State chief innovation officer says CitiStat and StateStat management programs helped pave the way.
Michael Powell has been Maryland’s chief innovation officer since August 2012. He joined state government after working as a consultant for IBM and spending three years as an analyst in Baltimore’s CitiStat Office, which helped pioneer the broad use of performance metrics for managing public agencies and programs. We talked to Powell at the NASCIO Midyear Conference earlier this year about the state’s strategy for using data and analytics.
I have three: Big data, open data and creating what the governor calls common data platforms. We’re seeing a convergence of all three.
When we talk about data — whether it’s crime, health or employment data — we want to have one place to go. It doesn’t matter if you’re the State Police, the Maryland Emergency Management Agency or a local fire department, you should be working off the same sheet of music. So we are building platforms in these areas.
One of the great ones is for a health information exchange. We’ve created a consortium of all 47 acute care hospitals in the state. Any admission, transfer or discharge is in the system, so if you show up at Mercy Medical Center one day and at Johns Hopkins a week later, Johns Hopkins knows your history of interactions. Can we start sharing diagnosis data through that system? Absolutely. Can we start integrated cost data, so we understand that better? We can do that too.
In local government you have data from your 311 system and a ton of crime data because the police department does that. The state is a little more limited. To a large extent we regulate and fund things, so the nature of the data is different.
But there are big opportunities. For example, we have data about every person who’s ever been unemployed in Maryland, largely because the state processes and pays claims for unemployment insurance. We’re starting to realize what an asset we might have in this data that gives us a picture of unemployment across the state. Analytics is where we want to go. We’re starting to look at this data as something that might be really valuable.
Yes, we launched an open data portal fairly easily because folks were used to collecting performance data about their agency, and we were able to put that data in the portal. Now the limitation is that we really think about things that way. We think about data that’s already been aggregated to a certain extent rather than really raw data and individual-level data.
We’re good at saying “County X” had 47 property crimes during a two-week period. But what I want to make available is the list of the actual crimes.
I think our solution is to aggregate the data to a certain level and report things that way. Instead of a list of every hospitalization over the past 12 months, we’ll count it up by census block group. We’re starting to work through those issues. I wish I could tell you I’ve solved them.