Validation of the Construct of Intrinsic Capacity in a Longitudinal Chinese Cohort

Abstract

Objectives

We examined the structure and predictive ability of intrinsic capacity in a cohort of Chinese older adults.

Methods

We used data from the MrOS and MsOS (Hong Kong) study, which was designed to examine the determinants of osteoporotic fractures and health in older Chinese adults. We analysed baseline and the 7-year follow-up data using exploratory factor analysis, confirmatory factor analysis (CFA), and mediation analysis.

Results

The study consisted of 3736 participants at baseline (mean 72.2 years), with 1475 in the 7-year follow-up. Bi-factor CFA revealed five sub-factors labelled as 'cognitive’, 'locomotor’, 'vitality’, 'sensory’, and 'psychological’ and one general factor labelled as 'intrinsic capacity’. The model fits the data well, with Root Mean Square Error of Approximation (RMSEA)=0.055 (90% CI=0.053–0.058) for the 5-factor model and RMSEA=0.031 (90% CI=0.028–0.035) for the bi-factor model. Significantly lower intrinsic capacity scores were found in older age groups, women, as well as those who had lower levels of education, lower subjective social status, reported more chronic diseases, or a higher number of IADL limitations (All p<0.0001). Intrinsic capacity had a direct effect in predicting incident IADL limitations at the 7-year follow-up (β=−0.21, p<0.001). The effect was larger than the direct effect of the number of chronic diseases on incident IADL limitations (β=0.05, not significant).

Conclusions

This study supports the construct and predictive validity of the proposed capacity domains of intrinsic capacity. The findings could inform the development of an intrinsic capacity score that would facilitate implementation of the concept of intrinsic capacity in clinical practice.

Introduction

Healthy ageing, defined by the World Health Organization (WHO) as ‘the process of developing and maintaining the functional ability that enables wellbeing in older age’, is fundamental to the science and practice across multiple disciplines. Rather than considering healthy ageing from the perspective of the presence or absence of disease, it is oriented around building and maintaining the ability of older adults to be, and to do the things they have reason to value (1, 2). At the core of functional ability is the intrinsic capacity of the individual, the environments in which they live and the interaction between the individual and these environments, where intrinsic capacity is defined as ‘the composite of all the physical and mental (including psychosocial) capacities that an individual can draw on at any point in their life’ (1). The WHO proposed a construct consisting of five domains: cognitive, locomotor, vitality, sensory (vision and hearing), and psychological (3, 4). These domains represent measurable indicators of ageing. Therefore, intrinsic capacity would be an outcome measure for healthy ageing that could be used to evaluate integrated care models for older adults living in the community recommended in the WHO guidelines on Integrated Care for Older People (ICOPE) (5).

However, there have been few studies validating this concept. Beard et al. explicitly described the components of each domain and examined the structure and predictive validity of the concept of intrinsic capacity (6). Using commonly measured biomarkers and self-reported measures, it was found that the intrinsic capacity score provided useful predictive information on an individual’s subsequent functioning as measured by activities daily living (ADLs) and instrumental activities daily living (IADLs). The authors also suggested that the five domains appear to operate at different levels. While the cognitive, locomotor, sensory, and psychological domains can be thought of as overt expressions of capacity, the vitality domain may represent the elements of the biological systems (e.g., nutritional, immune, and hormonal status) that underline the overt manifestations of capacity. Other studies have calculated an intrinsic capacity score as an outcome measure and associated it with biomarkers including C-reactive protein, homocysteine (7), and allostatic load (8). These studies provide fundamental information on the conceptual frame for the construct of intrinsic capacity. Nevertheless, further validation of the construct of intrinsic capacity with data from different populations, settings, and periods would be needed to fully appreciate the generalisability of the construct of intrinsic capacity. Such studies may lead to development of an intrinsic capacity score with reference/cut-off values that would facilitate implementation of the concept of intrinsic capacity in clinical practice.

Using a longitudinal Chinese cohort of 2000 men and 2000 women aged 65 years and older, we aimed to validate previous work by Beard et al. (6) by assessing whether a range of objective and self-reported measures might provide a useful estimate of intrinsic capacity and whether the intrinsic capacity score predicted incident functional limitations as measured by IADLs.

