The CDAT initiative is designing and developing a toolkit that will help clinicians, healthcare administrators and public health professionals to compare the effectiveness of different clinical practices, including preventive and treatment modalities.
The Cohort Development and Analysis Toolkit (CDAT) will allow healthcare professionals to better leverage the large amounts of healthcare data available in electronic health records, facilitating cohort development and analysis along the entire healthcare spectrum: healthcare providers, patients, systems and processes, and case management/discharge planning. By providing intuitive interfaces and workflows that are designed to support a medical practitioner rather than an IT or data expert, CDAT will make data mining and analytics more accessible to the medical community without the need for statisticians or data mining professionals.
The CDAT technology will provide automated cohort development using medical (MO) and clinical analytics (CAO) ontologies and can be mapped to different legacy healthcare systems, assisting clinicians, nurses and QI personnel in creating and analyzing cohorts based on medical datasets. The tool also provides intelligent search and retrieval, scouring the internet and CDAT documents to retrieve and summarize documents related to the focus cohort and data mining analyses. CDAT will glean study goals using the study’s attributes and comparative cohorts and will retrieve data and documents based on the same or similar study goals. The JackalFish® technology, KBSI’s natural language processing techniques and information retrieval tool, will be used to provide this capability.
The CDAT solution has the potential to significantly improve the level of healthcare service effectiveness while simultaneously lowering costs. In an era of tighter budgets and the dynamic evolution of service needs, healthcare providers must more efficiently leverage their resources. This includes understanding the effectiveness of different medical practices, demand patterns, resource and facility utilizations, customer preference patterns, insurance management, etc. CDAT’s data mining and cohort based analysis methods offer the potential for deciphering the service demand and consumer patterns that affect strategic long term and pragmatic short term planning.
CDAT will also help healthcare providers improve care via medical knowledge discovery. Knowledge, which is available in patient records but has not traditionally been easily mined, is a valuable resource for applying empirical knowledge that is hidden in clinical data systems. CDAT’s knowledge discovery and data mining methods offer benefits in (i) identifying significant clinical variables with respect to a diagnosis or therapy plan, (ii) modeling subjective and other contextual elements involved in medical decision making, (iii) modeling decision making rules, and (iv) revising and refining knowledge so that existing medical knowledge can help refine empirical knowledge elements hidden in medical records.