A Difficult Time
My Journey in Medicine	by Jerry Sobieraj, MD © 2001
Available at and Published by
Also available at:

Chapter 10: The Bean Counters

Documentation, documentation, documentation, was the mantra of the bean counters. After all, they were the ones who reviewed our office notes, and decided if the notes matched the billing codes we used for our office visits. One might think doctors would be careful in their documentation so that if they were sued, they could adequately defend themselves. This rationale was far from the case. Doctors documented according to standards set largely by the American Medical Association (AMA). The AMA developed a standard for coding services (Comprehensive Procedure Terminology, CPT). The fourth and current version, CPT-4 was released in 1992. This version was supposed to help physicians because it took into account the time and complexity of the care doctors provided. It also codified a documentation standard for reimbursement, which has been subsequently used by the bean counters to determine if the level of service we have billed has been adequately supported by our documentation in the office record (i.e. a patient’s chart).

The standards in CPT-4 were quite simply written, and thus, could be broadly interpreted. However, to interpret them accurately, one would need to be a physician, as an untrained person couldn’t discern whether a documented service met the medical complexity standards of CPT-4, or if the medical reasoning required to arrive at an assessment and plan of care was straight forward or not (this was another coding variable in CPT-4). Yet, such a monitoring system would be too expensive. An insurance company couldn’t afford to use MDs to audit claims to see if the documentation (and thus, the level of service provided) warranted the charge submitted. To avoid incurring such a cost insurance companies and corporate medicine devised another plan.

The clever answer of health insurers, which of course was endorsed by the fiscally conservative AMA (of which I have proudly never been a member), was to divide a typical office note into its component parts. Then the bean counters could use a relatively simple formula to count up the documented components, and translated this into a procedure code for the service provided. This might sound simple, but it was dauntingly complicated. This complexity of coding was not a trivial matter, but it was assumed that if you could count beans, you could count up the number of components documented in a note, and match it to a list that defined the appropriate billing code. Unfortunately, these documentation requirements have left countless clinicians spending ever increasing amounts of time documenting what they did. This in turn added to the difficulty of making time to actually talk to a patient during a visit, and thus, ensure that their issues were adequately covered.

The dogma of the bean counters was, “If it wasn’t documented in the chart it didn’t occur” (and thus, couldn’t be billed for, even if the service was actually rendered). Think about the implications of this. How effectively can you remember the nuances of your work day, such that after a 4 hour period, you could write down sufficient detail that included minor comments you made 2-3 hours ago? Furthermore, noting (in the patient’s chart) a comment made hours earlier, could determine whether you were paid $50 for the visit or $75. The value the customer (patient) received wasn’t affected by this, only your reimbursement. The care you provided a patient was completed with respect to the service you billed for when they left your office, however, your worth financially was based on how well you wrote it down afterwards. And this of course was the bottom line of the bean counters.

In reality, one could bill any level of service (i.e. code) one would like. However, the insurance companies had computer programs that identified billing practices that tended to “over bill” (that is, exceed the average level of service billed based on your specialty, location, etc.) Complicated stuff, eh? A health insurance company might consider a pattern of over billing if a practitioner consistently billed at the highest level of service, while the diagnoses used to justify the service were too trivial to support such extensive evaluations (e.g. a complete physical billed for every common cold seen in the office). After all, when a provider agrees to accept reimbursement from a health insurance company, he or she also agrees to their right to audit the records of their patients at any time. The health insurance company could decide after the fact, whether the documentation met the standards that currently existed (and not necessarily the standards that existed when one wrote his or her office note). This latter issue was a real concern, as Medicare had used the FBI in the late 1990s to investigate the billing practices of several AMCs. They used documentation standards developed in the 1990s to evaluate the adequacy of visits documented 1-2 decades earlier when documentation standards were rudimentary at best.

Medicine has a long tradition which has fostered predictability. This stability would be attractive to physicians who were risk aversive, and who wouldn’t want to take a chance with something like documentation. One’s inclination might be to exhaustively document patient visits to minimize the likelihood that an insurer would need to be paid back in the event a future audit suggested overcharging. The toll this takes on a practitioner who sees dozens of patients daily isn't trivial. A practitioner has at least 5-10 minutes of paperwork per patient visit. This is time required just to write office notes, with the primary focus of the notes to meet the standards of third party payers (i.e. insurance companies). These documentation requirements didn’t mean that the notes had more relevant information, or would help save a life. In fact, they often led to inordinate cluttering of information so the provider could be sure to include all points relevant to the billed level of service (i.e. trying to list as many of the billing components counted by the bean counters as possible). I would spend the equivalent of an eight hour day documenting my office visits each week. These eight hours didn’t include time I used for documenting phone calls, nor letters I sent to patients, nor the development of additional treatment plans based on data received after a patient’s visit.

