Does ICD-10 Pilot Forecast a Perfect Storm for Healthcare?

Let me concede from the outset that, in this blog post, I lean toward the negative—dire predictions, worst-case scenarios, a bit of doom and gloom, etc.

But I ask you, oh gentle, patient reader, how could I not?

Let’s go to the satellite. You can see warm air from a low-pressure system (Meaningful Use Stage 2, not changed dramatically by the one-year extension) collide with cool, dry air from a high-pressure area (the turmoil of Obamacare) and tropical hurricane moisture (ICD-10). Tell me you don’t see the Perfect Storm yourself.

And here we sit in our little fishing boat, waiting for the mighty ocean to consume us.

Overly dramatic? Certainly, but still not wholly inappropriate, I will argue.

Consider a recent report on the HIMSS/WEDI ICD-10 National Pilot Program collaborative that was created to, “…minimize the guess work related to ICD-10 testing and to learn best practices from early adopter organizations.”

Designed to ascertain the realities of the entire healthcare system adopting and using ICD-10, this pilot included an education and adoption program for all participants, followed by a set of “waves” in which diagnoses for the 100-200 most common medical conditions were actually coded and submitted using ICD-10.

The end-to-end testing approach …

…would encompass a number of medical test cases that mirror actual processing, including situations with multiple “hops” or “steps” between providers, clearinghouses, and health plans; the identification of high-risk medical test cases to help prioritize testing; the identification of available testing partners; and key reporting and sharing of test results. The test environment must mirror production.

And how did this pilot testing go? (Cue dark, foreboding music here …)

The average accuracy was in the 60 percent range with low scores around 30 percent.  Yes, some medical scenarios had nearly 100 percent accuracy, which is great. But very low accuracy accompanied a number of very common conditions. Not so great.

To be more specific:

  • 40%: Chest pain, unspecified
  • 33%: Closed fracture of unspecified part of femur
  • 45%: Coronary atherosclerosis of native coronary artery
  • 40%: Congestive heart failure, unspecified
  • 42%: Degeneration of the limb or lumbosacral intervertebral discs
  • 46%: Acute chronic systolic heart failure
  • 29%: Sebaceous cyst
  • 35%: Closed fracture of the intracapsular section of femur, unspecified

Predictably, the pilot identified coding-based challenges as the primary cause of low accuracy rates. Some are easily solved. Others, not so easily. You be the judge of this coding error best-of list:

  • Mixing up similar letters and numbers
  • Technical glitches with uploading and transmission of documents
  • Overworked coders
  • Incomplete EHR documentation
  • Coders forgetting key aspects of ICD-10 not present in the ICD-9 code set

Of course, all these errors require understanding the problem and tackling it within the context of process and team. If test subjects scored lower than 50 percent accuracy coding common diagnoses even after a well developed and implemented training program, what will mainstream providers achieve?  How much worse might they be?

I am speaking primarily of resource constrained provider organizations that are already on the edge financially, not Partners Healthcare or Mayo Clinic. For them, I think these pilot study figures portend a financial disaster: 50 percent coding accuracy means 50 percent claims denial and a precipitous decline in revenue. How will they make the needed changes to increase accuracy when organizations in the study could not?

According to the frank assessment offered by pilot study organizers, they will just need to focus.

The “perfect storm” will be quickly descending upon the healthcare system … All ICD-10 impacted organizations should act now to allocate as much time as possible for testing and remediation to protect their corporate bottom lines and cash flow to successfully achieve compliance.

While the pilot does not actually quantify the time and resources required for organizational change and ICD-10 compliance, a comment on the pilot offered by one physician speaks of an exhausted profession that can’t see a better day on the horizon.

As a practicing physician and using EHR (sic) for last 10 years, the last 2 with Epic both in office and hospital, I cannot image (sic) what this will mean. I now spend 11 to 15 hours a day, Monday through Friday, plus many hours on the weekend working on the computer. This ICD-10 sounds ridiculous to try to implement on top of everything else.

