Task Sharing to Support Paediatric Service Delivery in Low- and Middle-income Countries: Current Practice and a Scoping Review of Emerging Opportunities

Demographic and epidemiological changes have prompted thinking on the need to broaden the child health agenda to include care for complex and chronic conditions in the 0-19 years (paediatric) age range. The general and skilled paediatric workforce shortage especially in low- and middle-income countries (LMICs) will impede the provision of additional paediatric services. This paper examines experiences with task sharing as part of the solution to this human resources challenge in LMICs and specically looks beyond the provision of care for acute infectious diseases and malnutrition that are widely and historically shifted. Methods We (1) reviewed the Global Burden of Diseases study to understand which conditions may need to be prioritised; (2) investigated training opportunities and national policies related to task sharing (current practice) in ve purposefully selected African countries (Kenya, Uganda, Tanzania, Malawi and South Africa); and (3) summarised reported experience of task sharing and paediatric service delivery through a scoping review of research literature in LMICs published between 1990-2019 using MEDLINE, Embase, Global Health, PsycINFO, CINAHL and the Cochrane Library.


Abstract Background
Demographic and epidemiological changes have prompted thinking on the need to broaden the child health agenda to include care for complex and chronic conditions in the 0-19 years (paediatric) age range. The general and skilled paediatric workforce shortage especially in low-and middle-income countries (LMICs) will impede the provision of additional paediatric services.
This paper examines experiences with task sharing as part of the solution to this human resources challenge in LMICs and speci cally looks beyond the provision of care for acute infectious diseases and malnutrition that are widely and historically shifted.

Methods
We (1) reviewed the Global Burden of Diseases study to understand which conditions may need to be prioritised; (2) investigated training opportunities and national policies related to task sharing (current practice) in ve purposefully selected African countries (Kenya, Uganda, Tanzania, Malawi and South Africa); and (3) summarised reported experience of task sharing and paediatric service delivery through a scoping review of research literature in LMICs published between 1990-2019 using MEDLINE, Embase, Global Health, PsycINFO, CINAHL and the Cochrane Library.

Results
We found that while some training opportunities nominally support emerging roles for non-physician clinicians and nurses, formal scopes of practices often remain rather restricted and neither training nor policy seems well aligned with probable needs from high-burden complex and chronic conditions. From 83 studies in 24 LMICs, and aside from the historically shifted conditions, we found there is some evidence available for task sharing for a small set of speci c conditions (circumcision, some complex surgery, rheumatic heart diseases, epilepsy, mental health).

Conclusion
As child health strategies are further redesigned to address the previously unmet needs careful strategic thinking on the development of an appropriate paediatric workforce is needed. To achieve coverage at scale countries may need to transform their paediatric workforce including possible new roles for mid-level cadres to support safe, accessible and high-quality care.

