Response Rates by Condom Use Measurement Type Among Sexually Active Young Black Men Who Have Sex With Women

Measuring condom use accurately is important to understanding sexual behaviors; however, there is still no consensus on the criterion standard of condom use measurement. In this study, authors aimed to investigate if nonresponse rates were higher when absolute or relative frequency condom use questions were used.

This study used Check It study data of Black men who have sex with women aged 15 to 26 years in New Orleans, LA. All men underwent a survey of sexual behavior, which included questions on condom use. From January 2017 to August 2019, participants were asked for the number of sex acts during which they had used a condom (absolute frequency question), and after August 2019 until March 2020, this was changed to response options of "never," "sometimes," or "always" (relative frequency question). Participants could refuse to answer the question; the rate of refusal was tabulated by question type to determine if nonresponse rates differed.

There were 1907 participants included in the study, with data on 2957 partnerships. Of those asked the relative frequency question, 92.7% responded compared with 86.1% who were asked the absolute frequency. In generalized estimating equation modeling, those asked the relative frequency question had 2.42 (95% confidence interval, 1.42-4.13; P < 0.001) greater odds of response than those asked the absolute frequency question.

Relative frequency questions may be more likely to elicit a response, in research among this population.

There is no current criterion standard for measuring condom use in surveys. This study demonstrated that asking participants about their condom use qualitatively (i.e., whether they use condoms "never," "sometimes," or "always"), as opposed to asking them to provide the number of sex acts during which they used a condom, was associated with lower rates of nonresponse. Investigators should consider using qualitative condom use questions to maximize response rates.