Slashdot Mirror


Federally-Mandated Medical Coding Gums Up IT Ops

Lucas123 writes "The change over from a medical coding system in use since the 1970s to an updated version that adds more than 50,000 new 7-character codes is being compared to Y2K as an IT project that is nearly impossible to complete on time. ICD-10, which replaces ICD-9, adds far more granularity to medical diagnosis and treatment. For example, ICD-9 has one code for a finger amputation. In contrast, ICD-10 has a code for every finger and every section of every finger. An 'unfunded mandate,' the change over to ICD-10 codes is a multi-year project for hospitals, state Medicaid organizations, and insurance providers. The effort, which affects dozens of core systems, is taxing IT operational budgets at a time when shops are already under the gun to implement electronic health records."

30 of 254 comments (clear)

  1. Structured data makes this easier by Mjec · · Score: 4, Interesting

    Surely if the specification lists the data in a structured way, they don't need to be hard-coded. Can't you just stick them all in a database and do lookups? Can't the authority give that the requisite structure?

    --
    "But everyone should know everything." -markab
    1. Re:Structured data makes this easier by Registered+Coward+v2 · · Score: 2

      Surely if the specification lists the data in a structured way, they don't need to be hard-coded. Can't you just stick them all in a database and do lookups? Can't the authority give that the requisite structure?

      I'm not so sure of that - one of the big problems would be how do you integrate 9 and 10 digit data so historical records are accurate as well as what happens when systems expecting 9 digits now see 10? There's a huge set of 9 digit data that won't simply go away and systems need to be redone to account for two separate data types. Data entry also needs to still account for 9 digit codes as service delivery and data entry dates could be far apart.

      I guess you could simply add a digit to the 9 digit codes to make them 10 digit but I bet that would result in code overlap as well - or simply make all the 9 digit data no longer recognizable by systems.

      --
      I'm a consultant - I convert gibberish into cash-flow.
    2. Re:Structured data makes this easier by crashumbc · · Score: 2

      You don't work in the medical field do you? ICD codes are built into a lot of medical applications. For a famous slash dot car analogy, this is like trying to convert a U.S. car manufacturing plant over to use all metric.

    3. Re:Structured data makes this easier by tbannist · · Score: 3

      So it's something sensible that should have been done a long time ago?

      --
      Fanatically anti-fanatical
    4. Re:Structured data makes this easier by Richard_at_work · · Score: 2

      So thats it, its embedded in the core of the system so it is to never change again for the rest of mankinds existence?

      Uhm, no - you eat the cost of changing over from poor prior decisions and design a system that is resilient to change this time.

    5. Re:Structured data makes this easier by tbannist · · Score: 2

      Also the insurance company that has to compensate the guy who lost the finger(s) might like to know which one(s) were amputated and how much was amuptated. They pay different amounts for different fingers and different amounts depending on how many joints were lost.

      --
      Fanatically anti-fanatical
    6. Re:Structured data makes this easier by Anonymous Coward · · Score: 2, Insightful

      I don't see any "sense" in adding codes merely to tell the doctor which finger was amputated. All he has to do is LOOK and see for himself which finger is missing. This is typical government "make work" bureaucracy that makes no more sense than going-round and busting windows to boost construction jobs.

      It's also what bankrupted the treasury and led to the downfall of the Roman Empire (according to one historical theory).

      I had a finger amputated. Tell me which one. Go on, just LOOK, after all, that's obvious, right?
      Not like there's ever a situation where that information, individually or aggregated, might be useful to anyone in the spheres of research, information or analysis, without the patient in front of them, waving. And if it is, we can just line up all the amputees and they can look at them one at a time.
      Alternatively, maybe it'd be handy to code this stuff up.
      Just be grateful its ICD and not SNOMED CT, which contains over a million medical concepts.

