Frailty and Health-Related Quality of Life Among Residents of Long-Term Care Facilities

Frailty Assessment

Due to simplicity and ease of application, we chose walking speed < 1.0 m/s as the primary criterion of frailty. Walking speed was calculated from the time to walk 15 feet. Gait speed was recently endorsed by the Geriatric Advisory Panel as the most suitable instrument for quick, inexpensive, and reliable instrument to measure frailty in older people (Cesari et al., 2005; Rolland et al., 2004). We also measured unintended weight loss (> 10 lb in the preceding year), and grip strength by Jamar® handgrip dynamometer as other criteria for frailty.

Quality of Life Assessment

HRQOL was assessed with the Short Form 36 (SF-36, McHorney, Ware, Lu, & Sherbourne, 1994; McHorney, Ware, & Raczek, 1993). The SF-36 is a generic instrument widely used to measure health status with the physical component score (PCS) and mental component score (MCS). Scores range from 0 to 100, with higher scores indicating better HRQOL. This instrument has well-established validity, reproducibility, and prognostic value (McHorney, et al., 1994; McHorney, et al., 1993). The overall internal-consistency reliability (Cronbach’s α) is 0.93 for the physical functioning scale and 0.90 for the mental health scale. Importantly, the reliability is similar among older adults aged 75 or older; the Cranbach’s α is 0.92 for the physical functioning scale and 0.86 for the mental health scale (McHorney, et al., 1994).

Statistical Methods

Continuous variables are summarized by mean ± standard deviation, unless otherwise specified. Categorical variables are presented as group frequencies and percentages. Student’s t-test was used to test differences in age, body mass index, and SF-36 scores between groups. The Wilcoxon rank sum test was used to test the difference in the Charlson comorbidity index between groups. Group differences in categorical variables were tested using Pearson’s χ² test. To obtain age and sex-adjusted p-values for group differences in continuous variables, the variable of interest was modeled using linear regression with age, sex and frailty, and the p-value for the frailty parameter estimate is reported. To obtain age and sex-adjusted p-values for categorical variables, a logistic regression model for frailty was calculated with age, sex, and the categorical variable as covariates. A likelihood ratio test for the significance of the categorical variable in the model was reported. Linear regression was used to model the SF-36 standardized physical and mental component scores as functions of age, sex, comorbid burden (Charlson Index) and frailty.

Frailty and Long-Term Care Facilities

There was a high and similar prevalence of frailty and comorbid conditions among residents of either nursing homes or assisted living facilities. HRQOL as measured by the SF-36 questionnaire was similar between the two groups as well. Prior studies have demonstrated higher mortality, disability, increased hospitalization, and health services utilization in older adults (Fried, et al., 1998; Singh et al., 2002; Singh et al., 2007). These relationships, however, have not been well studied in residents living in long-term care facilities. In a study of 3,782 nursing home residents, severe functional impairment and worsening activities of daily living were found to be associated with an increased risk of subsequent hospitalization (Fried & Mor, 1997). Almost a third (30.2%) of 11,113 community dwelling and nursing home residents were found to be frail according to one of three frailty models tested by the Health and Retirement Study (Cigolle, Ofstedal, Tian, & Blaum, 2009). Frailty in older adults living in the community was a predictor for nursing home admissions following discharge from emergency department (Hastings, Purser, Johnson, Sloane, & Whitson, 2008).

As noted however, these studies used differing methods to measure frailty. Recently, gait speed has been proposed as a quick, inexpensive, and reliable measure of frailty (Abellan van Kan et al., 2008). Indeed, in a recent study, data was pooled from nine cohorts of community-dwelling older adults to assess whether gait speed was associated with survival. Within these cohorts, the prevalence of gait speed < 1.0 m/s ranged from 20% to 96% and the prevalence of gait speed < 0.6 ranged from 1% to 60%. (Studenski et al., 2011). The study concluded that gait speed was indeed associated with survival in older adults.

Similarly, in the present study, we used gait speed to measure frailty, and found that a large proportion of LTCF residents were frail. Even after redefining frailty as walking speed less than 0.6 meter/second, over 60% of the residents were frail (Guralnik et al., 2000). This high prevalence was in stark contrast to 7% in community-dwelling elders (Fried, et al., 2001), 27% in residents with coronary heart disease (Purser et al., 2006), and 30% in the Health and Retirement Study (Cigolle, et al., 2009). This suggests we may need to define frailty differently in this population. Recently, definitions for frailty that may be more applicable to nursing home residents have been proposed (Abellan van Kan, Rolland, Morley, & Vellas, 2008).

Quality of Life Measures and Frailty

There was an association between HRQOL measures and frailty among LTCF residents. Frailty was not only associated with a lower SF-36 PCS score, but also with comorbidity. Each component of age-associated impairment, including frailty, comorbidity, and disability, is known to influence long-term prognosis, and a significant overlap between them was previously demonstrated by Fried and colleagues (Fried, et al., 2004). Information on self-perceived health status may further our understanding of the association with frailty and comorbid conditions.

Our finding of poor HRQOL in frail LTCF residents is important for several reasons. First, data on the cross-sectional association between poor QOL and frailty are scant. Similar to our results, Masel and colleagues found that frailty status was significantly associated with lower scores on all physical and cognitive health related QOL scales in older Mexican American individuals (Masel, Graham, Reistetter, Markides, & Ottenbacher, 2009). In another small study, 24 residents with frailty were noted to have lower QOL, as well as mental well-being, independent of age, diabetes, macrovascular complication, kidney dysfunction, and depressed mood (Kanauchi, Kubo, Kanauchi, & Saito, 2008). Recently, another cross-sectional study also found an inverse association between frailty and dimensions of QOL among 239 community dwelling outpatients aged 65 and over (Bilotta et al., 2010).

Secondly, measuring HRQOL will likely improve our understanding of its complex interplay with already established age-associated variables. It will also lay the foundation for developing effective and targeted strategies for prevention and management of the frail older adult population with poor health status. It is important to note the clustering of risk factors in these vulnerable LTCF residents who not only have higher preponderance of traditional risk factors but also have higher burden of comorbid conditions, frailty, and now poor HRQOL. The study raises the possibility for potential interventions in the frail older adults living in LCTFs that can arrest or reverse frailty in order to improve HRQOL. Recently a prognostic index, incorporating age, sex, self-reported comorbid conditions, and functional measures, accurately stratified community-dwelling older adults into varying risks of death and those with poor functional status had worse long-term outcomes (Lee, Lindquist, Segal, & Covinsky, 2006). As the sample size is small and the study presents cross-sectional associations, additional studies are needed with longitudinal follow-up to define the prognostic value of these constructs.