Intervention mapping (IM), a systematic method for the development, implementation and evaluation of health interventions outlined by Bartolomew et al. [26
], has proven to be a useful way to construct programs grounded both in theory and empirical data [28
]. IM proceeds according to the following steps. Step 1 consists of a needs assessment through a review of the scientific literature to analyse the target population, environmental conditions, and determinants of health behaviour. In step 2 the determinants of the health behaviour are used to set objectives for behaviour change, divided in broad performance objectives and concrete change objectives in terms of what a person needs to learn to change his or her behaviour. In step 3, theoretical foundations and empirically evaluated methods and strategies for behaviour change are assessed. In step 4, the methods and strategies are translated into an organized intervention. In step 5, the adoption, implementation and sustainability of the intervention is planned. In step 6, an evaluation plan is provided for and carried out. The strategies used in this project for each of the six steps in Intervention Mapping are reported on in detail below.
For step 1 and 2, we carried out a comprehensive review of the literature on 'difficult' patients. The MEDLINE, CINAHL, and PsycINFO databases were searched for English articles published between 1979 and 2004, retrieving 94 eligible papers [7
]. Next we undertook additional research to describe the health behaviour and its determinants: a qualitative interview study among patients [30
], a survey among community mental health clinicians [17
], and a Delphi-exercise among scientists/policy makers/expert-professionals [16
]. We concluded with the formulation of the overall behavioural objective of the intervention, and the more concrete change objectives.
For step 3, we made a theoretical analysis of ineffective chronic illness behaviour [Koekkoek B, Hutschemaekers G, van Meijel B, Schene A: How do patients become to be seen as 'difficult'?: a mixed-methods study in community mental health care, revision submitted], which forms the foundation of the intervention program. We conducted a review of therapeutic methods available to change determinants (assessed in step 1), reach objectives (formulated in step 2), and confront ineffective behaviours of both patients and professionals [search strategy and results available from the 1st author]. Additionally, since empirical findings were limited, we collected data from current best practice sites. We visited three well-known national best practices, specialized in three important domains of long-term non-psychotic disorders (mood disorders, substance abuse disorders, and personality disorders) for data on possible effective practice-based strategies not yet described in the literature. Selection of these best practices took place by searching national scientific and professional journals, searching conference programs and reports, and inviting leaders in the fields (e.g. professors, directors, educators) to suggest best practices.
For step 4, we consulted an expert group of clinicians, scientists, and policy makers over an extended period of time (two years). Some of these experts were participants in one of the problem analysis studies in step 1, others were invited because of their expertise in a specific therapeutic method (for instance clinical case management or behaviour therapy).
In step 5, implementation was prepared with a steering group of scientists and managers in the psychiatric service the intervention was tested in. Before an agreement was reached, the intervention was first presented to a director, a research psychiatrist, and the psychiatrist of the team in which the intervention would be implemented. Next, the intervention and its evaluation were presented to the team members who all agreed to participate. After obtaining ethical permission and the final approval of the institution's chief director, the program was implemented.
In step 6, we designed a mixed-methods pilot study to evaluate the intervention program. This pilot study consists of quantitative and qualitative measurements of outcome and process variables, and is described in more detail later.
Ethical approval was obtained for the patient-related qualitative study and the pilot study from the Institutional Review Board of the organisation the 1st author is affiliated with. Informed consent was obtained from all participating patients in aforementioned patient-related studies.