Journal of Clinical Epidemiology 67 (2014) 176e183
Construct validity and reliability of a two-step tool for the identification of frail older people in primary care Janneke A.L. van Kempena,*, Henk J. Schersb, Rene J.F. Melisa, Marcel G.M. Olde Rikkerta b
a Department of Geriatric Medicine, Radboud University Nijmegen Medical Centre, P.O. Box 9101, 6500 HB Nijmegen, The Netherlands Department of Primary and Community Care, Centre for Family Medicine, Geriatric Care and Public Health, Radboud University Nijmegen Medical Centre, P.O. Box 9101, 6500 HB, Nijmegen, The Netherlands
Accepted 13 August 2013; Published online 1 November 2013
Abstract Objectives: To study the reliability and construct validity of the EASY-Care Two-step Older persons Screening (EASY-Care TOS), a practice-based tool that helps family physicians (FPs) to identify their frail older patients. Study Design and Setting: This validation study was conducted in six FP practices. We determined the construct validity by comparing the results of the EASY-Care TOS with other commonly used frailty constructs [Fried Frailty Criteria (FFC), Frailty Index (FI)] and with other related constructs (ie, multimorbidity, disability, cognition, mobility, mental well-being, and social context). To determine interrater reliability, an independent second EASY-Care TOS assessment was made for a subpopulation. Results: We included 587 older patients (mean age 77 6 5 years, 56% women). According to EASY-Care TOS, 39.4% of patients were frail. EASY-Care TOS frailty correlated better with FI frailty (0.63) than with FFC frailty (0.52). A high correlation was found with multimorbidity (0.50), disabilities (0.53), and mobility (0.55) and a moderate correlation with cognition (0.31) and mental well-being (0.38). Reliability testing showed 89% agreement (Cohen’s k 0.63) between EASY-Care TOS frailty judgment by two different assessments. Conclusion: EASY-Care TOS correlated well with relevant physical and psychosocial measures. Accordingly, these results show that the EASY-Care TOS identifies patients who have a wide spectrum of interacting problems. Ó 2014 Elsevier Inc. All rights reserved. Keywords: Frailty; Elderly; Assessment; Construct; Reliability; Family physician
1. Introduction There is international consensus that the care for the increasing population of frail older patients needs improvement [1,2]. Evidence showed that frail older patients would benefit from more proactive care delivery and the integration of multidisciplinary care systems [3,4]. Timely provision of suitable care for frail older patients first demands identification of their health and welfare problems [5e7]. However, the existing identification instruments often have been developed in the research realm and for use
Funding: This project was funded by grant 60-61900-98-217 of the Dutch National Care for the Elderly Program, coordinated and sponsored by ZonMw, The Netherlands Organisation for Health Research and Development. Ethical approval: The Medical Research Ethics Committee (MREC) of the region Arnhem/Nijmegen, The Netherlands, approved the study. Number of approval is 2009/223. All participants gave informed consent before taking part in the study. Conflict of interest: None. * Corresponding author. Tel.: þ31-24-3619465; fax: þ31-24-3617408. E-mail address: [email protected] (J.A.L. van Kempen). 0895-4356/$ - see front matter Ó 2014 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.jclinepi.2013.08.008
in a hospital setting. As a consequence, they have not been validated in a primary care setting or have shown to be impractical for use in primary care [6,7]. The two most often used frailty concepts are the ‘‘accumulation of deficits’’ [8] and the ‘‘phenotype of frailty’’ [9]. Fried’s phenotype of frailty intends to identify physically frail patients [9]. The other frailty concept, the accumulation of deficits, states that the more deficits someone has, the greater the risk on an adverse health outcome and the frailer the person. This concept is operationalized by means of a Frailty Index (FI), which is a ratio of the count of deficits against the total number of deficits screened for [6,8,10]. The aforementioned frailty instruments are extensively studied and validated, also in a primary care setting [11,12]. However, neither is specifically developed for the need with which family physicians (FPs) want to identify frailty, that is, selecting the target population for integrated care [5,13]. This may explain why these instruments for frailty identification are not widely implemented in primary care [6]. Hence, we concluded that the already existing instruments do not correspond with the needs of FPs [7,13,14].
J.A.L. van Kempen et al. / Journal of Clinical Epidemiology 67 (2014) 176e183
What is new? This report is one of the first that demonstrate the validity of a frailty identification instrument specifically developed for and with primary care professionals. The results of this study show that EASY-Care TOS has good construct validity as it correlated well with relevant physical and psychosocial measures and other already validated frailty measures. This means that EASY-Care TOS identifies patients as frail when they have multiple problems on various domains, which suits the need in primary care to identify patient groups that might benefit from more proactive and integrated care. With EASY-Care TOS, family physicians and other primary care professionals have an efficient, practical, and validated frailty identification method for primary care.
Consequently, we developed the EASY-Care Two-step Older persons Screening (EASY-Care TOS) with input of stakeholders [15]. The complete EASY-Care assessment is a tool for geriatric assessment in primary care and is extensively studied [16]. This tool considers frailty in the ‘‘broadest’’ sense of the concept as it comprises physical and psychosocial aspects of frailty and aspects of the care context of the patient. The EASY-Care TOS is derived from this EASY-Care assessment tool. Although, it is more efficient because it makes use of available prior knowledge of the FP. Furthermore, using the professionals’ appraisal as the frailty decision, instead of a cutoff score, it fits with everyday clinical reasoning of the FP [15]. A feasibility study confirmed that in the Dutch primary care context, with emphasis on longitudinal continuity and a strongly established doctorepatient relationship, only about 10% of the elderly population evaluated with EASY-Care TOS step 1 needed a home visit for the frailty appraisal. The home visits can be conducted by practice nurses. Dutch FPs highly appreciated the two-step approach of EASY-Care TOS and mentioned that it is time saving. Furthermore, the professionals considered their clinical judgment of patients’ frailty more important than a numerical score [17]. This report describes the construct validity of the EASYCare TOS by comparing it with the mostly used other frailty constructs [Fried Frailty Criteria (FFC) and Frailty Index (FI)] and other related constructs: multimorbidity, disability, cognition, mobility, psychosocial functioning, and quality of life. As we hypothesized that the EASYCare TOS identifies patients as frail when they have multiple problems on various domains, we expected that the overall judgment of frailty according to EASY-Care TOS
177
would correlate stronger with the FI than with the FFC as the latter only measures physical frailty, whereas the FI is a measure for frailty on multiple domains. Because the EASY-Care TOS includes multimorbidity and disability in the frailty judgment, we expected that there would be considerable overlap with these concepts.