This time commitment of documentation was exasperating, and served no purpose to benefit the patient, the doctor, nor the office staff. It only helped insurers decide if the incredibly low amounts they paid us for providing primary care were justified. One of the more distressing elements of this story has been the assimilation of these documentation practices by the physician leaders at major medical centers. The Chief of Medicine at my institution actually created a Compliance Office shortly after taking control in the mid-1990s. The people in the Compliance Office were to monitor our documentation to make sure it met third party standards. It was as though Big Brother had moved in next door. In addition, inordinate amounts of time were spent at our business meetings talking about the nuances of coding a visit for optimal reimbursement. The leaders have taken to groveling for every dollar they can get from insurance companies. The impact of this mentality, and the toll it took on the medical staff seemed to be viewed as inconsequential, but in reality was substantial.

Another aspect of documentation that was particularly bothersome to me was that no such standards existed for many non-clinical medical specialties. Radiologists, for example, had no standard way of documenting their interpretation of a chest x-ray. An x-ray report could state, “no interval change since the prior examination of (the prior date)”. This type of information was not only not helpful, but provided no objective evidence that the x-ray had been adequately evaluated. Why in primary care was it assumed that it didn’t happen if it wasn’t written down, whereas radiologists could omit all relevant detail, and we were supposed to assume it was read thoroughly and correctly. Not infrequently, I had to look at an x-ray myself to make sure the specific question I had asked was answered, as the radiologist hadn’t addressed it in the written report (e.g. a report of a shoulder x-ray stating there was no fracture of dislocation, but my question was, “Was there any arthritis?”). If standards did exist for the required elements in an x-ray report, they were not routinely implemented. I do not suggest that we should ask insurers to create an onus similar to that on primary care providers for radiologists also. I only wanted to stress the burden of documentation on primary care providers, and that similar documentation requirements did not exist across the board.

Standards and minimum requirements of documentation could be useful, but excessive detail required to facilitate monitoring of primary care services made no sense at all. There are times when a physician orders a test, but due to unexpected factors, the information resulting from the performed test is quite limited. A common example is an echocardiogram of a very overweight person. Changes in the chest wall due to the obesity prevent accurate measurement of some heart chambers. Thus, the effort required by the cardiologist to document the test result is less than expected, because only limited information resulted from the test. In such a case, the billing should reflect the deficient status of the data, and thus, the limited ability to make an interpretation. What I observed in such a case was a one paragraph echocardiogram report noting limited data, yet it was billed and compensated the same as one that comprehensively assessed all aspects of heart structure and function.

Office visit documentation requirements need to be lessened. As a quid pro quo, physicians would have a responsibility to not charge excessive amounts for simple, easy-to-provide services (there are some things that are easy to do which pay well, such as giving anti-inflammatory injections). Yet if anything, documentation practices have worsened. I received a notice from the Department of Health and Human Services dated 1 June 2000. It comes from Nancy-Ann Min DeParle, administrator at the Health care Financing Administration (HCFA), the people who bring us Medicare. She notes, "Today I want to emphasize the importance of close attention to billing requirements, especially for documenting services delivered and the reason for care, as a way to ensure you receive and Medicare makes proper payments".

She added, "For physicians, we will be focusing this year on two CPT codes used to report evaluation and management services-99214 and 99233. These codes accounted for a significant portion of the coding errors in the last two audits. In fact, documentation for many of these services was only found to be sufficient to support services more appropriately described by CPT codes 99212 and 99231". To translate, 99214 refers to a CPT-4 billing code for a detailed office visit. CPT, as described above, was devised by the AMA and insurance companies to permit standardized billing by physicians and Health care institutions. 99214 is sometimes called a level "4" service, as there were 5 potential codes used to bill an office visit with an established patient (defined as a person treated in the prior three years). 99214 is the fourth highest paying of the five visit codes. The memo implies that a number of claims billed at 99214 did not have adequate documentation to support that level of service (i.e. should have billed as a level 2 (99212) or level 3 (99213)). We don’t really know if there was an excess of level 4 visits billed, as not documenting enough data points for the bean counters may have been the only problem. Yet even if there were cases of billing for a level 4 visit when a level 2 service was really rendered, it is a different issue (an office visit for level 2 would be a straightforward common cold in a young person, a level 4 visit would require assessment and documentation of at least two significant medical issues such as heart disease and diabetes). The amount of effort expended to perform and document a level 4 visit (30-40 minutes) was excessive.

