Dunja Dreesens

127 To ensure the credibility of the Delphi-study (chapter 3) the results of each round were shared with all participants, quantitative as well as qualitative, including the decisions the project group made based on the results. They also had the opportunity to ask questions anonymously, regarding the process, method and the outcomes. Moreover, the chosen Delphi-method was copied from the RAND cooperation, which has a lot of experience with this method. Shortcomings were that although we invited participants from all domains in health care, some domains were limited or not represented, for example health insurers; their perspective was missed in the Delphi. As the number of participants was well above the required minimum we felt that we mitigated this limitation. Another limitation could have been the accuracy of the experts’ answers. That is why we chose a broad selection of participants with different backgrounds and expertise, instead of a small panel of experts, and shared all the participants’ responses during and at the end of the Delphi procedure. The experts participating in developing the framework of patient-directed tool types (chapter 4) were involved almost during the whole process. They developed the framework during the meeting and discussed it when finalising it via e-mail. Before publication they gave their approval. The elements of the framework were cross-checked with existing literature. A potential weakness was the extent to which the participants were involved in the subject. This was counteracted by establishing a group consisting of participants with different expertise and from various backgrounds, and by providing information beforehand so we started with a common ground. The GIN Working group AID Knowledge (chapter 5) held open invitation workshops in which they presented the issue at hand. The participants, delegates of GIN conferences from all over the world, were asked for their input and feedback. The minutes were shared and participants were invited to comment. One could say that because of the highbrow character of the subject, interest in the subject is limited and participants do not represent the average group of people involved. We counteracted that with a group that has representatives with theoretical, practice and policy backgrounds. Additionally, literature and earlier research was also used as a basis and for comparison. Credibility measures within chapter six (interviews with paediatricians) were member checking and independent coding by three 3 authors. Also, field notes were taken and discussed to reveal possible biases. Credibility could have been improved further by giving the participants the opportunity to comment on the analysis findings. Due to time constraints with the research team and the participants themselves this was deemed unfeasible. Transferability Transferability of the data is a slightly more complicated matter and is regarded limited. Partially because the research was mainly executed with tool types used in the Netherlands and a Dutch clinical practice guideline (respectively chapters two, three and six). Also, transferability is limited because part of the research was meant as a starting point to be followed through. However, moving from the scoping review to the Delphi-study more healthcare domains were included. The former covered curative and palliative care and the latter included long term care and public health as well. The newly added domains did not suggest many new tool types to be included in the Delphi. Which could suggest that the scoping review data sufficiently covered tool types used in long term care and public health for a large part as well. Outcomes Chapter 7