Opioid use disorder ICD-10 codes and MAT billing

In this article▾

1. What is the opioid use disorder ICD-10 code?
2. Is a patient on methadone or buprenorphine in remission?
3. What is the difference between F11.20 and F11.21?
4. How do you code opioid withdrawal?
5. Where does opioid dependence ICD-10 differ from opioid use disorder?
6. What trips up OTP and office-based MAT claims?
7. How do you code an opioid overdose?
8. What documentation supports an OTP bundle?
9. How the maintenance-therapy coding check can be automated
10. FAQ

TL;DR: The opioid use disorder ICD-10 code for a stable patient on methadone or buprenorphine is F11.20, not F11.21. Maintenance therapy is treatment, not remission. Coding a medicated patient as in remission is one of the fastest ways to undercut the medical necessity of the program treating them.

What this Blog covers

• What is the opioid use disorder ICD-10 code?
• Is a patient on methadone or buprenorphine in remission?
• Difference between F11.20 and F11.21
• How to code opioid withdrawal
• What trips up OTP and office-based MAT claims
• How to code an opioid overdose

Here is a question that costs opioid treatment programs real money. A patient has been on buprenorphine for fourteen months, has not used illicitly in a year, holds a job, and attends counseling weekly. What is the diagnosis code?

If your answer was F11.21, that instinct is understandable and it is usually wrong. The patient is doing well because they are being treated. Remission and successful ongoing treatment are different clinical states, and only one of them belongs on a claim for a program that is actively medicating someone.

What is the opioid use disorder ICD-10 code?

Opioid-related disorders sit in F11, with the same pattern structure as the rest of the block. F11.1 covers abuse, F11.2 covers dependence, and F11.9 covers unspecified use, with further digits describing intoxication, withdrawal, and induced conditions (Blueprint).

One naming note that confuses people searching for guidance. The code set uses dependence, while clinicians and DSM-5 use opioid use disorder. They describe the same territory, and the codes have not been renamed.

Is a patient on methadone or buprenorphine in remission?

Not by virtue of being medicated. F11.20 applies when a client meets dependence criteria without current intoxication, withdrawal, or induced disorders, which includes a stable client on maintenance therapy who is not experiencing acute symptoms but continues to meet dependence criteria (Blueprint).

Read that carefully, because it settles the case. Stability on maintenance medication is compatible with an active dependence diagnosis. The medication is why the patient is stable, which is an argument for continued treatment rather than evidence that treatment is no longer needed.

The billing consequence runs in one direction. A claim for an opioid treatment program bundle carrying a remission code invites the obvious question of why the patient needs the bundle. You have coded away your own medical necessity.

Why does this matter more in OTP settings? Opioid treatment programs bill weekly bundles built around ongoing medication and counseling. The diagnosis has to describe a patient who still needs that structure.

What is the difference between F11.20 and F11.21?

F11.20 says the dependence is active. F11.21 says the provider has determined it is in remission. The difference is a clinical judgment, documented in words, by someone qualified to make it.

F11.21 applies when clients previously met dependence criteria but currently show sustained recovery, and it requires clear documentation of remission duration and the absence of active symptoms (Blueprint). Chapter 5 of the ICD-10-CM Official Guidelines reinforces the point, holding that remission codes cannot be assigned without clear provider documentation specifying remission (ACDIS).

A coder cannot bridge that gap with inference. Twelve months of negative screens is a strong clinical signal and it is not a provider statement, so the active code stands until a provider writes otherwise.

Remission is also not the same as a history-of code. History-of describes something no longer treated or monitored, which does not fit a patient collecting a daily dose from your clinic (ACDIS).

How do you code opioid withdrawal?

Through F11.23, and this is the code that carries a detox or medically monitored admission. Withdrawal is the acute event a payer is approving a bed or an intensive service for, so it needs to appear as a diagnosis rather than only as a symptom list.

The documentation standard is the same one that applies across the block. A COWS score in a flowsheet is a measurement. A provider note reading "opioid dependence with withdrawal, COWS 22 on admission, managed with buprenorphine induction" is a diagnosis that supports a claim.

Facilities lose more money on this than on any coding rule, because the clinical care is usually excellent and the record of it lands in the wrong field. Nobody appeals a denial they believe is a coding technicality, and fewer than 1% of denied claims were appealed in 2024 (KFF).

Where does opioid dependence ICD-10 differ from opioid use disorder?

In vocabulary rather than substance. Clinicians write opioid use disorder because DSM-5 uses that language. ICD-10-CM kept the older abuse and dependence split, so the claim carries dependence terminology regardless of what the assessment says.

