Sexual Partners

We did all analyses with the complex survey functions of Stata (version 12.1) to incorporate weighting, clustering, and stratification of data. We present descriptive statistics by sex and age group. We used binary logistic regression to calculate age-adjusted odds ratios (aAORs) to investigate how reporting of key sexual behaviours and attitudes varied by three measures of social status: the National Statistics Socio-Economic Classification²⁸ (NS-SEC; individual level), education attainment defined according to school leaving age and academic qualifications obtained (individual level), and the Index of Multiple Deprivation²⁹ (area level). We then used ordinal and binary logistic regression to calculate aAORs to compare estimates between the three Natsal surveys for participants aged 16–44 years. We used Natsal-2 as the reference category to quantify the change between Natsal-2 and Natsal-3 and also to allow comparisons with Natsal-1. Finally, we obtained data from the three surveys to examine changes in the reporting of sexual practices with partners of the opposite sex and partner numbers by birth cohort and age at interview. We used an α of 0·05 in all analyses.

Full details of the methods used in Natsal-3 have been reported elsewhere.20 (link), 21 Briefly, we used a multistage, clustered, and stratified probability sample design. 1727 postcode sectors (geographical units used for sorting mail) throughout Britain were used as the primary sampling units and were randomly allocated to one of eight periods of fieldwork that took place between Sept 6, 2010, and Aug 31, 2012, with each period lasting about 3 months.
Within each primary sampling unit, 30 or 36 addresses were randomly selected and then assigned to interviewers from NatCen Social Research. To allow detailed exploration of behaviours in the age group at highest risk of some sexual health outcomes (eg, unplanned pregnancy and sexually transmitted infections), we oversampled individuals aged 16–34 years. We randomly allocated addresses to either the core sample (in which all individuals aged 16–74 years were eligible) or the boost sample (in which only individuals aged 16–34 years were eligible). Letters and leaflets giving background information about Natsal-3 were sent to sampled addresses before visits began.
Interviewers visited all sampled addresses, identified residents in the eligible age range, and randomly selected one individual to be invited to participate in the survey using a Kish grid technique. Participants then completed the survey in their own homes through a combination of face-to-face interviews with computer-assisted personal interview and a self-completion format with computer-assisted self-interview. Interviewers were present in the room while participants completed the computer-assisted self-interview and could provide assistance as necessary, but did not view responses. On completion of computer-assisted self-interviews, answers could not be accessed by interviewers. No names or other potentially identifying information was attached to the interviews. Interviews lasted about 1 h on average. Participants received a £15 gift voucher as a token of appreciation. An anonymised dataset will be deposited with the UK Data Archive, and the complete questionnaire and technical report will be available on the Natsal website on the day of publication.
These methods are broadly the same as those used in Natsal-1 and Natsal-2. However, in Natsal-1, paper was used (rather than computers) during face-to-face interviews and self-interview. Descriptions of the development stages of previous surveys are available elsewhere.9 , 10 , 11 (link), 22 (link), 23 , 24 (link) Most questions in Natsal-3 were identical to those in Natsal-1 and Natsal-2, including questions about age and circumstances of first sexual intercourse, sexual practices, experiences with sexual partners, details of most recent partners, and attitudes. Questions added in Natsal-3 included those about health problems that could affect sexual lifestyles, sexual function and satisfaction, a screen for depressive symptoms, non-volitional sex, and unplanned pregnancy.20 (link), 21 The questionnaire underwent thorough cognitive testing and piloting, as previously reported.²⁵
As in Natsal-1 and Natsal-2, we weighted Natsal-3 data to adjust for the unequal probabilities of selection in terms of age and the number of adults in the eligible age range at an address. After application of these selection weights, the Natsal-3 sample was broadly representative of the British population compared with 2011 Census figures,26 , 27 although men and London residents were slightly under-represented. Therefore, as in previous surveys, we also applied a non-response post-stratification weight to correct for differences in sex, age, and Government Office Region between the achieved sample and the 2011 Census (appendix).26 , 27
We compared data for participants aged 16–44 years in each survey. This age group was common to all three surveys. Information about variables that were compared was derived from identically worded questions. All three surveys had been weighted for differential selection probabilities. Natsal-1 was post-stratified to 1991 Census figures and Natsal-2 to 2001 Census figures, with procedures described for Natsal-3,20 (link), 21 which allowed us to make comparisons between the three surveys. However, there are minor differences from the weighting schemes used in previous reports.9 , 10 , 11 (link), 12 (link), 13 (link), 23
The Natsal-3 study was approved by the Oxfordshire Research Ethics Committee A (reference: 09/H0604/27). Participants provided oral informed consent for interviews.