Methods

Study design and participants

The MrOS and MsOS (Hong Kong) study, a cohort study on osteoporosis and general health in Hong Kong, began in 2001–2003. It involves a community sample of men and women aged 65 years and older. In the study, participants were volunteers, and the inclusion criteria were: 1) being ambulatory, 2) absence of bilateral hip replacements, and 3) competent to give informed consent. Participants completed an interviewer-administered questionnaire and a battery of clinical examinations during the baseline visit, and were invited for follow-up assessments at year 2, year 4, year 7 and year 14. Further details about the study were published elsewhere (9). In this study, data collected at baseline and year 7 was used (Figure 1). The study was approved by the Joint Chinese University of Hong Kong and New Territories East Cluster Clinical Research Ethics Committee in Hong Kong, which required informed consent to be obtained.

Figure 1.

Participant flow chart

Measures of the concept of intrinsic capacity

Cognitive domain

• •Cognitive function was assessed with the Mini-Mental State Examination (MMSE), which composes of 30 items (10, 11).

Locomotor domain

• •Walking speed was measured using the best time in seconds to complete a walk along a straight line six meters long in distance. A warm up period of less than five minutes was followed by two walks, and the best time was recorded.
• •Chair stands were used to evaluate muscle endurance in the lower extremities. Participants were asked to rise from a chair for a total of five times, as quickly as they could, with arms across their chest. The amount of time required to complete all five repetitions was recorded.
• •Dynamic balance was assessed using the best time in seconds to complete a narrow walking path (20 cm) over six meters. A warm up period of less than five minutes was followed by two walks, and the best time was recorded.

Vitality domain

• •Grip strength was measured using a dynamometer (JAMAR Hand Dynamometer 5030JO; Sammons Preston, Inc, Bolingbrook, IL). Two readings were taken from each side and the maximum value of right and left were used for analysis.
• •Adiposity to muscle ratio, the ratio of body fat to appendicular skeletal muscle mass (ASM), was measured by dual-engery X-ray absorptiometry (DXA) using Hologic QDR 4500 densitometers (Hologic Delphi, auto whole body version 12.4; Hologic, Bedford, MA). ASM was calculated as the sum of appendicular lean mass minus bone mineral content of both arms and legs.

Sensory domain

• •Binocular visual acuity was assessed using a Snellen 'Tumbling E’ chart (Clement Clarke, London, UK) at six meters. If the participants could not clearly see even the largest, they would move to three meters. Participants were asked to recognize the direction the “legs” of the E were facing, from the biggest to the smallest line that they could correctly recognize.
• •Stereopsis was measured by the Frisby Stereo test (Clement Clarke International, London, UK). Participants were tested at 40 cm away from the plate; if they failed to recognize, they would then view them at 30 cm. This test consists of three transparent plates (1.5, 3, and 6-mm thick) marked with four different pattern squares.

Psychological domain

• •Depressive symptoms were assessed with the 15-item Geriatric Depression Scale (GDS-15) (12, 13, 14).

Covariates

• •Covariables included chronological age, sex, education (primary or below, secondary, tertiary or above), subjective social status, and number of self-reported physician-diagnosed chronic diseases including diabetes, high thyroid problems, low thyroid problems, osteoporosis, stroke, Parkinson’s disease, hypertension, myocardial infarction, angina, congestive heart failure, chronic obstructive pulmonary disease (COPD), prostatitis, glaucoma, cataracts, gastrectomy, arthritis, kidney stone, and cancer.
• •Subjective social status was assessed using a 10-rung self-anchoring scale, which has been associated with physiological and psychological functioning (15). Possible scores range from 1–10, with higher scores indicating higher perceived social status. In this study, the scores were arbitrary grouped into 'low’ (scores 1–3), 'middle’ (scores 4–5), and 'high’ (scores 6–10).

