Another Failure to Launch?
After string of IT disasters, have state agencies learned any lessons?
Six years ago, at the direction of the Minnesota legislature, the Minnesota Department of Human Services (DHS) began planning to modernize its aging IT infrastructure. As the planning phase for the four major IT system modernizations comes to a close, the price tag for the project is coming into view. In FY2018, DHS plans to spend about $225 million on four major modernization efforts. Full funding for this multi-year project could top $500 million.
Can the state successfully pull off a project of this scale? The state faces long odds, considering its record of botching nearly every major IT project. Further lowering the odds
of success, DHS plans to continue using the same vendor responsible for the failed launch of MNsure—the state’s Obamacare insurance exchange.
State taxpayers deserve to know what will be different this time. When the legislature reconvenes in 2018 there are things they can do to help improve the chances of success. Oversight will be key. The legislature can also simplify elements of pubic programs to simplify the design requirements of the new software. Shifting to the federal insurance exchange IT infrastructure would also remove an ongoing IT headache.
DHS modernization plans
DHS makes a strong case for the need to modernize. Many of the IT systems supporting state human services programs are over 25 years old. The staff with the expertise to maintain these antiquated systems are themselves aging and retiring. The multiplicity of systems built up over the years use different platforms making it hard to share information. All of this creates technical challenges any time the legislature makes a programmatic change.
Another compelling reason to modernize now is the availability of enhanced federal funding. Major portions of the modernization plan qualify for a 90 percent federal funding match.
The plan focuses on four major modernization efforts. The first and most ambitious effort is the development of a new Integrated Service Delivery System (ISDS). The ISDS will provide one integrated platform to provide access to multiple services, including cash assistance, food stamps, health coverage, and other social service programs.
DHS also plans to modernize the troubled system MNsure uses for enrollment and eligibility determinations, the Medicaid Management Information System, and the Direct Care and Treatment system that provides direct care to over 10,000 people being treated for mental illness, developmental disabilities and chemical dependency.
Ultimately, these four IT modernization efforts aim to create “a streamlined, person centered delivery system by integrating several existing systems and re-thinking and improving social service delivery in Minnesota.”
The multiplicity of systems makes this task difficult from the start. But the numbers of people interacting with the system underscore the size and complexity of the task. Over 31,000 county, tribal and state workers use these systems to serve over 2.8 million people. More than 200,000 providers also work within these systems.
Whether or not MN.IT—the state’s IT agency—and DHS can successfully implement a project of this scale will depend on whether these agencies have learned from past mistakes.
History of failure
In recent years, the state of Minnesota botched nearly every major IT project it tried to implement. These projects include HealthMatch, MNsure, and, most recently, the Minnesota Licensing and Registration System (MNLARS).
HealthMatch proposed to automate eligibility determinations for Minnesota’s health care programs through a new web-based system. Like the planned ISDS, HealthMatch aimed to combine functions of existing IT systems. DHS documents explain how consumers were supposed to be able to easily apply for health care programs over the Internet and the processes for workers were to be streamlined. The project was the basis for advancing “a sound long-term vision for Minnesota’s publicly funded health care programs.”
But this vision never materialized. Instead, HealthMatch became a “perfect storm of IT failure,” according to the headline from Michael Krigsman’s article describing the project at ZDNet, a popular IT website.
HealthMatch started with a budget of $13 million in 2002 and a delivery date set for mid-2004. By 2007, the Office of the Legislative Auditor released a detailed report explaining how HealthMatch faced serious risks due to changes in project scope, underestimates of the projects complexity, delays in completing the design, unexpected staffing problems, shortcomings by the contractor, and overly optimistic timeline estimates. At that time DHS staff admitted they underestimated the project’s scope—in particular, the 16,000 eligibility rules the project needed to apply—but they remained committed to HealthMatch.
However, these problems could not be overcome and DHS ended up scrapping the entire project. The final bill for this failure amounted to $41.25 million after the state settled to pay an IT vendor $7.25 million in 2011 to finally close out the project.
Around the time Minnesota closed out HealthMatch, President Obama signed the Affordable Care Act (ACA) into law, which initiated the next big IT project. The ACA asked each state to create a health insurance exchange to help enroll people in public health care programs and private insurance coverage with federal tax credits. Minnesota agreed and, in 2011, began planning to launch MNsure on October 1, 2013.