2. Methods 2.1. Study population and data collection Six FP practices (with in total 15 FPs) in and around Nijmegen (The Netherlands) assessed their patients of 70 years and older with the EASY-Care TOS between February 2010 and August 2011 (n 5 1,159) and asked these patients to participate in the study. These practices were situated in urban (n 5 2), suburban (n 5 1), and countryside (n 5 3) areas. Patients who were too ill to be assessed (eg, patients in a palliative trajectory) were excluded. Patients were also excluded if they were under the treatment of a geriatrician or underwent a comprehensive geriatric assessment (CGA) in the past 3 months as the information of the geriatrician might influence the frailty judgment of the FP. Of this sample, 587 older patients gave informed consent and were included in the study. 2.2. Measurements 2.2.1. EASY-Care TOS FPs and primary care nurses received education on frailty and its identification with EASY-Care TOS. They were instructed on using the following conceptual definition of frailty: a frail older person has decreased reserve capacity because of multiple health, mental, or social problems, which make the person vulnerable for changes in the biopsychosocial context, especially when compensations are lacking [13,16,18e21]. In the first step of EASY-Care TOS, FPs made a frailty judgment of their older patients based on prior knowledge, for example, from the patients’ record. For this purpose, FPs used a 14-item checklist concerning physical and psychosocial functioning of the patient. At the end, the FP decided whether the patient was frail. This decision was not based on a standardized score but on clinical reasoning and tacit knowledge of the FP. The patients for whom the FP felt that he did not have enough information to make this judgment (‘‘unclear’’) were invited for the second step of EASY-Care TOS. This second step was an in-home assessment by a primary care nurse, who assessed functioning and well-being by applying the EASY-Care assessment instrument [22]. Afterward, the nurse discussed the information gathered from the assessment with the FP, and subsequently, they judged functioning of the patient (good, fair, and poor) on eight domains (diseases, medication, cognition, vision and hearing, functional status, mobility, mental well-being, and social context) and decided whether the
178
J.A.L. van Kempen et al. / Journal of Clinical Epidemiology 67 (2014) 176e183
Table 1. Characteristics of participants Characteristic Age, mean 6 SD*** Sex (women), n (%) Native country (The Netherlands), n (%) Educational level, n (%) Low Middle High Missing Comorbidity CIRS-G, mean 6 SD*** Polypharmacy Number of medications, mean 6 SD*** Disability KATZ-15, mean 6 SD*** Memory MMSE, mean 6 SD*** Mobility Walking speed (m/s), mean 6 SD*** Sensory problems Poor hearing, n (%)* Poor vision, n (%)*** Mental well-being GDS-15, mean 6 SD*** HADS-A, mean 6 SD*** Social context Loneliness, n (%)*** Sometimes Often Quality of life Grade of 0e10, mean 6 SD*** Care use Days of hospitalizations in the past year, mean 6 SD** Hr/wk home care, mean 6 SD*** Frailty measures Fried Frailty Criteria, mean 6 SD*** Rockwood Frailty Index, mean 6 SD***
Total (n [ 587)
EC-TOS not frail (n [ 356)
EC-TOS frail (n [ 231)
76.8 6 4.8 330 (56.2) 555 (94.5)
75.4 6 4.0 194 (54.5) 341 (95.8)
78.9 6 5.2 136 (58.9) 214 (92.6)
319 (54.3) 242 (41.2) 23 (3.9) 3 (0.5)
181 (50.8) 159 (44.7) 14 (3.9) 2 (0.6)
138 (59.7) 83 (35.9) 9 (3.9) 1 (0.4)
8.1 6 4.4
6.4 6 3.7
10.5 6 4.3
4.7 6 3.1
3.9 6 2.7
5.9 6 3.4
1.4 6 1.9
0.7 6 0.9
2.5 6 2.5
27.3 6 2.6
27.9 6 1.8
26.4 6 3.4
1.1 6 0.6
0.9 6 0.2
1.3 6 0.8
266 (45.3) 270 (46.0)
148 (41.6) 128 (36)
118 (51.1) 142 (61.5)
2.1 6 2.2 2.3 6 2.8
1.5 6 1.7 1.8 6 2.3
3.0 6 2.5 3.1 6 3.3
141 (24.0) 19 (3.2)
66 (18.5) 8 (2.2)
75 (32.5) 11 (4.8)
7.5 6 1.0
7.7 6 0.8
7.2 6 1.1
1.4 6 5.2
0.8 6 3.9
2.4 6 6.6
1.0 6 2.1
0.4 6 1.0
1.9 6 2.8
0.95 6 1.13 0.25 6 0.11
0.54 6 0.76 0.20 6 0.08
1.58 6 1.30 0.33 6 0.11
Abbreviations: EC-TOS, EASY-Care Two-step Older persons Screening; SD, standard deviation; CIRS-G, Cumulative Illness Rating Scale for Geriatrics; MMSE, Mini-Mental State Examination; GDS-15, Geriatric Depression Scale. P-value frail vs. not frail: *P ! 0.05; **P ! 0.01; ***P ! 0.001. Educational level: low 5 primary and lower secondary education, middle 5 upper secondary education, and high 5 tertiary education. CIRS-G: range 0e56, the higher the score, the more multimorbidity. MMSE: range 0e30, score of 24 is indicative for cognitive problems. KATZ-15: range 0e15, the higher the scores, the more (I)ADL disabilities. Walking speed: of 4-m walking test. GDS-15: range 0e15, score of 6 is indicative for depression. HADS-A: range 0e21, score of 7 is indicative for anxiety complaints.
patient was frail. For the complete EASY-Care TOS, see Appendix A at www.jclinepi.com.
2.2.2. Other measurements To study the construct validity of the EASY-Care TOS, additional data of the participants were collected by means of a CGA by a geriatrician and geriatric nurse, at the geriatric outpatient clinic of the University Medical Centre in Nijmegen (The Netherlands). The geriatrician and geriatric nurse were blinded for the results of the EASY-Care TOS. The maximum time between EASY-Care TOS and additional data collection was 4 weeks.
The FFC and FI were constructed from the CGA. Patients were frail according to the FFC when they had three or more of the following criteria: (1) unintentional weight loss (self-reported unintentional weight loss of 10 pounds/ year), (2) self-reported exhaustion, (3) weakness (measured with grip strength), (4) slow walking speed (measured with 4-m walking test), and (5) low physical activity (based on self-reported physical activity) [9]. We used methods and cutoff scores as mentioned in the article by Fried et al. [9]. We used a 45-item FI, of which the items are presented in Appendix B at www.jclinepi.com. Patients were considered frail according to the FI when they had a score of 0.25 or more [23].
J.A.L. van Kempen et al. / Journal of Clinical Epidemiology 67 (2014) 176e183
179
The other measurements that were included in the CGA were the Cumulative Illness Rating Scale for Geriatrics (CIRS-G) [24], Mini-Mental State Examination [25], KATZ-15 Activities of Daily Living (ADL) scale [26], Short Physical Performance Battery [27], Geriatric Depression Scale (GDS-15) [28], HADS-A anxiety scale [29], RAND-36 mental well-being scale, and RAND-36 social functioning scale [30]. Additionally, we assessed the number of medications, vision problems, hearing problems, loneliness, and social isolation by means of an interview and a physical examination. We defined multimorbidity as the presence of at least two domains with a score of 2 or more on the CIRS-G and disability as the presence of one or more ADL disabilities on the KATZ scale.
correlations were calculated to compare the strength of the association of EASY-Care TOS with these measures. For the interpretation of the strength of the relations, we used the following criterion: a correlation of 0.1e0.3 is a small, 0.3e0.5 medium, and 0.5 and higher a high correlation [31]. The construct validity of each of the eight EASY-Care TOS domains was studied by comparing the outcomes for the domains with corresponding measures with one-way analyses of variance and Pearson chi-square tests. For the interrater reliability, we determined observed agreement between the first and second judgment and Cohen’s kappa. We also compared the frailty judgments of the different FPs with a Pearson chi-square test.
2.3. Interrater reliability
3. Results
For determining the interrater reliability, we recruited a subgroup of 19 patients. These patients were assessed with EASY-Care TOS for a second time by an independent FP and nurse, within 4 weeks after the first assessment. The FP and nurse were blinded for the results of the first EASYCare TOS assessment.
3.1. Characteristics of the participants
2.4. Data analysis To study whether our study sample was representative for the general population, we compared key characteristics of included and nonincluded patients. To study construct validity, we compared the results of the EASY-Care TOS with the results of the measures for frailty and the other related constructs with an independent t-test and a Pearson chi-square test. Additionally,
The study sample was significantly younger (78.9 6 5.9 vs. 76.6 6 4.8 years) and less frail according to EASY-Care TOS (53.7% vs. 39.4% frail) than the nonincluded patient population. We found no differences regarding gender, quality of life, and health-care use. Among the study participants, frail participants scored significantly worse on most measures compared with those who were not frail according to EASY-Care TOS but not on educational level, problems with hearing, help in the case of an emergency, and days of hospitalization (Table 1). 3.2. Construct validity According to EASY-Care TOS, 231 (39.4%) persons were frail. According to the FFC, 58 (9.9%) and the FI
Fig. 1. Relation between frailty according to EASY-Care TOS, the Fried Frailty Criteria, and the Frailty Index, and the relation between frailty according to these three frailty instruments, multimorbidity, and disability (in Venn diagrams). EASY-Care TOS, EASY-Care Two-step Older persons Screening.
180
J.A.L. van Kempen et al. / Journal of Clinical Epidemiology 67 (2014) 176e183
258 (44.0%). A total of 396 (67.5%) patients had multimorbidity and 140 (23.9%) patients had ADL disability. Fig. 1 shows the interrelation of the three frailty constructs and their interrelation with multimorbidity and ADL disability. About 89.7% of patients who were frail according to FFC were also frail according to the FI and EASY-Care TOS. The patients who were frail according to FFC had multimorbidity, and in 62% of the cases, this was combined with the presence of ADL disability. We found overlap (n 5 170) between frailty according to EASY-Care TOS and the FI. However, 60 and 84 patients were frail according to EASY-Care TOS or FI only, respectively. Almost all patients who were judged as frail according to the FI or EASY-Care TOS had multimorbidity, 92% and 87%, respectively. Less than half of these patients had ADL disability, 40.7% and 39.4%, respectively. One hundred seventy patients had neither comorbidity nor ADL disability. Of these patients, 24 were frail according to EASY-Care TOS, 13 according to the FI, and none according to the FFC. The 24 patients who were frail according to EASY-Care TOS had problems on multiple domains other than morbidity or ADL disability. Two patients had problems on the psychosocial domain only. Table 2 shows the correlation of the frailty judgment of EASY-Care TOS with the results of the other frailty instruments (FI and FFC) and measures of other related constructs. The correlation between frailty according to EASY-Care TOS and FFC was 0.52, and 0.63 between EASY-Care TOS and FI. EASY-Care TOS correlated highly with multimorbidity (0.50), disability (0.53), and mobility (0.55) and moderately with polypharmacy (0.34), cognition (0.31), mental well-being (0.38), and self-perceived health (0.35). Table 3 shows the comparison of the judgment on the eight domains of EASY-Care TOS with the objective results of measures on corresponding constructs. The scores of the objective measures differed significantly between the different judgments (good, fair, and poor) on the corresponding EASY-Care TOS domains. 3.3. Interrater reliability We found 89% interrater agreement between the first and second frailty judgment, with a Cohen’s k of 0.63. Furthermore, we found no significant differences between the judgments of the different FPs.
