Service evaluation of an educational intervention to improve sexual health services in primary care implemented using a step-wedge design: analysis of chlamydia testing and diagnosis rate changes.

Background: Providing sexual health services in primary care is an essential step towards universal provision. However they are not offered consistently. We conducted a national pilot of an educational intervention to improve staff's skills and confidence to increase chlamydia testing rates and provide condoms with contraceptive information plus HIV testing according to national guidelines, known as 3Cs&HIV. The effectiveness of the pilot on chlamydia testing and diagnosis rates in general practice was evaluated.
Methods: The pilot was implemented using a step-wedge design over three phases during 2013 and 2014 in England. The intervention combined educational workshops with posters, testing performance feedback and continuous support. Chlamydia testing and diagnosis rates in participating general practices during the control and intervention periods were compared adjusting for seasonal trends in chlamydia testing and differences in practice size. Intervention effect modification was assessed for the following general practice characteristics: chlamydia testing rate compared to national median, number of general practice staff employed, payment for chlamydia screening, practice urban/rurality classification, and proximity to sexual health clinics.
Results: The 460 participating practices conducted 26,021 tests in the control period and 18,797 tests during the intervention period. Intention-to-treat analysis showed no change in the unadjusted median tests and diagnoses per month per practice after receiving training: 2.7 vs 2.7; 0.1 vs 0.1. Multivariable negative binomial regression analysis found no significant change in overall testing or diagnoses post-intervention (incidence rate ratio (IRR) 1.01, 95 % confidence interval (CI) 0.96-1.07, P = 0.72; 0.98 CI 0.84-1.15, P = 0.84, respectively). Stratified analysis showed testing increased significantly in practices where payments were in place prior to the intervention (IRR 2.12 CI 1.41-3.18, P < 0.001) and in practices with 6-15 staff (6-10 GPs IRR 1.35 (1.07-1.71), P = 0.012; 11-15 GPs IRR 1.37 (1.09-1.73), P = 0.007).
Conclusion: This national pilot of short educational training sessions found no overall effect on chlamydia testing in primary care. However, in certain sub-groups chlamydia testing rates increased due to the intervention. This demonstrates the importance of piloting and evaluating any service improvement intervention to assess the impact before widespread implementation, and the need for detailed understanding of local services in order to select effective interventions.