Objectives: The primary objective was to assess the utility of the number needed to treat (NNT) to inform decision-making in the context of paediatric oncology and to calculate the NNT in all superiority, parallel, paediatric haematological cancer, randomised controlled trials (RCTs), with a comparison to the threshold NNT as a measure of clinical significance.
Design: Systematic review DATA SOURCES: MEDLINE, EMBASE and the Cochrane Childhood Cancer Group Specialized Register through CENTRAL from inception to August 2018.
Eligibility criteria for selecting studies: Superiority, parallel RCTs of haematological malignancy treatments in paediatric patients that assessed an outcome related to survival, relapse or remission; reported a sample size calculation with a delta value to allow for calculation of the threshold NNT, and that included parameters required to calculate the NNT and associated CI.
Results: A total of 43 RCTs were included, representing 45 randomised questions, of which none reported the NNT. Among acute lymphoblastic leukaemia (ALL) RCTs, 29.2% (7/24) of randomised questions were found to have a NNT corresponding to benefit, in comparison to acute myeloid leukaemia (ALM) RCTs with 50% (3/6), and none in lymphoma RCTs (0/13). Only 28.6% (2/7) and 33.3% (1/3) had a NNT that was less than the threshold NNT for ALL and AML, respectively. Of these, 100% (2/2 ALL and 1/1 AML) were determined to be possibly clinically significant.
Conclusions: We recommend that decision-makers in paediatric oncology use the NNT and associated confidence limits as a supportive tool to evaluate evidence from RCTs while placing careful attention to the inherent limitations of this measure.
Keywords: clinical trials; leukaemia; lymphoma; numbers needed to treat; paediatric oncology.
© Author(s) (or their employer(s)) 2019. Re-use permitted under CC BY-NC. No commercial re-use. See rights and permissions. Published by BMJ.