Diaries beat monthly surveys for tracking PrEP and condom use in Hong Kong men

When it comes to measuring how men who have sex with men actually use HIV pre-exposure prophylaxis and condoms, the tool researchers choose may change the answer they get. A study conducted in Hong Kong, published as open access research in AIDS Research and Therapy, compared two ways of collecting the same behavioural information over a full year: a short-recall diary filled in close to the moment, and a monthly survey completed retrospectively. The conclusion was pointed. Monthly surveys appeared to over-estimate on-demand PrEP use, and the authors argue that where resources allow, the diary is the preferable instrument for variables that require continuous monitoring.

The work formed part of an HIV PrEP implementation study involving adult MSM in Hong Kong. Over a one-year period, each participant completed both a short-recall diary and monthly surveys that looked back over the preceding weeks. The research team, led by Rui Zhang with colleagues including Ngai Sze Wong, Tsz Ho Kwan, Grace Chung-Yan Lui, Denise Pui-Chung Chan, Krystal Chi-Kei Lee and Shui Shan Lee, set out to test how robust the short-recall diary was against the summarised monthly survey, which remains the more common approach in real-world programme settings even though a diary is considered ideal for capturing daily sexual activity and drug taking.

The scale of the dataset gives the comparison its weight. Between January 2020 and June 2023, the study collected 90,678 diary records and 2,022 monthly records from 287 MSM. For the head-to-head analysis, the investigators used 51,359 diary records and 1,988 monthly records from 268 participants. Consistency between the two methods in measuring PrEP use patterns and condom use frequency, along with compliance with survey completion, were the primary outcomes, quantified using the kappa statistic, a standard measure of agreement between two measurement approaches.

The first striking finding concerned compliance rather than measurement. Participants were far more likely to keep up the short-recall diary than the monthly survey: 86 percent versus 36 percent, a difference the authors report as statistically significant with P < 0.001. In other words, the method that demanded more frequent, smaller contributions was actually sustained better than the one that asked for a consolidated monthly recall.

The two instruments also told different stories about the same behaviour. Compared with the short-recall diary, monthly surveys reported a significantly higher proportion of on-demand PrEP use, 27 percent against 9 percent. Condom use showed the same direction of discrepancy: monthly surveys recorded inconsistent condom use at 52 percent versus 43 percent, and consistent condom use at 10 percent versus 1 percent, all differences significant at P < 0.001. The authors attribute the PrEP gap at least partly to the complexity of the 2-1-1 regimen, the on-demand dosing schedule whose event-timed structure is evidently harder to reconstruct accurately from memory a month later than to log in a diary at the time.

Agreement between the methods was not static across follow-up. The kappa value for condom use significantly decreased as the number of follow-up visits increased, with P = 0.022, suggesting that the two instruments drifted apart the longer participants were followed. One clinical variable was associated with better agreement: MSM who were diagnosed with syphilis were more likely to achieve a kappa value of 1, indicating perfect consistency between the two methods, with an adjusted odds ratio of 1.93 and a 95 percent confidence interval of 1.03 to 3.62. The study reports this as an association within its own cohort; it does not establish a causal mechanism, and the underlying reason for the link is not explained in the reported findings.

The study’s design and scope deserve emphasis when interpreting these numbers. This is a single implementation study among adult MSM in Hong Kong, running from 2020 to 2023, comparing two self-report instruments within the same participants rather than validating either against an objective standard. Both measures are self-reported, so the comparison shows where the two recall methods diverge, not which one matches true behaviour. The authors’ conclusion is correspondingly measured: monthly surveys might over-estimate on-demand PrEP use because of the complexity of the 2-1-1 regimen, and if resources are available, a diary is the preferable tool for measuring variables that require continuous monitoring.

For HIV prevention programmes and researchers, the practical message is that measurement method is not neutral. A monthly recall survey, the cheaper and more common option, appears to shift reported PrEP dosing patterns and condom use in consistent, predictable directions, inflating both on-demand PrEP use and reported condom use relative to near-real-time diaries. As PrEP programmes scale up and rely on behavioural data to judge adherence and risk, the choice between a short-recall diary and a summarised survey may materially change reported outcomes, and the diary’s higher completion rate in this cohort adds a further argument in its favour.

Subject of Research: Comparing short-recall diaries with monthly surveys for measuring PrEP and condom use among men who have sex with men in Hong Kong.

Article Title: Measuring the longitudinal HIV pre-exposure prophylaxis and condom usage pattern in men who have sex with men: short-call diary versus monthly survey

Article References: Zhang, R., Wong, N. S., Kwan, T. H., Lui, G. C.-Y., Chan, D. P.-C., Lee, K. C.-K., & Lee, S. S. (2026). Measuring the longitudinal HIV pre-exposure prophylaxis and condom usage pattern in men who have sex with men: short-call diary versus monthly survey. AIDS Research and Therapy. https://doi.org/10.1186/s12981-026-00970-6

Image Credits: AI Generated

DOI: 10.1186/s12981-026-00970-6

Keywords: HIV, pre-exposure prophylaxis, PrEP, condom use, men who have sex with men, diary, survey methodology, Hong Kong, 2-1-1 regimen, kappa agreement, self-report bias, implementation study