In 2011, Malawi became the first country to adopt a strategy for initiating treatment for all HIV-infected pregnant women, regardless of CD4 cell count, which later became known as the World Health Organization's (WHO) Option B+. Option B+ entails provision of lifelong antiretroviral therapy (ART) for all HIV+ pregnant and breastfeeding women to promote prevention of mother-to-child-transmission (PMTCT) of HIV. By 2013, 75% of HIV+ pregnant women were either already on ART or were being initiated on ART at their first antenatal care (ANC) visit, an increase from 40% coverage in 2010.1,2 Despite this remarkable increase in ART coverage, in 2013, Malawi's Ministry of Health (MoH) reported that the proportion of HIV affected mothers and children retained in care was 73% for mothers and 68% for children at 12-months, and 71% for mothers and 61% for children at 24-months.2 Suboptimal retention of MIPs in care threatens the goal to eliminate MTCT by 2020. These findings highlight the need for innovative interventions to improve retention of women and their infants throughout the PMTCT cascade.
Globally, different approaches have been explored to improve retention in care in-line with the WHO Strategic Vision for PMTCT elaborated in 2010.3 Successful approaches include male partner involvement, peer mentoring, mobile phone appointment reminders, support from community health workers, conditional cash transfers, efficient referral, and integration of health services.4–6
In particular, the concept of integrating care across service delivery areas and population groups offers promise for improving retention in care. Interventions integrating family planning and ART services into sexual and reproductive health services,7–9 as well as those where child services have been combined with PMTCT,10–12 have yielded positive outcomes. Integration of ANC and HIV care has resulted in increased enrollment rates, whereas integration of child health services with PMTCT has reduced lost-to-follow-up rates of babies at 6 weeks after delivery.13,14 However, acceptability and challenges of implementing integrated services, specifically through the end of the breastfeeding period, have not been evaluated in the context of Option B+.
The Promoting Retention among Infants and Mothers Effectively (PRIME) cluster randomized control trial was conducted in Malawi beginning in 2013 to assess the effectiveness of integrating care for HIV+ mothers and their infants within a single clinic (called MIP clinics) as a strategy to promote retention of mothers and infants throughout the PMTCT cascade. This article describes the implementation process and challenges experienced by the 20 health facilities that implemented the MIP clinic model.
The PRIME study compared 3 models of health care delivery: MIP clinics, MIP clinics plus short text messaging service (SMS) alerts to community health volunteers to facilitate tracing of defaulting MIPs, and the standard of care. This evaluation combines analysis of both intervention arms (MIP clinics and MIP clinics plus SMS) given that the integrated service model of the intervention was the same in both arms. Details of the PRIME study methodology, including study site selection and results, have been described.15
Between May 2013 and August 2016, 20 health facilities implemented MIP clinics as part of the PRIME trial. These facilities were drawn from the predominantly rural districts of Salima and Mangochi, in the Central and South regions of Malawi, respectively.
Implementation of the MIP Intervention
MIP clinics were designed to offer integrated services to HIV+ mothers and their infants on prescheduled days of the week. These services included both HIV and routine Maternal, Neonatal and Child Health (MNCH) services, which were provided on the same day and in the same physical location (ie, the same room within the clinic, whenever possible). In comparison, the standard of care offered the same HIV and MNCH services in separate locations, often at different times and on different days of the week.
Sites were assessed at baseline for infrastructure, availability of trained staff, and clinic flow to understand the support required to start implementing MIP clinics. No additional staff was deployed to support MIP clinic operations. At least 2 health care workers (HCWs) per site were trained on the MIP protocol, Option B+, data recording and management, and good clinical practices. Before the start of the study, communities, village leaders, and District Health Offices were sensitized about the possible benefits of integrated MIP clinics.
The MIP clinic schedule was determined by the facility staff based on staff workload, facility infrastructure, and expected patient volumes. Accordingly, MIP clinics could operate twice a week, weekly, twice a month, or monthly. Consequently, MIP clinics did not completely replace preexisting HIV and MNCH services but were offered by facility staff as additional, integrated clinics. MIP clinics operated during both the antenatal and postpartum period. Antenatal MIP clinics focused on integrating ANC and ART services for pregnant women, whereas postpartum MIP clinics additionally integrated MNCH services for the HIV-exposed infants. The intent was to have pregnant and postpartum women and their infants attend MIP clinics to allow for a cohesive and supportive environment for HIV+ women from the start of their pregnancy. In some facilities, staff rotation dictated how services were organized and delivered at the MIP clinics.
