HIStalk Quotes
HIStalk Interviews Kim Pederson, VP of Excellian, Allina Hospitals & Clinics
posted 05/23/2007
Kim Pederson e-mailed me a couple of weeks ago to let me know that Allina's last major Epic implementation was finished. I'd run rumors now and then that the project was in trouble: behind the timeline, over budget, and hitting a system performance wall.
I knew HIStalk readers were interested in the outcome, so I asked Kim for an interview. She agreed, for which I'm grateful.
Tell me about yourself and your role at Allina.
I'm the executive at Allina responsible for the implementation of electronic medical records for the 11 owned hospitals and 65 clinics. We call the project Excellian.
It has Epic Systems at the core, but as we implemented Epic, we integrated a number of systems, everything from ventilators and other biomedical equipment in the ICU to lab, PACS, imaging, and NaviCare for bed management. Excellian represents the total tool that we use. It’s beyond Epic, which is why we renamed it.
Describe the Excellian project.
We actually started probably six years ago with visits and board approvals. The goal was a single, integrated system across all hospitals and clinics, EMR as well as revenue cycle.
The patient could enter the system anywhere, data would be anywhere, and whatever questions they had could be answered. “One patient. One record.” is our tagline. We wanted a fully integrated system that tied patients together from all hospitals and clinics seamlessly.
How did you choose Epic?
We looked at seven vendors and narrowed it down to Cerner and Epic. We had over 700 people look at demos based on our own internal scenarios.
We came together as a project team in August 2003. We had a four-year implementation plan, a budget, and a structure around how we were going to do the work. We finished on time and within 2% of the original planned budget.
We implemented all the functionality that Epic had at the time, their enterprise package. That includes physician documentation, CPOE, radiology, OR, ED, all the rest of the revenue cycle, clinical documentation, and both ambulatory and inpatient orders.
Our 65th clinic was finished a couple of weeks ago. Our last big metro hospital finished May 1. By the end of that week, the percentage of orders completed online was 67%. We took the big bang approach there, the first time we included CPOE in the big bang.
The system was available on May 1 and the physicians had until June 1 to use it. Most chose to do it from the start. That hospital is rising to the Allina standard of 74% CPOE. Exceptions are pre-op orders, which we still take on paper while the patient is being set up, and chemo, since the system doesn’t support chemo very well.
If a physician sees less than 12 patients a year in any of our facilities, they don’t have to use CPOE. There’s a manual system. Otherwise, if you want to practice here, you have to use the computer. If not, medical staff privileges are taken away.
What drove the decision to buy Epic?
The integration. The “One patient. One record.” was the vision.
What were Cerner’s shortcomings?
Cerner didn’t really have a robust ambulatory product at the time we saw it. They had a hard time demoing. We also had hard time getting Cerner references, especially around revenue cycle. We found only one reference for billing.
So, it was a combination of not being able to demonstrate our workflows, their references, and some of the product functionality. We decided Epic was the better choice. There were seven criteria and Epic was the clear preference in six of those seven.
I heard the original project budget was $100 million and you ended up spending $242 million.
The original budget was $242 million. We’re spending $249 million.
What benefits are you projecting?
We’ve gone after significant benefits on the revenue cycle side. I’m a firm believer that putting in an IT system doesn’t drive benefits. It’s just a tool you use to enable better processes.
In revenue cycle, we’ve implement six or seven different front-end processes to get better demographics, payor information, and eligibility so we could have conversations with the patient ahead of time.
We deliberately did the project outside of IT because we knew it would be as much about business process and clinical redesign as it was about plugging in software. IT was not the best prepared to do it. The people who ran Excellian had done project improvements throughout Allina. The leadership was improvement focused.
When we did a cost-benefit analysis, we estimated that of about $64 million worth of ongoing benefits, two-thirds were related to cleanup of the revenue cycle. One-third was in decrease in practice variation, duplicate tests, and adverse drug events.
