State of Michigan profile
Long considered a national leader in its use of information technology to improve operations, the State of Michigan utilizes one of the nation’s most sophisticated Business Intelligence (BI) data warehousing solutions – developed by Bull Services – to manage most of its critical functions in Health and Human Services, as well as supporting activities in the Department of Treasury, the Courts, Corrections, and Unemployment. With more than 9,000 state government users, the BI data warehouse has become the critical backbone to support the informational and knowledge needs across Michigan state government, improve the efficiency and effectiveness of programs, deliver services to citizens, and help ensure that Michigan taxpayers are getting the most value for their money. Michigan’s BI solution, a model for other states to emulate, has received national recognition from some of the most prestigious industry analysts and organizations in the United States, including: the National Governors Association (NGA), the National Association of State CIOs (NASCIO), the The Data Warehousing Institute (TDWI).
Why did the State of Michigan launch a Business Intelligence (BI) project and how has it evolved?
Michigan first invested in a BI solution (data warehouse) from Bull Services in 1994 – during the early stages of BI technology – to help manage its Medicaid program (which provides health care services to needy citizens). As the Medicaid rules and regulations became more complex, the state implemented the data warehouse to monitor Medicaid claims for quality of care, as well as for overpayment, fraud and abuse. The solution has evolved with Michigan’s needs over the past decade.
Quickly, the BI data warehouse provided the Michigan Department of Community Health (DCH) with the ability to conduct in-depth analysis on the Medicaid program, reducing the time it took staff analysts to get answers from one-to-three weeks to 10-minutes to one hour. Workers were able to directly access mainframe-sourced data and bring it into spreadsheets on their PCs for rapid and comprehensive analysis.
One year after its implementation, the full capabilities of the BI solution were quickly put to the test, when, in 1995, Michigan was studied and ranked nationally by the U.S. Centers for Disease Control and Prevention (CDC) for its performance in childhood immunization rates. Michigan’s last-place finish was considered by state officials to be a powerful wake-up call, and they called for a united effort to address this critical issue. As one part of a comprehensive initiative to improve childhood immunization rates, officials explored the potential for the warehouse to provide important information on the subject. The warehouse was not only able to tell DCH the status of childhood immunization on a county-by-county basis, it also rapidly provided the data necessary for notifying parents and providers across the state about the need for timely immunization schedules. The BI data warehouse was one of many factors that helped Michigan exceed the national average in childhood immunization rates within two years and finish first in the country by 2000.
Since then, Michigan’s BI solution has evolved even more to better provide quality health care services, helping the state track standards for providers, manage contracts, analyze and measure the effectiveness of programs, and measure the performance of health plans. Today, as part of the “enterprise phase” of the BI data warehouse solution, and using the data warehouse as the backbone, Michigan, working with Bull Services, has integrated 21 separate health-related agency data sources into a single integrated environment. Many of DCH’s 1.4 million clients are enrolled in multiple programs supported by the department. However, prior to this integration project, it was virtually impossible for the agency to track and monitor services and costs associated with a single client through these separate agencies. Now, the BI data warehouse is helping DCH track fraud and abuse, conduct advanced data analysis, interpret patterns, determine which programs are most cost effective, reduce overall costs to taxpayers, and predict the state’s health care needs and priorities in the years to come.
The BI data warehouse now encompasses far more than health care data. It is on it’s way to becoming a true statewide enterprise solution – the best example in the country – that contains data from the Department of Human Services (providing cash and food assistance for needy citizens), the Treasury Department, the Michigan courts, Corrections, and Unemployment. More than 9,000 Michigan employees now utilize the BI data warehouse to perform their jobs. No other state shares data and information so seamlessly to deliver better services to citizens and reduce the costs borne by taxpayers.
What kind of strategy did the State put in place to meet its objectives?
Michigan realized several important factors early in its development of a BI solution. For example, the State had the vision that the greatest benefits – both financial and programmatic – would accrue as the system added more data and users. Thus, it was imperative that it first implemented an open, flexible, non-proprietary architecture that provided the best possible data quality; enabled the capability to add disparate data rapidly; had the capability to introduce the latest, best-of-breed, industry-leading tools without disruption; and was relatively easy to use. The State geared its implementation (and subsequent expansion) of its BI solution toward end users, recognizing that for BI to be truly effective, it needs to move beyond the realm of the IT department and provide answers and solutions to “business” problems. The Bull Services BI solution provided the flexibility, openness, high level of data quality, and ease-of-use to meet Michigan’s needs.
Next, Michigan recognized the importance of safeguarding sensitive data even as data is shared for the State’s benefit. Security is a paramount concern and the State has built in the necessary protections at many levels. For example, while data from several agencies resides on the BI data warehouse, only data that is intended to be linked together is shared. And, while some data sets allow access outside of the controlling agency, the data is controlled at the model, table, field, and even range levels. The data can be shared for a single analytic purpose, and access can be ended when the analysis is complete. The originating agency determines access to its data and the security, and the system tracks who accessed what; therefore data is protected via access security at the front end, and ability to monitor data access by individual user at the back end.
In addition, Michigan understood the importance of selecting and growing with a vendor that provided more than technology – one which could provide subject matter expertise, consulting services, mentoring, and a level of comfort working with state employees and sensitive state data. Bull Services’ lengthy state-and-local government experience, plus its Business Intelligence approach and expertise, provided Michigan with this assurance. For example, in the area of communications, Bull Services has worked with the agencies to provide information on the BI solution to state executives, develop “annual reports” to circulate across departments on the use of the system, and assisted the State in preparing the documentation for the various industry awards it has won for the BI data warehouse.
