RSSAll Entries in the "ICD-10" Category

Q&A: ‘Code first’ versus ‘in diseases classified elsewhere’

Don't get overwhelmed! Just ask us for help! Leave your question in the comments section below.

Don’t get overwhelmed! Just ask us for help! Leave your question in the comments section below.

Q: I can’t distinguish between “code first” and “in diseases classified elsewhere.” Both are used with manifestations and both can’t be sequenced as principal diagnosis, and both need etiology codes so what is the difference?

A: Technically, not all “code first” notes are mandatory, says Shannon E. McCall, RHIA, CCS, CCS-P, CPC, CPC-I, CEMC, CCDS, director of HCPro’s Certified Coder Boot Camp® programs, based in Middleton, Massachusetts. For example, ICD-10-CM category I50 includes a “code first” note but it is only used if applicable since heart failure can be a principal diagnosis.

The “code first” note informs us of two things, says Allen Frady, RN, BSN, CCDS, CCS, CDI education specialist at ACDIS. First, it informs you that two codes may be required. Second, it provides sequencing direction. There are some “code first” notes that are only applicable in certain instances, such as McCall mentions, at “heart failure” the note talks about assigning codes for hypertensive heart failure, pregnancy related heart failure etc., first, if applicable.  Secondly, if, for example the CDI specialist reviews a Parkinson’s manifestation, such as a dementia, and they see “code first Parkinson’s,” in the tabular list then first code Parkinson’s disease (G20) followed by the code for the dementia F02.80. In this example, the “code first” note is positioned next to the manifestation code to remind you to code the etiology first.

In contrast, the phrase “in diseases classified elsewhere” informs coders that two codes are required and means this code must be sequenced as the second code. If you see “in diseases classified elsewhere” in a code description, then you know you are looking at a manifestation code. These codes should never be used as a principal diagnosis and must be reported in conjunction with a code for the underlying cause/diagnosis.

While these terms may seem very similar, Frady says, the “code first” terminology represents an instructional note while “in diseases classified elsewhere” is actually a part of the code title itself. You would only see the “code first” note if you look up the code in the tabular list and review the instructional notes, whereas you would see “in diseases classified elsewhere” if you were simply reading the code title or description in the alphabetic index.

Interestingly enough, if you index Dementia in a current 2017 code book, these conventions are not used.  The index entry is Dementia (with) Parkinson’s disease:   G20 [F02.80]. In this case the formatting of code first [bracketed code second] provides the sequencing.

Caution is warranted, if you index this condition by looking up the keyword “Parkinson’s,” you get an entirely different code as you get an instructional note to “See Parkinsonism” and following that pathway in either a book or an encoder you will arrive at codes G31.83 and F02.80.

Editor’s note: This answer was provided based on limited information submitted to ACDIS. Be sure to review all documentation specific to your own individual scenario before determining appropriate code assignment. For information regarding coding or CDI Boot Camps visit


Thanksgiving mishaps? There’s an ICD-10 code for that!

It's Thanksgiving! Time to celebrate all the ICD-10 codes you (hopefully) won't actually encounter this holiday.

It’s Thanksgiving! Time to celebrate all the ICD-10 codes you (hopefully) won’t actually encounter this holiday.

As we prepare for the Thanksgiving holiday with family and friends and give thanks for all of our blessings, it is important to be ready with appropriate codes to accurately document any holiday mishaps.

Here’s a short list to help you quickly and efficiently communicate the information required to file a complaint claim:

For incidents with a fresh (live, not saucy) turkey:

  • W61.42 Struck by turkey
  • W61.43 Pecked by turkey
  • W61.49 Other contact with turkey

For general kitchen and meal prep actions:

  • Y93.G1 Activity, food prep and cleanup
  • Y21.2 Undetermined event involving hot water
  • Y93.G3 Activity, cooking and baking

For dealing with obnoxious Uncle Leo who insisted on pushing his way to the dessert table:

  • Y04.2 Assault by strike against or bumped into by another person

For your mother-in-law’s criticism of the lumpy gravy (which we know was not lumpy):

  • Z63.1 Problems in relationship with in-laws

For activities post-meal to work off effects of R63.2 Polyphagia (overeating):

  • W21.01 Struck by football

For Friday morning:

  • W72.820 Sleep deprivation

Editor’s note: The ACDIS office will be closed for the Thanksgiving holiday and will reopen on November 28. Please send along your most common documentation improvement opportunities either in the comment section or via email to

Guest Post (Part 2): Solving the documentation difficulties for sepsis-3

James S. Kennedy

James S. Kennedy

Note: This post is part two of four, excerpted from an article originally published in JustCoding. Read the first installment published on November 15. Click here to read the original.

by James S. Kennedy, MD, CCS, CDIP

In developing a CDI strategy for dealing with new sepsis-3 criteria, remember three environments by which we must consider disease terminology and supporting criteria. One cannot talk about sepsis, severe sepsis, or septic shock documentation without considering the environment in which such documentation is to be interpreted.

Clinical language
Physicians use a language in direct-patient care that communicates (easily translates) well with other physicians. Every physician knows what “urosepsis,” “unresponsiveness,” and “neurotoxicity” is; however, ICD-10-CM does not recognize these terms for coding purposes, thus we ask physicians to use different words so coders can assign the correct coding conventions.

Systematized Nomenclature of Medicine—Clinical Terms (SNOMED-CT) is a clinical language; so is sepsis-3. ICD-10-CM is not.

Not all physicians embrace sepsis-3, thus some may wish to label a patient has having sepsis even if they don’t have organ dysfunction, which makes clinical sense to them.

Coding language
In a landmark article published in the Journal of AHIMA in 2014, Sue Bowman, senior director of coding policy and compliance for AHIMA in Chicago, Illinois, makes it very clear that ICD-10-CM is not for clinical care but for administrative purposes.

“The standard vocabulary afforded by SNOMED CT supports meaningful information exchange to meet clinical requirements, Bowman says. “ICD-10-CM and ICD-10-PCS, with their classification structure and conventions and reporting rules, are useful for classifying healthcare data for administrative purposes, including reimbursement claims, health statistics, and other uses where data aggregation is advantageous,” she wrote.

Sepsis-3 amends clinical language only; however, for coding purposes we must still document using ICD-10-CM’s language, which still recognizes sepsis without organ dysfunction and sepsis with acute organ dysfunction (severe sepsis) and is based on the individual physician’s criteria.

Core measure language
Defining cohorts with core measures, such as SEP-1, is an abstraction based on clinical criteria and not necessarily based on what a physician writes.  For example, the definition of severe sepsis and septic shock is completely different in SEP-1 than Sepsis-3. We must remember, however, that in 2017, if a physician documents severe sepsis and R65.20 is coded, that record will be held accountable for the SEP-1 core measure even if it doesn’t meet the SEP-1 criteria. View this regulation here.

Editor’s note: Dr. Kennedy is a general internist and certified coder, specializing in clinical effectiveness, medical informatics, and clinical documentation and coding improvement strategies. The comments and opinions represent those of Kennedy and not necessarily ACDIS or its Advisory Board and advice given is general. Readers should consult professional counsel for specific legal, ethical, clinical, or coding questions.Contact him at 615-479-7021 or at

Guest Post (Part 1): Documentation and coding challenges abound for sepsis-3

James S. Kennedy

James S. Kennedy

Note: This post is part one of four, excerpted from an article originally published in JustCoding. Click here to read the original. The comments and opinions represent those of Kennedy

by James S. Kennedy, MD, CCS, CDIP

There are a number of coding compliance challenges with sepsis-3 and with sepsis or severe sepsis in general. In this article, I’ll review my top four concerns.

First, sepsis-3 states that patients with an infection meeting the new sepsis criteria should be coded as R65.20, severe sepsis. This is impossible in the United States, given that ICD-10-CM code R65.20 can only be assigned if the physician documents “severe sepsis,” not sepsis alone, or if the physician documents that an acute organ dysfunction is associated with sepsis, though many coders fail to assign R65.20 when these links are made. Its apparent that the sepsis-3 authors are not familiar with Coding Clinic for ICD-10-CM/PCS, the Department of Justice, or our friendly neighborhood recovery auditors (RA).

