Writing psychiatric treatment plans with measurable goals: why "patient will feel better" fails an audit and what to write instead

By Selam Koch, PMHNP-BC, founder of OneStep Scribe and practicing psychiatric nurse practitioner. Published August 2026.

Every psychiatric chart claims to have a treatment plan, and most of them contain goals like "patient will feel better" or "patient will improve coping skills." Those sentences are not goals. They are wishes, and in a chart review they behave like empty space: a reviewer cannot tell what the treatment is trying to change, how anyone would know it changed, or when the question will be asked again.

The stakes are practical. Payers reimburse treatment that is medically necessary for a documented condition and reasonably expected to improve it. A goal with no measure makes improvement invisible by construction: after six months of visits, the chart cannot show whether the treatment worked, because it never said what working would look like. That gap is exactly what utilization reviewers, agency auditors, and payer chart reviews probe, and expectations vary enough by payer and program that the safe assumption is that someone will eventually read the plan section closely.

This guide covers the anatomy of a measurable goal, worked rewrites, the diagnosis-goal-intervention chain, plan updates, review cadence, and what reviewers look for.

Why "patient will feel better" fails

Take the sentence apart and three failures show up:

• No measure. "Better" has no unit. No score, count, or observable behavior could confirm or refute it, so every later note is reduced to impression.
• No timeframe. Without a review date, the goal can sit in the chart unmet forever without ever being wrong. A reviewer sees a plan that was never designed to be tested.
• No link to the diagnosis. "Feel better" could sit under any diagnosis in any chart, so it does nothing to show that this treatment addresses this patient's documented condition, which is the core of medical necessity.

There is a clinical cost too. A vague goal gives the patient nothing to aim at and gives you nothing to titrate against. Deciding in advance what change you expect, and when, is what makes non-response visible early enough to act on. That is the logic of measurement-based care: a target you never defined can never be missed, or hit.

The anatomy of a measurable, time-bound goal

A goal that holds up in review has five parts. Frameworks like SMART capture roughly the same idea; the version below is tuned for psychiatric documentation:

1. The problem, in this patient's terms: the symptom or functional impairment the goal targets, specific enough that it could only belong to this chart.
2. A baseline: where the patient starts, as a scale score, a frequency count, or a described current behavior. Without it, progress has nothing to be measured against.
3. A target: the number or observable behavior that will count as progress.
4. A timeframe: when the goal will be evaluated, usually a number of weeks or a named review date.
5. A method of measurement: who measures, and with what: a repeated validated scale, a patient log, collateral report, or observation in session.

Two kinds of targets do most of the work. Scale scores from validated instruments give you a number that trends across visits; picking the instrument that actually maps to the treatment target is covered in choosing psychiatric rating scales. Observable behaviors cover everything a scale does not: attendance at work, panic attack frequency, nights of sleep, drinks per week, driving on the highway again. The strongest plans pair one of each, a number and a life change, because symptom relief that never becomes functional recovery is an incomplete outcome.

Worked rewrites: vague to measurable

The examples below are fictional. Each rewrites a goal as it commonly appears in charts.

The same translation works for almost any presentation:

One row is a deliberate exception: safety goals are process goals, measured by documented safety-plan review, updates, and reported use rather than by symptom counts, which is why that row alone carries no baseline or numeric target.

Tie each goal to the diagnosis, and each intervention to a goal

Measurability is only half of what a reviewer checks; the other half is connection. A defensible plan reads as one documented line from diagnosis to goals, goals to interventions, and interventions to the progress recorded at each visit. Break that line anywhere and questions follow:

• A goal with no diagnosis behind it raises the question of why this problem is being treated in this episode of care. If the goal targets insomnia, the chart should carry the diagnosis or documented symptom that makes insomnia part of this treatment.
• A diagnosis with no goal suggests a condition billed but not actively treated. If a diagnosis is on the claim, the plan should show what is being done about it, even if that is monitoring.
• An intervention with no goal is often the finding reviewers pursue hardest: a service, most commonly psychotherapy, delivered visit after visit with nothing in the plan saying what it is for. Every intervention line should name its goal: "weekly CBT targeting Goal 2 (panic frequency and avoidance)."

A quick self-audit: read your plan and draw the arrows. Diagnosis to goal, goal to intervention, intervention to the progress note that reports on it. If any arrow has nothing to point to, fix the document, or notice that the care itself has drifted.

