Browsing by Author "Franz Lauren"

Now showing 1 - 1 of 1
  • No Thumbnail Available
    Item
    Open Access
    Towards Naturalistic Data Collection in South Africa: Feasibility of Home-based Smartphone Recordings of Caregiver-Child Interactions for Coding with the Joint Engagement Rating Inventory (JERI)
    (2024) Ndlovu, Minkateko; de Vries, Petrus; Franz Lauren
    Naturalistic developmental behavioural interventions (NDBI) represent an evidence-based group of early interventions for autism. The COVID-19 pandemic forced autism interventions globally to switch to telehealth that utilised smartphone technology. Even though the evidence base for NDBI in low-and middle-income countries (LMIC) is very limited, early research has suggested that the tele-delivery of NDBI could be feasible in LMIC contexts such as in South Africa. The Joint Engagement Rating Inventory (JERI) is a behavioural coding system that has shown utility in measuring intervention outcomes in low-resource South African environments under controlled ‘laboratory' conditions. To date, no studies in LMIC have examined the feasibility of using smartphone recordings made by families in their own ‘naturalistic' home environments as data sources for coding with the JERI. In this study, we sought to answer two specific questions – first, to assess whether the home-based recordings of interactions between caregivers and their young autistic children had the necessary technical elements to be coded with the JERI; second, whether the JERI could be coded with confidence and whether satisfactory inter-rater reliability could be achieved when coding smartphone-recorded caregiver-child interactions. Methods Young autistic children (between 18-72 months) and their caregivers (≥18 years) were recruited as part of a larger project. Caregivers were provided with instructions to record 6- minute interactions with their child using their own smartphones and play materials available at home, pre-and-post 12 non-specialists delivered NDBI caregiver coaching sessions. Data were rated by two research-reliable JERI raters to assess 1) technical feasibility, and 2) raters' confidence, coding difficulty and inter-rater reliability using 16 pre-selected items of the JERI. Quantitative descriptive analyses were performed. Results Data were available on 18 smartphone recordings representing 108 minutes of data. All recordings had acceptable audio and visual quality and captured adequate data to allow coding on the JERI. In terms of rater confidence, the rater indicated being “sure of ratings” and “somewhat sure of ratings” for a majority of JERI items in the majority of the smartphone data (15-18/18 recordings). The rater experiences no difficulties coding five JERI items in most (12-15/18) smartphone recordings but reported difficulties coding eleven JERI items in 1-6 smartphone recordings. The JERI inter-rater agreement (within one scale point) ranged between 71-100% for all JERI items. Eleven of the 16 JERI items had weighted kappa values of one and observer estimated accuracy values of >99% (within one scale point). Conclusion Results of this study suggested that smartphone recordings of interactions between caregivers and their young autistic children were technically suitable to code with the JERI. The majority of smartphone-recorded caregiver-child interactions could be coded with only some difficulty and with good inter-rater reliability on the JERI. These findings suggest that smartphones could be used as naturalistic data collection methods to measure and track the impact of NDBI in a South African environment. Keywords: autism, NDBI, JERI, smartphones, naturalistic data collection