Home / Resources / Psychological Measures / MDQ
Mood Disorder Questionnaire:
Scoring and Interpretation
The Mood Disorder Questionnaire (MDQ) is a self-report screening instrument that asks about lifetime experiences of elevated or unusually irritable mood and classifies a respondent as screen-positive or screen-negative instead of producing a severity score (Hirschfeld et al., 2000). MDQ scoring under the original three-part rule raises three issues: why seven yes answers alone do not make a positive screen, how sensitivity drops from 0.73 among psychiatric outpatients to 0.28 in the general population, and why screen-flag syntax must leave incomplete forms unclassified.
What the MDQ Measures
The Mood Disorder Questionnaire (MDQ) is a screening instrument developed to identify features consistent with bipolar-spectrum disorders. It asks about experiences of elevated or unusually irritable mood and associated symptoms across the respondent’s history, then about whether those experiences occurred together and how much they affected functioning. It does not measure depressive symptom severity the way the PHQ-9 or the BDI-II does.
The original validation by Hirschfeld et al. (2000) evaluated the MDQ among 198 psychiatric outpatients against a structured diagnostic reference interview. The MDQ serves as an aid in deciding who needs further evaluation, not as a replacement for a bipolar-disorder diagnosis, and its test performance can be weaker for bipolar II and in less-selected populations.
Structure of the MDQ
- Symptom indicators: thirteen yes/no indicators (mdq1 to mdq13), coded 1 for yes and 0 for no.
- Concurrence question: one yes/no indicator (mdq_same_time, 1 for yes and 0 for no) asks whether the endorsed symptoms occurred in the same period.
- Impact question: one impairment indicator (mdq_impact), coded for this example as 0 none, 1 minor, 2 moderate and 3 serious. These are analyst-assigned database codes, not necessarily the values printed on the form.
- Recall period: lifetime history of mood and energy symptoms rather than a two-week severity interval.
- Subscales and reversals: none, and no sum of all 15 questions.
- Versions: the original standard three-part algorithm should not be silently replaced with a symptom-count-only rule, although modified symptom-count algorithms appear in some research.
The instrument itself is available from the National Institute of Neurological Disorders and Stroke common data elements and the original article.
How to Score the MDQ
- Check that the 13 symptom variables are binary 0/1 and code the two follow-up questions as specified in the structure above.
- Calculate mdq_yes_count, from 0 to 13, by summing the 13 symptom flags.
- Create mdq_positive as 1 only where at least seven symptom items are positive, mdq_same_time is 1, and mdq_impact is moderate or serious (2 or higher). Otherwise mark it 0 when all required data are present.
- The example requires all 15 fields for a valid classification. An instrument-specific missing-response rule is not established in the original abstract, and missing indicators must not be treated as negative responses.
COMPUTE mdq_yes_count=SUM.13(mdq1 TO mdq13).
COMPUTE mdq_positive=$SYSMIS.
IF (NMISS(mdq1 TO mdq13,mdq_same_time,mdq_impact)=0)
mdq_positive=(mdq_yes_count>=7 AND mdq_same_time=1 AND mdq_impact>=2).
EXECUTE.symptoms <- paste0("mdq",1:13)
data$mdq_yes_count <- rowSums(data[symptoms],na.rm=FALSE)
complete <- complete.cases(data[c(symptoms,"mdq_same_time","mdq_impact")])
data$mdq_positive <- NA_integer_
data$mdq_positive[complete] <- as.integer(
data$mdq_yes_count[complete]>=7 &
data$mdq_same_time[complete]==1 & data$mdq_impact[complete]>=2)Interpreting MDQ Scores
| Score band | Interpretation |
|---|---|
| 0 to 6 endorsed symptoms | Insufficient yes-count for the original positive screen |
| 7 to 13 endorsed symptoms | Count criterion met; evaluate concurrence and impact |
| Yes-count of 7 or more, concurrence yes and moderate or serious impact | Positive MDQ screen under the original algorithm |
Source: Hirschfeld et al. (2000); NIH/NINDS MDQ common data element.
The validation study reported sensitivity 0.73 and specificity 0.90 for the standard algorithm in its psychiatric-outpatient setting. These are sample-specific screening performance estimates; positive or negative results do not establish or rule out bipolar disorder.
The MDQ is a screen, not a diagnosis or a longitudinal mania-severity score. It has reduced sensitivity for bipolar II and hypomania and performs very differently outside psychiatric settings: in a US general-population sample the same algorithm had sensitivity 0.28 and specificity 0.97 (Hirschfeld et al., 2003). Clinical evaluation is needed when bipolar symptoms are suspected.
