PHQ-SADS: Somatic, anxiety, and depression scales

What this assessment is

PHQ-SADS places three established symptom scales together:

• the PHQ-15 for somatic symptoms
• the GAD-7 for anxiety symptoms
• the PHQ-9 for depressive symptoms

The name stands for Patient Health Questionnaire Somatic, Anxiety, and Depressive Symptom Scales, which describes the three symptom areas covered by the assessment.

The three scores stay separate. There is no standard PHQ-SADS grand total in the 2010 review that defined this grouping.

Three scales. Three results.

Who it was designed for

The PHQ scales were developed for adults in primary care and other medical settings, where physical symptoms, anxiety, and depression may overlap.

It supports screening and symptom measurement. It does not replace a medical review or mental health assessment.

Evidence from one age group, language, or setting should not be assumed to apply everywhere. Modified forms need their own validation.

How it is administered

The three components do not use one shared timeframe.

The PHQ-15 asks about physical symptoms during the past four weeks, while the GAD-7 and PHQ-9 ask about the past two weeks.

The response scales also differ. PHQ-15 responses contribute 0 to 2 points. GAD-7 and PHQ-9 responses contribute 0 to 3 points.

Do not blend them.

Some PHQ-SADS forms add a panic module and an overall impact question, but those additions are reviewed separately and are not added to the three scale totals.

What it measures

PHQ-SADS gives three views of recent symptoms:

• PHQ-15 records physical symptom burden.
• GAD-7 records generalized anxiety symptoms.
• PHQ-9 records depressive symptoms.

Overlap is common because sleep, energy, concentration, and physical discomfort can relate to more than one condition.

The pattern can guide follow-up questions. It cannot establish the cause.

How scoring works

Score each component with its own rules:

• PHQ-15: add 15 responses for a total from 0 to 30.
• GAD-7: add seven responses for a total from 0 to 21.
• PHQ-9: add nine responses for a total from 0 to 27.

Do not add these three totals together.

Keep them separate.

The systematic review uses 5, 10, and 15 as mild, moderate, and severe symptom thresholds across the component scales, while standard PHQ-9 reporting adds a 20-point boundary for its severe range.

Each component needs complete or source-supported missing-data handling. A missing response must not be silently replaced with zero.

How to interpret PHQ-SADS results

Read the three numbers as a profile.

A high PHQ-15 result can support a broader review of physical symptoms, but it cannot decide whether those symptoms have a medical, psychological, or mixed cause.

A high GAD-7 or PHQ-9 result shows greater symptom burden. It does not confirm an anxiety or depressive disorder.

The 2010 review found that evidence for sensitivity to change was strongest for the PHQ-9. Evidence for the GAD-7 and PHQ-15 was promising but less established at that time.

Do not confuse PHQ-SADS with the newer PHQ-SAD composite. A 2025 PHQ-SAD study tested a different 25-item form with one combined score.

A score does not replace a safety check

The PHQ-9 component includes a question about death or self-harm. Any endorsement needs direct review, even when the PHQ-9 total is low.

Safety comes first.

Safety overrides the profile.

Ask about current thoughts, intent, plan, means, protective factors, and available support. Do not use the PHQ-SADS profile as a substitute for a direct safety assessment.

What the scores cannot tell you

PHQ-SADS cannot diagnose a mental or physical condition. It cannot show which condition caused an overlapping symptom.

The scores cannot choose treatment or determine urgency by themselves because a proper assessment also considers function, history, medication, physical health, recent events, and the person's priorities.

Low results do not rule out a condition. High results do not prove one.

Evidence and limitations

The 2010 systematic review combined evidence from 9,740 patients in four multisite studies with other published research. It found useful measurement properties for all three components.

The components have different purposes and validation histories. Their performance changes with population, language, setting, and cutoff choice.

Report the version, language, timeframes, three separate totals, and any missing-data rule. That information is necessary for a result to be interpreted correctly.