    7. Re:Structured data makes this easier by Anonymous Coward · · Score: 5, Interesting

      They're versions, not digits. ICD-9 diagnosis codes (for some reason the International Classification of Diseases also has a set of procedure codes) use up to 5 digits in the form [0-9VE]##.##. ICD-10 is of the form X##.###X, except for some codes that have a "placeholder" to pad the middle of the code out to seven digits.

      The real problem here is that insurance claim submission is real design-by-committee bullshit of the highest degree. It's an ANSI standard, a submember of ASC X12 so half the shit in there is unused crap needed for the other things X12 is used for, like wholesalers restocking their shelves or boats reporting their cargo, because apparently code reuse is so damn important to these people that the claim form has a section (completely unused for claims) for reporting credit card details. Of course, the insurance companies all took this design under advice and did their own shit with it. There's a code that identifies whether the insurance company is blue cross, medicare, medicaid and so on (that you have to put on the claim just in case blue cross forgot, and they apparently forget a lot, since if the clinic forgets to set this code properly, the insurance company uses every excuse possible to refuse to pay). Simple enough, right? Well when I started, we had an insurance company that used blue cross's servers for processing claims, so even though they weren't blue cross, this code had to be blue cross or BC's servers would shit themselves. Setting that aside, there's a completely separate code for what kind of provider ID you're sending (blue cross, medicare, medicaid and so on) because of course people are going to bill blue cross using medicaid IDs. Fortunately, the NPI did away with that bullshit (for the most part... medicaid here still demands provider IDs for checkups because they refuse to give up any hoops for doctors to jump through, and one of those was that the doctor had to use a different provider ID for checkups versus treating someone sick. Their computer system apparently cannot sort these claims out themselves... and yet if you bill a checkup on the wrong provider ID, the computer system can easily reject it. Hmmm...)

      Ahem.... Anyways, instead of just adding a code to identify whether the diagnosis in question is using ICD-9 or ICD-10 (in the box the committee already created for the purpose of identifying the code being used), the committee got together and pretty much rewrote the whole damn thing. This is where IT got gummed up (it's getting better now). And believe me, you can talk about "legacy equipment" and other stuff til you're blue in the face, but claim submission is how doctors and hospitals get paid, that's where IT has been spending all of its time freaking out.

    8. Re:Structured data makes this easier by VoidEngineer · · Score: 2

      It's not just that a bunch of codes have been added or removed. The relationships between the codes have changed as well. What used to be a single code in ICD9 may have been replaced with a list of codes in ICD10 (or, worse, a tree structure of codes); and conversely, what used to be a dozen codes in ICD9 might have been replaced with a single code. The disjoint mapping of codes and their relationships is what's really gumming up the works.

  2. I bet you anything by Rosco+P.+Coltrane · · Score: 2

    the people who dreamt up the new coding system didn't even try to make it backward-compatible with the old one, hence the headaches and waste of money.

    If ICD-10 was a superset of ICD-9, in a way similar to how UTF-8 is a superset of ASCII, the transition would be perfectly seamless and painless...

    --
    "A door is what a dog is perpetually on the wrong side of" - Ogden Nash
    1. Re:I bet you anything by jhoegl · · Score: 4, Interesting

      True.
      As an IT guy who worked for a medical billing company I got to see the inner workings of the coding world.
      It is interesting to see that it requires another human being to code from a doctors notes. And then many things came into play, such as alife medical, a EMR system that codes based off of what it reads. They even converted TIFFs with OCR so they could read it. Now with things like NextGen you can put it all into the system and let the system do it for you. No more "interpreting" what the doctor wanted to say. However, with this system it will be tough to find doctors who overbill or put in information that is untrue.
      I would watch coders detect these things by finding a doctor attempting to charge for a procedure that did not even involve the issue. Can an EMR system do that? Hopefully, but it will probably be an after thought to the new coding.
      And what about these certified coders? Do they have to retrain and re-certify? Probably.

      That job was an eye opening experience into the Medical billing world. It was very interesting and I helped develop some of the very first medical billing methods.

