JESSICA

Hi. I’m Jessica Hill and I’d like to welcome you to “Inform Me, Informatics.” In today’s episode I speak with Dr. Gulzar Shah who is the chair of health policy and management at the Jiann-Ping Hsu College of Public Health at Georgia Southern University. Dr. Shah has worked in the field of public health informatics for over 20 years as both a practitioner and an academic researcher and I was very excited to speak with him about his work. One note about the audio quality. It’s summer break here in Georgia so we had to record our conversation over Skype. Just a heads up, the audio quality might not be the same as when two people are in the room together. That said, I was really excited to speak with Dr. Shah. He began by explaining a little bit about his work at Georgia Southern University.

GULZAR

Today is my first day as the chair of health policy and management department at the Jiann-Pink Hsu College of Public Health at Georgia Southern University. Prior to this, like, till yesterday, I was serving as the associate lead for research. I’m also serving as associate professor of health policy and management. As the associate lead for research, I was charged with promoting excellence in public health research that expands the existing evidence base for the purpose of public health and healthcare which in turn has the potential to improve health in communities and eliminate health disparities. In this capacity I have performed several activities in building research training, capacity building for faculty and students and as a faculty member, I write grants, publish peer reviewed articles and teach courses.

JESSICA

Congratulations on the new position. I didn’t realize today would be your first day so that’s exciting for us to be a part of this new transition for you. Thank you for speaking with us.

GULZAR

Of course.

JESSICA

So how did you first become interested in the field of public health informatics?

GULZAR

I became interested in the field of public health informatics…well, I think you will know that that is actually called public health informatics in mid 1990s when this field was getting some traction as the computer science and information technology were experiencing boom. I was working at the Utah State Department of Health which was among the leaders in state health agencies for health informatics including data innovations through probabilistic record linking, web-based data and interactive data query systems, smaller area analysis and geographic information systems were among the components of informatics used for various purposes including public health services.

JESSICA

So, you said that, you know, in the mid-’90s, all these fields were coming together such as computer science and information technology, plus healthcare. But then how did you enter that space?

GULZAR

Sure. That’s a good question. So, with the boom in computer science and related jobs, everyone was learning languages and software and so my training was, at master’s level, in statistics and at PhD level in sociology, demography. So immediately I learned some languages and my statistics background helped. Basically, it was more towards, you know, anticipation that computer science and informatics are gonna be important in the future. And so, I developed some skills on the job. And I also learned the probabilistic record linkage. That was also something that was…you know, training was available and it was…an opportunity came to me and I stepped up to the plate and it really, really helped.

JESSICA

Oh, wow. So, I wanna ask you about two things. Actually, to understand better what probabilistic record linkage is and also what web-based systems are.

GULZAR

Let’s start with probabilistic record linkage. So, you know, information systems and databases are maintained by many agencies in various formats. So, for instance, let me give you a quick scenario. A roadside accident happens, and information is collected on the victim and other people involved in the accident by at least 10 different agencies, early responders, if people have to go to inpatient or emergency so emergency department, inpatient department sometimes. They become part of the mortality data. And then the insurance. Every one of those information systems collect information in different formats. And particularly if there are no unique identifiers as far as the regulations, what you do is you take some other very, very semi-unique identifiers such as first name, last name, ZIP code and all, and then you apply an algorithm to compare records and link them. And probabilistic record linkage is a field that has evolved and now it’s a multimillion-dollar enterprise. But at that time, it was very cheap. So, we learned it and we applied it and we linked, you know, for example, mortality data with hospitalization data with Medicaid eligibility files and were able to research and test new hypotheses there otherwise are not possible.

JESSICA

I really learned a lot from that so thank you. But moving to the web-based query systems, how did they constitute health informatics and why did a state health department invest in it given all their other competing priorities?

GULZAR

So, web-based data query systems are…let’s call them WBQSs from now on.

JESSICA

Okay.

GULZAR

The WBQSs constitute health informatics in that a WBQS uses computer science, data analytics and health information systems to support public health practice which is a kind of generic definition of health informatics. We can improve efficiency of public health surveillance and evidence-based decision-making. So, for instance, a web-based data query system also sometimes called interactive query system is internet-based application that supports the dynamic query or on the fly query of databases by the users without having direct access to raw data and without having to have statistical samples where you analyze the data. To give you a concrete example, state health department in Utah and many other states collected information of all hospitalizations and there were many users requesting the raw data.

And so, replying to or responding to those queries took a lot of time because they were data use arguments involved and, you know, and the administrative parts involved. So by providing web-based data query systems most of those users fulfilled their information needs by just going online. For example, an insurance company trying to…or a hospital trying to figure out a market share they could actually go onto the internet and done some query systems and be able to get their answers without having access to the raw data. So, state health departments, local health departments use information systems for internal purposes for their own programs’ needs as well as external or the needs of external stakeholders such as those I mentioned.

JESSICA

And I wanna transition now to talking about your research around public health informatics at local health departments. And I know that you recently were the guest editor of a special issue of the Journal of Public Health Management and Practice and that that special issue really focused on what informatics looks like at local health departments. If you had to describe informatics capacity in three words, what it looks like at local health departments, what would they be?

