Data show that family physicians choose 99213 for about 61 percent of visits with established Medicare patients and choose 99214 only about 23 percent of the time for the same type of visit.1 So 99213 must be the correct code to use for a “routine” visit, right?
Not necessarily. Many of us may be shortchanging ourselves by reflexively coding a routine office visit as 99213 when the clinical circumstances of the encounter justify the higher-level code. We have developed coding habits based on the misconception that repetitive, routine clinical thought patterns must automatically translate into low-complexity medical decision making. We simply do not appreciate the value of our cognitive labor. The best defense against this form of undercoding is a basic understanding of the medical decision making required for 99213 and 99214 visits.
Calculating medical decision making
According to Medicare’s Documentation Guidelines for Evaluation and Management Services, a level-3 established patient office visit requires medical decision making of low complexity. Moderate-complexity decision making is required for a level-4 encounter. Before you can distinguish between the two, you must understand that the level of medical decision making in a patient encounter is based on three parameters: the problems addressed, the data reviewed and the level of risk.
The problems and data are evaluated using a system of weighted points depicted in the tables. These tables were developed by the Centers for Medicare & Medicaid Services and distributed to all Medicare carriers to be used on a voluntary basis; although widely used, they are not part of the official E/M guidelines.
HOW IT WORKS
Use these tables (
) to calculate your level of medical decision making. Your assessment of the problems addressed, the data reviewed and the level of risk will determine the overall level of complexity. Remember that two of three elements are required.
|MEDICAL DECISION MAKING||Problem points||Data points||Risk|
|Self-limited or minor (maximum of 2)||1|
|Established problem, stable or improving||1|
|Established problem, worsening||2|
|New problem, with no additional work-up planned (maximum of 1)||3|
|New problem, with additional work-up planned||4|
|Review or order clinical lab tests||1|
|Review or order radiology test (except cardiac catheterization or echo)||1|
|Review or order medicine test (PFTs, ECG, cardiac catheterization or echo)||1|
|Discuss test with performing physician||1|
|Independent review of image, tracing or specimen||2|
|Review and summation of old records||2|
|RISK||Presenting problem(s)||Diagnostic procedures||Management options|
|Minimal||One self-limited or minor problem (e.g., cold, insect bite, tinea corporis).|
An encounter earns points based on the number and type of problems addressed. For example, an encounter with a patient whose chronic illness is stable would be worth one “problem” point, while an encounter involving a patient with a new problem for which additional work-up is planned would be worth four points. The data table works similarly, with different numbers of points available depending on the type of data and the nature of the review. For example, reviewing or ordering a clinical lab test is worth one point, while reviewing and summarizing old patient records is worth two.
The risk table is identical to the one in the E/M guidelines. It only takes one element from any of the three categories listed in the table (presenting problems, diagnostic procedures and selected management options) to qualify for a particular level of risk. The documentation guidelines explicitly state that the physician should use the highest level of risk present when determining the complexity of the medical decision making. For example, an encounter with a patient who presents with one stable chronic illness would amount to a low level of risk. However, if the physician actively manages prescription drug therapy during the encounter, the risk level for the visit qualifies as moderate, because prescription drug management is associated with moderate risk.
After you determine the problem points, the data points and the level of risk, you can determine the complexity of the medical decision making. The “Medical decision making” table shows how the categories work together. The highest two of three elements determine the overall level of medical decision making.
Low complexity vs. moderate complexity
Distinguishing between low- and moderate-complexity decision making using the point system described above may seem awkward, but it is not difficult if you use a systematic approach. First, consider low-complexity medical decision making. Suppose you see a patient with osteoarthritis that was previously controlled on acetaminophen. The patient now says that the pain has gotten worse, so you decide to switch to over-the-counter ibuprofen and schedule a return visit in two months with routine labs.
Using the point system, this visit would add up to two problem points (for an established problem, worsening), one data point (for ordering labs) and moderate risk (due to the presence of a “mild exacerbation of one or more chronic illness”). Because two out of three factors must meet or exceed the requirements for any given level of medical decision making, it is easy to see that this encounter reflects low-complexity medical decision making, which would correspond to a 99213. Simply put, patients who are correctly assigned this code are not very sick. It is difficult to believe that the overwhelming majority of visits to primary care physicians fall into this category.
Now consider the cognitive labor required for a 99214 encounter, which calls for moderate-complexity medical decision making. Many physicians mistakenly believe that a patient needs to be in medical extremis to justify this level of medical decision making. If you break down the requirements, this misconception is easy to dispel.
Consider the patient above with osteoarthritis. If you add stable hypertension to the clinical scenario, the calculation of the medical decision making changes. In this case, you would garner three problem points (two points for the established, worsening problem of osteoarthritis and one point for the established, stable problem of hypertension). The data points would be unchanged (one point for ordering labs), and the risk would remain moderate (due to “mild exacerbation of one or more chronic illnesses”). Remembering that two out of three elements are required for any level of complexity, it now becomes apparent that the clinical circumstances justify moderate-complexity medical decision making.
This example shows that you can’t always rely on clinical intuition to predict the complexity of medical decision making. The hypertension may not make the patient seem much sicker to the physician, but that small clinical wrinkle pushes the medical decision making to the next level.
EXAMPLES OF MODERATE-COMPLEXITY DECISION MAKING
Patient has well-controlled diabetes and sub-optimally controlled hypertension. You increase lisinopril from 20 to 40 mg po qd. This visit contains three problem points: two for hypertension (established problem, worsening) and one for diabetes (established problem, stable). The risk qualifies as moderate due to either two stable chronic illnesses or prescription drug management.
