• Abstracts
   
• Biosketches
   
• Partners
   
• Photos
   
• Presentations
   
• Program
   
• White Paper
   
• Workgroup
  Questions
   

Adverse Events and Deaths Associated With Laboratory Errors at a Hospital Philadelphia, 2001

S. Chang^1,2, D. Selenic^1,2, M. Bell², P. Lurie³, K. Nalluswami³, C. Coventon³, C. Schaben⁴, J. Rankin³, J. Hersh³, D. Jernigan²

1) Epidemic Intelligence Service, Epidemiology Program Office, Centers for Diseases Control and Prevention;
2) Division of Healthcare Quality Promotion, National Center for Infectious Diseases, Centers for Disease Control and Prevention; 3) Pennsylvania Department of Health;
4) Philadelphia Department of Health

Background:

Prothrombin time (PT) and international normalized ratio (INR) are laboratory measurements of blood coagulation in patients, including those receiving warfarin, an anticoagulant drug with over 8 million prescriptions in the United States annually. On July 25, 2001, personnel at a Philadelphia hospital recognized falsely-low INR reporting from its laboratory.

Methods:

We observed laboratory processes and compared characteristics of patients tested during the period when errors were reported (June 4-July 25, 2001; N=843) with a pre-error period (April 15-June 3, 2001; N=835). We conducted a cohort study in patients having at least one high INR (>3.0), a level exceeding the usual therapeutic range, to determine possible associations between exposure to the errors and adverse events.

Results:

Insufficient quality control and incorrect programming of a laboratory information system (LIS) generated 2,251 falsely-low INRs in the error period. Patient socio-demographic factors did not significantly differ between periods. However, the proportion of patients with high INR level (10% vs 5%, p<0.005), the mean PT (15.5 sec vs 13.6 sec, p<0.005), and true INRs (2.3 vs 1.6, p<0.005) were significantly higher in the error period. Patients with PT/INR performed during the error period (RR=8.6, 95% CI: 1.3 12.4) and those with peak INR >7.0 (RR=2.0, 95% CI: 1.2 3.6) were more likely to have bleeding complications; three deaths from intracranial hemorrhage were attributed to the laboratory errors.

Conclusions:

Preventable errors due to inadequate laboratory quality control went unrecognized for 7 weeks leading to adverse events and deaths. Laboratories should verify automated LIS calculations and clinicians should suspect laboratory errors when clinical presentations are discordant with PT and INR results.