We’re very able to measure progress in revenue cycle. Clinical benefits are harder to measure since the primary measure is cost per case. We have seen that, but we also had other projects underway for productivity, reducing the supply formulary, and getting to more standard contracts. So, it’s difficult for us to say we’re achieving all we hoped to with the automated electronic record because other areas are contributing. Clinical benefits are there and are coming, but some are enabled by the system and some are not.
On clinical side, benefits come more slowly than expected. We planned to see those within six months, but it really takes about a year for clinicians to settle in, get over their learning curve, and get comfortable enough to see some of the benefits come through. That’s the biggest lesson we learned – benefits came more slowly than what we had projected and hoped for.
How do you separate benefits you could have achieved otherwise?
Transcription, for example. We’ve seen significant reductions across the organization. Reductions in adverse drug events, which have had an associated reduction in length of stay. We’ve seen a reduction in duplicate testing, which I don’t think we would be doing without the system.
We went after common, evidence-based order sets. On the DRGs where we’ve adopted standard order sets, we’ve seen cost per case go down, although we’re only tracking a handful. One of our facilities did not adopt the order sets and their costs continued to increase. We can’t really track specific dollar amounts, but tools and order sets developed with cross functional groups have seen an improvement.
Here are some numbers. At Abbott Northwestern, days in AR showed a 21% improvement. Claims denied, a 27% improvement. Wrong claims, a 50% improvement.
In the ED, we saw a 91-minute wait reduction. Dictation was improved 17%. ED to inpatient bed arrival was a 50% improvement.
Putting in a computer system doesn’t make that happen, but it’s a dual thing.
You had 350 FTEs working on the project. How many will be involved going forward?
At height of project, when we were doing Abbott Northwestern and running 11 concurrent clinic implementations, we had 350 people. We started at 100 and are now down to about 20 on the project because everybody else has moved into their position. We had the project position, then the ongoing support. We’re in the wind-down mode tying down loose ends.
We hired them into the project full time. Many of them, probably about half, came from inside Allina. Our focus on whole build of Epic was to have people do it who came from the bedside, who understood the workflow. About 120 people were 100% project full time. It wasn’t a borrow, it was a take. Those people have either gone back to similar jobs or moved to support, documentation, or upgrades.
We had 100 contractors. Epic skills are hard to find. I have little criticism of Epic, but they keep the supply of people down based on who they will train and certify. We had to buy those Epic skills. When you pull a nurse from the bedside, they don’t have Epic skills or project management skills.
Will the project transition to IT?
The project will remain outside, in the clinical side of the organization. We started out at support, which was going to be in IT. IT had a really hard time adjusting to the requirements of the clinician at the bedside needing resolution to their issues immediately.
We struggled and finally the organization decided that it would be best if the support was lined up with the project. Support came out of IT and got lined up with the project. We had been in the support role because it took a long time for IT to get ramped up and we were giving clinicians 24x7 support.
How mature would you say the Epic inpatient products are, module by module?
They’re a lot more mature now than when we got them [laughs]. We had a long time to work with Epic. They’re very interested in having their customers be successful. A lot of development has happened.
The ambulatory products, PM and EpicCare, are mature, good, solid products, not a surprise since that’s their sweet spot. The hospital products are less mature. ED and OR in the version that we had were the least mature of the bunch, although they’ve done so much development with us.
The next upgrade for us is a three-version upgrade from Fall '04. We retrofitted fixes, but didn’t stop to do an upgrade. Products are far more mature and have great development. Allina and Kaiser and other customers have had a lot do to with that.
They’re responsive and turn things around every rapidly, but the bad thing is they’re responsive and turn things every rapidly. In a big implementation, it’s hard to keep up with Epic. I would say that in the version of software we have, hospital products had a ways to come.
There were rumors of poor Epic scalability at Allina a year ago and then the same accusations at Kaiser. Fact or fiction?
Fact. We knew when we signed the contract that there was not hardware that could handle the size of Allina on a day-to-day basis. We didn’t know what we were going to do with that.
Epic, with our IT folks, monitored where we were with headroom and ultimately ended up with a solution. We went down to a benchmarking center and worked with Epic to simulate Allina’s load and database. They worked hand in hand with us and Cache’ to come up with a solution where there’s the main database server and application servers that tether in. People actually sign into the application servers, they go do the database and bring information back.
