Why Providence is Making the Switch to Cloud-Based Data Management

Tegria: What does Providence’s data-management model for EHR delivery look like today?

BJ: It’s a hybrid solution. Providence has three Epic instances in three different on-prem locations, all running exactly the same code base. We use three on-prem instances because there isn’t a single server that’s powerful enough to handle all the data.

But we’re also in the process of moving all of our primary and secondary instances to Microsoft Azure. One of our first steps toward that goal happened at Providence Alaska Medical Center in Anchorage. That hospital’s primary instance of Epic was implemented in Alaska. But we moved the hospital’s DR instance to an Azure data center in Washington state, and it’s performing well. If the primary in Alaska fails, the hospital will have a latency hit. So at least there’s geographic redundancy now. And we are working with Microsoft to create an Azure data center in Alaska, which will provide new opportunities for us.

Long term, we want to eventually get fully out of our on-prem data centers and make the switch to be entirely cloud-based.

Tegria: What is your rationale behind the transition to cloud-based data management?

BJ: We’re not going to “out data center” our competition. People don’t come to Providence because we run a better data center.

Tegria: What are some of the advantages of a cloud-based model?

BJ: One big advantage is cybersecurity, but there’s no security advantage to the cloud if you bring your sins with you. We let the cloud team provide its best-in-class security while we remain diligent in adherence to procedures, access levels, two-factor authentication, and so on.

Tegria: What about cost?

BJ: Switching from on-prem to cloud-based is a huge savings in capital investment. For on-prem, I have to buy very expensive hardware for my environments, and so I might as well leave those environments spun up. Whereas in the cloud, I can spin up an environment, use it for three to four hours, and spin it back down and only pay for what I need. That totally changes the game for us. And it’s the same thing if you have a particularly high load—Epic has scale-up components, and I can add application servers. So, if it’s at night or during low load, I can take those app servers out of the mix and lower my costs.

Disaster recovery is more manageable with cloud-based, too, because you can rely on Azure’s DR backbone versus having to create that yourself. And moving data around from cloud to cloud is easier than moving it from on-prem to cloud. That saves time.

Tegria: Does having a cloud-based strategy put healthcare organizations in a better position to take advantage of innovations in AI and analytics?

BJ: We’re already seeing that. One point I want to make is that almost all innovation is happening in the cloud, not in on-prem databases and software. Also, you need massive scale to do AI, which you don’t have with on-prem. So, if you are going to do AI or machine learning, you really have to move your data to the cloud.

Tegria: Do you see the cloud offering advantages in terms of being able to implement software and hardware more quickly?

BJ: Yes—cloud-based saves time and money because you don’t have to place an order, wait for the order to show up, rack the servers, apply the software, etc. We saw that during the pandemic at Providence when we had to spin up our Covid analytics work in March 2020. It happened in a single day. Without a cloud-based partner, it would have been May or June before we had servers racked to do that.

The agility works both ways: I can spin up instantly and a week later realize it was the worst idea ever and just turn it off. Versus if I experiment with on-prem hardware and it doesn’t work out, what am I going to do with all these servers? I’m stuck with them.

Tegria: Some people say that the #1 competitor for cloud hosting is a hospital’s own IT department. Why do you think some healthcare systems resist the move to cloud-based data management?

BJ: Cloud-based is new, and many of my CIO peers say they are hesitant due to questions they have about security and privacy. But there’s also just a huge change management resistance factor. For many people who built their careers in data centers over the last 20 to 30 years, cloud-based could put them out of a job. The message we conveyed at Providence was, essentially “If you hate change, you’ll hate irrelevance. So let’s help you not be the last IT pro in the U.S. who is a legacy data center expert. Let’s retrain you as an Azure-certified data expert because that skill set will be much more valuable.”

We found that about half of our people really embraced it, got Azure certification, and brought fresh life to their careers. Now they can deliver business value. That’s just a recognition of the fact that when it comes to EHR delivery, the future isn’t in on-prem data management—it’s in the cloud.