Blogs

Many state agencies have been modernizing their IT systems in the past decade, or are in the process of doing so — especially in health and human services.

Medicaid and public assistance eligibility and claims systems in particular have made local and state news amid growing pains in Medicaid Management Information Systems (MMIS) rollouts, showing the complexity of these programs and the equally complicated public-private contracting that makes them possible.

In North Carolina, the several-times-delayed implementation of the new Medicaid claims system NCTracks “avoided complete derailment, but problems persist,” as the North Carolina Medical Society said describing efforts by state officials to troubleshoot a variety of bugs doctors’ offices are experiencing — ahead of a pending federal audit of the system’s payment abilities.

In Louisiana, an MMIS implementation has been put on hold amid an investigation into possible conflicts of interest in the awarding of a contract that’s seen time and cost overruns — prompting media coverage portraying the state government and the contractor as somewhat inept.

Of course Medicaid information systems are a bit more complex than that, and the types of enterprise architecture and software-as-a-service approaches that have evolved recently in the private sector are perhaps more difficult to apply in public sector programs.

In Montana, though, the CIO is trying to avoid such kerfuffles with a modernization of Medicaid and public assistance.

Drawing form a background in the private and public sectors heavy in systems architecture, database management and systems analysis, Ron Baldwin became Montana’s CIO this year, after working as the CIO of the Department of Public Health and Human Services since 2009.

Government Health IT Associate Editor Anthony Brino spoke with Baldwin about how he has approached enterprise architecture in health and human services — trying to avoid the siloed models of old — and why he thinks the mainstream media horror stories of MMIS implementations are not necessarily exaggerations, but can be avoided.

Q: Working as health and human services CIO in 2009, as several IT systems were in the process of modernization, what were your implementation goals, given the recent histories of backlogs and overruns with both legacy and new systems?

A: My emphasis from a strategic planning purpose at the Department of Public Health and Human Services was to start a process of replacing the large human service systems, which are some of the biggest systems on the block in state government — the eligibility systems, the Medicaid management information systems and some of the others that need to be replaced, like child welfare and child support. Several years ago, money was appropriated to replace the eligibility systems for TANF, SNAP and Medicaid. The Medicaid system was fully replaced and put into production in 2009 on a completely new and modern platform.

What I set out to do there was take that as an opportunity — there was about $30 million available — to create an enterprise architecture approach that aligned with the department director’s vision of trying to take this very large organization of health and human services, this siloed and disjointed organization, and make it into an enterprise that feels like something that a customer can get into and not have to navigate a complex set of programs. That really dovetailed with what I wanted to do: create an enterprise architecture which allowed these systems to actually live within a community of systems that had their own self-awareness, the ability to share data and share functionality through a service-oriented architecture, with layers of functionality that can be shared between systems — make DPHHS really look and act like an enterprise, in other words, where there was an awareness between the programs, their staff, their function and the common population that they all serve.

I wasn’t going to just take the appropriated money to replace a system. I was going to replace systems in the context of an architecture that looked like I was an urban planner instead of just a building architect. I wanted to build buildings that are actually well-placed, work together and share common infrastructures and also have the ability to be built cost-efficiently and serve for the next two to three decades.

What went live last year was the SNAP and TANF eligibility components. When those other two systems went live in November, that completed the formal replacement of the old, combined eligibility system that ran on the mainframe. The company that built Medicaid eligibility was Northrup Grumman and the company that built the other two components for eligibility in SNAP and TANF, and put in place the enterprise architecture with some service layers, enterprise service bus, rules engine and all of that, was Deloitte Consulting.

Q: Were you following some examples of other states in crafting this enterprise architecture approach, or was it something you had been doing in the private sector?