4. Discussion As hypothesized, the EASY-Care TOS correlated well with constructs that measure physical and psychosocial frailty. Furthermore, EASY-Care TOS correlated to a higher extent with the FI, a count of deficits on various domains, than it does with the FFC, a measure for physical frailty. It identifies patients as frail when they have multiple
Table 2. Correlation among EASY-Care TOS, other constructs of frailty, and constructs that are possibly associated with frailty Construct Frailty Fried Frailty Criteria Rockwood Frailty Index Physical functioning Multimorbidity (CIRS-G) Self-perceived health Medication Polypharmacy (number of medications) Cognition MMSE (I)ADL KATZ-15 ADL (KATZ-15) IADL (KATZ-15) Mobility SPPB Mental well-being Depression GDS-15 Anxiety HADS-A RAND-36, mental well-being subscale Social context RAND-36, social functioning Quality of life Grade between 0 and 10 Care use Hospitalization (d/yr) Home care (hr/wk) Age Education
Correlation coefficient 0.52*** 0.63*** 0.50*** 0.35*** 0.34*** 0.31*** 0.53*** 0.40*** 0.51*** 0.55*** 0.38*** 0.25*** 0.30*** 0.28*** 0.27*** 0.17** 0.42*** 0.39*** 0.08*
Abbreviations: Easy-Care TOS, EASY-Care Two-step Older persons Screening; CIRS-G, Cumulative Illness Rating Scale for Geriatrics; MMSE, Mini-Mental State Examination; ADL, activities of daily living; SPPB, Short Physical Performance Battery; GDS-15, Geriatric Depression Scale. P-value frail vs. not frail: *P ! 0.05; **P ! 0.01; ***P ! 0.001.
problems on various domains, which suits the need in primary care to identify patient groups that might benefit from more proactive and integrated care. For all three frailty instruments, having physical problems was a prerequisite for being frail, and all three identified very few patients as frail based on psychosocial problems only. In this respect, EASY-Care TOS does not differ from the other two frailty instruments. However, the FFC differed in the frailty selection as they selected the patients who had the worst functional status. The differences between the populations selected by EASY-Care TOS and FI are less apparent and do not seem to be systematic. Accordingly, literature showed that different frailty instruments do select different but overlapping patient populations [32e34]. There is still no consensus about an optimal definition of frailty and which instrument to use. However, an aspect often overlooked in this ‘‘chase’’ for the ultimate definition of frailty is that different definitions and instruments may aim to identify frailty for very different goals, for example, some instruments are impractical in a clinical setting but are very useful for policy planning
J.A.L. van Kempen et al. / Journal of Clinical Epidemiology 67 (2014) 176e183
181
Table 3. Relation between outcomes on the eight domains of EASY-Care TOS and matching constructs Judgment on the EASY-Care TOS domain Construct Physical functioning CIRS-G, mean 6 SD*** Self-perceived health, n (%)*** Excellent Very good Good Reasonable Poor Medication Number, mean 6 SD*** Cognition MMSE, mean 6 SD*** Vision and hearing Poor vision, n (%)*** Poor hearing, n (%)*** (I)ADL KATZ-15, mean 6 SD*** Mobility SPPB, mean 6 SD*** Mental well-being GDS-15, mean 6 SD*** HADS-A, mean 6 SD*** RAND-36 mental well-being, mean 6 SD*** Social context RAND-36 social functioning, mean 6 SD*** Loneliness, n (%)*** Never Sometimes Often
Good 6.9 6 3.8
Fair 11.5 6 4.2 (0.0) (0.8) (33.1) (54.3) (11.8)
Poor 13.4 6 3.6
29 (6.5) 55 (12.4) 251 (56.4) 109 (24.5) 1 (0.2)
0 1 42 69 15
0 (0.0) 0 (0.0) 2 (13.3) 9 (60.0) 4 (26.7)
4.4 6 3.1
6.8 6 2.9
6.5 6 4.0
27.9 6 1.6
25.0 6 3.5
19.8 6 5.3
150 (37.6) 132 (33.1)
106 (62.0) 119 (69.6)
14 (82.4) 15 (88.2)
1.0 6 1.4
3.2 6 2.6
6.0 6 0.7
10.2 6 1.7
7.6 6 2.6
4.5 6 3.1
1.5 6 1.6 1.6 6 2.0 18.2 6 11.5
3.5 6 2.5 4.1 6 3.5 31.4 6 15.4
6.3 6 3.4 7.4 6 3.9 55.5 6 19.3
1.3 6 0.8
1.7 6 1.1
2.4 6 1.4
370 (78.2) 95 (20.1) 8 (1.7)
55 (50.9) 42 (38.9) 11 (10.2)
2 (33.3) 4 (66.7) 0 (0.0)
Abbreviations: Easy-Care TOS, EASY-Care Two-step Older persons Screening; CIRS-G, Cumulative Illness Rating Scale for Geriatrics; SD, standard deviation; MMSE, Mini-Mental State Examination; ADL, activities of daily living; SPPB, Short Physical Performance Battery; GDS-15, Geriatric Depression Scale. P-value frail vs. not frail: ***P ! 0.001. CIRS-G: range 0e56, the higher the score, the more multimorbidity. MMSE: range 0e30, score of 24 is indicative for cognitive problems. KATZ-15: range 0e15, the higher the scores, the more (I)ADL disabilities. SPPB: range 0e12, score of !9 is indicative for a high risk on loss of mobility. GDS-15: range 0e15, score of 6 is indicative for depression. HADS-A: range 0e21, score of 7 is indicative for anxiety complaints. RAND-36 mental well-being: range 5e30, the higher the score, the worse mental well-being. RAND-36 social functioning: range 1e6, the higher the score, the worse social functioning.
[14]. This means that there may be not one optimal instrument for the identification of frailty, but the optimal instrument is the one that best suits the needs of those using it [14,35]. The correlation between the EASY-Care TOS and the FI is high, and both intend to identify patients with multiple problems on various domains. Drubbel et al. [12] recently showed that an FI based on routine health-care data of FPs is predictive for the risk on adverse health outcomes. This means that the FI probably can be a useful frailty identification instrument in primary care. Despite the high correlation and the fact that the FI and EASY-Care TOS aim to identify patients as frail when they have multiple problems, there are some differences between the two instruments.
The FI uses a count of deficits, and the EASY-Care TOS uses the professional’s appraisal as frailty judgment. One of the main advantages of the EASY-Care TOS over the FI for use in primary care with the purpose of identifying the target population for integrated care is that the FP directly has an overview of the nature of the problems of the patient. A strength of this study is that it is unique in comparing a frailty instrument using the professional’s appraisal as frailty judgment with more objective measurement instruments. Furthermore, it is the first time that a frailty identification instrument is validated that was specifically developed for use in primary care with inputs from stakeholders.
