Xuxi Zhang

Social frailty  Furthermore, this research found that social frailty contributed to the ability of physical frailty  to explain physical HRQoL and to the ability of psychological frailty to explain mental HRQoL,  which was not reported by previous studies. Some studies reported that poor social contact  and support could influence HRQoL negatively. 10, 32  A qualitative study for older people in the  Netherlands found that ‘when participants’ health was poor, there was a shift from health to  social contacts as the most important aspect to quality of life’.  33  Other studies proved that  increasing social contact and social support were associated with better health behavior and  HRQoL. 34, 35  In frail people, where physical interventions are not practical, increasing social  contact or social support to reduce social frailty could be a proper choice to positively  influence HRQoL. 36  A previous study suggested that early identification and intervention can  enable frail people to maintain control over their HRQoL for longer. 21  Our findings suggest that  considering social frailty is important to improve both physical and mental HRQoL. They  implicate that health professionals and policy makers should pay more attention to social  frailty among older persons and could consider improving social support or social contact to  improve HRQoL of older people in Europe in the future.  Our study has some limitations. Although we made use of two validated questionnaires,  cultural differences in the interpretation of questions might still have caused some variation  between countries. In addition, the SF‐12 has been validated in UK, Greece, Croatia, The  Netherlands and Spain 37 , but the TFI has not been validated in all the five countries yet.  Currently, TFI is validated in the Netherlands 19  and Spain 38 . Nevertheless, our results indicate  that the TFI is a suitable screening instrument for assessing overall frailty as well as the three  domains of frailty in order to maintain or improve HRQoL. Secondly, we adopted cut points of  frailty and its three domains instead of exact scores to explore the association between frailty  and HRQoL which might cause information loss. However, we performed analyses on the  association between exact frailty scores and HRQoL (see Supplementary Table S3‐4). The only  difference was that the score of social frailty was negatively associated with PCS score in the  full model but no longer significant. All other significant results remained significant in the  same direction. Thirdly, relatively healthy participants may have enrolled to the study which  potentially caused selection bias. However, due to the inclusion of the rich data of 2327  participants at baseline, we do not expect that this limitation changed our findings. Finally,  the cross‐sectional design of this study did not allow to establish the causal relationship  between frailty and HRQoL. Our results support the need for further research on evaluating  the effects of frailty as well as the three domains of frailty on HRQoL.  2 39 Association between frailty and HRQoL