ObjectiveTo assess whether decentralising colposcopy services to a primary care facility in inner-city Johannesburg, South Africa raises access to colposcopy.DesignBefore–after study comparing 2 years before and 2 years after decentralisation, using clinical records and laboratory data on cervical cytology and histology.Primary outcomeThe proportion of all women attending Hillbrow Community Health Centre (HCHC) with an abnormal Papanikolaou (Pap) smear who had a colposcopy post-decentralisation.SettingCharlotte Maxeke Johannesburg Academic Hospital (CMJAH) has provided colposcopy services for several decades. HCHC, located about 3 km away, began colposcopy services in 2014.ParticipantsWomen, aged above 18 years, who had a colposcopy for diagnosis and treatment of precancerous cervical lesions following a Pap smear, from 2012 to 2016 at CMJAH or HCHC.ResultsPre-decentralisation at CMJAH, 910 women had colposcopy (2012–2014). Post-decentralisation (2014–2016), 721 had colposcopy at CMJAH and 399 at HCHC, the decentralised facility. The number who had a Pap smear at HCHC and then a colposcopy rose threefold post-decentralisation (114 vs 350). Post-decentralisation, 43 women at HCHC were referred to CMJAH for colposcopy, compared with 114 pre-decentralisation. Post-decentralisation, 47.3% of women at CMJAH waited >6 months for colposcopy, while 35.5% did at HCHC (p<0.001). Across all three groups, 26.9%–30.3% of women had cervical intraepithelial neoplasia III lesions or carcinoma on colposcopy. The proportion of invalid specimens was similar at CMJAH and HCHC (1.8%–2.8%). Of 401 women who had an abnormal Pap smear at HCHC post-decentralisation, 267 had colposcopy (66.6%).ConclusionDecentralisation can decrease the time to colposcopy and reduce the workload of tertiary hospitals. Overall, more women accessed services. Colposcopy coverage at HCHC is higher than other sites, but could be further improved. Decentralisation did not appear to undermine the quality of services and this model could be extended to similar settings in South Africa and elsewhere.

IntroductionSouth Africa is moving towards achieving elimination of mother-to-child transmission (eMTCT) but gaps remain in eMTCT programmes. Documenting successful outcomes of health systems interventions to address these gaps could encourage similar initiatives in the future.MethodsWe describe the effectiveness of a Quality Improvement Project (QIP) to improve HIV retesting rates during pregnancy among women who had previously tested negative by redesigning the clinic process. Eight poorly-performing clinics were selected and compared with eight better-performing control clinics in a subdistrict in North West Province. Over nine months, root cause analysis and testing of change ideas using Plan-Do-Study-Act cycles were used to identify and refine interventions. Analysis of patient flow showed that women were referred for retesting following their nurse-driven antenatal visits, and many left without retesting as this would have further prolonged their visit. Processes were redesigned and standardised, where a counsellor was charged with retesting patients before antenatal consults. Staff were mentored on data collection and interpretation process. Quality improvement nurse advisors monitored indicators bi-weekly and adjusted interventions accordingly.ResultsRetesting in intervention clinics rose from 36% in the three months pre-intervention to full coverage at month nine. At the end of the study, retesting in intervention clinics was 20% higher than in controls. Retesting also increased in the subdistrict overall.ConclusionService coverage and overall impact of HIV programmes can be raised through care-process analysis that optimises patient flow, supported by targeted QI interventions. These QI methodologies may be effective elsewhere for identifying new HIV infections in pregnant/breastfeeding women, and possibly in other services.