We know that lots of good things come in threes—like three-piece suits, the Three Stooges, the three French hens your true love gave to you, and the number of sheets to the wind you might be after happy hour. And now, that list includes our ICD-10 FAQ series. In our first two pieces (here and here), we covered a lot of ground and—hopefully—gave you a good mid-level understanding of ICD-10. This third installment dives into the nuances of using the new code set to help you address some of your trickiest questions. So, let’s get to it:
Does my treatment diagnosis have to match the diagnosis the referring physician sent? Sometimes my physicians don’t send codes with their referrals.
Your treatment diagnosis doesn’t necessarily have to match your referring physician’s diagnosis. With ICD-9, therapists typically used only the treatment diagnosis codes, leaving off the “true” diagnosis codes (i.e., medical diagnosis codes), because insurance companies only required treatment diagnosis codes for payment.
As the healthcare system shifts into a pay-for-performance environment—and with the introduction of alternative payment models—providers now must focus more on the complexity of the patient. And as far as claims go, the only way providers can communicate that complexity is through diagnosis codes. So, with ICD-10, it’s critical to include not only the treatment diagnosis codes, but also the codes for the medical diagnosis and comorbidities. This allows you to paint a complete, accurate, and detailed picture of the patient and his or her situation. Remember, though, that the primary code will still be your treatment diagnosis.
Thus, if your referring physician did not send a medical diagnosis, I would suggest reaching out to obtain one, as that will help ensure you code as thoroughly as possible.
When I get scripts with very generic instructions (e.g., “treat the legs”), is it okay for me to use an equally generic code (e.g., R29.9, Unspecified symptoms and signs involving the nervous and musculoskeletal systems)?
When you use generic—or in ICD-10 language, “unspecified”—codes, you risk payers inferring that your services are not medically necessary. So, it’s really up to you as to whether you’re willing to take that risk. We recommend either:
- reaching out to the referring physician to get more detail, or
- referring the patient to another specialist if, based on your clinical judgment, you believe the patient needs further evaluation that falls outside of your scope of practice.
In the ICD-10 code set, some codes have a dash ( – ) at the end of them. What does this mean?
The dash often is used in the note sections (e.g., Excludes1, Excludes2, and Code also). It indicates that the note applies to all of the codes in that series. That way, the codes don’t have to be listed out individually, which would take up a lot of space. So, to give an example, “T79.A-” would cover every single code that begins with “T79.A.”
Specific Coding Questions
How do I code for “X” patient with “Y” condition?
We’ve received numerous questions about how to code in very specific patient scenarios. However, because ICD-10 places such a strong emphasis on clinical judgment—which requires you, as the therapist, to select the most complete, accurate, and specific code(s) possible based on your assessment of the patient—we cannot provide coding advice for those scenarios. But, if you are able to provide supporting documentation, you can submit specific coding questions to this AHA portal.
What code(s) should I use to indicate decreased range of motion (ROM)?
While there are no codes for decreased ROM specifically, there are plenty of other codes that would apply to patients experiencing this symptom. For example, if a patient presents with decreased ROM in the knee, applicable codes may include those for difficulty walking or gait abnormality (R26.2 or R26.89), knee pain (M25.561 or M25.562), knee stiffness (M25.661 or M25.662), or knee effusion (M25.461 or M25.462).
I’m having trouble finding specific codes for conditions affecting the lumbar region (e.g., lumbar stiffness or decreased ROM in the lumbar spine). What are your suggestions?
While we are not trained coders, we can say that with ICD-10, you are encouraged to first code for the underlying condition causing such spinal symptoms. Here are a few resources that may help with coding spinal conditions:
What codes should I use to indicate muscle weakness in a specific anatomic location? The only muscle weakness code I can find is M62.81.
If the patient is experiencing weakness in one specific area, then the code for generalized muscle weakness (M62.81) likely would not be appropriate, as it indicates the weakness is present throughout the body. To code for weakness in a specific area, use the appropriate atrophy code (you’ll find many of these codes in the M62.5 code family). Additionally, keep in mind that the coding guidelines encourage you to code first for the underlying cause of the weakness, rather than the weakness itself. If you are able to do that, you can include the atrophy codes as secondary.
What ICD-10 code would you recommend for stiffness of the lumbar spine?
