Lack of services and supports for autistic individuals has been recognized by the World Health Organization as a global public health concern [1]. In low-and-middle-income countries (LMIC), where 95% of autistic people live, services and supports are critically scarce [2,3,4,5]. The recent Lancet Commission on future care in autism proposed the necessary coordination between healthcare and other sectors such as education to address the gap in access, and to promote programmes that can be personalized, taking into account the preferences, needs, and incurred costs (financial and otherwise) of families and their autistic children [6]. Furthermore, the Commission reiterated the importance of access to timely supports services and that no one, regardless of geographic location or resource availability, should wait for extended periods of time to start interventions that could improve child and family quality of life [6]. Early intervention for autism is important because it can support growth in receptive and expressive language, as well as cognitive abilities, social skills, and adaptive behaviours, with positive downstream effects on the developmental cascade [7, 8]. Current evidence-based practice in early intervention blend developmental and behavioural approaches and incorporate caregivers in intervention planning and delivery [9, 10].

Naturalistic Developmental Behavioural Intervention (NDBI) is a class of interventions delivered by trained therapists with active caregiver involvement, leverage the opportunity to support developmental growth of children at home and in their daily lives. Early childhood is a particularly sensitive stage of human development in which the brain undergoes rapid growth and maturation, offering a critical window for supportive intervention. Interventions that promote social and communication development during this period of rapid growth have been associated with a cascade of positive short- and long-term functional outcomes [11]. The Early Start Denver Model (ESDM) is an NDBI that promotes child social engagement by embedding social learning opportunities in child preferred routines, thereby heightening the reward value of engagement, and increasing child social attention [7]. Importantly, ESDM supports the development of communication abilities in whichever form is appropriate for the individual so that the child over time becomes able to express their needs, preferences and perspectives, and develop skills that enhance their quality of life. Through ESDM coaching, caregivers can be supported in understanding which social activities their autistic child prefers and how to join with their child in those preferred activities.

Meeting the needs of young autistic children in sub-saharan Africa

A high number of individuals with neurodevelopmental disabilities live in sub-Saharan Africa [5]. This is in part due to limited disability-specific services and supports for the growing number of children now surviving the first 5 years of life [5]. The African continent is undergoing an unprecedented demographic shift in child population size and is projected to reach one billion children over the next three decades, an increase in the under-18 population by two-thirds [12]. Considering sub-Saharan Africa embraces the highest proportion of the world’s children, this demographic shift highlights the importance of developing feasible, sustainable, and contextually anchored early interventions for autistic children that can be integrated into existing systems of care.

Meeting the needs of families who care for young autistic children in LMIC is hampered by significant ongoing systemic challenges. Although 95% of the global population of children and adolescents reside in LMIC, only 10% of mental health research - and even less on neurodevelopmental conditions—has been conducted in these settings [3, 4, 6]. Although the principle of ‘task-sharing’ has been widely promoted as a potential solution to meet demand, a lack of trained professionals and a dearth of specialist services has resulted in many families in LMIC with unmet needs. Evidence-based programmes for neurodevelopmental conditions (such as NDBI) that have been developed and evaluated in high-income countries, remain inaccessible due to inadequate resources dedicated to adaptation, implementation and evaluation of contextually adapted materials and procedures. High-quality research designed to evaluate a range of child (e.g. communication, cognitive, and social development) and caregiver (e.g. caregiver wellbeing and sense of parenting competency) outcomes, alongside an implementation evaluation (e.g. examination of fidelity or the degree to which the delivery of an intervention programme utilizes and adheres to the intended materials and procedures) is limited to non-existent in LMIC [13, 14]. Promising scalable non-specialist and caregiver-mediated services and supports, adapted for racially, ethnically, and linguistically diverse families in Africa, remains a significant gap to be addressed [6].

Meeting the needs of young autistic children in South Africa

Policies that prioritize early childhood development (ECD) are emerging in sub-Saharan African countries, like South Africa [15]. The enactment of policies that advance ECD goals affirms government commitment to the early childhood period, provides a mandate that supports funding for services, and identifies those accountable for providing care. The National Integrated Early Childhood Development Policy prioritizes children with disabilities, including autism, to ensure equitable access to services [15]. This policy recognizes the South African Governments’ “responsibility to ensure a sufficient number of appropriately qualified human resources, including non-specialist early childhood development practitioners and their supervisors” [15].

Although South Africa is an upper-middle-income country, it remains an economy with one of the highest, persistent inequality rates globally [16]. The imbalance in wealth distribution, along with the inequitable adoption and prioritization of anti-poverty policies aimed at advancing social determinants of health, have engendered stark disparities in access to health and education services [17]. In South Africa, these disparities are perpetuated by the legacy of apartheid with systematic exclusion and subsequent lack of intergenerational economic mobility. This has resulted in the needs of the vast majority of individuals, who may benefit from early intervention, going unmet [18,19,20].

