How can human resources for health interventions contribute to sexual, reproductive, maternal, and newborn healthcare quality across the continuum in low- and lower-middle-income countries? A systematic review

Background Well-trained, competent, and motivated human resources for health (HRH) are crucial to delivering quality service provision across the sexual, reproductive, maternal, and newborn health (SRMNH) care continuum to achieve the 2030 Sustainable Development Goals (SDGs) maternal and neonatal health targets. This review aimed to identify HRH interventions to support lay and/or skilled personnel to improve SRMNH care quality along the continuum in low- and lower-middle-income countries (LLMICs). Methods A structured search of CINAHL, Cochrane Library/trials, EMBASE, PubMed, SCOPUS, Web of Science, and HRH Global Resource Centre databases was undertaken, guided by the PRISMA framework. The inclusion criteria sought to identify papers with a focus on 1. HRH management, leadership, partnership, finance, education, and/or policy interventions; 2. HRH interventions' impact on two or more quality SRMNH care packages across the continuum from preconception to pregnancy, intrapartum and postnatal care; 3. Skilled and/or lay personnel; and 4. Reported primary research in English from LLMICs. A deductive qualitative content analysis was employed using the World Health Organization-HRH action framework. Results Out of identified 2157 studies, 24 intervention studies were included in the review. Studies where ≥ 4 HRH interventions had been combined to target various healthcare system components, were more effective than those implementing ≤ 3 HRH interventions. In primary care, HRH interventions involving skilled and lay personnel were more productive than those involving either skilled or lay personnel alone. Results-based financing (RBF) and its policy improved the quality of targeted maternity services but had no impact on client satisfaction. Local budgeting, administration, and policy to deliver financial incentives to health workers and improve operational activities were more efficacious than donor-driven initiatives. Community-based recruitment, training, deployment, empowerment, supportive supervision, access to m-Health technology, and modest financial and non-financial incentives for community health workers (CHWs) improved the quality of care continuum. Skills-based, regular, short, focused, onsite, and clinical simulation, and/or mobile phone-assisted in-service training of skilled personnel were more productive than knowledge-based, irregular, and donor-funded training. Facility-based maternal and perinatal death reviews, coupled with training and certification of skilled personnel, positively affected SRMNH care quality across the continuum. Preconception care, an essential component of the SRMNH care continuum, lacks studies and services in LLMICs. Conclusions We recommend maternal and perinatal death audits in all health facilities; respectful, woman-centered care as a critical criterion of RBF initiatives; local administration of health worker allowances and incentives; and integration of CHWs into the healthcare system. There is an urgent need to include preconception care in the SRMNH care continuum and studies in LLMICs. Supplementary Information The online version contains supplementary material available at 10.1186/s12960-021-00601-3.

improving the management, capacity, and motivation of existing HRH is vital to improving maternal healthcare quality [6].
A synthesis by Munabi-Babigumira et al. (2017) found that pre-service and in-service training, adequate staffing, supervision, incentives, leadership and management support, adequate equipment and supplies, and teamwork and collaboration improved the ability of skilled personnel to deliver quality IPC and PNC services [20]. Althabe et al. (2008) demonstrated that interactive workshops and reminders, educational outreach visits, audit and feedback, mass-media and patient-mediated interventions, financial incentives, and/or organizational and regulatory interventions had a moderately positive effect on healthcare provider performance and the quality of ANC, IPC, PNC, and neonatal care services in low-and middle-income countries [21]. Sibley et al. (2009) found that training in advice on ANC, IPC, and PNC; management of normal delivery; advice on breastfeeding; and timely detection and referral of women with obstetric complications by traditional birth attendants (TBAs) had a positive effect on the increment of timely referrals and reduction of maternal, perinatal and neonatal mortalities, and stillbirth rates [22]. However, there are no reviews examining HRH interventions' contribution to lay and/ or skilled personnel's performance to improve SRMNH care quality across the continuum in LLMICs. The review by Lassi et al. focused on skilled personnel only and lacked an examination of HRH interventions contributing to SRMNH care quality across the continuum [6]. We, therefore, undertook an updated and comprehensive review to examine HRH interventions and SRMNH outcomes.