While I’m in partial agreement with Dr. John Halamka of Beth-Israel Deaconess and Harvard on this one, I don’t think his suggestion of a 6-month ICD-10 extension is enough. What will be so significantly different in 6 months? In that timeframe, I think the challenges that exist now—Meaningful Use Stage 2, the upheaval of the Affordable Care Act–will pretty much be the same. While I don’t expect it to happen, I’d suggest we delay ICD-10 until innovation makes it less of a burden. I can’t say when that will be, but I do have faith that it will happen.

It’s not that ICD-10 is an inherently bad idea, or that hospitals and providers can’t meet the challenge with reasonable deadlines. But they have too many challenges right now, and we are forgetting that most of healthcare is small provider organizations, regional and county hospitals and critical access facilities. If Kaiser struggles with MU, the ACA and ICD-10 all at once, what is a county hospital in Kansas or Idaho, or New York or California, supposed to do? When their reimbursement rates fall, they will face bankruptcy, and vital healthcare services will disappear from the areas that can least afford to lose them.

Yes, it is a clichéd pop-culture reference, but we truly are looking at a healthcare perfect storm like no other next year. We expect that this confluence of challenges will eliminate some health IT companies, and we generally accept that the herd needs to be thinned anyway. But can we be so sanguine about the potential impact on healthcare itself when financial ruin and simple emotional overload seem highly possible, even likely?

Providers will be driven to the brink, and I cannot see how this ends well for American healthcare, which I thought was the original goal.

Edmund Billings, MD, is chief medical officer of Medsphere Systems Corporation, the developer of the OpenVista electronic health record.

51 replies »

  1. I do not know whether it’s just me or if everyone else experiencing issues
    with your site. It appears as though soke of tthe textt iin your
    content are running off the screen. Can someone else please comment
    and let mme know if thiis is happenig too them
    too? This may be a issue with my browser because I’ve
    had thuis happen previously. Thank you

  2. Wow! At last I got a blog from where I be capable of in fact obtain useful information regarding
    my study and knowledge.

  3. I don’t know whether it’s just me or if perhaps everybody else encountering problems with your blog.
    It appears like some of the text within your content are running off the screen. Can someone else please provide feedback and let me know if this is happening to them
    too? This could be a problem with my browser because I’ve had this happen before.
    Thank you

  4. 8. Preventing and detecting healthcare fraud and abuse….are you serious. The people who have found ways around the current system, WILL find there ways around this system. And it will only be worse!!! But in the end it is just going to give health ins, company’s another reason not to pay a claim. Invest in Health Ins. Company’s NOW!!!!! There 2015 profits are going to go though the roof!!!! I have had many health claims adjuster tell me, “why shouldn’t we deny the claim, what’s the WORST thing that can happen???? After you resubmit your claims 2 or 3 times and file a 1st and 2nd appeal….all we have to do is pay the amount we were first billed, if WE decide to pay at all. We just get to kept your money for 6-8 months longer, making money on it.

  5. That’s insane. It takes 20″ to decide there’s not a code for that injury, then what do you do?

  6. Slipped and fell on ice injuring tailbone doesn’t have a ICD10 code.

    You have a number to pick from:
    W00 Fall due to ice and snow
    W00.0 Fall on same level due to ice and snow
    W00.0xxA Fall on same level due to ice and snow, initial encounter
    W00.0xxD Fall on same level due to ice and snow, subsequent encounter
    W00.0xxS Fall on same level due to ice and snow, sequela
    W00.1 Fall from stairs and steps due to ice and snow
    W00.1xxA Fall from stairs and steps due to ice and snow, initial encounter
    W00.1xxD Fall from stairs and steps due to ice and snow, subsequent encounter
    W00.1xxS Fall from stairs and steps due to ice and snow, sequela
    W00.2 Other fall from one level to another due to ice and snow
    W00.2xxA Other fall from one level to another due to ice and snow, initial encounter
    W00.2xxD Other fall from one level to another due to ice and snow, subsequent encounter
    W00.2xxS Other fall from one level to another due to ice and snow, sequela
    W00.9 Unspecified fall due to ice and snow
    W00.9xxA Unspecified fall due to ice and snow, initial encounter
    W00.9xxD Unspecified fall due to ice and snow, subsequent encounter
    W00.9xxS Unspecified fall due to ice and snow, sequela

    You also have codes for Ice Skaters and Hockey injuries which could also have an injury to tailbone (or coccyx).

