HAM-D: Hamilton Depression Rating Scale

Overview

The HAM-D is a clinician-rated scale for depression symptom severity. Survey Doctor uses the 17-item form hosted by the University of Florida, which produces a total from 0 to 52. Other published forms and score bands differ, so record the edition and scoring convention with every result.

Which edition Survey Doctor uses

Hamilton published his depression rating scale in 1960 and described its further development in 1967. Later forms changed the number of items, interview guidance, response anchors, optional branches, and score interpretation. The name "HAM-D-17" therefore does not identify every detail needed for reproducible administration.

Survey Doctor uses the University of Florida HDRS form. Its anchors follow the version in the 1976 ECDEU Assessment Manual, a U.S. government manual whose editors believed it was the author's version. The University of Florida form uses gender-neutral wording and routine chores in place of hospital ward chores. It asks for ratings based on a structured interview, and its commentary describes the 17 items as covering the past week.

Item 16, loss of weight, is rated either from the patient's account or from weekly weight measurements, never both. The paper form also has a "Not assessed" box. The ECDEU manual reserves that box for information the rater made no effort to collect, so it marks missing information rather than a score of 3. Survey Doctor does not offer it, and a total needs a rating on all 17 items.

Because these details change interpretation, record the form, interview method, recall interval, rater training, and scoring convention with any result. Do not compare totals from unidentified editions as though they were interchangeable.

How HAM-D scoring works

Nine items are rated 0 to 4 and eight are rated 0 to 2. Survey Doctor adds the 17 ratings and labels the total with these severity ranges:

Zimmerman et al. (2013) recommended these ranges after studying 627 outpatients with major depressive disorder at one practice. Their 17-item totals were derived from a semi-structured diagnostic interview, and the 17 and 24 boundaries best separated clinicians' mild, moderate, and severe global ratings. The original Hamilton paper (1960) defined no bands, and other publications use different boundaries. The lowest range should not be relabeled as remission without a protocol that defines remission and assesses the person more broadly.

A total describes the sum of the current ratings. It does not diagnose major depressive disorder, establish functional recovery, select treatment, or explain why a score changed.

What the scale covers

The current form rates mood, guilt, sleep, work and interests, movement, anxiety, physical symptoms, weight, and insight. Some ratings rely on what the patient reports. Others depend on clinician observation and judgment.

This mixture is one reason rater training matters. Sleep, appetite, energy, anxiety, and physical complaints can reflect depression, medical conditions, medication effects, or more than one cause. The total does not separate those explanations.

Administration and reliability

The HAM-D is clinician-administered, not a patient self-report questionnaire. Use the same specified interview method and trained rating process when comparing results over time.

Williams developed a structured interview guide because agreement on individual items with less structured administration was often only fair or poor. Williams (1988) found that added interview guidance improved agreement. That finding supports standardized administration; it does not prove that every modern HAM-D form produces the same result.

In one primary-care study, a specified semistructured 17-item protocol produced an intraclass correlation of 0.95 among 42 patients with paired ratings. Morriss et al. (2008) applies to that protocol and sample, not to every edition or interview.

Item 3 and safety

Item 3 rates suicide on the form's five anchors, from absent to a suicide attempt. In a native Survey Doctor rating, any Item 3 rating above zero records a safety signal and shows crisis resources, whatever the total. Any nonzero rating needs direct safety follow-up. The item score does not establish current suicide risk, intent, or required level of care.

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.

Clinicians should follow their organization's direct suicide-risk assessment and emergency protocol regardless of the total. A manual or imported total-only result cannot show the Item 3 rating, so its safety status is unknown. Do not infer the item rating from a total.

When the HAM-D is useful

The HAM-D can document clinician-rated symptom severity in a specialist or research protocol that names an exact form and administration method. Repeated ratings can describe symptom change when the same edition, recall period, rater process, and scoring convention are maintained.

The scale should not replace a diagnostic interview or a broader review of functioning and safety. It also should not be used as a self-screen. A self-report tool such as the PHQ-9 collects different information and is not an interchangeable shorter HAM-D.

Interpreting change

Research studies often define response or remission for their own protocols. Those conventions vary and should not be imported into an individual result without naming the source, population, edition, and purpose.

The total cannot show what changed. A lower total means the rated items summed to less under the same scoring method, but it does not identify the cause or show whether important symptoms and functional problems remain. Review the item pattern, functioning, adverse effects, safety, and clinical course.

For a plain-language walkthrough, read what your HAM-D score means.

Important limits

• The 0 to 52 range applies to the 17-item form used here, and the severity ranges come from one outpatient study.
• The HAM-D is a clinician rating, not a diagnostic interview or self-report screen.
• Interview structure and rater training affect reliability.
• Total-only results cannot recover Item 3 safety information.
• Severity ranges do not prescribe treatment, level of care, response, or remission.