What Is the PROMIS Depression Scale?

The PROMIS Depression Scale is a set of questionnaire tools developed through a National Institutes of Health initiative to measure self-reported depressive symptoms using a standardized scoring system centered on a population average of 50, with each 10-point increment representing one standard deviation. It was designed to replace or complement older depression questionnaires by offering shorter, more precise assessments that work across diseases and clinical settings. The scale comes in several formats, from a fixed eight-item short form to a computer-adaptive version that tailors questions in real time, and research over the past decade-plus suggests it performs on par with established instruments while often requiring fewer questions and less time.

What PROMIS Is and Why It Was Built

PROMIS stands for Patient-Reported Outcomes Measurement Information System. It grew out of the NIH Roadmap for Medical Research, a large cooperative initiative aimed at creating better tools for measuring how patients feel and function across all diseases, not just one condition at a time.¹ Before PROMIS, researchers and clinicians relied on dozens of disease-specific or population-specific depression questionnaires, each with its own scoring system. If one study used the PHQ-9 and another used the Beck Depression Inventory, comparing results required conversion tables and guesswork. PROMIS aimed to fix that by building large item banks calibrated to a single common metric, so that depression scores from a heart-failure clinic could be directly compared with scores from an oncology ward or a primary-care office.²

The depression item bank itself is a pool of questions covering feelings like sadness, worthlessness, hopelessness, and loss of interest. Each item has been calibrated using modern psychometric methods so that the entire pool measures a single underlying dimension of depressive symptoms. The practical upshot is flexibility: clinicians can pull from that pool in different ways depending on how much time and technology they have available.

How the Scoring Works

All PROMIS instruments report results as T-scores. A T-score of 50 represents the average for the U.S. general population, and every 10 points above or below 50 equals one standard deviation from that average. So a person who scores 60 on the PROMIS Depression Scale is reporting depressive symptoms roughly one standard deviation above what is typical. A score of 40 means symptoms one standard deviation below the average, which essentially indicates very few depressive complaints.

Severity thresholds have been examined across multiple studies. Research in musculoskeletal patients found that a T-score of 53 or higher was the optimal cutoff to predict a DSM-5 depressive disorder diagnosis, achieving about 86% specificity and 79% sensitivity.³ Work in cancer patients similarly used a T-score of 53 as the threshold for mild depression and 60 for moderate depression, with specificity rising above 90% at the moderate cutoff.⁴ In hand surgery patients, scores above roughly 60 were categorized as indicating heightened depressive symptoms, and a change of about four T-score points was treated as a clinically meaningful difference between groups.⁵

A useful way to think about this: if your score lands in the low 50s, you may be experiencing mild symptoms worth monitoring. If it climbs into the high 50s or above 60, those symptoms are substantially above what most people report, and a conversation with a clinician is warranted. The specific thresholds can shift depending on the population being studied, so no single cutoff replaces clinical judgment.

Ways to Take the Test

PROMIS Depression can be administered in three main formats, each suited to different settings.

Computer adaptive test (CAT): Software selects questions in real time based on your previous answers, zeroing in on your symptom level with fewer items. In heart-failure patients, the median number of items completed was just four, and 77% of participants finished in six or fewer items.⁶
Short forms: Fixed sets of four, six, or eight items drawn from the item bank. Everyone gets the same questions regardless of their answers. These are easy to administer on paper or a tablet and don’t require special software.
Full item bank: The complete pool of questions, rarely used in routine care but available for research when maximum measurement precision is needed.

Head-to-head comparisons show that CAT slightly outperforms static short forms in precision, but the difference is often small. Short forms performed only marginally worse than CAT in simulation studies, and a two-stage branching approach (starting with a brief screener and then branching to targeted items) narrowed the gap further.⁷ For most clinical purposes, the short forms are good enough. The CAT is a luxury that saves a few questions and squeezes out a little extra precision.

Reliability across formats is strong. In a study spanning ethnically diverse groups, short-form reliability estimates exceeded 0.90 regardless of the estimation method used.⁸ And the way you take the test, whether on a computer screen, a paper questionnaire, or even an automated phone system, does not meaningfully change the score. A multi-mode study found no significant differences across administration methods, with parallel-forms reliability ranging from 0.85 to 0.93.⁹ A separate study among cancer patients comparing paper, tablet, and interactive voice response also found no meaningful differential item functioning across modes.¹⁰

How It Compares to Older Depression Questionnaires

The PHQ-9 is probably the most widely used depression screener in clinical practice, so comparisons with PROMIS Depression matter. Across three randomized trials, correlations between the two averaged about 0.72, and each one-point change on the PHQ-9 corresponded to roughly a 1.25-point change in PROMIS T-scores.¹¹ A three-point change on the PHQ-9, which is generally considered a clinically meaningful shift, corresponded to roughly a 3.8-point change in PROMIS T-scores. Researchers have also confirmed that the PHQ-9 can be calibrated onto the PROMIS metric, meaning scores from the two instruments can be translated back and forth with reasonable accuracy.¹²

Compared with the Beck Depression Inventory (BDI-II), the picture is a bit messier. In heart-failure patients, agreement between PROMIS-linked and observed BDI-II scores was moderate at baseline and improved over time, but individual-level discrepancies could be substantial, with mean differences of about three points at baseline.¹³ In deep brain stimulation candidates, PROMIS Depression crosswalk scores predicted elevated BDI-II scores with 94% specificity but only 63% sensitivity, meaning the PROMIS tool was better at ruling depression in than ruling it out in that population.¹⁴

In a study of people with multiple sclerosis, the PROMIS depression measure, the PHQ-9, and the CES-D-10 all correlated more strongly with each other than with scales measuring pain, sleep disturbance, or fatigue. This suggests that all three instruments are genuinely tapping into the same depressive construct rather than bleeding into other symptom domains.¹⁵