Background
Global strategies and initiatives for reducing child mortality and morbidity have previously focused on immunisation, acute infectious diseases, and nutrition as part of the Integrated Management of Childhood Illness (IMCI) and transmission of HIV/AIDS [1,2]. The transition to the Sustainable Development Goals (SGDs) prompted the global community to look forwards to broaden the agenda as part of "child health redesign" [3]. This includes care for complex and chronic conditions in the 0-19 years age range (referred to in this paper as paediatric care) that were previously neglected and that most health systems in low-and middle-income countries (LMICs) may not be well designed to address [3,4].
Expanding services to encompass complex and chronic conditions is threatened by workforce shortages. The World Health Organization (WHO) estimates a gap in the supply of 18 million health workers by 2030 mostly in LMICs [5] where there is likely to be a speci c challenge with the skilled paediatric workforce. The density of paediatricians in 2016 was 0.5 and 6 per 100,000 children in low-income countries and lower-middle-income countries respectively, as compared with a global mean of 32 [6]. Paediatricians also tend to work in tertiary hospitals or in the private sector, leaving few supporting primary or district-level public sector care [7]. In most Sub-Saharan African countries, non-physician clinicians and nurses ll the gaps and deliver over 80% of primary care [6]. This leads to either de facto task shifting or a lack of paediatric care.
Task shifting refers to "the rational redistribution of tasks among health workforce teams. Speci c tasks are moved, where appropriate, from highly quali ed health workers to health workers with shorter training and fewer quali cations in order to make more e cient use of the available human resources for health" [8]. Task sharing, in comparison, emphasises a teambased approach where different professionals work together to deliver services [9]. Task shifting and sharing (hereinafter referred to as "task sharing") have a long history [10]. For child health it is implicit in IMCI strategies [11]. More recently it is embedded in care for non-communicable diseases [12], mental health [13] and children and adolescents with HIV/AIDS [14]. Informal (or unsupervised) task sharing often occurs in rural and remote areas where mid-level clinicians and nurses perform procedures outside of their o cial (and sometimes legal) scopes of practice [15,16].
In this paper, we aimed to understand existing, sanctioned forms of task sharing and explore emerging opportunities for task sharing to support the delivery of care for complex and chronic paediatric conditions in LMICs. We conducted three parallel activities: (1) we explored which conditions have the highest disease burden for those aged 0-19 years; (2) we investigated the training opportunities and existing policy related to task sharing that might support expanded paediatric services in ve purposefully selected African countries; and (3) we conducted a scoping review of research examining task-sharing for child and adolescent health in LMICs with a speci c focus on conditions other than acute infectious diseases and malnutrition that are historically shifted. Finally, we triangulated and synthesised ndings to summarise the opportunities, evidence, gaps and implications for paediatric service delivery in LMICs.

Understanding burden of diseases using the Global Burden of Disease 2019
To understand what conditions in the 0-19 years age range might need to be prioritised in LMICs we extracted disabilityadjusted life-years (DALYs) of level 3 causes (diseases and injuries) for the age group "< 20 years" in 2019 from the Global Burden of Disease study [17]. We did this for countries de ned by the World Bank as: high-income, upper-middle-income, lowermiddle-income, and low-income; using the patterns in the rst two as an indication of how low-income and lower-middle-income countries may change alongside their economic development. For each category of income-level, we selected the top 20 causes ranked by DALYs, and highlighted those likely to require greater emphasis in developing accessible high-quality paediatric services.
Policy and document review of training opportunities and scope of practice Second, we examined national training policies and professional scopes of practice in ve East and Southern African countries (Kenya, Uganda, Tanzania, Malawi and South Africa). All have large gaps in the availability of skilled health professionals [18] and were the common location of research in our scoping review. We characterised the different professionals offering care, the extent of their pre-service paediatrics and child health training and opportunities for post-basic training in this eld. We focused on medical doctors, nurses and non-physician clinicians (clinical o cers, clinical associates, etc.) as the cadres of interest. We searched for documents or information (e.g. from websites) from approved training institutions, relevant regulatory councils and commissions. We reviewed schemes of service, relevant acts, task sharing policies, other broad and disease-speci c national strategic plans/policies to capture their scopes of practice.
Scoping review of research literature on task sharing and paediatric service delivery Lastly, we conducted a scoping review [19] of studies examining the design and practice of task sharing for paediatric services in all LMICs (Additional le 1: Scoping review protocol and PRISMA diagram). In summary, we conducted a systematic search using MEDLINE, Embase, Global Health, PsycINFO, CINAHL and the Cochrane Library to identify relevant articles. We combined terms and phrases related to paediatrics, task sharing, different cadres commonly involved in task sharing and the Cochrane low-and middle-income lter [20]. We included all study designs published between 1990-2019 in English. The inclusion and exclusion criteria for the screening process is shown in Table 1. After two stages of independent screening by two authors, we charted data from included papers and sorted them into three major groups based on the conditions they examined: acute infectious diseases and malnutrition; surgery (with sub-categories minor surgery and other complex surgery), emergency and intensive care; and chronic conditions (sub-categories complex and chronic conditions and mental health). For included papers we described speci c health services and procedures shifted/shared, study country, study design, cadres involved, major inputs and outcomes (health worker knowledge, skill, patient outcome) as originally reported in the included papers.