Methods
A deductive qualitative content analysis of HRH interventions and their effects on SRMNH care quality along the continuum was undertaken using an a priori conceptual framework to help direct and define the study and deliver practical insights for health policy and practice decision-making [23,24]. We used the WHO-HRH Action Framework [25,26] to define HRH interventions (Fig. 1). The framework identifies six action fields (Management, Leadership, Partnership, Finance, Education, and Policy) [25,26]. The Enhancing transparency in reporting the synthesis of qualitative research (ENTREQ) guidance was followed in this review [27], which is registered in the International Prospective Register of Systematic Reviews [28].

Search protocol
In consultation with two public health research librarians from the University of Technology Sydney (UTS), a Population, Interventions, Comparators, Outcomes, and Study design (PICOS) design, was applied to develop the review question: in LLMICs, how can HRH interventions contribute to SRMNH care quality across the continuum?
Included studies were those that described: (a) an HRH management, leadership, partnership, finance, education, and/or policy intervention (see definitions at Additional file 1: Table S1.); (b) one or more HRH interventions' effect on two or more consecutive quality SRMNH care packages across the continuum (a study reporting HRH interventions delivered in conjunction with one or more non-continuous SRMNH care packages was excluded since it was not a continuum); (c) the role of skilled and/ or lay personnel (see definitions at Additional file 1: Table S2.); and (d) primary research studies in English (other languages were excluded due to resource constraints) conducted in LLMICs (see inclusion criteria at Additional file 1: Table S3.). Studies that did not include any of the six HRH interventions were excluded. We defined a quality SRMNH care package as one that contained a safe, effective, timely, efficient, equitable, and people-centred package of interventions that comprised PCC, ANC, IPC, and/or PNC (maternal and/or newborn) services as described by the WHO [8,13]. As noted in the protocol for this review, the main outcome of the review is quality of care (defined according to WHO as care that is safe, effective, timely, efficient, equitable, and peoplecentered). Additional outcomes included were maternal and/or neonatal health outcomes directly related to the quality of care. These involve maternal and/or neonatal morbidity or mortality and uptake of the recommended and life-saving interventions facilitated by health workers (such as early initiation of breastfeeding, delayed bathing of newborns, and cord care) [28].

Search strategy
Six bibliographic databases (PubMed, Web of Science/Core Collection, SCOPUS, CINAHL, EMBASE/ OVID, and Cochrane Library/trials) were systematically searched in consultation with two public health research librarians from the UTS (Additional file 1: Table S4.). The search engine: HRH Global Resource Centre was searched for grey literature. Quantitative, qualitative, and mixed-method studies published between 01 January 2000 and 31 December 2019 were retrieved. This period was chosen to evaluate progress over time in SRMNH care quality along the continuum in relation to HRH interventions since the end of the Alma Ata Health for All Declaration in 2000, the Millennium Development Goals (MDGs) era, and in the first 5 years of implementation of the SDGs. The Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) 2009 guideline was used to outline the review process [29]. A total of 2157 studies were identified, including 1923 from six bibliographic databases and 234 from the HRH Global Resource Centre (Fig. 2), with 477 duplicates being identified and removed using the endnote de-duplication guidelines [30].
Two reviewers (MG and AD) independently used the Covidence online production tool for the title and abstract screening, full-text screening, data abstraction, and quality assessment. Differences in decisions regarding the final papers for review and quality assessment were resolved through a review by the third author (DS), and a consensus was reached. A total of 1437 and 219 articles were excluded during the title and abstract screening and full-text eligibility assessment, respectively, because they did not meet the inclusion criteria.

