jfa journal

AND option

OR option

FRAILTY IS ASSOCIATED WITH INCREASED MORTALITY IN OLDER ADULTS 12 MONTHS AFTER DISCHARGE FROM POST-ACUTE CARE IN SWISS NURSING HOMES

 

C. Fompeyrine1,2, L.A. Abderhalden2, N. Mantegazza2, N. Hofstetter2, G. Bieri-Brüning3, H.A. Bischoff-Ferrari1,2,4, M. Gagesch1,2

 

1. Department of Geriatrics and Aging Research, University Hospital Zurich, Zurich, Switzerland; 2. Centre on Aging and Mobility, University of Zurich, Zurich, Switzerland
3. Zurich Geriatric Services and Nursing Homes, Zurich, Switzerland; 4. University Clinic for Acute Geriatric Care, City Hospital Waid and Triemli, Zurich, Switzerland.
Corresponding author: Michael Gagesch, MD, Dept. of Geriatrics and Aging Research, University Hospital Zurich, Raemistrasse 100, 8091 Zurich, Switzerland, michael.gagesch@usz.ch

J Frailty Aging 2020;in press
Published online October 20, 2020, http://dx.doi.org/10.14283/jfa.2020.58

 


Abstract

Frail older adults with ongoing care needs often require post-acute care (PAC) following acute hospitalization when not eligible for specific rehabilitation. Long-term outcomes of PAC in this patient group have not been reported for Switzerland so far. In the present report, we investigated 12-month mortality in regard to frailty status upon admission to PAC in a nursing home setting. In our sample of 140 patients (mean age 84 [±8.6] years) 4.3% were robust, 37.1% were pre-frail, 54.3% were frail and 4.3% were missing frailty status. Mortality at 12-months follow-up stratified by baseline frailty was 0% (robust), 11.5% (pre-frail) and 31.6% (frail). Kaplan-Meier analysis stratified by frailty status showed a decreased probability of 12-months survival for frail individuals compared to their pre-frail and robust counterparts (P = 0.0096). Being frail was associated with more than 4-fold increased odds of death at follow-up (OR 4.19; 95% CI 1.53-11.47).

Key words: Frailty, long-term mortality, post-acute care, nursing homes.


Introduction

Health in older age comprises a broad spectrum from robustness to vulnerability and frailty (1). The latter is associated with multiple negative outcomes in various care settings, from general practice to acute hospital care (2). With their often complex health status, older adult patients frequently require longer lengths of stay in the hospital and often remain at an increased level of care, impeding a prompt discharge home after an acute illness (3). At the same time, standard rehabilitation programs do not always appear suitable for many frail older patients (4).
Post-acute care (PAC) programs aim to bridge the gap between acute care and returning home for older adults, not otherwise eligible for rehabilitation. While earlier studies from different countries and specific settings (i.e. heart-failure patients) have demonstrated positive effects of PAC, such as reduced readmissions and decreased mortality rates (5, 6), its potential benefits and outcomes, particularly in frail older adults are still understudied (7). In addition, healthcare systems and PAC programs appear to have major differences between countries, hampering direct comparisons (8).
In a prior analysis, clinically significant improvements of physical function and ADL were reported in robust and frail Swiss older adults after a PAC program in nursing homes with a mean duration of 31 days (9). However, no research on the long-term outcomes of PAC in Swiss nursing homes exists so far (4). Therefore, the aim of our study was to investigate the association of frailty status upon admission to PAC with 12-month mortality in a real-world sample of Swiss older adults.

 

Methods

Study Design

We conducted a one-year follow-up study at designated PAC units of three municipal nursing homes within the City of Zurich, Switzerland. Written informed consent was obtained before study enrolment. The competent ethics committee of the Canton of Zurich approved our study (BASEC 2016-01069).

PAC Setting and Patients

Our study recruited consecutive patients 60 years and older referred to a PAC unit after acute care hospitalization between August and September 2016. An interdisciplinary team under the supervision of a board-certified geriatrician completed a comprehensive geriatric assessment (CGA) for each patient within one week upon admission, performed the PAC program and held bi-weekly team meetings. PAC consisted of activating nursing care (i.e. goal-directed instruction and training of ADL), five sessions of individual physical therapy per week and additional occupational therapy as needed, based upon the initial CGA. The maximum length of stay at PAC units was usually limited to 10 weeks duration and the effective date of discharge was based on the accomplishment of specific goals, derived from the individual care plan (9).
For our follow-up investigation, we matched the initial list of PAC patients with the death registry of the City of Zurich at one year after discharge. Living status and mortality date (if applicable) was recorded. We utilized frailty status from CGA at admission to a PAC unit according to the Fried frailty phenotype (items: unintentional weight loss, fatigue, slowness, weakness, low activity level) (10). Among numerous proposed frailty definitions, the Fried frailty phenotype is one of the most recognized and highly cited concepts and has been validated in various healthcare settings (2, 11). Patients with zero positive criteria were classified as robust, patients with 1-2 positive criteria as pre-frail and patients with ≥3 positive criteria were considered frail (10). In addition, we utilized further patient characteristics recorded at admission (Barthel-Index, Short physical performance battery (SPPB), Mini-Mental State Examination (MMSE) score, number of drugs and number of diagnoses) to describe the functional status and comorbidity burden.

