CMS Proposed Rule: How to Stretch Without Breaking

By: Louis Wenzlow, Rural Wisconsin Health Cooperative Figure 1: Hospitals (I Believe) Likely to be Excluded from Incentives Let's Get Meaningful One thing we can all agree on: for the EHR incentive program to be meaningful, it needs to be designed to support our national goals of reducing healthcare costs and increasing healthcare quality. We're moving to electronic health records not for the technology's sake, but because we believe the technology is a means to actually help people and make things better. The main area where people disagree is whether or not the meaningful use bar has been set too high to facilitate the accomplishment of these common goals. Those who argue for a high bar believe that lower standards will lead to watered down benefits. Those who argue for a lower bar believe that unachievable standards will lead to dramatically fewer providers making EHR adoption (and accompanying quality and efficiency) gains. Mixed into all this are accusations from high-bar proponents that people who disagree with them are lazy whiners who should be focusing on meeting the standards rather than arguing against them, as well as suspicions from lower-bar proponents that advanced-EHR hospitals, systems, and provider groups have found a way to skim billions from the tax-payer trough for work that they have already done. Let's leave behind the name-calling and get meaningful! Stretch Don't Break According to ONC, the principle for determining the meaningful use bar is to find the appropriate balance between feasibility and urgency. National Coordinator Dr. David Blumenthal has recently said that he intends to "stretch but not break" the healthcare community in setting the threshold for meaningful use. I agree with this principle of "stretch don't break," but it's unclear to me how and even whether it is being applied. What seems to be getting lost in the discussion is that it is logically impossible to "stretch not break" hospitals and physicians that are at very different stages along a continuum of EHR adoption by using a single rigid meaningful use standard. If you stretch providers at advanced stages of EHR adoption, those at early stages will break. If you don't break providers at early stages of adoption, those at advanced stages won't stretch. Figure 1, at the top of this commentary, illustrates this issue. The colorful 7 stage grid includes HIMSS-provided percentages of critical access hospital (CAH) compared to prospective payment system (PPS) hospital EMR adoption statistics. I have added three text boxes to indicate (1) the CMS Stage 1 meaningful use threshold (black); (2) my assessment of the HIMSS stages that are least likely to meet these thresholds (red); and (3) my assessment of the HIMSS stages most likely to meet these thresholds (blue). In the Figure 1 comparison, 70% of CAHs are at stages that I believe are less likely to achieve meaningful use, compared to 46% of PPS hospitals. 48% of CAHs are at the two lowest stages of adoption, compared to 15% of PPS hospitals. We could (and should) do this same analysis with rural, small, disproportionate share, independent, and other categories of hospitals. The point of this is that when we talk about "stretch don't break," we need to clearly identify where providers are starting from, what timing requirements we are assuming are reasonably achievable, and what types of providers we are specifically referring to. Americans who live in communities where providers are likely to have lower levels of EHR adoption have a right know that ONC and CMS have decided to "break" their local rural providers in order to "stretch" the urban ones a hundred miles away. How to Stretch All Providers If we are truly committed to stretching without breaking, how do we do this for all of our providers, whether urban or rural, small or large, independent or system-owned, PPS or CAH? One way is to create more than one meaningful use bar, so that providers at different stages can all be incented to make meaningful EHR adoption strides with consideration paid to their starting points. Another way is to allow for flexibility. Instead of creating a one-size-fits-all, all-or-nothing meaningful use standard, why not allow providers to select the 90% of the requirements that are most suitable for their environments? If we force providers to move faster than what is a reasonable stretch we will in all likelihood see lower not higher quality. (See my blog on this issue at http://www.worh.org/hit/2010/01/cms-proposed-rule-threatens-care-quality-in-rural-communities/) A third way is to simply exempt certain types of providers from a portion of the meaningful use requirements, at least initially. If we know that it is unreasonable to think that 70% of CAHs and rural hospitals can implement CPOE in time to receive incentives that could go toward other important EHR adoption work (such as Pharmacy systems with contraindication checking capabilities and inpatient nurse documentation systems), then why are we requiring CPOE for these types of hospitals? There are those who have come to the false conclusion that there is a secret sauce, an EHR implementation recipe that, if only all providers follow the instructions in the exact same way, will somehow fix the problems of our healthcare delivery system. My experience, which I think is borne out by most of the existing research, is that no such single recipe exists. Rather, EHR implementation success depends on an organization's ability to ascertain the distinctive combination of interventions and strategies that will work within the organization's specific environment. According to AHRQ's Costs and Benefits of Health Information Technology, "HIT implementation consists of a complex organizational change undertaken to promote quality and efficiency. Studies of organizational change are fundamentally different from studies of medical therapies. Organizational interventions interact with a wide range of organizational system components. To be successful, they must address these components in a locally effective way. Thus, in a sense, these interventions are by nature not widely generalizable..." How are we accounting for this dynamic complexity by imposing the same rigid all-or-nothing meaningful use standard on every type of provider?