I’m not worried about ICD-10. Here’s why.

It is possible to categorize every human ailment, and assign every disease a code. This is called the International Classification of Diseases (ICD), which was first formalized as a short list of malaises at a meeting in Paris in 1900. Since then, this list has been revised ten times, getting longer each time, in an effort to aid epidemiological and policy matters around the world. The ninth edition (ICD-9) has been around since 1975 as a bible-thick book with about 13,000 codes. For example, 786.5 means chest pain. On the other end, some codes are as specific as E845.0: “accident involving spacecraft; occupant of spacecraft injured.”

The tenth edition, ICD-10, was supposed to be implemented throughout U.S. health care this year, but was recently postponed by an act of Congress until at least 2015. This 10th edition has 68,000 codes, with increased specificity over the ninth edition. For example: w5609XA: “other contact with dolphin, initial encounter,” or V95.45XD: “spacecraft explosion injuring occupant, subsequent encounter.”

Every doctor uses these codes every day. We have to write a code that corresponds to the problem, in order to get paid. (Doctors that don’t bill insurance companies, only take cash or work for free, don’t have to worry about all this.)

In training, I used to look up the most specific ICD-9 after every visit. It took me a couple minutes of flipping through a book or going to Codapedia. Now I have to pick a code up to 50 times a day, for every patient visit, lab order, home nurse authorization, request for incontinence diapers, x-ray order, and everything else imaginable.

Looking them up would waste an hour each day, and I’m not going to memorize 13,000 codes. So instead I have 20 memorized, and I pick from the same 20 for every patient. That’s right, most of family medicine, which includes many varied ailments, reduced to 20 codes. Does your wrist hurt? Forget 726.4 “enthesopathy of wrist,” I use 729.5 “pain in limb.” Do you have knee pain from arthritis? 729.5. Your big toe hurt because of gout? 729.5. Elbow pain because you were bit by a dolphin? 729.5.

Of course there are different codes for every type of big toe ailment, and insurance companies tell doctors to be specific with our coding. For example, 250.00 is “type 2 diabetes without mention of complication,” while 250.63 is “type 2 diabetes with neurological manifestations, uncontrolled,” and there are codes for every permutation in between. Although the suggestion is to be specific, I have absolutely no incentive to do so. So I use 250.00 every time I have a diabetic.

By no incentive, I mean: No insurance company argues with my 20 codes. They don’t allow 401 “hypertension” because it does not have enough significant digits, but I don’t waste time deciding between 401.0 “malignant hypertension” and 401.1 “benign hypertension,” I just use 401.9 “hypertension unspecified” every time. And my claims don’t get rejected from Medicare, Medicaid, Blue Cross, or ten other insurance companies I routinely bill.

In fact, I have a huge incentive to not use other codes. My 20 codes are tried and tested, I know they will get paid. But if I try a code that I usually don’t use, I sometimes do get rejected. As mentioned, if I use 401 instead of 401.9, I get rejected. Then I have to submit a new code in a time wasting, trial and error game of getting paid. That is to say, there is a big incentive for me to not be specific with my coding.

I have never had someone call me up and say “we notice your patients have a lot of pain in their limbs, could you please be more specific next time.”

This might make me sound like I am not playing by the rules. Like some epidemiologist somewhere will not be able to make a pie chart of big toe ailments, and advances in big toe medicine will suffer because of me. But guess what: Every other doctor does this also! Well many of them anyway.

Almost every outpatient family doctor has a “superbill,” a paper that, among other things, has a short list of 50 codes. Here is an example from a family medicine organization. Scroll down to page 2, and notice the phrase NOS, which stands for “forget your 13,000, I’m picking from my fave five.”

Of course when I really don’t know an appropriate code, I look one up. This happens a couple times a week for me, and probably the same for other doctors.

Paper superbills have been replaced by electronic systems that make it easier to pick more complex ICD-9 codes, but really, most family docs still default to a few memorized codes.

There is a lot of talk about the transition to ICD-10: A Y2k style fear of retooling electronic systems and workflows for the major change in coding methodology. Doctors are putting a lot of effort into opposing it; you can’t read about medical policy these days without seeing worries about ICD-10.

I am not worrying. I suspect I will have to memorize a different set of 20 codes. When I see a patient who got injured in a spacecraft explosion, you better believe I am going to use M79.609, “pain in unspecified limb”, unless my bills stop getting paid. And if they do, I will learn incrementally more codes to barely get by, but you will never find me looking up one of 68,000 codes, 50 times a day.

I’m pretty sure that most providers out there are going to do the same. Doctors who specialize in rotator cuff surgery will have their most common 20 rotator cuff codes memorized, and will look up other codes occasionally as needed. And the epidemiologists and policymakers who count these things will be left with similar quality data to what they have now: nonspecific in and nonspecific out.

Categorizing diseases to advance medicine is an important endeavor, but I am not going to add an hour to my already packed day of seeing patients. This is why I don’t care about ICD-10.

P.J. Parmar is a family doctor at Ardas Family Medicine and blogs at P.J.! Parmar.

Comments are moderated before they are published. Please read the comment policy.