Statistical Analysis

Three months and one year mortality rate after PAC discharge as well as further patient characteristics recorded at admission were calculated and stratified by level of frailty (robust, pre-frail, frail). Kaplan-Meier curves for visual representation were constructed for the overall sample to compare frail vs. robust and pre-frail at admission to PAC. Fisher’s exact test was used to evaluate whether mortality rate one year after discharge from PAC was independent of frailty status at admission. ANOVA and Chi-square test were used to evaluate whether there was a difference in mortality rate between frailty levels, as well as age and gender. Furthermore, a logistic regression model predicting mortality was evaluated to determine a possible association between frailty status upon admission to PAC and mortality rate on follow-up. The model was adjusted for age and gender. Statistical significance was determined as P<0.05 using 2-sided tests. All statistical analyses were performed using R v3.5.0 (The R Foundation for Statistical Computing, Vienna, Austria) and SAS v9.4 (SAS Institute, Inc. Cary, USA).

 

Results

Baseline Population

Our baseline sample consisted of n=140 patients, including 62.9% (n=88) women. Mean age at admission to PAC was 84 years (± 8.57). Mean length of stay at PAC was 31 days (± 16.5). In all, the most frequent diagnoses on admission to PAC were fractures (n=29), infections except pulmonary manifestations (n=18), mobility disorders (n=17), cognitive impairment (n=15), and heart disease (n=11), as reported earlier (12).

Mortality and Frailty Status

For n=139 patients, mortality status and mortality date at 3 and 12 months after discharge from PAC were applicable. At admission to PAC, 4.3% (n=6) of patients were robust, 37.1% (n=52) were pre-frail, 54.3% (n=76) were frail and 4.3% (n=6) were missing information on frailty status. The one-year mortality rate for the overall sample was 22.9% (32/140). One-year mortality rate stratified according to the different levels of frailty was 0% (robust, 0/6), 11.5% (pre-frail, 6/52) and 31.6% (frail, 24/76). Frailty status in relation to mortality, functional status and comorbidity burden is summarized in Table 1.

Table 1 Baseline characteristics and mortality after PAC stratified by frailty status

a. n=6 missing frailty status at admission; b. testing the difference between frailty levels; c. n=5 missing patients; d. SPPB, Short physical performance battery, n=10 missing patients; e. n=3 missing patients; f. n=1 missing patients; g. MMSE, Mini-Mental State Examination, n=7 missing patients; h. n=1 missing patient; i. n=2 deceased patients were missing frailty status at admission

 

For further analysis, we combined the group of robust and pre-frail patients, as none of the robust group deceased in the year following discharge from PAC. Our logistic regression model showed significantly increased odds of death for being frail (OR 4.19; 95% CI 1.53-11.47), and male gender (OR 3.19; 95% CI 1.28-8.0), but not for older age (OR 1.06, 95% CI 1.00-1.13 for each additional year).
Estimating survival with a Kaplan Meier analysis stratified by frailty status at admission to PAC showed a decreased probability of one-year survival for frail individuals, compared to patients classified as pre-frail or robust (P = 0.0096), Figure 1. In addition, each point increment on the frailty score at admission to PAC was associated with a decreasing one-year survival (P = 0.014).

Figure 1 Kaplan Meier estimates stratified for frailty status

 

Discussion

With more than one in two patients being frail and more than one in three being at risk for the condition (i.e. pre-frail) in our sample, frailty appears to be highly prevalent in Swiss older adults undergoing PAC in a nursing home setting. In comparison, the estimated prevalence of frailty in community-dwelling older adults in Switzerland is 5.8% (13). In our analysis, male gender and prevalent frailty were significantly associated with decreased survival at 12 months follow-up. In particular, frail patients had a greater than 4-fold increased odds for long-term mortality compared to their robust and pre-frail counterparts.
Our findings are in line with results from a prior study in older adults from Spain, where age, male gender and worse functional status were associated with higher 12-month mortality after acute illness (14). Our overall mortality rate of 22.9% is comparable to reports from earlier studies in former hospitalized geriatric patients from Germany and Italy (20.3% in Ritt et al. (15); 24.9% in Pilotto et al. (16)). However, those studies investigated one-year mortality after acute hospitalization without reporting on the utilization of PAC. Notably, patients in one of the aforementioned studies had a lower frailty prevalence at admission to acute care than our patient group (e.g. 43.3% vs. 54.3% in this study) (15).
When comparing the 12-month mortality rate of 31.6% in our frail patients with the aforementioned studies from acute care settings in Germany and Italy, it appears consistent with those reported by Ritt et al. (36.1%) and Pilotto et al. (24.9%) (15, 16). Of note, the higher mean age of patients in our study was more comparable to the first study (mean age >80 years), while Pilotto and colleagues investigated a sample with a mean age <80 years. Therefore, this difference is probably due to the influence of age in relation to the difference in mortality and warrants further investigation.
As a strength, our study is the first to report on the long-term outcomes of PAC in Swiss nursing homes and its association with frailty status. Further, we used a standardized operationalization of the Fried frailty phenotype, a derivation of the original version by Fried et al. (10). Our study also has its limitations. First, our sample size and short duration of patient recruitment limit the generalizability of our results. We also lack information on causes of death during follow-up. Furthermore, we had to cluster robust and pre-frail patients for our analysis, which might hinder comparisons to other studies. In addition, the frailty phenotype may not be the best frailty instrument to predict 12-month mortality in this patient group (15). Finally, our study did not include a control group of “standard” nursing home care residents to compare with our results regarding potential recovery time in the absence of specific interventions.

