Related Experiment Videos
A prototype Internet autopsy database. 1625 consecutive fetal and neonatal autopsy facesheets spanning 20 years
G W Moore1, J J Berman, R L Hanzlick
1Department of Pathology, Baltimore VA Medical Center, MD 21201-1524, USA.
Archives of Pathology & Laboratory Medicine
|August 1, 1996
Summary
A computer algorithm can generate accurate cause-of-death statements from autopsy data. This autopsy database, containing diagnostic terms, demonstrates a novel resource for medical research and public health.
Area of Science:
- Medical Informatics
- Pathology
- Public Health
Background:
- Autopsy reports contain valuable diagnostic information.
- A centralized, accessible autopsy database is lacking.
- Utilizing digital data for cause-of-death determination is an emerging field.
Purpose of the Study:
- To demonstrate automated cause-of-death statement generation from an autopsy database.
- To test the feasibility of a public, multi-institutional autopsy database.
- To identify challenges in creating and using such a resource.
Main Methods:
- Collected over 49,000 autopsy facesheets from a public internet database.
- Developed a computer algorithm to determine cause of death from diagnostic terms.
- Selected 1,625 fetal and neonatal autopsy records for a prototype study.
Main Results:
- The algorithm successfully determined causes of death in all cases.
- Automated cause-of-death statements required minimal to no modification by a pathologist in over 90% of cases.
- A functional, multi-institutional autopsy database was established online.
Conclusions:
- A large, multi-institutional autopsy database is now available on the Internet.
- This database can be freely accessed by researchers without compromising patient confidentiality.
- Automated generation of cause-of-death statements from autopsy data is feasible and beneficial for public health.