Objective: To test the use of a human resources-based classification of levels of care of ICUs; to evaluate the match between planned vs operative levels of care on a large sample of European ICUs.
Design: Analysis of the database of a multicentric, multinational, prospective cohort study, involving 89 ICUs from 13 European areas.
Setting: Database of EURICUS-I.
Methods: Provision of resources was measured as the number of nurses per ICU bed. Use of resources was measured by the daily use of a therapeutic index (nine equivalents of nursing manpower use score, NEMS) at patient level. Work utilization ratio (WUR) indicated the total number of NEMS points actually scored divided by the total possible NEMS score on each ICU. The planned level of care (LOC) or the mean number of patients to be assisted by one nurse (P/N ratio) made available to the unit was derived from the number of nurses and the number of beds in the ICU. The operative LOC or the actual mean number of patients who were assisted by one nurse (P/N ratio) was computed by dividing the number of NEMS points equivalent to the work of three nursing-shifts (46 points) by the mean daily NEMS score at ICU level. Severity of illness was evaluated by the new simplified acute physiology score. Kappa statistics, intraclass correlation coefficients, and interrater percentage of agreement were used to evaluate the reliability of the data collected for total NEMS score. X2 statistics and one-way analysis of variance were used when appropriate.
Main results: Data of 16,047 patients (74,383 patient-days) admitted to the ICUs were analyzed. With an overall value of 26.5±9.3, the mean NEMS score at ICU level varied significantly among European areas. These differences were not explained by the severity of illness of the patients. The mean WUR was 0.73±0.29, presenting also significant differences among ICUs and European areas that were not explained by severity of illness. There was a mismatch between planned vs operative LOCs on 68 ICUs (76%); on 65 (73%), the operative LOC was lower than the planned LOC. This loss of resources concerned particularly the 61 ICUs planned to operate at LOC 3.
Conclusions: The use of human resources-based classification of LOCs is an objective method for evaluation of the match between provision and use of resources in the ICU. This study has shown a large mismatch between planned and utilized LOC in a sample of 89 European ICUs. This mismatch, suggesting an important loss of invested resources, was more apparent in the ICUs that were planned to operate at a higher level of care.