 

Conclusion

Our study in 140 former geriatric inpatients 12 months after discharge from PAC suggests that male gender and frailty status upon admission to PAC are significantly associated with increased long-term mortality in this group of Swiss older adults. While in line with prior studies from other populations, our study adds important knowledge on the specific situation in Switzerland. More studies are needed to further investigate the impact of PAC programs on short and long-term outcomes in Switzerland, including older adults affected by frailty.

 

Acknowledgments: We like to thank Marion Thalmann and Thomas Tröster for performing the initial data collection. In addition we thank Dr. Wei Lang for his statistical advice. Furthermore, we would like to thank all involved staff members at the participating municipal nursing homes in the City of Zurich.
Conflicts of Interest: The authors declare no conflict of interest.
Ethical Standards: The authors declare that the study porcedures comply with current ethical standards for research involving human participants in Switzerland. The study protocol has been approved by the Cantonal Ethics Committee of the Canton of Zurich, Switzerland (BASEC 2016-01069);
Open Access: This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits use, duplication, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license and indicate if changes were made.
Funding section: Open Access funding provided by University Hospital Zurich.

 

References

1. Buchner DM, Wagner EH. Preventing frail health. Clin Geriatr Med. 1992;8(1):1-17.
2. Dent E, Kowal P, Hoogendijk EO. Frailty measurement in research and clinical practice: A review. Eur J Intern Med. 2016;31:3-10.
3. Covinsky KE, Palmer RM, Fortinsky RH, Counsell SR, Stewart AL, Kresevic D, et al. Loss of independence in activities of daily living in older adults hospitalized with medical illnesses: increased vulnerability with age. J Am Geriatr Soc. 2003;51(4):451-8.
4. Kone I, Zimmermann B, Wangmo T, Richner S, Weber M, Elger B. Hospital discharge of patients with ongoing care needs: a cross-sectional study using data from a city hospital under SwissDRG. Swiss medical weekly. 2018;148:w14575.
5. Feltner C, Jones CD, Cené CW, Zheng ZJ, Sueta CA, Coker-Schwimmer EJ, et al. Transitional care interventions to prevent readmissions for persons with heart failure: a systematic review and meta-analysis. Ann Intern Med. 2014;160(11):774-84.
6. Prvu Bettger J, Alexander KP, Dolor RJ, Olson DM, Kendrick AS, Wing L, et al. Transitional care after hospitalization for acute stroke or myocardial infarction: a systematic review. Ann Intern Med. 2012;157(6):407-16.
7. Roberts PS, Goud M, Aronow HU, Riggs RV. Frailty in a Post-Acute Care Population: A Scoping Review. PM & R : the journal of injury, function, and rehabilitation. 2018.
8. Abrahamsen J, Rozzini R, Boffelli S, Cassinadri A, Ranhoff A, Trabucchi M. Comparison of Italian and Norwegian postacute care settings for older patients in need of further treatment and rehabilitation after hospitalization. The Journal of Aging Research and Clinical Practice (JARCP). 2015.
9. Thalmann M, Troster T, Fischer K, Bieri-Bruning G, Patrick B, Bischoff-Ferrari HA, et al. Do older adults benefit from post-acute care following hospitalisation? A prospective cohort study at three Swiss nursing homes. Swiss medical weekly. 2020;150:w20198.
10. Fried LP, Tangen CM, Walston J, Newman AB, Hirsch C, Gottdiener J, et al. Frailty in older adults: evidence for a phenotype. J Gerontol A Biol Sci Med Sci. 2001;56(3):M146-56.
11. Hoogendijk EO, Afilalo J, Ensrud KE, Kowal P, Onder G, Fried LP. Frailty: implications for clinical practice and public health. Lancet. 2019;394(10206):1365-75.
12. Tröster T, Thalmann M, Fischer K, Bieri-Brüning G, Beeler PE, Bischoff-Ferrari HA, et al. Frailty, underweight and impaired mobility are associated with institutionalisation after post-acute care. Swiss medical weekly. 2020;150:w20276.
13. Santos-Eggimann B, Cuenoud P, Spagnoli J, Junod J. Prevalence of frailty in middle-aged and older community-dwelling Europeans living in 10 countries. J Gerontol A Biol Sci Med Sci. 2009;64(6):675-81.
14. Baztan JJ, Galvez CP, Socorro A. Recovery of functional impairment after acute illness and mortality: one-year follow-up study. Gerontology. 2009;55(3):269-74.
15. Ritt M, Bollheimer LC, Sieber CC, Gaßmann KG. Prediction of one-year mortality by five different frailty instruments: A comparative study in hospitalized geriatric patients. Arch Gerontol Geriatr. 2016;66:66-72.
16. Pilotto A, Rengo F, Marchionni N, Sancarlo D, Fontana A, Panza F, et al. Comparing the prognostic accuracy for all-cause mortality of frailty instruments: a multicentre 1-year follow-up in hospitalized older patients. PLoS One. 2012;7(1):e29090.