Otherwise healthy established patient complains of intermittent light-headedness. You perform an ECG and review the tracing, which is normal. You order a Holter monitor and schedule the patient for a follow-up visit in one week. This visit contains four problem points for a new problem with additional work-up planned and three data points (two for independent review of image, tracing or specimen, and one for ordering a medical test). The risk is moderate based on the presence of an undiagnosed new problem with uncertain prognosis.
Patient has diabetic nephropathy and chronic renal insufficiency, with creatinine stable at 1.5 mg/dL. Diabetes and blood pressure are both optimally controlled. Labs show that the microalbumin-to-creatinine ratio has gone from 140 to 320 mg/g. You increase the patient’s ARB dose and plan to recheck the renal profile and spot urine protein in three months. This visit contains three problem points (one each for diabetes, hypertension and renal insufficiency) and one data point for reviewing labs. The risk is moderate due to prescription drug management.
Documenting a 99214
Of course there is more to selecting the correct code than just evaluating the level of medical decision making. The history and exam you document must meet certain criteria as well (see the code selection table). Perhaps one reason physicians balk at the prospect of coding 99214 more often is the perception that the documentation is considerably more burdensome than for 99213. If you examine the E/M requirements for 99214, you will see that this is not the case. Remember that established patients require the documentation of only two out of three qualifying key components for any given level of care. Assuming that the medical decision making qualifies as being of moderate complexity and that medical necessity is clear, documenting either a detailed history or a detailed exam will support coding 99214.
|History: CC||History: HPI||History: ROS||History: PFSH||Exam||Medical decision making|
|99213||Required||1–3 elements||Pertinent||Not required||6–11 elements||Low complexity|
|99214||Required||4+ elements (or 3+ chronic diseases)||2–9 systems||1 element||12 or more elements||Moderate complexity|
Detailed history. According to the 1997 version of the documentation guidelines, a detailed history requires a chief complaint (CC), four elements of the history of the present illness (HPI) or the status of three chronic or inactive problems, a review of two to nine systems (ROS), plus at least one pertinent element from the past medical, family or social history (PFSH). Although this sounds like a lot of paperwork, the following example shows that it’s not that bad. Consider our patient with osteoarthritis and well-controlled hypertension:
CC: Follow-up osteoarthritis.
Interval history: The patient states his arthritis is no longer controlled on Tylenol. He complains of bilateral knee pain described as a dull ache, which has been worsening for the past two months. The pain is worse after walking long distances and is sometimes associated with swelling in both knees.
Medications: HCTZ 12.5 mg po qd, atenolol 25 mg po qd, acetaminophen 650 mg po q4h prn.
ROS: Musculoskeletal – negative for myalgias, proximal muscle weakness, or joint redness or warmth; and cardiovascular – negative for chest pain, orthopnea or PND.
Pertinent PFSH: Positive for HTN, which is well-controlled on current medications.
That’s all there is to a detailed history. The interval history contains five HPI elements (location, quality, duration, associated signs and symptoms, and modifying factors). The two systems of ROS (musculoskeletal and cardiovascular) and the PFSH (hypertension) are probative and informative. If you perform the documentation succinctly and precisely, the amount of information needed is not particularly onerous. The important thing is to document in an ethical manner by including only those elements that are reasonable and medically necessary for the clinical problems at hand.
Detailed physical exam: Instead of choosing to take a detailed history for the above 99214 encounter, some physicians may feel it would be more informative and within the bounds of medical necessity to perform a detailed physical exam. According to the 1997 E/M guidelines, this requires 12 bullets from any organ systems. Of course, each physician can perform whatever elements of physical exam he or she feels are clinically relevant, but a typical detailed exam for our patient with hypertension and osteoarthritis, for which nonsteroidal anti-inflammatory drug therapy is being considered, might look like this:
General appearance: No acute distress, looks about stated age, conversant.
Vitals: BP 130/80, HR 74, RR 20.
Eyes: Eye grounds clear with normal posterior segments.
Neck: No JVD or carotid bruits.
Lungs: Clear to auscultation and percussion.
CV: Regular rate and rhythm, no murmurs, rubs or gallops, and normal PMI.
Abdomen. Soft, non-tender, no HSM.
Extremities: Digits and hands show no active tenosynovitis or nodules; both knees have small effusions and demonstrate moderate crepitus and decreased range of motion; normal joint stability with no evident laxity; no peripheral edema, brisk pedal pulses bilaterally.
Skin: Normal turgor; no rash or levido reticularis.
This particular example contains at least 16 bullets, but the guidelines require only 12. Each exam element is clinically relevant and informative, satisfying the requirements for ethical documentation. As with the detailed history, the volume of data required for the detailed exam is not unreasonable.
Medical decision making. Whether you choose to fully document the history or the physical, it is important to remember to document the medical decision making as well. If a problem is uncontrolled, be sure to make that point clear in your assessment and plan. If the patient is starting new medications, don’t forget to mention it. Continuing our clinical scenario, here’s how one might reasonably document the medical decision making:
Start ibuprofen 400 mg po tid.
Continue current blood pressure medications unchanged.
Patient was educated about GI risks of increasing doses of ibuprofen, especially when combined with alcohol.
NSAIDS can also lead to worsening hypertension, so I asked the patient to monitor his blood pressure more frequently.
Return visit scheduled in two months with CBC and renal profile.
Giving yourself credit
The key to understanding when it is appropriate to code 99214 for a routine visit is to train yourself to recognize moderate-complexity medical decision making in your daily practice.
The ability to distinguish between level-3 and level-4 services is not an academic issue. In the current climate of shrinking reimbursement and increasing overhead costs, most doctors can’t afford to leave potential revenue on the table.
Pause for a moment before you code your next routine visit. Consider the medical decision making, including medical necessity, and let the intensity of the cognitive labor guide your code selection. Check to make sure your documentation is congruent with that code. You may be surprised how often 99214 is the appropriate choice.