Outcome measures

Functional limitations were measured with five Instrumental Activities of Daily Living (IADL) items including ability of 'walking 2–3 blocks’, 'climbing 10 steps’, 'preparing own meals’, 'doing heavy housework’, and 'doing own shopping’. Participants were asked to score these items on a 4-point scale (0=no difficulty, 1=some difficulties, 2=much difficulty, and 3=cannot do). Two scores were calculated: 1) the number of IADL limitations, with the scores ranging from 0 (no difficulty on any items) to 5 (difficulty on all five items, i.e., a score of 1 is recorded regardless of whether the participants rated the items at difficulty levels 1, 2 or 3), and 2) the degree of difficulty of the five items, with the scores ranging from 0–15, with higher scores reflecting a higher level of IADL difficulty. Incident IADL limitations were defined as an increase in the degree of difficulty of the five items at the 7-year follow-up.

Statistical analysis

Structural equation models (SEMs) including (a) traditional exploratory factor analysis (EFA), (b) confirmatory factor analysis (CFA) and (c) mediation analysis was performed.

Traditional exploratory factor analysis was performed to identify the sub-factors of the intrinsic capacity using maximum likelihood. Four items with highest communalities from MMSE (including 'orientation to date’, 'orientation to address’, 'registration of three objects’, and 'attending and calculation’ and four items with highest communalities from GDS-15 (including 'not feeling life empty’, 'not often getting bored’, 'not often feeling helpless’, and 'not feeling worthless’) were selected for subsequent analyses. Eigen value and scree plot were used to identify number of sub-factors.

Confirmatory factor analysis was performed according to the sub-factors identified from EFA. Models of one factor, second factor, correlated 5-factor, and Bi-factor (one general factor and five sub-factors) were conducted. Goodness of fit statistics including Root Mean Square Error of Approximation (RMSEA), Comparative Fit Index (CFI), Tucker-Lewis Index (TLI) were applied. A RMSEA of less than 0.6 and CFI and TLI values of greater than 0.95 indicate a very good fit.

Summary scores for general factor and specific sub-factors were generated from CFA. Individual items were standardized, with mean equal to 0 and variance equal to 1. The scores of sub-factors represented the contribution to specific domains. Construct validity were tested with different summary scores by analysis of variance (ANOVA).

Mediation analysis was adopted to access the predictive validity of the intrinsic capacity score on incident IADL limitations at the 7-year follow-up. Path analysis was used to demonstrate the relationship among incident IADL limitations, intrinsic capacity, multi-morbidity and participants’ characteristics including age, sex, education and subjective social status.

All statistical analyses were performed using the statistical package SAS v 9.4 (SAS Institute, Inc, Cary, NC). All statistical tests were two sided. A P value of less than 0.05 was considered statistically significant.

Results

The original sample at baseline comprised of 4000 participants, 3736 of whom had complete information on the variables for constructing the concept of intrinsic capacity and were included in the baseline study. The 3736 participants were 65 to 98 years of age (mean 72.2 years), 49.7% were female, 28.8% had secondary education or above, and 23.6% had IADL limitations. The baseline characteristics of the 3736 participants are presented in Supplementary Table 1. Compared with participants included in the baseline analysis, participants excluded were older (p<0.0001), reported more chronic diseases (p<0.0001), or a higher number of IADL limitations (p<0.0001) (Supplementary Table 2).

EFA, CFA (5-factor and bi-factor) and model fit

In the EFA, a 5-factor model was suggested, with five factors having Eigen values greater than one (i.e., 3.3, 1.9, 1.5, 1.3, and 1.1). These five factors explained 43.0% of total variance (data not shown). To confirm the 5-factor model, CFA was performed. A 5-factor model was found, with standardized coefficients ranging from 0.368 to 1.000 (All p-values <0.001). In addition, the bi-factor CFA revealed five sub-factors and one general factor, with standardized coefficients ranging from 0.323 to 0.935 (All p-values <0.001) for the five sub-factors and from 0.079 to 0.652 (All p-values <0.001) for the general factor. The five sub-factors were labelled as cognitive, locomotor, vitality, sensory, and psychological; and the general factor was labelled as intrinsic capacity (Table 1, Figure 2). The model fits the data well, with χ2=1123.3 (df=90), RMSEA=0.055 (90% CI=0.053–0.058), CFI=0.96, and TLI=0.95 for the 5-factor model and χ2=347.4 (df=75), RMSEA=0.031 (90% CI=0.028–0.035), CFI=0.98, and TLI=0.97 for the bi-factor model. The model fit information is presented in Supplementary Table 3.