PREVALENCE OF FRAILTY IN NURSING HOME RESIDENTS ACCORDING TO VARIOUS DIAGNOSTIC TOOLS

 

F. BUCKINX1,2, J.-Y. REGINSTER1,2, S. GILLAIN3, J. PETERMANS3, T. BRUNOIS1,2, O. BRUYÈRE1,2

 

1. Department of Public Health, Epidemiology and Health Economics, University of Liège, Belgium; 2. Support Unit in Epidemiology and Biostatistics, University of Liège, Belgium; 3. Department of geriatrics, University Teaching Hospital of Liège, Belgium; 4. Department of Sports Sciences, University of Liège, Belgium
Corresponding author: Fanny Buckinx, PhD Student, University of Liège, Belgium, Quartier Hôpital, avenue Hippocrate, 13, 4000 Liège, Belgium, Tel.: +32 4 366 49 33, Fax: +32 43 66 28 12, E-mail: fanny.buckinx@ulg.ac.be

J Frailty Aging 2017;6(3):122-128
Published online June 14, 2017, http://dx.doi.org/10.14283/jfa.2017.20

 


Abstract

Background: Although the theoretical foundations of frailty are well established in the literature, it remains an evolving concept lacking any unique definition or diagnostic criteria for use in clinical practice and epidemiological research. No consensus exists about the accurate prevalence rates of frailty. The various operational definitions of frailty can at least partly explain such discrepancies. Objective: To compare the prevalence of frailty, measured with different diagnostic tools, among elderly nursing home residents. Design: This is an analysis of baseline data collected among the SENIOR (Sample of Nursing home Elderly Individuals: an Observational Research) cohort. Setting: Nursing homes. Population: A total of 662 volunteer subjects from 28 nursing homes were included in this analysis. Among them, the mean age was 83.2 ± 8.99 years and 484 (72.5%) of them were women. Measurement: The percentages of frail and non-frail subjects were calculated according to 10 different definitions. Results: Prevalence of frailty varies from 1.70% (Frailty Index) to 76.3% (Groningen Frailty Indicator) depending on the tool used. Conclusions: The prevalence of frailty is highly dependent on the diagnostic tool used. It would be necessary to reach a consensus on which diagnostic tools to use if one wishes to have comparable data obtained in epidemiological studies.

Key words: Diagnostic tool, epidemiology, frailty, nursing homes, prevalence.


 

 

Introduction

With an ageing population, there is a growing interest in frailty. It may be regarded as a multidimensional geriatric syndrome of decreased resilience and resistance to stressors, resulting from cumulative decline across multiple physiological systems, causing vulnerability to adverse health outcomes such as falls, hospitalisation, institutionalisation and mortality (1). These adverse health effects in turn contribute to an increased demand for medical and social care and are associated with increased financial costs (2). Thus, one of the major challenges of geriatric medicine is to recognise these conditions as soon as possible and to halt (or slow) the downward spiral of increasing comorbidity and frailty. Although the theoretical foundations of frailty are well established in the literature, and the concept almost universally accepted, the practical effects and solutions remain controversial (3, 4). It remains an evolving concept lacking any unique definition or diagnostic criteria for use in clinical practice and epidemiological research (4). Multiple tools have been developed in recent years in order to diagnose this geriatric syndrome (5) and some of these tools have been widely used in epidemiological studies. Taking account of all such studies, the prevalence of frailty seems to increase with age, appears to be greater in women than in men and would appear to be more prevalent in people with  any combination of lower education or income, poorer health and higher rates of comorbid chronic disease and disability. However, no consensus exists about the accurate prevalence rates of frailty (6, 7). The various operational definitions of frailty used in these studies can at least partly explain such discrepancies (8). However, and to the best of our knowledge, no single study has investigated the impact of all these definitions of frailty on its prevalence in the same population. In nursing home populations, some studies have suggested that the prevalence of frailty is high, compared with non-institutionalised subjects (6, 7). The prevalence of frailty also depends on the countries (9). Indeed, a recent survey of 7510 community-dwelling older adults in 10 European countries found that the prevalence of frailty, according to frailty phenotype defined by Fried, was higher in southern than in northern Europe consistent with an unexplained north-south health risk gradient (10). African Americans are more likely to be frail than Caucasians (11). For these reasons, it is difficult to compare the results obtained in different studies, given the difference observed in the prevalence of frailty, which can be due to the inclusion of people living in different places, with different degrees of dependence or a different age range. However, it should be acknowledged that there is no specific operational definition of frailty validated for nursing home residents.  To the best of our knowledge, all existing tools to assess frailty have not been tested in this specific population. Indeed, only a few tools such as the frailty phenotype (12) or Clinical frailty Scale (13) have sometimes been used in studies performed in nursing homes, but a comparison between various tools has never been carried out. Therefore, the aim of this study was to compare the prevalence of frailty with regards to different diagnostic tools among elderly nursing home residents. Moreover, the differences in demographic and clinical characteristics of subjects diagnosed as frail according to the various definition of frailty are poorly understood and were also investigated in the present study.

