How often are potential patient safety events present on admission?

Area of Science:

• Healthcare quality measurement
• Patient safety research
• Administrative data analysis

Background:

• Present on admission (POA) data fields were added to hospital claims in 2007.
• AHRQ Patient Safety Indicators (PSIs) were developed using administrative data without POA information.
• The study examined the impact of incorporating POA data on PSIs.

Purpose of the Study:

• To evaluate how adding present on admission (POA) data affects the accuracy of Patient Safety Indicators (PSIs).
• To determine the validity of PSIs as measures of in-hospital patient safety events after accounting for POA status.

Main Methods:

• Analysis of California and New York HCUP state inpatient databases from 2003.
• Inclusion of 13 relevant PSIs where POA information was applicable.
• Comparison of PSI outcomes with and without POA data consideration.

Main Results:

• Suspect POA coding varied between New York (17%) and California (1%-2%).
• Post-exclusion, 92%-93% of secondary diagnoses were confirmed as POA.
• POA data reclassified many cases of decubitus ulcer, hip fracture, and pulmonary embolism/DVT as not in-hospital events.

Conclusions:

• Three of 13 PSIs assessed may not be valid measures of in-hospital safety events due to POA data.
• Ten PSIs remain potentially useful for patient safety measurement even without POA codes.
• POA data is crucial for accurately distinguishing comorbidities from in-hospital complications.