To identify and discriminate between patient and institutional determinants of investigation costs in the ICU.
All seven hospitals in the city of Winnipeg, Manitoba, Canada.
One hundred consecutive admissions to each of 11 ICUs. Two teaching hospitals (TH1 and TH2) each have three units (medical, surgical, and coronary care), the five community hospitals (CHs) have single combined units. TH1 operates an information-based management system.
Each admission was categorized as MEDICAL, SURGICAL, or CARDIAC. The frequency and cost of 17 laboratory or imaging procedures were collected for each admission. Demographic data included age, length of ICU stay, APACHE II (acute physiology and chronic health evaluation) score, therapeutic intervention scoring system (TISS) data, and ICU survival. The primary diagnosis on admission and acquisition of significant problems or complications after admission were collected.
Multivariate models revealed that length of stay, TISS score, and acquisition of a problem after ICU admission were strongly associated with increased costs in all categories (p=0.0001). Admission to TH2 was associated with greater costs in all categories (p=0.0001 MEDICAL and CARDIAC; p=0.0016 SURGICAL). Admission to a CH was associated with lower cost for SURGICAL admissions (p=0.0014), but costs at CHs were not significantly lower than at TH1 for MEDICAL (p=0.18) or CARDIAC (p=0.22) admissions.
ICU investigation costs vary significantly between institutions and are not always linked to patient-dependent factors. Acquisition of nosocomial and iatrogenic events during ICU admission increases costs dramatically. Costs are not necessarily greater in teaching hospitals.