Assessment

How to administer and score clinical assessments

PHQ-9, GAD-7, PCL-5, AUDIT and friends — and what to do with the score.

6 min read·5 steps· Updated June 10, 2026

Administer validated clinical assessments with instant scoring, subscale breakdowns, severity bands, and safety flags. Save results into the active session with an AI clinical summary.

Outcome measures move treatment from 'how are you feeling?' to 'are we actually moving?' They also strengthen any treatment plan or insurance appeal. Here's how to integrate them without making sessions feel like a survey.

Quick answer

Administering a clinical assessment well requires three elements: a clear rationale shared with the client, standardized administration (read items verbatim, no paraphrasing), and immediate scoring with the result fed back in plain language. Re-administer at consistent intervals (every 4 sessions or per protocol) so the data is comparable across time.

Key takeaways

Pick the right measure: Match instrument to presenting concern: PHQ-9 (depression), GAD-7 (anxiety), PCL-5 (PTSD), AUDIT/DAST (substance use), ORS/SRS (general).
Administer consistently: Same instrument, same interval (intake, every 4–6 weeks, termination).
Score immediately: Add up, compare to clinical cutoff, document.
Share the score with the client: Most clients find tracking validating.
Watch for safety items: PHQ-9 item 9 (suicidal ideation) requires a same-session response, every single time.

When to use this

Intake, every 4–6 weeks during treatment, and at termination.
Anytime you suspect a clinically meaningful change you can't yet articulate.
Before requesting additional sessions from a payer.

Steps

1
Pick the right measure
Match instrument to presenting concern: PHQ-9 (depression), GAD-7 (anxiety), PCL-5 (PTSD), AUDIT/DAST (substance use), ORS/SRS (general).
2
Administer consistently
Same instrument, same interval (intake, every 4–6 weeks, termination).
3
Score immediately
Add up, compare to clinical cutoff, document. Use any digital tool that auto-scores.
4
Share the score with the client
Most clients find tracking validating. Use it as a conversation starter, not a verdict.
5
Watch for safety items
PHQ-9 item 9 (suicidal ideation) requires a same-session response, every single time.

Example

Standard cutoff cheat-sheet

PHQ-9: 0–4 none, 5–9 mild, 10–14 moderate, 15–19 moderately severe, 20–27 severe. Item 9 > 0 → same-session SI assessment.
GAD-7: 0–4 minimal, 5–9 mild, 10–14 moderate, 15–21 severe.
PCL-5: ≥ 33 suggests probable PTSD; track total + cluster scores.
AUDIT: 8+ suggests hazardous drinking; 16+ suggests harmful; 20+ suggests dependence.
ORS: < 25 clinical cutoff (adults); reliable change ≥ 5 points.

Quick checklist

Measure matches presenting concern.
Same instrument re-administered at consistent intervals.
Score documented in the chart with date.
Safety items reviewed live and addressed.
Trend (not just absolute score) discussed with client.

Common variations

Measurement-based care (MBC)

Score every session; route results to a feedback dashboard the clinician reviews pre-session.

Routine outcome monitoring (ROM)

Brief ORS/SRS each session; longer measures monthly.

Evidence base

Lambert et al. and the OQ/PCOMS trials show measurement-based care reduces deterioration rates and shortens time to recovery, particularly for clients off-track early in treatment.

Deep dive

Why measurement-based care doubles outcome rates

Multiple RCTs and large naturalistic studies (Lambert, Harmon, Slade; Delgadillo; Fortney) have shown that routinely administering outcome measures — and acting on the results — roughly doubles the rate of clinically significant change versus treatment-as-usual. The mechanism is simple: deterioration that would otherwise go unnoticed becomes visible in time to course-correct. The clinical pivot, not the measure itself, is what produces the gain. If you score the PHQ-9 but never discuss the trend with the client, you have done the work without the benefit.

Choosing the right measure for the question

Pick measures based on what you need to know, not what is famous. Broad symptom screen: PHQ-9 (depression), GAD-7 (anxiety), DASS-21 (transdiagnostic). Functional change: WHODAS 2.0, Outcome Rating Scale. Working alliance and risk-of-dropout: Session Rating Scale at session 3 and session 6. Diagnostic clarification: PCL-5 (PTSD), MDQ (bipolar screen), AUDIT (alcohol), ASRS (adult ADHD screen). Don't stack — three measures per administration is the ceiling before client fatigue erodes data quality.

Feeding results back to clients in 60 seconds

A strong feedback moment has four beats: (1) name the measure and what it tracks, (2) state the score and the change since last administration, (3) interpret the clinical meaning in one sentence, (4) ask the client what they make of it. Example: 'Your PHQ-9 today is 9, down from 14 four weeks ago. Clinically, that's a real improvement — you've moved from moderate to mild depression. What's your read?' That 60-second exchange increases client engagement with the data and gives you a collaborative pivot point.

Tips

Trend matters more than absolute scores — a PHQ-9 dropping 18→11 is significant even though both are 'moderate'.
Pair a symptom measure with a functioning measure so improvement isn't only about mood.