Adoption is a major issue. The Mass General/George Washington University report shows that doctors are not moving to EHRs as quickly as previously thought.

Brailer: If there is anything that we try to do it is to have a physician realize that this is not an easy process. The issues that really matter here are changing how doctors do what they do, in collaboration with other parts of the healthcare system, from pharmacies and labs in hospitals to nurses, front desk clerks and others. That whole workflow, obviously, must change to achieve the kind of high performance healthcare system that we want, and it goes far beyond loading software.

Hilary Worthen: I am not surprised at the gap between the number of places with a partial implementation and those few with a full one. The most sophisticated parts of EHR functionality are those that support critical multidisciplinary workflows, often involving multiple departments. You can implement the more basic EHR functions and still have people happily working in their silos, but to do the hard stuff, people need to change their entire way of thinking to a much more collaborative, system-based approach. From what I see, this poses a much deeper challenge than just getting docs to use a computer.

We have implemented Epic's EpiCare system in the [outpatient] realm and we are just now launching it on the inpatient side. Convincing people to do the optimization [of business processes] and to help them understand [how EHRs work] are the big challenges we are facing now. The whole idea of process improvement is something that clinicians do every day. They are trying to improve the process of getting patients to take care of their diabetes, for example. But when you start to articulate it in business terms, it doesn't instinctively grasp clinicians' interests. We are working pretty hard to try to get doctors and administrators to understand each other's languages and to collaborate in the way to make those workflows affected by EHRs much more reliable and accurate.

Glaser: The 10 percent figure means that advanced features will lag in adoption as long as there is little incentive, such as reimbursements, for doctors to use those features and incorporate them into their daily practice.

But, the first thing we have to remember is that sometimes we label users as resisters and sort of cast them as being 4-year-olds who need a nap. Frankly, we are asking them to do tasks that they didn't do before. The way we get them through this is to engage them quite extensively in decisions of how we implement. You do as well as you possibly can on the software to make it as feature rich as possible. You try to make it as reliable as you can.

We've also begun to work with some insurance companies so that if you are a physician who is using this and you are using it well, there's additional money you will get from the insurance company that can sort of make the economics better.

As we see with Kaiser's EHR implementation, there are significant technical problems to overcome too.

Brailer: Major corporate change projects of any sort tend to be complex and full of surprises, and Kaiser's is no different. Someone has to pioneer the EHR in large corporate settings, and it might as well be Kaiser.

Worthen: I am not unbiased, since our organization is an Epic customer too. There are certainly a lot of examples out there of robust Epic implementations. When you add up the number of sites and the hours affected [by system downtimes] — given the scale of the implementation I am not sure anyone can do much better. That said, the scale of Kaiser's project is indeed enormous, and that level of complexity is likely to uncover unforeseen issues. I'm not sure it means that organizations tackling a less ambitious implementation need to alter course.