Update the plan as progress and setbacks are documented

A treatment plan written once and never touched is almost as weak as a vague one. Each progress note is evidence about the plan, and three moves keep it honest:

• Report against the goal, by name. "PHQ-9 today 11, down from 18 at baseline; Goal 1 on track at week 8 of 12" is a progress note doing its job. The repeated instrument behind that sentence is the machinery of measurement-based care; the goal is where the numbers point.
• Close or advance goals that are met. Say so, date it, and either move to a maintenance goal or set the next target. A met goal left open reads as an unread plan.
• Revise, do not ignore, goals that are missed. If the review date arrives and the target was not reached, document the reassessment: inadequate intervention, subtherapeutic dose, incomplete adherence, or a diagnosis worth revisiting. Then change something, and update the goal's target or timeframe to match. A setback with a revised plan is evidence of active management; the same setback with the original goal silently rolled forward is a chart on autopilot.

If the treatment target itself shifts, say from depression to newly prominent panic, reset the measurement to match: new instrument, new baseline, new goal, stated explicitly. Switching instruments without losing the trend is covered in the guide to choosing psychiatric rating scales, and the psychiatric progress note example shows where goal status sits in a full follow-up note.

How often should the plan be revisited?

There is no single required interval, and this is a place to be careful rather than confident. Community mental health programs, agencies, and some state Medicaid programs specify review intervals in regulation or contract, commonly somewhere between every 90 days and every 6 months, or whenever there is a significant clinical change. Commercial payers are usually less prescriptive but still expect the plan to be demonstrably current, and accreditation bodies add expectations in some settings. Verify the specific requirement with your payer, agency, or program, in writing where you can.

Clinically, a workable outpatient default is to touch the plan whenever something changes (a medication change, a met or missed goal, a new problem, a safety event) and to do a stated, dated review on a regular interval even when nothing has changed. The goal timeframes you wrote do this work for you: if every goal has a review date, the plan schedules its own maintenance.

What reviewers look for in the plan section

Across payer audit guidance and my own chart-review experience, the questions are consistent:

• Are the goals measurable and time-bound? A score, count, or observable behavior, with a baseline and a review date. "Feel better" fails on sight.
• Does each goal trace to a documented diagnosis, and does each active diagnosis have something in the plan addressing it?
• Is every billed service attached to a goal? Especially psychotherapy delivered alongside medication management: the plan should say what the therapy targets.
• Is there evidence the plan is alive? Progress notes referencing goal status, goals closed when met, revisions when missed, dated reviews at a defensible interval.
• Did the patient participate? Goals in the patient's own terms, with documented agreement or disagreement, read as care planning rather than paperwork.
• Is it this patient's plan? Individualized details are the strongest defense against a cloned-documentation finding; identical goal text across charts is one of the easiest patterns to spot.

Two or three well-built goals beat eight template goals in every review I have seen, and they take less time to maintain.

Frequently asked questions

Do measurable goals require rating scales?

No. Scales are the most convenient measure where a validated instrument maps to the target, but frequency counts, duration measures, and observable behaviors are fully measurable and often more meaningful to the patient. When a scale does fit, pick it deliberately: see the guide to choosing psychiatric rating scales.

What if the patient's own goal is vague, like "I just want to feel like myself again"?

Keep the patient's words, then translate them together. Ask what "feeling like myself" would look like on an ordinary Tuesday: the answers, cooking again, calling her sister, finishing a shift without leaving the floor, are the observable markers. Document the patient's framing as the anchor and the operationalized version as the goal.

How many goals should a plan have?

Enough to cover what is actively being treated, and no more. For most outpatient psychiatric episodes that is two to four. Every goal is a commitment to report on it in progress notes; goals nobody reports on become audit findings in waiting. If a problem is real but deferred, a monitoring statement is more honest than a goal you will not pursue this quarter.

Can an AI scribe help with treatment plan goals?

It can capture the raw material, which is where most of it gets lost. The baseline, the patient's stated goal, the agreed target, and the timeframe are usually spoken aloud in the visit and forgotten by the time the note is written. OneStep Scribe listens to the visit and drafts the plan with those specifics preserved, so goal status and score trends are easy to carry into follow-up notes. The clinician reviews, edits, and signs every note; the goals, and the judgment behind them, remain yours.

This article is educational and reflects one clinician's documentation approach. It is not clinical, legal, or billing advice for any specific patient or claim, and it does not guarantee any outcome, reimbursement, or audit result. The patients, scores, and goals described are fictional. Treatment plan requirements, review intervals, and documentation expectations vary by payer, state, program, and accreditation body and change over time; verify the specific requirements with your payer, agency, or MAC and a certified coder or compliance professional. Clinical judgment, including what goals are appropriate for any patient, stays with the treating clinician.