    Was it fractured: then you have other codes and it changes with each visit.
    S32.2 Fracture of coccyx
    S32.2xxA Fracture of coccyx, initial encounter for closed fracture
    S32.2xxB Fracture of coccyx, initial encounter for open fracture
    S32.2xxD Fracture of coccyx, subsequent encounter for fracture with routine healing
    S32.2xxG Fracture of coccyx, subsequent encounter for fracture with delayed healing
    S32.2xxK Fracture of coccyx, subsequent encounter for fracture with nonunion

    My doctor told my daughter who fell on the ice that she bruised her tailbone, that explained enough to me.

  7. No.

    The one that charges too much for coverage, sponsors all sporting events on TV and screws doctors by paying pennies on the dollar.

  8. We are in an agricultural area, so we do get collisions with cows, horses,
    sheep, etc.

  9. I assure you that the operatives to whom you send your data dread the advent of such coding as much as yourself.

    I would suppose that an inordinate frequency of collisions with sea lions, turtles and skaters, at certain locations, arguably bears looking into. Not by physicians, of course.

  10. That’s what happens in transcribed notes, archon41, assuming someone reads them. As for EHR, as I see it, the narrative portions can be lacking, which also makes them difficult to use for cases as you mentioned above, as well as work-related injuries. In my case, I transcribe and state specifically, the patient slipped and fell on ice injuring the tailbone area, etc. This type of description is critical to establishing a cause and effect.

  11. Unfortunately, the Bureau of Worker’s Compensation follows standard insurance protocols, so, yes we will probably have to start using them next year.

  12. Perhaps we should revert to Old School, and require providers to submit, to the people responsible for paying their fees, descriptive verbal narratives of treatment rendered, and the medical necessity thereof.

    The interest of the physician in the well being of his patient is hardly the sole legitimate interest involved here. The ER physician who took care to mention the specific locus of the vehicle accident resulting in injury was perhaps aware that auto insurers, and lawyers, can be most tiresome in confirming that the treatment in question was necessitated by a particular mishap, and no other.

  13. Jeff, how is ICD-10 going to be any more time consuming after implementation – once the initial training is accomplished? Yes, more codes but entering a code is entering a code.

    What is the dollar benefit, I don’t know, but what would you want as far as providing research material on the health of the nation? No data does not make better decision makers.

    “What we really need is an ICD-11 or ICD-12 ”

    Do you think docs will embrace that any better?

  14. Well, that’s good news. We in the U.S. have 5% of the world’s population, 25% of it’s prisoners, and 50% of the world’s lawyers.

  15. Sorry Arthur, revised code would be “a person showed up, I treated for everything, send money.

  16. I’ve just begun reading your latest book and will be citing/reviewing it on my REC blog (the link to which apparently remains banned here. Just
    Google “REC Blog” – #1 result).

    I will be citing this comment as well. Thank you.

  17. Call it a mandated transfer of wealth triggered by politics, not data-driven decision-making.

    One could essentially boil the entire ICD-10 debate down to this: The people who have to pony up cash for the conversion and then actually use ICD-10 don’t like the idea nearly as much as those who either bear less risk or even stand to profit from the mandate.

    Which would be all well and good if there were known, perhaps even piloted, benefits to the broader healthcare system. Where are those?

