Research Description
Major cardiovascular event (MACE– heart attacks, strokes and cardiovascular death) are the leading cause of disability and reduced life expectancy in older adults with type 2 diabetes (T2D). Patients may have varying response to the same drug class across subpopulations due to multiple coexisting conditions and comedications. Clinical trials have shown two newer classes of T2D therapies, sodium-glucose cotransporter 2 inhibitors (SGLT2i) and glucagon-like peptide-1 receptor agonists (GLP1RA), reduce MACE and SLGT2i particularly reduces hospitalized heart failure (HHF). Few studies directly compare the two or assess their combined use. Data from Medicare (diagnoses, drugs, tests . . .) is very detailed and comprehensive and generally available for people from age 65 until death. Dr. Tiansheng Wang will leverage this enormous database to determine individual patient characteristics associated with MACE and HHF comparing drugs. Using novel artificial intelligence methods, he will develop precision medicine approaches to predict in future care the ideal drug to use in subgroups of patients as well as develop models to make such predictions for individual patients. If successful, this approach promises to deliver the right drug to the right patient at the right time to improve outcomes (MACE and HHF). This innovation will enhance the value of these expensive medications and reduce the burden of cardiovascular diseases in elders with T2D.Research Profile
What area of diabetes research does your project cover? What role will this particular project play in preventing, treating and/or curing diabetes?My project covers precision treatment for type 2 diabetes in older adults. This particular project uses cutting-edge machine learning and epidemiological methods to analyze big data (e.g., US Medicare database) to discover subpopulation that may experience more benefit or adverse events for a given treatment, leading to a tailored treatment for a specific patient and, ultimately, better individual and population outcomes. This project will prevent serious adverse events (by identifying which subpopulation is more likely to experience a risk) and predict optimal treatment to maximize the beneficial treatment effect. If successful, this approach promises to deliver the right drug to the right patient at the right time to improve outcomes. This innovation will enhance the value of these expensive medications and reduce the burden of cardiovascular diseases in elders with type 2 diabetes.
If a person with diabetes were to ask you how your project will help them in the future, how would you respond?Each patient with diabetes has unique clinical characteristics such as age, gender, chronic diseases (e.g. cardiovascular conditions), and may take different medications for these chronic conditions. Patients may have varying response to the same drug class across subpopulations due to multiple coexisting conditions and comedications. Using novel artificial intelligence methods, I will develop precision medicine approaches to predict in future care the ideal drug to use in subgroups of patients as well as develop models to make such predictions for individual patients.
Why important for you, personally, to become involved in diabetes research? What role will this award play?I have developed a strong interest in diabetes since I worked as a retail pharmacist for Rite Aid pharmacy in 2009-2010. This award gives me the opportunity me to apply my skills in epidemiological methods and machine learning into analyze real-world data, which is the area I have great interest and passion with. This award is my first extramural funding, it is the key step for me to conduct research independently.
In what direction do you see the future of diabetes research going?Artificial intelligence (AI) and real-world data are transforming diabetes research by revealing data patterns that can be used to predict treatment outcomes for individual patients. Unlike conventional diabetes research where a hypothesis is evaluated with clinical trials on many patients, future AI-based computational research relies on large-scale clinical data collection to provide the power to assess heterogeneity, epidemiological methods to adjust confounding, machine-learning powered analysis to find patterns and generate predictions about treatment outcomes in subpopulations. Such data-driven methods will help researcher develop tailored, new treatment for individual patients.