Mentorship and support to clinic staff was provided by one trained midwife who had previously worked in district health facilities. The mentor used a checklist to (1) monitor the competencies of HCWs to implement integrated services; (2) document clinic organization and flow; and (3) verify whether mothers enrolled in the study were correctly identified and linked to MIP clinics. Feedback was provided to staff at the end of each mentorship visit, and a plan to improve areas of service delivery was mutually agreed.
Clinics were initially visited quarterly by the mentor in conjunction with government supervisory schedules. However, mentorship visits regularly highlighted the need for additional support at all sites. Therefore, in January 2015, one additional nurse mentor was hired, and a revised mentorship checklist incorporating a scoring system was introduced. Based on the facility's score, the frequency of mentorship visits was adjusted to allow increased support to sites with higher need. The mentorship score recorded select indicators of adherence to the MIP protocol to evaluate HCWs' performance in delivering integrated services. Indicators included the following: (1) proper and regular use of the appointment diary for mothers; (2) record of MIP visit attendance; (3) accurate and regular completion of maternal ART cards; (4) accurate and regular completion of infant HIV Care Clinic cards; and (5) initiation of tracing of defaulting participants. One point was given for each indicator, creating a mentorship score ranging from 0 to 5, with higher scores reflecting better performance.
Narrative mentorship reports that provided qualitative data from observations of mentors, and their interviews with HCWs, were produced twice in both 2013 and 2014, and once in both 2015 and 2016, because of the higher workload of mentors in 2015 and 2016.
Visit Alignment Data Collection
MIP clinic dates, follow-up appointment dates provided to mothers and infants by HCWs, and actual attendance dates for mothers and infants were recorded as part of MIP clinic implementation. Two indicators on visit alignment were created to develop proxy measures of staff willingness and ability to implement MIP services, measure participants' exposure to the MIP clinic intervention, and clinics' adherence to the MIP protocol. First, follow-up appointment dates given by HCWs to mothers and infants were compared with the health facility's MIP clinic dates. Second, actual attendance dates for mothers and infants were compared with MIP clinic dates. In both cases, if the dates matched, they were considered aligned, as illustrated in Figure 1.
Overall proportions were calculated for each indicator using the number of visits aligned in the numerator and the total number of all participants' visits in the denominator. These 2 alignment indicators were further analyzed over the 4 calendar years of the intervention period by log binomial regression using a generalized estimated equation that accounted for repeat clinic visits by participants, clustering of health facilities, and semiurban/rural stratification, and summarized by relative risks (RRs) and 95% confidence intervals (CIs). Regression analyses also included the mothers' antenatal or postnatal period as a covariate. Finally, on an individual level, proportions of each indicator were calculated out of the participants' total number of visits. These individual proportions were then summarized and displayed by facility in box plots that display the range, median, and interquartile range (IQR).
The PRIME study received ethical approvals from the University of Malawi's College of Medicine Research and Ethics Committee and the WHO Ethics Review Committee before implementation and for study amendments during implementation. Signed written informed consent was obtained from all study participants on enrollment.
Facility Baseline Characteristics
Table 1 displays facility-level baseline characteristics. Just over half of the facilities were classified as semiurban [semiurban was defined by the study team as facilities that were easier to access (ie, on tarmac roads or near trading posts)], and most were primary health care facilities. The Government of Malawi and the Christian Health Association of Malawi each owned and operated half of the study facilities. Most sites had power and water available (65% and 75%, respectively). More than half of the facilities decided to implement the MIP clinics weekly. Facilities had a median of 0 medical or clinical officers, 2.5 nurses, and 1 medical assistant on staff. All sites had a low ratio of facility-based health workers (clinical or medical officers, medical assistants, and nurses) for the served population, with a median of 2.1 per 10,000 (1.2–4.7). In comparison, the ratio of community-based health workers (health surveillance assistants and community HCWs) for the served population was higher with a median of 15.2 per 10,000 (6.1–33.8).
Mentorship Support for MIP Implementation
The number of mentorship visits per facility increased from once a quarter in 2013 to at least once a month in 2015, then reduced back to once a quarter for the remainder of the study in 2016. In total, 48, 67, 255, and 81 visits were conducted in 2013, 2014, 2015, and 2016, respectively. The mean (SD) mentorship score across all sites between January 2015 and June 2016 was 3.8 (1.4) of 5. Figure 2 shows that on average, all sites improved their mentorship scores between 2015 and 2016 (χ2 test for trend P < 0.001). The mean (SD) performance score for the 20 health facilities increased from 3.0 (1.5) in the first quarter of 2015 to 4.15 (1.2) in the first quarter of 2016. However, there was a clear division in mentorship scores, with 5 facilities performing very poorly and 5 other facilities performing very well from 2015 onward.
