Hospital Inpatient Mortality Indicators for California
California Office of Statewide Planning & Development
The California Office of Statewide Planning & Development (OSHPD) monitors the construction, renovation, and seismic safety of health care facilities. OSHPD also collects data, disseminates information about California health care infrastructure, and publishes health care outcomes.
OSHPD began to generate and publicly report on eight of the 15 of the Agency for Healthcare Research and Quality (AHRQ) Quality Indicators in 2006-2007. Since 2008, OSHPD has expanded public reporting of AHRQ Inpatient Mortality Indicators from eight to 12.
Scope of Report
Report’s Intended End Users
California consumers, health care purchasers, and health care providers.
Understanding the score
Hospital Inpatient Mortality Indicators (IMIs) report ranks hospitals “better” or “worse than expected” compared to the statewide observed rate. Hospitals may be better or worse on one of the sub-measures but not on the overall measure, or vice versa.
Hospitals are mandated to quarterly report their discharge data to OSHPD. This report is released annually. Data includes data from two consecutive previous years. For example 2014 report includes data from 2011 and 2012. The report offers state- and hospital-level data for each measure.
General acute care hospitals with:
- at least one patient eligible for inclusion in the IMIs and
- patients aged 18 or older for all measures, with exception for percutaneous coronary intervention (PCI) patients aged 40 or older and hip fracture patients aged 65 or older.
- Long Term Acute Care (LTACs) hospitals
- Hospice Care
- Pediatrics Facilities
- Hospitals reporting less than 3 patients for all AHRQ IMIs
- Obstetric discharges
- Transfers in or out to other hospitals
- Craniotomy with diagnosis of head trauma
- Ruptured abdominal aortic aneurism
- Acute pancreatitis for Pancreatic Resection
- AHRQ Quality Indicators
- California hospitals’ inpatient data including patient age, length of stay, gender, race, ICD-9-CM codes, and related information is submitted to OSHPD electronically
- All Patient Refined Diagnosis Related Groups (APR-DRG) “groupers” and associated “risk of mortality” categories were added to each patient record
- The coefficients are used in the risk-adjustment process and population rates, were constructed based on the 2010 National Inpatient Sample (NIS) compiled by AHRQ Healthcare Cost and Utilization Project (HCUP).
- The AHRQ IMI software produces numerators, denominators, observed rates, expected rates, risk-adjusted rates and calculates final Risk-Adjusted Rates for each IMI
- A weighted average is used to combine the two sub-categories’ mortality rates after patient’s expected mortality rates were calculated using appropriate coefficients
- For each IMI, uses a 95% confidence interval to identify hospitals whose performance differs significantly from the state average.
- Hospitals are rated “better than expected” or “worse than expected” if their risk- adjusted death rates were significantly better or worse than the statewide observed rate
It is important to note some of the differences between the previous OSHPD publication of the AHRQ IMIs (2006- 2011 data) and this report:
- Use Version 4.5 of the AHRQ software, while the previous reports used Versions 3.2a (2006 and 2007), 4.1a (2008 and 2009) and 4.4 (2010 and 2011).
- Report 5 newly stratified sub-measures in addition to 12
overall measures. The newly stratified sub-categories are
provided by AHRQ Version 4.5 for the AAA Repair, Acute Stroke and
Pancreatic Resection. OSHPD reports both the overall and
sub-measures for Acute Stroke and Pancreatic Resection as below:
- Acute Stroke Mortality Rate
- Hemorrhagic Stroke
- Ischemic Stroke
- Subarachnoid Stroke
- Pancreatic Resection Mortality Rate
- Pancreatic Resection, Cancer
- Pancreatic Resection, Other
- Acute Stroke Mortality Rate
- Combine Unruptured Open AAA Repair Mortality Rate with Unruptured Endovascular AAA Repair Mortality Rate using a weighted average after patient’s expected mortality rates are calculated using appropriate coefficients.
- The results for three of the 15 available overall IMIs were
not reported using OSHPD 2012 data for the following reasons:
- For coronary artery bypass graft (CABG) surgery, the OSHPD California CABG Outcomes Reporting Program (CCORP) already reports hospital and surgeon-level risk-adjusted mortality rates and quality ratings using data from a clinical registry expressly created for quality monitoring and reporting. This, along with other features of the data collected by CCORP, results in superior quality assessments to those obtained from the AHRQ CABG measure.
- For acute myocardial infarction (AMI), AHRQ IMIs include two measures: one includes all AMI patients and one excludes patients transferred to another acute care hospital. Upon advice from experts on its former TAC, OSHPD decided to report only the measure that includes transfer patients. Analyses show that transfer patients were, on average, less severely ill and experienced lower mortality rates than non-transfer patients, so hospitals that received large numbers of transfer patients were not disadvantaged by this decision.
- Finally, hip replacement was not included because it lacked National Quality Forum endorsement, had a very low mortality rate and OSHPD’s former TAC questioned its value as a hospital-level reported measure.
In- Patient Mortality that occurred in a specific population per 100 discharges
- Numerator – The number of inpatient deaths that occurred in a specific denominator population.
- Denominator – For each IMI, expert clinicians used ICD-9-CM codes to select patient discharge records with diagnoses or procedures that indicate a particular condition or procedure.
- Abdominal Aortic Aneurysm (AAA) Repair, unruptured – In-hospital deaths per 100 discharges with unruptured abdominal aortic aneurysm (AAA) repair coming open and endovascular
- Carotid Endarterectomy – In-hospital deaths per 100 carotid endarterectomy (CEA) discharges for patients
- Craniotomy – In-hospital deaths per 100 discharges with craniotomy. Excludes patients with a principal diagnosis of head trauma
- Esophageal Resection – In-hospital deaths per 100 discharges with esophageal resection for cancer
Pancreatic Resection – In-hospital deaths per
100 discharges with pancreatic resection, ages 18 years and
older. Includes metrics for discharges grouped by type of
diagnosis and procedure.
- Pancreatic Resection, Cancer – Discharges with any diagnosis for pancreatic resection and pancreatic cancer.
- Pancreatic Resection, Other – Discharges for all other pancreatic resection.
- Percutaneous Coronary Intervention (PCI) – In-hospital deaths per 100 percutaneous coronary intervention (PCI) discharges for patients 40 years and older.
- Acute Myocardial Infarction (AMI) – In-hospital deaths per 100 hospital discharges with AMI as a principal diagnosis
- Acute Stroke – In-hospital deaths per 100 hospital discharges with acute stroke as a principal diagnosis. Includes metrics for discharges grouped by type of stroke: Hemorrhagic, Ischemic, and subarachnoid
- Gastrointestinal (GI) Hemorrhage – In-hospital deaths per 100 hospital discharges with gastrointestinal hemorrhage as a principal
- Heart Failure – In-hospital deaths per 100 hospital discharges with heart failure as a principal diagnosis.
- Hip Fracture – In-hospital deaths per 100 hospital discharges with hip fracture as a principal diagnosis for patients ages 65 years and older.
- Pneumonia – In-hospital deaths per 100 hospital discharges with pneumonia as a principal diagnosis
Improvement tools and resources can be found at http://www.hqinstitute.org/tools-resources