Secondly, ICD-10-CM still has a multitude of codes for sepsis without organ dysfunction (e.g., A40-A41). The 2017 ICD-10-CM Official Guidelines for Coding and Reporting states that “the assignment of a diagnosis code is based on the provider’s diagnostic statement that the condition exists. It states:

The provider’s statement that the patient has a particular condition is sufficient. Code assignment is not based on clinical criteria used by the provider to establish the diagnosis.” (Emphasis added.)

Recent advice from Coding Clinic supports the concept that if an individual physician documents sepsis using his or her own criteria (that may differ from sepsis-3 or that of a RA), coders are obligated to code it. Therefore, if a physician documents sepsis, can we still defend the coding of an A40-A41 code if there is no documented organ dysfunction? I believe that the Guidelines and Coding Clinic say that we can, even if the RA doesn’t like it.

Thirdly, the ICD-10-CM table instructions for code R65.20, severe sepsis, tell us to use an “additional code to identify specific acute organ dysfunction.” If a physician documents severe sepsis based on the sepsis-3 criteria of a lactate over 2 milliequivalent per liter (mEq/L), or sepsis-3’s changes in the Glasgow Coma Scale, what is the organ dysfunction that should also be coded or queried for? Without an organ dysfunction documented and coded, a RA may claim that the severe sepsis code is invalid.

Finally, in my own personal review of the CMS 2015 MedPAR, approximately 45-55% of MS-DRGs 871 or 872 (septicemia or severe sepsis) do not have a code for severe sepsis, yet a number of patients have acute organ dysfunctions present on admission which I believe should have been linked to the patient’s sepsis to render the severe sepsis code.

RAs look at sepsis DRGs without R65.20, severe sepsis, or R65.21, septic shock, as opportunities to take money away from facilities who coded sepsis (e.g., A40-A41) as present on admission and sequenced it as a principal diagnosis without an additional R65.20 or R65.21 code. To take these records out of the RA data mining pool, CDI professionals must make every effort to query providers if the clinically valid indicators of organ dysfunction due to sepsis are present but the record does not have the documentation interpreting these indicators as to report R65.20 and R65.21 and their associated organ dysfunctions. This effort, however, must be coordinated with the SEP-1 or quality manager, given that any coding of R65.20 or R65.21 subjects the record to the SEP-1.

Editor’s note: Dr. Kennedy is a general internist and certified coder, specializing in clinical effectiveness, medical informatics, and clinical documentation and coding improvement strategies. The comments and opinions represent those of Kennedy and not necessarily ACDIS or its Advisory Board and advice given is general. Readers should consult professional counsel for specific legal, ethical, clinical, or coding questions.Contact him at 615-479-7021 or at

Membership Update: Important ICD-10-PCS survey

Do you have what it takes to become a CDI specialist?

Take our survey!

Notice to our ACDIS members: ACDIS has recently learned that some minor procedures coded in ICD-10-PCS lead to surgical DRG assignments and unexpectedly high payments. This has led to some facilities opting not to code these procedures, believing that it may result in future recoupments from CMS.

Please take a moment to answer this anonymous six question survey. ACDIS plans to alert regulatory coding authorities with these findings in hopes to attain clarification.

Click here to take the survey. 

Thank you.

Gold’s Gospel: Arrested development

Dr. Robert S. Gold

Dr. Robert S. Gold

By, Robert S. Gold, MD

Although we have gone through years of confusion regarding “cardiac arrest” and probably came to some conclusions about when it gets coded and when it doesn’t, this Ghost of CDI Past has come back to haunt us.

The code for cardiac arrest in ICD-9-CM was 427.5. In ICD-10-CM, it has expanded to include elements of the I46 series, where I46.2 represents cardiac arrest due to an underlying cardiac condition, I46.8 represents cardiac arrest from some other underlying condition and I46.9 is for “I have no idea why” cardiac arrest (also known as cause unspecified).

Well, first things first (again): In the process of dying of some chronic or acute disease or traumatic process, the heart stops. No cardiac arrest code is applied for these circumstances at all.

On the other hand, if the patient is not expected to be in the process of dying and something happens and the heart stops, whether it’s called sudden cardiac death or cardiac arrest, then the cardiac arrest code is assigned.

If the cause is known, or pretty clear even if it’s not known, then the specific code is assigned. Cardiac arrest due to ventricular fibrillation gets the I46.2 code as well as the specific code for ventricular fibrillation (I49.01) for the added specificity.

