Patient Safety Indicators for Judging Hospital Performance

Area of Science:

• Healthcare Quality Improvement
• Medical Informatics
• Patient Safety Research

Background:

• Patient Safety Indicators (PSIs) are utilized for quality of care screening, hospital ranking, and reimbursement adjustments.
• PSI data accuracy relies heavily on clinical documentation and coding.
• A critical gap exists in differentiating PSI events inherent to procedures from those indicating actual safety concerns.

Purpose of the Study:

• To evaluate the reliability of specific Patient Safety Indicators (PSIs) when used for hospital performance assessment.
• To determine the proportion of false-positive PSI events and assess their clinical significance.
• To examine interrater reliability in classifying PSI events.

Main Methods:

• A retrospective analysis of 136 cases with target PSIs from a single academic center over one year.
• Two patient safety officers independently reviewed cases to assess if PSI events were inherent to the procedure or clinically significant.
• Statistical analysis of agreement between reviewers for event classification.

Main Results:

• High false-positive rates were observed, with 11.8% to 33.3% of PSI events deemed inherent to the disease process and procedure, not safety issues.
• A significant proportion of events (11.8% to 30.4%) were found to be not clinically significant.
• Moderate interrater reliability was found for three specific PSIs, indicating variability in assessment.

Conclusions:

• Current Patient Safety Indicator reporting demonstrates high false-positive rates and insufficient reliability for hospital ranking.
• The inherent nature and clinical significance of PSI events require careful evaluation beyond automated data.
• Revisions to PSI methodology are needed to improve accuracy and ensure reliable quality assessment.