Resolve - Volume 1, 2012 - (Page 15)
inferences from that data, says Nabaa, modeling and optimization tools are wellsuited to streamline the delivery of healthcare. Sophisticated modeling tools already in use, for example, generate nursing schedules that account not only for shift preference, but also for individual tendencies to arrive early or late, for experience levels and for areas of expertise. The technology also exists for patients to electronically authorize a healthcare provider to access their entire medical histories with the swipe of a card, much as a supermarket verifies your credit history in a split second. Systems engineers can also identify and eliminate redundancies, says Nabaa, by enabling healthcare facilities to examine their entire operations in a “holistic” way. “Studies have shown that through the use of systems engineering principles, modeling techniques and analytical thinking, healthcare industries can improve efficiency, cut costs and treat more patients with a smaller staff.”
Getting a grasp on infinity
A company needs to deliver a day’s worth of parcels to their destinations while minimizing the number of miles its drivers log. An oncologist sets out to destroy a tumor with microwave radiation while restricting damage to healthy tissue. The first of these two optimization problems contains a finite number of variables, says Frank Curtis. The second is continuous. How will heat from the microwave radiation disperse across millions of cells? How will the individual temperatures of the various cells influence each other? How will these temperatures be affected by the amplitude and frequency the oncologist chooses? Curtis, assistant professor of industrial and systems engineering, develops algorithms, or mathematical methods, that aim to solve optimization problems with thousands or millions of variables and to do so as quickly and reliably as the existing methods that solve smaller problems. Software programs based on robust algorithms can solve the kind of continuous, large-scale, nonlinear optimization problem facing the radiologist, “ Computers will never be fast enough to solve says Curtis, but they do not continuous problems. We need better algorithms.” have the luxury of time when —Frank Curtis patients’ lives are at stake. Curtis, who has a threeyear, single-investigator grant from NSF, has developed an algorithm capable of solving large-scale continuous optimization problems in less than a quarter of the time required by contemporary methods. In one such problem, the challenge is to optimize the cost-effectiveness of computing server rooms by minimizing the air flow required for cooling. In modeling the bioheat transfer involved in the oncologist’s radiation problem, Curtis represents each point of interest — cell Curtis’s algorithm temperatures in this case — with a partial differential equation. solves large-scale “Our goal is to hit a target temperature for the tumor while keeping the temperature continuous probbelow the threshold for the noncancerous area and also monitoring the heat transfer. lems in a frac“Because the body is a continuous entity with many overlapping regions, we have to tion of the time model for space and time. We do that by turning one large-scale problem into a series of required by other finite-dimensional ones.” methods. Like an image that gains resolution in proportion to its quantity of pixels, says Curtis, a mathematical model achieves greater discreteness, and a more accurate rendition of reality, in proportion to the number of equations it generates. If an algorithm is too complicated, however, it can require more computer time to generate a solution. “Because of the complexity of the real world, you have to make sacrifices,” says Curtis, who tests his algorithms on software and then on applications. “You have to make your model simple enough for the computer to generate an answer.” Curtis’s goal is to develop algorithms that overcome computers’ limitations. “Continuous problems are infinite-dimensional,” says Curtis. “Computers will never be fast enough to solve them. To stay ahead of these problems, we need better algorithms.”
Improving the flow – and utilization – of data
The HSE program’s industry leadership board reads like a roster of Who’s Who in Medical Care. Members represent the region’s largest hospital networks – Lehigh Valley Health Network, St. Luke’s Hospital and Health Network and Easton Hospital – as well as the Mayo Clinic, Geisinger and Susquehanna Health; BAYADA Home Health Care, one of the largest providers of its kind; the Hospital and Healthsystem Association of Pennsylvania; the pharmaceutical risk management company ParagonRx; insurers such as Capital Blue Cross and Highmark; and the global consulting firm Towers Watson. “We are an accurate reflection of the healthcare industry,” ILB member Anne Baum said at an HSE panel discussion last fall. Baum is vice president of the Lehigh Valley region of Capital Blue Cross. “We are an unusual group of individuals with a set of common goals – to improve healthcare while cutting costs and making delivery more efficient.
LEHIGH UNIVERSITY • P.C. ROSSIN COLLEGE OF ENGINEERING AND APPLIED SCIENCE • 15
Table of Contents for the Digital Edition of Resolve - Volume 1, 2012
Resolve - Volume 1, 2012
If you would like to try to load the digital publication without using Flash Player detection, please click here.