The day MNsure launched, the stories began piling in from people who could not create accounts. What followed was a parade of IT failures, some of which have yet to be resolved today. In addition to those who could not create accounts, people were stuck in the system after they created accounts. Account records went missing. Insurance cards took months to deliver. Deadlines for sending out tax forms that people need to file taxes were repeatedly missed. Insurers received multiple applications from the same customer. Enrollment data needed to be manually inputted.
Government audits of MNsure continue uncovering serious problems. Recent federal audits identified “significant” security vulnerabilities and revealed Minnesota was one of only two states in 2016 that failed to submit data the IRS requires to verify tax credit claims. An audit by Minnesota’s legislative auditor found that 38 percent of the people they sampled were not eligible for the program MNsure enrolled them in and estimated these errors resulted in the state overpaying between $115 and $271 million over a five-month period. A more recent audit commissioned by DHS continued to find a 16.7 percent eligibility error rate.
These eligibility errors are made by the Minnesota Eligibility Technology System (METS), one of the four IT programs slated for modernization. Up until November 2015, METS had been referred to as MNsure by state officials. Apparently, state officials grew tired of talking about the ongoing failures with MNsure—a highly visible public program—and opted to pin the failure on a less politically charged acronym.
In May 2017, MNsure issued a Request for Proposal (RFP) that shows it’s ready to scrap key IT components. The RFP asks for proposals “that can replace the plan shopping and system of record technologies that currently support the individual market exchange.” The RFP also seeks to “identify solutions to replace the technology that currently supports the small business market.”
Couple this RFP to replace key MNsure components with the modernization plan to pump another $55 million into “addressing system defects and providing additional functionality” for MNsure and DHS appears to have completely given up on MNsure’s IT infrastructure. According to DHS documents, the tab for developing MNsure totaled $262 million through FY 2017. If DHS is indeed starting over from scratch, that’s yet another complete loss to state and federal taxpayers.
While MN.IT and DHS have been struggling to make MNsure work, the Minnesota Department of Public Safety’s Driver and Vehicle Services (DVS) has been struggling for an even longer time to launch MNLARS, a new web-based replacement for DVS’s 40-year-old system that manages driver and vehicle transactions.
DVS’s original plans projected MN- LARS would be completed and the old system retired between 2009 and 2012, at a cost of $48 million.
Like HealthMatch and MNsure, MNLARS was quickly beset by delays and problems with its vendor. It took two years from getting the go ahead from the legislature for DVS to publish a request for proposal from vendors. After initial negotiations with one vendor failed, it wasn’t until April 2012 when DVS finally signed a contract with Hewlett-Packard (HP) to start working on the project. Thus, it took nearly the entire time that DVS originally planned to complete the project to just sign a contract.
Then, in July 2014, DVS decided to terminate HP’s contract, citing a failure to meet deadlines and poor quality work. As the Star Tribune reported, an April 2014 audit “found a disconnect between the HP team and state employees working on the project, fueled by lack of communications and ‘incompatible goals and visions.’”
With HP fired, DVS then partnered with MN.IT to develop MNLARS in-house. Since then, they have been quietly working, maybe too quietly. A June 2017 review by the Minnesota Office of the Legislative Auditor concluded “project delays and vague communications, particularly about timelines, have eroded confidence in MNLARS and its development process.”
In July, DVS began the process of launching MNLARS to the public, which will soon confirm whether or not MNLARS provides a usable service.
The early news reports a much bumpier launch than expected. A week into using the new system, Minnesota Public Radio interviewed Suzanne Jensen, the president of the Minnesota Deputy Registrar’s Association. According to Jensen, “We definitely knew that this would be a difficult transition and it would be very challenging, but I never imagined it would be the way it is.”
Time will tell whether MN.IT and DVS can eventually fix MNLARS’s defects. Time, however, has already proven that DVS severely botched the MNLARS project. A project that was supposed to be completed in four years is now in its ninth year and still not complete. As of July 1, DVS had spent $79 million—about 65 percent more than originally budgeted.
