February 25, 2009

Matt Bailey
Management professor Matt Bailey

Please note: You are viewing an archived Bucknell University news story. It is possible that information found on this page has become outdated or inaccurate, and links and images contained within are not guaranteed to function correctly.

[X] Close this message.

By Julia Ferrante

LEWISBURG, Pa, -- Choosing the right time to begin drug therapy for patients with HIV can be as important as the treatment itself.

Doctors and patients may be inclined to begin drug therapy as soon as possible to fend off full-blown AIDS, noted Matt Bailey, an assistant professor of management at Bucknell University. But the common treatment for HIV – a combination of medicines known as "a cocktail" -- comes with its own set of side effects that diminish patients' quality of life, Bailey said. Waiting too long, on the other hand, can allow the disease to progress too quickly. || Audio clips

"When you're trying to decide when to start a therapy, you're really looking at tradeoffs," said Bailey, who specializes in clinical decision-making in health care. "You have a finite number of options. Once you've exhausted that finite number, you start repeating options with less effectiveness, and the virus eventually runs its course." 

Bailey has dedicated his research to examining the way hospitals are run and how patients are treated. Instead of conducting costly trials involving human subjects, the field of "decision-making under sequential uncertainty" uses mathematical formulas involving probability, statistics and optimization to determine the best courses of therapy and ways to make health care and hospitals more efficient.

Similar conclusions
Bailey and fellow professors Steven Shechter at the University of British Columbia and Andrew Schaefer at the University of Pittsburgh published a paper in the journal Operations Research, January-February 2008 issue, about the optimal time to initiate HIV therapy. Their conclusion that early and aggressive treatment is best was supported by a separate clinical trial involving real-life patients, published around the same time.

In the business of health care, patient studies noting the unanticipated effects of medication and treatment get the most attention. But statistical studies such as Bailey's and his collaborators' are gaining momentum.

"We are taking the tools of analytics and probability and optimization and saying, 'Here are some treatment options that may work better," Bailey said.
Research began in 2004
Bailey, who joined Bucknell last year, began the HIV research project about five years ago while teaching engineering at the University of Pittsburgh. He did his undergraduate work at Purdue University in mathematics and later refined his focus to management science. He received his Ph.D. from the University of Michigan in industrial and operations engineering.

Working with the Pittsburgh Medical Center, Bailey and his research team analyzed data about HIV-positive patients treated at a veterans' hospital, looking for various markers such as age, white blood cell counts over time and duration of treatment.

Weighing the benefits
One of the biggest questions when treating HIV-positive patients is when to start treatment and how aggressive to be, Bailey said. Once a patient starts therapy, the virus mutates and eventually begins to resist the drugs. Getting patients to comply and take all of their prescribed pills also can be difficult, because the treatment has some harsh side effects.

"You have to take into consideration quality of life and ask: Is this (treatment) detrimental to your quality of life?" Bailey said. "It is often said, 'One more day of life is a day of life. A day of life on the cocktail is different.'"

Conversely, if doctors delay treatment too long, the effectiveness of the treatment also diminishes, and the disease starts to win out.

"If you wait too long, you can't go back," Bailey said.

'Hit hard, hit early'
Historically, doctors have taken the approach of hitting HIV early and hitting it hard. There is another camp that says hit hard and hit selectively, Bailey said.

The National Institute of Health recommends treating HIV-positive patients with the cocktail when their white-blood cell count drops below 200. Bailey and his collaborators were looking for benchmarks to see if it would be beneficial to start treatment earlier.

"To do that, we had to go back and get a model of how the disease behaves over time," Bailey said. "It's like trading stocks. You want to determine a price at which you will buy the stock. With HIV treatment, once you get to the (200) count level, do you start therapy or don't you?"

The concept also is similar to that used for machine maintenance, Bailey said.

"With machine maintenance, you ask, when should I take that machine down to repair it?" he said. "If you do repairs too often or you start too early, then you end up spending too much money. … If you don't do it frequently enough, the machine gets to a point where you have poor quality products, the machine can't be repaired, or it's very expensive to repair … The exact same concept applies toward humans."

Contact: Division of Communications