Adaptation and cognitive testing of the African Children's Palliative Outcome Scale (African C-POS) with children and young people living with serious illnesses and their caregivers

Technical Summary: Adaptation and Cognitive Testing of the African Children's Palliative Outcome Scale (African C-POS)

Problem Statement
The development and evaluation of pediatric palliative care services in resource-limited settings, particularly in Africa, are hindered by a lack of robust, psychometrically sound patient-reported outcome measures (PROMs). Existing tools often fail to address the specific multidimensional symptoms and concerns of children with life-limiting illnesses in these contexts, or they lack validity for diverse developmental and cognitive ages. While the original African C-POS was developed to fill this gap, preliminary psychometric evaluations revealed limitations regarding developmental age differences, conceptual gaps (specifically in spiritual and psychosocial domains), and the clarity of items and response scales. There is a critical need to refine the instrument to ensure it is comprehensible, comprehensive, relevant, and acceptable to children, caregivers, and healthcare professionals.

Methodology
This study employed a mixed-methods design consisting of two distinct stages to adapt and validate the African C-POS, adhering to COSMIN taxonomy and Rothrock guidance for outcome measure development.

• Setting and Participants: The study was conducted in Uganda across three tertiary care centers (Mulago Hospital, Mildmay Uganda, and Kawempe Home Care). Participants included healthcare providers, academics, commissioners, children aged 7–19 years with life-limiting conditions, and parents/carers of children aged 0–19 years.
• Stage 1: Co-design Workshops: Two stakeholder workshops were held to refine the instrument. Drawing on existing evidence of priority needs, qualitative data from African children, and psychometric evaluation results, a diverse group of 47 stakeholders (including nurses, doctors, psychologists, and children with life-limiting illnesses) developed seven distinct versions of the scale. These versions addressed developmental/cognitive age differences and conceptual gaps.
• Stage 2: Cognitive Interviewing: Three iterative rounds of cognitive interviews were conducted using "think aloud" and verbal probing techniques. A total of 27 children, 19 proxy family caregivers, and 14 proxy health professionals participated across the seven versions. Interviews followed Tourangeau's four-stage model of survey response (comprehension, retrieval, judgement, and response) and utilized Willis's coding system to classify questionnaire problems.
• Data Analysis: Data were analyzed using Framework Analysis. Issues identified were mapped onto a matrix, and consensus decisions were made by the research team regarding changes to item wording, structure, and response formats.

Key Contributions and Adaptations
The study resulted in the development of seven specific versions of the African C-POS, moving away from a single measure to a suite of tools tailored to specific developmental stages and reporting contexts:

1. Child Self-Report Versions:

   • Flamingo (5–7 years): Uses a 3-point response scale and a recall period of "yesterday or today."
   • Crested Crane (8–12 years): Uses a 3-point response scale and a recall period of "past seven days."
   • Peacock (13–19 years): Uses a 5-point response scale and a recall period of "past seven days."
   • Note: Child versions utilize bird names chosen by a child advisory council to reduce stigma associated with chronological age labels.

2. Proxy Report Versions:

   • Parent Proxy A: For non-verbal children under 2 years.
   • Parent Proxy B: For children over 2 years unable to self-report.
   • Health Professional Proxy A: For children under 2 years with no family caregivers.
   • Health Professional Proxy B: For children over 2 years with no family caregivers.
   • Note: All proxy versions use a 5-point response scale and a "past seven days" recall period.

Key Results and Revisions
Cognitive testing revealed specific areas requiring modification to enhance face and content validity:

• Terminology Adjustments: The word "hurt" was replaced with "pain" for clarity. The term "anxiety" was removed as it was not understood by children and caregivers; it was replaced with "worry." The word "opportunity" in the context of asking questions about illness was removed as it was confusing.
• Item Refinement:
  • Feeding/Eating: Distinct terms were adopted: "feeding" for ages 0–7 and "eating" for ages 8–19.
  • Play/Crying: The item regarding "crying" was dropped due to ambiguity regarding the reasons for crying. The "play" item was rephrased from "felt like playing" to "being able to play" to focus on capability rather than desire.
  • Ordering: The order of items regarding "things you enjoy" and "things you usually do" was reversed to prevent perceived redundancy.
• Response Scales and Recall: The study confirmed that children aged 5–7 could reliably recall events from the past three days (or "yesterday or today") and utilize a 3-point scale, while older children could handle a 5-point scale and a 7-day recall period. The use of emoji-anchored cue cards was found to be acceptable and helpful for younger children.
• Acceptability: Content, recall times, and the length of the instrument were deemed acceptable by all participant groups.

Significance and Claims
The authors claim that the co-design and cognitive interviewing process successfully enhanced the face and content validity of the African C-POS. The resulting seven versions are described as addressing the developmental and cognitive age differences, symptoms, and concerns of children living with life-limiting conditions in Uganda.

The paper positions the refined African C-POS as a novel and robust outcome measure capable of bridging a critical health systems gap in pediatric palliative care in Africa. It asserts that the tool is underpinned by a child-and-family-driven framework and aligns with global imperatives for person-centred care. The authors state that the current versions possess robust psychometric properties suitable for driving person-centred care and quality improvement in children's palliative care. However, they modestly note that while the tool was developed using multinational data, the revised measure has only been tested in Uganda and requires further validation in other African countries before broader generalization.