In a cool, hushed room on East Magnolia Street, Maricopa County officials threw away more than $12 million in taxpayer money last week.
The county's Health Care Agency--with the blessing of the Board of Supervisors--began to dismantle a brand-new, multimillion-dollar computer system that was mostly paid for but had scarcely been used.
Pleading poverty, health officials contended that the system, within months of kicking into full gear, was simply too expensive to finish. Rather than invest several million more dollars to complete the project, they decided to scrap it altogether.
What had begun several years ago as an effort to help one of the county's largest agencies emerge from the technological dark ages instead ended as one of the most expensive financial blunders in county history.
As explained by county officials, and briefly reported in the daily newspapers, the decision to scrap the system was unavoidable. Looking for ways to cut $30 million from his budget--and contemplating laying off as many as 200 employees--acting agency head Ted Williams says he cannot afford to sink any more money into a computer system.
"We're going through the most aggressive reengineering in the history of a county department," says Williams, who took over the agency earlier this year. "This is a good management decision to live within the budget we've been given."
But the issue is hardly that clear-cut, say critics and county insiders, and under scrutiny, it is hard to discern a compelling reason for flushing an investment of more than $12 million.
According to numerous reports prepared by various consultants and county employees but not widely publicized, Williams' decision qualifies for enshrinement in the penny-wise, pound-foolish hall of fame.
The expected costs of the computer system have been known for years, and the project was coming in on budget, according to an internal county audit. Its final cost should not have been a surprise to anyone.
Virtually everyone agrees that the computer itself was fully capable of meeting the agency's technological needs, and, in fact, would have saved the county money when it was operating by improving patient billing and recordkeeping.
By scrapping a system that is already almost paid off, critics point out, the agency will now have to buy another one, but nobody has a clue what it will be or how much it will cost. They wonder how another new system can possibly be cheaper than one that's already all but paid for.
"They've got this much invested, and they're snatching defeat from the jaws of victory," says Steve Wilson, one of the numerous system planners who worked on the project. "Did we know how much this was going to cost? Absolutely. Has that changed? Absolutely not. Everybody approved all of this. Everybody knew what the cost was going to be."
The decision has left many involved in the project dumfounded. They acknowledge that the county is in severe financial straits, but see the decision as a shortsighted and capricious overreaction to a budget crunch.
Startled by the seeming abandon with which county management is throwing away a $12 million investment, Supervisor Ed King has asked the county's internal auditor to reconstruct the chain of events that led to the computer debacle.
But board members and their staffs, King and others say, are not free to explain why they signed off on the plan to scrap the system, because the County Attorney's Office has told them to keep their mouths shut in case lawsuits spring from the decision.
"The County Attorney's Office has advised that everyone be very prudent about what is said about the situation," says Tom Simplot, aide to Supervisor Betsey Bayless.
Whatever the merits of the decision, King says, it raises unsettling questions about the competency of county management when it comes to computer and automation planning.
"When it comes to electronic data processing in the county, it appears we have a headless horseman," King says. "Nobody's in charge."
@body:Buying a multimillion-dollar computer system is hardly an impulse purchase, especially for a government agency. There are the predicate consulting reports, meetings, committees and forests of projections and plans. Phrases like "platform," "applications" and "gateways" are bandied about, presumably by people who know of what they speak.
In the case of the Health Care Agency, according to many who were involved in the process, about five years were spent figuring out what kind of computer was needed and whom to buy it from.
Employees from all parts of the agency--the people who would actually have to use the computer--were called in and asked what they wanted and needed to do their jobs.
There was little question about the need for a new system. The agency's last big computer acquisition was a Honeywell system bought in the late 1970s. The running joke around the agency was that, aside from keeping track of dog licenses, the computer couldn't do much.
"The Honeywell is old technology," says Joe Martina, who headed the agency's health information systems center during a large part of the planning process. "It was fine for the day it was purchased, but those days are gone."
With its frequent crashes and bulky programs, the system was not up to the Health Care Agency's many tasks--running the county hospital, operating neighborhood health clinics and managing a long-term care program, among others.
Adolfo Echeveste, who headed the agency for ten years, until he was forced out earlier this year, says the health-care system "was losing its shirt" because it did not have adequate computer systems.
The winds of health-care management were already shifting to competition and prepaid care, he says. The creation of AHCCCS (Arizona Health Care Cost Containment System) and other programs that were setting standard reimbursement rates for care meant that health-care systems were going to have to get efficient to survive.
"When you're in a prepaid, competitive, managed-care environment, it is absolutely essential that you have the automation and the tools to define very quickly and in detail the costs for services--that you're able to handle and manage the system and collect data for billing purposes," Echeveste says.
By 1990, after five years of study, agency planners felt confident that they knew what type of system was needed to keep track of thousands of patient admissions and billing records, medical records, inventory data and the myriad other bookkeeping and records tasks attendant to being a health agency.
