Menu
Menu
Menu
Menu
Background: Health systems that deliver prevention of mother to child transmission (PMTCT) services in low and middle income countries continue to underperform, resulting in thousands of unnecessary HIV infections of newborns each year. We used a combination of approaches to health systems strengthening to reduce transmission of HIV from mother to infant in a multi-facility public health system in South Africa. Methodology/Principal Findings: All primary care sites and specialized birthing centers in a resource constrained sub-district of Cape Metro District, South Africa, were enrolled in a quality improvement (QI) programme. All pregnant women receiving antenatal, intrapartum and postnatal infant care in the sub-district between January 2006 and March 2009 were included in the intervention that had a prototype-innovation phase and a rapid spread phase. System changes were introduced to help frontline healthcare workers to identify and improve performance gaps at each step of the PMTCT pathway. Improvement was facilitated and spread through the use of a Breakthrough Series Collaborative that accelerated learning and the spread of successful changes. Protocol changes and additional resources were introduced by provincial and municipal government. The proportion of HIV-exposed infants testing positive declined from 7.6% to 5%. Key intermediate PMTCT processes improved (antenatal AZT increased from 74% to 86%, PMTCT clients on HAART at the time of labour increased from 10% to 25%, intrapartum AZT increased from 43% to 84%, and postnatal HIV testing from 79% to 95%) compared to baseline. Conclusions/Significance: System improvement methods, protocol changes and addition/reallocation of resources contributed to improved PMTCT processes and outcomes in a resource constrained setting. The intervention requires a clear design, leadership buy-in, building local capacity to use systems improvement methods, and a reliable data system. A systems improvement approach offers a much needed approach to rapidly improve under-performing PMTCT implementation programmes at scale in sub-Saharan Africa. © 2010 Youngleson et al.
The ethics of undertaking the systems improvement intervention and analysis and dissemination of the findings were evaluated by the Office of Human Research Ethics of the University of North Carolina at Chapel Hill. The Office determined that, given the nature of the intervention and the de-identified characteristics of the data, the study was exempted from further IRB review. Written consent was not obtained from individual patients as this was a population based health systems intervention to improve existing guidelines and protocols, and applied to the patient population without exception or exclusion. The ESD population (∼360,000) includes urban, peri-urban and rural communities. The birth rate is 20/1000 per annum (∼7,200/yr), and antenatal HIV prevalence is 17.4% (∼1,250 HIV exposed infants annually) [11]. The in-facility delivery rate in the district is high (92.9% in 2006 and 99.9% in 2008). 94% of HIV positive women on the PMTCT programme elected to exclusively formula feed from birth [12]. The majority of people living in ESD receive their medical care in the public facilities. In 2006, eight of the 15 primary care clinics offered antenatal care and all offered PMTCT infant follow up including PCR testing. Obstetric services are delivered in two birthing units. Although a proportion of women living in ESD receive antenatal care and deliver in adjacent sub-districts, the majority of women access post-natal infant follow up in ESD. PMTCT data was routinely collected in registers at the facilities and collated each month on paper-based forms that were submitted to the sub-district office for entry into electronic collation tables. These tables have built-in validation formulae to flag if numerators exceed denominators. Completeness and consistency of the data is actively monitored. Infant mortality data is compiled from death notifications by the Department of Home Affairs. During the period of the intervention there was a two-year delay before these data became available to the sub-district. Infant mortality data from 2006 were therefore only available for review in 2008. Maternal mortality data were not available for the sub-district during the study period. The health systems intervention was a partnership between the Western Cape Provincial Department of Health and Cape Town Municipality City Health departments and the Institute for Healthcare Improvement (IHI). Working closely with the health departments, IHI provided a systems improvement design and a quality improvement (QI) expert who introduced QI methods to the sub-district facilities, trained managers in QI methods and collaborated closely with senior health department officials on project design and execution. The project was implemented in two phases: a prototype phase in a sub-section of the sub-district (seven primary care clinics and the two birthing units) that formed a self-contained inter-facility PMTCT care and referral system, followed by a spread phase in which the systems improvement approach were disseminated through all 17 PMTCT-linked health facilities in ESD. The health systems intervention had three concurrent components: a quality improvement framework, policy and protocol changes, and targeted resource additions. The