Type of research: applied Duration from: 01/01/91. to 12/31/94. Papers on project (total): 34
Papers on project quoted in Current Contents: 2
Institution name: Medicinski fakultet, Zagreb (108) Department/Institute: Department of Statistics, Epidemiology and Informatics Address: Rockefellerova 4 City: 10000 - Zagreb, Croatia
Communication
E-mail: jkern@andrija.snz.hr
Summary: Possibilities of inductive learning methods, realized
through expert systems in the field of infectology, perinatology,
rheumatology and epidemiology of chronic-degenerative diseases, as well as
artificial neural network in laboratory evaluation of thyroid function,
were investigated. Validity of the systems was assessed through prognistic
accuracies, absolute accuracy being 61% through 90%, and relative accuracy,
not exceeding 88%. The best results were achieved in the field of
laboratory diagnosis, and the worst in epidemiological prognosis.
Development and possibilities of decision support systems in health care
systems were considered. Cognitive mapping was used as a method for problem
formulation. Time as a component should be necessary. Some topics relevant
for PHC, hospital system and public health institute were presented. Model
of health ecological geographical information system (HEGIS) was developed.
Its purpose is to ensure the visualization and spacial dispersion, as well
as statistical testing of differences between small areas in a large
region, different kind of epidemiological and ecological data. A standard
spreadsheet program was used for simulation of two problems: infectious
disease (shigelosis) dynamics in population, and dynamics of medical
doctors manpower in Croatia for next 20 years. Aiming to find out the
possibilities of making conclusion and decision on the basis of free text,
some different analysis of different examples has been done. Results of
free text analysis i.e. open ended interwievs, have been compared with
corresponding free text closed in a classic way, in order to find out some
differences, esspecialy in some kind of deficit in information closed in
some categories.
Keywords: Inductive learning, Expert systems, Neural network, Medical decision making, Decision support system in health care, Small area analysis, Geographical data bases, Simulation modelling, Free text analysis
Research goals: Goals of the research: - to investigate the method
of inductive learning in differentmedical and health care fields - to
develop the methodology for analysis of small areasaccording to
epidemiological, ecological, managerial and otheraspects - to simulate
dynamics of some phenomena in medical and healthcare systems - to find
out the possibilities of extracting information from open ended interview
in order to make decision Expected results: Development of models of
decision support systems in the healthcare system
COOPERATION - PROJECTS
Name of project
: CASE STUDY ZAGREB Name of institution: Agencija za posebni otpad City: 10000 - Zagreb, Croatia
COOPERATION - INSTITUTIONS
Name of institution
: Agencija za posebni otpad Type of institution: Institution whose primary activities ar Type of cooperation: Joint project City: 10000 - Zagreb, Croatia Other information about the project.