Scoping review of interventions to improve continuity of postdischarge care for newborns in LMICs

Introduction Neonatal mortality remains significant in low-income and middle-income countries (LMICs) with in-hospital mortality rates similar to those following discharge from healthcare facilities. Care continuity interventions have been suggested as a way of reducing postdischarge mortality by better linking care between facilities and communities. This scoping review aims to map and describe interventions used in LMICs to improve care continuity for newborns after discharge and examine assumptions underpinning the design and delivery of continuity. Methods We searched seven databases (MEDLINE, CINAHL, Scopus, Web of Science, EMBASE, Cochrane library and (Ovid) Global health). Publications with primary data on interventions focused on continuity of care for newborns in LMICs were included. Extracted data included year of publication, study location, study design and type of intervention. Drawing on relevant theoretical frameworks and classifications, we assessed the extent to which interventions adopted participatory methods and how they attempted to establish continuity. Results A total of 65 papers were included in this review; 28 core articles with rich descriptions were prioritised for more in-depth analysis. Most articles adopted quantitative designs. Interventions focused on improving continuity and flow of information via education sessions led by community health workers during home visits. Extending previous frameworks, our findings highlight the importance of interpersonal continuity in LMICs where communication and relationships between family members, healthcare workers and members of the wider community play a vital role in creating support systems for postdischarge care. Only a small proportion of studies focused on high-risk babies. Some studies used participatory methods, although often without meaningful engagement in problem definition and intervention implementation. Conclusion Efforts to reduce neonatal mortality and morbidity should draw across multiple continuity logics (informational, relational, interpersonal and managerial) to strengthen care after hospital discharge in LMIC settings and further focus on high-risk neonates, as they often have the worst outcomes.


INTRODUCTION
The newborn period (the first 28 days) of a child's life is when they are most vulnerable, with 2.4 million newborn deaths occurring globally in 2020. 1 Nearly two-thirds of newborn deaths in developing countries occur within the first 3 days of life. 2 A disproportionate burden is found in sub-Saharan

WHAT IS ALREADY KNOWN ON THIS TOPIC
⇒ Mortality rates for newborns in low-income and middle-income countries are a growing concern, with most deaths occurring within 30 days postdischarge.⇒ Continuity of care approaches have gained interest as a way of reducing neonatal mortality rates and achieving Sustainable Development Goal 3. ⇒ Previous scoping reviews have focused primarily on continuity of care in high-income countries.

WHAT THIS STUDY ADDS
⇒ Continuity of information provision was mostly attempted via home visits by community health workers (CHWs).⇒ Home visits were easier to implement in contexts where CHWs were well established and trusted within the community.⇒ Continuity of interpersonal relationships emerged as a salient consideration, highlighting the importance of relationships between family members, between healthcare workers or between people within the community.

BMJ Global Health
Africa (43% of newborn deaths in 2020). 1 Sustainable Development Goal 3 (SDG 3) aims for all countries to have a neonatal mortality rate (NMR) at least as low as 12 per 1000 live births. 3 4Between 1990 and 2017, the global NMR dropped from 36.6 deaths per 1000 to 18.0 deaths per 1000.However, improvements need to be accelerated in areas with persistently high NMR, such as sub-Saharan Africa. 5he WHO recommends at least four postnatal care visits in the first 6 weeks postdelivery for all babies, which can include home visits. 6The WHO also recommends that postnatal visits should include breastfeeding support and advice, assessment of danger signs, cord care and mental health support for mothers to create positive postnatal experiences. 68][9] Postnatal visits can be affected by income level, distance from facilities and mode of delivery. 8ddressing these barriers is critical, especially for sick newborns as postdischarge mortality and morbidity is a growing concern, 10 with most postdischarge deaths occurring within the first 30 days postdischarge. 11Robust follow-up systems have been highlighted as a necessity in LMICs for sick newborns. 12Continuity of care has recently been suggested as a way of reducing neonatal mortality for this group by linking pregnancy, delivery and postnatal care, as well as strengthening communication between households and hospitals, 13 14 which can lead to achievement of the SDG targets.Continuity of care is defined as 'the provision of coordinated care and services over time and across levels and disciplines, which is coherent with the patient's health needs and personal circumstances'. 15There are different approaches to care continuity, each underpinned by different assumptions or logics about how continuity might be achieved, and the levels of care and how the extent to which care is provided may differ over time.In the context of child health continuity, this has been characterised as involving time (the period from prepregnancy to childhood) and space (care from facilities to community-based care). 16ontinuity can have a positive effect on maternal and child healthcare (MCH), including improving breastfeeding initiation and duration, and outcomes related to maternal and newborn health status, such as improved maternal mental health. 17Several interventions have been implemented to improve continuity, for example, MCH booklets introduced in Angola. 18revious reviews on MCH in high-income countries have also emphasised the importance of strengthening continuity, highlighting this as a priority area for research in LMICs. 19articipatory approaches by engaging with the people who will ultimately interact with the intervention have drawn interest in recent years.These approaches aim to form collaborations with stakeholders and participants as opposed to viewing them as research subjects. 20articipatory approaches may lead to improved care continuity.For example, one review found using participatory methods to redesign forms with healthcare workers led to improved documentation. 21While participatory approaches hold promise, it is worth noting that few published studies mention power and control, which can hinder the egalitarian principles participatory approaches are based on. 22his scoping review aims to map and describe interventions and approaches that have been used in LMICs to improve care continuity for newborns, with a focus on the period after discharge from healthcare facilities.