There are significant socioeconomic barriers that limit access to private health care services for much of the South African population, and racial variation in expressive language abilities at time of diagnosis has been reported in children accessing public health care services in the Western Cape Province of South Africa [21]. In a 2-year retrospective case review of autistic children who attended a tertiary paediatric neurodevelopmental clinic in the Western Cape Province of South Africa, while 42% of White children were non-verbal at diagnosis, 77% of Coloured children (a South African term for mixed-race), and 94% of Black children were non-verbal at presentation [21]. Similarly, a recent study in the Education system documented a higher-than-expected proportion of autistic children from White racial groups and English-speaking families, when compared with Western Cape provincial demographic data. Structural inequalities impact access to diagnostic evaluation, affordable supports and services, and public awareness of neurodevelopmental conditions [19, 20, 22].

Non-specialist intervention in low-resource settings

While questions around the importance of cumulative intervention intensity are beginning to emerge, the vast majority of evidence-based early autism interventions are intensive and delivered by trained therapists [23, 24]. Both the intensity of intervention and reliance on highly trained therapists act as implementation barriers, particularly in low resource communities. The Lancet Commission report introduced the concept of a stepped care approach to service delivery as a framework for equitable resource distribution that supports improvement in outcomes for autistic individuals. This type of approach, where the least resource-intensive service such as low-intensity, non-specialist delivered interventions are offered first, may be particularly well suited to LMIC and other low resource settings [6]. In low-resource contexts, innovative solutions involving redistribution of intervention services to both caregivers and non-specialist providers (e.g., ECD practitioners) may address the service gap impacting young autistic children and their families [25]. Task-sharing may improve equity in access, and extend healthcare delivery, particularly in under-resourced contexts. This dual redistribution of roles from specialist to non-specialist providers, and from interventionist to caregiver, is responsive to the realities of low resource environments. Task-sharing may be a key implementation strategy that advances accessibility, through principles of sustainability and scalability.

Towards cascaded task-sharing to deliver intervention in sub-saharan Africa

Task-sharing in autism intervention promotes active caregiver involvement via caregiver coaching. Caregiver-implemented intervention, where caregivers are coached in strategies to support their child’s social and communication growth during everyday activities, may be utilized to overcome some service access barriers [26]. Increasingly, evidence both from high- and low-income countries support family-centered models of early intervention, which align with the NDBI approach [12]. Given that caregivers play a central role in early intervention, it is critical that the transactional process whereby caregivers can both impact intervention outcomes and be impacted by the intervention be recognized [27]. Furthermore, factors that impact adaptive family functioning, caregiver stress and self-efficacy such as poverty, limited social support, and stressful life events are more prevalent in low-resource settings, making these important contextual considerations in caregiver-implemented intervention [28,29,30,31].

Task-sharing the coaching role in caregiver-implemented intervention, from highly trained therapist to non-specialist provider, offers a parallel opportunity to increase access to services. In a meta-analysis of non-specialist delivered intervention, only two studies included autistic participants from LMIC [32]. In these two studies ‘non-specialists’ were certified teachers and therapists who provided intervention directly to the child without caregiver involvement. In a study conducted in India, lay health workers under the supervision of specialists delivered 12 coaching sessions of a developmental autism intervention [33]. While this study provided preliminary evidence of the feasibility and effectiveness of a non-specialist delivered early autism intervention in India, implementation determinants will likely differ in other regions of the world such as Sub-Saharan Africa.

In intervention research, there is growing attention to the concepts of outcome proximity (whether outcomes mirror intervention targets or skills in domains directly targeted by the intervention) and boundedness (whether an outcome is measured in a context that differs from the intervention context) [9]. In caregiver-mediated interventions, the vast majority of studies utilize behavioural coding on study-specific scales of intervention-specific skills. The limitation of this approach is that these types of measures may only identify transient and limited changes. A recommended approach to assess clinically meaningful child gains related to the intervention, is to use measures that do not only detect changes in intervention targets (i.e., caregiver strategies taught during coaching) but capture clinically meaningful change, and measures that are administered in a context different from the intervention (i.e., child’s skills assessed during interaction with a clinician vs. during a caregiver-child interaction) in order to assess generalization of child skills across different interaction partners.

