‘Putting data into action at Vermont Department of Health’
The text included in this article is a transcript from our recent webcast interview with Rich McCoy and Julian Tyndale Biscoe of InstantAtlas. You can also hear this interview by selecting the soundcloud box on the right hand column.
Hello, my name is Julian Tyndale-Biscoe and I’d like to welcome you to the third in our US public health webinar series. Today we are joined by Rich McCoy, Director, Center for Health Statistics at Vermont Department of Health. He is going to tell us about the project he has been working on. We wondered first off Rich if you could tell us why you think it is important to bring data into one place so it can be used as a central resource?
So until very recently our approach, our organizational model, was to leave it up to the staff to maintain their data sets individually and choose which tools they would use, which software packages they would use, for both the analysis and the reporting of these data. Over time that led to a decentralized model where we had many different software packages. The flavors and formats were quite different and it was becoming quite confusing, one for IT staff to support so many different tools and two, it made it difficult for staff to back each other up if someone was out and only they knew that one package. That made things quite difficult. Also, intertwining those data, being able to look at those data sets in combination was very, very difficult when you had to try to import it from different programs.
We made a strategic decision a couple of years ago that we would move towards one tool to combine and release these data for both internal and external partners. We would choose a package that would allow us to bring in various different data sets, that we would review and prepare and we would come up with one format, one presentation format, and make a tool that would allow the public, our partner organizations and our own staff to do queries, create graphs and charts, create maps and do reporting based on their interests, what they were looking for.
This had many advantages when we started looking at this model of centralizing to one depository, one tool. The advantages being, one, we would have all the staff trained on the one package rather than having everybody trained on four or five packages or only one or two individuals knowing each package. In this case we trained up a core group and then they went out and trained other staff. That has worked very, very well. Now we have a deep bench in case people are out sick or on leave, we have a number of people who can step in and work with the software, work with the various data sets.
The other advantage is our IT group very much appreciates that they have to support one package, one software, instead of three, four, five various types. It's a lot less work for them so therefore it saves time and money for them. And for our staff as well. The IT group has really been supportive of this model where we centralized the data, the data release, the data reporting.
I would say there's a third benefit to our model which is it really helps us standardize what we do. Instead of having procedures for different data sets and different tools we really have been able to bring our group together and agree upon standards that cut across. For example, we've been able to come to agreement on cell suppression standards for all these data sets for presentation through our one centralized tool, versus before we had different standards for cell suppression, for say with the cancer registry data and the cell suppression for vital records mortality data was different from the cancer registry and that was different from the behavioral risk factor surveillance survey data.
So by centralizing and going to one tool and combining and finding a format that will be consistent. It also forced us to review our policies and procedures and implement standards that all the staff are utilizing with the data. This makes it much easier for us to ensure that there's no errors or release of data that would concern us. It also makes it easier for us to communicate the limitations of these data to the public and to our partners. I think over the last two years as we've moved to this model it has definitely saved us incredible amounts of staff time and it's truly saved us a lot of money because we have dropped some of these support contracts for other software packages. We're very pleased with our progress on this front.
It is interesting to hear that the IT group has been supportive and I’m sure that this may not be the case in all departments. How important was it for you to have this support?
It's extremely important. I think the relationship between analytical staff, program management staff, epidemiologists, having the support of the IT staff is I think critical because they are the ones who will support and troubleshoot any problems with the software. They also are the holders of the databases, of the DBA, the database administrators. Very, very important person to have involved in these steps.
In our department we have a very, very small IT group at the department level. Most of our IT group has been moved above us to the Agency of Human Services so we do have to work with both groups when we are installing or updating software. For us it means having a good working relationship with two different IT groups.
I would say though, we are fortunate and maybe we are unusual in that we have IT staff that are as concerned about the accuracy, the protection of these data as we are. And they also are as committed as we are to releasing these data so they are useful for public health planning.. Not just for the health department but also for our partners outside in the field. I find it truly wonderful that they are always willing to come to the table when we are looking at making changes with these data or looking at changes in how we're going to present the data. I'm really, really happy that they often come to us with recommendations about what we could do more efficiently with the software, what we could do more efficiently with our reports and queries. It truly is a partnership and I'm not sure how common that is across other departments.
I feel very fortunate that this is a group that sees it as a collaboration. They've always said to me, 'Don't leave me out of the loop. We want to be part of the conversation as you moved ahead with these type of .' Particularly the database administrator I find incredibly insightful and very, very helpful. So I see part of my job as fostering those relationships with the IT group and bringing them together with my staff. It really is important for me to make sure that they have a healthy working relationship.
One of the challenges, of course, is our two worlds speak two different languages. Quite often our world, my staff, may speak about data and use certain terms that mean something different in the IT world. Again, that's why it's important to have a good collaboration with IT to make sure you're not talking past each other, or misinterpreting what you're asking for, especially when those terms may mean something a little bit different in your two worlds. I'm very, very pleased with our IT relationship.
I'm not sure how I can suggest that all departments move to having such a collaboration. I think it just takes lots of time. I think requires lots of face time. I encourage my staff to walk away from email and actually go and see the IT staff in person. I think that's really valuable. Email, I think, tends to be a barrier to creating collaboration because you really need to get to trust the other person and IT is one of those groups you really want to build trust with.
That all sounds very positive and it’s clear you have a good relationship with the IT group which is beneficial. Could we talk about how the data is being used now that it is all in one place? For example, whether the public and people outside the organisation recognise that the data is easier to find than it was before?