We have more than enough headroom for Allina and hopefully for outside affiliates if we extend it to physician offices and others we have clinical relationships with. We have what appears to be plenty of room to do that without having to go with Care Everywhere [software synchronization] like Kaiser has done. They’re bigger and have more of an issue than we do.
What advice would you give with regard to user training?
We had a long road with training. It was one of our most expensive parts of the project. We have trained over 28,000 users and with massive amounts of effort.
The biggest thing I would say is that teaching people how to push the buttons isn’t what they need. You need to teach them how to do things the new way. With each implementation, we would look at each aspect and decide what went well. Training was one of the places that evolved and evolved.
Epic was not a big fan of computer-based training and we were able to help them evolve their thinking. We train earlier, then have PCs in break rooms and make sure each organization makes time for people to practice. By the time the switch goes on, people have not only learned how the system works, but how their day will be different.
We were partially electronic in some of our facilities and it was harder to have happy users when we were going electronic to electronic. Suddenly the system they had before and didn’t like was great and the new system was the one they didn’t like [laughs]. We mapped what they had before to what they had now. People started to get used to the idea that we would lose some areas of functionality and some things we like, but would gain others.
If I had it to do over, I’d want to set up a live lab for people to practice in. I periodically round with clinicians and it’s one thing to sit in a quiet place and practice workflows, but completely another place to sit on the nursing unit with the phone ringing and people asking you questions and those distractions. On those first days and weeks, the sensory overload is something that people don’t prepare for. I’ve seen that months later with physicians.
Lots of hospitals are struggling with EMR selection and implementation. What advice can you give them?
I think it should become less expensive. Part of our $250 million was that we were learning along the way and Epic was learning along the way. If we’re all good about sharing our learning, which is part of our new mission in life, hopefully we don’t all have to make the same mistakes and pay for the same mistakes.
With Epic, they have now developed their model system, which is more of a canned EMR with choices made for you so there’s less design, less build. If you’re willing to let your technology drive your workflow instead of the workflow driving the technology, Epic and others are coming out with software that’s already built that will require a lot less intervention from a development perspective and about the same for workflow and training.
Each organization has to have their vision. Our vision drove our selection. We had a strong set of guiding principles from the beginning – how we were going to customized and manage scope. If you’re willing to take what you get and not be customizing all over the place and be focused on workflow, it’s possible to do it for substantially less than an Allina or a Kaiser.
How much of the chart is still on paper or outside Epic?
Zero. We get some things from the outside and those things end up imaged. Most of our EKGs feed into the system, but some places still have the paper strips and those end up imaged. Very, very little is on paper. The pre-op and chemo orders, just because the system doesn’t lend itself well to those workflows.
How did your total project cost break out?
The vast majority was for bodies to do the work – the design, build, validation, training, and support. Epic’s piece was a relatively small portion of it. The cost was about 40% on ambulatory. We had enough people to bring 11 sites live concurrently in 65 clinics over four years. In the beginning, it took five or six months to implement a clinic. At the end, it was about three months.
We made a significant investment in infrastructure. I bet 10 or 15% of that was spent on the network and redundancy and those kinds of things.
Software was the smaller cost. The larger cost was doing the work and making sure we had a backbone that could support it.
Are people stealing your employees with Epic experience?
They sure have tried. We’re probably retaining about half of the 350 in the permanent structure and those people are recruited all the time. We lost a lot of people during the project to consultants. There are always small startups that come after people.
There was a fair amount of picking off each other’s resources in the area since several hospitals went with Epic. We had an employment guarantee and from the height of 350 down to now, we have probably about 20 people who haven’t been placed.
What area did you come from and where do you go now?
I’ve been at Allina for 23 years. For the last 10 years, I have done one redesign or improvement project after another. I had asked to have this one.
Now I’m done. I’m going to be going out and looking for the next big challenge, as are most of the people in the project management office.