When the cause is hypercalcemia or hyperkalemia, then we have the I46.8 code plus the specific code for the electrolyte disturbance that led to the cessation of heartbeat. OK, that’s one. We code it when it’s appropriate to code it.

But, as I have heard in ICD-9-CM and am now starting to hear in ICD-10-CM, “We’re giving CPR, which is ‘cardiopulmonary resuscitation,’ and the patient’s oxygen saturation is 60%, so obviously the patient has acute hypoxic respiratory failure, right?” Wrong!

When the heart stops, whether while dying from lung cancer or responding to a massive ST elevation myocardial infarction at the origin of the left anterior descending coronary artery, breathing stops, renal function stops, the brain function stops—the patient dies—unless circulation can be restored. So all this advice of getting doctors to document “acute hypoxic respiratory failure” just because the oxygen saturations are low is bogus.

Does this mean that the two can never be coded together? Not at all! If the acute hypoxemic event precedes the cardiac arrest, as in drowning or smoke inhalation or acute pulmonary edema, and that is followed by a fatal arrhythmia or a myocardial infarction which stops the heart suddenly, then it is quite proper to look at having both events documented and coded. But you don’t shoot for a diagnosis just because of a lab result and your opinion that “it just makes sense.”

Editor’s Note: This post was originally published in Just Coding.

ICD-10 Tip of the Week: Systolic and diastolic heart failure

Don't let documentation for AMI give you a heart attack.

CDI specialists should look for evidence to support the type of heart failure. 

Once a physician documents heart failure, CDI specialists should look for evidence to support the type—either systolic or diastolic.

The terms systolic and diastolic heart failure describe the pathology of the heart failure, which can affect both the left and right sides of the heart, resulting in symptoms associated with either and/or both of these conditions. Consequently, it is difficult to diagnosis the pathological cause of heart failure by symptoms alone.

The CMS quality measure (HF-2) for inpatient care requires that any patient admitted with heart failure has documentation of a left ventricular systolic (LVS) function study in which the evaluation occurred before arrival or during hospitalization or is planned for after discharge. As such, it is likely that the medical record will contain this value, and that can assist with determining the type of heart failure.

Although codes cannot be assigned based on documentation from a previous visit, this information can be used as a clinical indicator for a specificity query. Remember that a new diagnosis cannot be introduced in the current record; the old record would only be a source to support the clinical indicators within a query.

Editor’s Note: For information, check out The Clinical Documentation Improvement Specialist’s Guide to ICD-10.

ICD-10 Tip of the Week: Using ICD-9 guidance in ICD-10

Use the query practice brief guidelines formulate effective queries.

Stop and reflect when using ICD-9 Coding Clinic guidance, and ask yourself if it in any way conflicts with the new guidelines.

The new AHA Coding Clinic for ICD-10-CM and ICD-10-PCS included some new language, which could help clear up confusion for CDI specialists and HIM/coding professionals regarding use of prior ICD-9 guidance.

By way of background, in 2010 the AHA stated that previous issues of Coding Clinic would not be translated to ICD-10-CM/PCS. The announcement led many to question if Coding Clinic advice could even be applied to ICD-10 coding practices, and if auditors will deny claims based on advice from AHA Coding Clinic for ICD-9-CM.

Previously, the AHA responded stating that past advice has focused on what documentation could be used and has not been specific to a coding system. However, in the December 2015 Coding Clinic, the AHA wrote:

In general, clinical information and information on documentation best practices published in Coding Clinic were not unique to ICD-9-CM, and remain applicable to ICD-10-CM with some caveats. For example, Coding Clinic may still be useful to understand clinical clues when applying the guideline regarding not coding separately signs and symptoms that are integral to a condition. Users may continue to use that information, as clues—not clinical criteria.

As far as previously published advice on documentation is concerned, documentation issues would generally not be unique to ICD-9-CM, and so long as there is nothing new published in Coding Clinic for ICD-10-CM and ICD-10-PCS to replace it, the advance would stand.

In response, Laurie Prescott, RN, MSN, CCDS, CDIP, AHIMA Approved ICD-10-CM/PCS Trainer, CDI Education Director at HCPro in Danvers, Massachusetts, says any ICD-9 Coding Clinic that describes clinical situations or general guidance that is consistent with the ICD-10-CM/PCS Official Guidelines for Coding and Reporting can likely be applied.

