PROMIS® Adult Health Profiles: Efficient Short-Form Measures of Seven Health Domains

Abstract

Background: There is a need for valid self-report measures of core health-related quality of life (HRQoL) domains.

Objective: To derive brief, reliable and valid health profile measures from the Patient Reported Outcomes Measurement Information System® (PROMIS®) item banks.

Methods: Literature review, investigator consensus process, item response theory (IRT) analysis, and expert review of scaling results from multiple PROMIS data sets. We developed 3 profile measures ranging in length from 29 to 57 questions. These profiles assess important HRQoL domains with highly informative subsets of items from respective item banks and yield reliable information across mild-to-severe levels of HRQoL experiences. Each instrument assesses the domains of pain interference, fatigue, depression, anxiety, sleep disturbance, physical function, and social function using 4-, 6-, and 8-item short forms for each domain, and an average pain intensity domain score, using a 0-10 numeric rating scale.

Results: With few exceptions, all domain short forms within the profile measures were highly reliable across at least 3 standard deviation (30 T-score) units and were strongly correlated with the full bank scores. Construct validity with ratings of general health and quality of life was demonstrated. Information to inform statistical power for clinical and general population samples is also provided.

Conclusions: Although these profile measures have been used widely, with summary scoring routines published, description of their development, reliability, and initial validity has not been published until this article. Further evaluation of these measures and clinical applications are encouraged.

Keywords: PROMIS®; health-related quality of life; information system; patient-reported outcome measurement.

Copyright © 2019 ISPOR–The Professional Society for Health Economics and Outcomes Research. Published by Elsevier Inc. All rights reserved.

Figures

Figure 1.

Test information functions by domain.

Figure 2.

Short form correlations with full banks by domain.

Figure 3.

Estimated power to detect a small effect (d = 0.2) in a general population for each of the short forms by domain. SF indicates short form.

Figure 4.

Estimated power to detect a small effect (d = 0.2) in a clinical population (1 SD below mean) for each of the short forms by domain. SF indicates short form.

Figure 5.

Mean T-scores by domain based on responses to the item: “In general, would you say your health is . .” The y-axis for PHF and SOR is on the right side of the panel, enabling a profile plot in which the upper part of each plot is consistently worse than the lower part of the plot. ANX indicates Anxiety; DEP, Depression; FAT, Fatigue; PAI, Pain Interference; PHF, Physical Function; SF, short form; SLP, Sleep Disturbance; SOR, Satisfaction with Participation in Social Roles and Activities.

Figure 6.

Mean T-scores by domain based on responses to the item: “In general, would you say your quality of life is … ” The y-axis for PHF and SOR is on the right side of the panel, enabling a profile plot in which the upper part of each plot is consistently worse than the lower part of the plot. ANX indicates Anxiety; DEP, Depression; FAT, Fatigue; PAI, Pain Interference; PHF, Physical Function; SF, short form; SLP, Sleep Disturbance; SOR, Satisfaction with Participation in Social Roles and Activities.