Tim Terrell, Managing Partner


For a brief time, it looked like the movement to value-based care was accelerating rapidly across the nation.  And then reality hit.  In 2014, only one fourth of all the Medicare Shared Savings ACOs received a shared savings payment.  Transforming an industry is harder than it looks.


Industrial Quality 101, as used in Lean and Six Sigma, teaches that there are only 3 ways to address quality.  I learned this using a rubber band powered catapult that launched small plastic balls toward an imaginary target.  My team’s first attempts were wildly inaccurate.  Then we began to apply process standardization to achieve a more consistent outcome.  We did test launches at different release points and with varying amounts of tension on the rubber band.  Slowly we narrowed in on even better results.  One of our balls was wobbly and inconsistent, so we removed it and got our best results.

It was a simple exercise but it proved out the 3 methods of improving quality and getting better outcomes.  First, standardize your processes.  Second, improve the processes.  And third, deal will non-conformance.

Mathematically we can think of this as a normal curve representing a range of outcomes (x-axis) and the number of observations at each level (y-axis).  When we standardize our processes, we lower the standard deviation of the range of outcomes.  This makes the normal curve taller and skinnier.  When we improve our processes, we lower the mean value and push the normal curve to the left.  And finally, no matter what we do, we may still need to address outliers and their causes.  A good example might be the HgbA1C values of diabetic patients.  We can close gaps in care and more closely monitor drug therapies to get a more consistent range of outcomes.  Then we start doing diet and exercise training for the patients and put them into group visit settings to help them empower themselves.  This should push the mean HgbA1C levels downward overall.  We will still have some non-compliant patients that may need individual counseling with a psychologist to get them to better outcomes.  Standardize.  Improve.  Deal with the outliers.

At each step along the way, information plays an important role.  You must identify the patients who are diabetic, and know which ones are not adhering to testing regimens or drug therapies.  You will want to maximize your new resources by targeting new approaches at the right patients.  And you need to follow your outcomes over time to make sure that process improvements are maintained.

This example illustrates the problem of transforming the health care industry.  Process change is involved and it is expensive.  In the health care industry, process change doesn’t happen by tweaking a machine, or standardizing an assembly line.   In health care, process changes must happen in thousands of ways across hundreds of disease states.  It often requires the application of human resources in new and inventive ways.  Care coordinators, nurse navigators, call center workers, dietitians, pharmacists, psychologists and case managers are becoming part of the new care team.  The health care industry is poorly equipped to handle the massive complexities.  We are beginning to think about populations of patients, but we still see them one at a time using a workflow process forced onto us by our electronic health record systems.

When I led an effort to put EHR into a large health system in 2005, our EHR system enabled a common workflow:  the patient comes to the waiting room and is checked in, then gets put in a room by the nurse, then is seen by the doctor, and finally checks out.  Some of our special clinics or procedural rooms or group visit sessions didn’t work this way and we had to create work around processes that would satisfy the new EHR system.  When the EHR forces us to use one process flow the tail is wagging the dog.

There are many signs that the standard workflows of the past may not be optimized for an industry in need of process transformation.   I recently ordered a Jeep Wrangler from my local dealer.  Since we couldn’t find the exact one I wanted, we built the order on-line.  This order kicked off an automated process of reconfiguring the automaker’s assembly line so that the car I wanted, in the “sunset orange” color I preferred, was built to my specifications.  I received it in a few short weeks and I was very pleased.  I named her Myrtle.

In health care, we need but don’t yet have this type of adaptive process.  We have trouble making a Myrtle.  Our information systems should know if a patient has drug compliance issues that need a pharmacy consult, or if the patient is under a care manager’s protocol.   The office visit “assembly line” should be configurable to the particular needs of the patient.  Should the patient see the nurse and doctor as has been standard for many years, or should we steer this patient to other care team members during the visit?

