Evaluation of an internet-accessed STI testing (e-STI testing) and results service in two London boroughs

Background: Internet-accessed STI (e-STI) testing is recommended in England to expand access to STI testing services, particularly among high-risk groups. Yet the evidence for the effectiveness of this testing modality is limited. This thesis aimed to address this evidence gap, by evaluating the effects of an e-STI testing and results service (SH:24) on STI diagnoses and STI testing uptake, when delivered alongside usual care in two London boroughs. Theoretical premise of thesis: In line with theory-driven approaches to evaluation, I developed an explanatory framework outlining hypothesised change processes that may be triggered by the intervention. I adopted a critical realist conceptualisation of causal mechanisms and I drew on the ‘candidacy’ lens to conceptualise change processes in relation to access and utilisation of STI testing services. I then used this framework to situate the evaluation findings. Methods: This evaluation was based on a randomised controlled trial (RCT). 2,072 individuals aged 16–30 years, resident in Lambeth and Southwark, with at least one sexual partner in the previous 12 months and willing to take an STI test, were recruited in community settings. Participants were allocated to an e-STI testing service (intervention) or to a website with signposting to local sexual health clinics (control). The e-STI testing service provided postal self-sampling kits for chlamydia, gonorrhoea, HIV and syphilis. Results were delivered via text message or phone, and participants were signposted to local clinics for confirmatory testing and treatment as necessary. Results: The published trial results are as follows: 1,031 participants in the intervention group and 1,032 control group were included in the analyses. At 6 weeks, 50.0% of the intervention group completed an STI test compared to 26.6% in the control group (relative risk, RR 1.87, 95% confidence interval 1.63 to 2.15, p<0.001). 2.8% of the intervention v 1.4% in the control were diagnosed with an STI (RR 2.10, 95% confidence interval 0.94 to 4.70, p=0.079). The effect on the proportion of participants treated was 1.1% in the intervention v 0.7% in the control (RR 1.72, 95% confidence interval 0.71 to 4.16, p=0.231). Secondary analyses (unpublished at the time of writing) demonstrated that the intervention was effective for uptake of STI testing among a subsample of participants who had never previously tested for STIs. Conclusions: The findings lend weight to national policies in England, which promote e-STI testing as a means to increase utilisation of STI testing services, particularly among groups who do not use conventional services. While the results support the candidacy model’s processual framing of health care utilisation, further research is required to understand how the construct of candidacy is recognised and negotiated within a digital service environment by different socio-demographic groups. In addition, larger trials are needed to assess outcomes later in the cascade of care.