 

Methods

Study design

This is an analysis of baseline data collected among the SENIOR (Sample of Elderly Nursing home Individuals: an Observational Research) cohort. The protocol was approved by the Ethics Committee of the University Teaching Hospital of Liège, under the number 2013/178.

Study subjects and setting

Residents of 28 nursing homes in the area of Liège, Belgium, were eligible for the study if they agreed to participate (i.e. informed consent). Subjects disoriented or unable to stand and walk (authorised technical support) were excluded from this research.

Data collection

Assessment of frailty

For each subject, frailty was measured using the 10 different diagnostic tools described below:
A) Clinical Frailty Scale (CFS) (14): this is based on a clinical evaluation in the domains of mobility, energy, physical activity and function, using descriptors and figures to stratify elderly adults according to their level of vulnerability. The score ranges from 1 (robust health) to 7 (complete functional dependence on others).
To measure the prevalence of frailty, all persons included in categories “terminally ill”, “very severely frail”, “severely frail”, “moderately frail” and “mildly frail”, were considered as “frail”.
B) Edmonton Frail Scale (EFS) (15): this samples 8 domains (Cognitive impairment, health attitudes, social support, medication use, nutrition, mood, continence, functional abilities). A score range between 0-3 is a robust state, 4-5 is a slightly frail state, 6-8 is a moderately frail state and 9-17 is a severely frail state.
All persons included in categories “severely frail”, “moderately frail” and “slightly frail” were considered as “frail”.
C) Frail Scale Status (16): this has 5 components: Fatigue, Resistance, Ambulation, Illness, and Loss of weight. Scores range from 0-5 and represent frail (3-5), pre-frail (1-2), and robust (0) health states.
D) Frailty index (17): this is expressed as a ratio of deficits present to the total number of deficits considered.  Frailty index includes 40 variables and the calculation was performed on the maximum number of deficits collected. Thus, participants were considered as frail when the ratio of deficits present to the total number of deficits considered was 0.25 (i.e. lowest quartile) or more (18, 19).
E) Frailty phenotype (7): this is a deficit across five domains. Thus, phenotype of frailty was identified by the presence of three or more of the following components: shrinking, weakness, poor endurance and energy, slowness and a low level of physical activity. The presence of one or two deficits indicates a pre-frail condition, and a total of three or more deficits indicates frailty while the absence of deficits indicates a robust state.
F) Groningen Frailty Indicator (GFI) (20): this consists of 15 self-report items and screens for loss of functions and resources in four domains: physical, cognitive, social, and psychological. Scores range from zero (not frail) to fifteen (very frail). A GFI score of 4 or higher was regarded as frail.
G) Sega grid (21): this establishes a risk profile of frailty and provides
reporting of problems and factors that may influence functional decline, including age, provenance, drugs, mood, perceived health, history of falls, nutrition, comorbidities, IADL, mobility, continence, feeding and cognitive functions. A score of 0, 1 or 2 is given for each item and a total over 11 points indicates a “very frail” condition, a score between 8 and 11 points indicates a frail condition while a score below 8 is a slightly frail condition.
All persons included in categories “frail” and “very frail” were considered as “frail”.
H) Share Frailty Instrument (Share-FI) (22): Using the five SHARE frailty variables (fatigue, loss of appetite, grip strength, functional difficulties & physical activity), D-Factor scores (DFS) were determined using the SHARE-FI formula and based on the DFS value, the subject could then be categorised as non-frail, pre-frail, or frail.
I) Strawbridge questionnaire (23): this defines frailty as difficulty in two or more functional domains (physical, cognitive, sensory, and nutritive). A score greater than or equal to 3 in more than one domain is considered vulnerable.
J) Tilburg Frailty Indicator (TFI) (24): The TFI consists of 2 parts. Part A contains 10 questions on determinants of frailty and diseases (multimorbidity); part B contains 3 domains of frailty (quality of life, disability, and healthcare utilisation) with a total of 15 questions on components of frailty. The threshold above which the participant is considered as frail is 5 points.
The objectives and the validation criteria of these various tools are shown in Appendix 1.

Other data collected

Other variables collected were socio-demographic data such as age or sex, anthropometric measurements such as weight, height, from which body mass index (BMI) was calculated, abdominal circumferences, type of institution, technical assistance for walking, drug consumption and medical history. The following clinical measurements were also collected:
–    Daily energy expenditure evaluated by the Minnesota Leisure Time Activities Questionnaire;
–    Cognitive skills assessed with the Mini Mental State Examination;
–    Nutritional status estimated by the Mini Nutritional Assessment;
–    Quality of life assessed by both the EQ-5D and the SF-36 questionnaires;
–    Activities of Daily Living estimated by the Katz index;
–    Comorbidities collected from the CIRS-G questionnaire;
–    Gait and body balance assessed using the Tinetti, the “Timed Up and Go” and the “Short Physical Performance Battery” tests  and gait speed

These data were collected during a face-to-face appointment with the patient. The same observer conducted all the tests in all nursing homes. The data were completed using the medical records.

