Optimizing Study Design for Evaluating Complex Interventions: An Example of a Feasibility Study in Person-Centered and Integrated Chronic Disease Care in Dutch General Practice

Background: Complex interventions are frequently used at different levels in healthcare. The main aim of this paper is to describe a method for conducting a feasibility study in preparation of an evaluation study for a complex intervention by substantiating several essential methodological choices. These choices are (A) establishing the most appropriate outcomes and instruments to measure them, including comprehensibility of questionnaires for study participants, (B) exploring the distribution and size of these outcomes in the patient target population and (C) quantifying key assumptions for the sample size calculation. We describe this method through the example of our feasibility study on a person-centered and integrated care (PC-IC) approach for multimorbidity and chronic conditions in general practice.

Methods: In 2021 we conducted a feasibility study in 7 general practices in three regions in the Netherlands. These practices replaced their standard disease management programs for diabetes mellitus type 2, cardiovascular disease, chronic obstructive pulmonary disease, and asthma with the PC-IC approach. Systematically selected questionnaires were administered to eligible patients at baseline and at 6 months, and comprehensibility of the questionnaires was assessed. We defined a composite outcome by comparing different scenarios for combining the questionnaire scores.

Results: The method for thoroughly designing an evaluation study for a complex healthcare intervention consisted of several steps. First, the measurement instruments for the feasibility study were chosen after a structured literature search, consulting experts, checking the questionnaires for comprehensibility by patients, and a consensus meeting with the project team. Next, the questionnaires were applied in the study target population for a period of 6 months. The results were then analysed to explore the distribution and size of these outcomes. Subsequently we assessed the most appropriate outcomes, which led to the creation of a composite outcome in our example. The final step was performing a sample size calculation based on the results of the feasibility study.

Conclusions: Using the described method, we conducted a feasibility study to prepare the evaluation of a complex intervention in Dutch general practice. Our paper is useful for other researchers preparing evaluation studies on complex interventions.