(Part I)

GUERRA: Most CIOs probably understand they have to check all a vendor’s products by going “under the hood,” not relying on the assumption that everything is integrated; but do you think IT directors at small hospitals are aware of that need also?

CLOUD: Yes, I think it’s largely driven by background. The CIO has a hard job because you have to be a business person, and then you have to be a technologist at the same time and, as you know, those two things don’t exist well in the same body quite often. I would say, probably the majority of CIOs know that, understand what you’re talking about, but unfortunately there are a lot that don’t. And especially in smaller shops, the temptation is to take the vendor’s word for everything because it’s simply less trouble, especially if you’re short-staffed and you don’t have professionals on your team to guide you in this regard. CIOs have to recruit well because they can’t know it all, and if they don’t have trusted advisors on their team who understand software and understand the vendor’s business, who understand development and all of that stuff, they can certainly get caught flatfooted and accept a lie.

GUERRA: It sounds like it can be critical to bring in some consultants who can help you make these decisions.

CLOUD: You got it. You either have to have consultants or some talented employees.

GUERRA: Otherwise, you could make a big mistake.

CLOUD: You could make a huge mistake. And I’ve certainly seen it during my 30-some years in this business.

GUERRA: Tell me about what you’re doing with Compuware.

CLOUD: To me, the story isn’t necessarily about Compuware; it’s in the problem, because I suspect it’s being faced largely by CIOs across the country.

Anyway, I took this post and quickly discovered that the appetite for automation was insatiable and a lot had been achieved. So we were essentially victims of our own success. I’m very, very pleased to be part of an organization that understands the value of information-based automation to improve the quality of healthcare and the efficiency of our operations. I don’t have to provide that level of education here, but I’ve had to do that in the past.

But that boon also had a bane, and the bane was our ability to keep up with demand. We were moving quickly toward full EMR automation, CPOE progress notes, electronically controlled meds administration – all of that required an IT infrastructure that was as solid as a rock. And ours was not. It required service levels that were high and ours were not. So to put it bluntly, I felt that this IT organization was going to be crushed under the weight of future demand.

So, what we had to do was figure out how to prepare this organization for what I saw coming. So I collaborated with my CEO. He understood the challenge, and we basically called a moratorium on new implementations. That doesn’t mean we didn’t do any, but rather that we did very few for a year. Our agreement was we would spend a year fixing IT and so, as we explored what methodologies we could employ, we knew that implementation of a strong service-management model was important.

But I am one that does not believe in talking industry jargon and wooing and wowing people with acronyms; I knew I had to make the whole service-management concept mean something. We translated it to our organization by saying that for one year, we are going to concentrate on four S’s: Stabilizing, Systems, Services, and their Satisfaction. We coined this term “Project S.”

What we did was determined what automation tools might be required to stabilize all of those things I talked about. I believe in integration for IT systems just like I believe in integration for clinical systems. We sought to determine how we could automate the workflow of IT, with as few tools as possible, for information to flow seamlessly from one mode to the next.

I’ll give you an example of that. I coined this term called the “Metamorphosis of IT Management Information.” A call comes in and is recorded on a call-tracking system. Data is moved to a ticketing system. The ticketing system data is moved to a project. The project produces the need to track time, to track status. So then you have a status reporting system: the project management system, a change control system.  

And so what we were looking for was something where the data moves within the same system and it changes states without requiring a change of systems. That’s what led us to the (Compuware) Changepoint/Vantage combination.

The other thing we needed was something to monitor the environment so we could have a predictive environment. What was killing me was that the way we found out things were broken or underperforming was through a call from a customer. That’s not the way to find out. And so, that’s where Vantage came in. So Changepoint really dealt with the internal operations of IS. It allowed us to automate our workflows, and Vantage gave us a strong monitoring environment that integrated with Changepoint, The data could flow between the two.

Project S was launched, and we set specific goals. So we tried to determine how you can best measure the improvements in stabilizations. We came up with some metrics for that. We knew that one thing we had to do was move more strongly into a service-level agreement environment, and so we divided service level agreement objectives into two parts – service level agreements with our vendors and service level agreements with the departments that we serve. And what we wanted to do is build all of that into a circular measurement of employee performance. So we have a system in which the managers set their goals at the beginning of the year, then we determine how we performed against those goals. Performance has an impact on your raise the following year.

