Use of AHRQ Patient Safety Indicators with Hospital Episode Statistics

1
Use of AHRQ Patient Safety Indicators with
Hospital Episode Statistics

• Min Hua Jen, Alex Bottle Paul Aylin
• Dr Foster Unit at Imperial College

2
Overview

• Counting harms AHRQ PSI background
• Challenges to using PSIs in the NHS
• Rates in England for eight PSIs
• Case-note audit of PSIs at 18 volunteer trusts
• Next steps

3
AHRQ PSIs

• Agency for Healthcare Research and Quality (US)
  PSIs are evidence-based
• Twenty indicators designed for use in routine
  data
• Complement purpose-built adverse event reporting
  systems such as the NRLS
• Include e.g. post-op sepsis, obstetric tears and
  foreign body left in patient
• Some need case-mix adjustment

4
Each PSI comprises

• Set of diagnosis, procedure and/or other codes in
  the numerator
• Set of diagnosis, procedure and/or other codes
  for the denominator
• Set of exclusions

5
(No Transcript)
6
List of 20 hospital level indicators

• Complications of anaesthesia (PSI 1)
• Death in low mortality DRGs (PSI 2)
• Decubitus ulcer (PSI 3)
• Failure to rescue (PSI 4)
• Foreign body left in during procedure (PSI 5)
• Iatrogenic pneumothorax (PSI 6)
• Selected infections due to medical care (PSI 7)
• Postoperative hip fracture (PSI 8)
• Postoperative haemorrhage or haematoma (PSI 9)
• Postoperative physiologic and metabolic
  derangements (PSI 10)
• Postoperative respiratory failure (PSI 11)
• Postoperative pulmonary embolism or deep vein
  thrombosis (PSI 12)
• Postoperative sepsis (PSI 13)
• Postoperative wound dehiscence in abdominopelvic
  surgical patients (PSI 14)
• Accidental puncture and laceration (PSI 15)
• Transfusion reaction (PSI 16)
• Birth trauma -- injury to neonate (PSI 17)
• Obstetric trauma -- vaginal delivery with
  instrument (PSI 18)
• Obstetric trauma -- vaginal delivery without
  instrument (PSI 19)

7
PSI adaptations required for NHS use

• Translation from ICD9 to ICD10 diagnosis codes
  was carried out using a look-up file from the
  internet
• Translation of the ICD9 procedure codes into
  OPCS4 was carried out by a trained clinical coder
  
• Conversion of DRGs into HRGs
• Validate in NHS naming and coding issues
• Liaison with Healthcare Commission

8
Rates in England and USA
Rates are per 1000 admissions
9
Median length of stay (days)
Total diff in stay in years (Denom num) n
/ 365
10
Validation first report

• All NHS trusts in England were sent in March 2007
  a report of their PSI rates with
• Crude and age- /- sex-adjusted rates
• Funnel plots showing their relative position to
  other trusts and national mean
• PSI specification
• Glossary explaining control limits etc
• Trusts were encouraged to ask questions

11
Validation case-note audit

• 18 trusts volunteered and completed audit of 2150
  records
• Case note audit of 2005/6 PSI numerator adms
  does HES record reflect reality?
• Note reason for difference, e.g. dx present on
  admission (not post-adm)

12
Results of audit (1)

• Obstetric tears were most reliable 94 for vag
  del w instrument, 82 for vag del w/o instrument
• 35 decubiti present on admission, 57 correctly
  coded
• Selected infections (12 miscoded) and post-op
  sepsis (21 miscoded) were considered gross
  underestimates

13
Results of audit (2)

• Death in low-mortality HRGs issue with HRG
  grouper giving priority to a minor op in ill
  patients
• PSI7 central line infections of more interest
• Foreign body often misinterpreted
• Post-op hip fracture rare better ways to capture
  severe in-hospital falls with HES?
• Some renaming of PSIs needed

14
Next steps (1)

• Amendment of PSI specs e.g. changing PSI7 to
  central line infections
• Calculation of rates for one other AHRQ PSI
  post-op DVT/PE
• Investigate specialty-specific indicators using
  HES

15
Next steps (2)

• Further use of HES to improve case-finding of
  existing PSIs e.g. admission for infection within
  12 months of orthopaedic procedure
• Flip side try to compare other data sources of
  PSIs e.g. bed sore register, Datix/NRLS, HPA with
  HES to find out how many HES misses and why