Burden of disease
Focusing on those conditions not typically covered by current strategies and initiatives, Table 2 illustrates the top-ranking conditions for which services will likely need strengthening if high-quality paediatric care is to be widely accessible in LMICs. For all countries, neonatal disorders (preterm, birth asphyxia and trauma, neonatal sepsis, etc.) are the highest-ranked cause. Malaria, lower respiratory infections and diarrheal diseases are the 2nd and 3rd top-ranked causes for low-income and lowermiddle-income countries respectively but covered by existing task sharing strategies. Congenital birth defects are ranked 4th and 5th for low-income and lower-middle-income countries respectively. Lower ranked but likely causes of substantial mortality and morbidity are road injuries, drowning, con ict and terrorism that require emergency and surgical care; and haemoglobinopathies and haemolytic anaemias, asthma, epilepsy and conduct disorder that are considered chronic conditions and require long-term multiple interactions with health services. Most of these conditions are also top-ranking conditions for upper-middle-income countries and high-income countries, which suggests that they will continue to be important needs as countries develop economically. Training opportunities and scope of practice Table 3 summarises the training opportunities and scope of practice related to child health for physicians, non-physician clinicians and nurses/specialist nurses in the ve African countries examined. The full list (by country and by cadre) is available in the additional le 2. This details speci c opportunities for child health training, existing child health scopes of practice (where de ned) and summaries of national policies and planning documents relevant to child health. In the countries examined physician training generally lasts 5 to 6 years and is followed by a one-to-two years pre-licensure internship that includes some months of supervised paediatric work within a hospital. All these countries offer further specialist training in paediatrics and child health and family medicine, however graduates of these specialist medical programmes are few. It is implicit in most policies that non-specialist physicians, even if junior, are expected to provide care for chronic and complex paediatric conditions with the exception of major surgery or intensive care. As such they may be expected to supervise, teach or receive referrals from non-physician clinicians and nurses offering primary care paediatric services in the absence of a specialist paediatrician.
For non-physician clinicians, most countries have 3-4-year entry-level diplomas or Bachelor's degrees that include some elements of paediatrics and child health (mostly 3-4 short courses). These diplomas and Bachelor's degrees also require several months of internship in paediatric wards pre-licensure with a relevant regulator. The scope of practice for non-physician clinicians usually includes prescription of common medication. In schemes of service documents some countries (Kenya, Tanzania, South Africa) also explicitly permit non-physician clinicians to perform certain typically minor surgical procedures. Three countries (Kenya, Uganda and Malawi) have advanced level courses on paediatrics for non-physician clinicians while all countries have advanced diplomas in other relevant specialties (most commonly family medicine, anaesthesia, ophthalmology, ear nose and throat [ENT]), however this training is not speci c to the paediatric age group and numbers of these specialist nonphysician clinicians are much smaller than generalists.
Nurses too receive some training in child health as part of entry-level training courses and most countries have advanced diplomas in paediatric nursing, while some also have Master's-level training which requires a Bachelor's degree for entry. This arrangement also applies to other relevant specialties that are not speci c to paediatrics (e.g. mental health/psychiatric, family medicine, critical care nursing). The scope of practice for nurses is more restricted than for non-physician clinicians as in most countries general nurses are usually not authorised to prescribe. However, in primary care settings because of de facto/informal task sharing nurses may prescribe and in some countries nurses are legally allowed to prescribe selected drugs for acute and chronic illness mostly related to HIV/AIDS and tuberculosis (Kenya) or at primary care level (Malawi). In some countries, malnutrition treatment and/or mental health counselling is within nurses' scope of practice while in Tanzania nurses are not (o cially) allowed to treat severe malnutrition at health centre level.
Scoping review on task sharing and paediatrics service delivery Table 4 shows the results of the scoping review on research evidence for task sharing and paediatric service delivery. A total of 83 papers were included for data charting. 84% of the papers were published before 2010. The included studies covered 24 countries, 20 of which were African, most commonly Malawi (n = 14), Kenya (n = 12), Uganda (n = 12), South Africa (n = 8) and Tanzania (n = 6). 49 studies assessed task sharing as a new intervention, and 34 studies reported task sharing as a norm, i.e. mentioned that services were routinely delivered by non-physicians but the study aim was not assessing task sharing. 65 studies used quantitative approaches (cross-sectional (n = 25), before-after (n = 13) and non-randomised trials (n = 9)). 10 used qualitative approaches either interviews (n = 5) or case study/review (n = 5), mostly investigating how task sharing initiatives were implemented and health workers' perspectives. Another 8 studies used mixed-method approaches. For the outcomes of care that were being shared assessed (n = 73 quantitative and mixed-method studies), 18 studies assessed the lower cadres' knowledge, 28 studies their skills, and 35 patient outcomes including mortality, length of hospital stay, follow-up and adherence rates and patient satisfaction. We now consider ndings organised by our grouping of studies based on the type and complexity of conditions. Acute infectious diseases and malnutrition 44 papers examined acute infectious diseases and malnutrition, mostly examining HIV/AIDS testing, antiretroviral therapy (ART), and neonatal disorders as addressed in IM(N)CI and Emergency Triage Assessment and Treatment (ETAT). As we are more concerned with other conditions we do not present their ndings here, but detailed characteristics of these studies are presented in eTable 2 of the additional le 1.