What we did was the agreements that we made with our departments and agreements that we made with our vendors – all of those somehow make their way back into the IS managers’ evaluation. So what that did was create a service environment where all the incentives were aligned. Even better than aligned, they’re pretty much the same. So that gives you a pretty good picture of what the two products contribute to our success in implementing a service-level environment.

Now, what was really quite nice is how Changepoint helps us automate workflows. So what we do is we chose four major workflows – project management is the clear and obvious one. And what I said was I wanted that automated to such an extent that an idiot could do it. That it would be impossible for them to depart from pre-established procedures and standards. And so we actually have a workflow engine that guides the project manager through every step of the process. And so I’m guaranteed to get repeatable quality.

GUERRA: You mentioned how important it is to have the right infrastructure. What exactly are we talking about when we use the term “infrastructure”?

CLOUD:  Great question. Think of wanting to add a data center but not being able to do so because you’re out of power, out of cooling, out of space.

GUERRA: Out of luck.

CLOUD: Right, I’m out of luck. I had to address the data center environment. We had systems that were not really monitored. So there was really not a good way for us to tell if the system was in stress or distress, filling up and all that. So we had no monitoring tools on the systems that were in this data center. That needed to be fixed. We did not do a good job of watching interfaces for failures. We had to fix that. It’s amazing we were as successful as we were. What happened was the organization grew faster than IS did, and so it was just a matter of taking the very good team I inherited and giving them a chance to take a breath, to get caught up.

Our disaster recovery was not what it needed to be. I spent a lot of sleepless nights recognizing that with a bad case of Murphy, I could end up not able to recover critical information. So we had to work on disaster recovery scenarios. We knew that virtualization was going to be very important if we were going to consolidate servers and improve testing environments because our testing environment was insufficient.

GUERRA: What about wireless?

CLOUD: Wireless too. As a matter of fact, we rolled out some wireless applications and we got pats on the back on the day of go live because people were glad to get them. Unfortunately we had a shock not long after because the bandwidth was insufficient. Our coverage was insufficient. Capacity was insufficient.

That’s an interesting study too because the wireless environment that was here when I got here was deployed based on a certain set of assumptions, established years ago. And it’s amazing how quickly technology can change and totally void decisions made only a few years prior.

What happened essentially was the wireless infrastructure that was in here was developed around the idea of COWs (Computer on Wheels), so you needed a certain density of antenna coverage. When we started going to handheld devices, especially in the area of pharmacy barcoded meds administration, a totally different density was required. It’s amazing how easy it is to miss that kind of thing. Since as we were moving into barcoded meds administration, it required us to totally revamp of wireless environment. WEP protocols used to be okay for security, and we had to go to WPA because we had a great demand to open up our networks to public access. So we have a guest network that patients and families can use without having to ask. You can’t do that without having the right security imports. So, security on the wireless network needed to be fixed.

GUERRA: Is it a good analogy to think of the infrastructure as the basement or the foundation of the house and the applications as additional floors, so if you don’t have a sound foundation, eventually the structure is going to crack and crumble?

CLOUD: That’s exactly how we look at it. If you remember back when we started our conversation, I said that during the assessment, we saw what was coming and we knew that IS would crumble under the demand. Demand doesn’t mean the number of requests, it means the types of things they were asking for.

So, yes, an insufficient infrastructure will sink an organization. I remember years ago when I was CIO in another organization, and one of the board member of a very large Fortune 500 company came up to me and said “Hey, let me give you some advice as a CIO that I gave my CIO. Never allow IS to become a constraint to business decisions.”  I never forgot that.

And so the whole idea is to have a flexible infrastructure of sufficient capacity that can respond quickly to shifts in business requirements, and that’s what we’re trying to do.

GUERRA: When did you start in this position?

CLOUD: I’ve been here five years.

GUERRA: As a final question, he gave you that great advice, do you have any advice for our readers, your colleagues?

CLOUD: Make assessments your friend. The temptation is to see a problem or an opportunity and run after it. Everything seems simple upfront but when you dig into it, you find out even the seemingly simplest things can be complex. Have a good assessment strategy, involve good consultants, develop good relationships with consultants from outside, develop your own internal assessment methodologies, and don’t start until you’ve well-defined the problem and also defined success.