Quality appraisal
Two reviewers (MG and AD) independently appraised the methodological quality of 25 studies that met the inclusion criteria to describe their methodological quality and ensure that there was enough methodological detail to ensure rigour to be included for the review. The Cochrane methods were applied to appraise the quality of randomized controlled trials (RCTs) (Additional file 1: Table S5) [31]. Each RCT was evaluated for internal validity and quality of reporting. Quality of quasi-experimental, prospective (pre/post), post-only and comparison, and post-only studies were appraised using the Joanna Briggs Institute (JBI) critical appraisal checklist for quasi-experimental studies (non-randomized experimental studies) (Additional file 1: Table S6) [32]. Studies collecting qualitative or mixed data were appraised using the United Kingdom's National Health Service Critical Appraisal Skills Programme (CASP) qualitative checklist (Additional file 1: Table S7) [33]. We included intervention studies with a methodologically low risk of bias or moderate to high quality for the review. One pre/ post-test study was excluded during the quality appraisal because it has a low-quality score (3 out of 9). Using the quasi-experimental studies' quality appraisal tool, a study with "yes" responses for the 9 signalling questions of less than 4 was considered low quality [34].

Data abstraction and synthesis
Data from the 24 studies were systematically extracted into tables using templates based on the Cochrane methods to integrate qualitative and implementation evidence within intervention effectiveness reviews [35]. A deductive qualitative content analysis of the extracted text related to each implemented HRH intervention from each included study was undertaken through coding texts according to emergent descriptions and labelling and structured along with the four categories of SRMNH care continuum [36,37]. Tables and concept maps were used to plot patterns and relationships across these categories, and robustness was assessed through critical reflection and discussion between the three authors.

PCC, ANC, IPC, and PNC continuum
Two cluster-randomized trials involving skilled personnel were identified [39,60]. Engineer et al. [39] described the effect of payment for performance (P4P) on the quality of maternal and child health (MCH) services in                  Afghanistan. Doctors and mid-level cadres were provided with quarterly bonuses based on the delivery of nine MCH related indicators. There was no direct communication with health workers about the bonuses, nor were there any demand-side interventions (raising or creating demand in communities). The intervention positively affected history taking and physical examinations, and client counseling quality of care indices across the SRMNH care continuum. The intervention, however, had no significant impact on equitable access to the targeted MCH services use between poor and rich families or on the adequacy of time spent with a client along the SRMNH care continuum. It did not also improve client satisfaction and the perceived quality of care along the SRMNH care continuum. Zeng et al. [60] investigated the effect of results-based financing (RBF), an approach to incentivize healthcare providers and operational activities based on performance, on SRMNH care quality in line with input-based financing (IBF), a traditional approach of increasing funding not tied to performance, in Zambia. Both the RBF and IBF interventions significantly improved the quality of injectable contraceptives, ANC, IPC, and PNC services, respectively, compared to their respective controls (without additional financing). Pregnant women and children in RBF districts gained 604 and 14,574 qualityadjusted life years (QALYs), respectively, while pregnant women and children in IBF districts gained 302 and 8,274 QALYs, respectively, as compared to the controls.