Research questions
We used the following research questions to guide the review: 1.What interventions, innovations or support mechanisms have been proposed to support continuity in postdischarge care for newborns in LMICs? 2. To what extent have participatory methods been used to develop these interventions? 3. What are the different logics (premises or assumptions) that underpin interventions to establish continuity? 4. What opportunities and challenges for improving continuity of newborn care within the period after birth have been identified in previous studies?

METHODS
We used the five-step scoping review method outlined by Arksey and O'Malley, 23 which included setting the research questions; identifying relevant studies; study selection; data charting; collating, summarising and reporting the results.The first author (GG) incorporated the updates by Peters et al, 24 hence the population, concept, context (PCC) framework was used to guide the inclusion and exclusion criteria.

Search strategy
Initially GG conducted exploratory searches to gain familiarity with the literature, define key terms and develop the search strategy.Search terms and information sources were selected based on keywords and databases used in previous reviews in similar topic areas. 15 25-27Two librarians reviewed the search strategy to confirm all relevant terms had been included.The full search terms can be found in online supplemental appendix 1. GG searched seven databases: MEDLINE, CINAHL, Scopus, Web of Science, EMBASE, Cochrane library and (Ovid) Global health.The search was limited to papers published in English between 2001 and 2021.

Eligibility criteria
GG screened the titles and abstracts of retrieved articles using Rayyan.One of the coauthors (AR) reviewed a random 10% sample of articles to refine the process of article inclusion and exclusion; disagreements in BMJ Global Health discussion with reference to the inclusion and exclusion criteria.Subsequently GG retrieved and read the full texts of all articles remaining after title and abstract screening to assess them for inclusion.Articles had to meet the PCC criteria: the population under investigation was newborn babies (defined as babies under 28 days of age), 28 the concept was defined as continuity of care after discharge and the context was LMICs.There were no limits on the study design if primary data were used.Articles were only included if there was a clearly defined intervention to improve continuity of care for newborns in LMICs.

