We have a physician utilization problem

Do we really have a looming physician shortage? We may, but even more acutely I believe we have a physician utilization problem, most particularly in primary care. After shadowing approximately 50 primary care physicians across the country and engaging physicians in conversation during 150 or so presentations on improving the delivery model of care, my observation is that 70-80% of the PCPs work output is direct waste: computer order entry, prescription processing, composing the billing invoice, filling out mammogram requisitions including typing in the date and location of the last mammogram, working an inbox full of random notifications and disconnected results, typing the visit note.

In many ways physicians have become the data entry clerks and data processors for convenience of back end auditors and distant researchers. And because most of this work is “off-the-clock” (there is no additional pay when the physician takes on work previously done by receptionists, transcriptionists, medical records clerks and pharmacists or new work created in the service of big data initiatives, such as ICD-10) it is free to the organization’s bottom line.

What percentage of time are health professionals optimally using their skills, i.e. working at “top of license” (TOL)? My guess is that at best only 20-30% of a primary care specialist’s day is spent on direct clinical care. It would be a fruitful to study top of license time (TOLT) in what I propose as a new field of practice science. (We spend billions of dollars each year on tests and treatments, but almost nothing on the basic science of the delivery model for those tests and treatments.)

TOLT study will help identify the hidden waste within the health care system, and begins to address my suspicion that we don’t so much have a workforce problem in primary care, but rather we have a workforce utilization problem.

Secondary questions follow: How does “top of license time” (TOLT) vary across physician specialties? My hypothesis is that TOLT is lowest in primary care, highest in procedural specialties, and generally tracks with RVU generating capacity across specialties.

What lives in the space between TOL of a nurse practitioner and a physician? This is a challenging question to raise, but explicitly considering it can shed light on the optimal use of MDs and NPs in healthcare.

How to define and measure TOLT? This is key, because many tasks that are commonly assigned to physicians do not need a physician’s level of training, but rather a skilled team member. I focus here on clerical tasks that are not TOL. Defining clinical tasks as not TOL may also be useful, but is tricky, because the sum of doing many lower level clinical tasks (managing blood pressure, evaluating acute knee pain, following up on depression) may be a strong patient-physician relationship that makes doing the highest level, highest value clinical work possible.

How TOL is defined will determine the utility of the findings. For example: prescriptions. A physician can make the decision with the patient to add a blood pressure medication. Communicating that decision to an assistant, such as an RN, takes 3 seconds. Putting that order into the computer and sending to the pharmacy make take one minute. Multiplied over thousands of prescriptions/renewals per year and that is a lot of waste. Deciding on the medication is TOLT, progressing through a series of drop down boxes to process a prescription is not.

What else is not TOLT? An OB/GYN in Michigan reports that her staff used to advise newly pregnant women about multivitamins; now the physician must now create an electronic order for multivitamins for each patient. A family physician in Wisconsin would previously simply ask her staff to do an ear wash for a patient, now she must create an electronic order for the ear wash.

Across the country physicians are spending a minute or more per patient creating the billing invoice, a task that previously took the physician a few seconds. Physicians at every task are taking on extra time to save a data entry clerk time.

In our state there is a new requirement that physicians complete the death certificate online. This sounds reasonable until you consider that allow it saves a data entry clerk time, this adds 7-12 minutes of physician time to the act of signing off on a death certificate.

A physician wrote recently “I am often working on charts at 4am until clinic and then after clinic until 10pm.” One large organization has data that 30% of their primary care physicians’ workday occurs between 7pm and 7am, and is spent doing EHR work (inbox and visit note documentation). Another organization has data that only 20% of their PCP’s 11 hour workday is spent with the patient. (And available data would suggest that at least a third of that is actually attending to the computer, not the patient.)

It could be that everyone is working at the top of their license except the physician, who is now spending most of her days at a desk, filling out paperwork and providing electronic signatures for patients she barely knows. In no other industry is the highest trained professional spending most of her time on work that others could perform.