Physical function trends and their association with mortality in postmenopausal women.

OBJECTIVE

Research is limited regarding the predictive utility of the RAND-36 questionnaire and physical performance tests in relation to all-cause, cardiovascular disease (CVD), and total-cancer mortality in older women.

METHODS

Data on the RAND-36 questionnaire, gait speed, and chair stand performance were assessed in 5,534 women aged ≥65 years at baseline. A subset (n = 298) had physical function assessments additionally at follow-up (years 1, 3, or 6). Multivariable Cox proportional hazards regression models estimated associations (HR) for a 1-standard deviation (SD) difference in baseline RAND-36 scores and performance tests (alone and combined) with mortality outcomes in the overall cohort and in models stratified by enrollment age (<70 and ≥ 70 y). The relative prognostic value of each physical function exposure was assessed using the Uno concordance statistic.

RESULTS

A total of 1,186 deaths from any cause, 402 deaths from CVD, and 382 deaths from total-cancer were identified during a mean follow-up of 12.6 years. Overall, each 1-SD unit higher baseline RAND-36 score was associated with significantly lower all-cause mortality (HR =0.90) and discriminatory capacity (Uno = 0.65) that was comparable to each performance exposure (HRs 0.88-0.91; Uno = 0.65). These findings were consistent in women aged <70 and ≥ 70 years. The associations of RAND-36 and performance measures with CVD mortality and total-cancer mortality were not significant in multivariable models nor in age-stratified models.

CONCLUSIONS

The RAND-36 questionnaire is a reasonable substitute for tracking physical functioning and estimating its association with all-cause mortality in older adults when clinical performance testing is not feasible.