Early care and support for young children with developmental disabilities and their caregivers in Uganda: The Baby Ubuntu feasibility trial

Background: Early care and support provision for young children with developmental disabilities is frequently lacking, yet has potential to improve child and family outcomes, and is crucial for promoting access to healthcare and early education. We evaluated the feasibility, acceptability, early evidence of impact and provider costs of the Baby Ubuntu participatory, peer-facilitated, group program for young children with developmental disabilities and their caregivers in Uganda. Materials and methods: A feasibility trial, with two parallel groups, compared Baby Ubuntu with standard care. Caregivers and children, aged 6–11 months with moderate-severe neurodevelopmental impairment, were recruited and followed for 12 months. Quantitative and qualitative methods captured information on feasibility (ability to recruit), acceptability (satisfactory attendance), preliminary evidence of impact (family quality of life) and provider costs. Results: One hundred twenty-six infants (median developmental quotient, 28.7) were recruited and randomized (63 per arm) over 9 months, demonstrating feasibility; 101 (80%) completed the 12-month follow-up assessment (9 died, 12 were lost to follow up, 4 withdrew). Of 63 randomized to the intervention, 59 survived (93%); of these, 51 (86%) attended ≥6 modules meeting acceptability criteria, and 49 (83%) completed the 12 month follow-up assessment. Qualitatively, Baby Ubuntu was feasible and acceptable to caregivers and facilitators. Enabling factors included community sensitization by local champions, positive and caring attitudes of facilitators toward children with disability, peer support, and the participatory approach to learning. Among 101 (86%) surviving children seen at 12 months, mixed methods evaluation provided qualitative evidence of impact on family knowledge, skills, and attitudes, however impact on a scored family quality of life tool was inconclusive. Barriers included stigma and exclusion, poverty, and the need to manage expectations around the child’s progress. Total provider cost for delivering the program per participant was USD 232. Conclusion: A pilot feasibility trial of the Baby Ubuntu program found it to be feasible and acceptable to children, caregivers and healthcare workers in Uganda. A mixed methods evaluation provided rich programmatic learning including qualitative, but not quantitative, evidence of impact. The cost estimate represents a feasible intervention for this vulnerable group, encouraging financial sustainability at scale. Clinical trial registration: [https://doi.org/10.1186/ISRCTN44380971], identifier [ISRCTN44380971]. A pilot feasibility single-blind, randomized controlled trial with two parallel groups across two study sites, one urban (Kampala) and one rural (Nakaseke), was undertaken. Neither site had existing formal support services for children with developmental disabilities, although referrals to specialist services including pediatric neurology and physiotherapy were possible (Supplementary Figure 1). Full details of the research methodology are described in the published protocol (33). Trial participants were infants aged 6–11 months with moderate-severe NDI [defined as a Griffiths Mental Developmental Scales (35) (quotient <70) and/or Hammersmith Infant Neurological Examination (36) (score <60)], and their primary caregivers (most frequently the mother however may be another relative or carer depending on each family’s individual circumstances), and from whom informed written consent was obtained. Exclusion criteria included: age ≥12 months; medical conditions requiring inpatient treatment; unwilling/unable to attend the full program; main residence outside a pre-defined geographic site criterion; non-Luganda or English speakers. Witnessed consent using a thumb print was available to caregivers with low literacy. At-risk infants in the community were screened for eligibility using the Malawi Developmental Assessment Tool (MDAT) (37). Comprehensive neurodevelopmental assessment using the Griffiths Mental Developmental Scales (GMDS) and HINE, was performed for those screening positive for developmental delay. Infants and their caregivers were randomized in a 1:1 ratio to the intervention or standard care (SC) arm using a random number generator prior to the commencement of the study, as previously described (33). The Baby Ubuntu program manual is freely available to download (footnote 1). The program is divided into ten modules, each lasting 2–3 h, covering understanding disability, positioning and carrying, feeding, mobilizing, communication, play, everyday activities, and the child within the community (34). Modules are delivered over 4–6 months, incorporating at least one home visit. The Baby Ubuntu groups are facilitated by an “Expert Parent,” themselves a caregiver of a child with developmental disability, with or without a healthcare professional. Facilitators receive 5 days of structured training with ongoing supervision and mentorship by a Baby Ubuntu “Master Trainer.” The program manual is freely available to download. Program groups of 6–10 participating families were selected based on locality. Standard care (SC) referred to existing local services which includes some limited access to physiotherapy and assistive devices, seizure management, audiology, ophthalmology, and nutritional support (Supplementary Figure 1), and this group were offered entry to the program following completion of the study. Feasibility was evaluated quantitatively as the total number recruited and randomized to each arm over a pre-specified time period (9 months). Acceptability amongst caregivers and healthcare workers was assessed by the protocol violation