Abstract
BACKGROUND: Engagement with and the potential impact of web-based interventions is often studied by tracking user behavior with web analytics. These metrics do provide insights into how users behave, but not why they behave as such. OBJECTIVE: This paper demonstrates how a mixed methods approach consisting of (1) a theoretical analysis of intended use, (2) a subsequent analysis of actual use, and (3) an exploration of user perceptions can provide insights into engagement with and potential impact of web-based interventions. This paper focuses on the exploration of user perceptions, using the chlamydia page of the Dutch sexual health intervention, Sense.info, as a demonstration case. This prevention-focused platform serves as the main source of sexual and reproductive health information (and care if needed) for young people aged 12-25 years in the Netherlands. METHODS: First, acyclic behavior change diagrams were used to theoretically analyze the intended use of the chlamydia page. Acyclic behavior change diagrams display how behavior change principles are applied in an intervention and which subbehaviors and target behaviors are (aimed to be) influenced. This analysis indicated that one of the main aims of the page is to motivate sexually transmitted infection (STI) testing. Second, the actual use of the chlamydia page was analyzed with the web analytics tool Matomo. Despite the page's aim of promoting STI testing, a relatively small percentage (n=4948, 14%) of the 35,347 transfers from this page were to the STI testing page. Based on these two phases, preliminary assumptions about use and impact were formulated. Third, to further explore these assumptions, a study combining the think-aloud method and semistructured interviews was executed with 15 young individuals aged 16-25 (mean 20, SD 2.5) years. Template analysis was used to analyze interview transcripts. RESULTS: Participants found the information on the Sense.info chlamydia page reliable and would visit it mostly for self-diagnosis purposes if they experienced potential STI symptoms. A perceived facilitator for STI testing was the possibility to learn about the symptoms and consequences of chlamydia through the page. Barriers included an easily overlooked link to the STI testing page and the use of language not meeting the needs of participants. Participants offered suggestions for lowering the threshold for STI testing. CONCLUSIONS: The mixed methods approach used provided detailed insights into the engagement with and potential impact of the Sense.info chlamydia page, as well as strategies to further engage end users and increase the potential impact of the page. We conclude that this approach, which triangulates findings from theoretical analysis with web analytics and a think-aloud study combined with semistructured interviews, may also have potential for the evaluation of web-based interventions in general.