Dunja Dreesens

133 Hoffman et al are joined in their plea by Ubel, and Elwyn et al. The former deems that next to patient decision aids, clinical practice guidelines should be preference-sensitive so that viable choices are not taken away from patients (314). The latter expressed their appreciation for trustworthy guidelines, but preferred so called ‘customized care tools’, translate existing evidence into tools to help patients and clinicians work together to decide next steps (144). Not everybody is convinced though; Wennberg for example prefers patient decision aids (not derived from guidelines) to CPGs when it comes to patient informed choice as CPGs (used to) over-rely on expert judgement (35, 315). Two initiatives that can contribute to making clinical practice guidelines, in the Netherlands often called evidence-based guidelines, more patient-centred and preference-sensitive are GIN and the GRADE Working Group. The Guideline International Network (GIN) “facilitates networking, promotes excellence and helps our members create high quality clinical practice guidelines that foster safe and effective patient care.” (316). One way to make guidelines more patient-minded is to include patients in guideline development. The Working Group GIN Public has developed a toolbox which addresses issues such as how to include (qualitative) research on patient views in CPGs and how to recruit and support patients in guideline development (317). GRADE stands for Grading of Recommendations Assessment, Development and Evaluation and the GRADE Working Group approaches the issue from a slightly different angle. Starting in the early 2000’s, the Working Group’s aim was and is addressing the shortcomings of grading systems in health care. Therefore, the Working Group has developed a transparent approach of grading quality (or certainty) of evidence and strength of recommendations. In short, when the graded evidence is of high quality (or certainty) for the main outcomes to be achieved this would lead to a so-called strong guideline recommendation. If the graded evidence is less relevant, certain, contradictory or unavailable, this would lead to a weak (or conditional) recommendation (298). Besides being transparent, the GRADE approach is also quite complex. The grading of evidence is not an easy feat, a lot of criteria must be factored in. The grading is therefore often done by guideline development methodology specialists, who are not most of the time healthcare professionals or patients (or their representatives). The approach consists of two main steps: 1/ rating the quality of evidence; 2/ determining the strength of recommendations. In step one the level of confidence is established, working with a hierarchy of evidence; randomized trials start at the top of the grading and observational studies at the bottom. The level can be downgraded or upgraded depending on certain criteria, such as: risk of bias or large effect. There is no mention of patient preferences, outcomes preferred by patients in step one. Nor clinical practice for that matter. In step two, the quality level of evidence is combined with other factors: benefits and harms, resources and cost, and patients’ values and preferences. For an innocent bystander, it could seem that the other factors are not considered evidence. A bit more informed bystander could conclude that the other factors (benefits and harms; resources and costs; and patients’ preferences and values could be part of the evidence but as these factors often cannot be or are not studied with RCTs their starting position in the GRADE approach could be regarded as less favourable (17). Chapter 7