P350 What is the optimum method for collecting robust data to understand a nation's sexual health needs?. (14th July 2019)

• Record Type: Journal Article
• Authors: Clifton, Soazig; Field, Nigel; Prior, Gillian; Aldridge, Robert; Bonell, Chris; Copas, Andrew; Gibbs, Jo; Macdowall, Wendy; Mitchell, Kirstin; Tanton, Clare; Thomson, Nicholas; Unemo, Magnus; Sonnenberg, Pam; Mercer, Catherine
• Is Part Of: Sexually transmitted infections. Volume 95(2019)Supplement 1
• Page Start: A181
• Page End: A181
• Publication Date: 2019-07-14
• Subjects: sexual behavior; Sexually transmitted diseases -- Periodicals; HIV infections -- Periodicals; 616.951005
• DOI: 10.1136/sextrans-2019-sti.457
• Languages: English
• ISSNs: 1368-4973

Abstract

Abstract : Background: Accurate information on a nation's sexual health is essential to plan and evaluate services, inform prevention, and contribute to societal understanding. In Britain, sexual health data arise from surveillance systems, convenience surveys of key populations, and the decennial National Surveys of Sexual Attitudes and Lifestyles (Natsal). Natsal has employed 'gold-standard' population survey methods: probability sampling, trained fieldworkers conducting detailed computer-assisted-personal-interviewing, and biosampling. However, this approach is resource-intensive and limitations include declining response rates and concerns about non-response bias. In designing Natsal-4, we reviewed whether alternative methods could meet the needs of data-users and the wider community. Methods: We evaluated methods used by major UK general population surveys and sexual health surveys internationally. Key considerations were: general population representativeness; sample size; breadth and depth of information collected; data quality; biosampling; the possibility for sub-group 'boost' sampling, and data linkage. Results: Five alternative methods were assessed (1) random-digit dialling phone surveys: considered unsuitable due to inadequate sample frame and response rate; (2) inviting participants from existing probability surveys to a follow-up sexual health interview: unsuitable because of additional non-response bias, difficulty achieving required sample size, and minimal cost-saving; (3) adding a sexual health module to existing probability survey(s), and (4) conducting a probability survey with fieldworker-selected individuals asked to self-complete a sexual health web-survey: both considered unsuitable due to much-reduced questionnaire; (5) 'web-first' mixed-mode survey, involving postal invitations to complete a web-survey with non-responding addresses followed-up by post and/or fieldworker visit: unsuitable due to concerns about response rate, unmeasurable and measurable response bias, and selection bias. Conclusion: Given major drawbacks of the alternatives examined, the design used for previous waves of Natsal is still considered the best option for achieving a representative sample, enabling detailed data collection, enhancing survey data with biological and routine data, and retaining Natsal's time-series; together maximising Natsal's utility and impact. Disclosure: No significant relationships.