Coders and CDI specialists may come across scenarios where guidance is consistent, and other situations where guidance is contradictory, Prescott says. Stop and reflect when using ICD-9 Coding Clinic guidance, and ask yourself if the ICD-9 guidance in any way conflicts with the new code set, coding conventions, and guidelines.

If there is a conflict, Prescott says do not rely on the ICD-9 Coding Clinic and instead follow the guidance within the ICD-10-CM/PCS Official Guidelines for Coding and Reporting. 

Editor’s Note: This post was based on an article originally published in CDI Strategies. To learn more, click here.

ICD-10 Tip of the Week: What’s next?


Many facilities are wondering what the next steps are for their CDI program.

This past weekend marked the one-month anniversary of ICD-10-CM/PCS implementation. The date that many feared would destroy their organization came and went—without much to-do. In fact, a recent poll on the ACDIS website indicated that 42% of respondents said their facility it handling ICD-10 implementation fine with no real problems, and 36 % are handling implementation okay with only a few minor documentation or coding hiccups. Less than 15% say implementation has been “not great.”

While it’s still too early to call ICD-10 implementation a success, some facilities are already asking “what’s next?”

Facilities may be eager to move ahead with the next project, but don’t rush. Depending on issues identified throughout this first month, ICD-10 response teams—IT, HIM, and CDI staff, at minimum— should continue to meet once or twice a week to address and answer questions that arise relating to documentation, CDI, and coding.

If your team hasn’t started one, create a log of questions, resolutions, and outstanding concerns. Develop a plan for continued education for physicians, coding, and CDI staff, which targets newly identified problem areas.

Be sure to include the physician champions or medical staff leadership in these discussions, and address physician concerns in a timely manner. The goal should be to resolve documentation issues as they occur, so physicians can learn and the problem doesn’t arise again.

B early 2016 response teams and/or leadership should start to identify system issues and a corrective action plan if problems exist. Run reports for an MS-DRG shifts and analyze for potential opportunities and improvements in areas such as case-mix index, DRG volumes, and MCC/CC capture. Determine financial impact of ICD-10 on outpatient services and use national coverage determinations for assessments.

Editor’s Note: This post was compiled using a number of ACDIS resources. For more information, check out the following:

ICD-10 Tip of the Week: Make implementation a fresh start

Do you have what it takes to become a CDI specialist?

ICD-10 is a fresh start for both physicians and CDI specialists.

Physician education has been a challenge for CDI specialists long before ICD-10 implementation. Now that we’re nearly three weeks in, some physicians may start to feel that they don’t need continuing ICD-10 education. The reality is ICD-10 will be an ongoing learning process for everyone. But, how do you keep physicians interested in what you to say?

Erica Remer, MD, FACEP, CCDS, clinical documentation integrity officer of University Hospitals in Cleveland, Ohio, said CDI specialists need to view this time as an opportunity to not only continue ICD-10-related education, but to embed standard CDI information to keep them engaged. “This is a great opportunity to go through the things that [the physicians] weren’t doing right in ICD-9 and reinforce it with ICD-10 education,” she said during the October 14 ACDIS Radio call. “This is a fresh start and one we should be capitalizing on.”

When placing a query, teach them why you had to query in the first place, Remer suggests. Educate them so you don’t have to query them and so they document properly—fewer queries is certainly an incentive to document correctly!

Physicians want to know what’s in it for them—what they get out of documenting more specifically, Remer says. When approaching physicians with an ICD-10 subject, reiterate things like quality scores, or physician ratings on sites like and Make sure they know how documentation affects not just the hospital but their individual jobs and reputations. This helps get physician buy-in, says Remer.

CDI specialists should also take ICD-10 implementation as a  “fresh start,” an opportunity to evaluate their own practices. At University Hospitals, CDI specialists make a point to query in a more timely fashion and stay on top of operative notes to make sure they are completed quickly, Remer says. “You can’t do procedures (PCS) without an op[erative] note,” says Remer. “It’s crucial now [for documentation]. What we would really like to prevent is having to query [the physician] twice.”

Editor’s Note: This post was written using a number of ACDIS resources, including: