ICD-10: Getting Sucked in and Surviving

During my fellowship I did some urgent care moonlighting. Sometimes when things got slow we would turn to the International Classification of Diseases, Ninth Revision (ICD-9) coding book for entertainment. It was a bit frightening to see how many codes there were for injury from a nuclear weapon-16, to be exact. My favorite code, though, was E844, “sucked into a jet (aircraft).” I've always wondered what exactly that means.

Jeffery C. Ward, MD

Today if I spend time with ICD-9, it is as likely as not to look at lymphoma codes. ICD-9 was developed by the World Health Organization (WHO) over 30 years ago. Back then, lymphoma nomenclature was based on the Rappaport classification first proposed in 1956 and then modified in 1965. Since then we have been through the Working Formula, Revised European-American Lymphoma (REAL), and WHO classification schemes. No oncologist or hematologist trained since 1982 has used the Rappaport nomenclature, but ICD-9 does, and every so often I have to go back to the time when reticulosarcomas roamed the neoplastic landscape.

Granted, updates have been made from time to time, and now there are codes to distinguish marginal zone and mantle cell lymphoma, but there is only room for 2 groupings of 10 each in ICD-9 code, and the WHO classification defines 80 lymphomas in 4 broad groupings. Simply put, when it comes to lymphoma, we are overdue for ICD-10.

A Little ICD-10 History

ICD-10 was endorsed by the World Health Assembly in 1990. It consists of two sets of codes: procedural codes for hospital inpatient use (ICD-10-PCS) and diagnostic codes (ICD-10, Clinical Modification, or ICD-10-CM).[1] ICD-10 was adopted by the WHO in 1993, and the developed world followed suit… with one exception. It has been 17 years since the United Kingdom adopted ICD-10. Canada, a slow adopter, began using it 11 years ago.

The United States has been the lone holdout for a long time, although our resistance has not been complete; our tumor registries have been using ICD-10 topography codes for 20 years. However, with the passage of the Health Insurance Portability and Accountability Act (HIPAA) of 2009, the Centers for Medicare and Medicaid Services (CMS) were ordered to make the switch. Initially set for October 1, 2011, implementation has been delayed twice, most recently on February 16, 2012. The newest line in the sand is October 1, 2014, and the Department of Health and Human Services (HHS) says that they really mean it this time.[2]

Switching from ICD-9 to ICD-10 has its dedicated proponents and vociferous opponents. Among the former are those who envision ICD-10 as the ticket to the promised land of a 21^st century healthcare system in which higher quality can cost less (the “less is more” argument is a topic for another day), and among the opponents are some who see a government plot that is good only for payers and Medicare Recovery Audit Contractors at the expense of patients and physicians. The truth is somewhere in between.

ICD-10: First, the Good News

There are many reasons to want something better than ICD-9, and ICD-10 has significant advantages over ICD-9. Whether “better” is actually “good” is a point of contention. It has been suggested that we should cobble ICD-9 together at this point and wait for ICD-11. Unfortunately, ICD-11 is at least another decade away. Medicine is moving too fast to wait that long. Here are the relatively good things about ICD-10.

First there is that thing about ICD-9 being full. It is a three- to five-character system. The first character is numeric or alpha, and the rest are numeric. A decimal point is placed after the third character. In the world of oncology, “neoplasia” is assigned codes 140 through 239. “Malignant hematology” gets 200 through 208, and “lymphoma” is allotted three of these. One is for “Hodgkin's disease” and the other two are for “non-Hodgkin's lymphoma,” although the actual term “non-Hodgkin's” is glaringly absent. The fourth character, the first after the decimal point, refers to the subtype of non-Hodgkin's lymphoma. Since we employ a base-10 counting system and the fifth character (always a digit) indicates the stage or site of disease, this leaves room for only 20 of the 80 subtypes of non-Hodgkin's lymphoma. There are many such examples of ICD-9 bursting at the seams.

ICD-10 has seven characters, with a decimal point after the first three. The first character is alpha, the second and third are numeric, and the other four can be alpha or numeric (with room for future expansion). The fact that there are 26 letters gives this system a whole lot more room before the decimal point, and the doubling of the number of characters following the decimal point gives it tremendous capacity for subtyping and descriptors, enabling it to accommodate both vertical and lateral growth of the diagnostic code.

Many consider this capacity for lateral growth to be critical as we move into the world of healthcare reform. Accountable care organizations cannot be accountable without the ability to measure the care that they are providing. For example, not only will all the lymphomas and their stages fit into ICD-10, but so will information about the context of the patient's encounter with the physician, such that the codes can distinguish between a new diagnosis, an ongoing problem, or a survivorship visit. It is argued that the lateral growth of the coding system will also greatly improve our ability to track public health and to conduct research, particularly in the health services arena.