Based on routine mentorship reports, provision for integrated HIV and MNCH services during MIP clinics were not consistently implemented within the facilities, despite systematic mentorship support provided. Furthermore, one facility failed to implement the MIP clinic at all because of the introduction of an alternate model of care by a partner after the study began. HCWs commonly reported experiencing various challenges including high workload, perceptions of increased workload due to the MIP clinic, lack of routine supervision from the MoH District Health Teams, and poor collaboration among HCWs themselves, especially between nurses and lay cadres. Other challenges documented by site mentors were expectations from HCWs to receive incentives to adopt the MIP clinic approach, lack of commitment or motivation of HCWs, and negative attitudes among staff such as reluctance to change or unwillingness to support one another. There were also high turnover of staff, which led to constant training needs of other staff. However, there were also issues beyond the control of the mentors, which adversely affected implementation of MIP clinics, such as infrastructure limitations in terms of inadequate space to operate MIP clinics as “one-stop shops,” lacking or broken equipment, and stock outs of key commodities. Further details of the challenges identified by mentors during their scheduled visits are available (see table, Supplemental Digital Content 1, http://links.lww.com/QAI/A995).
Visit Alignment Data
Overall Visit Alignment
The overall proportion of maternal follow-up appointment dates that aligned to scheduled MIP clinic dates was 47.0%, with the greatest improvement seen between 2013 and 2015 (RR 1.6, 95% CI: 1.4 to 1.9) (Table 2). The proportion of infant appointment dates that aligned with MIP clinics was 55.9%, also with the greatest improvement between 2013 and 2015 (RR 1.4, 95% CI: 1.2 to 1.6). The overall proportion of actual maternal attendance dates that aligned to MIP clinic dates was 41.7%, with significant improvement over both the intervention period and in the postnatal vs. antenatal period. In 2016, mothers' visits aligned to MIP clinics twice as often as in 2013 (RR = 2.1, 95% CI: 1.7 to 2.5) and around 50% more frequently in the postnatal period than the antenatal period (RR = 1.5, 95% CI: 1.3 to 1.7). The results were similar for infant visit alignment, with 51.2% of all attendance dates aligning to MIP clinic dates, also increasing over time from 42.5% in 2013 to 60.3% in 2016 (RR 1.5, 95% CI: 1.2–1.9). In a subanalysis, alignment between actual attendance dates and scheduled MIP clinic dates was examined when the appointment dates aligned with the scheduled MIP clinic date. Overall, 75.2% of these visits aligned for mothers and 77.8% for infants.
Visit Alignment for Each Mother
Mother's had a median (IQR) of 10 (6–12) appointment and attendance dates, median (IQR) of 4 (1–7) appointment dates that aligned to MIP visits, and median (IQR) of 3 (1–6) attendance dates that aligned to MIP visits. Figure 3A displays box plots for the proportion of mother's appointment dates that aligned with MIP clinic dates by facility. Figure 3B displays box plots for the proportion of mother's attendance dates that aligned with MIP clinic dates by facility. Only 3 facilities had medians ≥75% of actual attendance dates aligning to MIP clinic dates, and 3 more facilities had medians ≥50% of actual attendance dates aligning to MIP clinic dates.
The PRIME study assessed the effectiveness of integrating HIV and MNCH services for HIV+ women and their infants through MIP clinics, as a means of improving retention in care of HIV+ mothers and their infants. Over time, process indicators used to track and support integration of services, such as mentorship scores and visit alignment proportions, improved in most facilities.
As presented in Figure 2, although the mentorship scores improved between 2015 and 2016 in all facilities, there was a clear division among the facilities. Poor performing facilities improved their performance modestly over time but remained weak, and high performing facilities continued to perform well throughout the intervention period. These differences suggest the importance of intrinsic health facility attributes, which may include the quality of leadership and staff skills and motivation to implement the interventions. For example, in high performing facilities, HCWs who were already highly motivated may continue improving their performance after mentorship; whereas in other sites, poorly motivated HCWs may continue to perform poorly despite frequent visits and continuous support.
From the qualitative mentorship data, reasons for the inconsistent and variable practices in MIP clinic implementation included lack of space and functioning basic equipment, the complexity of making and sustaining significant changes in well-established patterns of practice, and attitudes and motivation of HCWs. Despite these challenges, several facilities were successful in integrating services through MIP clinics, as highlighted in high mentorship scores. However, many of these more successful sites also benefited from additional general support from partner organizations, which ensured better staffing, space, and equipment.