A level five visit (99215) is used for complete examinations in patients who have multiple medical issues (e.g. congestive heart failure and diabetes). In the Boston area in 2000, Medicare paid $95 for such a visit. This type of visit requires at least 40 minutes with the patient, if not an hour, and another 10-15 minutes to write a note to document the service provided. This latter 10-15 minutes was the most important in terms of compensation, yet, the patient likely considered the first 40-60 minutes most important, as that was when the care was actually provided. However, the emphasis has been placed on what we document in that latter 10-15 minutes, as it determines how much we received and thus, our productivity.

At business meetings for the General Medicine Section of which I was part, we regularly discussed coding. We would review the myriad criteria used by us to decide the level of service (e.g. level 2, 3, 4 or 5) we were providing (and presumably documenting). People from the Compliance Office would review billing criteria with us to help us maximize our billing. Unfortunately, coding our visits had evolved into a game. It was hard to imagine some of the questions my fellow doctors would ask at these meetings. For example, "If I ask them (the patient) if their mother had diabetes, does that count as family history?", as family history gave a physician points with the bean counters when one tried to "maximize" their reimbursement.

The most perverse outcome of these billing requirements is the indirect incentive given to physicians to prescribe medication, with all their potential for side effects and cost, as opposed to using non-pharmacologic means. That is, we actually have incentives to prescribe a drug to treat a problem as opposed to using life-style modification, because it permits us to bill at a higher level. This is quite unfortunate, as there are multiple impediments to using life-style modification in clinical practice. The first disincentive was the time it took to educate the patient. The first step was to get the patient to agree that there was a problem that required therapy (this occurred when prescribing medications also). Then a physician would need to apprise the patient of their treatment options, and how to manage a life-style change. In addition, the physician needed to give the patient tools to accomplish and monitor a life-style change, setting specific goals (just as one might also do with medication). Yet the goals were not just numeric (e.g. lower the blood pressure under 140), but also behavioral (e.g. walk a total of 10 miles a week). People who had mild hypertension (high blood pressure) and were inactive, didn't just jump up off the couch and get moving. They needed specific direction on how much exercise to start with and realistic goals to accomplish before a return visit a month or two later. Patients also needed to learn how to avoid becoming overly enthusiastic, and end up with all or none behavior (e.g. either they were very good or very bad).

When faced with counseling a patient about a behavioral change to manage a problem versus prescribing a medication, one could bill based on time spent face to face with the patient. A clinician needs to spend at least 25 minutes discussing these issues to bill at a level 4 (99214) or 15 minutes face to face to bill a level 3 (99213). If a clinician spent 25 minutes talking to a patient, he or she would get $60-70 reimbursement from a third party payer for the 99214 you billed. Many physicians have found it more productive to see two level 3 (99213) patients during this 30 minutes (15 minutes on average per patient), and receive over $100 in compensation. However, each 15 minute visit meant in and out. There would be little to no time to discuss details such as what the patient was eating, what environmental factors could be contributing to his or her condition, and/or how the patient would become more active.

How does the physician keep the visit to 15 minutes and still treat the condition? The physician prescribes a medication. It is quicker, and for many conditions, it often works. Yet, if time alone wasn’t enough incentive for a physician to prescribe a medication, the bean counters made it even better. When a provider prescribes a medication, the visit automatically rises from “low complexity” to "moderate complexity" (as discussed earlier, medical complexity is a billing component counted by the bean counters). The visit is considered low complexity when the clinician just talks about life-style modification. The levels of medical complexity established by CPT-4 are: straightforward, low, moderate and high. Straightforward implies that all physicians seeing such a particular clinical situation will in all likelihood arrive at only one outcome. (e.g. a person with decreased hearing due to wax in their ear; remove the wax and they will get better). A moderate complexity visit means if one’s documentation is at level 3 in terms of historical information and physical exam (i.e. you have enough pieces of data in your note to satisfy the bean counters for a level 3 visit), the clinician may upgrade the coding to a level 4 visit just by prescribing a medication. However, if a physician does the same history and physical and decides that using the DASH diet (the standard diet for lowering blood pressure) should be the first approach in managing a person's blood pressure, the visit remains a level 3. Thus, prescribing a blood pressure medication in this case increases a physician’s reimbursement for the visit by $15-20.

Documentation needs to be simplified so that primary care givers can record the core data needed to provide patient care. If we shift reimbursement to encourage physicians to talk to their patients by paying for time spent face to face, we can decrease the time spent documenting and learning about documenting, while enhancing the care we provide. This would require a major overhaul in current thinking. At the beginning of 2000, the Health care Finance Administration (HCFA), who manages Medicare for the US Government, put on hold new documentation requirements, which had been established in collaboration with the AMA. Believe it or not, they were much more onerous than those I describe above (which you surely wasn’t straightforward)! Thus, as with many things in Health care, it doesn’t appear we are heading in the right direction.

Table of Contents