The State also recognized the need to create a common identifier across disparate data sets so that, as new data sources are added, they can be easily linked to existing data to provide user access to all data related to a specific individual. This concept will be critical in identifying data relationships as the State continues to build its enterprise data warehouse.
What are “best practices” and “things to avoid” when implementing a major BI project?
Many of the “best practices” were explained in the previous section. It is important to stress that, from initial discussions and project planning, to implementation, to growth and expansion, the State and its BI vendor must maintain a deep level of openness, communication, and trust. These are complex and far-reaching projects, and a vendor is entrusted with some of the State’s most sensitive data; therefore an honest working relationship is critical. This is much different than the client-vendor relationship with a product vendor – the client-BI vendor must develop more of a partnership. Bull Services and the State of Michigan have developed such a partnership of success.
The State of Michigan and Bull did overcome some obstacles to achieve this success. The initial collection and cleansing of data brought with it a series of challenges, though this got easier as each new data set was added to the BI data warehouse. In addition, Bull Services and the State worked together to create a separate BI environment to supplement and complement the transactional environment (the system in place to process transactions, issue checks to recipients, etc.). BI is a querying environment built to conduct advanced analytics to transform data into knowledge; its entire structure and mission are different than a traditional transactional environment.
What kind of “business benefits” did you derive (and are you deriving)?
Michigan has achieved enormous financial and programmatic benefits from its BI solution, one of the most sophisticated in the United States. Industry analysts, publications, and other states have cited Michigan as a model to emulate in the BI space. In fact, the Michigan BI data warehouse, one of the first of its kind in the U.S., has served as a model for similar Bull Services’ implementations in Illinois, Minnesota, New York, and Utah.
As mentioned above, the data warehouse helps track and uncover fraud and abuse, conduct advanced data analysis, and interpret patterns. In addition to the “last-to-first” jump in child immunizations already mentioned, some of the business benefits Michigan has achieved include the following:
Estimated annual savings (MDCH) of nearly $200 million (advanced health care analysis, federal bonuses, avoiding federal sanctions, etc.), a figure that should continue to grow.
Medicaid administrative costs have been reduced by 25%
DCH has doubled its identification of fraudulent Medicaid activity
The State is better able to evaluate health care plans, payment rates, and policies, thus maximizing Medicaid program savings and sustaining quality care
The BI data warehouse enables the State to track and substantiate its Medicaid pharmaceutical costs
The solution helps identify coverage overlaps to reduce the number of instances in which individuals are obtaining similar or duplicate services across multiple programs.
Michigan’s BI solution helps eliminate inappropriate payments; for example, comparing eligibility with lists of deceased residents is an important fraud issue.
The BI data warehouse is used to analyze food stamp and cash benefits cases, tracking nearly $1 billion annually in taxpayer dollars. The Department of Human Services in Michigan uses the BI solution to validate and independently cross-check the Electronic Benefits Transfer (EBT) processing vendor that actually manages individual recipient and retailer accounts
Social service workers who visit clients are able to access the data warehouse remotely for the most up-to-date information, and then prepare their reports quickly and efficiently using pre-structured templates that help save time and resources.
In the Department of Treasury Bureau of Revenue, the BI solution has helped the state achieve a dramatic increase in number, fairness, and efficiency of audit reviews. With its old “paper system,” the Bureau was limited to 6,000 reviews per year for audit consideration. It is now able to review all 450,000 business taxpayers annually with the same staffing levels, helping it to select the best candidates to audit.
For these and other achievements, the Michigan BI solution has been recognized by the National Governors Association (NGA), the National Association of State CIOs (NASCIO), The Data Warehousing Institute (TDWI), and several national trade publications.
How will your BI project/application evolve in the future?
In Michigan, this evolution has already begun – at its core is the willingness and desire of more agencies and departments to add their data to the BI solution to improve their efficiencies and deliver better service.
The most recent major addition has been the Michigan Judicial Data Warehouse project sponsored by the Michigan State Court Administrative Office (SCAO), one of the most ambitious in the nation. Bull Services, ultimately will combine data from 251 different courts – circuit, district, probate, and family – loading criminal, civil, traffic, and juvenile cases onto the BI data warehouse. The project will standardize data across Michigan’s court system, allowing judges, clerks, and administrators to track an activity an individual has had within any court in Michigan – essentially “one version of the truth” – through a Statewide Name Search capability.
Once completed, judges, clerks, and administrators will have access to standardized, linked data with strong privacy and security safeguards. The court will have the ability to track individuals through the court systems, court operations are expected to be faster and more cost-effective, and courts will eventually have the ability to predict – not just report – the impact of changes to laws and policies.
Also under way is the aforementioned first-in-the-nation Child Day Care and Food Stamp/Cash Assistance fraud-fighting program undertaken by the Department of Human Services. Investigators have already identified millions of dollars in savings by linking cash assistance, child day care, wages, and unemployment data into the BI data warehouse for extensive analysis and comparisons.
These are just two examples of the widespread capabilities of the Bull Services BI solution for the State of Michigan. The State continues to put forth its innovative ideas, suggestions, and programs – and the Bull Services solution continues to have the flexibility and robustness to handle them all