Bundle interventions used to reduce prescribing and administration errors in hospitalized children: a systematic review

What is known and objective: Bundle interventions are becoming increasingly used as patient safety interventions. The objective of this study was to describe and categorize which bundle interventions are used to reduce prescribing errors (PEs) and administration errors (AEs) in hospitalized children and to assess the quality of the published literature. Methods: Articles published in English and Arabic between 1985 and September 2015 were sought in MEDLINE, EMBAS and CINHAL. Bibliographies of included articles were screened for additional studies. We included any study with a comparator group reporting rates of PEs and AEs. Two authors independently extracted data, classified interventions in each bundle and assessed the studies for potential risk of bias. Constituent interventions of the bundles were categorized using both the Cochrane Effective Practice and Organization of Care Group (EPOC) taxonomy of intervention and the Behavioural Change Wheel (BCW). Results and discussion: Seventeen studies met the inclusion criteria. All bundles contained interventions that were either professional, organizational or a mixture of both. According to the BCW, studies used interventions with functions delivering environmental restructuring (17/17), education (16/17), persuasion (4/17), training (3/17), restriction (3/17), incentivization (1/17), coercion (1/17), modelling (1/17) and enablement (1/17). Nine studies had bundles with two intervention functions, and eight studies had three or more intervention functions. All studies were low quality before/after studies. Selection bias varied between studies. Performance bias was either low or unclear. Attrition bias was unclear, and detection bias was rated high in most studies. Ten studies described the interventions fairly well, and seven studies did not adequately explain the interventions used. What is new and conclusion: This novel analysis in a systematic review showed that bundle interventions delivering two or more intervention functions have been investigated but that the study quality was too poor to assess impact.