Better Care Through HCIT 101: Part Three, Go-Live is the Springboard to Achieving Benefits - Choosing goals that deliver benefits | [node:field-byline] | Healthcare Blogs Skip to content Skip to navigation

Better Care Through HCIT 101: Part Three, Go-Live is the Springboard to Achieving Benefits - Choosing goals that deliver benefits

April 2, 2009
by Joe Bormel
| Reprints

Better Care Through HCIT 101: Part Three, Why Implement an EHR?

    • Go-Live should be the Springboard to Achieving Benefits
    • Choose goals that deliver quantifiable benefits
    • Meaningful use should be linked to quality/cost/access?

    What can your clinical information system do to improve health in your community?

    Ten years ago, you might have had to stop and think a bit.

    One year ago, your first sentence would probably contain the word "Safety" or the words "Patient Safety." The thought behind that would have included "integrated" medication management; its specifics would have included BCMA and CPOE.

    The discussion itself is problematic and inherently private -- no one is going on TV to tell the world that their institution had 10 avoidable deaths last year.

    Next year, your answer will likely be informed by ARRA. The vision: increase coverage and lower healthcare costs, using an EHR in which qualified providers can demonstrate meaningful use, and the EHR is certified. We're hopeful that meaningful use will tie back to last year's "Safety" through ePrescribing and it's in-patient analogue, CPOE.

    Okay, we're all, I presume, on the same page for vision.

    The immediate realities for clinical HCIT include everything around the basics - a problem list and it's cousins, the medication and allergy lists. Forget about the rest of clinical documentation for a moment.

    We all know how very specific test cases can create instant clarity on our real current state and the cost and timing of getting to a desired state. I've

    written previously that a powerful test case for EHRs is support for Coumadin management. This translates to probes for vendor selection, specific capabilities of flowsheets, order sets, and care pathways. No vendor I've seen would get a perfect score, or even a near perfect score, when you look at the TOC (total cost of ownership) for HCIT and Coumadin management. Oh, by the way, for the last decade Coumadin and related products have been the number one medication safety issue, at least in hospitals (see


    So, Coumadin covers the broad category of a test for drug safety. There are easy and obvious equivalents for diagnosis (service line, disease, condition) where Heart Failure serves well. And for process and workflow support, the pre-op surgical assessment is extremely effective. Combining clinical information completely in these examples continues to be challenging. Although several vendors have made huge strides forward in their 2007 and 2008 code, the TOC remains absurd for these common categories of clinical HCIT enablement.

    Today, I'd like to suggest diabetes detection and management as the best, patient-oriented test case for

    meaningful use of HCIT. I was inspired by the recent article,

    HbA1c Gains Legitimacy For Diagnosis of Diabetes, by Mitchel Zoler (Internal Medicine News, March 15, 2009.)

    Advances in the diagnosis of Diabetes Making HbA1c an accepted diagnostic test—let alone the preferred test—has been on the table for years. At a recent meeting sponsored by the ADA in New York, Dr. William C. Knowler spelled out the case in favor of using glycosylated hemoglobin, as well as the shortcomings of this approach. The strengths of HbA1c as a diagnostic tool include the following:

    ▸ A more standardized assay with less interlaboratory variability, compared with blood glucose measurements.

    ▸ A better index of overall glycemia.

    ▸ Consistency in using the same assay for diagnosis that's also routinely used to monitor patient treatment and to predict the risk for long-term complications.

    ▸ No need for fasting before the specimen is drawn.

    ▸ No effect from acute changes in blood glucose levels, such as those caused by illness.

    In short, there's a relatively new, legitimized way to screen for diabetes. It has lower costs than current testing. It has the potential to make increased population coverage more affordable for this huge segment (

    Diabetes has a 4% to 10+% prevalence, much higher in Medicare-age population, with huge opportunities for cost savings.). Subsequent diabetic management is well known to be considerably more cost-effective than the alternatives.




    I'm not aware that any organization has documented genuinely improved processes and outcomes with genuine cost savings for diabetes (hence CBO's projections). We have published early data on value-added processes for open-heart surgery. (Casale, A., R. Paulus, et al. (2007). "ProvenCareSM: A Provider Driven Pay-for-Performance Program for Acute Episodic Cardiac Surgical Care." Ann Surg 246(4): 613-21.) More recently, we have identified a 15% reduction in CABG process costs.

    What most distinguishes organizations that have reported such improvements in any domain is their organizational change and process re-design capabilities. (Of course, they support re-design with HIT.)

