Page 126 - Diagnostic delay of endometriosis
P. 126

number of responding panel members multiplied by the points appointed to a number 1 position (5). For each intervention, these three scoring criteria were calculated, combined and converted into three possible outcomes: ‘selected’, ‘rejected’ and ‘no consensus’. Intervention that fulfilled all three criteria were labelled as ‘selected’, interventions that met none were labelled as ‘rejected’ and interventions that met only one or two criteria were labelled ‘no consensus’. Potential differences between the patients and professionals, or between the subgroups of participants in the expert panel (patients, general practitioners, gynaecologists and other professionals) were analysed using separate calculations of the three criteria for each categories. In addition, the Mann-Whitney U test was used for comparison of scoring behaviour in patients and professionals and the Kruskal-Wallis 1-way ANOVA for comparison between the three subgroups of professionals. If an intervention met all three criteria in at least one of the four subgroups, the intervention was labelled as ‘selected by subgroup’ even if the analysis of the complete expert panel came to a different conclusion. This created the opportunity for interventions with a strong subgroup preference to be reappraised by the rest of the panel. Additional interventions suggested by panel members were discussed by two of the researchers (MZ and AN) and, when considered of possible added value and feasible, incorporated in the second questionnaire for appraisal by the complete panel.
Second questionnaire round
The second questionnaire started with an overview of the items that were rejected and selected by the complete panel or a subgroup of experts in round one. The individual panel members were asked whether they agreed or disagreed with the decision of the panel for each intervention. The remaining interventions, which were labelled as ‘no consensus’ in the first round, and the additional interventions suggested by the panel members were resubmitted to the expert panel. The distribution of the scores of the panel members in round one were visualised in box-and-whisker plots, showing the spread in appreciation in patients and professionals combined with his or her own score in round one. Panel members who did not participate in the first round were provided with the box-and-whisker plots for patients and professionals. An example of a personalised box-and-whisker plot is given in Figure 2. All panel members, including the non-responders of round one, were asked to rate the ‘no consensus’ interventions again on a 9-point Likert scale, and were encouraged to revise their opinion considering the provided information.. The last part of the second questionnaire consisted of the personal top-3
124 | Chapter 7































































































   124   125   126   127   128