Govert Veldhuijzen

158 Chapter 8 Bartlett’s Test of Sphericity was significant (Χ 2  = 2240.289, p<0.001), implying that our data were suitable for exploratory factor analysis. The subsequent exploratory factor analysis showed that questions could best be clustered into five clusters. 25 Compared to the original version of the GESQ, we created a new category: Skills and hospital were divided into skills and satisfaction. Hospital became a different cluster. 11 Three clusters remained the same: information before endoscopy, pain and discomfort during or after endoscopy and information after endoscopy. These titles were given to the clusters of questions after analysing which common content each cluster contained. The factor loadings are shown in table 2. The factor loadings represent the correlation between the question and the subcategory to which it belongs. An eigenvalue of >1 explains 64.9% of the total variance, both before and after rotation. Internal validity To determine the internal consistency reliability, we analysed the Cronbach α for each subcategory. These results are shown in table 3. including the item component correlation. The overall Cronbach α of the D-GESQ was 0.88. The Cronbach alpha for the subcategories ranged between 0.449 and 0.868. Information before endoscopy had a value of 0.848, skills and satisfaction were 0.868, pain or discomfort 0.831, information after endoscopy 0.724 and hospital 0.449. So, there was a high internal consistency, with an acceptable corrected item-total correlation, except for the corrected item-total correlation of the component hospital. Patient satisfaction Patient satisfaction scores were calculated; the overall mean satisfaction score was high (79.6 ±10.7 out of 100 maximum). Table 4. lists baseline characteristics that may have affected patient satisfaction. Patients who underwent endoscopy previously were more satisfied than patients who did not (p = 0.015). When patients had a history of ≤2 endoscopy procedures, they were more satisfied compared to patients who underwent <2 endoscopies (p = 0.019).