ANC, IPC, and PNC continuum
Thirteen studies focused on improving SRMNH care quality across ANC, IPC, and PNC continuum. Six studies featured interventions to improve the performance of CHWs. Okuga et al. and Waiswa et al. [48,49] evaluated the effect of community-based recruitment, training, deployment, supervision, modest financial and non-financial incentives for CHWs, and their integration into the healthcare system on maternal and newborn healthcare in rural Uganda. Health facility strengthening was undertaken at all facilities. Qualitative interviews with key stakeholders found that CHWs were positively received and used their social networks to identify and refer pregnant women and involve men in health education. Okuga et al. (2015) showed reduced delays in healthcare service delivery at health facilities; and improvements in compassionate and respectful care, and cord care. In addition, there were improvements in the early initiation of breastfeeding and feeding newborns on colostrum and delayed bathing of newborns. [48]. Waiswa et al. (2015) found improved maternal and essential newborn care practices among poorer families. Significant, positive impacts of the intervention were identified, including increased health worker home visiting during pregnancy and the first week after birth. There were improvements in early breastfeeding initiation, delayed bathing of newborns (≥ 24 h), and cord care [49].
In the study by Edwards et al. [50], village and community health workers recruited from their respective communities were given skills-based training on maternal and newborn health (MNH) and primary healthcare. They received monthly supportive supervision in rural Bangladesh. Health workers from villages to the general hospital and the community worked in collaboration and team. Confidential, no-blame perinatal and maternal death audit was also implemented. A higher proportion of poor women in intervention areas received ANC, skilled personnel-assisted birth (SPAB), caesarean section, and PNC services than poor women living in the non-intervention, nationally sampled study areas across the SRMNH care continuum. There was also a reduction in the gap in service use between the poorest and richest women in intervention areas than in the national sampled study areas along the SRMNH care continuum. In the study by Agarwal et al. [54], Accredited Social Health Activists (ASHAs) were trained to provide health education and connect women to healthcare facilities and providing home-based ANC and PNC. Exposure to ASHA program compared to non-exposure had no significant effect on completing all services across ANC, IPC, and PNC care continuum.
Mobile technology was the focus of study by Balakrishnan et al. [55] who examined the effect of a mobile phone app (mHealth platform) used by trained community-based frontline health workers (ASHAs, Anganwadi Workers and Auxiliary nurse-midwives) to track services delivered to women and their newborns in India. The intervention villages were found to have an increased uptake of ≥ 90 iron and folic acid tablets during pregnancy, early initiation of breastfeeding, and PNC home visits as compared to the non-intervention control areas. There was equity in the coverage of all quality indicators of SRMNH care across all casts (scheduled vs. others). However, there were no differences between intervention and control areas regarding the uptake of tetanus toxoid injections during pregnancy after a year of intervention. Maru et al. [43] evaluated the effect of a public-private partnership that involved developing an accountable care framework that integrated CHWs through companion and home visits to deliver health facility-based care in rural Nepal. The intervention showed an increase in ≥ 4 ANC visits by 6.4 pp, health facility birth by 11.8 pp, and postnatal contraception by 27.5 pp. Ninety-five percent of pregnant mothers received an ultrasound examination by month 8 or 9 of pregnancy. Infant mortality decreased from 18.3/1000 to 12.5/1000 live births.
Seven studies featured interventions to improve the performance of skilled personnel. Out of these, educational interventions were focused on five studies. Okawa et al. [53] examined the effect of doctors and mid-level cadres training and supervision on adequate contacts and SRMNH care quality in rural Ghana. The intervention had a significant, positive effect on the quality of PNC (p = 0.02). The intervention, however, did not significantly improve the quality of ANC or IPC. Having adequate contact with healthcare providers did not guarantee a high quality of care. In the study by McDougal et al. [57], community-based frontline workers were trained, mobilized, and empowered to improve the quality and effectiveness of home visits in India. The intervention had a significant, positive effect for nothing applied to cord after birth, kangaroo mother care, and postnatal contraception use. Ayalew et al. [41] examined the effect of multi-faceted interventions, including the Basic Emergency Obstetric and Newborn Care (BEmONC) training, supportive supervision, audit and site mentoring, and health facility-based quality improvement teamwork in Ethiopia. It had a significant, positive impact on healthcare provider performance during labour and birth (p = 0.002) and immediate PNC services (p = 0.001) compared to the comparison facilities. Magge et al. [45] studied the effect of clinical mentorship, training, and collaborative district-wide learning and leadership on maternal and newborn care quality in Rwanda. Post-versus pre-intervention outcomes showed pregnant women with premature rupture of membrane (PROM) treated with antibiotics of (38% vs. 24%); pregnant women with preterm labour treated with corticosteroids of (75% vs. 26%); waiting time to C-section in minutes (72 vs. 99); immediate kangaroo mother care (87% vs. 19%); and newborns checked for danger signs within 24 h of birth (98% vs. 47%). Rahman et al. [51] assessed the effect of an integrated packaged interventions, including community participation and onsite training on the management of deliveries and newborn complications on perinatal mortality in Bangladesh. Early pregnancy ANC home visits, caesarean section rates, early initiation of breastfeeding, colostrum as first newborn food, and delayed first newborn bathing were significantly, positively higher in the post-intervention period as compared to two years preintervention (p < 0.001). In intervention areas, perinatal mortality decreased by odds of 36%; less than 24 h timing of first newborn bath and preterm births significantly decreased.
Two studies examined financial and policy interventions. Binyaruka et al. (2015) and Kambala et al. [42,46] described the effect of the P4P program (Tanzania) and RBF (rural Malawi) on SRMNH care quality along the continuum. The P4P had a significant, positive impact on one of the eight targeted indicators: anti-malarial treatment during ANC visits (p = 0.001). However, there was no evidence of the effect of the P4P program on women's satisfaction with care. There was no significant effect on non-targeted services either (satisfaction with interpersonal care and waiting and consultation times) [46]. Kambala et al. [42] showed that the RBF had no significant, positive impact on women's perceptions of technical care, quality of amenities, or interpersonal relations during ANC, IPC, and PNC. Women reported instances of neglect, disrespect, and verbal abuse by health personnel while receiving care. The health personnel noted an increased workload due to the increased number of women seeking care at intervention facilities.