Whenever possible, you should first code for the underlying condition that is causing the symptom (which in this case is stiffness). Here is a resource that lists ICD-10 codes for common spinal conditions. As a secondary code, you could use M25.60, Stiffness of unspecified joint, not elsewhere classified (keep in mind that there is not a spine-specific option in that family of codes).
How do I code for impaired balance?
While there is not an ICD-10 code for impaired balance, there are several options that provide a much higher degree of specificity. These include the codes listed in the R26 (Abnormalities of gait and mobility) and R27 (Other lack of coordination) series. To select the code that best describes the patient’s condition, you’ll need to use your clinical judgment. You also may need to reference the coding synonyms provided in resources like this one.
Can I submit M54.5 (Low back pain) as a single, primary code?
Yes, M54.5 is a complete, billable code, and thus, you can use it as the primary. However, because it’s not a very specific code, you should only use it as the patient’s primary diagnosis code if there’s not a more specific code available to accurately describe the patient’s condition. If the patient has a confirmed underlying diagnosis (i.e., the condition actually causing the back pain), then you should code for that first. If not, then make sure you explain those details in your documentation.
I thought that the only seventh character options were A, D, and S, but I’ve noticed some codes have additional options (G, K, P, etc.). Is this a different seventh character system?
Yes. Codes for fractures use a whole different set of seventh characters:
- A: Initial encounter for fracture
- D: Subsequent encounter for fracture with routine healing
- G: Subsequent encounter for fracture with delayed healing
- K: Subsequent encounter for fracture with nonunion
- P: Subsequent encounter for fracture with malunion
- S: Sequela
You can learn more about fracture coding in ICD-10 here.
External Cause Codes
If a patient doesn’t know what caused his or her injury, or if there’s no evidence that a specific event (e.g., a past car accident) resulted in the patient’s current condition, then do I need to use external cause codes?
You cannot code for what you don’t know. So, if you don’t know what caused the patient’s injury—or if there’s no definitive correlation between a known causal event and the patient’s current condition—then don’t code for it. Remember, external cause codes (like those that denote accidents) are optional. Just make sure you accurately record any potentially relevant information within the patient’s documentation.
I heard that S codes (i.e., injury codes) and M codes (i.e., musculoskeletal codes) cannot stand alone. What should I do if I don’t know any of the details necessary to select external cause codes (cause, location, activity, etc.)?
There is no blanket instruction regarding M codes and S codes being unable to stand alone. While you are encouraged to submit external cause codes when possible, they are not required. And if you do not know the details necessary to select those codes, then you shouldn’t submit them (in other words, you cannot—and should not—code for what you don’t know). However, you may see notes at the top of specific code categories instructing you to submit additional diagnosis codes with the codes in those particular sections (e.g., “Code also” or “Code first”).
The official ICD-10 coding guidelines state that external cause status codes, activity codes, and place of occurrence codes are “used only once, at the initial encounter for treatment.” Does that mean you only submit these codes on the claim for your initial evaluation of the patient?
Yes, according to the guidelines, you only need to submit those codes once, and that’s at the patient’s initial encounter (i.e., you initial evaluation). However, we have yet to see what will happen if those codes are submitted on subsequent claims. Different payers may have different rules, and some may enforce them more strictly than others. We don’t anticipate claims being denied for having “extra” codes; that said, if and when ICD-10-related denials start to occur, we’ll keep our Blog readers informed on the most common reasons behind those denials.
The above-cited coding guidelines excerpt seems to conflict with the excerpt that reads, “Assign the external cause code, with the appropriate 7th character (initial encounter, subsequent encounter or sequela) for each encounter for which the injury or condition is being treated.” I’m confused; should I submit external cause codes at every visit or just at the first visit?
Although Chapter 20 of the ICD-10 code set is titled “External Causes of Morbidity,” this chapter actually contains four different types of codes:
- place of occurrence codes,
- activity codes,
- external cause status codes, and
- external cause codes.
So, while you should report the main external cause code at every encounter, you only need to report the other types of codes at the initial encounter. Furthermore, “When applicable, place of occurrence, activity, and external cause status codes are sequenced after the main external cause code(s). Regardless of the number of external cause codes assigned, there should be only one place of occurrence code, one activity code, and one external cause status code assigned to an encounter.”