Table 1.

Confirmatory factor analysis of 5-factor and bi-factor model

	5-factor model		Bi-factor model
	5 factors	Standardized Coeff.	5 factors	Standardized Coeff.	general factor	Standardized Coeff.
Orientation to date	Cognitive	0.573**	Cognitive	0.537**	IC	0.366**
Orientation to address	Cognitive	0.536**	Cognitive	0.484**	IC	0.402**
Registration of 3 objects	Cognitive	0.368**	Cognitive	0.327**	IC	0.381**
Attention and Calculation	Cognitive	0.377**	Cognitive	0.332**	IC	0.283**
5 Chair stand	Locomotor	0.421**	Locomotor	0.406**	IC	0.259**
Walking speed	Locomotor	0.879**	Locomotor	0.752**	IC	0.518**
Narrow walking speed	Locomotor	0.702**	Locomotor	0.576**	IC	0.533**
Grip strength	Vitality	1.000**	Vitality	0.598**	IC	0.652**
Body fat/ASM	Vitality	0.446**	Vitality	0.500**	IC	0.408**
Visual disparity	Sensory	0.371**	Sensory	0.323**	IC	0.122**
Visual acuity	Sensory	0.848**	Sensory	0.935**	IC	0.353**
Not feeling life empty	Psychological	0.737**	Psychological	0.739**	IC	0.079**
Not often getting bored	Psychological	0.710**	Psychological	0.702**	IC	0.144**
Not often feeling helpless	Psychological	0.472**	Psychological	0.459**	IC	0.148**
Not feeling worthless	Psychological	0.398**	Psychological	0.385**	IC	0.252**

IC: Intrinsic capacity; **p<0.001

Figure 2.

Bi-factor confirmatory factor analysis model of intrinsic capacity

Construct validity

To assess the construct validity of the concept of intrinsic capacity, the mean scores of intrinsic capacity and each sub-factor were compared between known-groups that are expected to differ due to known characteristics (i.e., age, sex, educational level, subjective social status, number of chronic diseases, and number of IADL limitations). Significantly lower intrinsic capacity scores were found in older age groups, where the mean intrinsic capacity scores in the 65–69 years and ≥75 year age groups were 0.984 ± 2.301 and −1.008 ± 2.510, respectively (p<0.0001). Women had a lower intrinsic capacity score compared to man (p<0.0001). Lower intrinsic capacity scores were also found in participants who had lower levels of education (p<0.0001), lower subjective social status (p<0.001), reported more chronic diseases (p<0.0001), or had a higher number of IADL limitations (p<0.0001) (Table 2).

Table 2.

