Databases Move Out Of The Back Room

Public can access drug-test results, bank financial info, and parolees' whereabouts

InformationWeek Staff, Contributor

September 10, 2004

7 Min Read

Clinical-drug-trial results, the financial performance of 8,000 federally insured banks, and the whereabouts of 250,000 criminal offenders on parole or probation. That's about as critical as information gets. Separate central data repositories for each of these areas are being built to consolidate data from many sources and make it available to many different parties.

Most databases today are locked away deep inside the IT systems of businesses, government agencies, and other organizations, safeguarding important information and restricting access to a limited number of people. But these new projects illustrate how database systems are becoming online data clearinghouses for sharing information--even with the general public. The clinical-drug-trial database will provide doctors and patients with the results of clinical tests of drugs on the market, including failed tests and adverse effects. The database of banks' quarterly financial data will help investors judge the financial health of federally insured banks. And law-enforcement agencies and even crime victims can monitor the cases of parolees and criminal offenders serving probation with a national database expected to be operational by year's end.

Last week, Pharmaceutical Research and Manufacturers of America, the industry's Washington trade group, said it's developing an online database with summaries of controlled clinical test results completed since October 2002 for drugs approved for marketing in the United States by PhRMA member companies, including Eli Lilly, Merck, and Pfizer. The database will be operational by Oct. 1.

PhRMA's move is in response to calls from doctors and medical journals for the creation of a mandatory central registry for all clinical trial data. Critics say drugmakers don't disclose enough information about potential problems with their products. Last month, GlaxoSmithKline plc settled a lawsuit brought by New York Attorney General Eliot Spitzer that charged the company with withholding information about possible adverse effects on teenagers of its Paxil antidepressant drug.

The database will include published and unpublished results of late-stage clinical tests, plus other information such as how the tests were designed, the number of participants, drug-label information, and bibliographies of articles published about drugs tested. But the database won't provide raw data from the clinical studies or data on drugs still under development, and it won't include information from some early-stage tests. Participation is voluntary.

Pharmaceutical companies will upload test results in PDF documents using the same format they use to submit data to regulatory agencies such as the Food and Drug Administration, says Dr. Alan Goldhammer, PhRMA's associate VP for regulatory affairs. Doctors and consumers can search test results by a drug's brand and generic names, manufacturer, and the disease it treats. Links to data posted elsewhere also will be provided. "It's really a rather simple project," Goldhammer says. He couldn't say how big the database ultimately will be.

Also disclosed last week was a central database being developed by the Interstate Commission for Adult Offender Supervision to help state and local law enforcement keep tabs on an estimated 250,000 parolees and criminal offenders on probation nationwide who have permission to move between states. The database, which is expected to be operational by the end of the year, and accompanying management tools also will provide victims with more timely, accurate information about their cases.

"Local law enforcement is interested because there's never been a national database of parolees and probationary offenders," says Terry Borjeson, VP of government operations for Softscape Inc. Softscape's CaseOne database- and case-management software will serve as the project's backbone.

The national database will rely on data downloads to the Softscape-hosted database from all 50 state parole and probation offices, plus those in Puerto Rico and the Virgin Islands. These offices will collect data from cities and counties and audit it for duplicate entries or inaccurate data. Ultimately, the national database will replace a painstaking paper process whereby local offices must mail heavy stacks of forms to get permission for an offender to move from one state to another. "Now it will all be done electronically in a matter of seconds," says Don Blackburn, executive director of the commission.

Victims will have access to publicly available information regarding offenders, such as the status of transfer applications, but not to offenders' personal information. In preparation for designing the database, Blackburn says, the commission studied law-enforcement data-sharing models such as the Homeland Security Department's Joint Regional Information Exchange System and the Regional Information Sharing Systems, a federally funded program administered by the Justice Department.

The commission's ultimate goal is to improve public safety, Blackburn says. "The database will let us know where people are and if they're supposed to be there."

Producers and consumers of information on the financial condition of banks stand to benefit from a database developed by the Federal Deposit Insurance Corp. that's due to go live early next year. The central data repository, built and managed by Unisys Corp. under a 10-year, $39 million contract, will house quarterly financial statements known as "call reports" submitted by each of the nation's 8,000 federally insured banks. The reports, which include balance sheets, income statements, and supporting schedules, will be produced and disseminated using the Extensible Business Reporting Language, an XML-based system for tagging elements of structured information in financial statements such as assets and depreciation. (Separately, the Securities and Exchange Commission also is testing XBRL as a format for public companies to use when filing financial statements with the agency.)

The system will allow call reports to be produced faster and more accurately and cheaply. The switch to XBRL entails little technology investment on the part of banks. Most of the business-process changes are being incorporated into reporting software they use from vendors such as Jack Henry & Associates Inc. For most banks, it will be their first experience with XBRL, although some large banks require corporate borrowers to file financial statements in XBRL format. Originally, the database was slated to go live this year, but that start was pushed back to give banks and their software providers more time to prepare.

The data repository exploits XBRL's ability to provide information in a more granular fashion than is possible using HTML, PDF, Word, and Excel formats. That standardization makes it easier for regulators to extract and compare information on different companies.

Use of XBRL will speed publication of call reports, quarterly financial statements submitted by banks, FDIC's Henning says.

By speeding the publication of call reports, XBRL will make them more useful for investment purposes. It now takes as long as 90 days after the close of a quarter before a call report gets published. By comparison, it takes 30 days for most companies, including banks, to publish quarterly earnings reports that may influence investment decisions. "The emergence of XBRL has enabled us to remove the friction in the data-supply lines," says Martin Henning, FDIC project director.

The database makes call-report data more accessible and transparent, says Mark Schnitzer, Morgan Stanley executive director. "We'll be able to cross-index information [from call reports and 10-Qs], and that's going to be important from an analytical perspective." The impact of the FDIC database will be similar to when Edgar, the SEC's financial-statement database, debuted years ago, he says.

Investors will be able to get information faster and more accurately without having to resort to screen-scraping, an error-prone process in which an HTML Web page is downloaded, then the relevant data is extracted and fed into another program. Says Henning, "Even if the data is accurate in HTML, it's difficult to ensure the data stays accurate while it's being screen-scraped."

Never Miss a Beat: Get a snapshot of the issues affecting the IT industry straight to your inbox.

You May Also Like

More Insights