The impact of rapid point-of-care sti testing on service delivery outcomes: a controlled interrupted time series study

Background

Meeting the increasing sexual health needs of communities in the context of reduced funding is an ongoing challenge for sexual health services (SHS) in England. Providers are increasingly driven to find efficiencies while maintaining quality of care, such as through introducing point of care testing (POCT). In 2018, Unity Sexual Health implemented a rapid POCT, diagnosis and treatment service for chlamydia and gonorrhoea through running the Hologic Panther System at Unity instead of the microbiology laboratory. We aimed to evaluate the impact of the new system on service delivery and resource needs.

Methods

A controlled interrupted time series approach was applied to a census of electronic patient record data from Unity and a control site over a two-year period centred at the intervention. Segmented generalised linear regression models were used to assess the effect of the rapid service on five outcomes: time to test notification, staff capacity and rates of gonorrhoea culture swabs, follow up attendances, and examinations. This was done separately for male and female treatment pathways. The models estimated the immediate change in outcomes at the time of the intervention (step change), as well as any changes in long-term trends (slope change), both adjusted for background changes in the control site.

Results

After an initial increase in time to notification for males (step change: +3.6 days, 95%CI +1.7%,+5.5%), we identified a significant long-term decrease (slope change: -0.2 days per week, 95%CI -0.3,-0.2). A similar effect was observed for rates of urethral gonorrhoea swabs (step change: +89%, 95%CI +37.1%,+160.9%; slope change: -3.2% per week, 95%CI -4.3%,-2.1%). In contrast, models for females showed significant decreases for both step and slope changes for time to notification (step change: -2.1 days, 95%CI -4.5,0.3; slope change: -0.1 days per week, 95%CI -0.2,0.0) and rates of cervical gonorrhoea swabs (step change: -40.8%, 95%CI -61.6%,-8.8%; slope change: -6.1% per week, 95%CI -7.8%,-4.5%). All four of these models estimated more than a 70% intervention-related decrease in outcomes by the end of the study period. There was no evidence of intervention-related changes in rates of follow-up visits or examinations. Staff capacity in clinics was maintained across the study period.

Discussion

The clear improvement in two of the outcomes, the fact that other outcomes showed no detrimental effects and that the new service was implemented without additional financial support suggests an overall net benefit of the rapid testing service, supporting the case for more widespread rollout in SHS.