Statistical analyses

Quantitative variables that were normally distributed were expressed as means ± standard deviation (SD), and quantitative variables that were not normally distributed were reported as medians and interquartile ranges (percentile 25, percentile 75). A Shapiro–Wilk test verified the normal distribution for all parameters. Qualitative variables were reported as numbers and frequencies (%). Participantss were defined as frail, or not, according to each of these 10 diagnostic tools. Then, the percentage of frail subjects for each definition was estimated. Afterwards, the degree of concordance between each definition was calculated by Cohen’s Kappa coefficient; the closer the value to 1, the better the concordance (i.e. k<0: disagreement, 0-0.2: very low agreement, 0.21-0.40: low agreement, 0.41-0.60: moderate agreement, 0.61-0.80: strong agreement, 0.81-1: excellent agreement). The percentage of pre-frail subjects was also assessed by 3 of these 10 definitions, that propose this intermediate state. The association between the different diagnostic tools and subject characteristics was assessed by multiple regression or logistic regression. All analyses were performed with Statistica 10 software and SAS Statistical package (version 9.3 for Windows). Results were considered statistically significant when 2-tailed p values were less than 0.05.

 

Results

Baseline characteristics of the population

A total of 662 subjects were included in this study. The mean age of the population was 83.2 ± 8.99 years and the population was predominantly women (72.5%). Participants’ demographic and clinical characteristics are shown in Table 1.

Table 1 Baseline characteristics of the population (n=662)

Table 1
Baseline characteristics of the population (n=662)

Prevalence of frailty according to different definitions

The prevalence of frailty varied from 15.2% (Frail Scale Status) and Frailty Index (83.7%) depending on the definition used. The percentage of pre-frail subjects varied from 28.0% (Clinical Frailty Scale) to 60.8% (Frailty phenotype) according to the definitions which propose this intermediate state. (Table 2).

Table 2 Number of frail subjects using the different definitions (n=662)

Table 2
Number of frail subjects using the different definitions (n=662)

Concordance between the different definitions of frailty

Table 3 presents the concordance between definitions. The concordance between the definitions was low (Overall Kappa Coefficient: 0.014 (-0.057 – 0.085)), with a Cohen’s Kappa coefficient which ranged from -0.77 (-0.85- -0.69), observed between Frailty Index and Sega gird, to 0.67 (0.61-0.73), observed between Frail Scale Status and Clinical Frailty Scale. Thus, participants diagnosed as frail with one definition are rarely diagnosed as frail with another definition. Nevertheless, reporting the Spearman’s correlation among the operational definitions without their categorization (i.e. continuous variables), these definitions follow similar patterns of increase in the risk of deficits. The correlations ranged between 0.13 (i.e. Edmonton frail scale and Strawbridge questionnaire) and 0.68 (i.e. Frailty Index and Frailty phenotype) and were all statistically significant.

Table 3 Concordance between definitions of frailty, estimated by Kappa Cohen’s coefficient (95% CI)

Table 3
Concordance between definitions of frailty, estimated by Kappa Cohen’s coefficient (95% CI)

A= Clinical Frailty Scale; B = Edmonton frail Scale, C= Frail Scale Status, D= Frailty Index, E= Frailty phenotype, F= Groningen Frailty Indicator, G= Sega Grid, H= Share Frailty Instrument, I= Strawbridge questionnaire, J= Tilburg Frailty indicator

Clinical characteristics of frail subjects

Depending on the tool, clinical characteristics of frail subjects appears to be different.Significant differences are observed regarding the age of participants, their sex, their walking support, their nutritional status evaluated by the Mini Nutritional Assessment, their quality of life assessed by the EQ-5D and by the SF-36, their functional abilities assessed by the Tinetti test, by the SSPB test and by gait speed (p<.0001 for all these data).

 