    I think what CBO assumes is that organizational change is remarkably difficulthence the inability (or refusal) of GM and Chrysler to do any of the obvious things everyone else knew they needed to do 20 years ago. Once the reimbursement is aligned to reward the creation and operation of value-added (improved quality, decreased cost) care processes, process re-design (or purchase of optimized processes) will accelerate with startling rapiditymy guess, starting in about 36 months.


    James M. Walker, MD, FACP
    Chief Health Information Officer
    Geisinger Health System

    Thanks Jim.  I know that you're right.  And, by the way, congratulations on your vendor's CEO appointment today!

    Jim and I correspond regularly, including during my research for this blog post.

    The CBO projections Jim speaks of relate to content that I chose to omit from the original post for brevity.  Honest!

    Since Jim brought it up, here's the story:

    I searched several recent Congressional Budget Office reports to see if my diabetes case had been studied.  Well, it had, but indirectly in the December 2008 Budget Options report.  In the first chapter of this report, CBO acknowledged that modeling and interpretation are hard.  Costs are often shifted by the programs elaborated in each option.  Benefits accrue over time.  They can be cost savings or quality improvements or both, and all of that is important from a policy perspective (of course).  So the option that covered Diabetes combined care improvement with the payment model for Medical "Home."  There's much more worthwhile detail in the report; I've copied the essentials below in blue:


    Medicare beneficiaries who have chronic illnesses, such as
    diabetes and congestive heart failure, consume more of
    the program’s resources than do beneficiaries who are in
    better health. Approximately 75 percent of Medicare
    spending pays for care for beneficiaries who have five or
    more chronic conditions and see an average of 14 differ-
    ent physicians each year. Research has shown that care for
    beneficiaries with chronic conditions is often fragmented,
    with little communication or coordination among the
    multiple physicians who care for them. Five percent of
    Medicare beneficiaries account for 43 percent of the pro-
    gram’s overall spending, and the costliest 25 percent of
    beneficiaries account for about 85 percent of outlays. An
    intervention that focused on coordinating care for those
    beneficiaries could both improve their health and reduce
    Medicare spending.

     ... Because the requirements to become a
    medical home would be demanding, only about 1 per-
    cent of physicians would be likely to meet the require-
    ments initially; however, in the Congressional Budget
    Office’s estimation, participation would increase over

    I think this example highlights the real-world challenge of defining meaningful use of EHRs from the perspective of projecting economic benefits.   These same challenges show up when projecting the economic benefits of reduced ADEs, test duplication and other benefits from clinical decision support.

    All that said, if we're going to translate the high level ARRA into explicit terms:  meaningful use (and expected benefit of that MU), and the related certification issues, I still think looking explicitly at care is an important option.

    To summarize why:

      We can better assess a meaningful multiplier - i.e. # of patients effected (Patients).

      We can better use existing, known cost data for that care today (Today cost), many non-tamperable sources for that).

      We can better project expected cost data based on evidence-based medicine (Tomorrow cost) using EHRs.

      and track Benefit achievement = (Patients) X (Today cost - Tomorrow cost)

    There are lots of benefit models out there.  Famously, the CTL models come to mind. 

    The appeal of this "diabetes and EHR" approach to defining meaningful use is that it maps directly to the two things that policy makers are consistently stating are the objective, increase covered lives and better manage costs.


    Jim Walker has been kind enough to contribute to my blog several times now. I've previously (in an earlier blog) introduced him. Once I read his latest article, I'll probably comment on it. In the meantime, here's the link and more!

    Health Affairs, 28, no. 2 (2009): 467-477
    doi: 10.1377/hlthaff.28.2.467
    © 2009 by Project HOPE

    From Tasks To Processes: The Case For Changing Health Information Technology To Improve Health Care

    James M. Walker and Pascale Carayon

    To deliver better health care at a lower cost, health informationtechnology (IT) should be redesigned to support improved, patient-centeredcare and not the isolated tasks of physicians and clinicians.This new approach has major policy implications: health IT canhelp mitigate the worsening shortages of physicians; it willrequire managers, clinicians, and patients to learn new skillsand behaviors; it will increase the need for clinically astutesystems analysts, business-process managers, and human-factorsengineers; and it will highlight the need to pay for processimprovements and improved patient well-being rather than themere purchase of health IT.

    For those who haven't notice, Health Affairs coverage of Health IT Systems has dramatically picked up; the entire last issue had this focus

    I've now had the benefit of attending the HIMSS/AMDIS Physician IT Symposium, where this topic was discussed several times from the perspective of Physician Executives who are reporting on their real-world experience in this space.