Proof-of-principle for cascaded task-sharing intervention in South Africa

Over the past eight years, the Center for Autism Research in Africa at the University of Cape Town has been studying various feasible approaches to early autism intervention ([14], p. 99–132). Our specific programme of research utilizes the Community-Early Start Denver Model (C-ESDM) materials. C-ESDM materials are open access, web-based, and designed to support families in low-resource contexts to learn NDBI strategies [34, 35]. Broadly, C-ESDM modules include strategies to: (i) increase child attention to people; (ii) increase child communication; (iii) create joint activity routines; and (iv) improve caregiver understanding of antecedents, behaviours, and consequence, to help teach new behaviours. Examples of specific intervention strategies that caregivers are coached in include: positioning (being in the child’s spotlight of attention); following the child’s lead; using gestures, sounds and speech to communicate; joining with the child in child-preferred activities; teaching the child to give, point, and show; setting up sensory-social play routines; integrating intervention strategies in everyday activities; and using antecedents, behaviours and consequences to teach new behaviours.

In the formative stages of our work we conducted five activities that set us up to complete the proof-of-principle pilot study of a cascaded task-sharing intervention in South Africa described in this manuscript (see Fig. 1).

Fig. 1

Formative activities that informed proof-of-principle cascaded task-sharing NDBI

First, we identified ECD practitioners, a non-specialist workforce supported by National policy and employed by the Western Cape Education Department, as non-specialist providers who could coach caregivers in NDBI strategies [18]. Second, we identified the Western Cape Education Department as an implementation partner [18]. Importantly, the Education Department oversees the ECD workforce, an alignment of non-specialist workforce and system of care which could support future scale-up efforts [36]. Coaching sessions were conducted at schools, with children on waiting lists for special education services, meaning participants were identified as autistic by the Western Cape Education Department but were not yet enrolled in school due to limited capacity to meet demand. Third, we identified caregiver preferences for early intervention, and examined whether joint activity routines, in which intervention strategies can be embedded, were applicable in low-resource, culturally diverse contexts in South Africa [30, 37]. Fourth, we adapted the training approach and session structure for non-specialist delivery. Modifications were made by an ESDM certified trainer and South African ESDM certified therapists, who were familiar both with intervention strategies and the South African context. A 4-day in-person training, led by South African ESDM therapists, was attended by ECD practitioners and their direct school supervisors. During the training C-ESDM provider materials introduced caregiver coaching concepts and core NDBI strategies, and ECD practitioners worked with a caregiver-child dyad to practice these strategies. The apprenticeship model for lay counsellor supervision in mental health informed ongoing ECD practitioner supervision [38]. Specifically, ECD practitioners received ongoing supervision by certified ESDM therapists, who reviewed the session plan with ECD practitioners pre-session and supported ECD practitioner reflection post-session. As ECD practitioners increased in their coaching competence, demonstrated by increasing implementation fidelity scores, the amount of supervision was scaled back.

Session structures for 12, one-hour, coaching sessions were created by the research team (see Fig. 2). A new intervention strategy was introduced with C-ESDM materials in each coaching session. The ECD practitioner then coached the caregiver in the new session skill across at least 2 caregiver-child activities. After each coaching activity, the ECD practitioner supported caregiver reflection. The session concluded with a discussion of the session skill and caregiver thoughts on how to practice the new skill across various caregiver-child routines.

Fig. 2

Caregiver coaching session structure of the proof-of-principle cascaded task-sharing NDBI

Fifth, we conducted a pre-pilot study with 2 caregiver-child dyads and completed a qualitative process evaluation of the adapted coaching approach which identified preliminary implementation determinants (barriers and facilitators) [39]. Efforts were then made to capitalize on facilitators and to mitigate barriers. Further adaptations to the intervention approach included: (i) creating and displaying simple visual aids during each coaching session to focus ECD practitioners and caregivers on session key points, (ii) focusing ongoing ECD practitioner supervision on specific coaching behaviours, including being ‘collaborative’ and ‘reflective’, (iii) school partners allocating coaching space and a laptop for sessions, and protecting ECD practitioner time to conduct sessions, (iv) downloading of C-ESDM modules onto school computers, given limited access to reliable internet (with the permission of intervention developers), and (v) building in flexibility into the coaching schedule to account for caregiver public transportation delays.

Building on our formative work, in this proof-of-principle pilot study, we set out to answer two specific objectives about our cascaded task-sharing NDBI approach that required exploration. The first objective was to determine whether our approach impacted fidelity of implementation of both non-specialist coaches and caregivers. In our first objective we specifically aimed to: (i) assess whether coaching by non-specialist providers resulted in improvements in caregiver use of intervention strategies with their young autistic child, and (ii) whether coaches were able to adhere to coaching procedures as outlined in fidelity checklists. The second specific objective was to assess whether coaching impacted key short-term child and caregiver outcomes. In our second objective we specifically aimed to: (i) assess whether signals of change were detected in child social and communication abilities, and (ii) caregiver stress and sense of competence. These specific objectives were important to answer as without information on implementation fidelity and potential child and caregiver impact, larger-scale clinical trials may not be warranted.