Thursday , August 16 2018
Home / Resources / Articles / Computer Predicts Effects Of Diabetes Treatments

Computer Predicts Effects Of Diabetes Treatments

Apr 26, 2005

For each treatment, the model can predict the rate of survival and death, amputations avoided, quality of life, costs, and cost per quality-adjusted life year. Computer modeling is a powerful new tool for treating diabetes. Researchers and physicians at Kaiser Permanente have developed and tested a simulation program to predict the effects of diabetes treatments.

Called the Global Diabetes Model, the program incorporates actual clinical data about thousands of diabetes patients over many years and data from hundreds of randomized clinical trials of drugs and other treatments. In effect, the GDM integrates most of the clinical science about diabetes and applies it to individual patients — or whole populations.

The model takes into account hundreds of variables, including patient histories, behaviors (smoking, diet, exercise), risk-factor levels (blood sugar, blood pressure, cholesterol), and complex treatment algorithms (aspirin, ACE inhibitors, beta-blockers, etc.). It also integrates data about the costs of treatments and other medical care.

Users are able to predict the effects of treatments on a wide range of diabetes-related complications. For each treatment, the model can predict the rate of survival and death, amputations avoided, quality of life, costs, and cost per quality-adjusted life year.

The model was developed by researchers at Kaiser Permanente’s Center for Health Research and physicians at Kaiser Permanente Northwest. It is an important tool because Type II diabetes is complex, costly and difficult to treat. Diabetes can cause heart disease and stroke, high blood pressure, blindness, kidney disease, nerve deterioration and many other problems. Genetic makeup, personal behavior and personal environment interact in complex ways to affect the disease’s progression.
Because there are now so many different medications and treatments, it has also become difficult to know which will be most worthwhile. In clinical studies, treatments are usually tested one at a time, not in the complex combinations that patients actually receive, and not usually in patients with the array of medical problems many people with diabetes have. Furthermore, most clinical practice guidelines are written for a single condition and often conflict with other disease-specific guidelines, whereas most patients have multiple conditions. No matter how smart and well-educated, no physician could remember and process all this information.

Fortunately, computers can. Running at night before patients arrive the next day, a large computer model can chew and digest hundreds of pieces of information about a patient and predict the 20-year effect of every possible treatment and dosage that might be tried. At this point, GDM has been used only in research.

The model has already been used by commercial scientists to plan clinical trials of new drugs and by Kaiser Permanente to study the relative value of blood pressure reduction, glucose control and lipid management in different kinds of patients.

The current world of document-bound, disease-specific clinical guidelines will become patient-specific, up-to-the-minute and interactive. Patients in health plans with electronic medical records will gain the power to see and rank all their treatment options, via the Web. They will be able to incorporate their own preferences for outcomes and their own insurance coverage for drugs and other treatments. Quality-of-care monitoring will move from crude process measures to comparisons of the health benefits of different types of care. Computer models themselves will increasingly compete against each other and become subject to strict quality monitoring.

Models of this kind represent the next great stage of what is broadly termed, "evidence-based medicine," the effort to provide medical care that really works to patients who really need it. The potential to lengthen life, motivate patients and reduce harmful and wasteful medical care are truly enormous.

Dr. Wiley Chan is director of guidelines and evidence-based medicine for Kaiser Permanente Northwest. Jonathan Brown, Ph.D., is senior investigator at Kaiser Permanente’s Center for Health Research

Diabetes is a Multi-Organ Disease
View an animation for better understanding about the role of incretin hormones in glucose homeostasis.;15837397;11269474;u


Professional CE
Attend the #1 obesity conference for 2005 learn more at