How Accurate Is It as a Depression Screener

For clinical screening, the key question is whether PROMIS Depression can reliably distinguish people who meet formal diagnostic criteria for depression from those who do not. The answer is that it performs essentially as well as the PHQ-9, the current go-to screener. Across three clinical trials, the average area under the curve for major depression was 0.91 for both PROMIS eight-item short forms and the PHQ-9. At a T-score cutpoint of 55, the PROMIS scales achieved about 81% sensitivity and 84% specificity, nearly identical to the PHQ-9 at its standard cutpoint of 10.¹⁶

In cancer patients, both the PROMIS CAT and the eight-item short form achieved AUC values above 0.80 compared against a structured clinical interview, which is the gold standard for diagnosing depression. At the mild-depression cutoff of 53, the CAT had 83% sensitivity and the short form had 80%, both with negative predictive values above 90%.¹⁷ That high negative predictive value is practically useful: if your score falls below 53, there is a strong chance you do not have clinical depression.

One finding worth flagging is that in musculoskeletal patients, PROMIS Depression met the bar for adequate specificity but fell short on sensitivity for DSM-5 depressive disorders at a cutoff of 53.¹⁸ This underscores that no questionnaire is a substitute for clinical evaluation. A low score is reassuring; a high score is a flag. Neither is a diagnosis.

Using PROMIS Depression in Routine Clinical Care

One of the practical selling points of PROMIS is that it was built to integrate into electronic health records. In one large-scale implementation across an orthopedic system, three PROMIS CAT instruments (physical function, pain interference, and depression) were administered at over 23,000 patient encounters using tablets. The median time to complete all three was about two and a half minutes, and results were imported directly into the electronic medical record in real time for use during the visit.¹⁹ Registration times for follow-up visits increased by less than 30 seconds on average.

In oncology, EHR-integrated symptom monitoring has also proven feasible. A pilot at a comprehensive cancer center assigned PROMIS-based assessments to patients and found that about 60% of assigned assessments were started, with the vast majority completed at home through the patient portal rather than in the waiting room.²⁰ Emotional symptoms including depression and anxiety were the second most commonly flagged concern, reported in about 28% of assessments. Integrating these tools into EHRs means clinicians can see a depression score trend over time alongside labs and imaging, rather than treating mood as an afterthought during a brief appointment.²¹

Pediatric Versions and Cross-Cultural Use

PROMIS Depression is not limited to adults. A pediatric short form has been validated in clinical samples of children and adolescents, showing strong correlation with established youth depression measures and acceptable accuracy for detecting a depression diagnosis, with area-under-the-curve values in the range of 0.70 to 0.77.²² There is also a caregiver proxy version for situations where the child cannot self-report, such as very young patients or those with cognitive limitations. In pediatric burn survivors, agreement between proxy and self-report was moderate to good, though caregivers tended to report slightly worse symptoms than children reported for themselves.²³ The recommendation from that research is to use self-report whenever possible and to treat proxy reports as an acceptable substitute only when self-report is impractical.

Cross-cultural applications of PROMIS Depression are expanding. A validation study comparing samples from Ghana and Saudi Arabia found high internal consistency and strong structural validity in both groups, with measurement invariance holding at every level tested. Differential item functioning analyses revealed minimal item bias, supporting the idea that the depression construct is interpreted equivalently across these culturally distinct samples.²⁴ A Dutch-Flemish CAT version has also been developed with results showing high efficiency and precise measurement of depression.²⁵ Translation work like this matters because one of the core promises of the PROMIS system is that scores should be comparable across settings and populations. If a T-score of 60 means something different in Accra than in Riyadh, the whole framework breaks down. So far, it appears to hold up.

Technical Robustness Under the Hood

A common concern with any instrument built on psychometric modeling is what happens when the statistical assumptions do not perfectly hold. In the case of PROMIS Depression, researchers specifically examined what happens when individual items do not fit the measurement model as well as expected. The result: item misfit had a negligible impact on T-score estimates and severity classifications in the general population.²⁶ This is reassuring because real-world data is always a bit messier than the conditions under which a model was built. The system appears resilient to small deviations.

Validation against clinical change over time has also been encouraging. In a three-month observational study, the PROMIS depression item bank tracked expected changes in symptoms, supporting its sensitivity to actual shifts in a patient’s condition rather than just its ability to capture a snapshot at one point in time.²⁷

Newer Frontiers for PROMIS Depression Scores

Two emerging applications are pushing the scale beyond traditional clinic visits. The first is ecological momentary assessment, where patients receive brief check-ins on their phones at random times throughout the day. Researchers have adapted PROMIS Depression short forms into momentary versions delivered via text message, finding acceptable reliability and moderate to excellent agreement with standard recalled versions of the same measures.²⁸ Early work has focused on tracking symptoms after concussion and around surgery, with postoperative depression and anxiety measured through thrice-daily assessments for 30 days.²⁹ The appeal is obvious: instead of relying on a patient’s recall of how they felt over the past week, you get real-time data that can capture the ebb and flow of symptoms across days.

The second application is health economics. PROMIS scores on their own tell you about symptoms, but health-policy decisions often need a single number representing overall quality of life to evaluate whether a treatment is worth its cost. Researchers have developed mapping algorithms that translate PROMIS-29 profiles, which include the depression domain, into quality-adjusted life-year values compatible with the EQ-5D-5L, one of the standard tools used in economic evaluations.³⁰ This means a clinical trial that already collects PROMIS data can feed those results into cost-effectiveness analyses without needing to administer a separate instrument. It is a quiet but consequential development that extends PROMIS well beyond the exam room.