Discussion

In this study it was found, as expected, that the prevalence of frailty is highly dependent on the diagnostic tool used.  However, the ratios observed differ very widely, ranging from 1.70% to 76.3%, and this could have important consequences for clinicians, researchers and public health decision-makers.
Clearly, the diversity and the breadth of definition of frailty criteria would appear to have contributed to the wide range of prevalence found (6). Indeed, there are two main kinds of definition for frailty (one broad and the other physical) and a recent systematic literature review showed that studies using a physical definition consistently reported lower prevalence of frailty than those using a broad frailty definition (6). Frailty measurements can be grouped into three categories: subjective (i.e. self-reported, reported by participant or by a researcher), objective (i.e. directly measured components) or mixed (i.e. subjective and objective combined) (25). This may also have an impact on the prevalence of frailty.
A systematic review highlighted that the prevalence of frailty in community-dwelling elderly adults varied from 4.0% to 59.1% according to the diagnostic tool used (6). Contrary to the systematic review that compared various tools but in different populations, the present study evaluates the differences in prevalence of frailty using the different operational definitions within the same population. The results presented in the review are somewhat different from those obtained in this study, which could be explained by the difference in the populations studied. Nevertheless, the results presented here are more consistent with a recent meta-analysis which showed that the mean prevalence of frailty in nursing homes differed widely from study to study, ranging from 19.0% to 75.6% (26). One study, published in 2015, compared how different frailty measures predict short-term adverse outcomes (27). The results highlighted that, over a time interval of 10 months and among a sample of community-dwelling elderly individuals, the Groningen Frailty Index predicted an increase in IADL disability, and the Tilburg Frailty Indicator predicted a decline in quality of life. Actually, no study has yet investigated the predictive value, in a nursing home setting, of different operational definitions of frailty for the occurrence of different adverse health outcomes, and, to our knowledge, no operational definition of frailty has been validated among institutionalised people. And yet this would seem to be an important aspect to be explored in prospective studies to identify the best operational definition adapted to this particular population. This definition could then be considered as the gold standard among nursing home residents and could be used in clinical practice and research to make studies more comparable.  It is also important to point out that gait speed at usual pace was found to be a consistent risk factor for disability, cognitive impairment, institutionalisation, falls, and/or mortality. (28) In the population under study here, gait speed seemed significantly different according to different operational definitions of frailty used. It would be interesting to clarify the predictive value of this variable in future prospective studies in a nursing home setting.
In the present study, the prevalence of pre-frailty was between 28% and 60.8% and similar with other studies (6). It is important to note that people included in this study were volunteers, not disoriented and had to be able to move. Because of this selection, the most frail people have probably not been included in the study and, therefore, the prevalence of frailty in this study may be underestimated. Anyway, it is important to identify pre-frail people because preventive intervention programs can be implemented, thus modifying the rates of associated events (9).
Otherwise, the agreement between the definitions was very low. This means that the people diagnosed as frail are different depending on the diagnostic tool used. Nevertheless, the definitions seem to be correlated with each other.  This means that the frailest subjects, according one definition, are also the frailest ones, according to the other definitions; but the threshold between frail and robust is different depending on the operational definition used. Moreover, significant differences were found regarding the clinical characteristics of frail subjects diagnosed according to these 10 definitions. Indeed, depending on the diagnostic tool used, it seems that significant differences are observed concerning the age of the participants. Also, nutrition status is different depending on the definition used, and this could be explained because the different definitions do not evaluate systematically nutritional status or, at best, do it differently (anamnesis, weight loss). In addition, the quality of life of frail subjects according to the different tools seems different. This can also be explained because the quality of life is not always considered in the various diagnostic tools for frailty or based on a simple question.
Investigators use multiple scales to assess frailty, all of which count deficits in health. Frailty scales differ in the nature and number of deficits they count, which could explain the heterogeneity of frail persons according to different definitions. Because the characteristics of frail subjects are different depending on the tools used for the diagnosis of frailty, the long-term clinical consequences of frailty may also differ. Therefore therapeutic strategies will not be easily evaluated and implemented as long as studies do not use the same diagnostic tool.
Consensus does not yet exist regarding the component element of frailty (29)  and there is no validated operational definition for nursing home residents. From a clinical and Public Health point of view, further investigations identifying the best model of frailty in this specific population are needed in order to obtain comparable data in epidemiological studies. In clinical practice, it would improve the management of frailty. An unambiguous definition of frailty is of great importance for clinicians to identify those at an increased risk of adverse health outcomes, but also for policy makers to make cost-effective decisions in health care. In conclusion, the prevalence of frailty is highly dependent on the definition used. In addition, the concordance between the different modalities of diagnosis is low and this research reveals that the clinical characteristics of frail subjects diagnosed with varied definitions are different. As long as no consensus has been reached about the operationalisation of frailty, clinicians and policy-makers should be aware that differences between definitions exist and that it should have important consequences, at least in epidemiological research.

 

Acknowledgments: We thank all nursing homes who agreed to participate in this study.
Funding: Fanny Buckinx is supported by a Fellowship from the FNRS (Fonds National de la Recherche Scientifique de Belgique — FRSFNRS. www.frs-fnrs.be).
Conflict of Interest: No conflict of interest to disclose.

