Lead Healthcare.gov IT Contractor Gets the Boot: Why Contractor Oversight and Proper Planning Are Key to Effective Government
by Nick Schwellenbach, 1/14/2014
Late last week, The Washington Post broke the news that the Centers for Medicare and Medicaid Services (CMS) was kicking CGI Federal off of its contract to develop and operate the complex Healthcare.gov website, which was wracked with problems when it went live on Oct. 1, 2013. At the end of 2011, the Montreal-based company had won a two-year contract to do most of the work developing the federal health insurance exchange website. There were three, one-year options for the federal government to extend the contract. But the federal government did not exercise those options. Instead, it is awarding a one-year, $90 million sole-source contract to Dublin-based Accenture. While CGI is not the only contractor working on the website, it has been the only one substantially singled out by government officials.
That the federal government is arguably switching horses midstream is understandable in the aftermath of the politically embarrassing rollout of the website, key to the implementation of President Obama’s landmark legislative achievement. But could things have worked out differently? The answer is yes.
The Government Implementation Team Needed More Expertise in Managing Complex Projects
In May 2010, David Cutler, a senior health care advisor to Obama, penned a letter to Larry Summers, at the time Obama’s director of the National Economic Council. Cutler wrote, “My general view is that the early implementation efforts are far short of what it will take to implement reform successfully. For health reform to be successful, the relevant people need a vision about health system transformation and the managerial ability to carry out that vision.”
A “revamped and enhanced implementation group” at CMS needed to have expertise in five areas, according to Cutler. Two of them were in “managing large and complex enterprises” and “information technology systems.” Cutler wrote that CMS did not have what it needed: “The agency is demoralized, the best people have left, IT services are antiquated, and there are fewer employees than in 1981, despite a much larger burden.”
His concerns did not lead to the changes he argued were needed. “Over the past three years five different lower-level managers held posts overseeing the development of HealthCare.gov, none of whom had the kind of authority to reach across the administration to ensure the project stayed on schedule,” according to a New York Times article last year. “As a result, the president’s signature initiative was effectively left under the day-to-day management of Henry Chao, a 19-year veteran of the Medicare agency with little clout and little formal background in computer science.” Furthermore, CMS “reserved the role of general contractor, or system integrator, for itself, even though it lacked the necessary in-house software engineering resources to handle such a task.”
CMS is not unique in this regard. The federal government gutted much of its in-house expertise – the people who had substantive knowledge and could plan projects and call out contractors on their rosy projections of performance – in the late 1990s. For instance, the Department of Defense (DoD) – responsible for more than two-thirds of federal contracts – went from having an acquisition workforce of 460,516 in 1990 to 206,653 in 2004. The DoD has tried recently to rebuild aspects of its acquisition workforce.
Even had the federal online marketplace been a simple undertaking, Cutler’s concerns would have still been valid. However, it is perhaps one of the most complicated federal information technology projects ever, communicating with at least 112 computer systems to do things like verify identities, check with dozens of state and federal agencies to determine eligibility for financial help and a credit company to verify income, and have substantial security measures in place.
Federal Officials Should Have Paid Closer Attention to Contractor's Past Performance
CGI did not have a great track record when it was picked.
When CGI was chosen in 2011, CMS officials did not have to consider that many high-level CGI executives “had come from another company that had mishandled at least 20 other government information technology projects more than a decade ago,” according to a December Washington Post story.
CGI had broken into the federal contracting scene largely by buying a smaller U.S. information technology contractor called American Management Systems, which was based in northern Virginia. The AMS track record “could well have knocked [CGI Federal] out of the competition, and probably should have,” one former CMS official told the Post.
However, in an evaluation of CGI for the Healthcare.gov IT contract, CMS officials found “no evidence of performance risk.” Curiously, although “the review found no performance risk for CGI Federal, agency officials had not always been satisfied with the company’s work,” the Post learned. “In 2010, CMS rejected CGI Federal’s bid under the umbrella contract to perform work on four health IT systems, in part citing unspecified performance issues in carrying out an earlier contract.”
But even had CMS focused more on CGI’s track record, the agency deemed past performance the second-to-least important factor when evaluating contractor bids.
Perhaps contributing to the less-than-robust review of CGI, the Post found that the decision to award to CGI – one of 17 companies on a list of vendors pre-approved in 2007 to bid on IT contracts at CMS, only four of which applied – may have been rushed. Although the Affordable Care Act was signed into law in March 2010, CMS did not begin soliciting bids until the second half of 2011. “There was very great time pressure,” a former CMS official told the Post.
A Complex Undertaking with Shifting Requirements Was Rushed without Proper Leadership and Planning
Accomplishing the goal of Healthcare.gov, even under the best of circumstances, is a complex and difficult undertaking. But it occurred in an environment that made a complex task almost inevitably a disaster. It is no surprise that a government oversight team without significant relevant experience and expertise, coupled with a contractor with a history of poor past performance, all without adequate resources, would botch implementation of such a difficult project.
Look no further than a McKinsey & Co.’s March 2013 “Red Team” assessment of the “trajectory of the federal marketplaces.”
While the federal website is not the only aspect that McKinsey’s team focused on, it was a substantial part of the problems they unearthed.