Minor surgery
Seven studies reported early infant and male circumcision performed by clinical o cers, nurses and midwives in Kenya, Uganda and Zambia [21][22][23][24][25][26][27]. This is a highly speci c "acute" service focusing on HIV/AIDS prevention that does not generally extend the professional role too far and only requires short training (e.g. 5 days didactic and hands-on training [22]) with limited need for ongoing supervision. Studies report a relatively low adverse event rate (from 0% [27] to 4.9% highest [21]) and high patient and/or maternal satisfaction rate [23,24]. One study reported minor burn services (wound care) provided by nurses at primary care while major burns were referred to secondary hospitals [28].

Other complex surgery
Five studies reported on amputation for some complex fractures, clubfoot corrective surgery, other orthopaedic surgery, burn surgery, ENT surgery and ventriculo-peritoneal (VP) shunting [29][30][31][32][33]. Three of these examined orthopaedic surgery delivered by clinical o cers in Malawi and they reported an acceptable mortality rate when performed unsupervised as compared with specialists [31] and high cost-effectiveness [32]. One non-randomised trial in Malawi suggested that when working together in central hospitals different cases were shared between clinical o cers and physicians: most burn surgery, foreign body removal cases and ventriculo-peritoneal (VP) shunt placement were performed by clinical o cers whereas general surgery, urology and congenital cases were more often performed by physicians, both groups had similar mortality and complication rates [29].
Another study focusing on VP-shunting in Malawi suggested that clinical o cers operating alone had a slighter higher mortality rate than with a surgeon present (6.6% vs. 5.9%), but comparable infection and shunting revision rates [33].

Emergency care
Five studies reported on "emergency care" in Kenya, Uganda and Ghana [34][35][36][37][38]. Task sharing for emergency care usually includes additional in-service training to build on non-physician clinicians and nurses' pre-service training and requires initial pairing with specialists. In one Ugandan study, nurses were trained for two years as emergency care providers (a new cadre) with the goal that they could perform assessment, diagnosis and initiate treatment independently without physician supervision. However, the mortality rate nearly doubled when they practiced unsupervised (5.04%) vs. supervised (2.90%), though for patients that were not severely ill there was no signi cant difference in mortality rate (3.09% vs. 2.17%) [36,37]. One study also examined continuous positive airway pressure for neonatal and paediatric patients in Kenya [34] and reported an overall 24% mortality rate when performed by nurses and clinical o cers. The other two studies reported only an increase in health worker knowledge of those taking on a new task [35,38].

