Q&A: Geisinger CIO Frank Richards
Frank Richards has helmed Geisinger Health System’s technology operations for 15 years – and has been an employee of the care delivery pioneer for nearly four decades.
While Geisinger’s $4 billion doesn’t make it a revenue giant among healthcare systems in the U.S., this vertically integrated healthcare organization is increasingly thought of as a go-to model. A thriving regional system, fast-growing Geisinger combines operational excellence with innovation.
Healthcare IT News recently spoke with Richards to learn how the long-tenured CIO sees the current and future state of healthcare technology. Here are six things to know about him and his work.
1. Thirty-eight years at Geisinger give him deep institutional knowledge. "I came to Geisinger in 1977 out of a graduate program at Ohio State with a degree in chemistry for an internship. At the end of that internship, they offered me a job. I worked then in the laboratory in the toxicology department. In 1981, they wanted somebody who had a computer background, which I had from college. That’s how I made the switch to IT. For a long time I was the sole clinical computing person, with a small operations staff. I did all the computer work for pharmacy, laboratory, pathology and the other clinical departments. I took over as CIO in 2000."
2. He's helped make Geisinger a technology pioneer. "The original Geisinger medical center was patterned after the Mayo Clinic, a multidisciplinary teams of specialists with a single-unit medical record. As we expanded in the ’80s and into the ’90s, the model of single-unit record could not work with paper. You can’t operate within the continuum in a paper world. Through the 1980s to the early 1990s, we were still using paper-based medical records. We tried many different kinds of things: faxing and document repositories and all kinds of things. We signed with Epic in 1995. We learned a lot of things the first two to three years, and the software got better. Now, I can’t imagine operating without it. The next iteration of electronic heath records needs to be smarter, culling through information, presenting providers information they need to act upon. I see the evolution from where nobody could find anything because it was on paper to our situation where we have so much information and need smarter systems presenting the users with actionable data."
3. The health system is heading toward real-time analytics. "We have had a sophisticated data warehouse for 10 years. This warehouse now services about 2,000 users. We look for patients who may not be receiving the care that they should be based on lab studies or other things that fall through the cracks across our multiple venues of care. We provide real-time alerting dashboards for inpatient critical care areas based on constant monitoring of information. We continue to expand our ability to analyze real-time data. We started with the emergency ICU. Now we go far beyond the ICU to some of our regular high-acuity beds. We’re going to need systems that can more effectively monitor all our clinical data in real time, getting information to the correct provider to take action. You need massive amounts of compute power to collect data and do something with it. That’s been one of our hurdles. A lot these systems generate tons of data from patients, but it has to go somewhere and be analyzed. There’s big promise in big data analytics, and we are just getting out there with Hadoop and other tools."
4. Integrating newly-merged business entities can take work. "To date, we’ve been able to convert all acquisitions to our entire suite of products, not just Epic, but lab, radiology, pathology and so on. We’ve been able to convert every one of our acquisitions to the Geisinger medical record scheme. This takes a certain amount of effort because we have to do a conversion of their existing medical records to our scheme. In some organizations, they haven’t standardized their medical records as much as we have; they live with that challenge even within their own systems."
5. Patient identity challenges continue to be an issue – but national ID isn't the answer. "There’s no national identifier and not even any set national parameters of data elements to identify someone. It continues to be a challenge for our health information exchange, which services 3,800 organizations in western Pennsylvania. As much as I would like it, a single national identifier would not make everybody’s life easier. I don’t even think it’s workable. My concern about having a standard national identifier is that it would get watered down and abused. What’s more interesting to me is creating standards for the common data elements we all agree on. These identifiers could create a positive enough ID so that I can at least return a list of possible people you might be looking for. Today, if I make a query to another health system and I don’t have enough information about the patient for them to return one and only one possible person that matches, the standard says don’t send any. That’s just not workable without some kind of positive ID."
6. Leadership means delegating. "The key is to hire people that you trust who can operate independently. You can’t be involved in everything. You can’t know everything that’s going on. The health plan’s got operations in several different states, and we’re talking of acquiring a system in New Jersey. So I really need to rely on the team and they need to rely on people on the front lines to know what the job is and to deliver. We have a lot of tenured people in IT who have been doing the EHR implementations. We have about 20 years of expertise here. People have to feel comfortable coming to you and telling you that there’s this problem or they’re running into problems and not wait until the last minute when things blow up. My staff is very good at giving me early warnings. Sometimes they turn out not to be anything, but that’s OK."