Psychometric Properties
| Property | Published finding | Source |
|---|---|---|
| Original clinical validation | 198 psychiatric outpatients across five clinics, with a blind structured diagnostic comparator | Hirschfeld et al. (2000) |
| Diagnostic sensitivity | 0.73 for the original standard positive-screen threshold | Hirschfeld et al. (2000) |
| Diagnostic specificity | 0.90 in the same psychiatric-outpatient study | Hirschfeld et al. (2000) |
| Reliability | The available original abstract does not provide a transferable Cronbach α for the three-part screening algorithm | Hirschfeld et al. (2000) |
| General-population validity | Sensitivity 0.28, specificity 0.97 in a community sample (n = 695 with interview); far lower sensitivity than in outpatients | Hirschfeld et al. (2003) |
| Construct and screening limitations | The MDQ is more effective at identifying classic mania than bipolar II; performance depends on setting | NIH/NINDS MDQ data element |
These are source-specific results and cannot be assumed to hold as reliability or accuracy estimates in a new sample.
Using the MDQ in a Study
In cross-sectional datasets, report the prevalence of the original full-algorithm MDQ-positive screen with its exact components and confidence intervals; Incidence & Prevalence Analysis covers that estimate. To analyze predictors, Logistic Regression can model MDQ-positive status, provided the events-per-parameter issue and the outcome prevalence are addressed. A symptom-count analysis is a different endpoint and should be labeled separately.
The MDQ is not a homogeneous severity scale, so Cronbach’s alpha for the entire 15-question form is generally not a useful replacement for validating its screening algorithm. When diagnostic interview data exist, estimate Sensitivity & Specificity (Diagnostic Accuracy) with predictive values and confidence intervals, and examine ROC behavior through ROC Curve Analysis. These estimates change with clinical case mix.
Related field: Clinical Psychology
Reporting the MDQ
Bipolar-spectrum symptoms were screened using the Mood Disorder Questionnaire (MDQ; Hirschfeld et al., 2000). The standard three-part algorithm defined a positive screen as endorsement of at least seven of 13 lifetime symptom indicators, endorsement of symptom co-occurrence, and moderate or serious associated impairment. Participants were classified as positive, negative, or missing according to [missing-data rule]; a positive screen was not considered a clinical diagnosis.
References
Hirschfeld, R. M. A., Holzer, C., Calabrese, J. R., Weissman, M., Reed, M., Davies, M., Frye, M. A., Keck, P. E., Jr., McElroy, S., Lewis, L., Tierce, J., Wagner, K. D., & Hazard, E. (2003). Validity of the Mood Disorder Questionnaire: A general population study. American Journal of Psychiatry, 160(1), 178-180. https://doi.org/10.1176/appi.ajp.160.1.178
Hirschfeld, R. M. A., Williams, J. B. W., Spitzer, R. L., Calabrese, J. R., Flynn, L., Keck, P. E., Jr., Lewis, L., McElroy, S. L., Post, R. M., Rapport, D. J., Russell, J. M., Sachs, G. S., & Zajecka, J. (2000). Development and validation of a screening instrument for bipolar spectrum disorder: The Mood Disorder Questionnaire. American Journal of Psychiatry, 157(11), 1873-1875. https://doi.org/10.1176/appi.ajp.157.11.1873
National Institute of Neurological Disorders and Stroke. (2024). Mood Disorder Questionnaire (MDQ): Common data element. https://cde-fe.ninds.nih.gov/ninds/noc-report/F1375/Mood%20Disorder%20Questionnaire%20(MDQ)
Frequently Asked Questions
Is a symptom count of seven enough for a positive MDQ?
No. Under the original standard algorithm, a yes-count of seven or more must be accompanied by co-occurrence of the symptoms and moderate or serious impairment before the screen is positive.
Does an MDQ-positive result mean bipolar disorder?
No. The instrument identifies people who may benefit from a structured clinical assessment, and sensitivity and specificity depend heavily on the setting in which it is used.
Is the MDQ suitable for repeated symptom-severity tracking?
The original measure is a lifetime screening tool, not a recommended short-interval severity outcome, so it is a poor fit for repeated-measures designs that track change in symptom severity.
Should I calculate alpha for the MDQ?
Alpha does not evaluate the three-part clinical screening algorithm. Diagnostic accuracy against an interview is more informative, where such reference data exist.
NEXT STEP
Need MDQ Screen Status
Scored and Checked?
OnlineSPSS is a statistical consultancy for researchers. Share your design and dataset, and a statistically trained Research Coordinator reviews it. The scope and one fixed project fee are confirmed before work begins.
No Commitment Required
Strictly Confidential
Fixed Project Fee
In-House PhD-Level Statisticians
Same-Day Response
Serving Researchers Since 2014