I'm not entirely sure that this belongs under good things, but it can be argued that the timing has never been better to convert to ICD-10. Medicine is rapidly, albeit belatedly, moving to computerization. By now we have all met the January 1, 2012 deadline for converting our billing software to Version 5010, which is necessary to support ICD-10 on the billing end of things. On the front end, paper charts and ICD-10 would be a disaster, but with computerized documentation, the level of detail necessary to allow for ICD-10 coding can be prompted within the physician's initial document. Software that can extract information (such as coding and Quality Oncology Practice Initiatives [QOPI] measurements) directly from physician documentation, as well as perform other expensive and time-consuming chores, cannot be too far in the future.

CMS would also have us believe that after we get over an initial hump, ICD-10 will result in improved efficiency in billing offices, with a decreased need for supporting documentation for claims and fewer denials and appeals. Of course, they also see ICD-10 as a tool for identifying currently undetected fraud and abuse that is able to blend in with appropriate claims in the underpowered ICD-9.

Payers, like CMS, see a lot of good things in ICD-10. They view it as a simple cost and opportunity equation: an expensive implementation that is worth the opportunity it provides to more clearly identify the necessity of care. Payers and providers would both like the costs required in time-consuming pre-authorization of services to fall by the wayside. ICD-10, coupled with a critical appraisal of a payer's policies and procedures, could allow for significant streamlining, to the benefit of payers and providers alike.

ICD-10: Now the Bad News

On February 2, 2012, Dr. James Madera, CEO of the American Medical Association, sent a letter to HHS Secretary Kathleen Sebelius urging her to “immediately halt the HIPAA-required implementation of ICD-10.”[3] Two weeks later to the day she did just that, although ultimately the “halt” proved to be a delay of just 1 year. The gist of Madera's letter, slanted toward the specific concerns of oncologists, is outlined in the paragraphs that follow.

ICD-10 is a massive undertaking that expands our current 13,000 codes to 68,000 codes. It will impact nearly all business processes in a physician's practice: verification of eligibility, pre-authorization, clinical documentation, research activities, public health reporting, quality reporting, and claim submission. It will require education, software implementation or updates, coder training, and coordination and testing with vendors and payers. The cost of implementation is projected to cost $83,000 for a 3-physician office and $285,000 for a practice of 10 physicians.[4] It can be anticipated that for oncology, with its already complicated billing systems, implementation will cost more than average. This is at a time when oncology reimbursement has already taken a significant hit as a result of the Medicare Modernization Act, and when sustainable growth rate (SGR) cuts continuously loom on our horizon. It is conceivable that many oncologists in private practice would find this to be the last straw.

The comparatively simple conversion to version 5010 billing software this year has not been without significant problems, resulting in disruptions in claims processing and delayed payments. It is noteworthy that these payer problems disproportionately hurt providers.[5] This is dramatically amplified when physician services include chemotherapy. Presumably small delays can result in cash flow problems that put a practice seriously in arrears with regard to financial obligations to distributors, employees, and landlords. Stories of practices that can no longer buy drugs on credit or that have had to rely on physicians' personal loans to meet financial obligations have fueled the migration from private to hospital-based practices. If the ICD-10 conversion is allowed to put more pressure on these stress lines, patients will suffer.

Staring down ICD-10 would be a lot easier if it didn't have a whole lot of company. Physician practices are already struggling with mandates for e-prescribing, electronic health record (EHR) meaningful use, and the Physician Quality Reporting System (PQRS). The struggle to keep up with all the various IT and reporting requirements leaves little time for oncologists to get engaged in finding their place in accountable care organizations or patient-centered medical oncology homes as envisioned by the Affordable Care Act. It was, perhaps, this argument and a chart showing the penalties associated with noncompliance with each mandate that carried the day with Secretary Sebelius.

However, there is more. Implementation of ICD-10 will have a huge impact on physician productivity. Even among proponents, there are misgivings about the demands that ICD-10 will place on physician documentation. Gone are the days of “breast cancer NOS.” A physician note that does not indicate that the breast cancer is an estrogen receptor–positive breast cancer in the upper outer quadrant of the patient's left breast without evidence of metastases and that it has been completely excised will be getting coding queries up to ying yang. Where physician financial productivity is tied to high volume and poor documentation, throughput will have to take second seat.

