Research Updates: Outcomes Studies

Dr. Gene Blackstone,
Director

The outcomes group of the AFIC grant, lead by Dr. Eugene Blackstone, has focused on several principal projects, such as developing mechanisms for evaluating what constitutes a successful AF ablation, developing techniques for exploring how streams of evolving outcomes can be coupled, in order to identify possible relations between such outcomes, and creating novel ways to represent clinical data to allow for more sophisticated data queries.

Temporal decomposition of longitudinal data: Dr. Blackstone's group has been developing novel algorithms to extract and graphically superimpose time-based changes in individual components of complex longitudinal datasets (patient data, in the present case). The goal of these efforts is to answer questions that previously seemed uncorrelated, such as "what is the temporal relationship between the return of mitral valve regurgitation and the occurrence of AF?", or "what is the correlation between AF burden and growth in the size of the left atrium of patients with heart conditions?". By finding answers to questions such as these, it is anticipated that we will be able to change our treatments to minimize adverse post operative outcomes, and to personalize treatment options. The algorithms have now been developed and tested; further testing and building of a software package that can interface with common statistical programs such as SAS are underway.

Semantics Database: The Semantics Database project is focused on new and innovative ways to represent clinical data. The database associates metadata tags to each element of data stored therein, allowing for powerful querying of data. A typical string search query on "heart attacks" would only find references with those words appearing somewhere in the body of the reference. The Semantic Database, with its metadata tags, has superior searching capabilities - the system, for example, would allow a physician or patient doing research on heart attack treatments to access information on "myocardial infarctions", even though such term was not part of the original search string.

Navigation screen for Semantic DB

The database intelligently can identify terms related to the actual search terms, and provide references that would otherwise be missed with current searching systems. This approach allows the database to grow in unique and powerful ways, and without the need of a database expert or computer programmer. The database and user interface have been developed, and data from nearly 200,000 patients at the Cleveland Clinic have been imported into it. Once fully completed and tested, this database will prove to be a powerful tool for biostatisticians, pharmaceutical companies, medical device companies, clinical research organizations, and other healthcare providers. Likewise, there are many opportunities outside healthcare for this robust system. To expedite development, in 2007, the Cleveland Clinic made a substantial financial commitment to Dr. Blackstone's group.Navigation screen for Semantic DB