The Center for Biomedical Informatics
State University of Campinas, Brazil
Research Abstracts
AN USER-PROGRAMMABLE TOOL FOR MEDICAL DECISION MAKING WITH
MICROCOMPUTERS
R.M.E. Sabbatini
Center for Biomedical Informatics, State University of
Campinas, Brazil.
The interest in the use of computers to support the
decision-making
process in clinical medicine has increased sharply in the last years,
mainly due to the availability of so-called medical expert systems. Its
practical applicability, however, has beenvery limited so far, because
they usually demand heavy computer power, are too large and complex,
and cannot be altered or programmed easily by the end user. With the
aim to produce an easy-to-use, field-programmable, small- scale system
for aiding healthcare workers, we have developed MEDTREE. This is a
stand - alone program, which runs in any 16-bit, IBM PC-compatible
micro computers, with a minimum of 256 Kbytes of RAM and a single
diskette drive, with the following main characteristics: a) it
implements a data-directed binary decision tree or table (yes/no
answers), traversed depth-first according to the answers given by the
user to simple questions (presence/absence of symptoms, signs,
indicators, conditions, etc.). b) it has provision for defining
numerical utility and probability functions at each level of decision,
in order to support weighing/decision- theoretical analysis; c) it has
facilities for explanation, summarization of running evidence,
comments, listing of decision targets and indicators, storage of
results, etc.; d) it has powerful, easy-to-use facilities for defining,
editing and listing decision- making procedures in any field, which are
stored as separate files onto disk; e) it has an advanced, highly
ergonomic user interface, with horizontal menus, context-sensitive help
windows, etc. All texts and messages within the program reside in an
auxiliary file and can be translated easily to any idiom, using a
common word processing program. The simplicity and effectiveness of
MEDTREE make it an ideal tool for health care health care applications
in developing countries, since the end users are able to define and
enter into the program their own decision-aid procedures, in an easy
way. Furthermore, MEDTREE is flexible enough to support several
algorithms of decision. Many medical decision procedures (diagnosis,
risk evaluation, therapeutic protocols, interpretation of laboratory
tests, etc.) have been implemented with MEDTREE, and are being tested
in realistic environments in Brazil.
Published in:
Conferencia Internacional de Informatica Medica, Congreso Internacional
de Informatica, La Habana, Cuba, Febr. 1988
Last Updated: March 2, 1996
renato@sabbatini.com