That year, contracts were signed with IBM to purchase a mainframe that would be the backbone of the system. Another contract was signed with IBAX--a partnership consisting of IBM and a second company--to provide the software for the system.
The system, according to IBM, is used by more than 100 hospitals across the country--including Samaritan Health Services and Tucson Medical Center in Arizona.
Installed as the system's heart was an IBM 3090, a string of big boxes that sits in an air-conditioned room and hums, serving as a repository for most any data health-care employees might need to tap into.
For almost its entire duration, the project moved along under Echeveste's leadership.
"Adolfo wanted this to be the system that would take the agency into the 21st century," says planner Steve Wilson.
There were problems with the project, although not unusual ones for an undertaking of such scope. Squabbles broke out over how to tailor the software for the specific needs of different branches of the agency.
Not infrequently, county records show, IBAX, IBM or others involved in the project complained that there was a lack of leadership, that no one was making hard decisions about how much time and money would be spent modifying the software for different uses.
In general, however, agency leaders and employees were confident they would soon have a whizz-bang computer helping them do their jobs.
(Months after new administrators decided to scrap the project, the county's internal auditors looked over the project's financial records and found that, while there had been some sloppy bookkeeping, the project was up to snuff in terms of its budget. The auditors "did not observe any expenditure of funds which were [sic] inconsistent with the terms of the contract.")
But then Echeveste, the project's patron, was run off as head of the agency by County Manager Roy Pederson, because the county hospital was hemorrhaging money.
Williams was appointed to fill the post temporarily, and charged with bringing the agency budget under control. From that day on, observers say, the county's investment was doomed.
@body:"Computers are really like ladies' hemlines. They go up. They go down. It depends what's fashionable," says an insider who observed the county's decision to abandon its computer investment.
Big mainframe computers that sit in their own quiet rooms have, of late, fallen out of fashion among those claiming to be on the cutting edge of the industry.
That disfavor explains why, earlier this year, IBM was forced to lay off employees for the first time in its history.
The trend, instead, is toward what are called distributive systems--smaller, more powerful computers linked by various networks. Instead of one, humming central repository for data, information is stored around and about, wherever it is the most useful, in computers that can talk to one another when necessary.
Simply put, it's a question of big boxes versus little boxes, and right now, the little boxes are winning. That is not to say that there is anything wrong with big boxes, particularly if they're already paid for.
There is little difference in what the two types of systems can do, but the newer, nonmainframe systems can be cheaper and easier to customize for each user's needs.
Had the county not already invested so heavily in a mainframe system, most planners agree, it might well want one of the newer systems. But the decision to go mainframe was made three years ago, and the money has been spent.
When Williams and other new administrators moved in to take over the Health Care Agency, several insiders say, they decided they wanted little boxes, not the big boxes that had already been purchased. They wanted to be fashionable.
The IBM/IBAX system was already being used by some parts of the agency, and within six months was expected to be fully up and running. Terminals had been parceled out to many of the agency's branches, and software designs were being finalized. The agency's Medical Records unit was already on the computer, as were some of the laboratory data and some claims processing.
Although the county had paid outright for much of the hardware and software, the IBM 3090 mainframe was on a lease-purchase program, and the county still owed more than $1 million for it.
Ditching the fledgling data center after $12 million had already gone out the door required some justification. That, the cynics say, is what consultants are for.
Shortly after taking over the agency, Williams ordered work on the system to stop, citing the need to reduce a budget shortfall estimated at about $30 million.
S.K. Ching & Associates, a small, San Francisco-based consulting firm, was brought in to study the full range of the Health Care Agency's financial woes. After a 45-day, fire-drill review of the agency, Ching & Associates turned in a host of recommendations, including walking away from the computer system.
The report did not question the need for the computer or the system's ability to do its job. In fact, the report praised the system--in a backhanded fashion.
"Certainly, it is a nice-to-have system and it can be justified in an up-market," the report concluded. "But in a down-market and given these economic times that require health delivery institutions to do much more with much less, some of the nice-to-haves may just have to go by the wayside."
The Ching report makes a reasonable point, but, critics contend, it ignores one rather salient fact: The IBM/IBAX system was already all but paid for.
The report, hotly disputed by a wide range of planners and county employees who actually helped design and implement the system, went on to estimate that the county could save $41 million over five years by scrapping the system.
Insiders say the report merely affirmed what Williams and other county administrators had already decided.
"The Ching report was the vehicle they used to go away from the centralized system," says one county insider. "They had already made their minds up."
Williams disputes the significance of the Ching report in the decision. With no cash to continue putting into the system, he says, the eventual cost of finishing or not finishing the system is meaningless. Whatever the cost, the agency does not have the money to pay it.
"It [the Ching report] has no relevance to the decision," Williams says. "Everybody's trying to make too much of a clear issue. I don't have the budget to put into the computer. It is difficult to go out and buy a new car if you don't have money in the bank."
