Track complete PHQ-9 results to describe how nine depression symptoms change across comparable two-week windows. A trend cannot prove treatment response, remission, relapse, or treatment failure by itself. Interpret it with item-level safety, functioning, context, administration conditions, and the patient's account.
Start with comparable results
Survey Doctor's native PHQ-9 requires all nine scored items and produces a total from 0 to 27. Each administration refers to the past two weeks. The original adult validation study established that scoring structure and the five descriptive bands used here.
Before interpreting a series, confirm the form and scoring method for each result. A manual total cannot show completeness, item 9, or the optional functional companion. Some imported summaries lack that evidence too.
Survey Doctor reports display saved numeric scores, including imported totals, without rescoring them. Charts and arithmetic changes can therefore include values that are not clinically comparable to native PHQ-9 results. Review the source record before applying native bands or treating a difference as symptom change. Numeric inclusion does not establish treatment response or equivalent scoring.
Keep these details stable when possible:
- The exact nine-item adult English form
- Complete item responses
- The two-week recall period
- Similar administration support and setting
- Clear dates and relevant contextual notes
If any of those differ, document the break in the clinical record. A continuous report line does not resolve a change in form, scoring, or administration.
Read the total and items together
The PHQ-9 bands are descriptive:
| Score | Severity |
|---|---|
| 0-4 | None or minimal |
| 5-9 | Mild |
| 10-14 | Moderate |
| 15-19 | Moderately severe |
| 20-27 | Severe |
A band crossing can help organize review, but it is not an outcome definition. A move from 15 to 14 is one point even though the label changes. A larger total change can also hide one symptom worsening while several others improve.
The plain-language PHQ-9 score guide explains these bands for patients. Use that page for score discussion, then return to the item pattern and clinical context.
Review which items moved and ask whether that pattern matches the patient's experience. Sleep, appetite, energy, concentration, and movement can change for reasons outside depression. The questionnaire cannot identify the cause.
The PHQ-9 is a screening and symptom-severity tool, not a diagnostic test. Repeated high scores do not establish a diagnosis, and repeated low scores do not rule one out or mean that care should stop.
Treat item 9 as a separate safety signal
Item 9 asks about thoughts of death or self-harm. Any answer above zero needs direct follow-up regardless of the total or trend. A falling total does not cancel a new item 9 endorsement.
The PHQ-9 is not a complete suicide risk assessment. Do not infer safety from the absence of other depression symptoms.
If you need help now. In the US, Call 988 or text 988. Call 911 if you are in immediate danger. Outside the US, contact your local emergency number or find support in your country.
Track functioning separately from the PHQ-9 total
Published PHQ-9 materials can include an unscored question about difficulty at work, at home, or with other people. Survey Doctor shows this optional companion when any scored symptom is above zero and stores the response separately.
The response does not change the 0 to 27 total or its trend. Historical, manual, and imported rows without companion data mean it was not collected. Ask about functioning at each clinically relevant review. A flat score with better function and a flat score with worsening function are not the same clinical picture.
Put change thresholds in context
No single point change or percentage applies across every population, baseline score, and purpose. Published estimates use different methods and samples. A rule that fits one quality measure or trial endpoint may not fit an individual treatment decision.
In a 2004 study of 434 older adults in the IMPACT trial, a five-point PHQ-9 difference was the estimated minimum clinically important change. That finding supports a review signal in a similar context, but it does not make five points a universal response rule.
Do not automatically label a fixed drop as response, a score below 5 as remission, or a small rise as relapse. Those terms require an explicit protocol and supporting evidence for the intended population. They also require clinical confirmation.
Retest timing is not universal either. Choose an interval based on the purpose of measurement, expected clinical course, treatment context, burden, and the decision the next result could inform. Repeating too soon may measure ordinary variation. Waiting too long may miss a change that matters.
Use a disciplined review sequence
For each new native result:
- Check completion and item 9.
- Compare the total with the last comparable result.
- Review item-level movement.
- Ask what changed during the two-week window.
- Review the optional companion when available, and assess functioning directly.
- Compare the score direction with the patient's account.
- Record any form or workflow change that limits comparison.
This sequence keeps the total useful without turning it into an automatic treatment rule. Treatment choices still depend on diagnosis, preferences, medical history, prior response, adverse effects, access, and current guidance. The practice-level outcome measurement guide applies the same comparability rule to aggregate reporting.
Handle discordance as information
A lower score can occur while a patient feels worse. The opposite can also happen. Neither observation should be dismissed. Look closer.
Check whether the questionnaire missed the main concern, whether symptoms fell outside the two-week window, or whether a few item changes drove the total. Ask about events, sleep, physical health, medication effects, and substance use when relevant.
The goal is not to make the interview agree with the score. It is to understand why they differ.
The clinical boundary
A PHQ-9 trend is structured symptom information. It becomes clinically useful only when the same form is compared honestly and the missing context is restored at review. Keep treatment decisions with the full assessment, not with a line on a chart.