THE ASSOCIATIONS OF MENOPAUSE AND HEALTH-RELATED QUALITY OF LIFE WITH HEALTH SERVICES UTILIZATION: RESULTS FROM THE STRIDE STUDY

Purpose: To quantify the impact of menopause and health-related quality of life on health services utilization in a cohort of midlife women.

Method: STRIDE is a longitudinal cohort of midlife women (aged 40-65 at baseline) enrolled from a single general internal medicine practice. Annual questionnaires include information about menopausal status and symptoms, medical comorbidities, health-related quality of life (RAND-36), as well as baseline age and race. For the first 3 years of the study, we abstracted information about all care received within the study site’s health system. Facility and professional fees were assigned based on the 2009 Medicare reimbursement schedule. We used multivariable linear mixed models to examine the associations among health services utilization (costs), menopausal status and symptoms, and the mental and physical health composites of the RAND-36 and to ascertain adjusted costs (in US dollars over 3 years) for each menopausal status and symptom group as well as each standard deviation of RAND-36 score below the mean. All analyses were adjusted for medical comorbidities, race, baseline age, and time.

Result: After removing women in the 99^th percentile of costs, 598 of 732 participants contributed at least one year of cost, menopausal status, and RAND-36 data. Compared to premenopausal women, those who reported ever having had a hysterectomy had significantly higher costs ($3,246, β=2.2, p=.03). No other menopausal status group had significantly different costs. However, women who reported hot flashes had significantly higher costs ($1,649, β=2.0, p=.04), as did women with worse physical, but not mental, health composite scores ($3250/ standard deviation, linear β=5.6, p<.001 and β=.6, p=.5, respectively). Compared to women with no medical comorbidities, those with ≥3 had significantly higher costs (β=2.3, p=.02). Non-white women had significantly higher costs than white women (β=2.4, p=.02). There was no cost difference based on baseline age (β=-.4, p=.7).

Conclusion: Contrary to prior reports, we did not find health services utilization to be affected by menopausal status, though utilization was affected by menopausal symptoms and health-related quality of life (and these, in turn, are impacted by menopausal status). Including symptoms and health-related quality of life as covariates provides better insight into determinants of health care utilization during mid-life and suggests that treatment of symptomatic hot flashes and attention to physical health limitations may decrease overall health care costs.