The Problem With The Problem List: when electronic medical records go rogue

Prior to the 1960’s, medical records were largely illegible, unorganized, and non-uniform. It was difficult for a doctor to keep track of his/her own patient’s histories and nearly impossible for a doctor to tell what was going on with a patient by reading another doctor’s notes. Enter Lawrence Weed, MD, a Professor of Medicine at the University of Vermont who invented the S.O.A.P note and taught generations of medical students to organize their hospital notes into the 4 sections of Subjective, Objective, Assessment, and Plan. He was also the father and champion of the  Problem-Based Medical Record, a way of organizing and filtering the various signs, symptoms, and diagnoses that a patient has in order to coordinate the diagnostic effort and to provide optimal longitudinal care of that patient.

When the Center for Medicare and Medicaid Services (CMS) wanted to encourage the widespread use of electronic medical records, it rolled out the “meaningful use” program as part of the HITECH Act of 2009. Physicians who adopted electronic medical records with several provisions were eligible for incentives of up to $44,000 to help offset the cost of purchasing and implementing an electronic medical record. Those electronic medical records had to:

  • Use computerized order entry for medication orders.
  • Implement drug-drug, drug-allergy checks.
  • Generate and transmit permissible prescriptions electronically.
  • Record demographics.
  • Maintain an up-to-date problem list of current and active diagnoses.
  • Maintain active medication list.
  • Maintain active medication allergy list.
  • Record and chart changes in vital signs.
  • Record smoking status for patients 13 years old or older.
  • Implement one clinical decision support rule.
  • Report ambulatory quality measures to CMS or the States.
  • Provide patients with an electronic copy of their health information upon request.
  • Provide clinical summaries to patients for each office visit.
  • Capability to exchange key clinical information electronically among providers and patient authorized entities.
  • Protect electronic health information (privacy & security)

And so, the problem list because a necessary part of every electronic medical record program in the country. On the surface, this sounds like a great idea – an easy way for multiple physicians using the same patient’s chart to quickly and easily see what is going on with that patient. However, the problem list rapidly morphed into a monstrous of list of often redundant words that became unusable for their originally intended purpose. So what happened to the innocent problem list?

Electronic medical record (EMR) companies designed the EMR so that the problem list was the central repository for all lists used in the chart. Elements of the past medical history fed into the problem list as did billing diagnoses, admitting diagnoses, and discharge diagnoses. Similarly, the problem list could be used to select the billing diagnoses, admitting diagnoses, and discharge diagnoses. Because the problem list was used to assign the billing diagnoses, all 69,000 diagnoses in ICD-10 had to be selectable in the problem list. And thus all of the weird and obscure ICD-10 codes suddenly became available to add to the problem list (such as V91.07XA: Burn due to water skis on fire, initial encounter).

What Is A Problem, Anyway?

Medical diagnoses are clearly problems. But there are other elements that are not so clear. Should the family history be included in the problem list? What about the social history? Should the patient’s list of surgeries be added to the problem list? What about non-specific symptoms – should the patient with I50.23 (Chronic systolic heart failure) also have orthopnea (R06.1), history of cardiomegaly (Z86.09), and pedal edema (R60.0) on the problem list?

Duplicate Problems

Each physician who sees a patient might use a different word to describe the same thing and this led to multiple duplicate symptoms on patient problem lists. For example, if a patient with breathlessness gets admitted to the hospital, the emergency room physician might add add the ICD-10 code “R06.00: Dyspnea“, the hospitalist might use “R06.02: Shortness of breath“, and the cardiologist might use “R06.09: Dyspnea on exertion“. By the end of the day, the patient’s problem list will include all three symptoms: dyspnea, shortness of breath, and dyspnea on exertion.

Multiple Layers Of Specificity

Similar to duplicate problems, many diseases can have many different ICD-10 codes for different degrees of specificity of that particular problem. And so “type 2 diabetes” can have a general code and also have many different subcodes, each of which gets its own place on the problem list. For the patient with diabetes who is seeing a family physician, a nephrologist, a cardiologist, an ophthalmologist, and a vascular surgeon, there can be 15-20 different diabetes-related problems on the problem list.

Immortal Problems

Unless someone actively cleans up the problem list, then signs, symptoms, and diagnoses that have resolved can stay on the problem list, causing further clutter. So, Right arm bruise, initial encounter (S40.021A) from an office visit in 2011 persists in perpetuity with Runny nose (R09.89) from an office visit in 2015. Similarly, the patient who had Hyponatremia (E87.1) during an office visit in June and then had Hypernatremia (E87.0) in July can end up with both hyponatremia and hypernatremia on their problem list. Old diseases don’t go away, they just stay around forever on the problem list.

Whose Problem Is It?

Most would agree that the primary care physician should be permitted to add problems to the problem list. But what about the specialist, the emergency room physician, or the hospitalist? Should advanced practice providers such as nurse practitioners, physician assistants, nurse anesthetists, or midwives be allowed to edit the problem list? What about nurses, pharmacists, psychologists, and dietitians? There is universal agreement about who the stewards of the problem list should be.

