Evidence from community level inputs to improve quality of care for maternal and newborn health: interventions and findings

Annually around 40 million mothers give birth at home without any trained health worker. Consequently, most of the maternal and neonatal mortalities occur at the community level due to lack of good quality care during labour and birth. Interventions delivered at the community level have not only been advocated to improve access and coverage of essential interventions but also to reduce the existing disparities and reaching the hard to reach. In this paper, we have reviewed the effectiveness of care delivered through community level inputs for improving maternal and newborn health outcomes. We considered all available systematic reviews published before May 2013 on the pre-defined community level interventions and report findings from 43 systematic reviews. Findings suggest that home visitation significantly improved antenatal care, tetanus immunization coverage, referral and early initiation of breast feeding with reductions in antenatal hospital admission, cesarean-section rates birth, maternal morbidity, neonatal mortality and perinatal mortality. Task shifting to midwives and community health workers has shown to significantly improve immunization uptake and breast feeding initiation with reductions in antenatal hospitalization, episiotomy, instrumental delivery and hospital stay. Training of traditional birth attendants as a part of community based intervention package has significant impact on referrals, early breast feeding, maternal morbidity, neonatal mortality, and perinatal mortality. Formation of community based support groups decreased maternal morbidity, neonatal mortality, perinatal mortality with improved referrals and early breast feeding rates. At community level, home visitation, community mobilization and training of community health workers and traditional birth attendants have the maximum potential to improve a range of maternal and newborn health outcomes. There is lack of data to establish effectiveness of outreach services, mass media campaigns and community education as standalone interventions. Future efforts should be concerted on increasing the availability and training of the community based skilled health workers especially in resource limited settings where the highest burden exists with limited resources to mobilize.


Background
Annually around 40 million mothers give birth at home without any trained health worker [1]. Consequently, most of the maternal, perinatal and neonatal morbidities and mortalities occur at the community level due to lack of good quality care during labour and birth. The poorest and fragile countries have the highest neonatal mortality rates and preventable deaths depicting the existing inequities [1,2]. The causes are multi-factorial, ranging from socio-economic aspects of poverty; poor health status of women; lack of autonomy and decision making authority; and illiteracy to health system related factors like poor antenatal and obstetric care; absence of trained birth attendant; inadequate referral system; lack of transportation facilities; and poor linkages between health centers and communities [1,3]. This burden could be averted by achieving universal coverage in skilled birth attendance and providing good quality care for all births. However due to paucity of trained human resource professionals in first-level health services and the reduced awareness of and accessibility to services for the deprived and marginalized populations, these are not accessible to the ones in need [1,4].
Community based delivery is now widely recognized as an important strategy to deliver key maternal and child survival interventions [5][6][7][8][9][10]. It has been instrumental in scaling up coverage of certain interventions, such as immunization, oral rehydration therapy for diarrhea, tuberculosis treatment and family planning initiatives. Interventions delivered at the community level have not only been advocated to improve access and coverage of essential interventions but also to reduce the existing disparities and reaching the hard to reach. Inputs at the community level involve programs based on training and consequent task shifting from healthcare personnel to mid-level health care personnel or lay individuals for local implementation at home, village or any defined community group. They focus on resources such as local knowledge, volunteers' time, community confidence and trust as channels for delivery. Community platforms can be used to deliver a full spectrum of promotive, preventive, and curative interventions including provision of basic antenatal (ANC), natal and postnatal care (PNC); preventive essential newborn care; breastfeeding counseling; management and referral of sick newborns; skills development in behavior change communication; and community mobilization strategies to promote birth and newborn care preparedness. These programs do not substitute for a formal health system, but provide a channel to reach far flung areas with knowledge, commodities and skills, thus attempting to reduce existing inequities in healthcare access and utilization. In this paper, we have reviewed the effectiveness of care delivered through community level inputs for improving maternal newborn health (MNH) outcomes. For this review, we have broadly categorized these interventions into four categories: outreach services (including home visitation and referrals); task shifting; training; and formation of support groups for community mobilization.

Outreach services and home visitation
Home-based strategies to support optimal maternal and neonatal care practices have emerged in the last decade to complement facility-based care and promote universal access to and utilization of health services with a specific focus on low-and middle-income countries (LMICs). These services mainly include ANC, skilled birth attendance and early PNC. Such programs involve standardized or individualized interventions for additional support provided during home visits, regular ANC visits, and/or by telephone throughout pregnancy by multidisciplinary teams of health professionals and trained lay workers. The major benefit of these programs is that the service is brought to the remote population in their own homes and allow care providers to deliver a more tailored health care service.