Complex and chronic conditions
18 studies examined care for rheumatic heart diseases (RHD), epilepsy, sickle cell, asthma, eye care and tumours across seven African countries, Brazil and Fiji. Six studies examined the shifting of RHD screening to clinical o cers, nurses, midwives and other cadres. With several days of additional training, these cadres achieved substantial agreement rates in RHD diagnosis using echocardiography as compared with specialists [39][40][41][42][43][44]. Two studies further reported on RHD treatment where health worker knowledge increased after training [45,46]. One reported good patient adherence rates for monthly prophylaxis after initial diagnosis and treatment at referral hospitals followed by task shifting to health workers in local clinics [45]. Five studies investigated epilepsy. Diagnosis and management by clinical o cers and nurses achieved better patient follow-up [47] and patient outcomes, e.g. mortality rate and seizure incidence [48] when care was decentralized rather than centralized in hospitals. In a study in Kenya, epilepsy treatment was shared from clinical o cers to nurses who received additional training, dosage and management charts and continuous on-site supervision from clinical o cers. Nurses showed moderate adherence to treatment protocols [49]. However, a qualitative study in Uganda showed that clinical o cers and nurses in primary care had inadequate supervision and multidisciplinary rehabilitation team support when providing epilepsy care and they gradually lost their skills [50].

Mental health
Of six studies four were from one set of work in South Africa. These included randomised controlled trials of two different posttraumatic stress disorder (PTSD) treatments delivered by nurses for adolescents with subclinical PTSD in schools accompanied by qualitative work [51][52][53][54]. After initial diagnosis by a psychiatric nurse and/or a clinical psychologist, patients received treatment from nurses who were completing a 1-year advanced psychiatry diploma. Nurses also received group supervision every week from one clinical psychologist. Task shifting in this study achieved satisfactory health outcomes (improved patient's PTSD score, depression and global functioning [51,52]) and was well accepted by patients and nurses despite the latter initially resisting supervision [53,54]. Two other studies in Ethiopia [55] and Uganda [56] respectively reported that health worker knowledge and skills improved after training for child and adolescent mental health.

Discussion
In this review, we explore for paediatric services in LMICs likely areas of considerable service need. We focus on current approaches to training non-physicians and nurses to support such care in ve African countries and summarise existing ndings from research on task sharing for provision of complex and chronic paediatric conditions. We discuss below the implications, potential opportunities and research gaps in work on task sharing and paediatric service delivery.
Task sharing for paediatric surgery, emergency and intensive care We found some training opportunities but rather limited policy opportunities for surgery, emergency and intensive care task sharing. Non-physician clinicians could receive post-basic training in surgery, anaesthesia and emergency medicine. While most previous research evidence on task sharing to non-physician clinicians or nurses focuses on adult and obstetrics services [57], research evidence on task sharing for paediatric surgery has emerged over the past decade on circumcision, burn surgery, orthopaedics and VP shunts. Surgery for more complex cases (e.g. congenital defects) seems restricted to the few trained physicians despite a high disease burden. Similarly, while there are advanced courses on critical care nursing and reasonably well-established short-courses for emergencies e.g. ETAT/ETAT + and helping babies breathe (HBB), these short courses do not aim to formally establish new professional roles or expand scopes of independent practice. In the few studies that are done on sharing complex surgery or emergency care the mortality rate of patients managed by unsupervised clinical o cers and nurses may be higher compared with patients managed by medical doctors or supervised clinical o cers/nurses. Given the general de cits in the medical workforce especially in paediatric surgery and emergency care specialists [6, 58], it would seem worth exploring a more deliberate effort to develop speci c paediatric task sharing roles at hospital-level as has been practiced for adults in Tanzania's assistant medical o cers [59].
Task sharing for paediatric chronic conditions Task sharing for these chronic conditions is likely to occur frequently in primary care to non-physician clinicians and nurses due to the shortage of physicians at this level [6]. Nonetheless, this is not clearly re ected in their training curricula and scopes of practices. Despite some examples of advanced paediatrics and family medicine training that covered most paediatrics subspecialties, the production of such professions is relatively small. For example, in 2018 there were only 255 clinical o cers and 119 nurses with higher diplomas or master-level paediatric quali cations in Kenya despite some of these courses being introduced in the late 1970s [7,60,61].
Research evidence on task sharing for chronic conditions is limited. Studies focus on mental health, RHD and epilepsy. Most were reasonably small in scale and examined either focused initial diagnosis (echocardiography for RHD diagnosis), or followup treatment in lower-level health facilities provided by clinical o cers or nurses alone. The implementation experiences reported for mental health and epilepsy treatment suggest successful task sharing requires sustained training and supervision, uninterrupted supplies of medications and sometimes support from specialised teams to meet complex medical and rehabilitation needs [50,53]. The challenges posed are similar to those for other non-communicable diseases and with the potential need for regular, scheduled follow-up countries need to consider how best to deliver this together with effective linkages between system levels.