improvement project was structured as a “Breakthrough Series (BTS) Collaborative”, a QI model that promotes change simultaneously across large parts of the system [13]. All healthcare facilities in the selected referring systems were linked into a learning network to accelerate peer-to-peer learning through setting common aims and goals, systematically improving the reliability of HIV care using QI methods, and sharing successful strategies for improving PMTCT. The improvement project was time limited (innovation phase of 21 months, and spread phase of 18 months). At six-monthly intervals, representatives from the clinical sites gathered at workshops (“Learning Sessions”) to learn how to use specific QI methods based on the Model for Improvement and the PDSA (Plan-Do-Study-Act) cycle [14]–[16]. These methods included setting aims, process mapping of the PMTCT care pathway, using routine data to identify of gaps in care, root cause analysis of these gaps, selection of change ideas to close specific gaps, and use of rapid-cycle change iterative methods to test improvement ideas. Between these learning sessions, clinics formed multidisciplinary improvement teams that applied these QI skills at individual facilities. In addition, project staff provided on-site mentoring in the use of QI methods. The intensity of support decreased over time from fortnightly on-site mentoring in the prototype phase to a more leveraged approach through monthly sub-district meetings in the spread phase as managers became familiar with the methodology. Facility-based improvement activities resulted in the development of local solutions to local problems. These solutions were spread to other sites through extension by the project staff, Department of Health (DOH) managers and through routine monthly meetings and the six-monthly BTS Collaborative Learning Sessions described above. Successful changes developed in the prototype phase were complied into a “Change Package” for testing and adaption by other facilities as the project spread (Table 1). Successful change strategies included: maximizing the use of existing resources through strategic redeployment of staff and services, reducing duplication, improving information transfer, ‘bundling’ interventions to fully utilize each of the patient’s visits, and developing patient-centered approaches such as minimizing the number of clinic visits, improving geographical access to care and introducing psychologically supportive changes to retain patients in care. In addition to QI methods, strategic additional resources and policy changes were introduced by the sub-district and province during the study period. The sub-district increased the number of antenatal clinics providing antenatal care through introducing Basic Antenatal Care (BANC) to two additional clinics using existing staff [17]. In an effort to increase local access to HAART, three new ARV service points were introduced into primary care facilities during the project, supplementing the two existing hospital-based ARV facilities. These new primary care ARV clinics used a QI approach to boost ARV capacity by redeploying existing staff from the subdistrict hospitals. Later these facilities received added staff from the DOH. Two major changes to the PMTCT protocol were introduced by the DOH during the study period. The timing of the introduction of AZT prophylaxis during pregnancy was advanced from 32 weeks gestation to 28 weeks gestation, and the age of HIV testing of infants was decreased from 14 weeks to 6 weeks. Both protocol changes were introduced simultaneously with uptake occurring over a period of a few months in the early part of the spread phase of the project (early 2008). From January 2006 to March 2009, facility staff and managers from the different branches of the DOH participated in this project attending a total of six Learning Session workshops in the BTS Collaborative design. Between these workshops, district managers and program coordinators analyzed performance data which was then used to assist frontline staff to test change ideas. Improvement ideas from the “change package” were also actively spread at monthly sub-district staff meetings and by program coordinators during routine clinic visits. Senior managers were kept informed through monthly project reports. A steering committee of district and DOH managers guided the process, sanctioned the spread of changes, and intervened to removed obstacles to improving PMTCT that were identified by the project. All data were obtained from routine monthly sub-district PMTCT data collected at antenatal, labour ward and infant follow up clinics and collated by the Municipal and Provincial DOH. Data measuring the performance of PMTCT processes of care and infant HIV infection rates were tracked over time. We used run chart and Shewhart control chart analysis (Chartrunner, Productivity-Quality Systems, Inc, and QI Charts, Scoville Associates, Chapel Hill, NC) to determine the effect of system changes on the processes and outcomes of PMTCT care as changes were made over time [18]. To determine whether the improvements in performance were statistically significant, we analyzed data variation and trends, using a time series analysis [19]–[21]. We used categorical analysis (chi-square and unpaired t-test) to compare change in performance before and after the interventions, or between different sub-districts.
