Conceptualizing patient-level adverse effects in implementation trials.

Identifying and monitoring adverse effects (AEs) are integral to ensuring patient safety in clinical trials. Research sponsors and regulatory bodies have put into place a variety of policies and procedures to guide researchers in protecting patient safety during clinical trials. However, it remains unclear how these policies and procedures should be adapted for trials in implementation science. As a starting point, we develop a conceptual model that traces causal pathways leading from implementation strategies to AEs, propose a definition and classification of such effects, and provide recommendations for monitoring and oversight. Main text We propose four major types of adverse effects for implementation trials. First, we characterize implementation strategies that lead to "proper use" of an intervention that align with AEs as conceptualized and reported in clinical trials. Second, we characterize a strategy's AEs mediated through "misuse" which involves inappropriate utilization of an evidence-based intervention (EBI). Third, we characterize a strategy which focuses on one EBI and may inadvertently cause the inappropriate discontinuation or "disuse" of other EBIs already in place, thus inducing AEs. Finally, we characterize strategies that may cause AEs by reducing the use of an EBI in the target population (i.e., "nonuse"). Based on these considerations, we propose an extended definition of adverse effects that includes harms that are causally related to implementation strategies, termed Implementation strategy Adverse Effects (IAEs). We recommend researchers, oversight committees, sponsors, and other stakeholders work together prior to trials to determine the best approaches for identifying, monitoring, and reporting IAEs. Conclusions In this paper, we develop a conceptual model to identify four types of AEs in implementation trials clarifying the mechanisms linking implementation strategies to patterns of use of the EBI and potential patient-level harms. We propose a new definition that links implementation strategies to AEs that can be used to guide conceptualization, monitoring, and oversight of potential harms in future implementation trials. Our work represents an important step towards understanding adverse effects in implementation trials and lays the groundwork for future advancement in the conceptualization of other types of adverse effects (e.g., harms to providers) encountered in implementation trials.