Task shifting
Task shifting for human resource management involves substituting specialized personnel with healthcare workers that are lesser trained but can perform some aspects of their tasks. A range of both skilled and semi-skilled health workers can play a major role in MNH service delivery. Community health workers (CHW) and traditional birth attendants (TBA) are considered semi-skilled workers, while mid-level health workers (MLHW) such as nurses, midwives, associate clinicians, medical assistants and nurse auxiliaries are skilled workers certified for their work. Health service delivery through these skilled and semi-skilled healthcare workers has been practiced in both high-income countries (HIC) and LMIC all over the world for the past several decades. Evidence suggests that they can contribute significantly in improving health of the populations. More recently, due to the growing human resource crisis especially in LMICs, task shifting has reemerged for extending services to hard-to-reach groups by substituting health professionals for a range of tasks [11][12][13][14]. Countries in south Asia and Africa have made a particular effort in recent years to reduce maternal and neonatal mortality and morbidity through deploying CHW [15,16]. The role of midwives and TBA in delivering MNH services has also received growing attention in the last few years, and a number of publications have described their role and documented the effects of such programs [17,18]. However, less attention has been given to assess the effectiveness of MLHW in delivering and improving health care delivery [19].

Human resource training
Globally there is a growing shortage of 7.2 million healthcare workers and around 90% of all maternal deaths and 80% of still births occur in countries that lack trained healthcare workforce [20]. Although skillsmix imbalances persist, advanced practitioners, midwives, nurses and auxiliaries are still insufficiently used in many settings. Many LMIC are increasingly facing difficulties in producing, recruiting and retaining health professionals as they tend to migrate to wealthier countries due to low salaries, poor working conditions, lack of supervision, low morale and motivation and lack of infrastructure [9,21,22]. To overcome the failure of providing birthing women with skilled attendance, poor countries are now investing on training MLHW to at least provide them with some sort of assistance instead of none at all. In countries like Malawi, Bangladesh, Pakistan and Guatemala, training and close supervision of TBA have been evaluated to improve MNH outcomes and have shown some potential in reducing harmful practices during delivery and childbirth and improving MNH outcomes [23,24]. These training courses include short and structured approach to equip lay workers with lifesaving tools like Newborn Life Support (NLS), Neonatal Resuscitation Program (NRP) and Essential Newborn Care (ENC) but might vary in origin, scope and target audience. Besides additional training of outreach workers, human resource training also includes conferences, lectures, workshops, group education, seminars and symposia.

Community mobilization and support groups
Soon after the Alma-Ata Declaration, it was recognized that community participation was important for the provision of local health services and for delivering interventions at the community level. Since then, it has been advocated to build its links with improving MNH [25]. Community support groups and women's groups are now increasingly becoming a core component of community service package comprising of community representatives for health promotion. The objective is to enable the community to provide support to pregnant women and families throughout pregnancy and delivery. Communities are encouraged to identify key barriers to care and select the most appropriate interventions for their situation. Community mobilization also helps educate about available services, identification of danger signs during pregnancy, and the importance of seeking care from skilled healthcare worker during obstetric and newborn emergencies. A range of promotive messages, quality care and scale up coverage for MNH can be delivered through community workers and women's groups [26].
In this review, we aim to systematically review and summarize the available evidence from relevant systematic reviews on the impact of the outlined community level inputs Figure 1). to improve the quality of care for women and newborns.

Methods
We considered all available systematic reviews published before May 2013 on the pre-defined community level interventions as outlined in our conceptual framework [27]. A separate search strategy was developed for each component using broad keywords, medical subject heading (MeSH), and free text terms: [(Community OR home OR "home visit*" OR outreach OR "task shift*" OR "human resource" OR "in-service" OR training OR mobilization OR "support group*"OR "women's group" OR "health worker*" OR "community health aides" OR "primary health care" OR "community health worker*" OR "lay health worker*" OR "mid-level health worker*" OR "community based interventions") AND (health OR maternal OR mother OR child OR newborn OR neonat*)] Our priority was to select existing systematic reviews of randomized or non-randomized controlled trials, which fully or partly address the a priori defined community delivered interventions for improving quality of care for MNH. We excluded reviews on home visits for prevention or screening for child abuse, maltreatment and childhood injury prevention as these were not included in the scope of our review. Search was conducted in the Cochrane library and PubMed and reviews that met the inclusion criteria were selected and data was abstracted by two authors on a standardized abstraction sheet. Quality assessment of the included reviews was done using Assessment of Multiple Systematic Reviews (AMSTAR) criteria [28] as detailed in paper 1 [27]. Any disagreement between the primary abstractors was resolved by the third author. For the pre-identified interventions, which did not specifically report MNH outcomes, we have reported the impacts on other health outcomes reported by the review authors. Estimates are reported as relative risks (RR), odds ratios (OR), risk differences (RD) or mean differences (MD) with 95% confidence intervals (CI) where available. For detailed methodology please refer to paper 1 of the series [27].