Projects fail for similar reasons
The many audits and reviews of the HealthMatch, MNsure, and MNLARS projects all reveal similar mistakes that, to varying degrees, caused severe delays, ballooning budgets, headaches for users, and, in the case of HealthMatch—and possibly MNsure—absolute failure. These mistakes include:
- Unclear communication, especially between state IT staff and vendors;
- Inexperienced program management;
- Little transparency in acknowledging challenges;
- Choosing a vendor that lacks adequate or compatible technology solutions;
- Poorly tested technology;
- Shifting project requirements; and
- A failure to appreciate the complexity of the project.
The fact that the state continues to make the same mistakes on major IT projects calls into question whether MN.IT and DHS are prepared for the largest IT project ever undertaken in Minnesota. Though learning from past mistakes and avoiding those same mistakes will be critical to DHS’s modernization efforts, planning documents reveal the agency may be once again picking the wrong vendor.
DHS plan relies on vendor that botched MNsure
Choosing the right vendor with the right expertise and the right technology is critical to the success of any large IT project. MN.IT and DHS have chosen to continue working with IBM Cúram, the same vendor responsible for creating MNsure’s dysfunctional IT backbone.
Why would DHS opt to stick with the vendor that failed to deliver on MNsure? That’s a question MN.IT and DHS need to be well prepared to answer.
There’s no question Cúram’s technology played a big role in MNsure’s failed launch. The proof lies in how Maryland’s exchange—using the same Cúram technology—failed in similar fashion.
It may be that Cúram’s insurance exchange failures do not reflect on their track record providing IT solutions to support other human services programs. In 2014, the Vancouver Sun reported on how “British Columbia is alone in wrestling with crashes, confusion and backlogs by its troubled social services computer system.” British Columbia chose Oracle Siebel software for its child welfare IT system when most other provinces chose Cúram.
However, shortly after the Vancouver Sun reported favorably on Cúram’s software, news outlets in Ontario began reporting how Cúram “bungled” the launch of the province’s new software for managing social assistance claims. A report prepared by the City of Toronto documented how the time spent on a new application grew from 15 minutes to 45 minutes under the new system. What once took four mouse clicks to enter client monthly income took 22 clicks under the new system.
It’s important to note that these sorts of failures are not uncommon. One study of IT projects by the Standish Group finds that only 10 percent of IT projects over $10 million are successfully delivered on time, within budget, and with required features and functions. Thirty-eight percent fail, either because they are canceled or never used. The high failure rate, however, also suggests a large IT project may not be the best IT solution.
Ultimately, there are only a few companies capable of supporting these large social services IT projects. Considering large IT projects have such a high failure rate, Cúram may be the best among this limited number of bad choices. MN.IT and DHS must, however, understand that picking the same vendor that botched MNsure will raise eyebrows and demand a full accounting of how they made that choice.
Legislative steps to help
When it comes to questioning MN.IT and DHS on their choices and their progress, that job largely rests with the state legislature. Stronger legislative oversight will be critical to the success of the DHS IT modernization plan. For prior IT projects, strict oversight came only after state agencies could no longer hide their failure. At that point, it was too late. When the legislature reconvenes in 2018, they should establish a new oversight committee focused on this IT initiative.
There are two more things the legislature can include in their 2018 agenda
to help the modernization process along. First, the legislature can reduce the complexity of the IT project by reducing the complexity of the eligibility system for public programs. Recall how one of the major missteps leading to the downfall
of HealthMatch was that both the vendor and DHS failed to appreciate the project’s complexity. At the time, Minnesota’s health care programs had 16,000 eligibility rules the project needed to apply. This complexity is the result of decades of legislative tweaks. Next year the legislature should work with DHS to identify eligibility rules and other elements of human services programs that can be simplified.
Second, the legislature should shift to using the federal insurance exchange’s eligibility system for commercial insurance. States have the option to choose a state partnership exchange, in which the state uses the federal government’s enrollment and eligibility technology. A smart partnership could allow Minnesota to offload MNsure’s role in qualifying people for tax credits and relieve a major IT headache.
DHS might not welcome all of this legislative “help,” but, in light of past failures with large IT projects, the agency needs to approach these IT projects differently. A more open and transparent process would be a good place to start.