Construct validity of intrinsic capacity and sub-domain score

Demographic and health characteristics	Mean (SD)
	Intrinsic capacity	Cognitive	Locomotor	Vitality	Sensory	Psychological
Age, years
65–69	0.984 (2.301)	0.265 (0.977)	0.556 (1.321)	0.175 (1.007)	0.327 (0.886)	0.055 (1.666)
70–74	0.309 (2.232)	0.115 (1.042)	0.108 (1.312)	0.046 (0.952)	0.124 (0.962)	0.100 (1.585)
75 or above	−1.008 (2.510)	−0.306 (1.378)	−0.527 (1.413)	−0.167 (0.911)	−0.328 (1.148)	−0.070 (1.738)
p-value	<0.0001	<0.0001	<0.0001	<0.0001	<0.0001	0.0345
Sex
Male	1.610 (1.908)	0.450 (0.792)	0.426 (1.373)	0.818 (0.572)	0.172 (0.991)	0.116 (1.607)
Female	−1.344 (2.073)	−0.377 (1.313)	−0.289 (1.367)	−0.777 (0.529)	−0.060 (1.061)	−0.051 (1.712)
p-value	<0.0001	<0.0001	<0.0001	<0.0001	<0.0001	0.0021
Education
Primary or below	−0.450 (2.411)	−0.193 (1.237)	−0.155 (1.356)	−0.116 (0.947)	-0.045 (1.060)	−0.072 (1.712)
Secondary	1.410 (2.019)	0.573 (0.681)	0.517 (1.397)	0.357 (0.942)	0.274 (0.926)	0.182 (1.573)
Tertiary or above	1.977 (1.891)	0.688 (0.562)	0.839 (1.398)	0.414 (0.919)	0.372 (0.892)	0.496 (1.328)
p-value	<0.0001	<0.0001	<0.0001	<0.0001	<0.0001	<0.0001
Social economic status ladder
1–4	0.092 (2.417)	0.099 (1.086)	−0.040 (1.405)	0.113 (0.952)	0.045 (1.019)	−0.278 (1.819)
5–6	0.425 (2.372)	0.139 (1.055)	0.190 (1.379)	0.043 (0.977)	0.109 (0.992)	0.249 (1.503)
7–10	0.424 (2.381)	0.016 (1.167)	0.345 (1.371)	−0.027 (0.992)	0.098 (1.097)	0.471 (1.299)
p-value	0.0002	0.0949	<0.0001	0.014	0.2039	<0.0001
Number of chronic diseases
0	0.623 (2.495)	0.047 (1.129)	0.357 (1.411)	0.162 (0.996)	0.249 (0.892)	0.187 (1.570)
1–2	0.227 (2.415)	0.050 (1.125)	0.136 (1.390)	0.026 (0.962)	0.058 (1.025)	0.137 (1.569)
3 or more	−0.288 (2.518)	0.017 (1.233)	−0.213 (1.416)	−0.054 (0.958)	−0.057 (1.105)	−0.243 (1.834)
p-value	<0.0001	0.7482	<0.0001	<0.0001	<0.0001	<0.0001
Number of IADL limitations
0	0.623 (2.315)	0.162 (1.064)	0.308 (1.331)	0.166 (0.953)	0.125 (1.003)	0.155 (1.569)
1–2	−1.169 (2.254)	−0.300 (1.312)	−0.526 (1.308)	−0.401 (0.890)	−0.132 (1.062)	−0.237 (1.811)
3–5	−3.136 (2.344)	−0.795 (1.516)	−1.873 (1.545)	−0.609 (0.764)	−0.407 (1.299)	−1.161 (2.116)
p-value	<0.0001	<0.0001	<0.0001	<0.0001	<0.0001	<0.0001

With regards to the sub-factor scores, increasing age was negatively associated with all five sub-factors (All p<0.05). Female sex (All p<0.01) and lower education (All p<0.0001) were associated with lower scores of each sub-factor. Lower subjective social status was associated with lower locomotor and psychological scores (All p<0.0001) and with higher vitality score (p<0.05). Participants who had reported more chronic diseases were associated with lower locomotor, vitality, sensory, and psychological scores (All p<0.0001). A higher number of IADL limitations was associated with lower scores of each sub-factor (All p<0.0001) (Table 2).

Pathways to functional limitations

Results of the path analysis revealed that intrinsic capacity had a direct effect in predicting incident IADL limitations at the 7-year follow-up (β=−0.21, p<0.001). The effect was larger than the direct effect of the number of chronic diseases on incident IADL limitations (β=0.05, not significant). With regards to the role of socio-demographic variables, increasing age and lower education had significant indirect effects through either the number of chronic diseases (age: β=0.15, p<0.001; education: −0.09, p<0.001) or intrinsic capacity (age: β=−0.22, p<0.001; education: β=−0.21, p<0.001) on the risk of developing IADL limitations at the 7-year follow-up. Female sex also had a direct effect on intrinsic capacity (β=−0.58, p<0.001), although its effect on the number of chronic diseases was not significant. Higher subjective social status had a direct effect on the number of chronic diseases (β=−0.05, p<0.05) and intrinsic capacity (β=0.05, p<0.05). The model explained 7.4% of the variance in incident IADL limitations (Figure 3).

Figure 3.

Path analysis of incident IADL at the 7-year-follow-up, standardized coefficient is presented in path

*p<0.05, **p<0.001

Discussion

This study demonstrates that the concept of intrinsic capacity introduced by the WHO could be defined as a composite of multiple domains of capacity. The composite intrinsic capacity score could also predict IADL limitations over seven years, even after adjusting for chronological age, sex, educational level, subjective social status, and the number of chronic conditions.