Findings
Our search yielded 310 potentially relevant review titles. Further screening of abstracts and full texts resulted in the inclusion of 43 eligible reviews: 17 for outreach services (home visitation and referrals), 6 for task shifting, 18 for human resource training and 2 for community mobilization ( Figure 2). The overall quality of the reviews ranged from 3 to 11 with a median of 9 on the AMSTAR criteria.

Outreach Services
We included 16 [29][30][31][32][33][34][35][36][37][38][39][40][41][42][43][44] reviews and 1 [45] overview of reviews pertaining to outreach and home visitation services with the median quality score of 7 on AMSTAR criteria. All reviews except one [33] reported MNH specific outcomes. The most commonly reported outcomes included maternal, newborn morbidity and mortality, immunization rates, breast feeding, referral, ANC and PNC utilization. Meta-analysis was conducted in six of the reviews. Reviews evaluating the impact of structured nurse-or midwife-based home visitation programs were from HIC while those focusing on service delivery through CHW were from LMIC. Table 1 summarizes the characteristics of the included reviews.        [39]. A review evaluating the effectiveness of emergency obstetric referral interventions in LMIC showed that community based interventions for generating funds for transport reduced neonatal deaths in India (OR: 0.48, 95% CI: 0.34-0.68) while maternity waiting home interventions in sub-Saharan Africa reported reductions in stillbirths (OR: 0.56, 95% CI: 0.32-0.96) [35]. Another review to assess the effects of a maternity waiting facility on maternal and perinatal health did not find any trial for inclusion [44].
Nurse-or midwives-based home visit programs did not report any significant impact on birth outcomes, hospital admission for complications and neonatal morbidity and mortality. However, some positive impacts were reported on immunization rates (RR: 1.67, 95% CI: 1.29-2.15); child health outcomes including mental and physical health; and injury prevention [29,30,38,45]. Home visits with a specific focus on post natal visit was found to be associated with improved immunization rates (RR: 1.40, 95% CI: 1.16-1.68) with non-significant impacts on child's mental and physical health [31]. Programs offering additional social support for pregnant women at high risk for preterm or low birth weight (LBW) delivery showed significant impacts on reducing the likelihood of antenatal hospital admission (RR: 0.79, 95% CI: 0.68-0.92) and cesarean birth (RR: 0.87, 95% CI: 0.78-0.97) when compared to routine care [34]. Specialist out-reach clinics did not have any impact on health outcomes but improved compliance to treatment (RR: 0.63, 95% CI: 0.52-0.77), patient-provider satisfaction (RR: 0.43, 95% CI: 0.29-0.62) and access [33].

Task shifting
We included six [41,[46][47][48][49][50] reviews pertaining to task shifting with a median score of 10 on AMSTAR criteria. Four reviews evaluating the impact of task shifting to CHW and midwives in LMIC reported MNH outcomes while the other two reviews from HIC focused on the impact of dietary counseling delivered through dietician versus nurses/doctors [50] and impact of nurses working as substitutes for primary care doctors [48]. Table 2 summarizes the characteristics of the included reviews.   Care provided by midwives was found to be associated with significant improvements in antenatal hospitalization (RR: 0.90, 95% CI: 0.81-0.99), episiotomy (RR: 0.81, 95% CI: 0.77-0.88), instrumental delivery (RR: 0.86, 95% CI: 0.78-0.96), initiation of breast feeding (RR: 1.35, 95% CI: 1.03-1.76) and hospital stay (MD: -2.00,95% CI: -2.15 to -1.85) [47]. Another review evaluating the effects of CHW interventions reported significant impacts on immunization uptake (RR: 1.22, 95% CI: 1.10-1.37), breast feeding initiation (RR: 1.36, 95% CI: 1.14-1.61), child morbidity (RR: 0.86, RR: 0.75-0.99) and TB cure rates (RR: 1.22, 95% CI: 1.13-1.31) compared to routine care [49]. Care delivered by MLHW versus non-MLHW was found comparable for antenatal hospitalization, antepartum hemorrhage, fetal loss/ neonatal deaths, induction of labour, spontaneous vaginal delivery, instrumental vaginal births, cesarean sections, perineal laceration requiring suturing, post-partum hemorrhage, preterm birth, LBW and admission to neonatal special/ intensive care unit. Furthermore, ANC provided by midwives alone gave comparable results on a range of MNH outcomes compared to care provided by doctors working in a team with midwives. These findings suggest that care delivered by MLHW can be safe and effective [46]. Improved patient satisfaction and recall was reported when nurses were substituted for primary health care provision in place of doctors; although the data was from HIC only [48]. In another review, dietary counseling given by dieticians was comparable to that by nurses or doctors [50].

Training of human resources
We found eighteen [23,39,[51][52][53][54][55][56][57][58][59][60][61][62][63][64][65][66] reviews on human resource training with median quality score of 8.5 on AMSTAR rating scale. Three reviews reported MNH specific outcomes including ANC, institutional delivery, cesarean-section rates-section rates, referrals, stillbirths, maternal, perinatal and neonatal mortality while other reported outcomes included knowledge, compliance, performance and patient satisfaction. Most of the reviews evaluating training programs for outreach workers, CHW, community midwives or TBA were conducted in LMIC while reviews on the training of other licensed healthcare professional like physicians, residents, fellows, and medical students were from HIC. Table 3 summarizes the characteristics of the included reviews.
In LMIC settings, training TBA (for providing basic antenatal, natal and postnatal care; preventive essential newborn care, breastfeeding counseling; management and referral of sick newborns; skills development in behavior change communication and community mobilization strategies to promote birth and newborn care preparedness) as a part of community based intervention packages showed significant improvement in referrals (RR: 1.4,   Educational outreach visits, defined as use of a trained person from outside the practice setting who meets with healthcare professionals in their practice settings to provide information with the intent of changing their performance.   [39]. TBA training also reduced peri-neonatal mortality however there was insufficient data to provide the evidence base needed to establish training effectiveness [23]. In-service training courses specifically directed to improve the management of critically ill newborns showed significant improvements in performance of initial resuscitation (RR: 2.45, 95% CI: 1.75-3.42) and reduced the frequency of inappropriate and potentially harmful practices (RR: 0.40, 95% CI: 0.13-0.66) [60] while in-service trainings for skilled birth attendants (doctors, nurses and midwives) were found to be associated with significant impacts on maternal mortality (RR: 0.57, 95% CI: 0.36-0.91) and institutional delivery (RR: 2.92, 95% CI: 2.09-4.06) [51]. The impacts on obstetric complication, ceasarean sections, ANC and referrals were non-significant. For outcomes other than MNH, educational outreach visits and meetings were associated with improved compliance (5.6%, Range: 3-9%), prescription (4.8%, Range: 3-6.5%), professional practice (6%, Range: 3.6-16%), and some of the patient healthcare outcomes [53,59]. The evidence for continuing medical education, problem based learning and clinical practice guideline implementation remained inconclusive [63,64,66]. The impact of critical appraisal teaching on physicians' behavior was mixed with positive impacts on improving knowledge (MD: 0.10, 95% CI: 0.06-0.14), skills, and attitude [55].

Community mobilization
We found two [39,67] reviews evaluating the impact of community mobilization strategies and formation of community support groups with median quality score of 8 on AMSTAR criteria. Both the reviews reported the impacts on MNH outcomes with one from HIC and the other from LMIC. Table 4 summarizes the characteristics of the included reviews.

