 |
 |
|
Streamlined social services
Governments are supposed to exist for the benefit of their constituents, particularly those most in need. Yet when it comes to social services, civil servants striving for this noble goal often run into an obstacle course of outdated and incompatible databases, repetitive paperwork, and frustrating inconvenience and inefficiency. That's where Dr. Athman Bouguettaya and Dr. Ahmed Elmagarmid come in. Backed by a three-year, $500,000 Digital Government grant, the researchers are investigating methods of consolidating the enormous variety of social service queries into a single, easy-to-use interface, without damaging or replacing the underlying databases. Their tools include metadata-based database interactions and new web service technologies.
The project, "Database Middleware for Distributed Ontologies in State and Federal Family and Social Services," employs a variety of strategies to pursue the simple goal of making life easier for citizens and government employees alike. One of the researchers' most significant innovations is a new XML-based language, written especially for this project, that can customize client-specific social service itineraries from bits and pieces of other programs. The researchers' work is based on the premise that today's social service systems cause delays for case workers and clients alike by requiring them to interact with multiple databases that are heterogeneous and distributed - that is, dissimilar and located in different places. "The current system is highly manual," says Bouguettaya, co-PI for the project and the program director of computer science at Virginia Tech. "You can only do one thing at a time. We don't have that one-stop shop where you can do everything." With the help of Indiana's Family and Social Services Administration (FSSA), the Indiana Department of Workforce Development and the U.S. Department of Health and Human Services, the researchers have spent the last two years investigating innovative solutions to this problem. What might seem the most obvious answer - standardizing the data in a single database or a few fully compatible databases - was off-limits from the start. Cost, effort and the sheer volume of data, which stretches into terabytes, made a wholesale transfer of all relevant social-service data from one type of database to another "almost unthinkable to entertain," Bouguettaya says. Privacy issues would also arise if highly sensitive information, such as Medicaid and welfare data, appeared in the same database as less sensitive, more easily accessible information. Moreover, centralization of data would create a system more "open to abuse," Bouguettaya says. Citing the example of a national registry of salaries, he says, "You wouldn't want to have it under one single roof with the other (databases) ... The moment you have everything centralized, you have one single authority (that) can do whatever they want to with the data." All of which leaves the key question still unanswered: How do you move the data from these legacy systems over to a new model without "breaking the system itself"? The answer, Bouguettaya says, is to "show people... that databases can talk to each other even though they're different." The researchers' latest result, an online demo called "WebDG," goes a long way toward achieving that goal. Although it contains just eight of the dozens of databases that would eventually be included and its mock data is sufficiently limited that following the guided tour is essential, the demo gives users a good taste of what an integrated social-services query interface might look like. The mock databases talk to each other in four ways, through:
CSSL would give case workers the ability to, in essence, create new social-service guides - tailored to meet the needs of each client - by combining individual parts of existing programs, according to Hao Long, a graduate student on Dr. Bouguettaya's research team. On WebDG, the CSSL feature is called "Advanced Programs," and it is one of two sections of the system that would be accessible only to caseworkers, not private citizens.
To use it, workers would have to be capable of writing their requests in the CSSL format, or of importing the CSSL text from a pre-existing source. Long says the researchers are working on a more "friendly" interface for those who don't know XML. The other "advanced" section, called "Advanced Inquiry," would allow caseworkers to interact directly with any database through a single interface. All databases, regardless of their schemas, would be accessible through that same web interface. Significantly, the databases in this section of WebDG are organized ontologically into three "domains of interest": families, blind people, and people with disabilities. A real-world version might contain dozens of ontologies and hundreds or thousands of programs, making clear organization extremely important for a caseworker trying to find relevant databases for a particular citizen. WebDG's other two sections, "Basic Inquiry" and "Simple Programs," would be accessible to private citizens as well as caseworkers. Citizens could use the "Simple Programs" search to find a social-service program by its name or its properties, and then look up their own personal information in that program's database. Or they could employ the intuitive "Basic Inquiry" web service for easy lookup of various types of information - such as pharmacies that take Medicaid and grocery stores that accept food stamps - that would otherwise be tucked away in separate databases. In a programming feat that graduate student researcher Brahim Medjahed describes as particularly groundbreaking, the team wrote an algorithm that would let citizens or case workers look up information on families seeking to adopt children while protecting those families' desired levels of privacy. Each family would be able to choose what information would be made public or kept private. The demo site provides an alternative search "without privacy protection," but in reality that option would not exist. All of WebDG's features are made possible not by merging databases or their underlying data, but by working with metadata that explains how to access each database. In no case is the schema of the underlying database altered. Some of the databases featured in the demo were written for Oracle, while others were written for Infomix, and one was written for ObjectStore, yet all work together seamlessly in WebDG. On the human side, the interface explored in WebDG could help simplify a social services system so frustrating and that some clients simply drop out without receiving their benefits, Bouguettaya says. Current benefits programs demand excessive paperwork and travel for even the simplest tasks, effectively sapping welfare recipients of valuable time and energy they could be using to become more self-sufficient, he says. The project's website cites the example of a hypothetical single mother who must make a total of six visits to four locations just to accomplish three simple tasks: registering for Medicaid, registering for food stamps, and finding a state-certified caretaker for her child. "It's a catch-22," Bouguettaya says. "You're trying to get out of welfare, but you can't, because you have to do all these things. It's an infinite loop." Social-service agencies and their employees suffer, too, losing valuable personal time they would rather spend with clients because they are busy navigating an obstacle course of incompatible information sources, and even traveling to multiple offices, Bouguettaya says. For FSSA, a particularly exciting aspect of Bouguettaya and Elmagarmid's research is the potential that it could be used across state lines.
As for Bouguettaya, he is already looking ahead to the next step forward. On this project, he says, "We've done what we we'd hoped to do, and more." But he hopes to apply for an extension of his Digital Government grant, which expires in 2003, to further explore the topic. Bouguettaya believes the future lies not in databases, nor even database ontologies, but Web services such as the CSSL-based "Advanced Programs" section of WebDG. He foresees an even more advanced, user-friendly XML service that collects basic information about a citizen's needs and immediately presents that citizen with the necessary information. The word "database" would never enter into the equation, at least as far as the citizen knows. "It's too difficult for people to understand data and where it's stored," Bouguettaya said. "We want to talk in terms of service." A citizen would never even have to request a certain type of information; it would be presented automatically. "You can think of all the universe of data that's out there as one single database," he said, "and I only see the part that I need." |
 |
This site is maintained by the Digital Government Research Center at the University of Southern California's Information Sciences Institute. |  |
" CONTACT " POLICIES | ||
| |
|||||
Using WebDG's "Simple Programs" feature, users can look up social-service programs by broad categories, such as "Food" or "Health," and by department, such as the Division of Mental Health or the Family and Social Services Administration.
Once they have found their program of choice, such as Medicaid, users can view their own personal information and make changes if necessary.