This creates the same translation gap that shows up across the block. DSM-5 grades severity by criteria count while ICD-10-CM sorts by pattern, and the two do not map automatically. A note reading "moderate opioid use disorder" leaves a coder without a pattern, and the resulting guess is invisible until a payer questions it.

The fix is a habit rather than a policy. Ask clinicians to name the pattern in the assessment line, so "severe opioid use disorder meets criteria for opioid dependence" reaches the coder completely.

What trips up OTP and office-based MAT claims?

Three things, and only one of them is exotic. The remission miscode is the expensive one. Unspecified codes on claims for structured programs are the common one. Diagnosis mismatch between authorization and claim is the one that hides.

Pull your last quarter of denied MAT claims and sort them by diagnosis code before you sort them by anything else. If F11.21 appears on claims for patients still receiving daily medication, you have found the leak.

Insurers upheld 66% of the appeals that were filed in 2024, which is the number that should shape your priorities (KFF). Appeals are a poor substitute for coding the chart correctly the first time.

How do you code an opioid overdose?

With two codes, not one. An overdose is a poisoning event, which lives in the T40 series, and the underlying disorder stays in F11. When the record documents both an acute poisoning and an underlying dependence, both codes are reported together.

The T40 codes carry their own detail. T40.2X1A is poisoning by other opioids, accidental, initial encounter, with the sixth character describing intent and the seventh describing the encounter type.

This trips up facilities that treat overdose as a severity marker on the F code. It is not. F11.20 with a T40 poisoning code tells the full story, and F11.20 alone tells a payer nothing happened.

Intent matters and it is a clinical determination. Accidental, intentional self-harm, assault, and undetermined each have a different sixth character, and undetermined is the honest choice when the record genuinely does not establish intent.

What documentation supports an OTP bundle?

The diagnosis, the medication, and the counseling, each recorded in a way that ties to the others. Opioid treatment program bundles pay for a package of services across a week, so the record has to show the package was delivered.

An active F11.2 diagnosis is the foundation. Without it, the bundle has no condition to treat, which is why the remission miscode is so costly in this setting.

Beyond that, reviewers look for evidence that counseling actually occurred rather than being scheduled. A note recording that the patient attended, what was covered, and how long it lasted does more for a bundle claim than any coding refinement.

Check your no-show handling as well. Bundles billed for weeks where the patient collected medication but attended nothing are a common audit finding, and the fix is a workflow rule rather than a coding one.

How the maintenance-therapy coding check can be automated

The remission question is a judgment call that has to be made repeatedly, quietly, across every patient in a program, by people under time pressure. That is the shape of the problem ambient software handles well.

Supa works inside your existing systems, so a claim carrying F11.21 for a patient with active medication orders gets flagged before it goes out. Supabill checks the billed diagnosis against the authorized one, catches unspecified F11.9 codes sitting on OTP bundle claims, and surfaces charts where withdrawal was measured but never diagnosed.

The provider still decides whether a patient is in remission. What goes away is the case where nobody noticed the claim said something the chart never did.

If you run an OTP or an office-based MAT panel and your denials are climbing, the diagnosis column is worth an hour. Book a demo and we will go through yours.

FAQ

Q: What is the ICD-10 code for opioid use disorder?

A: F11.20 for opioid dependence uncomplicated, F11.10 for opioid abuse uncomplicated, and F11.90 for unspecified use. ICD-10-CM uses dependence language while clinicians use opioid use disorder, but both point at the F11 family.

Q: Should a patient on buprenorphine be coded as in remission?

A: Generally no, not while they meet dependence criteria and remain on maintenance medication. F11.20 covers a stable maintenance patient. F11.21 applies only when a provider documents sustained remission explicitly.

Q: What is the ICD-10 code for opioid withdrawal?

A: The F11.23 series, describing opioid dependence with withdrawal. It is the code that supports medically monitored detox, and it needs a provider assessment naming the withdrawal rather than only a COWS score recorded in nursing documentation.

Q: Is opioid dependence ICD-10 different from opioid use disorder ICD-10?

A: They refer to the same codes. ICD-10-CM retained the abuse and dependence structure, while DSM-5 moved to a single opioid use disorder with severity levels. Claims follow the code set, so dependence is the term that appears.

Q: Can we code F11.21 after a year of negative drug screens?

A: Only if a provider documents remission. Negative screens support that clinical conclusion but do not substitute for it. Without a provider statement in the record, a coder assigning a remission code has created a documentation gap.

Q: Why would a remission code cause a denial?

A: Because it argues against the service. A claim for ongoing medication and counseling paired with a diagnosis stating the disorder is in remission gives a reviewer a straightforward reason to question whether the treatment is still necessary.