The visit alignment indicators presented in Table 2 and Figure 3 demonstrate that there was poor support for the intervention, as HCWs did not consistently schedule women's appointments on MIP clinic dates. These scheduling errors caused mothers and infants to return for appointments on days when MIP clinics were not operating. The median proportion of mother's attendance dates that aligned to MIP clinic dates was only 35.7%. This lack of sufficient exposure to the intervention may be a reason why the implementation of MIP clinics did not significantly affect the PRIME study's primary outcome of maternal retention.20
When participants were given a correct appointment date that aligned to an MIP clinic date, 75.2% of mothers and 77.8% of infants did attend on an MIP clinic date. This suggests that the intervention may be acceptable to participants. Finally, there was improvement in visit alignment over time, suggesting that with additional experience and practice, staff were better able to adhere to the model, and mothers were more likely to attend MIP clinic dates. Perhaps with more time, better implementation of MIP clinics could have been achieved, leading to higher retention in care.
The variation in baseline characteristics among the 20 clinics highlights some of the challenges for changing practices in rural settings. In 2006, WHO recommended 23 skilled HCWs (doctors, nurses, and midwives) per 10,000 inhabitants as the minimum threshold to enable the provision of 80% of basic health care services.16 Our results show that the number of staff available in each clinic varied substantially and remained insufficient with regard to the workload and size of the catchment population served. The quality of the consultation rooms and equipment available also varied considerably between sites. Limited availability of space to provide ART and MNCH services in the same physical location, at the same time, may have adversely affected the implementation of the MIP clinic model. The mentors concluded that inadequate numbers of trained HCWs and poor physical infrastructure impeded the willingness of HCWs to make changes in their service delivery approach in the short term, even if those changes may eventually result in better efficiency for both clients and HCWs in the long term.
Another challenge encountered by study mentors was that some HCWs requested incentives to support the project implementation because the integrated MIP clinics were viewed as additional work. These requests are partly because of prevailing practices in Malawi (where research activities or new ideas supported by partners are seen as sources of much-needed additional income) and highlight the difficulty of conducting implementation research in low-resource settings. In agreement with the MoH, no additional clinic staff were provided to implement MIP clinics, as the study design sought to implement changes in practice in a way that could be sustainable and scalable countrywide.
Overall, the study results also show that a simple monitoring tool, such as a mentorship checklist, can aid supervisors to regularly assess and monitor staff responsiveness and that increased mentorship can have a powerful impact on performance, as the highest mentorship scores were reached when the intensity and frequency of mentorship visits increased in 2015. These data suggest that the mentorship score seems to be an efficient tool to assess proper implementation of integrated HIV and MNCH services. Furthermore, high mentorship scores were found to be significantly correlated with the PRIME study's primary outcome of maternal retention in care (Pearson correlation coefficient = 0.610, P = 0.004), suggesting that the tool may have clinical utility as a predictor of the quality of the services that are being provided. Some of the limitations of the study were that perspectives of mothers attending the clinics were not captured in a systematic way, and no sophisticated measures of “service integration” were used.
Integration is not a simple solution for some of the challenges facing health systems that aim to provide comprehensive coverage of essential HIV and MNCH interventions. Careful design of a service delivery approach is needed that is both acceptable to clinic staff and addresses the local realities. The integrated service model needs to be understood by local staff so their expectations are consistent with those of health managers, and there is a shared belief and commitment to the new approach. Similarly, the means of introducing an integrated approach to local HCWs and devising a plan to scale up the approach needs to be planned and budgeted for both in terms of financial resources and staff time. Relatively simple tools, based on site-specific data, can be used for both tracking progress and for identifying sites needing additional help. However, such feedback is only meaningful if there are mechanisms for resolving problems, including systems for escalating specific difficulties to district managers who have the authority and funds to remedy problems. Implementing new systems, even relatively simple interventions such as diaries to organize follow-up appointments, are not necessarily easy to establish and sustain.
Consistent with literature from southern Africa,17–19 efforts to achieve integrated health service delivery in resource-constrained settings, with insufficient human resources, management and oversight, and poor infrastructure are not guaranteed to achieve positive results. Despite overall improvement because of mentorship, this study found that integrated MIP clinics could not be easily implemented by HCWs even with regular mentorship and support. Such findings highlight that appropriate design and investment are critical to achieving success of integrated MIP clinics.
The authors acknowledge Frank Chimbwandira and Michael Eliya for their technical support during implementation, the WHO technical support from Geneva and in the Afro-region, and the in-country WHO team for their operations support. The implementation support from District Health Offices and staff in Mangochi and Salima and help from CHAI staff Leslie Berman, Nurse Nyambi, Emily Kobayashi, and George Pro are very much appreciated. The authors also acknowledge technical inputs from Christiane Horwood and Sifiso Phakathi who substantially contributed to the development of the mentorship scoring tools.