Data charting and synthesis
GG extracted descriptive information for each of the articles included in the review using an Excel spreadsheet (see table 1, online supplemental appendix 2).A second reviewer (AR) charted data from a random 10% sample of articles to get a second opinion on data abstraction; differences were resolved through discussion.The table provides an overview of study characteristics such as author, title, year of publication, study location, study design, aim, type of intervention, methods and sample size.
We prioritised a set of 28 core articles (which provided rich descriptions of intervention, results, contexts and outcomes) and used these as a starting point for iterative analysis of the 65 articles.Guided by the review questions, a more detailed analysis of the articles was underpinned by the continuity of care framework by Haggerty et al 29 which distinguishes between informational, management and relational continuity.Informational continuity is defined as information being shared and used appropriately; management continuity as managing an illness in a consistent and coherent manner; and relational continuity as the ongoing relationship between a patient and one or more providers.We treated these different facets of continuity as distinct logics, that is, assumptions underpinning how continuity might be achieved.In reviewing these papers, the lead author sought to interpret how different logics of continuity underpinned and informed different interventions in each of the studies.Multiple continuity logics could underpin a single intervention, for example, informational continuity might be attempted as part of an intervention while explicitly or implicitly supporting relational continuity.The logics underpinning continuity were not always clearly articulated by study authors, therefore, interpretation and judgement was needed.The first author conducted an inductive analysis of extracted data to compare and contrast findings across studies, interventions and continuity logics and produce summaries, which were then reviewed and discussed with coauthors.The analysis of core articles resulted in the addition of 'interpersonal continuity' as a new category, highlighting the importance of relationships between family members, between healthcare workers or between people within the community.
To address the third research question, the methods used in papers were assessed to see if they were participatory.Participatory studies are defined as involving the community 30 31 and designing interventions with people, not for them. 32However, there are different levels of participation, for example, who is included (community members vs end-users vs stakeholders) and when they are involved (problem definition through to assessment of the project). 33We applied the classification by Vaughn and Jacquez 20 to differentiate the participatory studies.The authors defined 'empowerment' of participants as the community leading research decision-making, while 'informing' participants was defined as information being provided to the community and 'involvement' of participants as researchers working directly with the community.
To address the fourth research question, we defined opportunities as structural or behavioural levers to continuity of care in the postdischarge period, and challenges as structural or behavioural challenges that are barriers to continuity of care in the postdischarge period.GG extracted information on opportunities and challenges from the studies that were considered pertinent to the implementation of the intervention, then conducted an inductive thematic analysis.

Description of the studies
The database search produced 5541 records, of which 3137 were removed as duplicates and 2404 were screened by title and abstract.204 articles were included in fulltext review, resulting in 65 articles in the final dataset.A Preferred Reporting Items for Systematic Reviews and Meta-Analyses extension for Scoping Reviews diagram in figure 1 summarises the search results and screening processes for this study.

Study characteristics
All articles were published from 2003 onwards, with fluctuations in the number of articles per year, as shown in figure 2. Figure 3 shows the countries where the studies were conducted.Just under two-thirds of the articles were conducted in Asia and just under a third in Africa, with two articles each from South America and the Middle East.
The final dataset included 44 quantitative articles (68.2%) and 7 mixed-methods articles (10.6%).The 14 remaining articles (21.2%) were participatory and described how the community was engaged.Of the participatory studies, only one study was qualitative, using interviews, observations and focus-group discussions.Two studies had qualitative components, including interviews or observations and focus group discussions (FGDs).Quantitative studies mostly consisted of cluster randomised controlled trials (RCTs) (27/65; 41.5%) or RCTs (14/65; 21.5%).The seven mixed-methods studies
Figure 3 Countries where studies on continuity of newborn care were conducted.

BMJ Global Health
visits, which could include administration of antibiotics or resuscitation, management of newborns or referrals to care. 37 48-58][70][71][72][73][74] Most of the 65 interventions (61.5%) involved CHWs conducting home visits as mentioned earlier.The majority of studies (36.9%) targeted pregnant women, 59 60 62 63 65-67 80 followed by mothers of term babies (35.3%). 37 42 44 46 53 54 58 60 64 80-9066 Ten studies (15.4%) focused on the implementation process itself. 35 54 55 59 63 65 75 77 98 99rticipatory studies Of all studies, 16.9% (11/65) were defined by study authors as participatory (see table 2), and 3.1% (2/65) were not defined as participatory, but used participatory elements in intervention design to different degrees, for example, consulting with key stakeholders, such as district health management, before implementing KMC interventions and essential newborn care. 40Participatory learning and action cycles were used most to facilitate participation (6/11; 54.5%), 61-63 65-67 with only two studies (2/11; 18.2%) using PDSA cycles. 59 60The participatory learning and action cycle activities led to interventions such as disseminating information via counselling women in their homes or in community centres, 59 bicycle ambulances 60 and health education on postnatal care. 59 60 small number of studies (2/11; 18.2%) focused on community mobilisation. 61 64takeholder participation varied: 6 of the 11 studies defined by authors as participatory (54.5%) included women from villages where the intervention was taking place, that is, pregnant women, recently delivered women or women of reproductive age. 60 62-67 Two studies aimed to 'inform' participants via online social networks about breastfeeding 93 and community health education about danger signs via women's groups, farmers groups and village coordination committees. 64Two studies aimed to 'involve' participants, one by designing a behaviour change intervention via input from participants such as current newborn care practices, designing behaviour change messaging, optimising available resources and providing viable alternatives. 57Another study set up community support groups to implement activities supporting postnatal care, neonatal care counselling and referrals.Results are presented under each of these categories below.Studies with more detail were included as an example from each continuity logic to illustrate the application and outcomes of the logic.

