Written and maintained by CASRAI Editorial Board
Last updated
The PHQ-15 is a 15-item self-report questionnaire that measures somatic symptom severity on a 0 to 30 scale. Researchers use it to quantify bodily symptom burden in clinic and population samples and to compare groups or change over time. This page covers how it is scored, what the 5, 10 and 15 cutpoints mean, a worked scoring example, the validity evidence behind it, and how to compute and report it. It is a measurement reference for study design and analysis, not a tool for diagnosing yourself or anyone else.
What is the PHQ-15?
The PHQ-15 is the somatic symptom severity scale derived from the Patient Health Questionnaire (PHQ). Kroenke, Spitzer and Williams introduced it in 2002 (Psychosomatic Medicine 64(2):258-266), describing it as 15 somatic symptoms from the PHQ, each scored from 0 (“not bothered at all”) to 2 (“bothered a lot”). In the PHQ/GAD-7 Instruction Manual, the instrument is described as being used to derive somatic symptom severity scores and to flag potential somatization or somatoform presentations, not to make provisional diagnoses.
It belongs to the same family as the PHQ-9 depression scale; if you need the depression instrument, see the CASRAI guide to the PHQ-9.
PHQ-15 items and scoring (0-30)
The instrument has fifteen items, each scored 0 to 2, so the total ranges from 0 to 30. According to the instruction manual, the items are of two kinds:
- 13 somatic symptom items (items 1a to 1m on the full PHQ): scored 0 = not bothered at all, 1 = bothered a little, 2 = bothered a lot.
- 2 items taken from the depression module (sleep and feeling tired): scored 0 = not at all, 1 = several days, 2 = more than half the days or nearly every day.
The scale can be derived from page 1 of the full PHQ or administered as a stand-alone PHQ-15 or within the PHQ-SADS. The total score is the sum of the fifteen item scores. Because the sleep and tiredness items use a frequency anchor while the other thirteen use a bother anchor, a data dictionary should record both response sets, even though both map to 0, 1, 2.
PHQ-15 severity cutoffs: 5, 10 and 15
Kroenke et al. (2002) reported that scores of 5, 10 and 15 represented cutpoints for low, medium and high somatic symptom severity. The 2010 systematic review by Kroenke and colleagues (four multisite cross-sectional studies, 9,740 patients) restated the cutpoints as mild, moderate and severe, the same labels used for the PHQ-9 and GAD-7.
| Total score | 2002 label (low/medium/high) | 2010 review label | Instruction-manual flag |
|---|---|---|---|
| Below 5 | Below the low cutpoint | Below the mild cutpoint | None |
| 5 to 9 | Low | Mild | None |
| 10 to 14 | Medium | Moderate | 10 or higher: “yellow flag” (possible clinically significant condition) |
| 15 to 30 | High | Severe | 15 or higher: “red flag” (active treatment probably warranted) |
The label sets differ between publications, so state which one you are using. The yellow-flag and red-flag thresholds are screening heuristics from the instruction manual, applied across the PHQ-9, GAD-7 and PHQ-15; in a study they define analytic categories, not diagnoses.
Worked example: scoring one respondent
This is an illustrative calculation with invented responses, not data from any real participant. Suppose a respondent answers the 13 somatic items as follows (0 = not bothered at all, 1 = a little, 2 = a lot): 1, 0, 2, 1, 0, 1, 2, 0, 1, 1, 0, 2, 1.
- Sum the somatic items: 1 + 0 + 2 + 1 + 0 + 1 + 2 + 0 + 1 + 1 + 0 + 2 + 1 = 12.
- Score the sleep item. The respondent reports problems “more than half the days”: 2.
- Score the tiredness item. The respondent reports “several days”: 1.
- Total = 12 + 2 + 1 = 15.
A total of 15 sits exactly on the highest cutpoint, so it falls in the “high” (2002) or “severe” (2010) band and meets the manual’s 15-or-higher red-flag threshold. Boundary scores are where band labels change, so pre-specify in your analysis plan whether categories are lower-inclusive (10 to 14, 15 and above), as the manual’s wording implies.
Interpretation and validity evidence
The validity case for the PHQ-15 rests mainly on external criteria. In the 2002 study, administered to 6,000 patients in eight general internal medicine and family practice clinics and seven obstetrics-gynecology clinics, higher PHQ-15 scores were associated with a stepwise decline in functional status (SF-20 subscales), more sick days, more clinic visits and more symptom-related difficulty, with similar results in both samples. Somatic and depressive symptom severity had differential effects on these outcomes, which supports treating somatic burden as a construct related to, but not interchangeable with, depression. That is a convergent-versus-discriminant argument; see convergent and discriminant validity for how such evidence is organised.
