The Hamilton Depression Rating Scale (HAM-D, also called HDRS or HRSD) has been the dominant measure in depression research since Max Hamilton first published it in 1960. It remains the most widely used clinician-administered depression rating scale and the benchmark against which other measures are validated.
But the HAM-D was designed for research, not clinical practice. Using it effectively in modern psychiatric settings requires understanding both its strengths and its limitations, and knowing when other instruments might serve you better.
What the HAM-D measures
The HAM-D assesses depression severity through clinician ratings of patient symptoms. The original 17-item version covers mood symptoms (depressed mood, guilt, suicidal ideation, work and activities), anxiety symptoms (psychic and somatic anxiety), sleep disturbance (early, middle, and late insomnia), somatic symptoms (gastrointestinal, general, genital, weight loss), and other features (agitation, retardation, hypochondriasis, insight).
The scale uses a mixed scoring system: nine items use a 5-point scale (0-4) including depressed mood, guilt, suicide, work and activities, retardation, agitation, and both anxiety items. Eight items use a 3-point scale (0-2) covering the insomnia items, somatic symptoms, genital symptoms, weight loss, and insight. Total score range is 0-52 for the 17-item version.
Severity thresholds
| Score | Severity |
|---|---|
| 0-7 | Normal / clinical remission |
| 8-13 | Mild depression |
| 14-18 | Moderate depression |
| 19-22 | Severe depression |
| >=23 | Very severe depression |
Treatment response is defined as a 50% or greater reduction from baseline score. Remission means achieving a score of 7 or below. Partial response is 25-49% reduction.
HAM-D versions
The 17-item version (HAM-D17) is standard. The 21-item version adds diurnal variation, depersonalization, paranoid symptoms, and obsessive-compulsive symptoms. These identify depression subtypes but don't change the total score.
The HAM-D6 (a 6-item core symptom scale covering depressed mood, guilt, work and activities, retardation, psychic anxiety, and general somatic symptoms) has shown superiority to HAM-D17 in terms of scalability, transferability across patient groups, and sensitivity to change. Systematic reviews have found the HAM-D6 is as sensitive to treatment effects as the full 17-item version while being more focused on core symptoms.
The GRID-HAM-D provides structured anchor points for each symptom level and achieves excellent inter-rater reliability (ICC 0.93-0.99) even among inexperienced raters.
Administration
The HAM-D is clinician-rated. You score items based on clinical interview and observation, which differs fundamentally from self-report measures like PHQ-9.
Unstructured administration (15-20 minutes) means conducting your usual clinical interview and rating items afterward. This approach is faster but less reliable.
Structured administration (20-30 minutes) using the Structured Interview Guide for the HDRS (SIGH-D) or GRID-HAM-D improves reliability substantially. Trained raters achieve ICC values above 0.90, while untrained administration produces ICC in the 0.57-0.73 range. The investment in training pays off in consistent, meaningful scores.
Setting the time frame matters. HAM-D assesses the past week. Be explicit about this with patients. Cover all items systematically rather than skipping those that seem irrelevant. Rate observation-based items (agitation, retardation) from what you observe throughout the interview, not patient report. Score somatic symptoms carefully, rating only symptoms that appear related to depression rather than independent medical conditions.
Scoring and interpretation
Sum all 17 items for the total score. But understand what that score includes:
Heavy insomnia weighting: Three items assess insomnia. A patient with severe sleep disturbance but otherwise mild depression may score as moderately depressed.
Anxiety components: Psychic and somatic anxiety contribute substantially. Patients with anxious depression may score higher than their depressive symptoms alone would suggest.
Somatic emphasis: Multiple somatic items may inflate scores in patients with comorbid medical conditions.
For treatment monitoring, calculate both absolute change (current minus baseline) and percent change. A patient starting at HAM-D 22 who improves to 10 shows a 12-point absolute change and 55% reduction (meeting response criterion), but a score of 10 doesn't meet remission criterion (7 or below). Track both response and remission. Both matter clinically.
Beyond totals, examine item-level patterns. Any elevation on the suicide item (item 3) requires clinical attention regardless of total score. High insomnia scores suggest addressing sleep specifically. High anxiety scores may indicate adjunctive treatment or anxiety-focused interventions. The retardation vs. agitation pattern informs medication selection: activating antidepressants for retarded presentations, sedating options for agitated ones.
Strengths and limitations
Six decades of research validate HAM-D as a depression severity measure. The FDA recognizes HAM-D endpoints, and virtually all antidepressant trials use it as an outcome measure. This creates extensive normative data and comparability with published literature.
Clinician judgment integrates information from interview, observation, and clinical knowledge, which is more accurate than patient self-report when patients minimize symptoms, have limited insight, have cognitive impairment, or over-report for secondary gain.
However, the 20-30 minute administration time limits how often HAM-D can be used in busy practices. Despite thorough coverage of some domains, the scale underrepresents hopelessness and cognitive symptoms while overweighting neurovegetative symptoms. Designed in 1960, it reflects mid-20th-century understanding of depression.
The mixed scoring system (3-point and 5-point scales) complicates interpretation. Clinician-administered scales also introduce rater bias. Clinicians may rate patients they're treating as improving more than they actually are.
When to use HAM-D
HAM-D adds value in specific contexts:
Treatment-resistant depression evaluation: When standard treatments have failed, detailed clinician-administered assessment confirms severity, identifies specific symptom patterns, and documents the case for specialized treatments. The FDA's approval of esketamine (Spravato) for treatment-resistant depression defined TRD as non-response to at least two antidepressants at adequate dose and duration, situations where thorough assessment matters.
Key medication decisions: Initial evaluation, major medication changes, treatment response plateaus, and pre-authorization for expensive treatments all benefit from detailed symptom assessment.
Patients with questionable self-report: When patients minimize symptoms, over-report for secondary gain, have cognitive impairment, or limited insight, clinician-rated assessment provides more accurate data.
When to use alternatives
For routine monitoring during treatment, self-report measures are more practical. PHQ-9 takes 2-3 minutes for patients to complete, correlates well with HAM-D (r = 0.72), and enables assessment at every visit without consuming clinician time. The PHQ-9 actually has higher measurement accuracy in distinguishing depression severity than HAM-D17.
For measurement-based care programs requiring frequent, scalable assessment, self-report measures work; clinician-administered measures don't.
When you need better coverage of hopelessness and cognitive symptoms, consider the MADRS (Montgomery-Asberg Depression Rating Scale). It's also clinician-administered but designed specifically to be sensitive to treatment-related changes. MADRS and HAM-D have similar effect sizes in clinical trials (ES 0.49-0.53), though MADRS has less somatic emphasis. For a self-report option, CES-D-R provides strong cognitive symptom coverage.
If a particular domain is the clinical focus, like sleep, anxiety, or functioning, use domain-specific measures: HAM-A or GAD-7 for anxiety, dedicated sleep measures for insomnia.
Integrating HAM-D into practice
A practical approach: use HAM-D at initial evaluation to establish baseline severity and symptom pattern. Monitor with PHQ-9 at each visit for efficient tracking. Return to HAM-D at key transitions, such as treatment changes, suspected resistance, and treatment completion.
If implementing HAM-D regularly, invest in formal training for all raters, establish reliability procedures with periodic rating comparisons, use structured interview versions, and monitor for rater drift over time. Record total scores, individual item scores, time frame assessed, and administration method to enable meaningful longitudinal tracking.