Facebook
Twitter
LinkedIn
Email
 |
Checking for direct PDF access through Ovid | |
Excerpt
Method: A data set was distributed to the 77 hospitals performing cardiac surgery in Germany. Nine hospitals responded. All patients (n = 7149) undergoing coronary artery bypass grafting (CABG) and/or heart valve surgery between 1 January and 30 June 2000 were included. From this data the score was calculated. Missing items entered the score as zero. ANOVA was used for statistics. The performance of the score in this patient population was assessed by the receiver operator curve (ROC).
Results: The number of patients per hospital ranged from 60 to 1704. The mean age of all patients was 66 ± 10 years. 27% of all patients were female. The included operations were CABG (71%), valve replacement (19%) and CABG + valve replacement (10%). The mortality rate was 1.5% for all patients (1.0% CABG, 1.6% valve, 4.5% combined procedure). The mean Higgins score for each hospital ranged between 2.1 and 3.5 (P < 0.05). The mortality increased steadily from 0.3% in the lowest risk class up to 10.1% in class with 10 or more points (see figure). The occurrence of individual risk factors in the hospitals was quite scattered. The occurrence of emergency cases ranged from 1.8% to 27% (P < 0.05). Cardiac re-operations differed from 2.4% to 13.3% (P < 0.05). Elevated creatinine (>141 μg L−1 and <168 μg L−1) occurred from 1.2% to 3.6% of all patients (P < 0.05). The area under the ROC curve for the data was 0.77.
Discussion: The results demonstrate that the Higgins score is valuable for comparing the case-mix severity for patient populations. Using the German cardiac-anaesthesia data set it is possible to calculate the risk score. However, creating this data set is a difficult task, which is shown by the participation of just nine out of 77 hospitals. Problems of coding still exist as demonstrated by the wide variability of coded emergency cases (1.8-27%). For measurable factors like cardiac reoperations and elevated creatinine, the difference is reliable and reflects the variation in the patient population among hospitals in Germany. The area of 0.77 under the ROC curve indicates that the Higgins score is a valuable measure for the case-mix severity for each hospital.
