Adverse events and in-hospital mortality: an analysis of all deaths in a Norwegian health trust during 2011.

BACKGROUND: The estimated number of in-hospitals deaths due to adverse events is often different when using data from deceased patients compared with that of a population experiencing adverse events. METHODS: The study was conducted at three hospitals in the Bergen Hospital Trust, including a 950-bed university hospital. The objective was to study the reported deaths and investigate the probable number of deaths caused by adverse events. Information about all patients who died in the hospitals during 2011 was retrieved from the electronic patient data management system and the medical records. All deaths were classified into two groups according to Norwegian law based on whether or not the death was sudden and/or unexpected. The cause of death in the latter group was further classified as being due to either natural or unnatural causes according to national requirements. An expert review panel screened the patient records for information regarding adverse events and possible (≥ 50%) preventability. Age, length of hospital stay, and Charlson Comorbidity Index were also registered. RESULTS: There were 59,605 unique patients admitted in 2011 and 1185 registered deaths (1.98%). The mean and median ages of the deceased were 73,8 and 78 years, respectively, and the median length of stay was 5.6 days (range). Of these deaths, 290 (24.5%) were considered sudden and/or unexpected and 218 were considered to be due to natural causes. Of the 72 unnatural deaths, 16 (1.4%) were classified as preventable or probably preventable. For 18 deaths (%) it was impossible to confirm or rule out preventability. CONCLUSIONS: Using this method, we identified a small proportion of hospital deaths that could be classified as unnatural. Furthermore, there was a ≥ 50% chance or more that 34 deaths (2.9%) were due to causes that could have been prevented.

[Medical records, DRG and intensive care patients]. / Epikriser, DRG og intensivpasienter.

BACKGROUND: In order to control the quality of the medical report after a hospital stay with regards to the stay in the intensive care unit (ICU), and to cheque for correct DRG grouping, this study of 428 patients treated in our ICU in 2003 was conducted. MATERIAL AND METHODS: All ICU patients from 2003 were found in our database, which includes specific ICD-10 diagnosis and specific ICU procedures. The medical record summarising the hospital stay (epicrisis) was retrieved for each patient from the hospital's electronic patient files and controlled for correct information regarding the ICU stay. DRG groups for each patient were retrieved from the hospital's administrative database. All stays were re-coded, with all information about the ICU stay was also included. The new DRG codes were compared with the old ones, and the difference in DRG points computed. RESULTS: The description of the stay in the ICU was missing or very insufficient in 46% of the records. In the DRG control we found that an additional 347.37 DRG points (18.4% of the original sum of all DRG points) were missing, corresponding to a loss to the hospital of 6.2 million NOK. In addition we discovered missing codes for tracheostomy corresponding to 2.8 million NOK, giving a total loss of 9 million NOK. CONCLUSION: This study confirms that an adequate description of the stay in the ICU is insufficient in a large number of medical records. This also leads to incorrect DRG grouping of many patients and significant financial losses to the hospital.