ANC and IPC continuum
Six studies focused on improving outcomes across ANC and IPC continuum. Five studies featured interventions to enhance the performance of skilled personnel. Of these, financial and policy interventions were the focus of the two studies. In the study by Basinga et al. [44], the Rwandan government launched a national P4P scheme in health centres based on 14 key MCH-care quality indicators. Quarterly audits of care were made at each health centre based on direct observation and a review of patients' records using a standardized assessment tool on an unannounced, randomly chosen day. The intervention had the greatest effect on indicators that only had the highest payment rates and needed the least effort from the service provider. The intervention had a significant, positive impact on standardized total ANC's quality, the number of high-risk pregnancies referred to district hospital for delivery during pregnancy, and the number of emergency transfers to hospital for obstetric care during delivery. The intervention, however, didn't improve the uptake of tetanus toxoid injections during ANC visits. Bonfrer et al. [40] assessed the effect of the performance-based financing (PBF) policy and quarterly quality assessment by local regulatory authorities on the quality of pregnancy and IPC in Burundi. The PBF policy is the one with financial incentives for healthcare providers based on their performance, excluding operational activities. The program had a significant, positive effect on the coverage of ≥ 1 tetanus toxoid injections during ANC; and SPAB among the non-poor (p < 0.028) but the poor in provinces with PBF program as compared to those without it.
Two studies focused on partnership and management interventions. Mwaniki et al. [58] evaluated the effect of a collaborative health worker advisory grouping in improving maternal healthcare quality in rural Kenya. The intervention, at the end line compared to the baseline, had significantly increased ANC contacts per month with standardized care (p < 0.001), and pregnant mothers actively referred from the community to health facilities for ANC and IPC services (p = 0.012). Duysburgh et al. [61] reported on the effect of a computer-assisted clinical decision support system and performance-based financial and non-financial incentives in Burkina Faso, Ghana, and Tanzania. The intervention showed a significant improvement in the number of lab proteinuria examinations during ANC, history taking on vaginal bleeding during pregnancy, births with correctly completed partograph, blood pressure monitoring during childbirth, and oxytocin for women after childbirth. Larson et al. [47] focused on educational and management interventions in rural Tanzania. Providers from 12 primary care clinics received in-service training, mentoring, supportive supervision and infrastructure support, and community members received peer outreach services. The intervention had a significant, positive, and negative effect on ANC's quality and obstetric care cost, respectively.
The TBA educational and management interventions were featured in a study by Satti et al. [59] in rural Lesotho. In this research, one hundred women, mostly TBAs, were trained and provided with incentives to identify pregnant women in the community and accompany them to a health centre for ANC and delivery. A nurse-midwife was deployed to the health centre to provide care and supervise the TBAs. Pregnant women from isolated communities were accommodated at a maternal waiting room two weeks before delivery. The intervention resulted in that the HIV testing, syphilis testing, and haemoglobin testing during ANC visit increased, as did the number of women with complications who were successfully transferred to the district hospital for obstetric care during delivery. There were no maternal deaths among the women in the program.

IPC and PNC continuum
Three studies focused on improving outcomes across the IPC and PNC continuum that featured educational and management interventions. Ghosh et al. [56] studied the effect of a multi-faceted, onsite nurse-mentoring and simulation program on nurses' and midwives' skills in the diagnosis and management of intrapartum asphyxia and postpartum haemorrhage (PPH) in India. The mentoring had a significant, positive effect on the diagnosis of asphyxia and PPH and management of asphyxia per additional week of mentoring. Gomez et al. [52] investigated the impact of an onsite, lowdose, high-frequency, and clinical simulation-assisted midwives' BEmONC training in Ghana. The intervention significantly reduced the intrapartum stillbirth and 24-h newborn mortality rates by 52% and 70%, respectively, during the 7-12 months of implementation compared to the baseline. Pirkle et al. [62] evaluated the effect of maternal death reviews in 32 referral hospitals on the quality of obstetrical care in Mali and Senegal. Women treated at intervention hospitals had, on average, higher quality of care scores than those treated at control hospitals.