    To date no one — and certainly not CMS — has clearly outlined the case for exactly what benefits can be reaped by anyone. Yes, proponents can rattle off lists packed with vagaries about analytics, Big Data, population health, granularity from 7-digits that will ultimately cure cancer, eradicate diabetes, and make sure every mother in the remotest village is perfectly safe during childbirth.

    Remove those smudged-lens glasses. Refocus instead on Oct. 2014, Oct. 2015, even 2016, and the potentially overwhelming provider percentage that achieves only minimum compliance and doesn’t intend anytime soon to harness ICD-10 codes for anything other than reimbursement, let alone running the manner of sophisticated analytics on those specific codes, aka big data, to identify patterns in their patient populations and then based on that information somehow glean a way to alter treatment regimens such that outcomes are hereby improved.

    I’d be all for that, of course, but building on Dr. Billings point about the WEDI/HIMSS National Pilot, is WEDI’s latest readiness survey that found “the industry has slipped further behind key ICD-10 compliance milestones.” Since Feb. 2013, that is. And those results were published this week!

    Providers have to pay for and then use the new codes without seeing near- or long-term benefit. WEDI found that 1 in 5 vendors have yet to provide updates, while months of testing still has to happen between organizations and the federal government which is dragging its feet on such testing — I’m just, you know, speculating but perhaps because it would prove how difficult ICD-10 will be and force its own hand to push back the deadline again, no? And then of course all those small and mid-size providers that literally cannot afford ICD-10 on top of meaningful use, omnibus HIPAA, payment reform — so to continue the weather analogy how many more elements would it take to rightfully consider ICD-10 a perfect storm?

  18. No, the coding can be brief

    V9542XA: Spacecraft crash injuring occupant, initial encounter
    W5922XA: Struck by a turtle
    W5612XA: Struck by a sea lion, initial encounter
    W5609XA: Other contact with dolphin, initial encounter
    S30867A: Insect bite (nonvenomous) of anus, initial encounter
    V0001XD: Pedestrian on foot injured in collision with roller-skater, subsequent encounter.

    All of the above (except the insect bite) could be coded as trauma. Whether you are struck by a space craft, sea lion, turtle or Subaru it is all trauma.

    Do you think it is useful to demand that auto accidents be coded for what intersection they occurred at (like in my anecdote)? Or do you think that an auto accident needs to be coded for the type of vehicle? (What year Ford Taurus was that?)

  19. That’s ridiculous. Why require that the patient be sick? Are you suggesting that providers get stiffed when treating patients that only believe they are sick?

  20. Peter-

    To be socially utilitarian about this for a minute, what are these nine claimed benefits worth in quantifiable dollars and what will they cost our society to get them? I realize it all cannot be reduced to dollars, but if there is no definable ROI to society, we shouldn’t be shamed into doing it because “everybody else is using it”. The proper frame to compare benefits and costs of ICD-10 is at the societal level.

    Having said that, most clinicians I know are already spending, at max, half their time taking care of patients, and the rest of the time feeding their insatiable EHR’s. ICD-10 is merely the straw that breaks the camels back. We’re ALREADY squandering countless millions of clinical person hours documenting and checking boxes

    What we really need is an ICD-11 or ICD-12 which frames the clinical risks of the patient at presentation, not just an absurdly detailed and really useless characterization of their presenting diagnosis.

  21. I guess we only need one code – patient sick, I treated, send the money.

    If docs don’t like the codes they can opt out and go the cash only route.

  22. Peter1,

    What delusional world do you live in where you think epidemiological data would be used to make health care decisions?

    The bureaucrats demand to collect the data that the politicians will not use.

  23. Dear Colleagues,
    I spent several years dissecting the unintended consequences of successive modifications of health policy and health financing since LBJ’s unholy compromise with the AMA that created CPTs which begot DRGs, the RUC and the like:

    As a result, the US entered the 21st Century with the fox in the hen house.
    Maybe the ACA has caused the fox to cower, for a bit. But we if we are moving US “health care” out of the hen house, it’s into the asylum.

