Home  |   Subscribe  |   Resources  |   Reprints  |   Writers' Guidelines

August 15, 2011

EHRs and Personalized Treatment Programs
By Maura Keller
For The Record
Vol. 23 No. 15 P. 8

There’s much that can be done to help keep patients healthy and lower their risk for many common yet serious illnesses. A healthful diet, steady exercise, regular health screenings, and avoiding unhealthful habits such as smoking all can help patients steer clear of cancer, heart disease, and other conditions.

While many healthcare providers use general guidelines to determine a patient’s likelihood of having a heart attack or a stroke, others are coming around to the idea of analyzing EHR data to construct specific individualized treatment guidelines.

Personalization at Its Core
In terms of the practical aspects of individualized treatment guidelines, Archimedes, Inc, a healthcare modeling group in San Francisco, has developed an application called IndiGO (Individualized Guidelines and Outcomes) that can be inserted into clinical information systems. It can also be adapted to programs that have data warehouses or the ability to merge legacy information systems.

“In such systems, the deidentified information on each person can be automatically downloaded for calculations, requiring no extra work by physicians, patients, or other personnel,” says Peter Alperin, MD, vice president of medicine at Archimedes.

Results are accessed through an EHR or another computer-based application. Because individualized guidelines take into account a broad spectrum of information, including demographics, physical findings, lab values, and past and current medications, they are best used in systems that have some form of electronic records.

“Individualized guidelines would be more difficult to use in old-fashioned paper-based systems that would require manually inputting the data,” Alperin says.

IndiGO’s cardiovascular (CV) risk tool was built using data from dozens of publicly available datasets and clinical trials. Archimedes scientists analyzed the data to create mathematical equations that predict a patient’s CV risk.

“The equations use information that is readily available in electronic health record systems—eg, age, gender, blood pressure, laboratory data, medication use, family history, etc—to accurately determine a patient’s CV risk,” Alperin says. “Furthermore, IndiGO assesses the potential benefit that a patient would see if he or she were to take steps to improve their health, such as quitting smoking, losing weight, or taking medications to lower cholesterol or blood pressure. The tool was validated by testing the predictions of the tool against data that wasn’t used to build it.”

Patricia B. Wise, RN, MS, MA, FHIMSS, vice president of healthcare information systems for HIMSS, witnessed a similar concept used at her physician’s office. ”My physician was an early adopter of EHRs in 1996 and had written his own program that he utilizes with patients to discuss health risks. The program utilizes your height, weight, smoking history, activity levels, and blood levels of cholesterol and triglycerides and calculates your chance of having an MI [myocardial infarction] in the next five years,” she says.

Using an iPad, the physician enters data while in the exam room with the patient. ”There is no doubt it is your data,” Wise says. ”You then might learn that you have a 35% chance of an MI in the next five years. As you watch, he changes the data telling you, ‘Let’s say you quit smoking. You do nothing different except quit smoking. Where does that leave us?’ He then indicates that your smoking has stopped and as you watch, the 35% chance of an MI becomes 20%. He then suggests you start exercising only 15 minutes a day. The 20% chance turns to 12%. And so on. It is a very powerful tool in patient education and attempting to get patients to change their lifestyle.”

According to Alperin, current guidelines tend to focus on particular risk factors and use sharp thresholds to determine who should be recommended for treatment. For example, if a patient’s systolic blood pressure is greater than 140 mm Hg, then treatment is suggested.

When estimating the likelihood of a patient developing an MI or a stroke, IndiGO evaluates various data elements, including demographics (eg, age, sex), laboratory data (eg, cholesterol and glucose levels), biomarker information (eg, blood pressure, weight), history, medications, and behaviors (eg, smoking status).

“The data is run through the calculator which then returns the specific risk of MI or stroke for that individual,” Alperin says. “It is the combination of the input data and algorithms in the calculator that determines the risk of the events.”

The tool goes one step further by calculating the potential benefit of different interventions, including medications and behavioral changes, to reduce that risk.

“This is important because this provides the information that a patient needs to take action to reduce his or her risk of heart attack or stroke,” Alperin says.

An Archimedes white paper showed that compared with not using guidelines, individualized guidelines can deliver the same benefit as the national hypertension guideline (called JNC-7) in terms of reducing heart attacks and strokes at a 67% reduction in cost. When spending the same amount of money as the JNC-7 hypertension guidelines, individualized guidelines could achieve a 43% greater benefit without an increase in cost.

“Individualized guidelines can be delivered through clinical information systems or other Web-based applications,” Alperin says. “They provide patients and physicians with quantitative information about each patient’s risks of important events such as heart attacks, strokes, diabetes, and its complications.”

They also provide patients and physicians with information about how each of those risks can be reduced through numerous treatments, one by one and in all possible combinations. “This enables physicians and patients to move from simply saying, ‘You should stop smoking’ or ‘You should continue to take your hypertension medication’ to being able to say things like this hypothetical example: ‘If you stop smoking, your risk of a heart attack or stroke in the next five years will go down from 20% to 10%.’”

Alperin explains that evaluations of individualized guidelines in clinical settings have shown that this type of quantitative information is more apt to result in behavior change than the qualitative statements delivered by current guidelines. “The information can be used by patients on their own, as in a health risk appraisal application; by physicians and patients together to make shared decisions about appropriate treatments; by care managers to identify the highest priority patients and treatments for outreach programs; and by physicians and health plans to optimize the management of a group of patients,” he says.

Are individualized guidelines going to become commonplace?

“If the program such as the one my physician developed is a personalized model, I would say they are very compelling, and physicians interested in changing patient lifestyles to improve the quality of care could have a powerful tool,” Wise says. “Utilizing this model could add time to the visit, but the outcomes for this quality time spent have tremendous potential. I would hope that physicians, through information technology tools, are able to develop and utilize personalized treatment plans.”

— Maura Keller is a Minneapolis-based writer and editor.