182
J.A.L. van Kempen et al. / Journal of Clinical Epidemiology 67 (2014) 176e183
This study also has some limitations. First, our study population is not a completely accurate representation of the general population. Compared with the screened population, the participants were younger and were less frail according to the EASY-Care TOS. This probably is because participants had to come to the hospital for the geriatric assessment, which could be too aggravating for the severely frail patients. This will give an underestimation of frailty and related problems in our study population. However, the prevalence of frailty in our study is comparable with that in other studies. Collard et al. [36] found a weighted average prevalence of frailty according to the FFC of 9.9%. The Health and Retirement study using the FI found a prevalence of frailty of 32%. In this study, however, the included patients were aged 51 years and older [32]. Second, EASY-Care TOS identifies more than one-third of patients as frail, which is a large selection. This would mean that all these patients are in need of integrated care. We can imagine that the selected patients are not all in need of integrated care but will all benefit from proactive care. Further study is needed to find out whether these frail patients really are in need of integrated care or that we need a stricter selection. Third, we did not use validated measurement instruments for some domains, for example, not for medication and vision and hearing. Therefore, we used information from the interviews and physical examination of the geriatric assessment. Finally, only a small study population was available for determining the interrater reliability, and most patients in this subpopulation appeared to be not frail. Furthermore, the time between the first and second interval had a maximum of 4 weeks. Preferably, the second assessment was conducted at the same moment. This was because of organizational reasons, which made it not possible. However, we asked the FPs whether things changed in the health status of the patient in the weeks between the first and second assessment. We only included patients with no change in health status. A former study already established the feasibility of the EASY-Care TOS [17]. Additionally, this study shows the construct validity and interrater reliability of EASY-Care TOS. This instrument enables the FP to identify patients with multiple problems on physical, mental, and social domains. A planned prospective follow-up study will have to evaluate whether EASY-Care TOS indeed identifies those patients as frail who will benefit from integrated care. In conclusion, we think that EASY-Care TOS suits the need of FPs for an efficient, practical, and validated frailty identification method for primary care.
Acknowledgments The authors thank Hanny Hordijk, Leny Theunisse, Marlies Hoogsteen, Marieke Perry, Sarah Robben, and
Brechtje Lubberink for their help with the data collection and George Borm for his help with statistics. Appendix Supplementary data Supplementary data related to this article can be found at http://dx.doi.org/10.1016/j.jclinepi.2013.08.008. References [1] Iliffe S. Family medicine in an aging society. Can Fam Physician 2009;55(5):463e7. [2] Boeckxstaens P, De Graaf P. Primary care and care for older persons: position paper of the European Forum for Primary Care. Qual Prim Care 2011;19(6):369e89. [3] Wilhelmson K, Duner A, Eklund K, Gosman-Hedstrom G, Blomberg S, Hasson H, et al. Design of a randomized controlled study of a multi-professional and multidimensional intervention targeting frail elderly people. BMC Geriatr 2011;11:24. [4] Eklund K, Wilhelmson K. Outcomes of coordinated and integrated interventions targeting frail elderly people: a systematic review of randomised controlled trials. Health Soc Care Community 2009; 17(5):447e58. [5] De Lepeleire J, Iliffe S, Mann E, Degryse JM. Frailty: an emerging concept for general practice. Br J Gen Pract 2009;59(562):e177e82. [6] Lacas A, Rockwood K. Frailty in primary care: a review of its conceptualization and implications for practice. BMC Med 2012;10:4. [7] Pialoux T, Goyard J, Lesourd B. Screening tools for frailty in primary health care: a systematic review. Geriatr Gerontol Int 2012;12(2): 189e97. [8] Rockwood K, Mitnitski A. Frailty in relation to the accumulation of deficits. J Gerontol A Biol Sci Med Sci 2007;62(7):722e7. [9] 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):M146e56. [10] Mitnitski A, Song X, Skoog I, Broe GA, Cox JL, Grunfeld E, et al. Relative fitness and frailty of elderly men and women in developed countries and their relationship with mortality. J Am Geriatr Soc 2005;53(12):2184e9. [11] 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 Geriatr 2010;10:57. [12] Drubbel I, de Wit NJ, Bleijenberg N, Eijkemans RJ, Schuurmans MJ, Numans ME. Prediction of adverse health outcomes in older people using a frailty index based on routine primary care data. J Gerontol A Biol Sci Med Sci 2013;68:301e8. [13] Gobbens RJ, Luijkx KG, Wijnen-Sponselee MT, Schols JM. Toward a conceptual definition of frail community dwelling older people. Nurs Outlook 2010;58(2):76e86. [14] Sternberg SA, Wershof Schwartz A, Karunananthan S, Bergman H, Mark Clarfield A. The identification of frailty: a systematic literature review. J Am Geriatr Soc 2011;59(11):2129e38. [15] van Kempen JA, Schers HJ, Jacobs A, Zuidema SU, Ruikes F, Robben SH, et al. Development of an instrument for the identification of frail elderly as a target population for integrated care. Br J Gen Pract 2013;63(608):e225e31. [16] Nourhashemi F, Andrieu S, Gillette-Guyonnet S, Vellas B, Albarede JL, Grandjean H. Instrumental activities of daily living as a potential marker of frailty: a study of 7364 community-dwelling elderly women (the EPIDOS study). J Gerontol A Biol Sci Med Sci 2001;56(7):M448e53. [17] Keiren SM, van Kempen JA, Schers HJ, Olde Rikkert MG, Perry M, Melis RJ. Feasibility evaluation of a stepped procedure to identify
J.A.L. van Kempen et al. / Journal of Clinical Epidemiology 67 (2014) 176e183
[18]
[19]
[20] [21]
[22] [23]
[24]
[25]
[26]
community-dwelling frail older people in general practice. A mixed methods study. Eur J Gen Pract 2013. In press. Strawbridge WJ, Shema SJ, Balfour JL, Higby HR, Kaplan GA. Antecedents of frailty over three decades in an older cohort. J Gerontol B Psychol Sci Soc Sci 1998;53(1):S9e16. Bandeen-Roche K, Xue QL, Ferrucci L, Walston J, Guralnik JM, Chaves P, et al. Phenotype of frailty: characterization in the women’s health and aging studies. J Gerontol A Biol Sci Med Sci 2006;61(3): 262e6. Rockwood K, Hogan DB, MacKnight C. Conceptualisation and measurement of frailty in elderly people. Drugs Aging 2000;17(4):295e302. Schuurmans H, Steverink N, Lindenberg S, Frieswijk N, Slaets JP. Old or frail: what tells us more? J Gerontol A Biol Sci Med Sci 2004;59(9):M962e5. Richardson J. The Easy-Care assessment system and its appropriateness for older people. Nurs Older People 2001;13(7):17e9. Rockwood K, Andrew M, Mitnitski A. A comparison of two approaches to measuring frailty in elderly people. J Gerontol A Biol Sci Med Sci 2007;62(7):738e43. Salvi F, Miller MD, Grilli A, Giorgi R, Towers AL, Morichi V, et al. A manual of guidelines to score the modified cumulative illness rating scale and its validation in acute hospitalized elderly patients. J Am Geriatr Soc 2008;56(10):1926e31. Folstein MF, Folstein SE, McHugh PR. ‘‘Mini-mental state’’. A practical method for grading the cognitive state of patients for the clinician. J Psychiatr Res 1975;12(3):189e98. de Rooij SE, Govers AC, Korevaar JC, Giesbers AW, Levi M, de Jonge E. Cognitive, functional, and quality-of-life outcomes of patients aged 80 and older who survived at least 1 year after planned
[27]
[28]
[29] [30] [31] [32]
[33]
[34]
[35] [36]
183
or unplanned surgery or medical intensive care treatment. J Am Geriatr Soc 2008;56(5):816e22. Guralnik JM, Simonsick EM, Ferrucci L, Glynn RJ, Berkman LF, Blazer DG, et al. A short physical performance battery assessing lower extremity function: association with self-reported disability and prediction of mortality and nursing home admission. J Gerontol 1994;49(2):M85e94. Yesavage JA, Brink TL, Rose TL, Lum O, Huang V, Adey M, et al. Development and validation of a geriatric depression screening scale: a preliminary report. J Psychiatr Res 1982;17(1):37e49. Zigmond AS, Snaith RP. The hospital anxiety and depression scale. Acta Psychiatr Scand 1983;67(6):361e70. Hays RD, Sherbourne CD, Mazel RM. The RAND 36-item health survey 1.0. Health Econ 1993;2(3):217e27. Cohen J. A power primer. Psychol Bull 1992;112(1):155e9. Cigolle CT, Ofstedal MB, Tian Z, Blaum CS. Comparing models of frailty: the Health and Retirement Study. J Am Geriatr Soc 2009; 57(5):830e9. van Iersel MB, Rikkert MG. Frailty criteria give heterogeneous results when applied in clinical practice. J Am Geriatr Soc 2006; 54(4):728e9. Sourial N, Wolfson C, Bergman H, Zhu B, Karunananthan S, Quail J, et al. A correspondence analysis revealed frailty deficits aggregate and are multidimensional. J Clin Epidemiol 2010;63:647e54. Martin FC, Brighton P. Frailty: different tools for different purposes? Age Ageing 2008;37:129e31. Collard RM, Boter H, Schoevers RA, Oude Voshaar RC. Prevalence of frailty in community-dwelling older persons: a systematic review. J Am Geriatr Soc 2012;60:1487e92.