The 2010 review concluded that the PHQ-15 is equal or superior to other brief measures for assessing somatic symptoms and screening for somatoform disorders. On sensitivity to change, it described the evidence as well established for the PHQ-9 but only emerging for the PHQ-15 and GAD-7, and the instruction manual likewise cites preliminary evidence of responsiveness in treated somatoform or high-somatization samples. If you plan to use change scores as an outcome, treat responsiveness as something to justify in your protocol rather than assume.
| Question | What the cited sources support |
|---|---|
| Does it track function and service use? | Yes: stepwise decrements in SF-20 function, more sick days and clinic visits (Kroenke et al., 2002) |
| Is it distinct from depression severity? | Somatic and depressive severity had differential effects on outcomes (Kroenke et al., 2002) |
| Is it responsive to change? | Emerging, not yet definitive (Kroenke et al., 2010); preliminary evidence per the instruction manual |
| Do the same cutpoints apply across instruments? | 5, 10, 15 mark mild, moderate and severe on the PHQ-9, GAD-7 and PHQ-15 (Kroenke et al., 2010) |
The instruction manual also notes that few translations have been validated psychometrically against independent structured interviews, so a translated version warrants its own reliability and validity evidence (see split-half reliability and test-retest reliability).
Assumptions and common mistakes
- Summing across a mixed anchor. The sleep and tiredness items use a frequency anchor; do not recode them to the bother anchor.
- Mixing label sets. Reporting “moderate” while citing the 2002 “medium” label, or the reverse, makes bands hard to compare across studies.
- Treating the total as interval-scaled without checking. A 0 to 30 sum of three-point items is bounded and often right-skewed in general populations; inspect the distribution and consider floor effects (see floor and ceiling effects) before using a mean-based test.
- Reading a score as a diagnosis. The manual describes severity scores, not provisional diagnoses.
- Undeclared missing-data handling. Decide in advance how incomplete forms are treated and report it.
How to calculate the PHQ-15 in Excel, R or Python
Each snippet assumes one row per respondent and fifteen columns already coded 0, 1 or 2.
Excel (items in columns B to P, first respondent in row 2):
=SUM(B2:P2)
=IF(Q2>=15,"severe",IF(Q2>=10,"moderate",IF(Q2>=5,"mild","below mild")))R (data frame d with columns q1 to q15):
items <- paste0("q", 1:15)
d$phq15 <- rowSums(d[, items])
d$band <- cut(d$phq15, breaks = c(-Inf, 4, 9, 14, Inf),
labels = c("below mild", "mild", "moderate", "severe"))Python (pandas data frame df):
items = [f"q{i}" for i in range(1, 16)]
df["phq15"] = df[items].sum(axis=1)
df["band"] = pd.cut(df["phq15"], bins=[-1, 4, 9, 14, 30],
labels=["below mild", "mild", "moderate", "severe"])Applied to the worked example above, each snippet returns a total of 15 and the “severe” band.
How to report the PHQ-15 (APA style)
Name the instrument, the score range, the band labels you used and their source, and give the descriptive statistics. A template with placeholder values:
Somatic symptom severity was measured with the PHQ-15 (Kroenke et al., 2002), a 15-item scale scored 0 to 30, with cutpoints of 5, 10 and 15 indicating mild, moderate and severe symptoms (Kroenke et al., 2010). In this sample, scores averaged M = [mean] (SD = [SD]); [n] participants ([%]) scored 10 or higher.
Replace the bracketed values with your own results, and report the internal consistency (for example Cronbach’s α) observed in your sample; see Cronbach’s alpha in SPSS.
FAQ
What does the PHQ-15 measure?
It measures the severity of somatic (bodily) symptoms, using 13 somatic symptom items plus sleep and tiredness items, each scored 0 to 2 for a total of 0 to 30.
What do PHQ-15 scores of 5, 10 and 15 mean?
They are the cutpoints for low, medium and high (2002) or mild, moderate and severe (2010 review) somatic symptom severity. The instruction manual also treats 10 or higher as a yellow flag and 15 or higher as a red flag.
Is the PHQ-15 a diagnostic test?
No. The instruction manual describes it as a way to derive severity scores, and its flags are screening heuristics. Interpretation of an individual’s result belongs with a qualified clinician.
How is the PHQ-15 different from the PHQ-9?
The PHQ-9 measures depressive symptoms; the PHQ-15 measures somatic symptoms. Their cutpoints (5, 10, 15) are shared, but the item content and constructs differ, and the 2002 study found that somatic and depressive severity affected outcomes differently.
Sources
- Kroenke, K., Spitzer, R. L., and Williams, J. B. W. (2002). The PHQ-15: Validity of a new measure for evaluating the severity of somatic symptoms. Psychosomatic Medicine, 64(2), 258-266. PubMed 11914441 (read 2026-11-08).
- Kroenke, K., Spitzer, R. L., Williams, J. B. W., and Löwe, B. (2010). The Patient Health Questionnaire somatic, anxiety, and depressive symptom scales: A systematic review. General Hospital Psychiatry, 32(4), 345-359. PubMed 20633738 (read 2026-11-08).
- PHQ/GAD-7 Instruction Manual, hosted by The Joint Commission: instructions_PHQ.pdf (read 2026-11-08).
