Objective: To evaluate and synthesize interpretability metrics, including minimal important change (MIC), minimal important difference (MID), and minimal detectable change (MDC), across PROMIS and related systems (Neuro-QoL, TBI-QoL, SCI-QoL) in rehabilitation populations.
Data sources: Comprehensive searches of electronic databases (MEDLINE, EMBASE, PsycINFO, HaPI, CINAHL, Cochrane Library, Web of Science) and clinical trial registries (ISRCTN Registry, ClinicalTrials.gov) were conducted from inception through March 23, 2024, in consultation with an information specialist.
Study selection: Eligible studies assessed interpretability metrics in rehabilitation populations using PROMIS, Neuro-QoL, TBI-QoL, or SCI-QoL. Studies of pediatric or non-rehabilitation populations, abstracts, posters, or consensus statements were excluded. A total of 202 studies met inclusion criteria.
Data extraction: Two independent reviewers extracted study characteristics, interpretability metrics, and analytic methods following COnsensus-based Standards for the Selection of Health Measurement Instruments (COSMIN) guidelines.
Data synthesis: MIC, MID, and MDC values varied widely across populations and domains. PROMIS mental health domains (e.g., depression, anxiety, fatigue) demonstrated relatively consistent estimates, whereas physical function domains were more variable, particularly in chronic and geriatric groups. PROMIS Computer Adaptive Testing (CAT) measures showed fewer floor and ceiling effects than short forms, indicating enhanced sensitivity to change. Limited data were available for SCI-QoL and TBI-QoL.
Conclusions: Standardizing interpretability metrics and expanding research on SCI-QoL and TBI-QoL are critical to improving the clinical utility of these measures in rehabilitation. Future work should incorporate response-shift considerations and establish population-specific cut-points to support patient-centered care and evidence-based practice.
Objective: To examine the associations between obesity, defined using the spinal cord injury (SCI)-specific (>22 kg/m2) and the standard (≥30 kg/m2) body mass index (BMI) thresholds, and determine which index better discriminates against BMI-related cardiometabolic, physical, and psychosocial associations reported among the general population in chronic traumatic SCI (TSCI).
Design: Multicenter cross-sectional study.
Setting: Sixteen SCI Model System (SCIMS) sites.
Participants: Adults with TSCI (n=1523, 78.7% male, age 45.7±15.9 years, 56.7% tetraplegia, 8.5±10.5 years post-SCI). Participants were stratified into groups using BMI>22 (n=1,123) and BMI≥30 (n=376), based on available height and weight, follow-up data, and complete outcomes data from the 2016-2020 SCIMS database.
Interventions: Not applicable.
Main outcome measures: Prevalence and odds of self-reported cardiometabolic (diabetes, hypertension, hyperlipidemia), physical (arthritis, pressure injuries [PI], urinary tract infections [UTI], falls, rehospitalizations), and psychosocial (Patient Health Questionnaire-9, Resilience Short Form, Satisfaction with Life Scale, Self-perceived Health [SPH]) measures.
Results: Obesity prevalence was 73.7% using the SCI-specific threshold and 24.7% using the standard threshold. Individuals classified as obese by either definition had higher odds of diabetes, hypertension, and hyperlipidemia, with consistent findings across all neurological impairment categories. Arthritis was more prevalent among individuals with than without obesity, but increased odds were observed only for those with a BMI≥30. UTIs and PI were more common among participants with a BMI>22, while poorer SPH was associated with a BMI≥30. No significant associations with psychosocial outcomes were found using either threshold.
Conclusion: The SCI-specific BMI classified more persons as obese than the standard threshold, yet both thresholds were associated with cardiometabolic risk. Patterns diverged for other outcomes (arthritis, SPH at ≥30), suggesting common obesity-health risk patterns may not generalize to SCI. These findings highlight the complexity of obesity in SCI, suggesting that despite BMI's common use, more accurate, clinically accessible measures are needed to improve risk identification.