Implications and future considerations
Countries with very few specialists in paediatrics or family medicine and that rely on thse cadres to extend access to paediatric care for more complex and chronic conditions might take decades to achieve this given the challenges of training capacity, duration and cost. Task sharing to cades with shorter training could be one solution to this human resources gap. However, several issues need to be highlighted. Providing such paediatric services requires a system-approach with integrated models of care spanning healthcare organisations, communities, patients, and sometimes other stakeholders [62]. For example, long-term disability requires sustained interactions with the medical and rehabilitative services [4,50,63]. Careful, strategic thinking on the mix of cadres, their roles, regulation, nancing and training and supervision and management of teams and services are needed [64-66]. To inform this much more might be learned from better evaluation of existing experience.
Task sharing strategies should also be mindful of professional identities and hierarchy [66,67]. The planning needs to be context-speci c, based on countries' existing structures, available resources, previous experiences of task sharing and future planning for universal health coverage. To this end, better research is needed on the outcomes, quality of care and costs associated with task sharing if it is to be a means of improving coverage and quality of care rather than associated with the provision of "second-rate" services [68, 69].

Limitations
Our study is not without limitations. Due to data and resource availability, we present secondary data on disease burden for 2019 instead of predicting the DALYs for the future. For the training opportunities and scope of practice review, we only examined ve East and Southern Anglophone African countries. Paediatrician density is lowest in Sub-Saharan Africa [6] and non-physician clinicians are more common in this region as evidenced by the fact that most identi ed research was from these countries. For the scoping review, we are only able to search and synthesise evidence reported in the research literature, in some circumstances task sharing may already happen and become the norm, and therefore may not be reported in research papers. We also focused exclusively on task sharing to professionals in the health sectors although it is well-known that other carers play a huge role in service delivery for chronic conditions.

Conclusion
The child health redesign agenda provides an ambitious outlook for children and adolescents in the SDG era, however addressing the human resources gap is a key challenge to further expand service provision. Our review summarised the current practices and emerging opportunities for task sharing to support paediatric service delivery in LMICs. While training opportunities for expanded services exist, non-physician clinicians' and nurses' training opportunities and scopes of practice are rather restricted. Aside from the historically shifted care of acute infectious diseases and malnutrition, there is limited research evidence on outcomes and quality of care for other forms of task sharing. Service delivery arrangements for other priority conditions (congenital anomalies, major injuries, other chronic conditions e.g. cancers, hemoglobinopathies) should be the subject of future research. To achieve coverage at scale countries may need to transform their paediatric workforce including possible new roles for mid-level cadres to support safe, accessible and high-quality care.

Consent for publication
Not required.

Availability of data and material
All data relevant to the study are included in the article or uploaded as additional les.

Competing Interests
The authors declare no competing interests. Authors' Contributions YZ and ME conceived of the analysis. YZ and CH contributed to study selection, data charting and collation for the scoping review. YZ wrote the rst draft of the manuscript. ME, CH, DG and NS provided critical feedback on the rst draft of the manuscript. All authors read and approved the nal manuscript.