Discussion
There is a greater body of existing evidence on the effectiveness of community based inputs for improving MNH outcomes in LMIC compared to district and facility level inputs (discussed in papers 3 and 4) [68,69]. At community level, home visitation, community mobilization, women's support groups and training of the CHW and TBA have shown maximum impact on a range of MNH outcomes. Community based generation of funds for transportation has also shown significant impacts in resource limited settings of India and sub-Saharan Africa. Interventions delivered through midwives, CHW and MLHW have not only demonstrated comparable outcomes when compared to routine non-MLHW care delivery but also reported better results for some of the outcomes. Specialized outreach clinics, continuing medical education, problem based learning, clinical practice guideline implementation and critical appraisal showed inconclusive and mixed results. Although the process pathways for the effectiveness of community delivered interventions are uncertain, they seem to influence community awareness, behavior change and practices, such as accessing skilled birth, use of clean delivery kits, breastfeeding and care seeking for maternal and newborn illnesses. Our overview findings greatly add to the global evidence base of intervention and delivery strategies that may improve MNH outcomes. It implies that within the community level inputs, three interventions have unparalleled significance: first, CHW who provide primary health care, can mobilize community members and impart knowledge; second, training of and linkages to TBA can provide basic prenatal and obstetric care, as well as referrals where skilled birth attendants are absent; third, community support groups, especially women's groups, can empower communities and help in problem solving and planning to improve opportunities for women's health, as well as care for mothers and newborns.
Countries in Asia and sub-Saharan Africa are facing critical shortages of healthcare workers despite of bearing 25% of the world's diseases burden [20]. Reasons behind the migration of professional healthcare force to richer countries is suggested to be lack of incentive, poor working conditions and fewer opportunities for promotions [70]. There is also an existing pull from HIC to recruit health workers from LMIC. Given the shortage of care providers and functional health facilities, and the deeply entrenched practices, there is much interest in community-based interventions and strategies for care. Increasing the number of skilled health workers, training and educating them, providing them with incentives and improving the infrastructure is what needs to be done in all the LMIC to make progress towards achieving the Millennium Development Goals (MDG) 4 and 5.The findings from this overview testifies that increasing the availability and training of the skilled health workers including TBA and CHW for adequately recognizing and managing obstetric complications can significantly reduce maternal and neonatal morbidity and mortality especially in the resource limited settings of Asia and Africa where the highest maternal mortality burden exists with limited resources to mobilize. The challenge is to incentivize these programs and link them with formal health systems to increase retention. In many countries CHW are not linked to national health systems and are expected to work as volunteers which is a major drawback. Another existing challenge is the variation in prerequisites, recruitment, training, supervision and workload of various cadres of community workers including CHW, TBA and midwives. There is a need to streamline their functioning and delegate the activities to achieve efficient implementation and maximum impact.
With the established effectiveness of task shifting and training of CHW, future studies should focus on the factors affecting the sustainability and cost effectiveness of these interventions when scaled up [46]. There is a dearth of information on costs and equity aspects of community based programs as only a few studies have reported the actual costs incurred for saving lives or averting deaths with the use of these strategies. Researchers should now facilitate cost-effectiveness studies and consequent meta-analysis by collecting and reporting cost effectiveness data in a standardized format [39]. Further work is also needed on nurse-based care delivery models including length and frequency of contact, type of approach (e.g. individual or group, behavioral therapy or instructional techniques), level of training of practitioner, patient satisfaction and initial characteristics of patients to establish equivalence in care with the physician-based model and also for program replication [50]. Formal monitoring and evaluation of programs especially referral interventions are also necessary to develop better understanding of how referral interventions work. There is lack of data to establish effectiveness of mass media campaigns and community education as single strategies.
Outreach services may confer the most benefit to access and health outcomes in rural and underprivileged settings hence there is a need for good comparative studies in resource deprived settings rather than in HIC [33]. Among the outreach workers in the LMIC, the role of TBA is pivotal as they remain a major maternity care provider and their services expand from birth attendance to include newborn and post natal care like bathing and massage, domestic chores, and provision of care during postnatal period. Despite of that, TBA training remains controversial in relation to the global 'Safe Motherhood Initiative' as there is insufficient data to provide the evidence base needed to establish training effectiveness [23]. Hence methodologically rigorous evaluations with an adequate sample size are needed to measure the magnitude of the impact of TBA training on maternal and neonatal mortality.
Community based interventions have promising potential to provide range of services throughout the continuum of care and also reach the hard to reach population groups. Current evidence emphasizes that effective community based strategies exist to deliver a range of preventive and promotive interventions to improve the quality of care delivered for MNH and many of these interventions have the potential to reduce maternal, perinatal and neonatal morbidity and mortality. There is now a need to implement them on a larger scale throughout the LMIC. These interventions exist within the current health systems in most of the LMIC and hence policies are needed to integrate and sustain various task shifting and training interventions with the maternal health programs within their health systems. All stakeholders including governments, communities and donors need to work together to form these policies and develop models of health care to suit the needs of their own population. Still more work needs to be done in areas of recruitment, deployment and retention of the community based health care workers in the rural and underprivileged areas and improve the working conditions for them.