    The ICD was a tool designed by epidemiologists to get some handle on the prevalence of disease. It was always a blunt instrument, suffering from lack of reliability and of validity of the categories, a lack that was both temporal and geographic. Why anyone could ever imagine that such soft categorizations could be harnessed to capture more than the crudest measure of the experience of illness is baffling. And anyone who imagines that most illness experiences can be parsed into reliable and valid components is deluded into thinking that “health care” can be managed with the same tools that make Sam’s Club and Costco so successful. Stocking cans of vegetables is as concrete as illness narratives are abstract. All we will do trying to codify illness into its components is create cadres of coders and auditors, administrators and regulators, data generators and analyzers, and … so many that we’d be better off letting the inmates run the asylum.

    The ACA has much that is noble in intent and much that deserves celebration, not the least of which is universal access regardless of prior illness. But treating health care as a “free market” is delusional: There is no caveat emptor since the patient is insured for much that is unnecessary and is blamed for poor outcomes: If the can of vegetables is spoiled, Costco will replace it, apologize and take the issue up with its supplier. Try that in the US medical system. You’ll end up in a tort.


    In addition to discounting caveat emptor, there is no governor on cost in the ACA, even in PCORI:


    Unfortunately, I see no way to head off the impending chaos. I see escalating expense, blame and culpability, lay-offs and bankruptcies, all as the price of the billing scheme that drives the implementation of the ACA, a billing scheme that pivots on ICD-10.

    There are other approaches that are far more rational and that take the patient and the patient’s experience of illness as the primary raison d’être:

    For example: https://thehealthcareblog.com/blog/2010/02/04/the-health-assurance-%E2%80%93-disease-insurance-plan/

    Maybe such will be the Phoenix.

  24. You must work in Chapel Hill in the all-glass floating building on 15-501

  25. I like Consumer injured in collision with web-site, subsequent encounter

  26. I like Pedestrian on foot injured in collision with roller-skater, subsequent encounter.

    Obviously the guy didn’t learn the first time to stay out of the way of a roller

  27. Dave, did you look up the meaning of “sycophant” or do you just like to use it inappropriately cause it sounds smart?

  28. I agree. The financial risk in on the providers.

    Better data is not inherently bad. It is a question of how it is captured. There is no time for physicians to give that won’t take away from their care.

  29. It is all just matter of time and focus. It’s a zero sum game. There is value in granular diagnostic and procedure data. But, physicians are on a treadmill due to our service based health care system. There is only one place they can take the time from and that is direct patient care and engagement. Their focus should be on direct care and quality outcomes, not on point and click documentation of ICD 10. The disturbing news is that this pilot forecasts that physicians may be pushed to the edge or over it by the confluence of storms in 2014.

  30. We have all heard the ICD10 horror stories about systems integration issues, training, finding enough coders (good coders that is), clearing houses (remember how simple 5010′s were meant to be?). However, I do believe the greatest concern has to be with the health plans, both government and private. After all they pay the claims primarily based on the combination of CPT and ICD codes and we know how much they enjoying paying claims.

    Here’s an idea for the Health Plans. Create a cross walk for your plan builders and allow the new ICD10′s to link to the ICD9 in your claims system. If your ICD9 code edits says to pay the claim then pay it (and fix your ICD10 claim logic). Yes, it’ll be a little extra work but the potential mess that could be created by going straight to ICD10′s and not having it mapped correctly is truly staggering.

    Having worked on the dark side previously I know this can be done.

  31. Peter1, I’m honestly not sure if anything you’ve written is satire or if you truly are some lefty sycophant.

  32. Hopefully, we will continue to track infectious diseases by more effective methods than the retroactive analysis of insurance claims processed by high school dropouts.

  33. Obviously Legacy, if you live in Oklahoma there would be no need for those marine codes, so you would not have to use them. However in a coastal city turtles, sea lions and dolphins may be a reason for that code.

    As for the roller-skater, the information could be used to ban the friggen things on sidewalks if too many incidents occur.