My billing system limits the number of codes I can submit. If I cannot submit all of the external cause codes that apply to a particular patient, how should I choose the one(s) I do submit?
Per the official reporting guidelines, “If the reporting format limits the number of external cause codes that can be used in reporting clinical data, report the code for the cause/intent most related to the principal diagnosis.”
Should I always use an aftercare code for post-op patients?
Per the official ICD-10 coding guidelines (which you can find here), “The aftercare Z codes should not be used for aftercare for conditions such as injuries or poisonings, where 7th characters are provided to identify subsequent care. For example, for aftercare of an injury, assign the acute injury code with the 7th character ‘D’ (subsequent encounter).” Based on those guidelines, if the surgery resulted from an injury that allows for the seventh character, you would use the original acute injury code with the seventh character “D.”
What is the difference between the aftercare codes Z48.89 (Encounter for other specified aftercare) and Z47.89 (Encounter for other orthopedic aftercare)?
Z48.89 applies to all types of surgical aftercare, whereas Z47.89 applies to orthopedic procedures specifically.
Can I submit a post-surgical aftercare code (e.g., Z51.89, Encounter for other specified aftercare) as the patient’s primary diagnosis code?
Your treatment diagnosis should be the one that best supports the medical necessity of your therapy services, and if you submit more than one treatment diagnosis code, you should order them according to significance. The ICD-10 coding guidelines seem to suggest that the order may change based on the specific context of the patient’s situation. The guidelines read: “Z codes may be used as either a first-listed (principal diagnosis code in the inpatient setting) or secondary code, depending on the circumstances of the encounter.” In other words, it’s up to you (and your clinical judgment) to determine which code best represents the main reason the patient needs therapy treatment.
If I am providing post-surgical rehab therapy treatment, and the patient’s original condition (i.e., the one that prompted the surgery) is no longer present, which codes can I submit besides the aftercare code?
If the condition that caused the patient to undergo surgery no longer exists, then you should not code for it. Instead, you can submit any applicable pain/symptom codes relevant to the patient’s treatment. Ultimately, the codes you use, and the order in which you submit them, are up to you and your clinical judgment. Just make sure your documentation clearly supports your coding choices.
Considering the degree of specificity ICD-10 affords, it seems strange that there are not surgical aftercare codes for specific surgeries (e.g., rotator cuff repair or ACL repair). How, then, should I code for these post-op patients?
While there is not an aftercare code for every single surgery, in many cases, the proper way to designate the phase of treatment (i.e., indicate that the patient is receiving aftercare) is to code for the original acute injury and add the appropriate seventh character (which in this case, would be D). So, if, for example, the patient originally strained his or her right rotator cuff, you would indicate that you are providing aftercare by using the code S46.011D, Strain of muscle(s) and tendon(s) of the rotator cuff of right shoulder, subsequent encounter. For more information on coding for aftercare, check out this blog post.
If I submit multiple diagnosis codes on a single claim, which one should I point to on my service lines?
You can submit up to four diagnosis code pointers per service line, which means you can point to up to four diagnoses.
If I have to resubmit a claim for a date of service prior to October 1, 2015 (meaning the claim contains ICD-9 codes), do I need to change the codes to ICD-10?
Because the transition to ICD-10 goes by date of service—not the date of claim submission—you should continue using ICD-9 codes for all claims with dates of service on or before September 30.
We bill electronically, and it is my understanding that we should be able to submit up to 12 diagnosis codes per claim. However, my billing software only sends the four top codes through to the claim. Why is that?
Because the standard HCFA claim form allows space for only four diagnosis pointers per service line, some billing systems only pull the four most relevant diagnosis codes through to Box 21 on the claim form. Those codes then become the origins for the four diagnosis code pointers that appear in Box 24E. We recommend looking for a billing system that allows you to send up to 12 codes through to Box 21; however, regardless of whether your system limits the number of diagnosis codes you can submit, it’s crucial that you order your codes according to significance, with those codes that most strongly support the medical necessity of your treatment appearing first.
If ICD-10 is keeping you up at night, this post should help give you the guidance you need to get a few more hours of shut-eye. But, if our third ICD-10 FAQ wasn’t quite the charm, posit your ponderances to us using the form below, and we’ll do our darndest to answer them.