Evidence synthesis methods for overcoming data scarcity in child health research: a focus on antiemetic medications for preventing chemotherapy-induced nausea and vomiting in children

Background

A lower disease incidence in children, compared to adults, means child evidence is often lacking. This is the case for antiemetic medications for preventing chemotherapy-induced nausea and vomiting (CINV). To address this evidence scarcity, this thesis applies evidence synthesis methods that borrow strength from indirect evidence, populations and/or outcomes, to improve understanding of intervention efficacy. The practicalities and usefulness of applying evidence synthesis methods in child health more broadly are discussed.

Methods

Randomised evidence on antiemetic medications recommended to prevent CINV in children was identified using a systematic review and synthesised using:
• Bayesian Network Meta-Analysis to determine relative treatment effects.
• Information-sharing models to determine whether adult evidence can improve the precision of treatment effects in children.
• Multivariate synthesis to improve understanding of poorly reported outcomes.
• Individual Participant Data network meta-regression to determine whether patient-level characteristics alter antiemetic efficacy.
Children and families with lived experience of CINV were involved to prioritise outcomes and interpret and disseminate findings.

Findings

Synthesis of 19 clinical trials of 3,523 children and 42 clinical trials of 16,838 adults confirmed:
• Regimens containing olanzapine and/or NK1 antagonists better prevent CINV but cause more side effects.
• Differences in treatment effect between populations meant broadly, adult evidence did not improve the precision of child estimates.
• Vomiting outcomes did not improve understanding of poorly reported nausea outcomes, due to insufficient correlation in treatment effects.
• Older children benefit more from regimens containing olanzapine and/or NK1 antagonists.
• Patient-important (quality-of-life-related) outcomes were poorly collected and reported.

Conclusions

The broad aim of improving understanding of antiemetic efficacy in children was achieved, indicating the potential usefulness of evidence synthesis methods in child health. However, resource requirements were high, and precision gains depended on similarities in treatment effects between populations and/or outcomes, conditions which should be considered when applying these methods in child health more broadly.