1. Government of Malawi
Ministry of Health. Integrated HIV
Program Report, October–December 2010. Available at: http://http://www.hiv.health.gov.mw
/index.php/our-documents. Accessed March 2, 2017.
2. Government of Malawi
Ministry of Health. Integrated HIV
Program Report, October–December 2013. Available at: http://http://www.hiv.health.gov.mw
/index.php/our-documents. Accessed March 2, 2017.
3. World Health Organization. PMTCT Strategic Vision 2010–2015. Available at: http://http://www.who.int
/pub/mtct/strategic_vision/en/. Accessed January 22, 2017.
4. Brusamento S, Ghanotakis E, Tudor Car L, et al. Male involvement for increasing the effectiveness of prevention of mother-to-child HIV
transmission (PMTCT) programmes. Cochrane Database Syst Rev. 2012:CD009468.
5. Ambia J, Mandala J. A systematic review of interventions to improve prevention of mother-to-child HIV
transmission service delivery and promote retention. J Int AIDS Soc. 2016;19:20309.
6. Geldsetzer P, Yapa HMN, Vaikath M, et al. A systematic review of interventions to improve postpartum retention of women in PMTCT and ART care. J Int AIDS Soc. 2016;19:20679.
7. Obure CD, Jacobs R, Guinness L, et al. Does integration
and sexual and reproductive health services improve technical efficiency in Kenya and Swaziland? An application of a two-stage semi parametric approach incorporating quality measures. Soc Sci Med. 2016;151:147–156.
8. Obure CD, Guinness L, Sweeney S, et al. Does integration
and SRH services achieve economies of scale and scope in practice? A cost function analysis of the Integra Initiative. Sex Transm Infect. 2016;92:130–134.
9. Kennedy CE, Spaulding AB, Brickley DB, et al. Linking sexual and reproductive health and HIV
interventions: a systematic review. J Int AIDS Soc. 2010;13:26.
10. Grépin KA, Reich MR. Conceptualizing integration
: a framework for analysis applied to neglected tropical disease control partnerships. Plos Negl Trop Dis. 2008;2:e174.
11. Atun R, de Jongh T, Secci F, et al. A systematic review of the evidence on integration
of targeted health interventions into health systems. Health Policy Plan. 2010;25:1–14.
12. Briggs CJ, Garner P. Strategies for integrating primary health services in middle- and low-income countries at the point of delivery. Cochrane Database Syst Rev. 2006:CD003318.
13. Turan JM, Onono M, Steinfeld RL, et al. Implementation and operational research: effects of antenatal care and HIV
on elements of the PMTCT cascade: results from the SHAIP cluster-randomized controlled trial in Kenya. J Acquir Immune Defic Syndr. 2015;69:e172–e181.
14. Sibanda EL, Weller IVD, Hakim JG, et al. The magnitude of loss to follow-up of HIV
-exposed infants along the prevention of mother-to-child HIV
transmission continuum of care: a systematic review and meta-analysis. AIDS Lond Engl. 2013;27:2787–2797.
15. Mwapasa V, Pro G, Chinkhumba J, et al. Mother-infant pair clinic and SMS messaging as innovative strategies for improving access to and retention in eMTCT care and Option B
+ in Malawi
: a cluster randomized control trial (the PRIME study). J Acquir Immune Defic Syndr. 2014;67(suppl 2):S120–S124.
16. World Health Organization. Working Together for Health the World Health Report 2006. World Health Organization. Available at: http://http://www.who.int
/whr/2006/whr06_en.pdf?ua=1. Accessed January 22, 2017.
17. Wiegert K, Dinh TH, Mushavi A, et al. Integration
of prevention of mother-to-child transmission of HIV
(PMTCT) postpartum services with other HIV
care and treatment services within the maternal and child health setting in Zimbabwe, 2012. PLoS One. 2014;9:e98236.
18. Maharaj P, Cleland J. Integration
of sexual and reproductive health services in KwaZulu-Natal, South Africa. Health Policy Plan. 2005;20:310–318.
19. Scott V, Chopra M, Azevedo V, et al. Scaling up integration
: development and results of a participatory assessment of HIV
/TB services, South Africa. Health Res Policy Syst. 2010;8:23.
20. Mwapasa V, Joseph J, Tchereni T, et al. Impact of mother-infant pair clinics and short-text messaging service (SMS) reminders on retention of HIV
-infected women and HIV
-exposed infants in eMTCT care in Malawi
: a cluster randomized trial. J Acquir Immune Defic Syndr. 2017;75(suppl 2):S123–S131.