said. The state explained that caseworkers provide services to the children and families, and then record the information in the system.
Unfortunately the design limited the SACWIS's functionality because caseworkers could not enter information into the system after an event happened, the GAO said, and the state had to redesign the system to correct the design flaw.
Finally, designing a system that earns SACWIS compliance from the HHS entails overcoming troubling technical challenges. The GAO said 36 of the 50 states responding to its survey encountered technical issues, including matching state data element definitions to HHS's data categories.
North Carolina officials told the GAO that state policy requires every location where a child resides to be counted, including hospital stays, but one set of federal regulations stipulated that hospital stays and other short-term placements should not be included when counting foster care placements.
In cases where state and federal policies differ, states must carefully reformat their data to meet federal reporting requirements. Texas officials, for example, told the GAO that although a federal review of Texas' SACWIS instructed the state to modify its system to collect, map and extract data on guardianship placements, the state itself does not support guardianship arrangements.
This puts states in an awkward position, the GAO said. They must make sure they're reporting accurate data to the federal government but not contradicting state policies.
The dearth of states with a fully compliant SACWIS is somewhat surprising, said Lynda Arnold, director of the National Resource Center for Child Welfare Data and Technology (NRC-CWDT), but not completely unexpected because of the complicated nature of such systems.
"The issue with achieving tier-one status is just more complex than people had thought," said Arnold, who served as Child Welfare director in the Oklahoma Department of Human Services prior to joining the NRC-CWDT. During her tenure, Oklahoma became the first state to implement a federally compliant SACWIS.
The NRC-CWDT spends most of its time helping states with SACWIS data issues.
"Our main requests [from states] have been about converting data from their old system to their new system and not losing data quality in that process," she said. "A lot of our emphasis has been on mapping the data to meet federal requirements."
Perhaps the most misunderstood aspect of a SACWIS is that it's a child welfare management system, she said, not an information system.
"You've got to have that involvement from child welfare staff and from the leadership in child welfare," Arnold said. "Sometimes, getting the buy-in from the child welfare staff -- and getting that priority within the child welfare system when they're dealing with so many other issues -- is very difficult."
Early SACWIS development efforts appeared to be driven by the IT piece alone, she said, which is one reason why states ran into trouble designing systems that worked well while capturing and reporting data in ways that met federal guidelines.
One SACWIS development model that produced positive results and pleased the federal government was used by Dynamics Research Corp. (DRC) to design, develop and roll out a SACWIS for New Hampshire in 1997 and for Colorado in 2001.
DRC and New Hampshire rolled out New Hampshire Bridges, a system modeled on an earlier SACWIS originally developed in Oklahoma. DRC took the Oklahoma system from the public domain and adapted it to fit New Hampshire's organizational and operational needs, said Kathleen Perras, DRC's vice president and general manager of State and Local Programs.
"New Hampshire was one of the early states to adopt a SACWIS," Perras said. "The state wanted to take advantage of the 75 percent federal matching funds. A few states had gone after brand-new systems, but the federal government, at that