Abstract
BACKGROUND: Implementation science groups change methods into two categories: (1) clinical, behavioral, or biomedical intervention targeting recipient's health outcomes and (2) implementation strategies targeting the delivery system. Differentiating interventions from strategies based on their intended functions is critical to accurately attributing their effects to health or implementation outcomes. However, in coordinating 200+ HIV implementation research projects and conducting systematic reviews, we identified change methods that had characteristics of both interventions and strategies that were inconsistently categorized. To alleviate confusion and improve change method specification, we propose that implementation science should adopt an extant but rarely used term-adjunctive interventions-to classify change methods that are distinct from the common intervention/strategy taxonomy. MAIN TEXT: Adjunctive interventions as change methods that target recipients (e.g., patients, participants) of a health intervention but are designed to increase recipients' motivation, self-efficacy, or capacity for initiating, adhering to, complying with, or engaging with the health intervention over time. In two of our published reviews on implementation of HIV interventions, 25 out of 45 coded change methods fell into this gray area between strategy and intervention. We also noted instances in which the same change method was labelled as the intervention ("the thing"), as an adjunctive intervention, or an implementation strategy in different studies-further muddying the waters. Adjunctive interventions are distinguished from other change methods by their intended targets, desired outcomes, and theory of action and causal processes. Whereas health interventions target recipients and have a direct, causal effect on the health outcome, adjunctive interventions enhance recipients' attitudes and behaviors to engage with the intervention and have an indirect causal link to the health outcome via increasing the probability of recipients' utilization and adherence to the intervention. Adjunctive interventions are incapable of directly producing the health outcome and will themselves require implementation strategies to effectively impact sustained uptake, utilization, and adherence. Case examples, logic modeling, and considerations (e.g., relationship to consumer engagement strategies) for adjunctive intervention research are provided. CONCLUSION: Conceptualizing adjunctive interventions as a separate type of change method will advance implementation research by improving tests of effectiveness, and the specification of mechanisms and outcomes.