Construct validity and reliability of a two-step tool for the identification of frail older people in primary care Janneke A.L. van Kempena,*, Henk J. Schersb, Rene J.F. Melisa, Marcel G.M. Olde Rikkerta b
a Department of Geriatric Medicine, Radboud University Nijmegen Medical Centre, P.O. Box 9101, 6500 HB Nijmegen, The Netherlands Department of Primary and Community Care, Centre for Family Medicine, Geriatric Care and Public Health, Radboud University Nijmegen Medical Centre, P.O. Box 9101, 6500 HB, Nijmegen, The Netherlands
Accepted 13 August 2013; Published online 1 November 2013
Abstract Objectives: To study the reliability and construct validity of the EASY-Care Two-step Older persons Screening (EASY-Care TOS), a practice-based tool that helps family physicians (FPs) to identify their frail older patients. Study Design and Setting: This validation study was conducted in six FP practices. We determined the construct validity by comparing the results of the EASY-Care TOS with other commonly used frailty constructs [Fried Frailty Criteria (FFC), Frailty Index (FI)] and with other related constructs (ie, multimorbidity, disability, cognition, mobility, mental well-being, and social context). To determine interrater reliability, an independent second EASY-Care TOS assessment was made for a subpopulation. Results: We included 587 older patients (mean age 77 6 5 years, 56% women). According to EASY-Care TOS, 39.4% of patients were frail. EASY-Care TOS frailty correlated better with FI frailty (0.63) than with FFC frailty (0.52). A high correlation was found with multimorbidity (0.50), disabilities (0.53), and mobility (0.55) and a moderate correlation with cognition (0.31) and mental well-being (0.38). Reliability testing showed 89% agreement (Cohen’s k 0.63) between EASY-Care TOS frailty judgment by two different assessments. Conclusion: EASY-Care TOS correlated well with relevant physical and psychosocial measures. Accordingly, these results show that the EASY-Care TOS identifies patients who have a wide spectrum of interacting problems. Ó 2014 Elsevier Inc. All rights reserved. Keywords: Frailty; Elderly; Assessment; Construct; Reliability; Family physician
1. Introduction There is international consensus that the care for the increasing population of frail older patients needs improvement [1,2]. Evidence showed that frail older patients would benefit from more proactive care delivery and the integration of multidisciplinary care systems [3,4]. Timely provision of suitable care for frail older patients first demands identification of their health and welfare problems [5e7]. However, the existing identification instruments often have been developed in the research realm and for use
Funding: This project was funded by grant 60-61900-98-217 of the Dutch National Care for the Elderly Program, coordinated and sponsored by ZonMw, The Netherlands Organisation for Health Research and Development. Ethical approval: The Medical Research Ethics Committee (MREC) of the region Arnhem/Nijmegen, The Netherlands, approved the study. Number of approval is 2009/223. All participants gave informed consent before taking part in the study. Conflict of interest: None. * Corresponding author. Tel.: þ31-24-3619465; fax: þ31-24-3617408. E-mail address: [email protected] (J.A.L. van Kempen). 0895-4356/$ - see front matter Ó 2014 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.jclinepi.2013.08.008
in a hospital setting. As a consequence, they have not been validated in a primary care setting or have shown to be impractical for use in primary care [6,7]. The two most often used frailty concepts are the ‘‘accumulation of deficits’’ [8] and the ‘‘phenotype of frailty’’ [9]. Fried’s phenotype of frailty intends to identify physically frail patients [9]. The other frailty concept, the accumulation of deficits, states that the more deficits someone has, the greater the risk on an adverse health outcome and the frailer the person. This concept is operationalized by means of a Frailty Index (FI), which is a ratio of the count of deficits against the total number of deficits screened for [6,8,10]. The aforementioned frailty instruments are extensively studied and validated, also in a primary care setting [11,12]. However, neither is specifically developed for the need with which family physicians (FPs) want to identify frailty, that is, selecting the target population for integrated care [5,13]. This may explain why these instruments for frailty identification are not widely implemented in primary care [6]. Hence, we concluded that the already existing instruments do not correspond with the needs of FPs [7,13,14].
J.A.L. van Kempen et al. / Journal of Clinical Epidemiology 67 (2014) 176e183
What is new? This report is one of the first that demonstrate the validity of a frailty identification instrument specifically developed for and with primary care professionals. The results of this study show that EASY-Care TOS has good construct validity as it correlated well with relevant physical and psychosocial measures and other already validated frailty measures. This means that EASY-Care TOS identifies patients as frail when they have multiple problems on various domains, which suits the need in primary care to identify patient groups that might benefit from more proactive and integrated care. With EASY-Care TOS, family physicians and other primary care professionals have an efficient, practical, and validated frailty identification method for primary care.
Consequently, we developed the EASY-Care Two-step Older persons Screening (EASY-Care TOS) with input of stakeholders [15]. The complete EASY-Care assessment is a tool for geriatric assessment in primary care and is extensively studied [16]. This tool considers frailty in the ‘‘broadest’’ sense of the concept as it comprises physical and psychosocial aspects of frailty and aspects of the care context of the patient. The EASY-Care TOS is derived from this EASY-Care assessment tool. Although, it is more efficient because it makes use of available prior knowledge of the FP. Furthermore, using the professionals’ appraisal as the frailty decision, instead of a cutoff score, it fits with everyday clinical reasoning of the FP [15]. A feasibility study confirmed that in the Dutch primary care context, with emphasis on longitudinal continuity and a strongly established doctorepatient relationship, only about 10% of the elderly population evaluated with EASY-Care TOS step 1 needed a home visit for the frailty appraisal. The home visits can be conducted by practice nurses. Dutch FPs highly appreciated the two-step approach of EASY-Care TOS and mentioned that it is time saving. Furthermore, the professionals considered their clinical judgment of patients’ frailty more important than a numerical score [17]. This report describes the construct validity of the EASYCare TOS by comparing it with the mostly used other frailty constructs [Fried Frailty Criteria (FFC) and Frailty Index (FI)] and other related constructs: multimorbidity, disability, cognition, mobility, psychosocial functioning, and quality of life. As we hypothesized that the EASYCare TOS identifies patients as frail when they have multiple problems on various domains, we expected that the overall judgment of frailty according to EASY-Care TOS
177
would correlate stronger with the FI than with the FFC as the latter only measures physical frailty, whereas the FI is a measure for frailty on multiple domains. Because the EASY-Care TOS includes multimorbidity and disability in the frailty judgment, we expected that there would be considerable overlap with these concepts.
