We propose to design, implement and test a computer-based workstation to assist clinical geneticists in their tasks of consulting with patients concerning genetically-based diseases. The proposed tool will help the counselor to collect relevant data about the consultant and his family, to interpret those data according to accepted principles of probabilistic analysis, and, using a data base of relevant information from the genetics literature, to focus counseling effort toward helping the patient make difficult and value-laden decisions, to help produce accurate and informative reports and records, and to help organize and aggregate the wealth of pedigree data collected during counseling sessions for input into later research studies. Having already developed and begun to test a prototype program that captures family pedigrees via a graphical interface and calculates genotype probabilities for single-locus Mendelian disorders, we propose to extend it to deal with the wide range of genetic issues that arise in clinical practice. Some of the required extensions include being able to handle more than two alleles at a site, incorporate evidence other than phenotype and family structure into the probabilistic calculations, allow the analysis of multi-locus linkage, especially to exploit the growing availability of RFLP data, to allow the use of different a priori gene frequencies for different individuals based on their ethnic background, and to incorporate age-dependent penetrance models. We also plan to allow the program to help the counselor determine the likelihood of various inheritance patterns in a specific pedigree. In addition, we plan to extend the input and output capabilities of the program to produce appropriate on-screen and printed reports, and to accept information in flexible ways to encourage the capture of already- available genetic data. We will test the usability of the program at the Genetics Department of the Harvard Community Health Plan, and later assess its utility for helping genetic counselors perform consultations. We will also develop a limited version of the program that contains only a genetic model for familial breast cancer and perform early tests of this version in collaboration with Memorial Sloan Kettering Cancer Center.

Agency
National Institute of Health (NIH)
Institute
National Library of Medicine (NLM)
Type
Research Project (R01)
Project #
5R01LM005296-02
Application #
2237745
Study Section
Biomedical Library and Informatics Review Committee (BLR)
Project Start
1993-09-30
Project End
1996-09-29
Budget Start
1994-09-30
Budget End
1995-09-29
Support Year
2
Fiscal Year
1994
Total Cost
Indirect Cost
Name
Massachusetts Institute of Technology
Department
Type
Organized Research Units
DUNS #
City
Cambridge
State
MA
Country
United States
Zip Code
02139
Szolovits, P (1995) Uncertainty and decisions in medical informatics. Methods Inf Med 34:111-21
Szolovits, P; Kohane, I (1994) Against simple universal health-care identifiers. J Am Med Inform Assoc 1:316-9