Sure. I think for public health it's always a slow process in terms of publicizing and marketing the good things that we do. We can't really advertise but we certainly try to promote the new things that we're doing. So in this regard it's been a little slow but what we've been doing is utilizing our local health offices. We have district offices that cover specific sections of our state and so we're working through those district offices who have the relationships with the communities, who have the relationships with the decision makers and influencers to show them here's the new resource that we have made available to you on the web that allows anybody to access these data and create maps, tables and charts utilizing all these different data sources.
Part of it is just training and education of those district offices, so the staff understand what's behind it, how it works, what the limitations are. Then going to meetings with the district office and their community partners. There's a part here where it's a bit of a demonstration show going around, whenever we can, to demonstrate this. We have put out press releases but I'm not sure how much attention a press release gets. Every time we release results from a new report we always mention these results are also available through our web-based tool, to try to drive traffic to it.
I would certainly say we have seen a decrease in the number of data requests that our staff receives as a result of making much more of this information publicly available. So they can create their own tables and maps. It does decrease the amount of time we spend. Now the counter-argument is sometimes it leads to new types of requests, new types of questions. They may utilize our data through the tool we make available but it may lead to new questions. So it balances out a little bit but right now I would still say it saves us a little time on not having to answer the same type of questions over and over and over. For example, mortality data. There are certain questions we get about mortality data, it's the same question every single time. Here's a way to dramatically reduce that.
The feedback that we've had so far has been very positive. The district offices love it. The community partners that have started using and accessing the data have been very, very pleased. Their primary use so far has been for obtaining data for grant applications. Many of our communities apply for a variety of federal, state non-profit grants. For example, some communities apply for grants regarding healthy spaces. Parks, physical activity, things of that nature. Well, with our tool they're able to go to a central place, obtain data about their community that relates to obesity, physical activity, access to fruits and vegetables, things of that nature and then take the data and apply it into their grant application.
In the past they always had to call our office so we would get lots and lots of these data requests from communities to get them these data. Now we really don't have to do much of that anymore. They're able to create it themselves and insert it into their grant applications. That has been a wonderful thing for both us and our community partners.
Also internally, above our level at the agency level and at the governor's level there's a strong, strong interest in dashboards. Dashboards go by a variety of names. 'community indicators', 'dashboards', 'buckets', all sorts of terms. At the heart of it is a tool that shows you a list of priority areas that you want to be able to quickly see what is the progress, how are we doing? For a community indicator it may be, how are we doing towards driving our community towards physical activity? For our agency it might be a dashboard that shows how are we doing towards decreasing the number of emergency room visits for asthma.
Again, having these data set up as we are, using the tool that we are, we're able to help them create these dashboards, these community indicators. Again, it reduces the workload on our staff and it makes the data much more quickly available to both these internal and external partners.
I would say we still have a ways to go to get the message out there. I think just the normal person, the normal resident living in a town is completely unaware of these tools. So I think we're doing well in terms of getting the news out to the organizations and the community groups but the general populace, I don't think is aware of them yet. It's a struggle. In public health you don't have a lot of resources for the promotion part and so that's why we look for every opportunity to talk about it as much as we can.
I wondered, since you travel around and talk to other directors, do you find they are in a similar position, or do you think you are leading the field?
I do work very closely with many of my counterparts in other states and jurisdictions and we talk about this topic quite a bit. I would say that here in Vermont, I wouldn't say that we're at the front of the pack. I would say we're in the top 10 or so, top 10% or 20% in terms of the steps we've been taking. I think there are some states and jurisdictions that are still doing their same old process which is static charts and tables that are published in an annual report and publish the report on the web and that's it. I think there are still lots of places like that. But there are other states and jurisdictions that are doing things like we're doing and in some cases are ahead of us.
For example, we have not yet incorporated our hospital discharge and emergency department data into this model. Our plan is to, but we have not yet done it. I know there are certain states that have gone ahead and done that. I would say we're in the front but not quite as far as maybe a couple of other states.
Sometimes that's a good place to be because you can look around at what people are doing and borrow.
It is. There have been times when Vermont has volunteered to be the first to pilot test something and it's always a huge amount of work pilot testing something, working out the bugs. Sometimes it's not so bad being the second or the third state to try something.
You’ve given us a great flavour of what you have been doing, but could also give us an insight into your plans for the future? Where would you like the project to be in a year’s time?
We have a couple goals for the end of this year. One is to add additional years of historical data especially from the vital records world, so mortality data, natality data. I really would like our staff to go back in time and add several more years. That definitely is one of our goals. My preference is to have at least ten years of data for any data source available. I'd prefer to go back further if I can and in some cases we might go back to the 1990s. I think this is important. I think trend analysis is always important but Vermont is a small state, we have a small population so adding these additional years will make these data and the tool that much more important and more reliable. That really is one of our primary goals, is to add additional historical years.
In terms of other data sets, yes, we definitely have been talking about adding much more not just youth risk behavior survey, not just the adult behavioral risk factor survey but we're also talking about adding our adult tobacco survey data. We do that survey every two years. We're also talking about can we add immunization data. We have an immunization registry. Can we add immunization data? We've already done that to a certain extent with our cancer registry data but we have not yet done that immunization data. However, I think it would be very valuable if we did.
We do have some lofty goals but I think these are good goals for us to have and I think some of it is based on the feedback that we have received from our partners. There's a real interest in having many more years of the vital records data and certainly we've had a lot of feedback that they want access to the immunization data.
Rich, thanks for your time today, it’s been very interesting to talk to you and I’m sure we’ll catch up with you again soon. If listeners would like to follow up on any aspect of the project Rich would be more than happy to answer them, so please forward your requests to firstname.lastname@example.org
Thanks again Rich
Thank you very much