Implementing health information technology processes into a clinician’s workflow is not child’s play. It’s hard work and if not done correctly, it can be detrimental to an organization’s ability to treat a patient.
As deputy chief medical officer at Atrius Health, a Newton, Mass.-based health system, Joe Kimura, M.D. knows this. He understands the importance of properly embedding health IT tools into a clinician’s workflow so that things are done the right way. Since the late 1990s, Atrius has been ensuring its clinicians aren’t met with the kinds of IT stumbling blocks that hold up clinical workflows, something which occurs at even the best of healthcare organizations.
“In tomorrow’s medicine, if not today, everything we do clinically will be touched by some kind of HIT capability,” Dr. Kimura says.
Kimura recently previewed some of the talking points in a Q&A with Healthcare Informatics Senior Editor Gabriel Perna. Below are excerpts from that interview.
As a doctor, what have been some of the challenges you have seen in trying to bring the electronic health record (EHR) into a clinical workflow?
In our organization, the EHR and any HIT tool is absolutely embedded through almost all of the clinical workflows that we do. It’s our transactional, operational, and documentation system, it’s all of that. When I’m in clinic, the second I’m logged in, all the activity I do, communication with nurses, specialists, all of the orders, the bills I input, it’s all through the EHR. Everything we do is fed through that.
We’ve been on Epic since 1998, and we were on a homegrown EHR before that. From my perspective of using generic HIT elements in the clinical workflow, within our organization it’s bread and butter. There is not a lot of concern about getting people to adopt the HIT systems, especially around the basics of clinical workflow. When you’re trying to do new clinical tasks, say around population health or care management activities, with new HIT tools, then it becomes a more robust conversation around innovation and use that can be more challenging around diffusion. In tomorrow’s medicine, if not today, everything we do clinically will be touched by some kind of HIT capability
Atrius has been doing this for a while and has a background that most organizations lack. From an industry overview, it just hasn’t been as smooth. What can be done to ensure it is “bread and butter” and an organization is not dealing with workflow challenges?
The core idea behind it is there is no perfect EHR. Even the ones out there currently need more work. Until we get to later generations, the EHR will not integrate as well into what I’d call “natural clinical workflow.” The challenge for the technology is that there is such as heterogeneity around the clinical workflow processes. Even though we may all be primary care physicians and we may all be treating the same kinds of patient communities, there is an individual stylistic element for most physicians on how they actually do their work. It’s the art side of practicing medicine. A lot of that has to do with how they are communicating, how they like talking with patients. The challenge is EHR systems can’t be configured to every individual. You need to come up with some kind of standard. Clearly, vendors are trying to come up with the ability to customize things—most of it is trivial though—and they start to get fixed on how the system supports the functions you need to do to complete a visit. That’s where you start to get some tension.
When clinicians feel like the system can’t slow them down, a lot of it is around their own expectations. If you are a three doc practice and you have a custom EHR, it could be great. You may say, “This thing does not slow me down. It does exactly the thing it needs me to do when I want to do it.” But for most of us using big box vendor EHRs, there is a challenge that there is a global set of standard workflows that we’re trying to get people towards, which will force you to change the order of how you do things. It could make you re-learn own workflow to utilize the power of the EHR. That’s hard. But is it any harder than clinician a being taught to do their lung exam a better way? It’s hard to change that workflow because you have a way of doing that exam, but the evidence points you in a different way and you know it’s better. Clinicians have the ability to adjust and change to make it work. Where you hit the stumbling block is when the system is making you do stuff without out that benefit or making it worse. If the new way of doing the lung exam gives you less information and you’re forcing them to do it, that’s when people get irritated. That’s when HIT implementations are not done in a thoughtful way. It can be problematic.
So how do you ensure physicians are engaged fully in using the EHR to its fullest capacity?
As with any generic change management effort, you need have some balance of frontline engagement with management leadership. Certain things can’t be fixed, even if 100 percent of frontline physicians voted for something. If it can’t be done, it can’t be done. But if you are having a conversation and collectively developing things and being transparent, you’ll be surprised at how much physicians understand limitations and become willing to work with you to accept new realities. It’s the process of transparency and a collective design approach, which we do a lot here. It’s slower and less efficient but probably leads to better adoption going forward. You have people’s engagement at the get go, not at the very end.
What is the ideal governance structure for bringing technology and data into the clinical workflow?
Data governance is a structured element. If you are trying to make sure there is a good system in the organization around not only what is supposed to be the “source of truth” in terms of raw data, but the aspects around metrics and how those metrics are presented and in what context, then there has to be a consistent view of that, regardless of where that information is coming from.
That ideal has, at a minimum, three levels. There is an executive level. You need executive buy-in to ensure everyone is on the same page….to define that standard. At the second level, you need directors, people whose business lines and direct accountabilities are responsible for those particular metrics. The reason we get 17 different definitions for readmissions is because the people responsible for that unit are making the metric as most helpful to them. Again, each hospital may have a different view of what that helpfulness sounds like and that’s how they came up with different definitions. To get back to a standard definition, you have to get those accountable directors to give up some of that specificity in order for entire organization to be using a standardized system. That’s usually a big political battle. The final level is the raw data element. You absolutely need a core process, starting from your ETL [extract, transform, and load] process all the way to the populations of your EDW [enterprise data warehouse], all of those levels need a fair amount of data quality checking. That’s a technical element. That’s at the ground floor of data and analytics. Governance requires all three levels.
And it requires them to work together.
Yeah but the executives don’t want to hear that the ETL was broken. There is an assumption at each level that the level below it is trying to optimize its function. The data people and ETL shop are making sure the ETLs are correct. The directors are assuming it’s correct and calculating the metrics the same way. And the top level executives are saying as along as metrics are the same and the data is OK, we should be moving together whenever we enter a new domain of value analytics.
To hear more from Joe Kimura, M.D. register here for the Health IT Summit in Boston.
