One of the most complex, multifaceted issues facing American society today is the rising cost of health care. The amount of data available in the health policy and economics sector is unprecedented. In the Medicare system alone, over six billion transactions have been made since the program’s inception. It’s a volume of information no human could wrap their brain around. Luckily, the 21st century technology that makes this huge wealth of data available about patients’ care and clinical status can be used to parse that data.
Carl Kesselman, a professor with the USC Viterbi School of Engineering, specializes in the creation of cyber infrastructure to render vast datasets usable. A director of medical informatics at USC Viterbi’s renowned Information Sciences Institute, Kesselman is a leading authority on the development of computational grid systems, notably with the Globus Alliance. Kesselman has teamed up with Professor Dana Goldman — an expert in health policy and economics with the USC Price School of Public Policy — to apply innovative analytic technology to the health policy and economics sector through a project partially funded by the National Institutes of Health.
As Kesselman puts it, the project is rooted in innovating new means for the application of “analysis and data driven science from these communities . . . part of it is dealing with the data, and part of it is ultimately about transforming the way that questions are asked and answered.”
Indeed, this is the first time such massive datasets have been concurrently analyzed in this field. And though data’s mere existence does not necessarily make for knowledge or insight in and of itself, there is great potential for discovery.
Medicare is an entitlement program, and in our society, that word is loaded with a whole host of judgments. Thus far, the national discussion surrounding health care has been largely centered around debating how and how many dollars should or should not be spent. The work being undertaken by Kesselman and Goldman to apply grid technology to the six-billion-record-strong Medicare dataset has the potential to elevate this discussion.
Central to this shift is the crucial differentiation between cost and value. For instance, if a patient with an injured knee seeks care, providers may respond by recommending two aspirin and rest, or they may go so far as to take an MRI. The aspirin costs pennies and the MRI thousands of dollars. If the patient’s knee is simply sprained, the vastly lower cost aspirin actually offers more value. However, if something more serious is wrong, such as a torn ACL, the MRI becomes more valuable. Value is a reflection of the benefits received from something, and cost reflects the resources required to produce it. The key is improving our ability to recognize that distinction. As Goldman puts it, “Medicare is a program to improve the quality of life for a large segment of the American population and that has a lot of value. The issue is just how do we make it more efficient? And that’s what we’re trying to do with our work.”
The move toward a macro view of the data has already begun, but thus far it has only been possible to analyze one dataset at a time, such as Medicare, Medicaid, or commercial claims. With grid technology, all of these can be examined simultaneously to explore much more difficult questions because larger datasets provide better statistics. As Goldman and Kesselman continue to develop tools that allow for the extraction of practical insights from a meta-analytic macro view, their work will open additional avenues to optimizing the efficiency of the health care system. Ultimately, they are pursuing a systematic application of this innovative approach to how care is provided.
A key issue arises regarding privacy and who owns the medical information that researchers collect about people. Insurance companies could use the information to charge people more, or employers could use it to screen out potential job applicants. Moving forward, it will be crucial for researchers to be mindful of differing motives in the sector, how data is used and that appropriate permissions are acquired for the work being done. Failure to do so could, of course, hinder research in the long term.
Prior to their collaboration with USC Viterbi, Goldman’s team had working methods in place, “but they were time consuming both in terms of CPU time and the amount of effort people had to put into doing these analyses.” Working with Kesselman has allowed “better tools . . . and better algorithms . . . to scale up [the] research enterprise, be able to ask much broader questions, and be able to answer them. The fact of the matter is that there are very few places in this country where [one] can get health policy people, working with economists, working with engineers to help solve these problems.”
This speaks to the uniquely open interdisciplinary environment that exists at USC. In this case, a world class health policy and health economics group that is open to working with the Viterbi School, which spurs the advancement of both disciplines. Cooperative application of resources is absolutely crucial to devising workable solutions to the complex problems that society must confront in the 21st century.
Published on April 1st, 2013
Last updated on August 5th, 2021