Informational continuity
Informational continuity refers to information being shared and used appropriately, which means considering patient preferences, values and context, as well as medical records. 29This approach to organising care was the most dominant, used in 81.5% of studies (53/65).Informational continuity was evident in interventions that focused on sharing information via education sessions between CHWs, who had received training, and the community (23/53; 43.4%).The interventions attempted to improve neonatal outcomes by increasing knowledge on essential newborn care, 44 48 51 57 63 64 66 95 danger signs, 49 54 exclusive breast feeding (EBF), KMC, 74 79 86 92 the importance of postnatal visits 78 and providing referrals to care, 52 55 56 87 which could include referral cards which mothers were encouraged to take to each visit. 55 56portunities These interventions intended to improve the way information was shared, and studies reported they were more accepted in regions where CHWs were well established. Information sharing was also reportedly improved by these interventions in settings where seeking care at facilities, for example, for deliveries or postnatal care, was uncommon. 48

BMJ Global Health
information to families where previously they would have received nothing.

Challenges
Sharing information was less easy in circumstances with patriarchal gender norms because men or mothers-in-law were considered the custodians of women's access to health. 63 64 86 95However, two studies focused their education sessions on men, with the aim of increasing their knowledge and improving their decision-making ability 95 or improving healthy behaviours and communication between couples. 78or example, Colbourn et al 60 combined quality improvement at facilities with participatory women's groups via a two-by-two factorial design to see if these influenced perinatal and neonatal mortality.Groups followed an action cycle to identify and prioritise maternal and neonatal health problems, pick local solutions, implement and evaluate them; groups were facilitated by volunteer facilitators.Half the groups chose to focus on maternal and neonatal health knowledge and thus had maternal and neonatal health taskforces added to the groups.The taskforce aimed to promote delivery at facilities, postnatal care and provide health education.Quality improvement at facilities consisted of small tests of change using PDSA cycles and training healthcare workers on specific clinical areas such as neonatal resuscitation drills.There was a 22% reduction in neonatal mortality in the combined intervention clusters compared with the control arm.

Relational continuity
Relational continuity refers to the ongoing relationship between a patient and one or more providers, with the understanding that consistency of core staff can provide a sense of coherence. 29This approach to organising care was evident in interventions that focused on repeated interactions (13/33; 39.4%), that is, home visits by care providers 35 44 48 50 51 54 56 57 69 74 83 86 92 because they attempted to build trust with families and improve access to care for communities who would not otherwise receive a postdischarge follow-up.

Opportunities
These interventions intended to improve the patientprovider relationship and it was reported they were easier to implement in communities where there was greater acceptance of CHWs as they were carefully selected from and resided within the community, 44 48 for example, communities showing appreciation for CHWs by providing transport or money. 54Some studies reported that the patient-provider relationship was improved by home visits by CHWs, as these communities reported a common practice of confining mothers and babies at home for 4-6 weeks after delivery, which is the period when babies are most likely to experience issues. 57 92allenges However, it was difficult to improve the patientprovider relationship in communities that faced difficulties physically accessing facilities due to geographical barriers such as mountains, 50 distance to the facility 56 and weak transport links. 86 99Relational continuity was also disrupted by women returning to their parental home for delivery, which may be far from their antenatal care facility. 66or example, de Vries et al 35 report impacts of relational continuity in neonatal care.This study describes how the Ministry of Health and two civil society organisations in Palestine trained nurses and midwives on the national postnatal care protocol including home-based postnatal care, neonatal care, postpartum complications and behavioural change communication.Nurses and midwives were given kits for home visits and conducted the first visit within 48-72 hours postdischarge.Relational continuity was built as the next two visits could be conducted at home or in a clinic with the same healthcare provider.The number of home visits for registered births increased from 5.4% to 12.9%.NMR remained the same as preintervention but qualitative interviews with both mothers and healthcare workers revealed that they thought the home visits led to increased support and thus better outcomes, though not in terms of NMR.Qualitative data showed that women felt encouraged to breastfeed and learnt about the benefits of breast feeding; EBF rates increased by 8.6% after the introduction of the programme.Healthcare workers said the visits made them more empathetic, and the mothers said they felt valued and cared for.

