Each year, researchers release an annual report on the status of cancer in the United States, providing a key look into trends related to cancer’s diagnosis and the mortality rate.
The report has long relied on data collected throughout the country, including by the National Cancer Institute’s cancer registry and its many feeder registries, all of which were once using different systems and software to process and submit their data.
In Iowa, for instance, the cancer registry computerized the basics of their records, but kept the richer details of each case on paper. For the staff tasked with ensuring the data’s quality, checking back records meant sifting through archived records stored in dozens of filing cabinets.
That changed with the arrival of a new data management system built by a pair of Washington area contractors at the behest of the Bethesda-based National Cancer Institute. With the faster, more automated system, Iowa’s staffers suddenly had time to go back and convert paper records into digital ones, making these complete records much easier to access.
The project may not sound cutting edge, but it is an example of the largely unsung work done by the Washington area’s sprawling information technology community.
Typically, it is the failure of such projects in the federal sector that draw headlines. Less attention is paid to those that succeed.
At the National Cancer Institute and the regional registries upon which it relies, the new data management system has helped standardize the data reaching the central registry and cut labor costs, all at a price, officials say, that is lower than what the registries might have spent upgrading their own unique systems.
The Iowa registry is one of more than a dozen NCI-funded regional databases — some represent an entire state and some a region or portion of a state — that send their data to the National Cancer Institute and are used by researchers to track trends and changes in cancer incidence and survival. (The latest version of the report, which came out this month, found that overall cancer death rates continue to decline).
Back in the late 1990s, the multiple feeder registries each had their own software and processes for collecting information on tumors.
Each registry’s “job is to collect information and know about every tumor that’s diagnosed in a person” in their region, said Scott Depuy of health technology company Information Management Services, which helped develop the new system. “Every one had their own independent information system ... It was really totally up to them about what software they used.”
Over time, these registries’ hardware and software had become outdated, and NCI officials judged that modernizing each one could be wasteful and likely risky.
The National Cancer Institute turned to a company called Information Management Services, which had been handling some registry work. The two decided to create a single system that could be used by all.
To date, IMS, which has offices in Calverton and Rockville, and the National Cancer Institute have installed the data management system in 11 locations. New Jersey is in the process of switching over now.
The software put together by IMS is hosted on computer hardware managed by another vendor named NetApp, which has its local office in Tysons Corner. Together, their system seeks to provide a straightforward way to manage complex data from numerous sources, said Carol Kosary, program manager for the NCI registry, which is known as the Surveillance, Epidemiology and End Results — or SEER — program.
“If you’re diagnosed with cancer, you’re probably going to see a surgeon, you’re going to have tissues sent in for biopsy ... you’re going to maybe be in the hospital, but maybe you’re going to just see an oncologist,” Kosary said. “To get various pieces of information on a patient who is diagnosed, we’re going to need to get information and records from the pathology department, the surgical department ... oncology, radiology, chemotherapy. Just hundreds of pieces of information that all have to come together.”
The National Cancer Institute’s cancer registry collects about 450,000 newly diagnosed cancer cases annually — or about one-quarter of the cancer cases diagnosed in the United States. (The remaining cases are collected by the Centers for Disease Control’s National Program of Cancer Registries, according to Kosary.)
Each NCI-funded registry operates a little differently, but all must make sense of disparate data they collect from hospitals, pathology departments and other sites.
The collected data may include patient demographics, the site of a tumor and its diagnosis stage, the treatment and the follow-up results.
Kosary estimated that it would have cost about $25 million to $30 million total (in 1999 dollars — when the assessment was done) to modernize the regional registries; she said the NCI has paid less than half that over the course of a decade to centrally upgrade.
IMS, which was founded by two mathematicians in the 1970s, had been working with the registry for decades, starting with a focus on statistics and reporting, according to Depuy, director of computing operations at the company.
Creating a new system — which required remote support for registry locations all over the country — was not an easy task for the small business.
“It put a pretty big challenge on IMS’s shoulders,” Depuy said. “We have to administer and be responsible for these IT systems that are scattered around the country, and we had never done that before.”
IMS brought in NetApp, which specializes in data storage, to help.
“At the core of any institute or agency [is] the data,” said Sam Omar, account manager for health and human services at NetApp’s U.S. public sector business. “Where you put that data and how you protect that data, how you secure that data and how you share that data is really where we come in.”
NetApp sought to offer a computer server-based storage system that would be reliable and easy to change when reporting requirements for the registry changed.
Building the system was the first step. Teaching people to use it around the country was the next. It has taken about a year each — including training and testing— to migrate data from one region’s system to the new version.
David A. Powner,who focuses on information technology for the Government Accountability Office, said it can take a significant amount of time to implement a complex program.
But, Powner — who has not specifically studied the SEER program — said federal IT programs tend to work best when developed incrementally and with user feedback.
When programs are developed and enhanced over time, “you’ve got less that you need to manage and that scope is controlled,” he said.
A registry based in Detroit served as the beta test site, helping IMS and NCI refine the system, said Ann G. Schwartz, the registry’s director.
“That process was difficult for the staff because we would be working iteratively and making changes all the time,” Schwartz said.
Today, she said, the system — which has been improved over the years — allows the registries to be more connected.
In Iowa, it used to take the health registry three to five days to receive and process an “abstract” — or a mini-medical record. Now, it takes about a day.
“We can do more work,” said Kathleen M. McKeen, director of Iowa’s cancer registry, who added that the registry has reduced its clerical staff.
Kosary reported that the new model has helped the National Cancer Institute improve the consistency of its data. “Being able to collect your data correctly and with integrity and high quality is important,” she said.
Better standardized data also saves time. “It decreases the fixing of the data, the cleaning or the tidying up of the data,” Depuy said.