Alcohol use disorder ICD-10 codes that get paid

In this article▾

1. What is the alcohol use disorder ICD-10 code?
2. Why does DSM-5 severity not tell anyone which code to use?
3. The four words that can solve it
4. 5 Common pitfalls that you can avoid
5. When is F10.10 actually the right call?
6. What a payer is actually reading in your notes
7. Where the diagnosis meets the authorization
8. How Supa handles the translation
9. FAQ

TL;DR: Alcohol use disorder ICD-10 coding breaks at the handoff between the clinician and the coder. DSM-5 collapsed abuse and dependence into one disorder with severity tiers. ICD-10-CM kept them as separate codes. So a note saying "moderate alcohol use disorder" does not tell anyone whether to bill F10.10 or F10.20, and the guess shows up later as a denial.

What this Blog will cover

• What is the alcohol use disorder ICD-10 code
• 5 Common pitfalls that you can avoid while using the code
• When is F10.10 actually the right call
• What a payer is actually reading in your notes
• Where the diagnosis meets the authorization

Here's a small thing that costs the field an enormous amount of time. You finish an assessment, you're confident in your formulation, and you write "moderate alcohol use disorder, five criteria met." Precise, defensible, exactly what your training asked for.

Then someone in billing has to decide whether that's F10.10 or F10.20, and nothing in your sentence tells them. They pick one. You find out which one they picked, if you ever find out, when a claim comes back.

What is the alcohol use disorder ICD-10 code?

It depends on the pattern you documented. F10.10 is alcohol abuse uncomplicated, F10.20 is alcohol dependence uncomplicated, and F10.90 is alcohol use unspecified. Complications like intoxication or withdrawal add further digits.

The pattern digit carries the weight: .1 for abuse, .2 for dependence, .9 when nobody said (American Psychological Association Services).

That last column isn't in any code book. It's what the person reviewing your authorization request is thinking while they read.

Why does DSM-5 severity not tell anyone which code to use?

Because the two systems were built on different logic. DSM-5 defines alcohol use disorder as one condition graded by how many of 11 criteria a person meets: mild is 2 to 3, moderate is 4 to 5, and severe is 6 or more (Alcoholism: Clinical and Experimental Research, 2021).

ICD-10 has two separate patterns, abuse and dependence, with no criteria count attached to either. There's no published rule converting five criteria into a .2.

What changed in DSM-5? It merged the old abuse and dependence categories into a single alcohol use disorder using the same 11 criteria, adding craving and removing legal problems (Alcoholism: Clinical and Experimental Research, 2021).

So the gap is structural, not sloppiness on anyone's part. You're documenting in the newer system and billing in the older one.

The four words that can solve it

Name the ICD-10 pattern alongside your DSM-5 severity. That's the whole fix.

Weak: "Moderate alcohol use disorder, five criteria met."

Better: "Moderate alcohol use disorder, five criteria met; meets criteria for alcohol dependence."

Four extra words, and the ambiguity is gone forever. Nobody downstream guesses, nobody calls you, and the code on the claim is the code you intended.

Some agencies handle this with a blanket policy instead, mapping moderate and severe to dependence and mild to abuse. That works most of the time and it's still a local convention rather than a rule from the guidelines. If your organization uses one, write it down and apply it consistently, because an auditor will ask how you decided.

5 Common pitfalls that you can avoid

The history and assessment disagree. Your history says the client abused alcohol in his twenties. Your assessment says dependence now. Both true, and a coder may bill both. Add "history of" explicitly to the past-tense one.

Coding down to be gentle. F10.10 on a client who meets dependence criteria feels kinder and reads to a reviewer as a facility billing above its own diagnosis. It's also the fastest way to lose the residential authorization you wanted for them.

Coding up to protect the authorization. The reverse is worse, because it moves a payment problem into compliance territory. Code what you assessed.

Leaving F10.90 in place after the evaluation. Unspecified is honest at intake, before you've the picture. It stops being honest the moment you've it.

Forgetting the complication digits exist. A client in active withdrawal coded F10.20 is described as calmer than they were, and the record won't support the care you provided.