Discussion
This study represents the first systematic review to examine HRH interventions and their impact on SRMNH care quality across the continuum in LLMICs. Our findings identified the key elements to successful HRH interventions and the areas for future research regarding HRH interventions.

Respectful, woman-centred care as a criterion of RBF
The RBF, based on a pre-defined and communicated set of indicators and its policy, significantly improved the quality of targeted SRMNH care services across the continuum given the increasing human resources, equipment, and supply demands are fulfilled [39,40,42,44,60]. However, these interventions had no significant, positive effect on clients' perception of quality or equity in care [39,40,42,44,46]. This signals the critical importance of including respectful, woman-centred care. A systematic review suggests that continuous, personalized care provided by the usual midwife and delivered within a family or a specialized setting, generates the highest level of satisfaction [64]. Womancentred care fosters the woman's psychological and physiological recovery, often surpassing clinical action, and is associated with lower intervention rates [64,65]. A systematic review from low-and middle-income countries states a positive effect of P4P on history taking, physical examination, blood pressure measurement, and blood and urine testing during ANC visits, and on provider's adherence to explaining medicine intake for children under five years, children are given medicines and children follow-up treatment. However, there was weak evidence for P4P's positive effect on maternal and neonatal health outcomes and out-ofpocket expenses [66].

Local budgeting and administration to deliver financial incentives
This review revealed that local budgeting, administration, and policy to deliver financial incentives and improve operational activities, were more effective than mobilizing direct funds from donors for financial incentives, operational activities and incentivizing users, and its policy in improving SRMNH care quality along the continuum and maternal and/or neonatal health outcomes [39,40,44,46,57,59,61]. Reliance on donor funding and donor-driven initiatives may reduce the responsibility for the delivery of care and increase dependence on outside funds. Another study found that despite extensive investment from donors, in-service training and supportive supervision to improve health workers' performance in providing quality antenatal and sick child care in seven countries in sub-Saharan Africa (SSA) did not improve the quality of care [67].

Training, health worker empowerment, and the integration into the healthcare system
Community-based recruitment, training, deployment, empowerment, supportive supervision, technology assistance, and modest financial and non-financial incentives for CHWs in rural hard-to-reach areas had a significant, positive effect on the quality and equity of SRMNH care services across the continuum. The social networks that CHWs had were a valuable asset enabling the building of rapport and relationships with pregnant women to facilitate referral to nearby health facilities and the provision of home-based life-saving ANC and PNC services [43,48,54,55,57,59]. This finding identifies the importance of training for CHWs, and policy initiatives to ensure their empowerment and integration into the healthcare system. A Cochrane trial's review demonstrates that TBA training significantly reduced stillbirth and perinatal and neonatal deaths [22]. There is a need for policy to recognize the important cultural and social roles CHWs fulfill in their local communities that can positively affect maternal and neonatal health outcomes [68].

BEmONC training, supportive supervision, teamwork, and collaboration for skilled personnel
In-service BEmONC training, supportive supervision, mentoring, audit, quality improvement advisory team, collaboration with CHWs, and community participation for skilled personnel had a significant, positive effect on quality ANC and IPC, and decreased perinatal mortality and preterm births [41,45,51,58]. In a review from Africa, implementing comprehensive interventions that strengthen the health system's different components, both in the community and at the health facilities, directly or indirectly improved the quality of maternal healthcare and morbidity and mortality outcomes [69].