    2. Re:I bet you anything by VoidEngineer · · Score: 2

      Unfortunately, the old ICD9 is considered broken, so it's not suitable for supersetting. As an example, say that the skull is considered part of the head in ICD9, but the skull is considered part of the skeleton in ICD10. The code for skull has moved from 'head' to 'skeleton'. How do you superset that kind of relationship change? Worse, what if the original ICD9 relationship is considered no longer scientifically valid? What if a disease of an organ is reclassified to be a disease of blood? Or of the nervous system? Supersetting wouldn't fix those broken relationships.

    3. Re:I bet you anything by geekoid · · Score: 2

      Except it's not possible. I suggest looking into why it isn't.

      --
      The Kruger Dunning explains most post on /. http://en.wikipedia.org/wiki/Dunning%E2%80%93Kruger_effect
    4. Re:I bet you anything by geekoid · · Score: 2

      It's almost always impossible to super set when you are doing a refinements of a dataset of that reflects real world messy 'objects'

      Most people don't understand that. Sadly, A lot of 'software people' don't get it either.

      --
      The Kruger Dunning explains most post on /. http://en.wikipedia.org/wiki/Dunning%E2%80%93Kruger_effect
  3. Makes a lot of sense by swbozo · · Score: 2

    It may be an "unfunded mandate", but it will probably help eliminate thousands of medical errors (mistaken amputations, incorrect medicine given to patients, etc.)

    1. Re:Makes a lot of sense by Anonymous Coward · · Score: 4, Informative

      No. Not at all.

      Coding, is just a manner of recording in a concise format, what diagnosis was made, and what treatment was given. This way the data can be used for billing and statistics.

      I have never, ever seen medical codes (be them ICD9, ICD10, SNOMED) ever used by doctors, medical technologists, or anyone with any direct influence over patient care. They aren't used in X-ray equipment (so there's no risk of incorrect examinations or incorrect interpretation), blood-work analysis machines, etc.

      Being an MD, and part-time software developer, I've done a lot of work with medical databases. I looked at coding of educational files, and being able to import selected educational cases directly out of the hospital's electronic record system, in anonymised fasion, into an educational system. The first thing I noticed was that the coding (in this case ICD9) was extremely inaccurate, and often wildly misleading - fewer than half the codes I looked at were correct. The reason was that the coding was done by clerical staff, who read through the charts, and worked out what codes to use. These clerical staff weren't doctors, nurses, or other people trained in medical diagnosis. As a result, they would frequently misinterpret the charts and the wrong codes would be sent to billing and for government health statistics.

  4. meh. by Ephemeriis · · Score: 4, Informative

    The effort, which affects dozens of core systems, is taxing IT operational budgets at a time when shops are already under the gun to implement electronic health records.

    For the most part, this isn't my problem.

    ICD-10 has been on the radar for a while now. At least a couple years. And it's Federally mandated. So we didn't have any problem adding the necessary funds to our budget this year.

    Each of our HIS vendors has already got ICD-10 stuff ready to go. We'll have to pay them for their time, or a software release, or whatever... But, as far as my own labor is concerned, it'll basically involve giving them remote access or throwing a disc in the drive.

    Most of the labor involved is in our coding department. They're going to have to send folks out to get (re)trained in the ICD-10 stuff. They are, understandably, a little stressed. But they've been working on this for a while, too.

    --
    "Work is the curse of the drinking classes." -Oscar Wilde
  5. Procrastination hurts by RKThoadan · · Score: 5, Informative

    While this is definitely a huge pain, I have little sympathy for those complaining about the timing of this when the standard was finalized in 1992.

  6. Cry me a river by Enry · · Score: 4, Interesting

    ICD-10 has been out for nearly 20 years. There was a 5 year timeline to get ICD-10 implemented, and there was likely a few years of discussion with major Medi* billers before that to let them know this was coming along.