2. Methods 2.1. Study population and data collection Six FP practices (with in total 15 FPs) in and around Nijmegen (The Netherlands) assessed their patients of 70 years and older with the EASY-Care TOS between February 2010 and August 2011 (n 5 1,159) and asked these patients to participate in the study. These practices were situated in urban (n 5 2), suburban (n 5 1), and countryside (n 5 3) areas. Patients who were too ill to be assessed (eg, patients in a palliative trajectory) were excluded. Patients were also excluded if they were under the treatment of a geriatrician or underwent a comprehensive geriatric assessment (CGA) in the past 3 months as the information of the geriatrician might influence the frailty judgment of the FP. Of this sample, 587 older patients gave informed consent and were included in the study. 2.2. Measurements 2.2.1. EASY-Care TOS FPs and primary care nurses received education on frailty and its identification with EASY-Care TOS. They were instructed on using the following conceptual definition of frailty: a frail older person has decreased reserve capacity because of multiple health, mental, or social problems, which make the person vulnerable for changes in the biopsychosocial context, especially when compensations are lacking [13,16,18e21]. In the first step of EASY-Care TOS, FPs made a frailty judgment of their older patients based on prior knowledge, for example, from the patients’ record. For this purpose, FPs used a 14-item checklist concerning physical and psychosocial functioning of the patient. At the end, the FP decided whether the patient was frail. This decision was not based on a standardized score but on clinical reasoning and tacit knowledge of the FP. The patients for whom the FP felt that he did not have enough information to make this judgment (‘‘unclear’’) were invited for the second step of EASY-Care TOS. This second step was an in-home assessment by a primary care nurse, who assessed functioning and well-being by applying the EASY-Care assessment instrument [22]. Afterward, the nurse discussed the information gathered from the assessment with the FP, and subsequently, they judged functioning of the patient (good, fair, and poor) on eight domains (diseases, medication, cognition, vision and hearing, functional status, mobility, mental well-being, and social context) and decided whether the
178
J.A.L. van Kempen et al. / Journal of Clinical Epidemiology 67 (2014) 176e183
Table 1. Characteristics of participants Characteristic Age, mean 6 SD*** Sex (women), n (%) Native country (The Netherlands), n (%) Educational level, n (%) Low Middle High Missing Comorbidity CIRS-G, mean 6 SD*** Polypharmacy Number of medications, mean 6 SD*** Disability KATZ-15, mean 6 SD*** Memory MMSE, mean 6 SD*** Mobility Walking speed (m/s), mean 6 SD*** Sensory problems Poor hearing, n (%)* Poor vision, n (%)*** Mental well-being GDS-15, mean 6 SD*** HADS-A, mean 6 SD*** Social context Loneliness, n (%)*** Sometimes Often Quality of life Grade of 0e10, mean 6 SD*** Care use Days of hospitalizations in the past year, mean 6 SD** Hr/wk home care, mean 6 SD*** Frailty measures Fried Frailty Criteria, mean 6 SD*** Rockwood Frailty Index, mean 6 SD***
Total (n [ 587)
EC-TOS not frail (n [ 356)
EC-TOS frail (n [ 231)
76.8 6 4.8 330 (56.2) 555 (94.5)
75.4 6 4.0 194 (54.5) 341 (95.8)
78.9 6 5.2 136 (58.9) 214 (92.6)
319 (54.3) 242 (41.2) 23 (3.9) 3 (0.5)
181 (50.8) 159 (44.7) 14 (3.9) 2 (0.6)
138 (59.7) 83 (35.9) 9 (3.9) 1 (0.4)
8.1 6 4.4
6.4 6 3.7
10.5 6 4.3
4.7 6 3.1
3.9 6 2.7
5.9 6 3.4
1.4 6 1.9
0.7 6 0.9
2.5 6 2.5
27.3 6 2.6
27.9 6 1.8
26.4 6 3.4
1.1 6 0.6
0.9 6 0.2
1.3 6 0.8
266 (45.3) 270 (46.0)
148 (41.6) 128 (36)
118 (51.1) 142 (61.5)
2.1 6 2.2 2.3 6 2.8
1.5 6 1.7 1.8 6 2.3
3.0 6 2.5 3.1 6 3.3
141 (24.0) 19 (3.2)
66 (18.5) 8 (2.2)
75 (32.5) 11 (4.8)
7.5 6 1.0
7.7 6 0.8
7.2 6 1.1
1.4 6 5.2
0.8 6 3.9
2.4 6 6.6
1.0 6 2.1
0.4 6 1.0
1.9 6 2.8
0.95 6 1.13 0.25 6 0.11
0.54 6 0.76 0.20 6 0.08
1.58 6 1.30 0.33 6 0.11
Abbreviations: EC-TOS, EASY-Care Two-step Older persons Screening; SD, standard deviation; CIRS-G, Cumulative Illness Rating Scale for Geriatrics; MMSE, Mini-Mental State Examination; GDS-15, Geriatric Depression Scale. P-value frail vs. not frail: *P ! 0.05; **P ! 0.01; ***P ! 0.001. Educational level: low 5 primary and lower secondary education, middle 5 upper secondary education, and high 5 tertiary education. CIRS-G: range 0e56, the higher the score, the more multimorbidity. MMSE: range 0e30, score of 24 is indicative for cognitive problems. KATZ-15: range 0e15, the higher the scores, the more (I)ADL disabilities. Walking speed: of 4-m walking test. GDS-15: range 0e15, score of 6 is indicative for depression. HADS-A: range 0e21, score of 7 is indicative for anxiety complaints.
patient was frail. For the complete EASY-Care TOS, see Appendix A at www.jclinepi.com.
2.2.2. Other measurements To study the construct validity of the EASY-Care TOS, additional data of the participants were collected by means of a CGA by a geriatrician and geriatric nurse, at the geriatric outpatient clinic of the University Medical Centre in Nijmegen (The Netherlands). The geriatrician and geriatric nurse were blinded for the results of the EASY-Care TOS. The maximum time between EASY-Care TOS and additional data collection was 4 weeks.
The FFC and FI were constructed from the CGA. Patients were frail according to the FFC when they had three or more of the following criteria: (1) unintentional weight loss (self-reported unintentional weight loss of 10 pounds/ year), (2) self-reported exhaustion, (3) weakness (measured with grip strength), (4) slow walking speed (measured with 4-m walking test), and (5) low physical activity (based on self-reported physical activity) [9]. We used methods and cutoff scores as mentioned in the article by Fried et al. [9]. We used a 45-item FI, of which the items are presented in Appendix B at www.jclinepi.com. Patients were considered frail according to the FI when they had a score of 0.25 or more [23].
J.A.L. van Kempen et al. / Journal of Clinical Epidemiology 67 (2014) 176e183
179
The other measurements that were included in the CGA were the Cumulative Illness Rating Scale for Geriatrics (CIRS-G) [24], Mini-Mental State Examination [25], KATZ-15 Activities of Daily Living (ADL) scale [26], Short Physical Performance Battery [27], Geriatric Depression Scale (GDS-15) [28], HADS-A anxiety scale [29], RAND-36 mental well-being scale, and RAND-36 social functioning scale [30]. Additionally, we assessed the number of medications, vision problems, hearing problems, loneliness, and social isolation by means of an interview and a physical examination. We defined multimorbidity as the presence of at least two domains with a score of 2 or more on the CIRS-G and disability as the presence of one or more ADL disabilities on the KATZ scale.
correlations were calculated to compare the strength of the association of EASY-Care TOS with these measures. For the interpretation of the strength of the relations, we used the following criterion: a correlation of 0.1e0.3 is a small, 0.3e0.5 medium, and 0.5 and higher a high correlation [31]. The construct validity of each of the eight EASY-Care TOS domains was studied by comparing the outcomes for the domains with corresponding measures with one-way analyses of variance and Pearson chi-square tests. For the interrater reliability, we determined observed agreement between the first and second judgment and Cohen’s kappa. We also compared the frailty judgments of the different FPs with a Pearson chi-square test.
2.3. Interrater reliability
3. Results
For determining the interrater reliability, we recruited a subgroup of 19 patients. These patients were assessed with EASY-Care TOS for a second time by an independent FP and nurse, within 4 weeks after the first assessment. The FP and nurse were blinded for the results of the first EASYCare TOS assessment.
3.1. Characteristics of the participants
2.4. Data analysis To study whether our study sample was representative for the general population, we compared key characteristics of included and nonincluded patients. To study construct validity, we compared the results of the EASY-Care TOS with the results of the measures for frailty and the other related constructs with an independent t-test and a Pearson chi-square test. Additionally,
The study sample was significantly younger (78.9 6 5.9 vs. 76.6 6 4.8 years) and less frail according to EASY-Care TOS (53.7% vs. 39.4% frail) than the nonincluded patient population. We found no differences regarding gender, quality of life, and health-care use. Among the study participants, frail participants scored significantly worse on most measures compared with those who were not frail according to EASY-Care TOS but not on educational level, problems with hearing, help in the case of an emergency, and days of hospitalization (Table 1). 3.2. Construct validity According to EASY-Care TOS, 231 (39.4%) persons were frail. According to the FFC, 58 (9.9%) and the FI
Fig. 1. Relation between frailty according to EASY-Care TOS, the Fried Frailty Criteria, and the Frailty Index, and the relation between frailty according to these three frailty instruments, multimorbidity, and disability (in Venn diagrams). EASY-Care TOS, EASY-Care Two-step Older persons Screening.
180
J.A.L. van Kempen et al. / Journal of Clinical Epidemiology 67 (2014) 176e183
258 (44.0%). A total of 396 (67.5%) patients had multimorbidity and 140 (23.9%) patients had ADL disability. Fig. 1 shows the interrelation of the three frailty constructs and their interrelation with multimorbidity and ADL disability. About 89.7% of patients who were frail according to FFC were also frail according to the FI and EASY-Care TOS. The patients who were frail according to FFC had multimorbidity, and in 62% of the cases, this was combined with the presence of ADL disability. We found overlap (n 5 170) between frailty according to EASY-Care TOS and the FI. However, 60 and 84 patients were frail according to EASY-Care TOS or FI only, respectively. Almost all patients who were judged as frail according to the FI or EASY-Care TOS had multimorbidity, 92% and 87%, respectively. Less than half of these patients had ADL disability, 40.7% and 39.4%, respectively. One hundred seventy patients had neither comorbidity nor ADL disability. Of these patients, 24 were frail according to EASY-Care TOS, 13 according to the FI, and none according to the FFC. The 24 patients who were frail according to EASY-Care TOS had problems on multiple domains other than morbidity or ADL disability. Two patients had problems on the psychosocial domain only. Table 2 shows the correlation of the frailty judgment of EASY-Care TOS with the results of the other frailty instruments (FI and FFC) and measures of other related constructs. The correlation between frailty according to EASY-Care TOS and FFC was 0.52, and 0.63 between EASY-Care TOS and FI. EASY-Care TOS correlated highly with multimorbidity (0.50), disability (0.53), and mobility (0.55) and moderately with polypharmacy (0.34), cognition (0.31), mental well-being (0.38), and self-perceived health (0.35). Table 3 shows the comparison of the judgment on the eight domains of EASY-Care TOS with the objective results of measures on corresponding constructs. The scores of the objective measures differed significantly between the different judgments (good, fair, and poor) on the corresponding EASY-Care TOS domains. 3.3. Interrater reliability We found 89% interrater agreement between the first and second frailty judgment, with a Cohen’s k of 0.63. Furthermore, we found no significant differences between the judgments of the different FPs.