Interpersonal continuity
Interpersonal continuity refers to relationships between family members, between healthcare workers or relationships within the wider community.Building these relationships outside of the 'patient-provider' dyad can foster support systems and lead to improvements in care.This approach to organising care was evident in interventions that focused on healthcare workers working together (2/24; 8.3%) because they attempted to improve communication between different cadres of staff. 52 54nterpersonal continuity was seen in interventions that used CHWs to foster supportive relationships within families (7/24; 29.2%) to promote gender equity via shared household chores, 69 supportive relationships from husbands 34 95 and joint decisionmaking between couples.Interpersonal continuity was encouraged by interventions which involved community mobilisation (8/24; 33.3%), which aimed to improve discussions within the wider community on localised issues related to maternal and newborn health. 49 51 59 63-65 74 93MJ Global Health Challenges Although these interventions intended to improve interpersonal relationships, findings show that they were less able to do so in communities with patriarchal cultures.This was reportedly due to women's low levels of education and perceived low social standing, 64 women's access to health being gatekept by men (and therefore often not prioritised) 95 and sociocultural norms against women seeking care. 66portunities More participatory than non-participatory studies reported improvements among interpersonal relationships within the community, though this was only evident in studies that involved communities, as opposed to a focus on key stakeholders only.Improving relationships within the community via participatory methods was easier as facilitators were selected from the community to guide contextual intervention creation, and members of the community who were part of the intervention would often help those who were not. 63 65or example, Bich et al 34 designed a year-long community-based educational intervention targeting fathers at antenatal, delivery and postnatal periods for supporting breastfeeding practices up to 6 months in Vietnam.The intervention was integrated into routine healthcare services, messages were played on loudspeakers, and flyers, mugs and calendars were also used to provide guidance for fathers so they could better understand breast feeding and be motivated to encourage continuity of EBF after early initiation.CHWs were trained to run monthly group health education and counselling, individual counselling at birth and at four follow-up home visits and monthly social public activities.These activities were geared to sensitising fathers to breastfeeding and encouraging them to be supportive partners for early initiation and continuation of breastfeeding.A contest was held to see which father retained the most knowledge and was the most supportive; the event was open to fathers who participated in the intervention and their family to make paternal engagement in EBF normalised, thereby improving interpersonal continuity.Early initiation of breast feeding (1 hour after birth) was 48.6% in the intervention group and 35.7% in the control group, and EBF remained higher in the intervention than in the control group up to 6 months.

Management continuity
Management continuity refers to managing an illness in a consistent and coherent manner, which is particularly important when there are multiple care providers and there is a need for complementary and timely provision of care. 29This approach to organising care was evident in interventions that focused on introducing algorithms for CHWs based on national or WHO guidelines on when to treat or refer sick newborns 49 55 87 or introducing protocols on the management of illness for healthcare workers 52 54 99 (6/15; 40.0%).These interventions attempted to standardise and make consistent the care that newborns would receive.

