salvador, Brazil.

Table 4
Adjusted odds ratio of expressing an expected attitude toward sexuality, citizenship, and gender issues: adolescents participating in sex education programs compared with controls*, in three cities.Rio de Janeiro, Belo Horizonte, and Salvador, Brazil.
* Exposure variable (yes: sex education group; no: control group).** These variables were divided into groups according to the median.Note: significant odds are highlighted in bold.

Table 5
Proportion of participants and control students who reported being sexually active, in each of the three cities.Rio de Janeiro, Belo Horizonte, and Salvador, Brazil.