Evaluating completion rates between kiosk and smartphone computer-assisted self-interviewing at a sexual health clinic in Melbourne: a before-and-after implementation study.

Abstract

Background: Digital self-reporting tools are increasingly used in sexual health clinics for individuals to disclose sensitive information. In December 2023, the Melbourne Sexual Health Centre replaced its touchscreen kiosk system with a smartphone-enabled Digital Front Door platform for administering computer-assisted self-interviewing (CASI) to collect sexual history data before clinical consultation. This study aimed to evaluate how this change impacted response completeness.

Methods: We conducted a before-and-after study using routine data collected from clients aged ≥16 years who attended Melbourne Sexual Health Centre from January to June 2023 (kiosks period) and from January to June 2024 (smartphone period). We compared the CASI completeness (i.e. the proportion of questions for which each client provided a response, based on the total number they were expected to complete) between the two periods. Linear regression with generalised estimating equations was used to examine associations between completeness and individuals' characteristics.

Results: Among 56,095 eligible consultations (28,093 kiosks; 28,002 smartphones), the mean questionnaire completeness was significantly lower in the smartphone group (36.2%) compared with the kiosk group (54.0%; P < 0.0001). The proportion of clients who did not respond to any single question (i.e. 0% completeness) increased from 45.1% (12,676/28,093) during the kiosk period to 57.4% (16,065/28,002) during the smartphone period. Smartphone use was associated with a 15.31 percentage point reduction in completeness (β = -15.31, 95% CI: -16.12 to -14.49) after adjusting for other confounding factors. Reduced completeness was also independently associated with older age, birth outside Oceania, Indigenous identity and people living with HIV. First-time attendees demonstrated substantially higher completeness.

Conclusion: The transition from kiosks to smartphones for administering CASI was associated with a modest, but consistent, decline in data completeness. Incomplete CASI data may reduce clinic efficiency by requiring clinicians to collect missing information during consultations, posing challenges in high-demand settings. Further research is needed to assess whether response completeness improves as users become more familiar with the new digital system.