Do we even need this patient to come to our office, or should we meet the patient where they are, in their homes or at their workplaces?  E-visits and remote monitoring are a few of the ways this is happening today.  Systems for eVisits and remote monitoring exist, but most are poorly integrated to the EHR record.  We have barely skimmed the surface of what automation can do to create a 360-degree view of our patients.  The future promises such things as analysis of food buying habits against known medical conditions, genetic testing to maximize drug therapies by patient, and analytics of daily exercise showing how exercise contributes to better outcomes over time.

I believe that our EHR systems need to be redesigned in the following ways:

1)   Analytically informed doctor-patient relationships

  • Chronic conditions should be understood as continuums with known treatment pathways to follow along this continuum
  • Therapies must be understood and judged for effectiveness across types of patients
  • Lack of adherence to standards must be rooted out
  • Care pathways must be enabled with the right information reaching the right care team member at the right time
  • Doctors must understand in real-time where they trail their peers in quality and how the patient in front of them represents an opportunity to improve their quality metrics
  • Patient risk profiles must inform care decisions

2)   Adaptive visits

  • Patients should be able to control their own access to the visit using on-line self-scheduling tools or eVisits
  • Each patient visit should be automatically configured for maximum benefit across a range of care team members based on what is known about the patient
  • The visit should be configured to achieve maximum results, not maximum billing as in the past
  • Real-time interactive educational materials should be available on common platforms such as smartphones to help the patient understand their physiology or disease state

3)   Works smoothly across care settings

  • The EHR must not be a tool solely suited for care mainly in a doctor’s office setting
  • Access from home, SNF, hospital, or patient work settings should be easy and natural for a range of providers
  • It should support eVisits and on-line chat
  • The EHR needs to present the right information for the setting of care
  • The EHR should be configurable to the care team member

4)   Patient interaction and engagement

  • Patient views of the EHR record should be allowed
  • Interactive educational tools should be delivered to patients on mobile platforms (smart phones)
  • On-line interactive scheduling, with algorithmic pre-assessments, inform the visit
  • Patient chat rooms and social communities connect patients to one another
  • Patient alerts from remotely monitored data avert problems

5)   Smooth information interchanges

  • Uploadable patient data is synthesized and analyzed to make it useful
  • Genetic information is integrated into treatment decision trees
  • Best practice treatment protocols and order sets are integrated into the visit
  • Alerts from claims systems about precertification for hospitalization and high cost procedures occur in real time
  • Alerts and workflows around transitions of care setting prevent errors of omission

6)   Smooth application interconnection

  • Quick launch of third party applications with the ability to exchange data nimbly creates extensibility of function
  • Easy addition of new applications and technologies (without expensive development to connect) supports on-going innovation

The above list is all about using process change to standardize and improve the health care industry.  It is about using the EHR to enhance these process changes.  It’s about making our health care industry like other industries that have figured out how to make a perfectly configured sunset orange Jeep on demand.

Current EHR systems are ill prepared to support these types of changes.  They are stuck on old platforms, or are too busy just keeping up with Meaningful Use.  It is time for some enterprising entrepreneur to rethink and redesign the EHR system of the next 25-year cycle of care delivery.  This new EHR will be process driven, adaptive, and analytically informed.  It must be because there are only three ways to improve quality no matter what the industry.

Medical groups are embarking on the hard work of value-based care and the management of populations.  This is leading to an explosion of technology-based solutions and apps.  There are some really good ideas and products being built.  However, these systems are largely fragmented and disconnected and physicians balk at using them if they aren’t EHR connected.  It is time for someone to reinvent the EHR.  A process-driven EHR would create great value in the new process-driven health care industry.  How about you?  It can’t happen too soon.

Tim is Commonwealth Health Advisors’ expert in population health, risk contracting and value-based care.  He is the former CIO of Cornerstone Health and most recently served as the CIO of CHESS, a managed services organization dedicated to helping other medical groups make the move to value-based care models.