62 Chapter 4 The result of the multilevel analysis is presented in table 5. From these results, it is clear that none of the four postulated associations was (strongly) supported by the data, although for AEs, preventable AEs, and Medication related AEs the estimated odds ratios were smaller than 1. This is in line with the expected smaller risks. Table 5 - Odds ratios for higher EMRAM score calculated with a multilevel model Patients with: Odds ratio (EMRAM>=3/EMRAM<3) 95% confidence-interval (CI) Lower Upper Adverse events (AE)* 0,84 0,87 1,63 Preventable AE* 0,86 0,74 1,80 Medication related AE* 0,41 0,65 1,86 Unplanned readmissions* 1,10 0,87 1,41 Dichotomized: EMRAM012 = 0 and EMRAM3456 =1 * adjusted for: over-representation of deceased patients, type of hospital department and hospital types For length of stay, the estimated association is also not significant. The estimated difference in mean length of stay is numerically in the expected direction, amounting to 0,24 days (95% confidence interval -1,4 to 0,9 days) shorter for hospitals with higher EMRAM scores. Table 6 - Influence of EMRAM score on length of Stay (LOS) in the hospital (days). Length of Stay (LOS) in the hospital (days) * 95% confidence-interval (CI) Lower Upper Mean LOS in hospitals with EMRAM <3 6,0 5,0 7,0 Mean difference with hospitals with EMRAM >=3 -0,24 -1,4 0,9 EMRAM012 = 0 and EMRAM3456 =1 * adjusted for: over-representation of deceased patients, type of hospital department and hospital types
RkJQdWJsaXNoZXIy MjY0ODMw