A: I would say that in the area of health and human services, Montana was a leader in this area, and I wasn’t necessarily following anybody else’s lead and it wasn’t just my idea, although it was something I brainstormed from a technical perspective. I brought in some skills I had garnered in the area of building more open systems that use web-based services. Enterprise architecture really became a significant topic of discussion and even became a discipline. What I was able to use were enterprise architecture standards and templates that were starting to be created out there about four or five years ago — for example, MITA, that Medicaid Information Technology Architecture, which is a requirement that CMS has now. You cannot build a new MMIS system without a having a MITA-compliant plan in place.

What was really driving me was coming from industry trends in using service-oriented architectures and open systems, building systems that share functionality and data, and the trends coming out of federal agencies and national standards.

Q: So the new eligibility system went live last November, and the contract for the new MMIS — $57 million, over 35 months, being performed by Xerox’s ACS Healthcare — is set to go live in 2015. There are some horror stories in the press about MMIS roll outs gone awry. Some of those may be exaggerations, but there clearly have been problems. Since you planned one recently, how did you design it to avoid those issues?

A: Basically, I have two comments about the press on MMIS systems. Most of them, in my opinion, are probably not an exaggeration. MMIS systems are notoriously complicated and they take a long time to build and there’s various factors, some of which have been project management problems, some of which have probably been staff, resource and technical problems. But MMIS systems are some of the most notorious systems out there, because they’re the most costly, complex systems you can build and there’s a lot they’re responsible for in paying out — sometimes billions of dollars in claims.

Knowing all of that and paying attention to other states — cost overruns and time-overruns, or in the case of the state of Maine, where the project just failed outright — we did two things. We worked extremely diligently on the RFP, making sure it was extremely detailed in regards to the requirements, making sure all of our expectations would be met, both from a functional perspective and from a technical architectural perspective. The other things to make it successful was the contract. We made sure that the contract had enough strength and provisions in it that it would allow Montana to be successful in getting what it paid for. Our contract is structured so that it has discrete deliverables that are developed within an agile project management style, and those deliverables are actual demonstrable pieces of code or modules that will actually do things, and when those modules are delivered then payment is made. The contract payment schedule is heavily backloaded, to make sure that Montana pays only for things that it does get when it gets them, according to the payment schedule. There are also built-in penalties for missing the dates.

When a vendor responds, they are responding to a very discrete, detail-specific set of requirements that they can sign up to and put together a viable plan and pricing for. Those requirements flow directly into a management repository and will be tracked and traced throughout the project, and against the contract. The requirements fall into two major areas. One was technical architecture and the other one was functional components — those were based on modules, like claims. There were over 3,000 discrete requirements that the vendor had to respond to.

Q: Now that you’re the CIO of the whole state of Montana, what are some of your goals for IT in state government?

A: Here I come in in January, appointed to state CIO by Governor Bullock. I really wanted to continue the things that I was doing at HHS, but do it on a broader scale. Now I have a bigger canvas to work with. Now we’re talking about all the state agencies and trying to create a set of services and an infrastructure that everybody wants to participate in, share and make use of. I had to shift my thinking at a technical level from systems thinking to more infrastructure thinking — but still following those basic principles of, ‘Let’s do this in a way that supports the enterprise of the state of Montana. Let’s not have agencies by agencies build their own system that does similar things. Let’s promote the shared use of systems and infrastructure.’ There’s a lot of gravity behind that right now.

One particular focus area is electronic content management. That’s going to be major focus of mine in the next couple of years. I’m putting together a plan and a platform for centralized enterprise electronic management and enterprise content management for the state of Montana. Another major area we’re looking at, and looking at carefully, is cloud computing. And a third major area — third, last, but certainly not least — is data protection. We were appropriated money is the last session to continue our efforts in enterprise-level data protection, which involves identity management services at the enterprise level and additional protections that help with our general security infrastructure here in the state.

I want to make sure that agencies can meet their individual objectives, goals and missions for how they deliver services to the state. But now I’m over here in this role, looking at all of the opportunities where we can make all of those agencies be as efficient as possible.