  • http://EasyOpinions.blogspot.com/ Andrew_M_Garland

    Px801.25red: Convicted of Political crime against epidemiology, second encounter, 5th degree of resistence, with re-education specified.

    • pjp

      thats funny

  • SteveCaley

    First point: I disagree with your statement:
    It is possible to categorize every human ailment, and assign every disease a code. This is called the International Classification of Diseases (ICD), which was first formalized as a short list of malaises at a meeting in Paris in 1900
    The assumption is fundamentally flawed. Therefore, everything that derives from it is flawed in the same way.
    The naming of things is not an off-hand matter. In medicine, we name things to the degree specified by the presentation of the matter of concern, and its essential similarity based on etiology to a generalizable condition.
    Classification and labeling is a powerful tool, but it often cuts too bluntly to be useful. I have all sorts of clinical observations over years which do not fit neatly into the corpus of the experimental literature, but are rules of thumb. Similarly, I have concepts which enclose fuzzy as well as precise processes – they are of equal merit, for they exactly describe a certain thing, to the degree that it is manifested in the patient.
    For example – I was called on a consult by the surgeons on “Acute Alzheimer’s.” The actual phrase is nonsense, of course; but the use of the phrase was valuable. The surgeon was saying – to me, this looks like Alzheimer’s but it’s acute – I’m puzzled (hence the consult.) Yet, there is no ICD-XX for acute Alzheimer’s. Any codist can suggest a litany of similar phrases; but the actual one used was the most valuable.
    Diagnosis is similar to Linnaeus’s work – to see an individual, and to proceed to gather it into a class, and name it. The order of taxonomy is a welcome thing in science, and probably inspired our friends in the ICD meeting. However, species are, with few exceptions, discrete, finite and enumerable; ill humans are none of these. Some things are worthy of naming from their phenomenology; some are not.
    What is the ICD-10 code for a crying woman?
    The author has surrendered to the necessity of constructing a pidgin language from the mother tongue – a twenty-word vocabulary. If, for instance, you travel to Mexico City, you may be able to get around with 20 words of Spanish – but not attend a seminar on the lexical undertone of Cervantes. Me no speakum so good, in English Pidgin.
    It is a surrender of the core culture of medicine to the bureaucratic mentality of pigeon-holing; it is fundamentally inhuman. It is coming; I am probably leaving.

  • ninguem

    You will care when you find your income plummet because you stop getting paid for “errors” in your ICD.

    Although, looking at your web site, I have to suspect that the financials of your practice do not depend on such errors.

    Or someone else is doing the coding for you.

    • pjp

      I do my own coding, and examine ERA/EOBs for reject reasons. I may have been exaggerating a little when I suggest using just 20, since I can easily look up other codes, but I am not exaggerating when I suggest that there are many practices that use paper superbills like the AAFP example I linked to. That one has about 150 codes, but most docs still using superbills probably only pick from 50 of them. At least that is how the majority of docs I trained with worked.

  • betsynicoletti

    Thank you for this. It injects a modicum of common sense into the discussion. Physician services are paid based on CPT code. The diagnosis code establishes the medical necessity. As Dr. Parmar points out, almost any diagnosis code will be sufficient to get paid for an office visit.

    For a diagnostic test or surgery, a more specific diagnosis code is typically needed. I’m sure Dr. Parmar supplies those when ordering a CT, or it wouldn’t be “allowed” and paid.

    Dr. Parmar, if you enter into risk based contracts, like ACO contracts, more specific diagnosis codes may increase your revenue at the time of the contract year reconciliation. And, documenting and coding the patient’s underlying medical conditions may increase your payment on year end reconciliation.

    I wrote an article about ICD-10 “The ICD-10 emperor has not clothes” which was published on kevindmd. http://www.kevinmd.com/blog/2014/04/icd10-emperor-clothes.html

    And, thanks for mentioning Codapedia. I’m the co-founder.

    • pjp

      An honor to hear from you, Codapedia is very useful for many reasons. Great posting of yours (your link just above).

      I’m a self-employed outpatient family doctor, and not part of any contracts based on ICD specificity, all of which probably allows simpler coding than some other settings. Still, I use the same handful of codes even when ordering CTs, insulin syringes, and almost anything else.

      An employed doctor friend of mine (ER), who is already on ICD10, added that when he gets more specific, his hospital coding staff leaves him stickies full of time-wasting clarifying questions. So he plays it safe by dumbing it down to some less specific code that they can’t question, which is the same experience I have with payers.

      • querywoman

        Labeling helps clarify lots of things. Of course, you use the codes when ordering supplies.

  • eqvet2015

    At a recent appointment, my doctor and I were working on the list of ICD codes to try to get my insurance to pay for some expensive genetic testing for suspected mitochondrial disease (E88.40). As she worked on the list of codes to send to the insurance company, since you can’t use the code of the thing you are looking for, it hit me that it’s like being a district attorney trying to nail a crook at trial. You charge them with as many crimes as you can think of, so that even if the murder charge doesn’t stick, maybe they’ll still go away for money laundering, conspiracy, or littering Too bad there’s conflict of interest on the side of the jury.

Most Popular