These findings support the findings from Beard et al. that has examined the structure and predictive value of intrinsic capacity among older adults aged 60 or above.(6) As Beard et al. has shown, the structure of intrinsic capacity comprised of the five domains including cognitive, locomotor, vitality, sensory, and psychological. These domains could be aggregated into a composite measure of healthy ageing that may be relevant to identify deviations from normal ageing — a similar approach to the way child development charts guide paediatric practice. In the path analysis examining incident IADL limitations, intrinsic capacity had a direct relationship with the outcome. This pattern points to the key role of intrinsic capacity and stresses the importance of facilitating functional capacity even in the presence of chronic diseases, as outlined in the 2015 WHO World Report on Ageing and Health (1). In this context, this will require changes to health care model, which is currently often disease-oriented and inadequate to meet the needs of the increasing number of older adults.

The findings from the path analysis also suggest that women were more susceptible to less than optimum intrinsic capacity than men. The sex differences may stem from differences in comorbidities/chronic conditions and the complex interaction of pathophysiological, behavioural, and environmental factors. For example, women tend to experience more chronic conditions that are relatively less lethal and do incur a burden of disability (e.g., arthritis, depressive symptoms, and dementia) (16, 17). Sex differences in inflammation as indicated by elevated C-reactive protein and fibrinogen (18) and physiological dysregulation as indicated by elevated allostatic load in response to environmental stresses (19, 20) may also contribute to the sex differences in intrinsic capacity, where inflammatory markers and allostatic load have been considered as a biological substrate to intrinsic capacity (7, 8). In addition, the findings found a social gradient in intrinsic capacity, where a lower educational level or a lower subjective social status was associated with lower intrinsic capacity scores. The sex differences and the social gradient in intrinsic capacity observed in this study further support the validity of the intrinsic capacity score. The finding also highlights the association of social/ socioeconomic environment with intrinsic capacity, which has major implication for policy/strategies to tackle inequalities in healthy ageing. Nevertheless, a lower subjective status was associated with higher vitality scores. Analysis of the data at baseline revealed that participants with lower subjective social status were more physically active (with a significant higher Physical Activity Scale for the Elderly (PASE) score, data not shown) leading to a higher grip strength and lower body fat to ASM ratio, which are sub-factors of the vitality domain.

Although this study adopted a very similar approach as with Beard et al.’s study (6) in order to reconfirm the conceptual frame for the construct of intrinsic capacity in a Chinese population, there are some methodological differences in regards to the selection of sub-factors (i.e., components of each domain) between the two studies that should be noted, particularly the vitality domain. This study used grip strength and body fat to ASM, an adiposity to muscle ratio while in Beard et al.’s study, grip strength as well as cellular level characteristics including DHEA, IGF-1, haemoglobin and FEV were used. Beard et al. has suggested that the five intrinsic capacity domains operate at different levels and that the vitality domain describes variance in the biological systems, in particular in the nutritional, immune, and hormonal dimensions, that underlie the overt manifestation of capacity of other domains. In this regards, adiposity to muscle ratio was considered as a proxy of nutritional and/or hormonal status that may underlie the manifestation of cognitive and locomotor decline. This proxy has taken adiposity into account while measuring muscle mass as examining muscle mass alone without consideration of the associated adverse effect of increasing fat mass may lead to erratic observation, particularly in overweight individuals. A previous study has also demonstrated that adiposity to muscle ratio had an independent relationship with physical limitations (21). Therefore, this study supports the conceptual frame for the construct of intrinsic capacity and the notion that characteristics that capture information on nutritional, immune, or hormonal status could be considered as the vitality sub-factors.