When is F10.10 actually the right call?

When the person meets abuse criteria and not dependence, and the care you're providing matches that intensity. Outpatient treatment for someone whose drinking is causing real problems without tolerance, withdrawal, or loss of control is exactly what this code describes.

There's nothing second-rate about it. The instinct to reach for dependence because it sounds more serious does your client no favors, and an accurate lighter code is more defensible than an inflated heavier one.

There's also a moment where the conservative code protects you. If someone is admitted for evaluation and dependence isn't yet established, code what the chart supports today and update it when the picture resolves. Coding forward to what you expect to find is how charts end up contradicting their own claims.

What a payer is actually reading in your notes

They're asking one question: does this diagnosis describe someone who needs what you asked for. F10.20 with a withdrawal complication answers it. F10.90 answers nothing.

Insurers denied 19% of in-network claims on marketplace plans in 2024, with rates ranging from 3% to 36% depending on the insurer (KFF). Behavioral health sits at the rough end, and denial rates have been climbing.

You'll rarely see the denial yourself. It reaches whoever manages denials, who calls you for records, and you spend forty minutes reconstructing a session from four months ago. That's the real cost to you, and it's prevented by a clause you write while the client is still in the room.

When a denial does come back marked medical necessity, the diagnosis specificity is almost always where it started.

Where the diagnosis meets the authorization

One more thing worth knowing, because it surprises clinicians. If your organization obtained an authorization citing one diagnosis and the claim goes out with a different one, some payers treat that as an authorization failure rather than a correctable coding error.

That means a diagnosis you refined in good faith between the authorization call and the claim can create a mismatch nobody notices until the remittance. It is not a reason to avoid refining the diagnosis. It is a reason to tell whoever handles your prior authorizations when you do.

Fewer than 1% of denied claims were appealed in 2024, and insurers upheld 66% of the appeals that were filed (KFF). Almost nothing gets contested, and most of what does gets lost.

How Supa handles the translation

The gap between DSM-5 and ICD-10 is a handoff problem, and handoff problems are where quiet errors live. Nobody makes a mistake. The information just doesn't survive the trip.

Supa works inside the systems you already use. It notices when your note records DSM-5 severity without naming an ICD-10 pattern, when an unspecified code is sitting on a chart headed for a residential authorization, and when the diagnosis on a claim no longer matches the one on the authorization.

You keep making the clinical judgment. What goes away is the version where your judgment was sound and the record didn't carry it.

If your clinicians are fielding records requests about sessions they barely remember, that is the loop worth closing. Book a demo and we will look at a few of your charts together.

FAQ

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

A: It depends on the pattern documented. F10.10 covers alcohol abuse uncomplicated, F10.20 covers alcohol dependence uncomplicated, and F10.90 covers unspecified use. Withdrawal and intoxication add further digits, as in F10.230.

Q: Does moderate alcohol use disorder mean F10.10 or F10.20?

A: The code set doesn't answer that. DSM-5 severity and ICD-10 patterns are separate systems, so "moderate AUD" leaves the pattern undocumented. Add a clause naming abuse or dependence and the ambiguity disappears.

Q: Is alcohol abuse still a valid diagnosis?

A: In ICD-10, yes. F10.1x remains current even though DSM-5 merged abuse and dependence clinically. Claims follow the code set, so the distinction still matters for what you write.

Q: Can I use F10.20 for a client in early recovery?

A: Yes, while the dependence is active and being treated. Moving to F10.21 requires you to document remission explicitly. Time in program and negative screens support that judgment without substituting for it.

Q: What if I refine the diagnosis mid-treatment?

A: Refine it. Then tell whoever manages the authorization, because a mismatch between the authorized diagnosis and the billed one can deny as an authorization problem rather than as a coding correction.

Q: Why would a residential claim deny with an alcohol diagnosis on it?

A: Usually specificity. An unspecified code rarely supports residential intensity, and a mismatch with the authorized diagnosis can deny for a different reason entirely. Both look like unrelated denials on a remittance.