    Notable among them:

    Physician's IT Symposium (HIMSS/AMDIS, focused on CMIOs and Physician Adoption and Clinical/Mgmt Benefits)New Tools for the Medical Home by Steve  Clemenson, MD


    The Reality of Pay-for-Performance by Scott E. Hessen, MD)

    Other powerful presentations from this symposium listed in subsequent post.

    [I'm sharing this detail of presentation and speaker names, recognizing that you can purchase the PDFs and Audios for either, both, or all presentations I reference.  I have no relationship to disclose with; they have a table in the South Hall and offer this commercial service.]

    I also had an opportunity to curb-side the medical director of one of the largest public health systems, on the topic of this post.

    In addition, I went to the microphone in the general session and asked the five thousand physicians (okay, I'm exaggerating, I'll change that number if Bill Bria or Rich Rydell choose to clarify for me; I've emailed them both on this.)  I pointed out the above CBO estimates and asked if anyone knew of other relevant, published estimates or health plan experiences.

    Bottom line is this - there are saving from managing various conditions.  The CBO is correct in their introductory chapter of the previously cited December 2008 report, that 

    The estimates in this volume rely on CBO’s current anal-
    ysis of and judgment about the response of individuals,
    firms, and providers to changes in the health care system.
    These estimates are informed by the available economic
    and health research literature and by CBO’s own work
    and analysis in this area. Many options rely on CBO’s
    Health Insurance Simulation Model to generate the esti-
    mate.1 When possible, CBO has estimated an option’s
    impact on the number of individuals (presented as an
    average for a particular year) who have health insurance
    or who would participate in federal programs.

    Every option involves some trade-off or redistribution
    between parties, whether individuals, businesses, provid-
    ers, or the government (state or federal). Some options
    would yield significant federal savings but would repre-
    sent dramatic changes for individuals or providers—for
    example, reducing the cost of a federal program but at the
    expense of its beneficiaries. Conversely, spending options
    might expand access to health insurance but do so ineffi-
    ciently or with secondary effects on the system that
    policymakers might find unacceptable.

    Based on the input I've received, for diabetes and just for previously identified diabetics, real savings start in about year four.  I would not be so bold as to extrapolate the national implications.  I'm glad that we have the CBO's Health Insurance Simulation Model referred to above.

    Per the input of Dr. James Walker above, more data on other specific conditions is forth coming.

    Per other input, there is data from other centers coming shortly on Asthma.

    If "Meaningful Use" is in the context of Meaningful Results, and if we're trying to professionally assess expected economic impacts of that Meaningful Use, clearly this kind of approach will be essential.   Although some of the data to draw these estimates exists, the kind of assumptions laid out in this post should, in my opinion, should be explicitly stated for any Meaningful Use definition.

    Thanks for your post. But you have to admit it's a bit more complex than you've allotted to your words thus far. And that includes Jim's comment and your response.

    There is a point in time where (and it's here) where're we are going to need experts like you to flesh out the cost-benefit relationship of EMR/EHR. Those of us who are ignorant in terms of how to "pitch•bCrLf the EMR/EHR proposition in the broader sense, may not be able to do so using a simple Diabetes example . . . simply because a lot of other people with maladies just want to stay alive. Could you please elaborate?

    Is the "cost•bCrLf of ARRA worth the "cost•bCrLf in terms of patients? How do we not use the PR people to spin this so that we can efficiently and cost-effectively implement EMRs to achieve a universal/ubiquitous EHRs from which we can all benefit?

    I would greatly appreciate your input and expertise. I think you must admit this is all quite perplexing.

    I've reviewed you blogs for quite some time. Anthony has called you one of the smartest people he knows. You don't need to give us references or links, give us your guidance! Thank you.

    Thanks Jack, both for the kind words and bringing me back to reality!

    In short, you're absolutely right. Healthcare IT is complex. And, that's in the context of how we actually deliver care, and our healthcare policy.

    These topics are necessarily complex, and, for everyone I know, a bit overwhelming.

    I'm only going to refer to the CBO green text above - I can't say it any better.

    If you only read one sentence, this is the one:

    "Some options would yield significant federal savings but would repre-
    sent dramatic changes for individuals or providers"

    I think that's an understatement. All options will have effect, as well as consequences on Quality and Safety. Some predictable, others not. The ARRA is not a pilot project.