Don’t Touch My Problem

Many physicians are very possessive about their additions to the problem list and can get angry if another physician deletes one of their problems without talking to them. So, the patient who was rude to the office staff at the surgeon’s office might have gotten Negative attitude (F60.2) added to the problem list; when the same patient was a delight to the staff in the endocrinologist’s office, the endocrinologist might have deleted F60.2 from the problem list and then the surgeon later gets mad because he wanted that reminder to the staff that the patient was a handful during the last office visit.

Forcing Review Of The Problem List

For many electronic medical records, an outpatient visit cannot be closed until the physician attests that he/she has reviewed the problem list. Typically, an alert box will pop up when trying to close the encounter reminding the physician to review the problem list. By that time, the patient has left the office and the physician just wants to get done with the day’s charting so the physician will typically scroll down to the bottom of the problem list and click “problem list reviewed”, without actually looking at the problem list. Problem list fatigue is just as much of a challenge as alert fatigue in our offices and hospitals.

More Is Less

In the spirit of Dr. Weed, some physicians will organize their progress note in the electronic medical record by importing the patient’s entire problem list into their daily office note. This can result in progress notes that are the length of a small novel making the note unnavigable to any reader and making it more difficult for other physicians to figure out what the author of the note was actually trying to say.

Too Many Problems

Every physician who sees a patient adds her/his problems to the problem list. When I see a patient whose primary care physician and other specialist physicians use a different EMR than I do, I am the only person adding to the problem list and so that problem list is small, relevant to my practice, and manageable. However, the more physicians using a common EMR a patient sees, the longer the problem list becomes. List of 30, 40, or 50 problems are not uncommon for patients with several chronic medical problems seeing multiple specialists or for patients with several hospitalizations.

Erroneous Problems

Errors in the problem list are rampant. For example, a patient fills out a pre-visit past medical history form and checks the review of systems box for “impotence”. The nurse reviews the form and transcribes the checked boxes into the past medical history and accidentally types the first 4 letters as “impr” rather than “impo”. The first diagnosis that comes up under “impr” is imprisonment and so it gets entered into the past medical history that then feeds into the problem list and now “imprisonment” shows up on the patient’s problem list. There are a myriad number of ways that erroneous problems find their way into the problem list and once they are there, they often stay there indefinitely.

High Maintenance Costs

The currency of work effort in ambulatory medicine is the number of mouse clicks necessary to perform any given task. A typical office visit lasts about 15 minutes and there is a lot of things that have to be done in those 15 minutes. Reviewing and editing the problem list costs time and mouse clicks and usually falls down to the bottom of the priority list of things to accomplish during those 15 minutes. Spending an extra 30 seconds per patient to truly review and update the problem list means that over the course of the day, that doctor will see one less patient than he/she otherwise would have seen. One solution that has been suggested is to partner with the patient so that the patient reviews the problem list to help identify resolved or erroneous problems. However, trying to explain what “suprascapular entrapment neuropathy of left side” means to the patient with a sore shoulder can often add 2 minutes to that office encounter.

“Need for…” Diagnoses

Some tests and vaccinations cannot be ordered unless they are associated with a “need for” diagnosis. For example, if you are going to order a flu shot, you have to use the diagnosis of “Need for influenza vaccine” (Z23). Some insurance companies will not pay for a bone density study unless the order is accompanied by the diagnosis of “At risk for osteoporosis” (Z91.89). If you want to see if a patient could have lead poisoning, you need to include “Need for lead screening” (Z13.88). These diagnosis codes often find their way into the problem list and do not add anything to the long-term management of that patient.

Diagnoses Required For Ordering Tests

Similar to the “need for…” diagnoses, many tests cannot be ordered unless the physician uses a specific ICD-10 code or one of a group of codes. These are so-called “allowable codes”. Many electronic medical records will present an alert box if a non-allowable diagnosis code is associated with an ordered test, such as a glucose level or a cardiac echo. The physician will then either keep adding diagnoses to the diagnosis list until she/he comes up with a diagnosis code that the insurance will accept or they will scroll through the list of allowable diagnoses in the alert box and pick a diagnosis code that is covered, whether or not it perfectly matches the real indication for that test. These diagnoses are necessary for the physician to order the test but may be irrelevant to the longterm management of that patient; nevertheless, the codes often end up on the problem list permanently. Physicians hate the alert boxes – they cost mouse clicks and time. Consequently, physicians, like Pavlov’s dogs, undergo operant conditioning and to avoid those alert boxes, they find a diagnosis that insurance companies accept for any given test and always use that diagnosis whenever they order that test. Thus, there are legions of physicians who always use “Other forms of dyspnea” (R06.09) for every cardiac echo that they order because they know that diagnosis works.

When used right, the problem list is a great tool for patient management. But a number of unintended consequences of electronic medical records have resulted in problem list monstrosities that are often irrelevant or a hinderance to medical care. As physicians, we are the main users of the problem list and it is up to us to truly make “meaningful use” of the problem list. We have met the problem of the problem list and the problem is us.

October 8, 2019