Sexual behaviour is a normal part of human development, and sexual health is a central component of physical and mental health. However, sexual activity can also lead to adverse health consequences including sexually transmitted infections (STI) and unintended pregnancies, whose impact may be profound and long lasting. Intimate relations may also be the setting for violence between partners with wider health consequences for families. In the UK and elsewhere the above adverse outcomes are common and may be increasing, in part reflecting deficiencies in intervention policies. The UK Government has specified improving sexual health as an overarching priority. Sexual behaviour associated with increased risk of adverse sexual health outcomes may be influenced by the dynamic interplay of biological, individual, social, cultural, environmental and political processes across the life course. In particular, sexual development takes place in the context of developing social, peer and romantic relationships. It follows that an understanding of the development of young adult sexuality and other risk-taking behaviours and the association between these behaviours and outcomes such as STI requires longitudinal data about early and middle childhood experiences, as well as accurate data about the onset of romantic, sexual, and risk-taking behaviour from late childhood onwards. Such understanding is key to the design of appropriate public health interventions to improve sexual health. If the causal roots of risky sexual behaviour and its adverse sexual health consequences lie early in the life course then this is where effective intervention should be targeted. However, it may be that risk of adverse sexual health outcomes is substantially determined by proximal factors essentially independent of the pathways that precede them. This possibility has different implications for prevention. It may also be the case that most risk of adverse sexual health outcomes is not primarily attributable to a definable syndrome of risky sexual behaviour. Finally it is important to investigate the predictors of positive sexual health outcomes as these may also have implications for effective health promotion.

Investigation of these hypotheses, and their different implications for prevention, calls for general population based studies of adolescents at an age where a sufficient proportion is sexually active with prospective data both on sexual behaviour, its correlates and antecedents and on sexual health outcomes. ALSPAC is one of the few studies internationally, and the only study in the UK, currently able to support such investigation

Objective 1: To measure patterns and prevalence of sexual behaviour and related risk behaviours at age 24

Participants attending the clinic will be invited to complete a validated computer assisted self-interview (CASI), covering sexual behaviour, attitudes and lifestyle (including age at coitarche in non-virgins, circumstances of first intercourse, whether activity was regretted, condom and other contraceptive use, pregnancy and parenthood, use of pornography, other 'romantic relations', intimate partner violence, and positive sexual experiences. Questions used will be the same as in previous assessments to maintain consistency and allow estimation of incidence with more detailed questions on current sexual lifestyle appropriate for young adults also included.

Objective 2:To estimate the age-specific prevalence and incidence of genital Chlamydia, Mycoplasma, and Gonorrhoea in a birth cohort of UK adolescents.

ALSPAC participants attending the age 24 clinic assessments will be invited to submit a first catch urine specimen, which will be tested for the presence of the above infections, and a blood sample, which will be tested for the presence of chlamydia antibodies. Permission will also be sought to test stored blood and urine samples collected at previous clinic assessments between the ages of 11 and 15.

Objective 3: To investigate pathways to both positive sexual health outcomes and risky sexual behaviour and adverse sexual health outcomes from early life to adolescence in a birth cohort of UK adolescents

Objective 4: To investigate life course influences on use of sexual health services and participation in Chlamydia screening

Objective 5: To investigate the validity of serological testing as a measure of lifecourse exposure to genital Chlamydia trachomatis

Objectives 4 & 5 will use data obtained from linkage between ALSPAC and the PHE CTAD database.

Objective 6: to collect information on sexual networks to inform modelling of the transmission dynamics of common STI.

We will examine four broad hypotheses

a.) Risky sexual behaviour is an important independent risk factor for adverse sexual health outcomes.

b.) Other problem behaviours and outcomes (substance use, antisocial behaviour, low educational attainment) are important independent risk factors for risky sexual behaviour.

c.) There is a causal pathway from early life adversity and childhood problems to risky sexual behaviours, mediated through other problem behaviours in adolescence (substance use and/or antisocial behaviour and/or low educational attainment).

d.) Early life adversity, childhood problems, and other problem behaviours in adolescence are associated with STI infection.

A key issue in making inference about causal relations between variables such as those in the above diagram is clarification of the extent of socio-economic confounding. Many associations between different adverse health and social outcomes are likely to reflect their covariance with social disadvantage, rather than a distinct causal pathway. Some of these relations, however, may mediate the association between social disadvantage and poorer health. We will add, in turn, potential confounding variables related to (1) early life adversity and childhood socioeconomic circumstances; (2) childhood psychosocial problems; (3) other variables relating to problem behaviours in early adolescence; and finally (4) behavioural variables measured concurrently with sexual behaviour and sexual health outcomes. In considering the choice of possible confounding variables we will follow the approach of Hernan and colleagues who point out that strategies based on knowledge of the causal structure are more appropriate than those based on statistical strength in choice of confounding variables. We will use multilevel models as appropriate, to include repeated measures on each individual (for example to reflect changing socio-economic circumstances).