rate, e.g., participants in the intervention arm being treated as if in the control arm or vice versa, and by pre-specified criteria for “satisfactory attendance” (≥6 modules). A number of outcome measure tools were piloted and used to examine for early evidence of impact on child and caregiver outcomes (33). These included; Family quality of life (QoL) assessed using the scored Pediatric Quality of Life Family Impact module (PedsQL) (38); Child motor functioning assessed by the mobility score of the Pediatric Evaluation Disability Inventory (PEDI-CAT) (39); Child cognitive function as assessed by the Griffiths Mental Developmental Scales (GMDS) (35); Child growth and nutritional status assessed by weight, height, occipito-frontal head circumference and estimation of hemoglobin (HemoCue AB, Angelholm, Sweden); Caregiver mental health as assessed using the Self-Referral Questionnaire (SRQ) and the Parenting Stress index (PSI) (40); Caregiver-child attachment using the Maternal Infant Responsiveness Instrument (MIRI) (41); and Quality of the home environment assessed using the Infant Toddler-Home Observation for the Measurement of the Environment (IT-HOME) (42). For quantitative data, participants in both arms were assessed by study staff blind to trial allocation, at three time points; pre-intervention (age 6–11 months), at program completion (age 12–17 months), and at 12 months post-completion (age 18–23 months). Outcome measure tools were administered by trained study staff including nurses, doctors, physiotherapist, occupational therapist, medical clinical officer, and a clinical psychologist. A social scientist conducted in-depth interviews (IDIs) with 20 randomly selected primary female caregivers in the intervention and SC arms at baseline and endline, and nine intervention arm male caregivers at endline. Participants selected for IDI were contacted and interviewed in the local language. Interviews were conducted at the study site, and later transcribed into English by the social scientist for analysis. To further develop our understanding of program feasibility, acceptability, impact and scale-up, focus group discussions (FGDs) with female caregivers (two FGDs per site) and healthcare workers (one FGD with healthcare workers from both sites) were conducted. In addition, a stakeholder workshop for investigators, study staff, program facilitators and caregivers was held. A cost analysis was conducted to examine program costs including set-up (training, equipment and furniture, pre-program expenditure) and running costs (staff, building, supplies, transport refunds, home visits, outreach) over a 1-year time period. Costs relating to the trial, as opposed to implementation of the intervention, were not included in provider costs. Information was gathered from financial data recorded by the project implementation team and facility and program staff interviews. Costs were allocated according to the implementation activities of the program: recruitment, education sessions, and home visits. Costs were inputted into an Excel-based costing tool, in prices in the currency that the cost was incurred; British pounds (GBP) and Ugandan shillings (UGX), and annualized to obtain the economic costs. Costs were incurred in 2018 then inflated to May 2022 prices based on the consumer price index of the currency of the initial recorded cost.2 Finally, all costs were converted to 2022 US dollars (USD). The trial aimed to recruit 126 children and their caregivers; 63 per arm. Allowing for a 20% dropout rate, this sample size gives 90% power to detect a minimal relative difference of 20% on PedsQL Family Impact score between the intervention and control arms, at 5% significance level, assuming a mean PedsQL score of 65 in the SC arm and SD of 20 in both arms (based on data from a pilot pre- and post-evaluation study). The provider cost analysis was performed using all participants completing the program in the intervention arm (n = 56). Feasibility of participant recruitment and randomization was assessed quantitatively by the total number recruited and randomized to each arm, with feasibility demonstrated if the target sample size of 126 was achieved within the 9-month recruitment period. Acceptability was assessed quantitatively by (i) calculating the protocol violation rate and (ii) summarizing the number of program sessions attended with 6 or more defined as acceptable. Analyses compared outcomes between intervention and control arms at the end of the program, and again 6 months later. On the advice of the DSMB and following CONSORT reporting guidelines for pilot and feasibility trials, we did not plan any formal statistical tests due to the preliminary nature of the trial; instead confidence intervals provide inference around the possible range of effect sizes. Regression models were used to adjust comparisons for baseline measures of the outcomes. Analysis was done on an intention-to-treat basis and missing data were not imputed. The DSMB did not instigate any interim analyses or stopping guidance. Qualitative data were analyzed using a thematic framework approach around the topics of feasibility, acceptability, impact and scale-up. Two social scientists reviewed the interview transcripts to identify the codes and themes based on the study objectives and other interesting themes that emerged from the data. We described the experiences of children and caregivers relating to the intervention received including the impact of the disability, parental confidence level, inclusion in community life and experiences of stigma and discrimination. We examined changes in these domains over the follow-up period and explored attributions of change. In addition, we performed social mapping of parent networks and group discussions with staff on their perspectives and experiences of using the program.