In essense, there was a perfect storm of factors that made the federal website development a mess. Requirements kept changing, and there wasn’t enough time to do adequate testing and implement fixes before rollout. Indeed, significant testing was to occur after the website went live, which is an enormous problem given the website was going live “at full volume” on Oct. 1, 2013 when traffic was predictably going to be very high.
McKinsey identified the “most critical risks to marketplace implementation efforts.” Root causes of system failures included “limited end-to-end testing prior to launch” and not enough capacity to deal with web traffic, among other causes.
Leadership would have been one of the top things that could have changed to enable better implementation. “Fast, targeted, locked-down decisions are needed for implementation effort,” the assessment states. But there were “no clear roles, responsibilities and processes for making change” and “no single empowered decision-making authority.”
Locking down “version 1.0” of the site and fully testing it and making final decisions on several requirements by the end of April were things that needed to happen, according to the consultants.
But months later, McKinsey's concerns were still valid. An exchange published by The New York Times illustrates the dysfunctional management of the project.
Henry Chao, the CMS manager of the program, did not seem to have enough control of CGI and other contractors. The Times article states:
One evening last summer, he called Wallace Fung, who retired in 2008 as the Medicare agency’s chief technology officer. Mr. Fung said in an interview that he told Mr. Chao to greatly simplify the site’s functions. “Henry, this is not going to work. You cannot build this kind of system overnight,” Mr. Fung said he told him.
“I know,” Mr. Chao answered, according to Mr. Fung. “But I cannot talk them out of it.”
Days before the site was supposed to go live, a test showed that it “could not handle more than 500 simultaneous users.”
The management of Healthcare.gov was dysfunctional on multiple levels. However, it is far from the only case. As Federal News Radio noted:
David Powner, GAO's director of IT management issues, testified complex federal IT projects can be successful if agencies follow seven acquisition best practices, including accountability, transparency, oversight, expertise and program management.
He said too often when federal IT projects fail, many of these best practices aren't used, which seems to be the case with HealthCare.gov.
Political Warfare: The Legislative and Executive Branches and the States and Feds Were Not Working Together
The managerial debacle of the online federal marketplace took place against a backdrop of not just partisanship, but political warfare that impeded the effective working of government. Imagine if a different set of circumstances existed. For instance: if constructive congressional oversight happened well before the online exchanges went live and highlighted the problems detailed above with the goal of fixing them, those problems might have been dealt with or at least mitigated more than they were. Also, if Congress would have adequately funded implementation along with proper oversight, things might have worked much better by Oct. 1.
Congress has both the power of the purse and an appropriate oversight and legislative role in our constitutional system of government. There can and should be some degree of debate and friction between the legislative and executive branches of government, but for any complex and significant federal government effort to succeed, there must be some mutual willingness to work together for success. But especially after the House of Representatives flipped over to Republican control in the wake of the 2010 midterm elections, many in Congress began working to throw monkey wrenches into the implementation of the Affordable Care Act.
Notably, while the Affordable Care Act included $1 billion to implement the law, Congress’s budget office projected much greater funding needs of $5 billion to $10 billion over a decade, according to a May 2013 Washington Post story. The Department of Health and Human Services (HHS) “has repeatedly requested additional funds from Congress to assist in the implementing but has been turned down,” the Post notes. “And because many states have refused to partner with the federal government in setting up the law, the burden on HHS has grown.”
The budget limitations and delays hampered proper implementation (note: HHS and CMS did manage to scrounge up some more money, although not enough, by defunding other needs). In management-speak, according to the McKinsey review: “Budget uncertainty and timing prevents execution of plan” since “many functions are contractor dependent”; “core contracts not awarded due to budget”; and it “hampers ability to hire/resource critical path activities.”
The confusion and embarrassment caused by the poor performance of Healthcare.gov in its early days is now increasingly in the rear view mirror, but not entirely. The website works far better now, thanks in part to a focused effort under the leadership of Jeffrey Zients and Todd Park, tasked by the White House to get the site working the way it should have on Oct. 1, and the assignment of another contractor, QSSI, as the general contractor to handle the day-to-day management and coordination long needed.
The sidelining of CGI isn’t a surprise. It’s being done by simply not exercising extensions of its contract, a fairly simple way of getting rid of them. However, legitimate questions about CGI’s ability to adequately perform on contracts should be asked by agencies before they decide to award to the company in the future.
Furthermore, the award to Accenture is being sole-sourced, meaning it is being done without competition on the basis that there are “time constraints.” There are some good reasons to think Accenture might work out better. Accenture was behind California’s state-run online health insurance exchange, which has worked far more successfully that the federal site. However, it has a record of some of its own troubled projects. Moreover, it is unclear why the administration could not have run a relatively time-limited competition. After all, CGI faced off against three other contractors when it originally bid for the Healthcare.gov contract. Another month or two operating under CGI in a short-term contract extension probably would not have adversely affected Healthcare.gov considering many of its major problems have been brought under control.
More fundamentally, though, the government needs to do a better job of managing its contractors and complex projects, whether it is setting up federal online health care marketplaces or buying complex weapons systems. Gutting the government of expertise is a sure-fire way to ensure that it cannot operate effectively. We need a government workforce that has both experience and expertise, not simply at the very top, but in the middle rungs with people who are not only driven but thoughtful and realistic about the tasks at hand. Unfortunately, many federal policies – such as years of pay freezes – that are the result of political warfare have driven away some of the best and brightest who used to work in government and are keeping younger talent at bay.