4. Discussion As hypothesized, the EASY-Care TOS correlated well with constructs that measure physical and psychosocial frailty. Furthermore, EASY-Care TOS correlated to a higher extent with the FI, a count of deficits on various domains, than it does with the FFC, a measure for physical frailty. It identifies patients as frail when they have multiple
Table 2. Correlation among EASY-Care TOS, other constructs of frailty, and constructs that are possibly associated with frailty Construct Frailty Fried Frailty Criteria Rockwood Frailty Index Physical functioning Multimorbidity (CIRS-G) Self-perceived health Medication Polypharmacy (number of medications) Cognition MMSE (I)ADL KATZ-15 ADL (KATZ-15) IADL (KATZ-15) Mobility SPPB Mental well-being Depression GDS-15 Anxiety HADS-A RAND-36, mental well-being subscale Social context RAND-36, social functioning Quality of life Grade between 0 and 10 Care use Hospitalization (d/yr) Home care (hr/wk) Age Education
Correlation coefficient 0.52*** 0.63*** 0.50*** 0.35*** 0.34*** 0.31*** 0.53*** 0.40*** 0.51*** 0.55*** 0.38*** 0.25*** 0.30*** 0.28*** 0.27*** 0.17** 0.42*** 0.39*** 0.08*
Abbreviations: Easy-Care TOS, EASY-Care Two-step Older persons Screening; CIRS-G, Cumulative Illness Rating Scale for Geriatrics; MMSE, Mini-Mental State Examination; ADL, activities of daily living; SPPB, Short Physical Performance Battery; GDS-15, Geriatric Depression Scale. P-value frail vs. not frail: *P ! 0.05; **P ! 0.01; ***P ! 0.001.
problems on various domains, which suits the need in primary care to identify patient groups that might benefit from more proactive and integrated care. For all three frailty instruments, having physical problems was a prerequisite for being frail, and all three identified very few patients as frail based on psychosocial problems only. In this respect, EASY-Care TOS does not differ from the other two frailty instruments. However, the FFC differed in the frailty selection as they selected the patients who had the worst functional status. The differences between the populations selected by EASY-Care TOS and FI are less apparent and do not seem to be systematic. Accordingly, literature showed that different frailty instruments do select different but overlapping patient populations [32e34]. There is still no consensus about an optimal definition of frailty and which instrument to use. However, an aspect often overlooked in this ‘‘chase’’ for the ultimate definition of frailty is that different definitions and instruments may aim to identify frailty for very different goals, for example, some instruments are impractical in a clinical setting but are very useful for policy planning
J.A.L. van Kempen et al. / Journal of Clinical Epidemiology 67 (2014) 176e183
181
Table 3. Relation between outcomes on the eight domains of EASY-Care TOS and matching constructs Judgment on the EASY-Care TOS domain Construct Physical functioning CIRS-G, mean 6 SD*** Self-perceived health, n (%)*** Excellent Very good Good Reasonable Poor Medication Number, mean 6 SD*** Cognition MMSE, mean 6 SD*** Vision and hearing Poor vision, n (%)*** Poor hearing, n (%)*** (I)ADL KATZ-15, mean 6 SD*** Mobility SPPB, mean 6 SD*** Mental well-being GDS-15, mean 6 SD*** HADS-A, mean 6 SD*** RAND-36 mental well-being, mean 6 SD*** Social context RAND-36 social functioning, mean 6 SD*** Loneliness, n (%)*** Never Sometimes Often
Good 6.9 6 3.8
Fair 11.5 6 4.2 (0.0) (0.8) (33.1) (54.3) (11.8)
Poor 13.4 6 3.6
29 (6.5) 55 (12.4) 251 (56.4) 109 (24.5) 1 (0.2)
0 1 42 69 15
0 (0.0) 0 (0.0) 2 (13.3) 9 (60.0) 4 (26.7)
4.4 6 3.1
6.8 6 2.9
6.5 6 4.0
27.9 6 1.6
25.0 6 3.5
19.8 6 5.3
150 (37.6) 132 (33.1)
106 (62.0) 119 (69.6)
14 (82.4) 15 (88.2)
1.0 6 1.4
3.2 6 2.6
6.0 6 0.7
10.2 6 1.7
7.6 6 2.6
4.5 6 3.1
1.5 6 1.6 1.6 6 2.0 18.2 6 11.5
3.5 6 2.5 4.1 6 3.5 31.4 6 15.4
6.3 6 3.4 7.4 6 3.9 55.5 6 19.3
1.3 6 0.8
1.7 6 1.1
2.4 6 1.4
370 (78.2) 95 (20.1) 8 (1.7)
55 (50.9) 42 (38.9) 11 (10.2)
2 (33.3) 4 (66.7) 0 (0.0)
Abbreviations: Easy-Care TOS, EASY-Care Two-step Older persons Screening; CIRS-G, Cumulative Illness Rating Scale for Geriatrics; SD, standard deviation; MMSE, Mini-Mental State Examination; ADL, activities of daily living; SPPB, Short Physical Performance Battery; GDS-15, Geriatric Depression Scale. P-value frail vs. not frail: ***P ! 0.001. CIRS-G: range 0e56, the higher the score, the more multimorbidity. MMSE: range 0e30, score of 24 is indicative for cognitive problems. KATZ-15: range 0e15, the higher the scores, the more (I)ADL disabilities. SPPB: range 0e12, score of !9 is indicative for a high risk on loss of mobility. GDS-15: range 0e15, score of 6 is indicative for depression. HADS-A: range 0e21, score of 7 is indicative for anxiety complaints. RAND-36 mental well-being: range 5e30, the higher the score, the worse mental well-being. RAND-36 social functioning: range 1e6, the higher the score, the worse social functioning.
[14]. This means that there may be not one optimal instrument for the identification of frailty, but the optimal instrument is the one that best suits the needs of those using it [14,35]. The correlation between the EASY-Care TOS and the FI is high, and both intend to identify patients with multiple problems on various domains. Drubbel et al. [12] recently showed that an FI based on routine health-care data of FPs is predictive for the risk on adverse health outcomes. This means that the FI probably can be a useful frailty identification instrument in primary care. Despite the high correlation and the fact that the FI and EASY-Care TOS aim to identify patients as frail when they have multiple problems, there are some differences between the two instruments.
The FI uses a count of deficits, and the EASY-Care TOS uses the professional’s appraisal as frailty judgment. One of the main advantages of the EASY-Care TOS over the FI for use in primary care with the purpose of identifying the target population for integrated care is that the FP directly has an overview of the nature of the problems of the patient. A strength of this study is that it is unique in comparing a frailty instrument using the professional’s appraisal as frailty judgment with more objective measurement instruments. Furthermore, it is the first time that a frailty identification instrument is validated that was specifically developed for use in primary care with inputs from stakeholders.
