Nearly three years in the making, the Total Cancer Care Survivorship Initiative at Tampa-based Moffitt went live in late August with a patient Web portal, starting with colon cancer survivors. The goal is to follow survivors for the rest of their lives to reduce recurrence of cancer through personalized follow-up care and lifestyle changes, and to catch any recurrences early. The program also will provide patients with a detailed summary of their care to educate them and other providers about the treatment history and any risk-related effects.
Its really an opportunity to let patients, once their treatment is complete, get on with their lives and make lifestyle changes they want to make, says Paul Jacobsen, Ph.D., chair of the Department of Health Outcomes and Behavior at Moffitt. But the treatment summary also will support continuity of care, Jacobsen adds. For instance, if cancer returns, many patients dont remember much about their initial treatment except that they had chemotherapy.
An application in a data warehouse behind the portal will enable the building of care plans for each survivor with recommendations for what tests should be taken and at what intervals. The recommendations initially will be based on general evidence-based guidelines, but will become more personalized as researchers find new best practices.
The generic recommendations are one-size-fits-all, such as getting a colonoscopy every six months, Jacobsen says. The hope is to over time make the recommendations pertinent to each patient by including such factors as pain, sleep and fatigue problems, and refer patients to specialists or trusted reference information on the Web. Using data to better identify personalized traits also can show if family members should have genetic counseling to identify their cancer risks.
Proponents hope to have some personalized treatment guidelines and provide help finding additional information starting in the summer of 2010, depending on available funding, says Lance Caranante, director of business and finance systems at Moffitt. We want to present them data that theyre likely to go looking for anyway on the Web, he notes. Were focusing on trusted sites and valuable information, such as age-specific treatment for certain drugs, that they might not find. Were going to give them the data we hope will be relevant to them.
The program also will include a physician Web portal to share best practices and other information with patients primary care physicians. It will take two or three years to compile adequate data to support valid research.
The forthcoming personalized treatment based on genetic and lifestyle considerations, combined with a portal to bring information gleaned from the data warehouse down to cancer survivors and their physicians, is revolutionary, Caranante contends. Were on the forefront of developing a personalized portal to a level that doesnt exist today. This first piece is very important because it gives the means to support the portal. Nearly 40,000 patients have consented to having their data stored in the warehouse. The possibilities are endless.
How it Works
The Total Cancer Care program started with a data warehouse holding information on 38,500 patients, which has taken nearly three years to collect. The warehouse holds data from Moffitts electronic health records system and a cancer registry. The registry, with data from Moffitt and patients other providers, will aid in following patients throughout their lifetime. The cancer center annually contacts patients and their providers to get updated information via telephone, e-mail, postal mail, fax or electronic health records.
The data warehouse also will support continuing research at Moffitt and 17 other treatment facilities across the nation to identify the most appropriate treatments based on best practices, patient genetic characteristics and other factors. We try to find the underpinnings of the molecular mechanism of cancer, and reveal genetic signatures predictive of diagnoses, therapeutic interventions and prognosis of the disease by bringing together molecular and clinical data, says David Fenstermacher, Ph.D., a biomedical and genomics researcher, and chair and executive director at Moffitts Department of Biomedical Informatics.
Moffitts data warehouse sits on a platform from Oracle. Its data marts are managed with Microsoft SQL Server software.
Microsofts Amalga Life Sciences data integration and aggregation software combines data from multiple sources and organizes it, and the vendors ProClarity software facilitates the building, analysis and reporting of data sets pulled from the warehouse. In essence, Amalga and ProClarity provide some of the tools to analyze data, Fenstermacher says. Moffitt is installing the latest version of Amalga, which includes a dynamic view builder that will enable researchers to quickly select what data elements they want, such as smoking history and other risk factors, rather than wade through thousands of different elements, Fenstermacher notes. Well be able to design on-demand data for researchers and tell them the strengths and limitations of the data.
Hes also excited about Microsofts recent acquisition of certain genomic data analysis technologies of the Rosetta Biosoftware unit of Rosetta Inpharmatics LLC. Microsoft will embed the technologies in the Amalga software. Well take it and look for genes turned on and off to help us understand the molecular mechanisms of cancer,
Fenstermacher says. Its going to provide a rich set of data from diagnosis through treatment and survival.
While the program didnt start for patients until late August, the data warehouse and analytics software has already provided significant benefits for Moffitt researchers, who previously had to look in 25 different information systems for data. Many of the legacy systems have been turned off and the warehouse gets data feeds from 14 active systems. The time savings are huge, Fenstermacher says. We can get the data in a minute rather than doing chart abstractions which can take weeks or months.
He notes, however, that today only 40 to 50 percent of the data he needs comes from the warehouse, but expects big gains during the next three years as about 100 patients daily consent to their data being put in the warehouse. Well never get to the point where all data is in the data warehouse. If I could get 80 percent, thats when Id say success.
The warehouse also is starting to yield the rich data needed to build algorithms to assess best treatment practices, Fenstermacher says. For instance, two clinical trials are using molecular markers to determine appropriate treatments. One of the trials, covering ovarian cancer, is using cell lines from tumors to unveil the molecular changes in cells based on chemotherapeutic treatments. The goal is to learn which types of cell lines individual cells captured from a tumor that grow and divide in Petri dishes until there are enough for research respond best to which types of drugs. If the cells die, we can look at the molecular pathways that enable cell death and then translate that knowledge to ongoing clinical trials, Fenstermacher says.
Even in the early stages, the amount of data being collected and analyzed is huge. Many data sets we work with on a regular basis are many terabytes in size, Fenstermacher explains. Its not unusual for us to work on a project involving five to seven terabytes and its only getting worse. The warehouse has the infrastructure to support three or four petabytes of data. A terabyte is one trillion bytes of data and a petabyte is one quadrillion bytes.
An important benefit of the warehouse and cancer registry will be significant shortening of the process of identifying and recruiting patients for clinical trials based on their molecular traits. Were looking for patients with specific cancers who have failed conventional treatments, Fenstermacher says. Youll take a search for eligible patients from months to years literally down to weeks. He believes Moffitt and its cancer treatment partners are two to three years away from having a database large enough to do such searches routinely.
Advancing data warehouse and analysis technologies are dramatically increasing the speed by which researchers can identify the most effective cancer treatments based on the type of cancer, the patients genetic makeup and lifestyle, and other factors. The missing piece, however, is a set of tools to bring this information out of the research settings and down to physician offices and patient homes.
Fenstermacher hopes this capability will be available in two or three years, when Moffitt plans to use the physician Web portal to enable physicians to access what has been learned. We eventually have to push this data down to the physicians and patient, he says. The best thing we can do is empower the patients to be an active part of treatment by providing tools to help them understand their disease, their options and to make decisions with their physician.