Implications

These findings have a number of implications. From a preventive perspective, incorporating the concept of intrinsic capacity in community settings will identify an individual’s needs for the purpose of care planning and interventions to slow down or reverse declines in intrinsic capacity, address frailty, and optimize functional ability. Although a recent global analysis reported that healthy life expectancy increased in most of the counties and territories over the past 20 years (22), there is evidence that older persons in recent generations had poor physical and cognitive functioning or a higher level of frailty than did their earlier counterparts (23, 24). This situation is partly a result of the lack of awareness and knowledge on the concept of healthy ageing among older persons, carers, services providers and the general public as well as the insufficiency of services to optimize the functional ability of older adults. In many high-income countries and territories, the primary care system appears to correspond to chronic medical illness (e.g., diabetes, hypertension, pain etc.), which are only partial components of the multifaceted concept of healthy ageing, and neglects those without 'significant’ conditions. To address the complexity of health states in older age, services will need to be designed around older adults’ intrinsic capacity in addition to managing chronic diseases and their risk factors. In addition, identified loss of intrinsic capacity should be followed-up by evidence-based interventions.

Trajectories of intrinsic capacity could be used as an indicator of the effectiveness of actions/interventions to promote healthy ageing. It is also essential to facilitate alignment of health systems to provide person-centred and integrated care to older adults, as proposed in the WHO’s Policy Framework for Healthy Ageing (2). For example, data around intrinsic capacity/trajectories of intrinsic capacity collected could be linked and shared among health and social organizations to enable coordinated and efficient health management. In this context, an operational definition of an intrinsic capacity score summarizing the intrinsic capacity levels and the development of reference/cut-off values are particularly essential. As acknowledged by the WHO, a metric for overall health status and/or healthy ageing rather than the presence or absence of disease is desirable (2), as this could be used to measure and monitor healthy ageing across an individual’s life-course, and to enable the maintenance of functional ability and well-being of older adults. Nevertheless, the scores calculated by Beard et al. and in this study were subjected to the distribution of the variables included within the study population that would thus not necessarily apply to other populations. Further studies in different populations (e.g., young-old, populations from different income settings) are required to identify the reference/cut-off values of an intrinsic capacity score for indicating less than optimum trajectory.

From a clinical perspective, incorporating assessments of intrinsic capacity in hospital settings may benefit community practice and be able to promote healthy ageing. For example, intrinsic capacity assessments in combination with clinical assessments conducted in hospital settings would be used to develop a holistic health care/discharge plan. While hospitals should continue to provide comprehensive geriatric assessments and rehabilitation services, proactive and preventive approach (e.g., assessments of intrinsic capacity) could be adopted to maintain intrinsic capacity, manage frailty, and address functional impairment of older adults.

Strengths and limitations

Strengths of this study include 1) the longitudinal nature of the study design which allows us to examine the direction of causality, 2) the use of both objective and self-reported measures in defining intrinsic capacity, and 3) the bi-factor model scores that represents a pure measure of the underlying latent trait of interest, after controlling for all five specific sub-factors. The limitations include 1) the lack of hearing measures in the sensory domain, 2) data on cognitive and psychological capacities was obtained through self-report, and 3) the loss to follow-up bias.

Conclusions

In conclusion, this study supports the construct and predictive validity of the proposed capacity domains of intrinsic capacity and demonstrated how these domains could be measured to reflect the overall health status in a longitudinal cohort of Chinese older adults. The proposed domains of intrinsic capacity and their sub-factors could be used to monitor and maximize intrinsic capacity. Further studies are warranted for consensus about which sub-factors should be considered in the construct of intrinsic capacity, how to measure intrinsic capacity, dynamic interactions among domains of intrinsic capacity, trajectory of intrinsic capacity, normative cut-offs for indicating less than optimum intrinsic capacity, interventions focused on optimization of intrinsic capacity, and particularly, ways to incorporate the concept of intrinsic capacity as part of a routine health assessment in both community and hospital settings.

Conflict of Interest

None.

Funding

The work was supported by the National Institutes of Health R01 Grant AR049439-01A1 and the Chinese University of Hong Kong Research Grants Council Earmarked Grant CUHK4101/02M.

Ethical standards

The study was approved by the Joint Chinese University of Hong Kong and New Territories East Cluster Clinical Research Ethics Committee in Hong Kong, which required informed consent to be obtained.

Electronic supplementary material

Supplementary material is available for this article at https://doi.org/10.1007/s12603-021-1637-z and is accessible for authorized users.

Supplementary material, approximately 32.3 KB.