Medicare Transactions

he Health Care Financing Administration, the agency that administers Medicare, spent six years, from 1991 to 1997, and at least $50 million developing the Medicare Transaction System.

Today, HCFA's 82-page strategic plan for information technology over the next five years makes no mention of the ill-fated system, instead only making oblique references to the errors of "earlier HCFA IT investment efforts." HCFA Chief Information Officer Gary Christoph, who was hired after MTS development was canceled, says the agency is moving on.

"Some of the concepts in MTS were really good concepts. Where we fell down was the implementation," Christoph says. "When you're growing up, you first learn to crawl, then walk, then run, then ride a bicycle. You don't just jump into a Ferrari."

The history of HCFA's computer problems is typical of federal agencies: Congress creates numerous programs over the years under the agency's domain. The agency builds a new information system for each new program, using various programming languages, platforms and contractors for each. The systems tend to be proprietary (making it hard to make updates or get them to share information) and written in languages popular in the 1950s and 1960s, such as COBOL.

Then, sometime in the 1980s, it dawns on the agency that its systems are becoming unmanageable. So it decides to build one large, integrated system to handle all the agency's programs-a solution expected to reap huge cost savings over time (after a significant up-front investment, of course). Overseeing one system is obviously less expensive than managing multiple systems. Technology managers picture that glorious day about five years down the road when they will cut the ribbon, unsheathe the new system and bask in the applause of the agency administrator, doting members of Congress and cheering taxpayers.

But you know what happens instead.

IG Criticism

In HCFA's case, the agency issued a request for proposals for its mega-system on Sept. 17, 1992. Even before the call for bidders hit the street, the Health and Human Services Department inspector general recommended that HCFA pause and "follow a more strategically oriented approach to streamlining, consolidating and integrating Medicare claims processing."

Four months after the request for proposals went out, the inspector general suggested that HCFA needed to more clearly define what exactly the agency wanted the Medicare Transaction System to do. The IG also recommended that HCFA make sure contractors understood that Medicare is a constantly changing program that needs flexible, adaptable software. In response, HCFA established 11 work groups to examine various aspects of Medicare processing and provide information to the contractor. Later, the number of work groups would be expanded to 24.

In the meantime, HCFA awarded GTE Government Services the MTS contract in January 1994. In April 1994, HCFA hired Intermetrics Inc., a small IT company in Virginia now known as AverStar, to independently monitor the contract's progress.

"This new system will bring to reality a new era of customer service and convenience for Medicare beneficiaries," beamed HHS Secretary Donna Shalala in a January 1994 press release. "It uses state-of-the-art technology to replace the forms and the hassles that have characterized Medicare in the past." The system would be completed by late 1998, the release said.

GTE began planning its strategy to help HCFA combine into one system Medicare's 14 systems at 60 sites operated by more than 70 contractors, while the work groups tried to make sense of the agency's systems architecture.

Danger Signs

The danger signs were all around. In December 1994, the inspector general again warned HCFA that it wasn't giving GTE enough information to help the contractor plan for the complexities of Medicare processing. The following year, HCFA assured the IG that it had begun providing ample information to GTE. The system was on track for full implementation by September 1999, HCFA said.

In November 1995, the General Accounting Office highlighted several concerns it had with MTS in a report to the House Government Reform Committee, including unclear contract requirements, a tight schedule and a lack of reliable cost-benefit information.

Then-HCFA Administrator Bruce Vladeck responded: "We are well aware that this is a high-risk venture. We are conducting the project with a keen appreciation of the need to manage the risks involved."

In mid-1996, Intermetrics warned HCFA and GTE that the MTS development schedule had too many overlapping milestones, but that warning went unheeded, an Intermetrics official would testify at a House hearing in May 1997.

A month before that hearing, HCFA ordered GTE to stop work for 90 days on all but one of the segments of MTS to evaluate the reasons for the project's increasing cost overruns and missed deadlines.

By May 1997, GAO was more adamant that MTS was in big trouble. HCFA had not planned the project's schedule well enough, did not have performance measures to gauge progress, and had failed to use "cost-benefit analyses and other tools to continually track and assess whether funds spent on MTS will contribute to a return on this investment," GAO said (T-AIMD-97-91). In addition, the projected costs of MTS had mushroomed from $151 million to $1 billion. GAO estimated that HCFA had spent $80 million on the project so far, though HCFA says the total spent on MTS was $50 million.

In response, HCFA, GTE and even Intermetrics vowed that MTS would be completed, albeit probably not until 2000.