Opportunities
Although these interventions intended to improve the way illnesses were managed, publications reported they had high rates of acceptance in circumstances where relationships had been previously built with the community because CHWs had been well established in the community for over 20 years. 48An example of ready acceptance is the community appreciating the care received by CHWs, and expressed this by providing gifts, lifts or money. 54n addition, interventions were reported to be easier to implement when researchers had been collecting data and running programmes for years. 87or example, Darmstadt et al 55 trained CHWs on pregnancy surveillance, essential newborn care, neonatal illness surveillance and management of illness based on a clinical algorithm adapted from Integrated Management of Childhood Illness in order to examine effects on NMR and the factors associated with coverage of postnatal assessments and compliance with CHW referrals.CHWs identified pregnancies and were notified when the woman was in labour so they could either be present or visit early in the postnatal period.CHWs conducted postnatal visits where they completed a standardised newborn assessment form, to identify serious illnesses requiring referral and made referral to the hospital according to the algorithm.CHWs' classification of neonates with illness had high validity compared with physicians' classification.NMR was 55.3 per 1000 for babies who had never been assessed by a CHW and 12.3 per 1000 for those who had been assessed at least once.

DISCUSSION
This scoping review aimed to map and describe interventions used in LMICs to improve continuity of care for newborns by examining which interventions have been used; to what extent participatory methods were employed to develop these interventions; how interventions attempted to establish continuity (ie, through what logics and assumptions) and what opportunities and challenges they encountered.Our findings show that the literature has primarily focused on quantitative studies, and studies in Asia, whereas the burden of neonatal mortality is greatest in Africa. 5Home visits were the most common intervention, and usually focused on providing education for mothers or community-based care for newborns.Other reviews have also recognised the importance of community-based care. 100 101Most interventions targeted pregnant women; however, one study suggested the optimal time to begin the continuum of care is prepregnancy. 102Few studies targeted mothers of babies who were high risk, such as those born preterm, LBW or sick.
Our analysis used the framework by Vaughn and Jacquez 20 to distinguish different types of participatory BMJ Global Health approaches.Studies enacted participation by families, healthcare workers and other stakeholders to different degrees.Seven studies used PDSA, participatory learning or action cycles to 'empower' communities to come up with solutions to local issues. 59 60 62 63 65-67These studies involved representatives and/or women from the community in participatory learning and action cycles or PDSA cycles, where groups identified problems, came up with solutions, implemented their ideas and assessed whether they worked.Previous research has shown that women's groups are an effective way of reducing NMR. 101 103While these studies involved members of the community in participatory learning and PDSA cycles, a review focused on RCT studies in high-income countries found that none of them involved participants in coproduction of study design or intervention. 104In our review, studies did not always engage mothers, 59 which is problematic when interventions are described as participatory yet do not include the primary carers of newborns.Furthermore, the process of engaging community members as opposed to key stakeholders in interventions can foster continuity of care by improving interpersonal continuity.The level of participant engagement varied across studies, and it was not always clearly described, with the lowest level of engagement focusing on how to 'inform' participants, for example, providing information to participants.While informational continuity was the most applied logic, the earlier example by Colbourn et al 60 highlights that participatory studies can be used to engage with participants to understand their information needs before implementing an intervention.Detailed information about participant engagement is required to understand its role in study design and coproduction of interventions.Across papers, detailed information for replicability is lacking; how specific interventions were developed by groups 66 67 and participatory involvement processes were not well described. 57 64These findings echo the call for future research papers to focus on the process involved with participatory approaches, especially in LMICs. 21n our analysis, we used the continuity framework by Haggerty et al 29 which distinguishes between three types of continuity logics (informational, relational and managerial).The literature on postdischarge care emphasised informational continuity more by focusing on information flow between the health system and carers, that is, education of mothers and families via CHWs.Sharing information with mothers and families contributed to improvements in knowledge on breast feeding and danger sign recognition.However, informed carers are not all that is required for good outcomes, and access to quality, affordable care must remain a priority.The focus on information sharing aligns with a review of continuity of care in high-income countries, which suggested further research into when and how to share information with mothers, as they received inconsistent advice and were not confident in care provision. 17In addition, information sharing and communication between healthcare workers and mothers was highlighted as a major issue during antenatal visits in four developing countries. 105though the Haggerty framework covers relational continuity in terms of patient-provider relationship, this does not extend to relationships between healthcare workers, family members and mothers and the wider community.These relationships may impact continuity of care in LMICs, where men and mothers-in-law often have the final say in careseeking. 63 64 86 95Recently literature has focused on people outside the mother-child dyad, with growing recognition that including men is important. 106 107Involving men in postnatal care is recommended by the WHO 108 and can help tackle stereotypical gender norms. 109Therefore, we have extended the continuity framework by adding interpersonal continuity as a new category to emphasise that newborn care involves multiple complex relationships which need to be supported.Although a previous framework on continuity of care also recognises the importance of co-ordination between healthcare workers, 15 the importance of family members in decision-making and care-seeking has been less explored.

