Validity and reliability in reporting sexual partners and condom use in a Swiss population survey

Abstract

Objectives To examine the validity and reliability of indicators of sexual behaviour and condom use in annual telephone surveys (n=2800) of the general population aged 17 to 45 for the evaluation of AIDS prevention in Switzerland. Methods A test-retest study with additional focused interviews was conducted on a subsample (n=138) of the respondents aged 17 to 22 years. Results The subsample included more French speaking respondents (OR: 1.7, CI: 1.1-2.5) and more people in a stable relationship (OR: 2.2, CI: 1.5-3-3) than the initial sample but did not differ in any other way, although no data is available on their attitudes towards sex. The reliability of the indicators considered was high: number of lifetime, casual sex partners in the last 6 months and condom use with them, acquisition of a new steady partner during the year and condom use with this partner, condom use at last intercourse. However, the focused interviews raised questions about the validity of some of these indicators, presumably due to imprecise wording of the questionnaire items. Among sexually active respondents, 12.5% (95% CI: 4.7-25.5) of the men included non-penetrative sex in the definition of 'sexual intercourse', but only 1.9% (95% CI: 0.1-10.3) of the women. The propensity for men of counting acts or partners with whom no penetration had taken place in the total reported sex acts or partners was not significantly associated with any socio-demographic variables. In addition, among the 15 respondents who had reported consistent condom use with casual sex partners at interview, 40% (95% CI: 16.3-67.7) admitted at reinterview that sometimes they also had unprotected sex. Conclusions The reliability of reports on sexual behaviour and condom use in this Swiss evaluation survey is good. The indicators derived from the annual surveys are robust measures and the monitoring of trends seems to be based on reliable measurement. However, more research is required on the validity of the data.