Three months later, HCFA canceled its contract with GTE.

'Too Much Confidence'

"We had one major mistake in the life of [the MTS] experience, and that was probably having too much confidence in the ability of the contractor to deliver," former administrator Vladeck says.

Until mid-1998, HCFA continued to work on reducing the number of systems it relied on. Then it suspended all modernization efforts to make sure that 50 million lines of software code were free of potential Y2K glitches.

HCFA didn't come out of the MTS project completely empty-handed. The agency did cut down the number of contractors and sites managing Medicare processing systems, and integrated its 14 systems into six. In addition, the project gave agency officials a better understanding of how their systems, which are controlled primarily by contractors, operate.

"HCFA came out of [MTS] with a much clearer short- to medium-term strategy for managing claims processing," Vladeck says.

In fact, Vladeck says that MTS probably still could have been salvaged after GTE's failure, except for two factors. First, Congress and the Clinton administration failed to consider many of the benefits that would justify the cost of an expensive single, integrated system, including better fraud detection and improvements in processing efficiency. Second, the project fell victim to the still-prevalent notion, first championed by former Office of Management and Budget Director Franklin Raines, that single "big-bang" procurements just don't work.

"At the time the executive branch was so dominated by Director Raines' opposition to single, comprehensive procurements that even though you could still make an intellectual argument for MTS, we didn't think we had any chance of convincing OMB," Vladeck says.

The project failed to realize the vision espoused by Shalala in 1994. There is no single system for Medicare processing. According to HCFA's new five-year information technology plan, its current systems still "clearly reflect business and system design philosophies of an era when, for example, claims processing was largely a paper-handling function. . . . Medicare processing systems and HCFA central office IT infrastructure remain substantially similar to the technology implemented 10 to 15 years ago, despite operating on newer equipment."

Lessons Learned

HCFA is not much better off now than it was in 1991, in part because the technology staff has spent the past two years resolving Y2K problems. But HCFA did glean lessons from the effort, which are reflected in the agency's new IT strategic plan.

• Set realistic milestones. HCFA's contract with GTE essentially gave the contractor three years to submit a final package. An adjustment to the contract requiring incremental delivery of software came too late, and even then HCFA allowed milestones to overlap and did a poor job of monitoring GTE's progress. HCFA now plans to use "modular contracting," which separates modernization efforts into distinct phases under separate contracts.
• Know what you want before you ask for it. Not only did HCFA not have a complete grasp on how its systems operated, officials kept revising their requests to GTE as it became more apparent that the modernization effort was very complex.
• Avoid mission creep. New requirements kept getting tacked onto MTS. "There are must-haves and wanna-haves. You can't let the wanna-haves outweigh the things you must have," Christoph says. "You don't waste your energy on things that are niceties."
• Have an investment review process. The MTS effort was kept under close scrutiny by an internal management board, an independent validation contractor, the inspector general's office and the General Accounting Office. But despite the layers of oversight, there was no structured investment review process. HCFA has since created such a process to calculate return on investment, establish performance measures, tie technology efforts to agency business objectives, identify risks that could derail investments, set out a detailed schedule for accomplishing all phases of a system and ensure that IT purchases are consistent with the agency's overall technology architecture.

"We developed an architectural vision that includes not just claims processing," Christoph says. "We want to build an enterprise-wide information structure that is data-centric at its core. Then we will have enterprise databases that are accessible to all our applications that need the data."

• Manage expectations. HCFA promised the moon but delivered far less. The job of sorting out the agency's poorly documented and ad hoc software systems was undeniably enormous. HCFA's new five-year plan explains the difficulties of recovering from years of poor IT management. In addition, the five-year plan describes a slow and steady approach to modernizing the agency's systems.
• Reduce risk. HCFA's new five-year plan calls for performance-based contracting for IT projects, with incentive payments tied to meeting quality and schedule requirements. The agency will also require contractors to provide monthly reports describing any variances in cost, schedule or performance.

HCFA learned more IT project management lessons from its successful battle with the Y2K bug, including how to engage program managers in technology projects and prioritize. Y2K was a do-or-die situation, but Christoph hopes to build on the confidence restored by successfully dealing with the millennium bug.

"We've won the war, and now it's back to peacetime. There are a lot of competing priorities to deal with," Christoph says. "When you can build support for products that is real because they satisfy the business's needs, then you have a much better chance of being successful. I'd like to think we're taking these lessons to heart."