Interventions aiming to change multiple healthcare professional behaviors: A systematic review

Background: Healthcare professional’s (HCPs) provision of care is characterized by having to engage in an array of multiple different clinical actions. Implementation interventions sometimes single-out a given clinical action for intervention. However, changing practice to keep up with evidence-informed care can also involve supporting HCPs in changing more than one clinical behavior at a time. We aimed to identify and synthesize the characteristics of multiple behavior change interventions in HCPs.Methods: Five electronic databases were searched up to November 2023. We included randomized trials of interventions aiming to change more than one clinical behavior, where at least one of the clinical behaviors focused on the provision of behavior change advice. Two reviewers independently screened studies and extracted data. Quality was assessed using Cochrane’s Risk of Bias 2 tool and content of interventions was coded using Behavior Change Ontologies. Evidence was narratively synthesized.Results: Seventeen studies were included; all studies aimed to change multiple HCPs behaviors simultaneously, and most targeted physicians. Clinical behaviors identified in the included studies varied widely and had different levels of specificity. Based on their purpose, behaviors were categorized into seven upper-level categories: ‘provision of behavior change advice’, ‘screening/assessment/measurement’, ‘prescription of medication’, ‘referrals’, ‘treatment’, ‘diagnosis’ and ‘past medical history taking’. Interventions targeting multiple clinical behaviors were mostly brief in duration (median = 5 days) and delivered through a combination of face to face and at-a-distance modes of delivery (41%) and in a group setting (47%). Only 3 trials reported using theory to inform intervention development (18%). Each intervention used at least one behavior change technique (BCT; 22 BCTs across all studies). The most frequent BCTs included ‘set behavior goal’ (90%), ‘add objects to the environment’ and ‘present information from credible influence’ (63% each). Conclusions: Current trials of interventions aiming to change more than one HCP behavior are highly heterogeneous in the clinical behaviors targeted for change and approaches to their measurement. Greater recognition of the multiple behavior nature of such interventions, and on reporting their content accordingly, may serve to accelerate progress in understanding whether and how best to intervene on more than one clinical behavior at a time.