Objective: We examined the relationship between major ICU characteristics and labour cost per patient.
Design: Four-week prospective data collection, in which the hours spent by each physician and nurse on both in-ICU and extra-ICU activities were collected.
Setting: Eighty Italian adult ICUs.
Measurements and results: The cost of the time actually spent by ICU staff on ICU patients (labour cost) was computed for each participating unit, by applying to the average annual salaries the proportions of in-ICU activity working time for physicians and nurses. Multiple regression analysis was used to identify ICU characteristics that predict labour costs per patient. Labour cost per patient was positively correlated with ICU mortality and patients average length of stay (slopes =0.67, p =0.048 and 0.09, p <0.0001, respectively). Labour cost per patient decreases almost linearly as the number of beds increases up to about eight, and it remains nearly constant above about twelve beds. The number of patients admitted per physician (not per nurse) increases with the number of beds (Spearman correlation coefficient =0.567, p <0.0001).
Conclusions: Our findings suggest that ICUs with less than about 12 beds are not cost-effective.