Sampling frame for sexual minorities in public health research

Extract

Dear Editor,

Hickson et al.¹ use a large community-recruited sample to describe differences in mental health indicators within men who identified as gay or bisexual. This is much needed research, as there is still a relatively underdeveloped evidence base describing inequalities in health for sexual minority groups, compared to the wider population and when considering differences within sexual minority groups as done here. The results show that we need to better understand variation within sexual minority groups, in addition to how minority groups differ from the majority groups in the wider population.

As the authors acknowledge, most population health research with sexual minorities has used convenience sampling as done here. We agree that general population surveys produce small numbers of sexual minorities, but disagree with them on two points. First, the reason why sexual minority health research has historically used convenience samples was not because ‘representative general population surveys only recruit a small absolute number of people in sexual minorities’—it is that such data were simply not available. This has now started to change. Since 2008 surveys of population health and other kinds of surveys in the UK have included measures of sexual orientation, which can be defined in terms of identity, behaviour and/or attraction.² These data are freely available to the research community stored at the UK Data Service. Second, we disagree that there is ‘no sampling frame for sexual minorities’. Population health surveys are already recording sexual orientation identity using a standardized question recommended by the Office of National Statistics (ONS).³ Other surveys have recorded same-sex behaviour and same-sex attraction.