    Much like the FCC and HDTV, health care companies must have ignored the mandates until it was too late, whined and cried about how they couldn't meet such a strict deadline and pretty please can we extend it for another 5 years. Repeat until our health care records system is completely unusable.

    Though, wow, I would have thought VistA would have ICD-10, but it's being bolted on now. Strange.

    1. Re:Cry me a river by crow_t_robot · · Score: 2
      It appears to me that most hospitals/medi orgs were dragging their feet to see how it would play out:

      "Quite frankly, the hold up is it's a big undertaking and it took them a while to get under way. Everybody's started, but a large percentage of hospitals are in the heavy analysis stage or they're just starting," said Casey Corcoran, vice president commercial solutions for healthcare at General Dynamics Information Technology, a vendor offering ICD-10 consulting services.

      "heavy analysis": heavy procrastination

    2. Re:Cry me a river by Gideon+Wells · · Score: 3, Insightful

      More and more I keep hearing "unfunded mandate" or "harsh deadline", but experience is translating it as "I waited till the last minute and now I'm screwed."

      --
      by Anonymous Coward: I, for one, welcome the shift from car analogies to pizza analogies. um.. overlords?
  7. Re:Good. by Anonymous Coward · · Score: 5, Informative

    The misattribution of the reason for rising costs in health care is unfortunate. The fact that government as a payor is in this system is one of the main reasons costs rise as such a rate.

    I work on the revenue side at a fairly large health system, and due to our population we have approximately 50% of our patients privately insured, and close to 50% are government insured. (There are very few people coming in the doors who truly have no coverage, despite what the politicians would have you believe). The privately insured generally repay approximately what's billed, but for the government paid accounts, we would do well to recover 50%, 40% is a better guess. This is because the government solves its own budgetary problems by withholding increases to its medical payouts. In this area, we're still being paid at 2002 rates in 2011. We never withhold a needed service, so we just eat the cost.

    But we have to make the books balance somehow. We're non-profit, but we still have to keep the lights on. Charge rates have to go up across the board, and the money lost on literally every government paid patient who walks in the door is then made up by the privately insured. You hear about $8 tylenol being billed, well, tylenol has to cost $8 because you're actually buying a whole bottle and sharing it with everyone else.

    I think we're going to continue to have a huge gap of misunderstanding while people continue to emote over "big business" and "fat cats" and "obscene profits" without understanding what the real financial issues are on the ground.

    Posted as AC due to PHB concerns.

  8. Re:It doesn't exactly sound like a waste of time by crashumbc · · Score: 2

    FYI, this really doesn't affect the diagnostics. 98%+ of it has to do with billing. Just because there was only one code for amputating a finger doesn't have anything to with the doctor doing his job. He does cuts it off describes what was done in the chart(or EMR) then after the person is discharged a person in a little room called a "coder" goes through the chart and enters the billing code(ICD-9 currently) in the bill that gets sent off to the insurance company.

  9. Total Bullshit by Saerko · · Score: 4, Interesting

    As pretty much everyone else has already said, if you don't have a system that can quickly and easily update from ICD-9 to ICD-10, you're so far behind the IT implementation curve that you should be drug out into the street and shot.

    It's 2011. They've had many, many years to upgrade, and now they're poised to paid by the government to do so. Hell, my employer stands to gain $50 million dollars over the next couple years from implementing key portions of the HITECH provisions in ARRA. For those reading, that's more than half of my (quite large and well-funded) health system's annual budget.

    For our part, we just slapped down a couple hundred thousand for a product that hot-swaps our ICD-9 coding for ICD-10, and also tosses in a problem list that physicians can use that's tied to these coding schema, potentially improving efficiency and accuracy as well. The only excuse, and I mean ONLY excuse for ICD-10 being a problem is poor IT leadership within the health system/hospital--a failing which is incredibly, unbelievably common.

    I'm lucky, I work in an IT shop that actually has its shit together.

