Zhan C, Miller M; AcademyHealth. Meeting (2003 : Nashville, Tenn.).
Abstr AcademyHealth Meet. 2003; 20: abstract no. 699.
AHRQ, Center for Quality Improvement and Patient Safety, 6011 Exec. Blvd. 200, Rockville, MD 20852 Tel. (301)593-0317 Fax (301) 594-2155
RESEARCH OBJECTIVE: Developing systems that identify medical errors and facilitate studies of risk factors and interventions has been the major focus of patient safety research since the publication of the Institute of Medicine report, To Err is Human, in late 1999. This study explores the use of administrative data and multivariable matching in epidemiological studies of medical errors. STUDY DESIGN: We used the Agency for Healthcare Research and Quality (AHRQ) Patient Safety Indicators (PSIs) to identify medical injuries during hospital stays, using AHRQ Healthcare Cost and Utilization Project (HCUP) data. Multivariable matching was used as an alternative to regression-type risk adjustment to examine the excess length of stay (LOS), charges and mortality attributable to medical injuries. Each medical injury case was matched to up to 4 controls on hospital identifier, DRG, sex, white or minority race/ethnicity, age within 10-year difference, and with difference in risk of death due to comorbidity less than 1% to control for variation in coding and practice across hospitals, principal diagnosis, sociodemographics and comorbidities. Excess adverse patient outcomes were derived as differences in outcomes between matched pairs. Paired t-test was used to test statistical significance. POPULATION STUDIED: A total of 7.45 million hospital discharges from all 994 short-term acute care hospitals in 28 states in 2000. PRINCIPAL FINDINGS: Event rates from 20 PSIs ranged from 0.004 (transfusion reactions) to 224.21 (obstetric trauma - vaginal with instrument) per 1000 discharges. The percentages of cases matched to at least 1 control varied from 21% (failure to rescue) to 99% (obstetric trauma - vaginal without instrument and C-section). Excess LOS attributable to PSIs ranged from -1.75 days (failure to rescue) to 9.42 days (post-operative wound dehiscence); excess charges from $220 (obstetric trauma - vaginal with instrument) to $57,727 (post-operative wound dehiscence); and excess mortality 2.16% (technical difficulty with medical care) to 21.96% (Septicemia) (p<0.01). CONCLUSIONS: Excess LOS, costs and mortality attributable to medical injuries during hospitalization are significant and vary by types of injuries. IMPLICATIONS FOR POLICY, DELIVERY OR PRACTICE: Despite of limited clinical details, various coding problems, low sensitivity, moderate specificity, and limited risk adjustment capacities, administrative data for its large volume and continuous availability provides a viable alternative for patient safety research while systems with richer clinical data are developed.
Publication Types:
Keywords:
- Case-Control Studies
- Costs and Cost Analysis
- Databases as Topic
- Databases, Factual
- Diagnosis-Related Groups
- Epidemiologic Studies
- Female
- Health Services Research
- Hospitalization
- Hospitals
- Humans
- Length of Stay
- Logistic Models
- Medical Errors
- Risk Adjustment
- economics
- injuries
- mortality
- statistics & numerical data
- hsrmtgs
Other ID:
UI: 102275668
From Meeting Abstracts