[575] Death Certification at an Academic Medical Center: A Re-Evaluation

Cherie Paquette, Steven Shapiro, Nicholas J Hardin. Fletcher Allen Health Care, Burlington, VT; University of Vermont, Burlington, VT; Office of the Chief Medical Examiner, Burlington, VT

Background: Death certificate data are central to epidemiologic research, safety data, and population monitoring; however, certification is poorly taught, often yielding inaccurate or irreproducible certificates. A prior 50 patient retrospective study at this institution found a high death certification error rate. The results prompted a five-year period of procedural and educational changes including new 'death packets' and an online tutorial. Recommendations not implemented included an annual refresher exercise and discussion of certification on rounds; additionally, death certification was removed from the intern orientation. In this study, we evaluate an additional 50 death certificates to assess for progress.
Design: Fifty death certificates from non-pathologists were randomly selected from 2009. One author (CP) abstracted summaries from the medical chart, and two authors (NH and SS) used the summary to create mock certificates while blinded to the original. Discrepancies were resolved by consensus, and errors were identified. The error framework from the prior study was used. (Table 1)

Error Types

0	No error
Ia	Incomplete fields
Ib	Abbreviations or illegible
Ic	Part 1 diagnoses not in logical order
II	Minor missed comorbidities
III	Major missed comorbidities
IVa	No acceptable cause of death in Part 1
IVb	Wrong cause or manner of death

Results: Thirty-six (72%) certificates contained at least one error, compared to 96% in the 2005 study. Type Ia was most common (52%). The percentages of errors of Type III, IVa and IVb were 22%, 28%, and 4%, respectively. There was an increase in Type 0 from 4% in 2005 to 28% in 2009. We identified two forensic errors (failure to notify the medical examiner of a non-natural death).
Conclusions: Death certification education is inconsistent at this medical center, and correspondingly it appears that the error rates have not improved over the past five years. Most worrisome are two unreported forensic deaths. While a literature review suggests that there are short-term improvements in certification after education, the effect is not sustained. We attribute the errors to multiple factors including sporadic use of certification skills, time constraints, and perceived irrelevance of the document. We recommend random or even total review of inpatient certificates by autopsy pathologists as a mechanism for quality improvement, as well as increased consultation of the medical examiners.
Category: Education

Tuesday, March 20, 2012 9:30 AM

Poster Session III # 132, Tuesday Morning