PROMIS Global Health v1.2: physical and mental health scores

What this assessment is

PROMIS Scale v1.2 Global Health is a 10-item adult self-report measure of broad physical, mental, and social health.

It produces two main results: a Global Physical Health T-score and a Global Mental Health T-score.

Four responses contribute to the physical score. A different four contribute to the mental score. The other two responses remain useful individual ratings but do not enter either component sum.

There is no standard 10-item total.

Keep both scores.

This page covers adult version 1.2. Earlier adult versions use different coding for some responses and must not be mixed with it without the specified recoding.

Versions matter.

The scores are not diagnoses.

Use version 1.2.

Who it was designed for

PROMIS, the Patient-Reported Outcomes Measurement Information System, was developed to measure health consistently across diseases and populations.

Adult Global Health is intended for a broad view rather than a detailed assessment of one condition. It is used in clinical care, research, population studies, and outcome monitoring.

The early PROMIS testing program evaluated global items alongside physical, mental, and social health domains in a large adult sample.

The person reports their own health. Pediatric, parent-report, and parent-proxy global health measures are separate forms with different scoring.

How it is administered

The adult scale contains 10 short ratings and usually takes less than a few minutes.

The questions do not all use the same time frame or response scale. Use the complete version 1.2 form and keep its item identifiers intact.

Do not rewrite the questions, make a custom short form, or substitute items from a PROMIS bank. Standard T-score conversion depends on the exact items and response coding.

Use the same administration mode and language when comparing scores over time.

What it measures

The Global Physical Health score summarizes broad physical function, pain, fatigue, and overall physical health.

The Global Mental Health score summarizes quality of life, mood, satisfaction with social activities and relationships, and overall mental health.

These are broad component scores. They do not provide separate validated subscale scores for pain, fatigue, depression, or social function.

The two responses outside the component sums concern important aspects of health but remain separate ratings under the standard method.

Do not add them.

Physical and mental health can diverge. Read both scores rather than treating one as the person's overall result.

Score them separately.

How scoring works

Code responses according to adult version 1.2.

Add the four responses assigned to Global Physical Health. Add the different four assigned to Global Mental Health. Each raw component sum ranges from 4 to 20 when all four responses are present.

Convert each raw sum with the matching version 1.2 table or an approved response-pattern scoring service. The output is a T-score.

PROMIS T-scores use a reference mean of 50 and a standard deviation of 10. Higher Global Physical Health and Global Mental Health scores indicate better reported health.

Fifty is not a cutoff.

The available T-score range is not 0 to 100. It is limited by the selected items. The exact range also differs between the physical and mental components.

The current PROMIS Global Health scoring manual prefers response-pattern scoring, especially when an answer is missing. Raw-sum lookup tables require complete component responses.

Do not prorate a missing response unless the scoring method being used explicitly supports it.

How to interpret PROMIS Global Health scores

A score of 50 is the mean of the reference population used for the calibration. Ten points represent one reference standard deviation.

That does not make 50 a pass mark. It also does not make every score below 50 clinically abnormal.

Higher scores indicate better reported global health. A physical score can be lower than the mental score, or the reverse. That pattern can help direct follow-up questions.

There are no universal mild, moderate, or severe bands that diagnose illness across all populations. If a study or service uses thresholds, it should name the condition, population, and purpose that support them.

For repeated assessment, keep the exact version and scoring method constant. Review the standard error when deciding whether a small difference is likely to reflect more than measurement uncertainty.

What the scores cannot tell you

PROMIS Global Health cannot diagnose a physical or mental health condition.

The mental component is not a depression or anxiety test. The physical component cannot identify the source of pain, fatigue, or functional difficulty.

A broad score can hide one severe problem. Someone may report good overall physical health while still having an important symptom that needs assessment.

The two components cannot identify which treatment caused a change. Health, support, expectations, medication, and life circumstances can all affect the result.

The scores do not replace a history, examination, focused symptom measure, or conversation about the person's priorities.

Evidence and limitations

Research on the PROMIS global physical and mental health scales supported the use of two brief component scores and examined their relationships with established health measures.

The scale's main strength is efficiency. It gives a common broad metric across different conditions and services.

Broad means brief.

That efficiency limits detail. Four items cannot describe every part of physical or mental health, and the component scores should not be treated as full diagnostic assessments.

Detail can be lost.

Version control is a practical risk. Some response coding changed between adult versions 1.0 or 1.1 and version 1.2. A mislabeled dataset can produce an incorrect score.

T-scores also depend on the correct item calibrations. Always report the exact instrument version, scoring service or table, respondent, and assessment date with the result.

Label every result.