Maximizing Quality from IT
, October 13, 2009
KICHAK: Everything we've talked about has a play in the 10%, and that's why I think we should have put the bar a little bit higher, even if we failed.
VAUGHN: Voluntary CPOE models don't work. It takes forever to move from 10%, and then all you'll do is you'll move to 30%, and you'll stay at 30% until it becomes mandatory. We told our physician leaders that
we believe this needs to be mandatory or we're going to get no benefits. You're going to get no efficiencies because you've got half your chart on paper. Our med execs all agreed, they passed it, and now when we go live, we go live mandatory.
Mining the data channels
HEALTHLEADERS: What are the challenges in how healthcare IT organizes and structures the data being collected?
PAUL: Right now, the desire for publicly reported quality measures has moved ahead of our ability to efficiently capture the data, which are still largely abstracted and gathered by hand. Consumer and patient advocates rightly want to see quality-related information from our hospitals and doctors, so they've been driving this ahead of the ability of the system to collect the data. We have a chance right now with the push toward electronic health records. If we could sync that back up so that the EHR that is implemented is one that can actually populate those measures by design, then physicians could start pulling out quality-related information and improving in real time. That data misalignment now makes doctors and hospitals just crazy.
VAUGHN: There's not enough structured documentation to describe a complex clinical scenario with five or six problems and a crashing patient. There are five or six fields that are critical but the majority of it is going to be some artistic description of subtleties that tells a story that we all want to be able to exchange because we're human beings. So you've got to find a balance between a tool that does not require me to log in two different ways or to do two different activities, but allows me to do structure and nonstructure in the same encounter, at the same time, with tools that are seamless. Nobody's there yet.
MORLEY: Any discussion of structured data is primarily organization-dependent. But the implications of that dialogue affect the vendor community. For example, a typical vendor may have 2,000 customers in the U.S. Helping develop those unique capacities to support that individual organization is a significant demand. So the sooner the standards get resolved, the quicker the vendor community will be able to support its customers' initiatives.
HEALTHLEADERS: What are some steps that hospitals can work on now to improve gathering the right data at the right time?
PAUL: Linking directly with monitors and equipment in the hospital, such as biomedical equipment like blood pressure cuffs and cardiac monitors.
MORLEY: Our business is high-velocity, but the area in which we are observing the fastest growth, and therefore providing the greatest support, is information technology infrastructure. We are also heavily involved in the entire point-of-care experience, which includes a variety of information technologies that can aid providers in their quality improvement initiatives.
VAUGHN: What's going to be interesting is if we're looking at safety, it's going to take more than just that one side of technology like the smart pump. It also is going to look at how we're packaging our pharmaceuticals, so that you can't plug something into the pump unless it's exactly what you are telling the pump it is.
Spending on infrastructure
HealthLeaders: Why is healthcare so far behind in data warehousing?
KICHAK: Healthcare has never spent the money in IT. Its spend rate has always been at 2%. So we put in that 2% in operational systems—get the lab system, get the pharmacy system, the nursing, the CPOE, et cetera. Only recently with key clinical systems being up and running are we starting to get a little funding to do the decision support from a datawarehousing standpoint. There are two types of decision support. There's the one that all of us are interested in, which is clinical decision support, but there's also decision support coming off of a data warehouse that helps you build the next treatment patterns or it looks at your ordering patterns.
MORLEY: The effectiveness of data warehousing tools is dependent on two things. One is the data that providers accumulate and two is the data they're trying to extract. If providers have a clear understanding of the data they need, then these tools can be an asset. We're not aware of many organizations that are using data warehousing to support outbound needs like clinical decision support. They may, however, use it internally to satisfy unique intelligence needs, specifically for discrete financial and clinical data.
KICHAK: Go to any MBA textbook and try to find a data model on healthcare. The first thing that's needed to build a database from a pure technical standpoint is to have a data model. Data models for healthcare exist in plethora. If you've seen one, you've seen one.
PAUL: I'm curious what your recommendation would be to community hospitals out there in the country that don't have EHRs. Should they buy an EHR now or later?