Skills-based, on-site, regular, and clinical simulation training for nurses and midwives
On-site, short, frequent, regularly run, and clinical simulation-assisted training and mentorship of nurses and midwives had a significant, positive effect on the quality of ANC, IPC, and PNC along the continuum. This, in turn, significantly reduced the intrapartum stillbirth and 24-h newborn mortality rates [45,52,56]. On the other hand, in-service training, orientation, mentoring and supportive supervision, and peer outreach services implemented in two studies in support of the continuumof-care package had only a modest effect on the quality of ANC, IPC, or PNC [47,53]. Analysis of the Service Provision Assessment data from seven countries in SSA shows that in-service training and supportive supervision had a modest effect on ANC and sick child care quality. Providers were observed performing fewer than half of the recommended clinical actions for pregnant women and sick children on average [67]. Accordingly, skills-based, regular, focused, onsite, and clinical simulation and/or mobile technology-assisted in-service training of skilled personnel are more effective than the knowledge-based, irregular, and ineffective donor-funded training or supervision in LLMICs.

Maternal and perinatal death audits in all health facilities
Maternal death reviews in 32 referral hospitals in Mali and Senegal, coupled with training and certification of skilled personnel, had a significant, positive effect on ANC and IPC's quality across the continuum [62]. Confidential, no-blame perinatal and maternal death audits in rural Bangladesh improved the equity for poor women in receiving life-saving maternal healthcare services along the SRMNH care continuum [50]. A systematic review of effective non-drug interventions for improving outcomes and quality of maternal healthcare in SSA also found that facility-based clinical audits and maternal death reviews supported by demand and supply-side financial incentives, health systems strengthening interventions, community mobilization and peer-based programs, and task shifting directly or indirectly improved quality of maternal healthcare and morbidity and mortality outcomes in SSA [69].

Involving skilled and lay personnel in primary care
In primary care, HRH interventions involving both skilled and lay personnel were more effective than those involving either skilled or lay personnel alone on improving quality or health outcomes [39, 40, 47-49, 55, 61]. A systematic review that focused on IPC found that increasing the use of skilled personnel where TBAs are providers of birth care suggested that deploying midwives closer to communities, financial incentives for providers and users, mobilizing the community, and partnering with TBAs decreased maternal mortality [70].

Comprehensive HRH interventions targeting different components of the healthcare system
There is no single effective HRH intervention: ≥ 4 comprehensive HRH interventions had a more positive effect on quality SRMNH care continuum and/or health outcomes than those studies with fewer HRH interventions [39, 41, 45, 47-50, 52, 55, 57-59, 62], suggesting a cumulative effect. A systematic review from SSA also indicates that comprehensive HRH interventions implemented at the healthcare system's different components, both in the community and health facilities, improve maternal health [69].

Inclusion of PCC in the SRMNH care continuum
The PCC, any intervention to optimize a woman's health before pregnancy to improve maternal, newborn, and child health outcomes, effectively improves pregnancy outcomes: including smoking cessation; increased use of folic acid; breastfeeding; greater odds of obtaining ANC; and lower rates of neonatal mortality [71,72]. This vital component of the SRMNH care continuum lacks our studies reviewed. This review suggests an urgent need to include PCC in the SRMNH care continuum and studies in LLMICs.
The review has several limitations. Despite the inclusion of methodologically moderate-to high-quality intervention studies retrieved from seven databases, some studies may have been missed as our search terms may not have been sufficient to retrieve them. The use of a deductive qualitative content analysis guided by the WHO-HRH action framework may have resulted in the loss of contextual nuances that may have provided insight into the processes that enabled the successful delivery of HRH interventions across the SRMNH care continuum. On the other hand, the number of studies with HRH interventions having a positive effect on SRMNH care quality across the continuum was more than those reporting negative effects, and hence there might be a potential publication bias.

Conclusions
Policy-makers in LLMICs should include respectful, woman-centred care as a key part of RBF initiatives and ensure healthcare workers' incentives are locally administered and CHWs are integrated into the healthcare system. Skills-based, regularly run, effectiveness focused, onsite, and clinical simulation and/ or mobile phone-assisted in-service training of skilled personnel are needed. Besides, facility-based maternal and perinatal death audits in all health facilities, involving skilled and lay personnel, the implementation of ≥ 4 HRH interventions that target the different components of the healthcare system, and the inclusion of PCC in the SRMNH care continuum and studies are recommended.