HPTN 061 was a multi-site study conducted in Atlanta, Boston, Los Angeles, New York City, San Francisco, and Washington, DC to determine the feasibility and acceptability of a multi-component HIV prevention intervention for Black MSM.
From July 2009 to October 2010, men were recruited directly from the community or as sexual network partners referred by index participants. Index participants were: (1) HIV infected but unaware of their infection; (2) previously diagnosed with HIV infection but not receiving HIV care and having unprotected sex with partners of negative or unknown HIV status; or (3) HIV-uninfected. Recruitment methods were developed at each site and included: community outreach; engagement of key informants and local community-based groups; print advertising; and online strategies. At each site, enrollment of community-recruited HIV-uninfected participants was capped at 200 participants; enrollment of community-recruited participants who had a prior diagnosis of HIV infection and were already in care, or reported only having unprotected anal intercourse (UAI) with HIV-positive partners, was capped at 10 participants.
Men were eligible to participate in the study if they: self-identified as a man or male at birth and as Black, African American, Caribbean Black, or multiethnic Black; were at least 18 years old; reported ≥1 instance of UAI with a man in the prior six months; resided in the metropolitan area; did not plan to move away during the time of study participation; and provided informed consent for the study. Men were ineligible if they were enrolled in any other HIV interventional research study, had been a participant in an HIV vaccine trial, or were a community-recruited participant in a category that had already reached its enrollment cap. Prescreening to determine eligibility was performed either in person or over the telephone.

Participants (N = 277, 55% female, mean age = 36.8 years, SD = 12.21 years) were recruited through Prolific in May 2020, as part of a larger project examining wellbeing during the coronavirus pandemic. Most participants (n = 267, 96%) were in a romantic relationship with an opposite-sex partner. An additional 57 participants completed the protocol but were excluded based on pre-registered exclusion criteria (i.e., taking less than a third of the median response time, failing a data check item, stopping halfway through, showing invariant responding across 2 or more screens, or answering that they spoke English “not well” or “not at all; see https://aspredicted.org/blind.php?x=HSL_CMW).

The dependent variable for the study was reported HIV self-testing. Self-reports on HIVST have been used in previous studies, either with or without verification that the tests were actually conducted as reported [48 (link)–56 (link)]. Prior to data collection, the meaning of HIV self-test was explained to the participants and it included a test that allows people to take a HIV test and find out their result in their own home or other private locations. Participants were then requested to indicate whether they had ever undertaken a HIV self-test prior to the study.
The independent variables included socio-demographics: Age, place/country of birth, religion, education level, employment status, marital status and income. While it is possible that there were false reports, checks were put in place to reduce the possibility of wrong reporting by excluding responses with inconsistencies in all the cascades. The questionnaires were also pre-tested and refined before the main data collection.
Besides the above, the following risk factors data were collected: Identify as MSM/MSW, number of sexual partners, preferred sex position (top or bottom), condom use, condom use after alcohol/hard drugs and type of lubricant. The following variables were of importance in measuring the acceptability and factors that facilitates the uptake of HIVST: Frequency of HIV/AIDS test, use of protection or condoms with partner if HIV/AIDS negative, ever heard of HIV self-test, would prefer to use the Oral or Blood self-test kit, cost of the HIV self-test kits, ever tested for HIV, type of facility for the last HIV test, reason for most recent test, ever heard of window period, and have you ever experienced stigma from healthcare providers.

Individuals were asked about exposure to lifetime traumatic events including physical assault or abuse as an adult from an intimate partner and sexual assault or rape in adulthood, as described above. In addition, individuals also completed the Demographic and Health Survey (DHS) domestic violence module which captures four distinct domains of IPV: controlling behavior and emotional, sexual, and physical IPV [37 ]. This tool is a modified version of the Conflict Tactics Scale and asks respondents if they have experienced 15 separate behaviorally-specific types of IPV [38 (link)]. In our prior work, we noted discrepancies between reports of physical and sexual violence on the DHS domestic violence module and the modified version of the Life Events Checklist for DSM-5 used in the assessment of lifetime PTEs in this analysis [39 ]. Thus, we considered individuals who reported physical IPV with their most recent partner on the DHS domestic violence module, but did not report physical assault or abuse from a partner in adulthood on the PTE assessment to have experienced physical assault or abuse in adulthood from a partner in this analysis. Similarly, we considered individuals who reported sexual IPV with their most recent partner on the DHS domestic violence module but did not report sexual assault or rape as an adult on the PTE assessment to have experienced sexual assault or rape in adulthood.