Evaluation of Sexual Orientation Items in Population Health Surveys Among Canadians: A Mixed Methods Approach

A single item sexual orientation question has been asked in Statistics Canada health surveys to identify health inequalities in Canada. Using a mixed methods study with convenience sample of Canadians, we evaluated this question in comparison to a set of US questions that includes domains of sexual identity, behaviour and attraction. The single item question had an 85.8% sensitivity and 100% specificity in capturing sexual minorities from the broadest definition (n = 311). Chance-corrected agreements with sexual identity, 12 months and life time sexual behaviour were 0.89, 0.39 and 0.48 respectively. Both trans and cisgender people revealed that there were problems with the question, although trans people were more likely to be unclassifiable by the single item question. Findings suggest the need to further refine this question for more accurate identification patterns of health in Canada.