Can Data Analytics Reduce Infant Mortality?
Indiana is launching an enterprise analytics initiative designed to improve performance across state executive branch agencies.
In late March, Indiana Gov. Mike Pence released an executive order creating a Management and Performance Hub (MPH), a coordinated effort among the state's agencies, the Indiana Office of Technology (IOT), and the state Office of Management and Budget. MPH will provide centralized data sharing, correlation and analysis for the state in areas where multiple agencies must work together.
The first target? Improve Indiana's infant mortality and child fatality rates through data analytics. To do this, state agencies will use historical data to make predictions and better decisions on unfolding events, according to state CIO Paul Baltzell.
For years Indiana's infant mortality rate, a measurement of infant deaths within the first year of life and an important public health indicator, has ranked near the bottom nationally. In 2011, Indiana had the sixth highest rate in the country, or 7.7 deaths per 1,000 live births, according to the State Department of Health.
"It was an easy target to pick because it's something everybody can get behind," said Baltzell. "Directly affecting childs' lives and outcomes in their lives--that's as good a cause as it gets."
But tackling infant mortality and child fatality rates are just the start. Indiana is working to automate agencies' key performance indicators (KPIs) and revamp the state's Web site to allow the public to search these indicators. The state will also work to improve outcomes on other individual targets in addition to infant mortality and child fatality rates, Baltzell said.
To assist agencies with performance-based management, the IOT is automating the process of regularly compiling and updating each agency's KPIs. This project is underway and is being rolled out in stages, with the IOT working with 22 cabinet-level agencies to automate their processes. When this stage is completed in August, the IOT will expand the project to the rest of the executive branch, according to Baltzell.
The project will also provide greater transparency with a revamped MPH public site. Once the site goes live in May, Baltzell said the public can view top-level KPI results on a larger scale and, using interactive graphs, they can manipulate that data to get the information they are seeking. Some examples of searchable information will include health and recidivism statistics and disaster relief response.
To enhance policy development and decision-making, the state is narrowing in on individual targets and using analytics across agencies to better understand the causes of problems and also get a clearer idea of possible solutions.
Baltzell said IOT is using detailed data from areas that may be contributing to or causing infant mortality and children fatality, like access to health care, demographics and family criminal history. By analyzing a wide variety of sources that span state and federal agencies -- anywhere from 50 to 60 databases -- the state can get a complete picture of the problem.
The goal is to use analytics to give users -- in this instance, caseworkers -- access to veteran knowledge. Better understanding complex situations based on years of experience is something the state hopes to tap into by building algorithms that share some of this analytical information with those who are less experienced.
The project, which began late last year, is large in scope, bringing together many data sets, according to Graig Lubsen, IOT's communications and marketing director.
"We're talking about tying in the entire executive branch's systems. Not just one or two agencies or three. We're talking about 90 plus agencies," Baltzell said.
Although Indiana is one of a few states to have a centralized data center, which lends itself to data analytics, it had only used analytics within individual agencies. MPH is the first analytics project at the enterprise level.
Building a system at this scale requires tracking and protecting the data, as well as identifying which data is needed, according to Baltzell. He also said the project has required business analysis for the targeted cases, data scrubbing to ensure data accuracy and bringing in scientists to correlate all the data into algorithms.
To aid in this complex process, the state has brought in a combination of technologies, including real-time analytics with SAP HANA, and the open source analytics tool Hadoop. While Hadoop will give users answers in a short period of time, saving on cost, SAP HANA will allow users access to real-time analytics.
HANA will help Indiana aggregate data and identify root causes to decrease Indiana's infant mortality rate, according to Dante Ricci, senior director of public services for SAP.
Going forward, MPH will also include a public-private advisory panel to ensure best practices in data analytics and security, according to the governor's executive order. Baltzell said the formation of such a panel reaffirms the size and importance of the project to the state.
Indiana recently received a $500,000 grant from the Lilly Endowment, Inc., for cutting-edge technology enhancements to the project, according to the governor's press release. However, the majority of project funds will come from IOT's regular budget.
As for answers to improving the state's infant mortality rate, Baltzell expects predictive analytics to give results by the summer to help its agencies make better decisions.
Baltzell said the state has chosen to tackle one subject area at a time to ensure success. But after using the tools to analyze infant mortality and child fatality rates, some possible contenders include offender recidivism and fraud.