Small-scale implementation with pragmatic process evaluation: a model developed in primary health care

Kirsti Malterud, Aase Aamland, Kristina Riis Iden

Research often fails to impose substantial shifts in clinical practice. Evidence-based health care requires implementation of documented interventions, with implementation research as a science-informed strategy to identify core experiences from the process and share preconditions for achievement. Evidence developed in hospital contexts is often neither relevant nor feasible for primary care. Different evidence types may constitute a point of departure, stretching and testing the transferability of the intervention by piloting it in primary care. Comprehensive descriptions of aims, context and procedures can be a more useful outcome than traditional effect studies. We present a model for small-scale implementation of relevant research evidence, monitored by pragmatic evaluation. The model, which is applicable in primary care, is supported by Weiner's theory about organizational readiness for change and consists of four steps: 1) recognize the problem - identify a workable intervention, 2) assess the context - prepare for inception, 3) pilot the intervention on site, and 4) upscale and accomplish the intervention. The process is evaluated by exploring selected relevant aspects of experiences and outcomes from the first to the last step. Process evaluation is a logical precondition for outcome evaluation - attempting to assess either the efficacy or the effectiveness of a "black box" intervention makes no sense. We argue why evidence beyond effect studies and evaluation beyond randomized controlled trials may be adequate for science-informed evaluation of a small-scale implementation project such as is often conducted by primary health care practitioners. The model is illustrated by an ongoing project, in which a strategy for upgrading the management of depression in nursing homes in Norway is currently being implemented. A flexible and manageable approach is suggested, in which the inevitable unpredictability of clinical practice is incorporated. Finding the appropriate middle ground between rigour and flexibility, some compromises must be made. Our model recognizes the skills of practical knowing as something other than traditional medical research, while maintaining academic values such as systematic and transparent reflection, using adequate tools. Considering the purpose and context of our model, we argue that these priorities, emphasizing relevance and feasibility, are strengths, not limitations.