182
J.A.L. van Kempen et al. / Journal of Clinical Epidemiology 67 (2014) 176e183
This study also has some limitations. First, our study population is not a completely accurate representation of the general population. Compared with the screened population, the participants were younger and were less frail according to the EASY-Care TOS. This probably is because participants had to come to the hospital for the geriatric assessment, which could be too aggravating for the severely frail patients. This will give an underestimation of frailty and related problems in our study population. However, the prevalence of frailty in our study is comparable with that in other studies. Collard et al. [36] found a weighted average prevalence of frailty according to the FFC of 9.9%. The Health and Retirement study using the FI found a prevalence of frailty of 32%. In this study, however, the included patients were aged 51 years and older [32]. Second, EASY-Care TOS identifies more than one-third of patients as frail, which is a large selection. This would mean that all these patients are in need of integrated care. We can imagine that the selected patients are not all in need of integrated care but will all benefit from proactive care. Further study is needed to find out whether these frail patients really are in need of integrated care or that we need a stricter selection. Third, we did not use validated measurement instruments for some domains, for example, not for medication and vision and hearing. Therefore, we used information from the interviews and physical examination of the geriatric assessment. Finally, only a small study population was available for determining the interrater reliability, and most patients in this subpopulation appeared to be not frail. Furthermore, the time between the first and second interval had a maximum of 4 weeks. Preferably, the second assessment was conducted at the same moment. This was because of organizational reasons, which made it not possible. However, we asked the FPs whether things changed in the health status of the patient in the weeks between the first and second assessment. We only included patients with no change in health status. A former study already established the feasibility of the EASY-Care TOS [17]. Additionally, this study shows the construct validity and interrater reliability of EASY-Care TOS. This instrument enables the FP to identify patients with multiple problems on physical, mental, and social domains. A planned prospective follow-up study will have to evaluate whether EASY-Care TOS indeed identifies those patients as frail who will benefit from integrated care. In conclusion, we think that EASY-Care TOS suits the need of FPs for an efficient, practical, and validated frailty identification method for primary care.
Acknowledgments The authors thank Hanny Hordijk, Leny Theunisse, Marlies Hoogsteen, Marieke Perry, Sarah Robben, and
Brechtje Lubberink for their help with the data collection and George Borm for his help with statistics. Appendix Supplementary data Supplementary data related to this article can be found at http://dx.doi.org/10.1016/j.jclinepi.2013.08.008. References [1] Iliffe S. Family medicine in an aging society. Can Fam Physician 2009;55(5):463e7. [2] Boeckxstaens P, De Graaf P. Primary care and care for older persons: position paper of the European Forum for Primary Care. Qual Prim Care 2011;19(6):369e89. [3] Wilhelmson K, Duner A, Eklund K, Gosman-Hedstrom G, Blomberg S, Hasson H, et al. Design of a randomized controlled study of a multi-professional and multidimensional intervention targeting frail elderly people. BMC Geriatr 2011;11:24. [4] Eklund K, Wilhelmson K. Outcomes of coordinated and integrated interventions targeting frail elderly people: a systematic review of randomised controlled trials. Health Soc Care Community 2009; 17(5):447e58. [5] De Lepeleire J, Iliffe S, Mann E, Degryse JM. Frailty: an emerging concept for general practice. Br J Gen Pract 2009;59(562):e177e82. [6] Lacas A, Rockwood K. Frailty in primary care: a review of its conceptualization and implications for practice. BMC Med 2012;10:4. [7] Pialoux T, Goyard J, Lesourd B. Screening tools for frailty in primary health care: a systematic review. Geriatr Gerontol Int 2012;12(2): 189e97. [8] Rockwood K, Mitnitski A. Frailty in relation to the accumulation of deficits. J Gerontol A Biol Sci Med Sci 2007;62(7):722e7. [9] 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):M146e56. [10] Mitnitski A, Song X, Skoog I, Broe GA, Cox JL, Grunfeld E, et al. Relative fitness and frailty of elderly men and women in developed countries and their relationship with mortality. J Am Geriatr Soc 2005;53(12):2184e9. [11] 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 Geriatr 2010;10:57. [12] Drubbel I, de Wit NJ, Bleijenberg N, Eijkemans RJ, Schuurmans MJ, Numans ME. Prediction of adverse health outcomes in older people using a frailty index based on routine primary care data. J Gerontol A Biol Sci Med Sci 2013;68:301e8. [13] Gobbens RJ, Luijkx KG, Wijnen-Sponselee MT, Schols JM. Toward a conceptual definition of frail community dwelling older people. Nurs Outlook 2010;58(2):76e86. [14] Sternberg SA, Wershof Schwartz A, Karunananthan S, Bergman H, Mark Clarfield A. The identification of frailty: a systematic literature review. J Am Geriatr Soc 2011;59(11):2129e38. [15] van Kempen JA, Schers HJ, Jacobs A, Zuidema SU, Ruikes F, Robben SH, et al. Development of an instrument for the identification of frail elderly as a target population for integrated care. Br J Gen Pract 2013;63(608):e225e31. [16] Nourhashemi F, Andrieu S, Gillette-Guyonnet S, Vellas B, Albarede JL, Grandjean H. Instrumental activities of daily living as a potential marker of frailty: a study of 7364 community-dwelling elderly women (the EPIDOS study). J Gerontol A Biol Sci Med Sci 2001;56(7):M448e53. [17] Keiren SM, van Kempen JA, Schers HJ, Olde Rikkert MG, Perry M, Melis RJ. Feasibility evaluation of a stepped procedure to identify
J.A.L. van Kempen et al. / Journal of Clinical Epidemiology 67 (2014) 176e183
[18]
[19]
[20] [21]
[22] [23]
[24]
[25]
[26]
community-dwelling frail older people in general practice. A mixed methods study. Eur J Gen Pract 2013. In press. Strawbridge WJ, Shema SJ, Balfour JL, Higby HR, Kaplan GA. Antecedents of frailty over three decades in an older cohort. J Gerontol B Psychol Sci Soc Sci 1998;53(1):S9e16. Bandeen-Roche K, Xue QL, Ferrucci L, Walston J, Guralnik JM, Chaves P, et al. Phenotype of frailty: characterization in the women’s health and aging studies. J Gerontol A Biol Sci Med Sci 2006;61(3): 262e6. Rockwood K, Hogan DB, MacKnight C. Conceptualisation and measurement of frailty in elderly people. Drugs Aging 2000;17(4):295e302. Schuurmans H, Steverink N, Lindenberg S, Frieswijk N, Slaets JP. Old or frail: what tells us more? J Gerontol A Biol Sci Med Sci 2004;59(9):M962e5. Richardson J. The Easy-Care assessment system and its appropriateness for older people. Nurs Older People 2001;13(7):17e9. Rockwood K, Andrew M, Mitnitski A. A comparison of two approaches to measuring frailty in elderly people. J Gerontol A Biol Sci Med Sci 2007;62(7):738e43. Salvi F, Miller MD, Grilli A, Giorgi R, Towers AL, Morichi V, et al. A manual of guidelines to score the modified cumulative illness rating scale and its validation in acute hospitalized elderly patients. J Am Geriatr Soc 2008;56(10):1926e31. Folstein MF, Folstein SE, McHugh PR. ‘‘Mini-mental state’’. A practical method for grading the cognitive state of patients for the clinician. J Psychiatr Res 1975;12(3):189e98. de Rooij SE, Govers AC, Korevaar JC, Giesbers AW, Levi M, de Jonge E. Cognitive, functional, and quality-of-life outcomes of patients aged 80 and older who survived at least 1 year after planned
[27]
[28]
[29] [30] [31] [32]
[33]
[34]
[35] [36]
183
or unplanned surgery or medical intensive care treatment. J Am Geriatr Soc 2008;56(5):816e22. Guralnik JM, Simonsick EM, Ferrucci L, Glynn RJ, Berkman LF, Blazer DG, et al. A short physical performance battery assessing lower extremity function: association with self-reported disability and prediction of mortality and nursing home admission. J Gerontol 1994;49(2):M85e94. Yesavage JA, Brink TL, Rose TL, Lum O, Huang V, Adey M, et al. Development and validation of a geriatric depression screening scale: a preliminary report. J Psychiatr Res 1982;17(1):37e49. Zigmond AS, Snaith RP. The hospital anxiety and depression scale. Acta Psychiatr Scand 1983;67(6):361e70. Hays RD, Sherbourne CD, Mazel RM. The RAND 36-item health survey 1.0. Health Econ 1993;2(3):217e27. Cohen J. A power primer. Psychol Bull 1992;112(1):155e9. Cigolle CT, Ofstedal MB, Tian Z, Blaum CS. Comparing models of frailty: the Health and Retirement Study. J Am Geriatr Soc 2009; 57(5):830e9. van Iersel MB, Rikkert MG. Frailty criteria give heterogeneous results when applied in clinical practice. J Am Geriatr Soc 2006; 54(4):728e9. Sourial N, Wolfson C, Bergman H, Zhu B, Karunananthan S, Quail J, et al. A correspondence analysis revealed frailty deficits aggregate and are multidimensional. J Clin Epidemiol 2010;63:647e54. Martin FC, Brighton P. Frailty: different tools for different purposes? Age Ageing 2008;37:129e31. Collard RM, Boter H, Schoevers RA, Oude Voshaar RC. Prevalence of frailty in community-dwelling older persons: a systematic review. J Am Geriatr Soc 2012;60:1487e92.
