Xuxi Zhang

Table 3 Multivariate linear regression model (frailty and HRQoL)  PCS  MCS  Items  Crude  Model  Adjusted  Model   Crude  Model  Adjusted  Model   Frailty Yes vs. No ‐11.69 c ‐8.49 c ‐9.47 c ‐7.30 c Age  ‐0.17 c 0.13 b Sex   Female vs. male ‐1.55 b ‐1.18 a Country  The United Kingdom versus Spain  ‐5.42 c ‐1.87 a Greece vs. Spain  ‐0.12  ‐1.86 a Croatia vs. Spain  ‐4.58 c ‐6.35 c The Netherlands vs. Spain  ‐5.43 c 0.19  Education level  Primary or less vs. tertiary or higher  ‐1.95 a ‐2.50 b Secondary or equivalent vs. tertiary or  higher  0.34  ‐1.35  Living situation Living alone vs. living with others  1.22 a 0.98 a Life style Alcohol risk vs. no alcohol risk   1.34 b 0.71  Exercise once a week or less vs. more  than once a week  ‐7.50 c ‐3.71 c Smoking vs. not smoking  0.97  ‐0.25  Multi‐morbidity Yes vs. No ‐4.64 c 0.09  Adjusted R 2 , %  23.2  38.2  19.3  27.2  Note: Abbreviations: PCS, Physical Component Suammary summarized by the SF‐12; MCS,  Mental Component Suammary summarized by the SF‐12.  The Crude Model is the unadjusted model with frailty as determinant.  The Adjusted Model is the adjusted model with frailty and the covariates as determinants.   a p <0.05,  b  p <0.01,  c  p <0.001, significant P‐values in bold.  Table 4 presents the multivariate linear regression models for the domains of frailty and  HRQoL. Physical frailty had the strongest association with physical HRQoL. In the adjusted  models, the mean PCS score of physically frail participants was 9.94 lower than that of not  physically frail participants ( p < 0.001). The mean PCS score of psychologically frail participants  was 3.21 lower than that of not psychologically frail participants ( p < 0.001) and the mean PCS  score of socially frail participants 2.54 lower than that of not socially frail participants ( p  <  0.001). Among the three adjusted models, the amount of variance explained was largest for  physical frailty (42.6%).  2 35 Association between frailty and HRQoL