Despite Williams' protestations, the Ching report's findings have fueled widespread perceptions that ditching the system will save the county more than $40 million, and helped convince the Board of Supervisors to affirm the decision.
Among the witches' brew of numbers and cost estimates now swirling about, Ching's $41 million figure has come to hold sway, parroted with little challenge in virtually every news account of the computer system's demise.
"We call it the 'I Ching Report,'" says former county computer planner Wilson. "It turned into a religious document. Nobody questioned it. At least the board didn't."
The savings predicted over five years by Ching is wildly inflated, say many of those involved in the project.
Right off the bat, they point out, the entire budget for the agency's computer operations was only about $7 million per year before the project was halted, including everything from salaries to copying machines and newspaper subscriptions. That means that wiping out every dime of money for information services could only save $35 million over five years.
"That right there tells you that the Ching numbers are bogus," says one county insider.
Further, critics point out, the Ching report recommended abandoning one system without providing any specific estimates of what it will cost the county to buy another one to replace it.
In a seeming contradiction of logic, the report contends that the county can throw away $12 million of investment, buy another completely new, unspecified system and still save $41 million.
That assertion boggles minds throughout the hierarchy of planners and employees who have invested years trying to put the IBM/IBAX system in place.
"It hasn't made any sense to me from the beginning," says Tom Rowan, the agency's technical support manager. "They [Ching] had specific dollar figures, but we have no idea how they were arrived at. We don't even have a replacement system specified."
(S.K. Ching & Associates referred questions about its report to Ted Williams at the Health Care Agency.)
In response to the Ching report, IBAX and several county employees proffered their own assessments. Most estimated that finishing the system would cost somewhere between $1 million and $3 million--depending on how many bells and whistles were included in the final product--and after that, the county would actually start to save money.
For instance, just by reducing the agency's accounts receivable--that is, billing people and getting them to pay in a more timely fashion--IBAX and county employees estimated that the new system could net the county $10 million over four years.
Overall, IBAX reported, finishing the system would ultimately save the county $18.4 million over four years. IBM also offered to help the county through its cash crunch by extending the lease terms for the system's hardware so the payments could be lowered.
But those reports, insiders say, were overwhelmed by Ching's gloomy assessment, and members of the Board of Supervisors were left with a lopsided view of the world when they approved the agency's decision to scratch the IBM.
Carefully attempting not to comment directly on the decision, Ed King says he believes the incident indicates that the Board of Supervisors has not been fully informed about what the county's management is doing with the money it is given for computer systems.
County management, he says, seems to have an attitude that the board does not want to know the nuts and bolts of what county employees are doing.
"The staff feels the Board of Supervisors is not interested in helping them solve their computer problems," King says.
@body:Although the decision was made months ago, the final bullet was fired into the IBM/IBAX system last week. The new system is now being shut down, and left awaiting the tractor-trailer trucks that are supposed to come in and haul it away.
So what is the health agency using as a computer? To a large extent, it is once again relying on the 14-year-old Honeywell system that everyone agrees can't do the job.
In its report, Ching & Associates concluded that the Honeywell could limp along until another, new system is found and put in place. But critics point out that the Honeywell, with all due respect for its service over the years, is well past its prime. In fact, no one knows how much it will cost to keep nursing the thing along while more consultants, committees, reports and planners decide what to do next.
The longer it is in use, experts say, the more the Honeywell will cost, and the greater the risk of a major failure that could wipe out crucial agency records.
"I've told the folks here the Honeywell system is held together with spit and gum," says technical support manager Rowan. "We're looking for a disaster."
It is not unusual for the Honeywell system to crash on any given day, Steve Wilson says, and the entire system must be shut down for hours every Thursday for maintenance. While there has been no significant loss of data in any of the crashes to date, Rowan says, the risk is constant.
"We've been taking disk-drive hits," Rowan says, in what constitutes apocalyptic language among computer types. He means that in some crashes, files and information are just disappearing. Sometimes they can be retrieved, sometimes not. Such disasters can be expected to continue so long as the ancient Honeywell system is in use.
Desperate to maintain its computer capabilities after the IBM/IBAX system was shut down, Rowan and others say, the agency had to come up with a quick fix to keep operations going.
If you like this story, consider signing up for our email newsletters.
SHOW ME HOW
You have successfully signed up for your selected newsletter(s) - please keep an eye on your mailbox, we're movin' in!
So last week, in preparation for the departure of the IBM 3090, agency computer technicians were rushing to prop up the moribund Honeywell system.
The agency bought another IBM mainframe, this one much smaller than the 3090 that is being replaced. Williams says it cost the county about $80,000.
So until the county decides what will replace the IBM mainframe it no longer wants, it will use another IBM mainframe, along with the ancient Honeywell system that almost everyone agrees should have been scrapped years ago. And $12 million-plus of taxpayer investment will be rolled out the door and hauled away.
In the words of one planner involved in the project, who asked not to be named for fear he would join the many others who have lost their county jobs: "It doesn't strike me as particularly logical.