Featured News - Current News - Archived News - News Categories
OncoMed: Software is Rx for bad handwriting
By TRACEY DRURY
tdrury@bizjournals.com | 716-541-1609
Deciphering a doctor's scribbled prescription can be a challenge, and that goes not only for the patient but also for pharmacists trying to get it right.
Researchers at the University at Buffalo are working with an oncology pharmacy firm to improve those chances, using handwriting analysis software originally designed for the U.S. Postal Service.
"It becomes kind of a second set of eyes," says Burt Zweigenhaft, CEO of OncoMed - The Oncology Pharmacy, a Long Island-based company with operations on the Buffalo Niagara Medical Campus. "It's amazing the number of deaths from dosing errors that are made in cancer."
OncoMed pharmacists are working with researchers to develop the new systems at the Center of Excellence for Document Analysis and Recognition (CEDAR) at the UB Center of Excellence in Bioinformatics & Life Sciences. The company also provided $150,000 in funding.
Researchers on the project are Venu Govindajaru, associate director of CEDAR and a professor of computer science and engineering, and Srirangaraj (Ranga) Setlur, principal investigator at the center. Though there's been a push in the medical field toward digital or online prescriptions, many physicians prefer the handwritten form. The goal with this project is to allow them to continue to practice in the way that works best for them. It also works in tandem with the pharmacists.
"This does not take the human out of the loop," Setlur says. "The angle we're introducing is the safety aspect. It brings the element of a second check."
The project includes first building a database of oncology drugs, along with the likely range of doses based on patient age, weight and possible negative interactions with other drugs. The "intelligent" system is also designed to build a knowledge base about each physician as prescriptions are scanned into the system, so it will learn to recognize their handwriting and the fax number from which the prescription comes.
According to OncoMed, a large percentage of oncology prescriptions come in the form of a fax to the pharmacy. And even for trained pharmacists, being able to read them can be difficult. The biggest appeal to this type of program, Zweigenhaft says, is it allows doctors to use their time most efficiently.
"Doctors in oncology don't write a single prescription as most doctors do," he says. "They're using combinations of drugs and multiple days in a cycle, and not always the same drugs in the same days of a cycle. It's a lot more complex than writing for a cholesterol drug or a diabetic."
OncoMed is also working with OnCallData in Gaithersburg, Md., to develop electronic prescriptions and refill systems. Zweigenhaft says both efforts will help create efficiencies and also drive safety.
The efficiency side is vital, he says, especially as the population of individuals with cancer continues to grow, along with the demand for oncologists.
"We hope to cut the manual time in entering all the data by 90 percent, which means our clinicians would have a lot more time to focus on intervention and care management," he says. "Anything we can do to make the oncologist more efficient opens up more avenues to care."
Researchers say the system will leverage expertise in handwriting technology, including capitalizing on other projects CEDAR has completed in health care. They include a system to read surveys collected from people participating in flu vaccine drives, allowing handwritten data to be converted and analyzed and provide immediate, real-time information to the state Department of Health.
Another project, funded by a $500,000 contract with the National Science Foundation, involved building a system to read reports written in the field by emergency-response professionals from 911 calls. The reports are scanned into the system, allowing health officials to search how many people made emergency calls in a particular month for acute asthma or allergies, or whether individuals taking a particular medication are experiencing symptoms.
"All the data is read in real time, so anyone can ask these questions," Govindajaru says. "All of these advances are enabled by expertise in handwriting technology."


