‘Big Data’ for Cancer Care; Vast Storehouse of Patient Records Will Let Doctors Search for Effective Treatment; Some 1.6 million Americans are diagnosed with cancer every year, but in more than 95% of cases, details of their treatments are “locked up in medical records and file drawers or in electronic systems not connected to each other,”

Updated March 26, 2013, 8:46 p.m. ET

‘Big Data’ for Cancer Care

Vast Storehouse of Patient Records Will Let Doctors Search for Effective Treatment

By RON WINSLOW

A major oncology group is launching an ambitious project to collect data on the care of hundreds of thousands of cancer patients and use it to help guide treatment of other patients across the health-care system.

Cancer doctors would be able to consult the database, much like doing a GoogleGOOG +0.34% search. They would get advice on treatment strategies that might work for their patients based on how similar patients fared in practices around the U.S.

Details of the plan are being unveiled Wednesday by the American Society of Clinical Oncology, a nonprofit professional association.

It is the latest example of emerging efforts in medicine to harness the power of “Big Data” to improve care. It reflects growing expectations that information gleaned from huge clinical databases can speed learning about benefits and harms of treatments, support efforts to improve quality of care and hasten development of new medicines.

Some 1.6 million Americans are diagnosed with cancer every year, but in more than 95% of cases, details of their treatments are “locked up in medical records and file drawers or in electronic systems not connected to each other,” said Allen Lichter, chief executive office of ASCO. “There is a treasure trove of information inside those cases if we simply bring them together.”

Gathering and analyzing huge amounts of medical data is fraught with challenges. ASCO says it has overcome one of the big ones by developing software that can accept clinical information from almost any electronic health record—a major hurdle in the long-running effort to computerize medical records. A prototype system has collected about 100,000 breast-cancer records so far from 27 oncology groups, many of which use different types of electronic records.

But researchers still are working on ways to turn the data into useful reports that oncologists can act on. Issues of patient privacy and the limitations of such data—which don’t involve randomized trials and so can introduce bias—are other hurdles.

It likely will take 12 to 18 months before the first components of the project are rolled out for use by doctors, ASCO said.

ASCO is joining a Big Data movement that is well under way across medicine, including other initiatives in cancer and in cardiology. The Institute of Medicine, an independent body that advises the U.S. government on medical issues, believes such databases eventually will become a “health-care utility” to generate knowledge for treating a variety of diseases.

The ASCO project is “recognition that big data is an imperative for the future of medicine,” said Lynn Etheredge, a consultant with the Rapid Learning Project at George Washington University in Washington, D.C.

Currently, less than 5% of adult cancer patients participate in randomized trials, the gold standard for treatment. Such studies enroll a carefully defined group of patients and often don’t reflect the patients oncologists see in day-to-day practice.

In the ASCO project, called CancerLinQ, almost all patients would, in effect, become part of a clinical study. The system would collect data that doctors routinely record in a patient’s files, such as age, gender, medications and other illnesses, along with the patient’s diagnosis, treatment and, eventually, date of death. Once the outcome of a sufficient number of patients is pooled, doctors could tap the database for help in developing treatments for other patients.

For instance, consider a 77-year-old man with stage 3 colon cancer, heart failure and diabetes, said W. Charles Penley, a cancer doctor at Tennessee Oncology in Nashville. “That patient wouldn’t have been included in a clinical trial, but those are the folks we take care of in the real world all the time,” he said.

Using the database, Dr. Penley could see how the top three chemotherapy regimens for similar patients performed, and how age, heart failure and diabetes might have affected the treatment.

“If you can query a database in real time for a patient in your office, this is a potential game changer,” Dr. Penley said.

The database also would “give us more evidence for the treatments we actually use,” said Sandra Swain, ASCO president and a breast-cancer specialist at Medstar Washington Hospital Center in Washington, D.C.

For instance, a common treatment for patients with so-called HER2 positive breast cancer but very small tumors is chemotherapy plus the drug Herceptin. But it isn’t clear from clinical studies that the therapy helps in that stage of the disease. A randomized trial would be too expensive and time-consuming, but data in CancerLinQ might lead to an answer, she said.

The database also could help with very rare cancers, such as those an oncologist might see only a few times during a career, ASCO said.

As envisioned, only doctors and researchers would have access to the database. A portal is planned to allow patients to contribute information on their response to treatment, but not surf the database.

ASCO said it is taking into account national and state laws on patient privacy. In the breast-cancer prototype, patient information was sent to a third party that stripped medical records of individual identifiers before it was entered into the database.

Lee Newcomer, senior vice president, oncology, at medical-benefits companyUnitedHealth Group Inc., UNH +1.44% who isn’t involved in the project, said it has “great potential” to enable “very-real time monitoring and assessing of what’s going on with therapies in oncology.” It could be valuable, he said, in identifying and potentially reducing variation in treatment and in research on cost-effectiveness of cancer care.

But he said data used to guide specific treatment decisions needed to be used “cautiously” because such databases can yield misleading information.

ASCO’s Dr. Lichter said the sheer volume of patients should overcome some weaknesses of the data, and that statisticians are working on methods to strengthen the validity of the findings.