None of us that were part of the project management office had any experience implementing anything of this magnitude and only a couple had experience implementing anything at all. We were told by all kinds of IT people that we couldn’t do it, but we figured it out. Most of us are not IT people by background.
What’s the structure going forward?
That’s part of why I’m leaving. I like beginning and an end [laughs].
The people that we have hired to carry forward the next phase of the project are really committed to make it better and better. We have about 160 people carrying forward the future. Fifty of those will be doing upgrades and close to 100 will be doing support. The rest of that will be focused on optimization -- functional optimization, but also workflow optimization, using our user groups to identify priorities, measurements, and analysis where we’ve lost productivity or have big cost issues.
Recently we decided that maybe we were seeing some revenue slippage at one of our sites, which turned out to be a budget problem, but found where auditors were contacting doctors during concurrent coding review. They’d put a note on the paper chart and if the physician thought that was appropriate, would document or say it wasn’t appropriate. With the electronic record, that’s disappeared. I’ve been leading coders and a group of docs to get back to the workflow and prompting. Using the tool better and better is our focus.
What happens next at Allina?
The first big thing is the three-version upgrade. Work will begin in July. The goal is to bring it up in early to mid-2008. Nobody has every done a three-version Epic upgrade, according to Epic, all at once.
Our philosophy at implementation was to do it good enough, but not perfectly. Now will begin to optimize use and make sure we’re getting value from our investment, like making sure that training stuck and adding on new features.
comments (9)
1. The PACS Designer left...
05/23/2007 8:21 pm
Mr. HIStalk, you have really allowed all of us to be fully informed about
how much work goes into a large software project such as Epic with your
excellent questions. Since the Epic software cost is smaller than the rest
of the costs it sends a clear message to anyone anticipating a similar
rollout that they need to spend significant resources on getting ready
before this new paradigm is turned over to users! Thank you for posting
this interview on your site.
2. jimmer left...
05/23/2007 8:31 pm
$249 M in HIS for EPIC...Amazing...EPIC is NOT battle tested. Why the hell
would any CEO, CFO, CIO put her/his ass on the line for this product???
3. HITSupporter left...
05/23/2007 11:29 pm
Good interview. Congrats to both the Allina and Epic teams. Quietly
getting the job done - nice work. People don't always realize the
magnitude of some of the projects like Allina and Kaiser. Very nice to get
an inside view from someone who led it. Well balanced and honest
assessments like this help lead others through these very challenging
projects.
4. VJ left...
05/24/2007 12:24 am
Excellent Article - Very helpful to understand the market dynamics
5. Art Vandelay left...
05/24/2007 8:02 am
Great interview - exactly the reason why you are one of the best overall
blogs out there. I totally agree with TPD's comments. It is interesting
to see this done outside of the IT department. Not that I fully agree with
the comments that IT Doesn't Matter but I do agree with the need to be
flexible and more linked with the workflow. Configurable systems allow
this paradigm to occur and it'd be great for Allina to continue to share
their experiences.
6. Epic_Escapee left...
05/24/2007 10:20 am
It's interesting to hear Allina's perspective re: Epic experienced
employees. When the Allina project first got rolling, many people w/ Epic
experience from the other Twin City health care organizations quit and went
to Allina. I personally worked with one client where almost half of their
Epic project team went to Allina.
7. Lacey Underall left...
05/24/2007 1:18 pm
Excellent interview. It's great to hear from our side of the fence
(customer side) for a change. Any other CIO's want to give it up?
8. Epic Implementer left...
05/24/2007 1:30 pm
Thank you very much for the informative article! As a past contractor at
the Allina Excellian Project and as a current full time employee of
Kaiser's HealthConnect Ambulatory project, I found it very interesting to
hear the lessons learned from a high level perspective. Thanks for such a
great interview. Of course Epic is still learning... I think they have
grown very large in a short time. I know they have their problems but I
think its only natural for a company with that kind of growth to have
growing pains. People point them out more because they are strange - they
are privately held by a very "eccentric" CEO, and they are also very
successful. Maybe if Judy started charging nonprofits 250 bucks a pop to
come hear her pontificate at a mansion people would like them more??? :)