When all is said and done, it is not a given that ICD-10 will make good on its promises. ICD-10 is heavy into anatomy and light on histology. It is much more interested in knowing what lobe a metastatic lung cancer began in than in descriptors such as cancer cell type, stage, and histologic grade, which are far more important in determining what drug to use. The improvements in ICD-10 do more for primary care, ambulatory care, mental health, and preventive medicine encounters than they do for specific disease states such as cancer. Time will tell, but ICD-10 may prove to be little more than a very expensive stop gap to an ICD-11.

Taking Ownership and Getting Prepared

Short of burning tires in the streets (and maybe even then), ICD-10 seems bound to happen, and whether you herald it as a dramatic improvement or the lesser of two evils, putting off preparation would be a mistake; better to take ownership of the problem and command of the situation.

The first step in preparation is to take an appraisal of where you are at now, and that means not only your practice but outside entities such as your EHR and other software vendors, claims clearinghouses, outsourced billing services, and payers. If they aren't ready, you aren't either.

In the process of this stock-taking, you will identify diagnosis coding tools, superbills, public health reporting tools, compliance plans, and documentation templates, to name a few, that will need to be updated and translated from ICD-9 to the new Greek of ICD-10. It is important to recognize that this project is much more than a billing issue-you must be sure to identify all those who will need training in ICD-10 and what that training will entail. Once the size of the task is understood, a budget and timeline can be established and oversight assigned to practice management. A physician champion, if you can roust one, will prove invaluable.

Staff training will be an arduous task, but the rewards of a job well done will be self-evident. When Canada converted to ICD-10, it reported a 55% reduction in productivity that took 6 months of on-the-job training to reverse.[1] It may well be that the most critical and resistant group to train will be physicians who are reluctant to tighten up their notes. The goal is not to turn doctors into coders, as they should focus on being good clinicians; however, good documentation is good clinical practice.

These steps should all be accomplished well before the deadline for compliance, to allow time to conduct in-depth internal testing, so as to ensure that you can generate transactions with ICD-10 coding accurately and efficiently. Then turn to your vendors and payers to make sure that ICD-10–coded transactions can be sent and received.

Even then, be prepared for possible delays in payments and negative impacts on revenue cycles. If cash flow could be a short-term problem, establish lines of credit ahead of time. Keep in mind that on a longer-term basis, industry estimates ominously suggest that it could take up to 5 years post ICD-10 to stabilize cash flow.[6]

Finally, prepare to specifically monitor ICD-10 coding and billing on the back end for at least 6 to 12 months. Identify best practices in the clinic and bring staff members who lag behind back for additional training sooner rather than later. Identify high-frequency diagnostic problems and fix them. At the end of the monitoring program, integrate your findings into your compliance program going forward.

Choosing to Survive

I'm still not clear what “getting sucked into a jet (aircraft)” means. But given that it is a diagnostic code, I figure that it means that it is possible to survive it. After a hard look at ICD-10, the good things and the bad things, and trying to put it all into the context of 21^st century healthcare and my oncology practice, I don't think that ICD-10 is a jet (aircraft). Whatever it is, though, we are getting sucked into it, and we might as well plan to survive.

Financial Disclosure:The author has no significant financial interest or other relationship with the manufacturers of any products or providers of any service mentioned in this article.

References:

REFERENCES

1. American Academy of Professional Coders. ICD-10: the history, the impact, and the keys to success. March 2012. [Internet] Available from: http://www.aapc.com/icd-10/icd-10-white-paper.aspx.

2. Centers for Medicare and Medicaid Services Publication. ICD-10-CM/PCS: an introduction. April 2010. [Internet] Available from: http://www.cms.gov/ICD10.

3. American Medical Association Press Release. HHS announces intent to delay ICD-10. February 16, 2012. [Internet] Available from: http://www.cms.gov/ICD10.

4. MGMA Press Room. New study finds ICD-10 mandate hardship for health care providers. October 14, 2008. [Internet] Available from: http://www.mgma.com/press/article.aspx?id=22612.

5. American Medical Association. Version 5010 electronic administrative transactions. [Internet] Available from: http://www.ama-assn.org/ama/pub/physician-resources/solutions-managing-your-practice/coding-billing-insurance/hipaahealth-insurance-portability-accountability-act/transaction-code-set-standards/version-5010-electronic.page.

6. RAND Corporation Technical Report Series. The costs and benefits of moving to the ICD-10 code sets. March 2004. [Internet] Available from: http://www.rand.org/pubs/technical_reports/2004/RAND_TR132.pdf.