Implementing remote data collection methods in an address-based probability sample for a population-based sexual health survey (Britain’s fourth National Survey of Sexual Attitudes and Lifestyles, Natsal-4)

Background The National Surveys of Sexual Attitudes and Lifestyles (Natsal) are decennial probability sample surveys, representative of the British population and have previously been conducted face-to-face. The fourth Natsal survey was under development when the COVID-19 pandemic began in early 2020. Pilot fieldwork was delayed while remote data collection options were considered. Methods In 2020/21, we evaluated three potential remote data collection models by considering their ability to deliver Natsal’s key survey design features: sample quality, selection of one participant per household, boost samples, a long interview (~60 minutes) with complex routing, a self-completion element, biological sampling, data linkage and time-series maintenance. Our chosen design prioritised in-person interviewing, with a remote option where necessary or preferred, rather than participant-led choice or experimental allocation, to minimise differences with previous Natsal surveys and maximise response to biological sampling and data linkage. Two pilot studies were carried out (summer 2021 and spring 2022), incorporating the adapted study design and procedures. Results The requirements of the study at the time were best met by a model that used an interviewer-led doorstep contact approach, prioritised in-person interviewing and provided an option of remote data collection (by telephone or video). The implementation of a remote mode required significant adaptation to the existing study design. Remote interviews included interviewer-administered questions, an online self-completion questionnaire, biological sampling, and data linkage consent. Across the two pilot studies 30% (n=79/261) of interviews were conducted using remote methods. Response to biological sampling and data linkage consent were lower in remote interviews (34% and 61%, respectively) compared to face-to-face (56% and 80%, respectively). Mainstage Natsal-4 address-based probability sample fieldwork retained a remote option, but emphasised in-home interviewing as ...