Validating adverse events in administrative healthcare data in ireland: a retrospective chart review study.

BACKGROUND

Pneumonia, urinary tract infections, pressure ulcers and delirium are adverse events that affect older inpatients. Accurate administrative data are key to improving patient safety and healthcare quality. The aim of the study was to validate Hospital In-Patient Enquiry (HIPE) data on the occurrence of pneumonia, urinary tract infections, pressure ulcers and delirium in older patients discharged from an acute hospital in Ireland through retrospective chart review.

METHODS

A cohort of one thousand randomly selected admissions of inpatients aged over 65 from a university, tertiary hospital in 2022 were reviewed using a two-stage retrospective chart review. The researchers, healthcare professionals and patient representatives co-produced a study-specific chart review protocol and data collection instrument. HIPE data were validated by comparing the chart review data to the HIPE data. Since HIPE only codes the presence of the respective adverse event once, the comparisons between the HIPE data and the chart review data were carried out at admission level.

RESULTS

Of the 1,000 admissions reviewed, 231 (23.1%) contained at least one adverse event. At event level, 373 adverse events were identified including 133 pressure ulcers in 71 admissions, 101 delirium episodes in 100 admissions, 84 pneumonia episodes in 79 admissions and 55 urinary tract infections in 52 admissions. Of the 302 adverse events found in chart review on admission level, 96 (31.8%) of these were coded in the HIPE data and flagged by the Hospital Acquired Diagnosis indicator. Compared with chart review data, the overall sensitivity of the administrative data was low, and the specificity was high. The positive predictive values varied, and the negative predictive values were generally high. In HIPE data, 42 adverse events were found that were not identified in the chart review.

CONCLUSIONS

The results demonstrate that HIPE data may not accurately represent these specific adverse events as experienced by older patients. Improving the accuracy of these data may facilitate benchmarking of adverse events across hospitals and countries and provide opportunities for improvements in patient safety.