Marson, Ben A.
(2022)
Development of a core outcome set for childhood lower limb fractures.
PhD thesis, University of Nottingham.
Abstract
Purpose
The selection of relevant outcome domains and the tools to measure them is a critical step in designing interventional studies. There are no agreed outcomes or outcome sets. Lower limb fractures are common in children contributing 20-30% of the inpatient fracture burden in the UK. There is an urgent need for more research to identify the most effective treatments, but it is unclear what outcome domains or tools should be used.
A core outcome set would offer researchers an agreed set of outcome domains and outcome tools to compare interventions. A core outcome set is a minimum group of outcomes that need to be measured across all trials in a condition. The aim of this thesis was to address the lack of agreed relevant outcome domains following childhood lower limb fractures, principally by developing a novel core outcome set using robust, international methodology.
Aims and objectives
This principal aim of this thesis has been separated into four objectives, each of which have formed the basis for the study design in each of the chapters. These four objectives were:
• Identification of outcome domains currently in use by experts and reported in the scientific literature
• Evaluation of outcome domains relevant to families.
• Development and agreement of the core set of outcome domains for lower limb fractures.
• Selection of the best outcome tools for one of the core outcome domains.
Methods
The initial phase of developing a core outcome set requires an agreed “what to measure”. A core set of outcome domains was developed through Chapters 2,3 and 4.
In Chapter 2, a provisional list of candidate outcome domains was established through a systematic review of the literature and identification of outcome domains reported in previous trials. Outcome domains were coded using the World Health Organization International Classification of Functioning, Disability and Health (WHO ICF).
In Chapter 3, an exploration of the experience of limb fracture from the viewpoint of children and their parents was conducted using a critical realist analysis. Interviews with child-parent dyads were completed and transcribed. A systematic review of qualitative literature was completed to provide provisional codes which were expanded and analysed to form empirical themes. Themes were subsequently abducted onto previous theories to contextualise the findings from these interviews.
The agreement of the core set of outcome domains is described in Chapter 4. The provisional list of outcome domains from Chapter 2 was expanded with an analysis of the child-parent transcripts. This expanded list was prioritised with a three-round Delphi study. The Delphi survey results were used as the basis for a face-to-face Technology of Participation consensus conference, with brainstorming, clustering, naming and resolving stages. A confirmatory vote with a 70% threshold was used to confirm the agreed core set of outcome domains.
“How to measure” was investigated in Chapter 5 for one of the core outcome domains, “return to baseline activities of daily living”. Three candidate patient reported outcomes were identified to measure this domain: 1) Activity Scale for Kids (ASK), 2) PROMIS Mobility (parent -reported) and 3) PROMIS Mobility (child-reported). Content validity, initial measurement properties including structural validity and reliability and outcome tool feasibility was assessed through professional review of candidate tools and a multi-centre validation of outcome tools. Measurement properties were compared to COSMIN standards for endorsement of measurement tools for core outcome sets.
Findings and Results
The systematic review of trials in Chapter 2 identified 100 eligible studies including 28 trials of lower limb interventions. From these, 554 distinct outcomes were extracted which mapped onto 52 domains in the WHO ICF framework. Results for lower limb trials were reported through 41 WHO ICF domains. B280 sensation of pain, s750 structure of lower extremity and s580 health services, systems and policies were reported in more than half of trials. The ASK score was also the most common patient reported outcome instrument for lower limb trials (n=4, 16.7%). The 41 WHO ICF domains were combined with the 9 additional domains reported in upper limb trials to form the preliminary set of candidate outcome domains.
In Chapter 3, the 28 higher order codes from the qualitive systematic review were expanded to 395 codes in the initial critical realist coding frame for the transcripts. The following 6 core themes were identified from the data:
• Limitations in activities and opportunities,
• Emotional drain,
• Treatments to make things better,
• How did this happen?
• Something is wrong,
• Needing help from others.
Content analysis of child-parent dyad interviews identified 911 outcome codes that were mapped onto 76 WHO ICF domains. This represented 45 new WHO ICF domains that had not been identified through analysis of previous trial reports.
The consensus study in Chapter 4 started with the candidate outcome list of 68 outcome domains. A total of 51 outcome domains met the consensus threshold as important outcomes following lower limb fractures. At the consensus meeting, the following core set of outcome domains met the final voting threshold:
• Pain and discomfort (100%)
• Return to physical and recreational activities (100%)
• Emotional & psycho-social well-being (100%)
• Recovery of mobility (100%)
• Complications from the injury and its treatment (96.7%)
• Return to baseline activities of daily living (96.7%)
• Participation in learning (93.3%)
• Appearance & deformity (93.3%)
• Time to union (80.0%)
In Chapter 5, none of the three candidate outcome tools had universal content validity. The ASK score was found to have three domains with a S-CVI>0.78 less than the 85% threshold proposed by COSMIN. These were relevance to children with lower limb fractures, relevance for clinical trials and response options are appropriate. Comprehensiveness was acceptable with a CVI of 0.86. PROMIS Mobility had one domain with a S-CVI>0.78 which was relevance for clinical trials. Comprehensiveness for this domain was low with a CVI of 0.17.
Structural validity was calculated for the ASK score. Unidimentionality was not confirmed with three potential factors in principal component analysis and 12 items with poor model fit in Rasch analysis. Intraclass correlation coefficient was 0.773 for ASK, 0.808 for parent-reported PROMIS mobility and 0.618 for child-reported PROMIS mobility. Minimal detectable change was 12.6 for ASK, 12.7 for parent-reported PROMIS mobility and 16.9 for child-reported PROMIS mobility.
High (>20%) ceiling effects were observed at all time points for all outcome scores with 34-53% ceiling effect at 10-12 weeks. The most conservative of the calculated minimal important clinical improvement was 10.1 for ASK, 7.9 for parent-reported PROMIS mobility and 6.1 for child-reported PROMIS mobility.
Conclusions
This thesis has developed understanding of relevant outcomes following childhood limb fractures through the development of a core set of outcome domains. The study has broadly followed international standards methodology and has also explored novel territory by evaluating the child reported experience of limb fracture using a critical realist framework and through the assessment of measurement properties for ASK and PROMIS mobility.
The primary limitations of this study include the restriction of systematic review and interviews to English language. Individual interviews with stakeholders other than paediatric orthopaedic surgeons who publish trials and families were not sought until the Delphi and consensus meeting which may have failed to identify additional relevant outcome domains. The meeting was hosted in the UK, and while the group was mindful of an international audience for the core set of outcome domains, additional work is required to validate this set for children in other countries. The validation study is slightly underpowered for the reliability analysis with 46 rather than the target 50 participants, though it is hoped that this can be expanded as future work.
Key findings
• Anatomical structure is the most frequently reported outcome domain in childhood lower limb fracture trials
• The experience of sustaining a limb fracture as a child seems to be moderated by previous experiences of children and caregivers.
• Different aspects of recovery have different weighting over time from the initial acute event to full rehabilitation
• A core set of outcome domains has been identified and agreed for children with lower limb injuries
• None of the currently available outcome tools are ideal for the measurement of return to baseline activities of daily living.
• A new outcome tool may be required to measure this and other outcome domains for the core set.
Actions (Archive Staff Only)
 |
Edit View |
Tools
Tools
![[img]](/style/images/fileicons/application_pdf.png)