TL;DR: Adjustment disorder ICD 10 coding starts at F43.20 for the unspecified form, with subtypes running through F43.29 based on the dominant symptom picture. The code is inherently time-limited, which creates a documentation obligation most diagnoses do not: you have to show what happened when the stressor resolved.
Key takeaways
- F43.20 is "Adjustment disorder, unspecified" and it is billable, but a subtype is almost always more defensible.
- The subtype follows the dominant symptom, not the stressor.
- The diagnosis is time-limited by definition, so a chart carrying it for a year invites questions.
- An identifiable stressor has to be named in the record, not implied.
- ICD-10 references this category as "transient adjustment reaction" in exclusion notes elsewhere, which tells you how the classification thinks about it.
Adjustment disorder is the diagnosis clinicians reach for when something real happened, and the response is out of proportion, but nothing else quite fits. A redundancy, a divorce, a diagnosis in the family. The client is struggling in a way that is understandable and still clinically significant.
It is also the diagnosis most likely to sit on a chart long after it stopped being true.
What are the ICD-10 codes for adjustment disorder?
The category is F43.2, and the unspecified code F43.20 carries the official descriptor "Adjustment disorder, unspecified" (AAPC). It is a billable code. Subtypes extend through the F43.2 range, with the commonly used ones covering depressed mood, anxiety, mixed anxiety and depressed mood, and disturbance of conduct.
The structural logic is worth internalising. You are not coding the stressor. You are coding the predominant symptom response to it. Two clients made redundant on the same day can carry different subtypes, because one presents with low mood and withdrawal while the other presents with worry and physical tension.
That is a different mental model from most of the F-chapter, where the code follows the syndrome. Here the syndrome is defined by its relationship to an event, and the subtype is defined by what the response looks like.
One telling detail from elsewhere in the classification. The exclusion notes attached to panic disorder refer to this category as "transient adjustment reaction (F43.2)" (AAPC). Transient is doing real work in that phrase. The classification treats adjustment disorder as something that resolves.
How do you choose between the subtypes?
Choose by asking which symptom cluster is doing the most damage, not which one the client mentions first. Distress after a stressor is usually mixed, so the subtype is a judgment about dominance rather than exclusivity.
A practical sequence that holds up well:
- Confirm an identifiable stressor within the recent past, and name it in the record.
- Confirm the response exceeds what the context would ordinarily produce, or that it impairs functioning.
- Rule out a better-fitting diagnosis, particularly major depressive disorder and generalized anxiety disorder.
- Identify which symptom cluster predominates, and code that subtype.
- Default to F43.20 only when the picture is genuinely mixed in a way the subtypes do not capture.
Step three is the one that gets skipped, and it is the one a reviewer will test. Adjustment disorder is sometimes used as a softer alternative to a depressive diagnosis, either to avoid stigma or because the clinician is not yet certain. The instinct is understandable, and the documentation consequence is not: a chart describing a full depressive syndrome coded as adjustment disorder reads as either under-diagnosis or code selection driven by something other than the clinical picture.
Why does the time-limited nature matter for your notes?
Because adjustment disorder makes a claim about trajectory, and trajectory is the one thing a single note cannot show. The diagnosis says this response is tied to a stressor and expected to resolve as the client adapts. Your record has to eventually address whether it did.
Three things follow.
Name the stressor explicitly and date it. "Client experiencing situational stress" is not a stressor. "Client's mother received a terminal diagnosis in March" is.
Track the trajectory. Notes should show movement, or explain its absence. A client six months in with no improvement is telling you something diagnostically, and the record should reflect that you noticed.
Resolve or revise. When the stressor passes and the symptoms lift, the record should show the diagnosis being reconsidered. When symptoms persist beyond what adjustment disorder describes, that is a prompt to reassess rather than to keep coding the same thing.
This is the discipline that separates a defensible adjustment disorder chart from a lazy one, and it is mostly about revisiting rather than writing more. For the mechanics of carrying a diagnosis through the record, our guide to progress note templates covers the structure.
What are the common coding mistakes?
Four recur, and all of them are documentation failures rather than knowledge failures.
No named stressor. The diagnosis requires one. A chart that never identifies it cannot support the code.
Defaulting to unspecified. F43.20 is billable and legitimate, but a subtype communicates more and is usually available. Habitual use of the unspecified code across a caseload is a pattern worth checking.
Using it to avoid a heavier diagnosis. If the client meets criteria for a depressive disorder, code the depressive disorder. Diagnostic softening shows up clearly in a chart audit and it does the client no favours when their record needs to justify a level of care.
Leaving it in place indefinitely. The most common of the four. The code is transient by design, and an indefinite adjustment disorder is a contradiction the chart is quietly carrying.
When would you use an adjustment disorder code, and when would you not?
Adjustment disorder fits a clinically significant response to an identifiable recent stressor that does not meet criteria for another disorder. Those last nine words carry the weight. If the presentation satisfies the criteria for a depressive or anxiety disorder, that diagnosis takes precedence, whatever the trigger was.
