The sensitivity of the MOS SF-12 and PROMIS® global summary scores to adverse health events in an older cohort

To compare the predictive validity of two self-reported outcome measures, the Patient-Reported Outcome Measurement Information System (PROMIS) Global Health measure and the 12-item Health Survey (SF-12).

Data were obtained from 1286 persons (55% female) aged 61–77 responding to a longitudinal survey. Inter-correlations of the SF-12 and PROMIS physical and mental summary scores were examined. ROC and AUC analyses were conducted to compare mental health score sensitivity to high levels of depression symptoms. Multiple regression was used to assess physical health score sensitivity to adverse health events over 12-month follow-up.

All scores displayed negatively skewed distributions. The respective SF-12 and PROMIS physical (r = .78) and mental (r = .62) health scores displayed strong associations. Mental health scores provided useful discrimination of persons reporting high depression symptoms (AUC_SF12 = 0.90; AUC_PROMIS = 0.84), although the SF-12 provided better case discrimination. Decreases in physical health over time were associated with recurrent falls (B_SF12 = − 1.62; B_PROMIS = − 1.14) and hospitalisations (B_SF12 = − 1.69; B_PROMIS = − 1.11).

The SF-12 and PROMIS brief measures of physical and mental health assess related but distinct health constructs. However, they display comparable sensitivity to adverse health outcomes. Results from studies utilising the SF-12 and PROMIS global health measures should be compared with sensitivity to differences in the content and scoring of these measures.