  34. “how can it track public concerns”

    It looks like it can more accurately track infectious disease – for the general publics health – not just YOUR patient. If there was an outbreak of a certain infectious disease I would have a concern as to whether I could be in danger and those trying to isolate and treat it would have better information.

  35. “This will allow for more accurate claims payment and coverage decisions”

    Translation: ICD-10 provides us with enormous new opportunities to delay and deny payment. Of course Aetna loves it.

    ICD-10 is a set of frigging billing codes: how can it track public concerns, or any of the other Herculean labors you listed above?

  36. Nine advantages of ICD-10:

    1 Measuring the quality, safety and efficacy of care

    2 Designing payment systems and processing claims for reimbursement

    3 Conducting research, epidemiological studies, and clinical trials

    4 Setting health policy

    5 Operational and strategic planning and designing healthcare delivery systems

    6 Monitoring resource utilization

    7 Improving clinical, financial, and administrative performance

    8 Preventing and detecting healthcare fraud and abuse

    9 Tracking public concerns and assessing risks of adverse public health events

    One example of what an insurer is doing to implement;

    What is Aetna’s position on the final rules?

    “We plan to meet all applicable timeframes for compliance. We will work closely with providers and clearinghouses as they also strive for compliance.

    We feel that the more detailed coding is helpful and brings the United States in line with countries that have already adopted ICD-10. The ICD-10 codes allow for more accurate information to accompany a claim. This will allow for more accurate claims payment and coverage decisions.”

    Don’t you want to anything PCD?

  37. Below are some of the more interesting ICD-10 codes:

    V9542XA: Spacecraft crash injuring occupant, initial encounter
    W5922XA: Struck by a turtle
    W5612XA: Struck by a sea lion, initial encounter
    W5609XA: Other contact with dolphin, initial encounter
    S30867A: Insect bite (nonvenomous) of anus, initial encounter
    V0001XD: Pedestrian on foot injured in collision with roller-skater, subsequent encounter.

    Why on earth would anyone collect this sort of information? Are we planning on doing epidemiological studies on injuries caused by marine animals?

    This reminds me of a story that was told by a former colleague of mine. One of the ER docs at his hospital used to provide “clinical information” about the location of the auto accident – “MVA at Main St. and Oak St.” as if the location made any difference in the diagnosis or treatment.

    primary care doc is right – this is just more bullshit and another (uncompensated) hurdle to jump over. And another reason to deny payment or ding a doc for “billing fraud”. And another reason for doctors and their billing departments to curse the Feds and Insurance companies.

  38. “All ICD-10 impacted organizations should act now to allocate as much time as possible for testing and remediation to protect their corporate bottom lines and cash flow”

    I don’t think it’s the insurers who will see their cash flow harmed by the massive claim denials they will send out beginning on day 1 of ICD-10.

    “It’s not that ICD-10 is an inherently bad idea”

    But it is an inherently bad idea, in that there is no benefit accruing to those who do the work and shoulder the expense.

  39. “It’s not that ICD-10 is an inherently bad idea, or that hospitals and providers can’t meet the challenge with reasonable deadlines. But they have too many challenges right now…”

    “The Department of Health and Human Services announced a one year delay of ICD-10 implementation which is now scheduled for October 1, 2014.”

    “The ICD-9 code set is over 30 years old and has become outdated. It is no longer considered usable for today’s treatment, reporting, and payment processes. It does not reflect advances in medical technology and knowledge. In addition, the format limits the ability to expand the code set and add new codes.
    The ICD-10 code set reflects advances in medicine and uses current medical terminology. The code format is expanded, which means that it has the ability to include greater detail within the code. The greater detail means that the code can provide more specific information about the diagnosis. The ICD-10 code set is also more flexible for expansion and including new technologies and diagnoses. The change, however, is expected to be disruptive for physicians during the transition and you are urged to begin preparing now.”

    Dr. Billings, aren’t you like the college student who only needs “just one more week” to complete their paper?