Use an adjustment disorder code here
A client's marriage ended three months ago. She is tearful most days, sleeping poorly, and has withdrawn from friends. She is still working, still eating normally, has no anhedonia, and denies hopelessness or any thoughts of self-harm. She does not meet criteria for a major depressive episode.
This is the case the diagnosis exists for. Named stressor, recent onset, real distress, functional impact, and no better-fitting diagnosis. Code the depressed mood subtype and record the stressor with its date.
A client relocated for work eight weeks ago. He reports racing thoughts, difficulty settling, and physical tension, all beginning after the move. There is no history of anxiety before it.
The anxiety subtype fits. The response is tied to the change, the timeline is short, and there is no pre-existing anxiety pattern that would point elsewhere.
Do not use an adjustment disorder code here
A client lost his father four months ago. He reports mood 2 out of 10, five hours of sleep with early waking, a 6kg weight loss, anhedonia, worthlessness, and passive thoughts of being better off dead.
Code a depressive disorder. The stressor is real, and the response meets full criteria for a major depressive episode, so the depressive code takes precedence. Using adjustment disorder here softens the record in a way that will not support the level of care this client needs.
A client presents with anxiety and low mood but cannot identify anything that changed. Nothing at work, at home, or in her health.
Adjustment disorder requires an identifiable stressor. Without one, the diagnosis has no foundation, and a chart that never names the stressor cannot support the code under any review.
A client has carried an adjustment disorder diagnosis for fourteen months. The original stressor, a job loss, resolved a year ago when she started new work.
Reassess rather than continue. The classification treats this category as transient, and a diagnosis that has outlived its stressor by a year is describing something else. Either a different disorder is present, or treatment should be concluding.
A client is grieving a death six weeks ago in a way that is painful but proportionate, with no functional impairment.
Consider whether any diagnosis is warranted. Distress after a loss is not automatically a disorder, and coding one to justify a session is a habit that erodes the meaning of the record.
How the review cycle can be automated
Adjustment disorder is unusual among diagnoses in that the documentation risk is not what you wrote. It is what you did not revisit. That makes it a good fit for automation, because remembering to look again is precisely the kind of task humans are worst at and software is best at.
Here is the mechanism, in plain terms.
The stressor gets anchored at intake. When the initiating event and its date are captured as structured detail rather than buried in prose, they stay retrievable. Every later note can be read against that anchor.
The trajectory becomes visible. If each session's draft carries the working diagnosis and the functional detail alongside it, the arc across eight sessions is legible without anyone manually reading back through the chart.
The review point surfaces. A diagnosis that is expected to resolve should be reconsidered on a schedule. That is a date-driven prompt, and dates are something a system can hold reliably while a clinician with a full caseload cannot.
Supanote drafts each session note with the working diagnosis and the clinical detail in place, so the chart accumulates a readable trajectory rather than a series of disconnected entries. You review, edit, and sign every draft.
The honest limit is the same one that applies throughout. Software should not tell you the client has adapted, or that the subtype should change from anxious to depressed mood. Those are clinical readings that depend on being in the room. What it can do is make sure the evidence for whichever reading you reach is actually in the record, and make sure the question gets asked again at a sensible interval instead of never.
There is also a case where this is the wrong purchase. If you carry a small caseload and review every chart monthly anyway, the marginal gain is thin and you should keep your money.
Want the diagnostic thread to hold across the whole episode? Try for Free.
FAQs
Q: What is the ICD-10 code for adjustment disorder?
A: The category is F43.2, with F43.20 as the unspecified code, officially described as "Adjustment disorder, unspecified." Subtypes within the F43.2 range cover the dominant symptom picture, and choosing one is generally more defensible than defaulting to unspecified.
Q: How long can someone carry an adjustment disorder diagnosis?
A: The diagnosis is time-limited by definition, so an indefinite one is a contradiction. If symptoms persist well beyond the stressor and its aftermath, that is a signal to reassess rather than to continue coding the same thing.
Q: Do I have to name the stressor in my notes?
A: Yes. The diagnosis depends on an identifiable stressor, so a record that never names one cannot support the code. Date it as well, since the timeline is what distinguishes this from a primary mood or anxiety disorder.
Q: What if the client has both a stressor and meets depression criteria?
A: Code the depressive disorder. Adjustment disorder is the diagnosis when the response does not meet criteria for another disorder. Using it as a gentler label for a client who meets depression criteria creates a record that will not support the care you provide.
Q: Can adjustment disorder become another diagnosis over time?
A: Frequently, and the record should show that transition rather than a silent code change. Document what you observed that prompted the reassessment, which turns a bare code edit into a defensible clinical decision.
Q: Is F43.20 billable?
A: Yes, F43.20 is a billable ICD-10-CM code. A subtype usually communicates more clinical information, so reserve the unspecified code for presentations that are genuinely mixed rather than using it as a default.
