This study explored how quality of life (QoL) and frailty are related among 17,529 English individuals over a 16-year period. The relationship between QoL, measured by CASP-12, and frailty, assessed by FFM, showed a consistent inverse pattern, almost linear. Despite statistically significant cross-lagged coefficients between CASP-12 and FFM, indicating mutual influence over time, the practical longitudinal impact was seen as minimal.

Some of the results presented in Table 1 are consistent with previous studies that have utilized ELSA data. For example, Marshall, Nazroo [30] reported summary statistics for a sample of ELSA participants at wave 1, which showed an average age of 65, 54% of females, and a frailty score of 0.16. Similarly, Niederstrasser, Rogers [31] reported summary statistics for a sample of ELSA participants at wave 2, which showed an average age of 67, a distribution of wealth categories similar to our findings, and a frailty score of 0.16 at wave 2. Differences in cohort samples and frailty measures may have contributed to these slight differences. The analytical samples of these studies do not match our analytical sample, which means that comparison of results with theirs might not be appropriate. Most sample studies did not include refreshment samples in later waves. Marshall, Nazroo [30], excluded younger participants (i.e., < 60) [5] or presented the summary statistics for the sample by classified them by a variable group, such as gender [32] or survival [5]. In our study, long-term conditions (LTCs) were defined based on 16 health conditions, which differs from other studies that use ELSA. For instance, Nguyen, Chua [32] included 26 health conditions in their study at wave 2 in ELSA and reported that around 80% of participants had two or more LTCs.

There are few studies that investigate the two-way relationship between frailty and quality of life (QoL) in observational studies [15]. A cross-lagged panel model (CLPM) is one method that can be used to examine this kind of investigation [16, 17] although biased estimations might occur when an individual’s characteristics are not distinguished from those of the sample group. Our findings suggest that there is a minimal but significant bidirectional relationship between frailty and QoL, indicating that neither has a dominant effect on the other. The study revealed that the connection between frailty and QoL, as influenced by social background, is relevant under current conditions. This challenges the idea that one factor causes changes in the other and suggests that both contribute to the level of successful ageing at any given time. The consistent negative correlation between frailty and QoL over nine different time points supports this conclusion. Consequently, social care professionals may inquire with individuals in need of assistance to determine the specific support they require. It appears that women typically sought support to maintain or improve their QoL, while men tended to seek support to prevent severe frailty.

Additionally, it was observed that factors such as gender, age, net wealth, and multimorbidity had a significant impact on the relationship between Quality of Life (QoL) and frailty at a group level but were not as noticeable at the individual level. This indicates that although these factors play a role in the average relationship between QoL and frailty, they vary more when considering individual experiences over time. This could be due to limited available information, such as the few observations per individual. The long intervals (two years apart) between measurements and sample attrition caused by unobserved reasons like hospitalization or death could be other reasons. As far as we know, this is the first study to adjust the bidirectional association between frailty and QoL based on gender, age, net wealth, and LTCs.

A recent study showed that CASP-12, a shorter version of CASP-19, more independently captures quality of life [7]. It is critical to consider age ranges − 50–59, 60–69, and 70+ - to avoid measurement errors when using CASP-12. So, applying multi-group analysis is recommended if the range age of participants is beyond one of these three age groups [33].

The results of this study have important clinical and research implications. Clinically, the findings suggest that interventions aimed at addressing frailty and quality of life (QoL) may need to be more adaptable and dynamic, tailored to different stages of aging. Early interventions like physical rehabilitation or social support could potentially help prevent temporary declines in QoL due to frailty, and vice versa. However, once a certain level of frailty is reached, the relationship between frailty and QoL may stabilize or diminish, requiring a focus on broader health determinants [34, 35]. From a research perspective, longitudinal studies are essential to understanding how the relationship between frailty and QoL evolves over time, aiding in the prediction of outcomes and the development of long-term care strategies [15]. Additionally, these findings can inform public health policies, emphasizing the need for community-based programs focused on maintaining or improving QoL among older adults, particularly those at risk of or experiencing frailty.

The present study has notable strengths. It is the first study to investigate the reciprocal relationship between quality of life (QoL) and frailty over 16 years with a considerable sample size using the LCM-SR method. The quality of the dataset was excellent. The analysis was adjusted for several crucial factors, including sex, age, wealth, and long-term conditions (LTCs).

One of the limitations of this study is the presence of missing data in CASP-12 items. We addressed this issue by utilizing two imputation methods sequentially and we employed FIML to handle unbalanced samples across the nine waves under the structural equation model framework. Also, observations that were dropped out did not have any CASP-12 items, accounting for around 9% of the target sample. These individuals were more frail, older, and had less wealth. The impact of the bias was reduced by applying multiple group analysis. The time between measurements was two years, so more immediate impacts are not captured. The sample had a higher proportion of wealthier individuals although our analyses demonstrated that the results remain robust against net wealth differences. Although the sample size was large, we cannot assume generalizability in this work since participant weighting was not adjusted for representativeness. Another limitation is the reliance on self-reported measures of frailty without a physical test. While self-reported frailty and QoL can introduce bias due to inaccurate reporting or recall issues [36, 37], for instance, in studies relying solely on self-reported data, there is a risk of measurement error and bias that can affect the validity and reliability of the findings. On the other hand, it also serves as a strength by offering direct insights into the participants’ experiences, which is particularly relevant for studies focusing on quality of life and subjective health measures.

In addition, one factor that could influence the bidirectional relationship between frailty and quality of life is loneliness, especially among older adults in Great Britain. Research highlights the importance of considering diverse patterns of loneliness when developing policies and interventions to combat social isolation and improve the quality of life for older adults living in the community [38].

To summarize, a bidirectional relationship between QoL and frailty is close to linear and inversely proportional over time. Although the bidirectional cross-lagged for CASP-12 and FFM coefficients were statistically significant, the magnitude of the effects are small. Even when we considered factors like gender, age, wealth, and having multiple health conditions, we noticed some differences in the overall results between different people, but less within the same person over time. The study provides empirical evidence supporting a bidirectional association between QoL and frailty in older individuals who reside at home, providing valuable insights for healthcare providers to modify their services accordingly.