posted on April 1, 2010
George Towfic, Ph.D., associate professor of computer science at Clarke College, and his colleagues and collaborators are busy developing efficient tools to interrogate medical datasets. Through their current work with complex HIV data, they believe that data can speak – and that it can tell medical professionals how to use knowledge embedded in patients’ medical datasets to analyze patients’ responses to diseases and treatments.
Towfic, along with Judy Munshower, Ph.D., associate professor of mathematics at Clarke, and Samira Towfic, Ph.D., assistant professor of computer science, recently received a second grant from the Grow Iowa Values Fund in the amount of $84,721 to continue their work in developing a set of tools that can be used to improve treatment efficacy by implementing a sound medical ontology to enable knowledge embedded in stored electronic medical records to communicate with each other.
In recent years, the first grant allowed Clarke’s research group to develop prototype mathematical models, expert systems and graphical tools to analyze patients’ reactions to HIV treatments provided by clinicians in the State of Iowa and the State of Wisconsin. The group was also able to design and implement a Web portal that provides relevant queries and analysis results for clinicians and medical researchers.
Throughout the 2007 project, the research group worked with a number of project partners – Stanford College’s medical informatics department, the College of Iowa Hospital and Clinics and the College of Wisconsin-Madison Hospital and Clinics, as well as Harvard Medical School and John Viner, MD, a Dubuque-area leader in the area of HIV and infectious diseases.
In this most-recently funded grant, the research group plans to use their developed software, hardware, and research tools to design and implement a set of Iowa clinical ontology resources, beginning with Dubuque. Dr. John Viner and his colleagues from the Dubuque Internal Medicine will provide valuable medical expertise to design the proposed medical ontology and to validate the resulting medical analyses.
Ultimately, this statewide resource will:
- feature a detailed clinical vocabulary database for all health problems and associated treatments used by different hospitals and clinics in the state.
- establish relations and rules between different patients’ health problems. The relations will categorize and identify patients with common health problems and provide major relations between one health problem and other related health problems.
- establish treatment rules under which a given treatment has provided successful or limited results. These rules will be generated using machine learning algorithms based on the developed clinical vocabulary database to identify patients and environmental conditions that show successful or limited treatment efficacy.
- enable patients and clinicians to obtain statistical analysis related to patient health conditions. Such analysis will take advantage of the electronic medical records, currently available in the Iowa clinics, to provide patients and clinics with comparison charts that show patient reaction and adherences to treatments in comparison to patients with similar health conditions.
Yes, the data can speak. And, a Clarke’s research group and their collaborators are listening, processing and making sense out of it.
For more information, contact the Clarke College Marketing and Communication Office at (563)588-6318.