Strengths and limitations
In conducting this review, we searched a large number of databases to identify a range of different types of interventions aimed at improving continuity in postdischarge care.Beyond mapping interventions and their characteristics, we used and extended an established theoretical framework on care continuity, to delve deeper in the analysis and interpret how interventions attempt to establish continuity, as well as providing a critical assessment of participatory methods used.One of the limitations of this review is that grey literature was not included which means we may have missed some interventions not described in academic publications.In addition, there were missing descriptions of approaches used, such as how participatory methods were employed.Another challenge was the diversity of study types and reported outcomes which precluded direct quantitative comparisons, but enabled us to perform rich interpretive analysis, which was the aim of the review.

Implications for practice and research
This scoping review may be of interest to policymakers, researchers and funders, when considering how different continuity interventions may best meet specific context(s) and what assumptions need to underpin these interventions.For example, home visits were the most popular intervention in the literature, but best outcomes were achieved in contexts where CHWs were well established and trusted within the community (ie, using informational and interpersonal continuity logics).In addition, some interventions faced challenges with implementation in contexts with patriarchal gender norms, hence more effort may be required for community engagement or interventions should be designed to include male-identifying community members in these settings.Some studies suggested focusing on babies at higher risk post discharge to improve outcomes 10 110 and to be cost-effective. 111Further research is required on how to develop interventions to specifically meet continuity needs of high-risk newborns.Qualitative research could provide in-depth insights on contextual acceptability, which would BMJ Global Health enhance the ability to translate interventions across contexts.We are currently conducting a review of grey literature on continuity interventions in LMICs to understand if there is additional learning to be generated from studies that have not been published in academic journals.

CONCLUSION
Most reports in the academic literature that describe research on the continuity of postdischarge care for newborns in LMICs are about home visits by CHWs, who educate mothers on newborn care practices (ie, assuming that improved information flow will result in better care) or provide communitybased care.Most studies were quantitative and focused on infants who were born at full-term and healthy.This study adds to Haggerty and Reid's continuity of care framework by including aspects of interpersonal continuity that are important in settings where relationships between family members, between healthcare workers and within the wider community are likely to influence intervention success.The review highlights that future efforts to reduce neonatal mortality and morbidity should focus further on high-risk neonates after hospital discharge in LMIC settings, as they often have the worst outcomes.
Twitter Michuki Maina @mitchmaina Contributors The conception and design of the review was conducted by all authors.Data collection, analysis and interpretation were conducted by all authors.Our thanks go to Eli Harriss and Carolyn Smith, the librarians at the University of Oxford who supported refinement of key terms and initial searches.AR screened titles and abstracts and full text papers during data collection.Writing was led by GG, with CP and SSF contributing to a series of drafts.All authors have contributed revisions, read and approved the final manuscript.GG is acting as the author guarantor.

Figure 1
Figure 1 PRISMA chart demonstrating the sources of data selected for review.ANC, antenatal care; PRISMA, Preferred Reporting Items for Systematic Reviews and Meta-Analyses.
Interventions face challenges with implementation in contexts with patriarchal gender norms, hence more effort may be required for community engagement or interventions should be designed to include men.
MIGHT AFFECT RESEARCH, PRACTICE OR POLICY ⇒ Establishing CHW programmes within communities, although challenging, may aid community-based care.⇒

Table 1
General characteristics of interventions included in the review Relational HCW, healthcare worker; KMC, kangaroo mother care; LBW, low birth weight; NMR, neonatal mortality rate; TBA, traditional birth attendant.Table 1 Continued BMJ Global Health used surveys combined with in-depth interviews or focus group discussions.

Table 2
4951 52 57 63 6674 86Publications reported information sharing was easier because home visits provided General characteristics of participatory studies included in the review