Appendix 1

References

1.    Bauer JM, Sieber CC. Sarcopenia and frailty: a clinician’s controversial point of view. Experimental gerontology. 2008;43:674-8.
2.    Lally F, Crome P. Understanding frailty. Postgraduate medical journal. 2007;83:16-20.
3.    Buckinx F, Rolland Y, Reginster J-Y, Ricour C, Petermans J, Bruyère O. Burden of frailty in the elderly population: perspectives for a public health challenge. Archives of Public Health. 2015;73:19.
4.    Bergman H, Ferrucci L, Guralnik J, et al. Frailty: an emerging research and clinical paradigm–issues and controversies. J Gerontol A Biol Sci Med Sci. 2007;62:731-7.
5.    Cesari M, Gambassi G, van Kan GA, Vellas B. The frailty phenotype and the frailty index: different instruments for different purposes. Age and ageing. 2014;43:10-2.
6.    Collard RM, Boter H, Schoevers RA, Oude Voshaar RC. Prevalence of frailty in community-dwelling older persons: a systematic review. Journal of the American Geriatrics Society. 2012;60:1487-92.
7.    Fried LP, Tangen CM, Walston J, Newman AB, Hirsch C, Gottdiener J, et al. Frailty in older adults: evidence for a phenotype. J Gerontol A Biol Sci Med Sci. 2001;56:M146-56.
8.    Castell MV, Sanchez M, Julian R, Queipo R, Martin S, Otero A. Frailty prevalence and slow walking speed in persons age 65 and older: implications for primary care. BMC family practice. 2013;14:86.
9.    Jurschik P, Nunin C, Botigue T, Escobar MA, Lavedan A, Viladrosa M. Prevalence of frailty and factors associated with frailty in the elderly population of Lleida, Spain: the FRALLE survey. Archives of gerontology and geriatrics. 2012;55:625-31.
10.    Santos-Eggimann B, Cuenoud P, Spagnoli J, Junod J. Prevalence of frailty in middle-aged and older community-dwelling Europeans living in 10 countries. J Gerontol A Biol Sci Med Sci. 2009;64:675-81.
11.    Xue QL. The frailty syndrome: definition and natural history. Clinics in geriatric medicine. 2011;27:1-15.
12.    Gonzalez-Vaca J, de la Rica-Escuin M, Silva-Iglesias M, et al. Frailty in INstitutionalized older adults from ALbacete. The FINAL Study: rationale, design, methodology, prevalence and attributes. Maturitas. 2014;77:78-84.
13.    Matusik P, Tomaszewski K, Chmielowska K, et al. Severe frailty and cognitive impairment are related to higher mortality in 12-month follow-up of nursing home residents. Arch Gerontol Geriatr. 2012;55(1):22-4.
14.    Rockwood K, Mitnitski A. Frailty defined by deficit accumulation and geriatric medicine defined by frailty. Clinics in geriatric medicine. 2011;27:17-26.
15.    Rolfson DB, Majumdar SR, Tsuyuki RT, Tahir A, Rockwood K. Validity and reliability of the Edmonton Frail Scale. Age and ageing. 2006;35:526-9.
16.    Morley JE, Malmstrom TK, Miller DK. A simple frailty questionnaire (FRAIL) predicts outcomes in middle aged African Americans. J Nutr Health Aging. 2012;16:601-8.
17.    Searle SD, Mitnitski A, Gahbauer EA, Gill TM, Rockwood K. A standard procedure for creating a frailty index. BMC geriatr. 2008;8:24.
18.    K. Rockwood AM. How might deficit accumulation give rise to frailty? J Frailty Aging 2012;1(1):8-12.
19.    Mitnitski A, Collerton J, Martin-Ruiz C, Jagger C, von Zglinicki T, Rockwood K, et al. Age-related frailty and its association with biological markers of ageing. BMC medicine. 2015;13:161.
20.    Baitar A, Van Fraeyenhove F, Vandebroek A, et al. Evaluation of the Groningen Frailty Indicator and the G8 questionnaire as screening tools for frailty in older patients with cancer. Journal of geriatric oncology. 2013;4:32-8.
21.    Schoevaerdts didier bs, Malhomme brigitte, Rezette céline, Gillet jean-bernard, Vanpee dominique, Cornette pascale, Swine christian. Identification précoce du profil gériatrique en salle d’urgences : présentation de la grille SEGA. La Revue de Gériatrie. 2004;29:169-78.
22.    Romero-Ortuno R, Walsh CD, Lawlor BA, Kenny RA. A frailty instrument for primary care: findings from the Survey of Health, Ageing and Retirement in Europe (SHARE). BMC geriatrics. 2010;10:57.
23.    Strawbridge WJ, Shema SJ, Balfour JL, Higby HR, Kaplan GA. Antecedents of frailty over three decades in an older cohort. The journals of gerontology Series B, Psychological sciences and social sciences. 1998;53:S9-16.
24.    Gobbens RJ, van Assen MA, Luijkx KG, Wijnen-Sponselee MT, Schols JM. The Tilburg Frailty Indicator: psychometric properties. Journal of the American Medical Directors Association. 2010;11(5):344-55.
25.    Bouillon K, Kivimaki M, Hamer M, et al. Measures of frailty in population-based studies: an overview. BMC geriatrics. 2013;13:64.
26.    Kojima G. Prevalence of Frailty in Nursing Homes: A Systematic Review and Meta-Analysis. J Am Med Dir Assoc. 2015.
27.    Coelho T, Paul C, Gobbens RJ, Fernandes L. Frailty as a predictor of short-term adverse outcomes. PeerJ. 2015;3:e1121.
28.    Abellan van Kan G, Rolland Y, Andrieu S, et al. Gait speed at usual pace as a predictor of adverse outcomes in community-dwelling older people an International Academy on Nutrition and Aging (IANA) Task Force. J Nutr Health Aging. 2009;13:881-9.
29.    Buckinx F, Rolland Y, Reginster JY, Ricour C, Petermans J, Bruyere O. Burden of frailty in the elderly population: perspectives for a public health challenge. Archives of public health = Archives belges de sante publique. 2015;73:19.
30.    Rockwood K, Song X, MacKnight C, et al. A global clinical measure of fitness and frailty in elderly people. CMAJ. 2005;173:489-95.