GULZAR

Okay, so those three words would be challenges, variation and training-needs. Now you would say that training and needs are the two words but actually it’s…I would consider it one hyphenated word.

JESSICA

Okay, we’ll hear the hyphen. Our listeners are hearing the hyphen in training-needs.

GULZAR

But really it would be kinda unfair to try to summarize the findings for the entire issue into three words so…

JESSICA

That’s fair, that’s fair. And of course, here on “Inform Me, Informatics” we always want to be fair to our guests, but I think, you know, the idea of capacity, informatics capacity at local health departments all over the country just feels really overwhelming. So how about, like, three main takeaways?

GULZAR

Sure. Now that’s fair. So, the three main takeaways would be that informatics infrastructure and workforce capacities are still poor for many local health departments and they vary, the informatics capacities vary by size, location such as rural, and type of coordinates with respect to health department coordinates. In other words, health departments that are parts or the units of state health department. They have slightly higher capacities. And often compared to the local health departments that are independent or locally offered. And the rural and small local health departments have bigger challenges in health informatics capacity. So that’s the first takeaway.

The second one is that connectivity of health departments with the electronic health records is still very poor, of the 90% health departments that provide clinical services. In other words, 9 out of 10 health departments said that they provide clinical services. Roughly one third still use paper records. So, in addition, 60% of those health departments who provide clinical services, their nonclinical data are still stored in paper records. So that would tell you that there is a lot of room for improvement.

So, the third takeaway would be with respect to interoperability of information systems used by or maintained by local health departments. When asked, only 7.4% of health departments said that most or all of their systems were interoperable. Actually if I further dissect that 7.4% of health department, actually it was 1.4% of the health departments said that they have all systems that are interoperable and 6% said that they have most of their systems being interoperable. And the rest of the health departments which amounts to 93%, they said that only few or none of their information systems are interoperable.

JESSICA

So, it seems like interoperability itself and whether or not systems are interoperable is actually an indicator of informatics more generally. And I know in one study you really dug in on this idea. So, I was wondering if you could explain, like, why you chose to focus on interoperability and what does interoperability tell us about informatics at the local level?

GULZAR

A lot of times data are collected by different programs and a lot of times those records are about the same people. So, in order to make any meaningful chart out of multiple components of the health department or multiple programs of the health departments collecting information on the same clients, what it requires is to put together those separate pieces of data together. And if those information systems don’t talk to each other…in other words, if they cannot exchange information and information cannot be exchanged without any major intervention, human intervention, then there are duplications of effort. And so, one way to explain that is, the WIC program, chronic conditions program and Medicaid program, hospitalization data…they are all kept separately about the same clients for instance, and the demographic information which should be the same, is entered separately, for one, that causes a lot of complication. And secondly, still it’s difficult to put together those data sets without standardizing those unique or semi unique identifiers.

And not only that. I remember when I was working for the state health department, we had a major challenge with even calling out the data. So, in one information system initially we followed that gender was coded as M and F. In another one it was coded as 1 and 2. And so, all the small, small differences in entering information in an unstandardized way can create a lot of work for the health departments and therefore, you know, understanding how interoperable information systems are, was important.

JESSICA

So, I just wanna pause there for the moment just to really make sure I understand this point because I think it’s really important. So, you know, in public health we talk a lot about just the reality of funding siloes and how that sometimes results in siloes between programs. But as sort of a consequence of that, it might be that within a health department, their maternal child health program might keep a database but their coding could be different than the chronic disease. And so, if we wanna find out, you know, connections between women in care and chronic health disease indicators, even within the health department, those systems might not be able to seamlessly match. Is that right?

GULZAR

Yes. That is so true.

JESSICA

I really appreciate all the time that you’ve spent with us today and we’ve even been talking longer than we had anticipated but it’s been really interesting. And so, I hope you have a little more time for us to ask this final question because it’s the question that we ask all of our guests which is, how do you define informatics?

GULZAR

So that’s a good question to wrap up our interview. There are several definitions but my favorite one is systematic application of information analytics, computer science and technology to support the day-to-day work on public health including surveillance, reporting and health promotion. So, it is both our infrastructure and how people within the health department use information in their public health work.

JESSICA

Many thanks to Dr. Gulzar Shah for taking time out of his busy schedule to speak with us. I learned so much from our conversation. And listeners, if you’re interested in learning more about Dr. Shah’s research, check out the special issue of the Journal of Public Health Management and Practice from November, December 2016. Dr. Shah was a guest editor and an author of many of the papers about the current strengths and challenges in the field of public health informatics. We’ll have a link to it on our website. We’ll also have a link to the healthcare cost and utilization project which is also called HCUPnet. It’s a project of the agency for healthcare research and quality and Dr. Shah said it’s one of his favorite web-based data query systems.

This podcast is a project of the Public Health Informatics Institute and the Informatics Academy. Visit phii.org to learn more about all of our informatics work. You could also find us on Facebook or follow us on Twitter @PHInformatics.

This episode included songs composed by Kevin MacLeod. Thanks also go to our production team, especially our producer, Piper Hale who’s on a mission to make this podcast interoperable with the world. I’m Jessica Hill and you’ve been informed.

BUTTON

JESSICA

He said this WDQS. This is so hard to say.