  10. Re:And it slows things down by BlackHawk · · Score: 3, Informative

    As someone who's working with this stuff right now, I can say if it's slowing you down, you're not taking advantage of the available tools. They're out there. Keep looking. Moreover, "data entry" is one way of looking at it. A different way to call it is "documenting what they're doing with sufficient detail". That was the entire point of these kinds of standardized coding systems: to (as best as we can) remove the fuzzy documentation in the systems before, and to remove the idiosyncrasies from medical records. With the proper coding systems in place, a patient in Allentown who moves to Duluth can have his PHI moved to the new caregiver and be (for the most part) confident that the Iowans will be able to understand what the Pennsylvanians did for him before. Yeah, there's going to be transitional pain. There always is. But as has been pointed out in other posts, it's not like ICD-10 ambushed anybody. Frankly, if you haven't been moving toward ICD-10-capable systems for at least 2 years, you've been slacking. There's a penalty for that at crunch time.

    --

    Believe nothing, not even if I say it, if it violates your sense of reason -- Buddha

  11. Re:What is the ICD-10 code by weeboo0104 · · Score: 2

    What is the ICD-10 code for being a dickless idiot. They could save time by automatically adding it to the record of all Federal employees. OK - maybe not the FBI - honestly I didn't mean you, I was just sayi.....

    The ICD-10 codes for being a dickless idiot would be:
    Q56.4 Indeterminate sex
    F79.1 Unspecified mental retardation with significant impairment of behaviour requiring attention or treatment

    --
    It is easier to build strong children than to repair broken men. -Frederick Douglass
  12. Unfunded mandate? by hellfire · · Score: 2

    I thought the 2009 stimulus package had $30 Billion which was to be paid to hospitals to update their systems? The only backup I have for this is an I, Cringley article I read over the weekend but I'd feel he'd of all IT writers would be a reasonably reliable and impartial source. Definitely a far more reliable source than the average Slashdot submission.

    --

    "All great wisdom is contained in .signature files"

  13. Re:Good. by iluvcapra · · Score: 4, Insightful

    This is a misconception. The United States has the ighest share of private spending per individual on health care, yet it has the highest costs and only middling health outcomes.

    Every system in the developed world has a private spending component, usually through insurance and copays just like here, but most also have a government payer either as the single payer or more commonly as a backstop, last resort payer; all systems more firmly regulate costs and practices and all systems at least have independent boards to assure efficacy of treatments (we call those "death panels.") There's no empirical evidence that people make rational decisions about their own health care spending.

    --
    Don't blame me, I voted for Baltar.
  14. Re:What is the ICD-10 code by nitehawk214 · · Score: 2

    What is the ICD-10 code for being a dickless idiot.

    S38.2 Traumatic amputation of external genital organs
    F73 Profound mental retardation

    The fun part about ICD codes are you get 3 replies with 3 different codes for the same problem.

    --
    I'm a good cook. I'm a fantastic eater. - Steven Brust
  15. Re:Good. by geekoid · · Score: 2

    ". The fact that government as a payor is in this system is one of the main reasons costs rise as such a rate."
    no, it is not. In fact, it's cheaper.

    Do you know why Canada gets cheap Meds? because the government is the payor? do you know why the VA gets meds cheaper then Canada? the government is the payor.

    "(There are very few people coming in the doors who truly have no coverage, despite what the politicians would have you believe)."
    Wow, you don't see the fallacy there? really?

    What nonsense.

    "You hear about $8 tylenol being billed, well, tylenol has to cost $8 because you're actually buying a whole bottle and sharing it with everyone else."

    No, it's because it is being administer by highly payed professionals.

    Normal I don't respond to AC, but your post is so full of crap regards a hot topic issue, it needed to be pointed out you are full of crap.

    --
    The Kruger Dunning explains most post on /. http://en.wikipedia.org/wiki/Dunning%E2%80%93Kruger_effect