Re: Walmart Health: Just had a great dental visit this morning, which was preceded by helpful reminders from Epic, and…
By Andy Spooner, MD
“A 10-system Review of Systems was performed and found to be noncontributory.”
Billing compliance auditors get queasy when I put the above language in an electronic note.
Should they? I really did do a review of systems!
The documentation quoted above is not by itself non-compliant. The passive voice is used skillfully in the E & M coding rules to describe the complete review of systems:
“At least ten organ systems must be reviewed. Those systems with positive or pertinent negative responses must be individually documented. For the remaining systems, a notation indicating all other systems are negative is permissible.”
–1997 Documentation Guidelines
for Evaluation and Management Services
U.S. Center for Medicare and Medicaid Services
If you had — in your head — reviewed 10 organ systems during your history, then the above “noncontributory” statement would be compliant. So why the queasiness?
It’s worrisome because it seems like one of those “easier said than done” situations. It’s easy to insert a line of text that describes a complex thought process. But did you really think of all 10 organ systems? The solution to the queasiness is to call for a list of exactly everything you asked the patient about.
Why should we care how many organ systems someone thought about?
We have obligated ourselves (via the E & M guidelines to which we all subscribe in the U.S.) to show that we thought about multiple organ systems in the case of complex patients — if we want to get paid for complex care. There is nothing wrong with this concept, but then we have the problem of how to show that we performed this thinking. The most common way to indicate review-of-symptoms thinking — the method that seems safest to the compliance auditors — is the symptom checklist, where we enumerate everything the patient doesn’t have.
The irony of the checklist solution (there’s always irony when it comes to compliance) is that it tends to transform a valuable thought process (a physician’s internal review of his or her total knowledge of human pathophysiology) into a litany of irrelevant information that we care very little about. We see that performing this checklist process as being beneath us. We begin to care so little about “doing a review of systems” that we gladly detach this process from the act of history taking. We isolate it in several ways:
- We make it a separate part of the chart, as if reviewing systems can be performed independently of taking a history. It can be done separately, but why bother? Thoughtful coding consultants will tell you that the review of systems does not have to be a separate section, but even if it is embedded into the HPI, it still needs to be in some form where one can count “bullets” to assign to the canonical list of organs.
- We delegate this task (via the E & M guidelines) to absolutely anyone else who wants to “do” a review of systems. That’s not to say that a checklist produced by a medical student or nurse or a medical assistant isn’t accurate. The information is usually just fine. But unless it is integrated logically with the history of what is going on, what use is it?
- We gladly accept a patient-completed questionnaire for the information-gathering task. There’s nothing wrong with patient input, but if “doing” a review of systems is supposed to reflect the doctor’s thought process, how does a patient questionnaire do that?
- We work to ignore this separate blob of information. A study published recently by Clarke et al. on the information needs of ambulatory physicians suggests that the review of systems is usually regarded as superfluous — part of the noise. I get feedback from referring physicians that the thing they would most like omitted from letters sent to them by consultants is the review of systems (followed quickly by the past/family/social history and physical exam).
Some medical students buy laminated cards that spell out a review of systems in the form of a giant checklist. The result is what you’d expect:
ENDOCRINE: No blood sugar problems, cold intolerance, growth excess, heat intolerance, abnormal hair growth, impotence, increased thirst, increased appetite, frequent urination, skin discoloration, sweating, excess thirst, increased urination, or weakness
I always love to see that in the chart of a four-month old with bronchiolitis.
What’s going on here? Is it a bad idea to review a patient’s systems? Of course not. The goal is to make sure that we think of disease processes that fall outside our preconceived notions of what the patient has. Since all that wheezes is not asthma, the skilled clinician wants to be sure not to miss one of those unusual causes like a bronchial foreign body, vocal cord dysfunction, or cystic fibrosis. So why can’t the skilled clinician simply say that? We could even have the computer generate a differential based on documented findings, and then we could simply check a box that says something like “yes, I considered all of that.” (Or “yes, I considered all of that, and did not bewilder my patient by asking about cold intolerance or how many pillows he sleeps on because that’s just not relevant here.”)
E & M coding rules are based on the assumption that we are using paper, and that every additional bit of information we record costs us a little bit of energy. The argument goes that if we want to get paid more, we will be more willing to spend the energy to fill the paper with information in proportion to the complexity of the patient’s situation. With electronic systems, this calculus of documentation energy no longer applies. We can create long documents with very little energy. Since the paper-based rules assume a symptom checklist (paper is great for checklists), that’s what we make our electronic systems create for us. The subsequent “review of systems” is almost always meaningless.
If our purpose on reviewing systems is to assure that we consider broad possibilities in the diagnostic evaluation of the patient, why can’t our computer systems help us with that directly? We might be able to design our EHRs to be more useful if we could just let go of the assumption that the unit of analysis is the document, rendered as if on paper.
Ultimately, the rules about how to document are based on skepticism — perhaps a healthy skepticism — that we are going to do the intellectual work required to deal with complex clinical situations. This skepticism is here to stay, but the model of responding to it with bizarre lists of symptoms does not have to. Our clinical systems are capable of recording our efforts at creating a differential diagnosis. That intellectual work should count.
Andy Spooner, MD, MS, FAAP is CMIO at Cincinnati Children’s Hospital Medical Center. A general pediatrician, he practices hospital medicine when he’s not enjoying the work involved in keeping the integrated electronic health record system useful for the pediatric specialists, primary care providers, and other child health professionals in Cincy.