Association of automated data collection and data completeness with outcomes of intensive care. A new customised model for outcome prediction.

BACKGROUND: The Finnish Intensive Care Consortium coordinates a national intensive care benchmarking programme. Clinical information systems (CISs) that collect data automatically are widely used. The aim of this study was to explore whether the severity of illness-adjusted hospital mortality of Finnish intensive care unit (ICU) patients has changed in recent years and whether the changes reflect genuine improvements in the quality of care or are explained by changes in measuring severity of illness.
METHODS: We retrospectively analysed data collected prospectively to the database of the Consortium. During the years 2001-2008, there were 116,065 admissions to the participating ICUs. We excluded readmissions, cardiac surgery patients, patients under 18 years of age and those discharged from an ICU to another hospital's ICU. The study population comprised 85,547 patients. The Simplified Acute Physiology Score II (SAPS II) was used to measure severity of illness and to calculate standardised mortality ratios (SMRs, the number of observed deaths divided by the number of expected deaths).
RESULTS: The overall hospital mortality rate was 18.4%. The SAPS II-based SMRs were 0.74 in 2001-2004 and 0.64 in 2005-2008. The severity of illness-adjusted odds of death were 24% lower in 2005-2008 than in 2001-2004. One fifth of this computational difference could be explained by differences in data completeness and the automation of data collection with a CIS.
CONCLUSION: The use of a CIS and improving data completeness do decrease severity-adjusted mortality rates. However, this explains only one fifth of the improvement in measured outcomes of intensive care in Finland.