Article Text

Download PDFPDF
Ageing trajectories of health—longitudinal opportunities and synergies (ATHLOS) Healthy Ageing Scale in adults from 16 international cohorts representing 38 countries worldwide
  1. Elena Critselis1,2,
  2. Dimitris Panaretos1,
  3. Albert Sánchez-Niubò3,4,
  4. Iago Giné-Vázquez3,
  5. José Luis Ayuso-Mateos4,5,6,
  6. Francisco Felix Caballero7,8,
  7. Javier de la Fuente5,6,
  8. Josep Maria Haro4,
  9. Demosthenes Panagiotakos1,9
  1. 1 Department of Nutrition and Dietetics, School of Health Science and Education, Harokopio University, Athens, Greece
  2. 2 Proteomics Facility, Center for Systems Biology, Biomedical Research Foundation of the Academy of Athens, Athens, Greece
  3. 3 Research, Innovation and Teaching Unit, Parc Sanitari Sant Joan De Deu, Sant Boi De Llobregat, Spain
  4. 4 Centro De Investigación Biomédica En Red De Salud Mental (CIBERSAM), Instituto De Salud Carlos III, Madrid, Spain
  5. 5 Department of Psychiatry, Universidad Autonoma De Madrid, Madrid, Spain
  6. 6 Hospital Universitario de La Princesa, Instituto de Investigación Sanitaria Princesa (IIS Princesa), Madrid, Spain
  7. 7 Department of Preventive Medicine and Public Health, Universidad Autónoma de Madrid/Idipaz, Madrid, Spain
  8. 8 Centro de Investigación Biomédica en Red de Epidemiología y Salud Pública, CIBERESP, Madrid, Spain
  9. 9 Faculty of Health, University of Canberra, Canberra, Australia
  1. Correspondence to Demosthenes Panagiotakos, Department of Nutrition and Dietetics, Harokopio University, Athens 17671, Greece; dbpanag{at}hua.gr

Abstract

Background Uniform international measurement tools for assessing healthy ageing are currently lacking.

Objectives The study assessed the novel comprehensive global Ageing Trajectories of Health: Longitudinal Opportunities and Synergies (ATHLOS) Healthy Ageing Scale, using an Item Response Theory approach, for evaluating healthy ageing across populations.

Design Pooled analysis of 16 international longitudinal studies.

Setting 38 countries in five continents.

Subjects International cohort (n=355 314), including 44.4% (n=153 597) males, aged (mean±SD) 61.7±11.5 years old.

Methods The ATHLOS Healthy Ageing Scale (including 41 items related to intrinsic capacity and functional ability) was evaluated in a pooled international cohort (n=355 314 from 16 studies) according to gender, country of residence and age group. It was also assessed in a subset of eight cohorts with ≥3 waves of follow-up assessment. The independent samples t-test and Mann–Whitney test were applied for comparing normally and skewed continuous variables between groups, respectively.

Results The ATHLOS Scale (range: 12.49–68.84) had a mean (±SD) value of 50.2±10.0, with males and individuals >65 years old exhibiting higher and lower mean scores, respectively. Highest mean scores were detected in Switzerland, Japan and Denmark, while lowest in Ghana, India and Russia. When the ATHLOS Scale was evaluated in a subset of cohorts with ≥3 study waves, mean scores were significantly higher than those of the baseline cohort (mean scores in ≥3 study waves vs baseline: 51.6±9.4 vs 50.2±10.0; p<0.01).

Conclusions The ATHLOS Healthy Ageing Scale may be adequately applied for assessing healthy ageing across populations.

  • Ageing
  • epidemiology of ageing
  • public health
  • research methods
  • screening

Statistics from Altmetric.com

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

INTRODUCTION

The demographic shift towards an ever-ageing population has increased the health burden of age-related disorders and related disability,1 2 incurring 51.3% of the total global disease burden among adults and between 137.8 (128.9–148.3) and 265.9 (251.0–280.1) disability adjusted life years (DALYs) per 1000 adults.3 As of such, the promotion of healthy ageing is an emerging public health priority for ensuring both individuals’ health outcomes and quality of life, as well as diminishing healthcare-associated costs.4 5

The processes of ageing (namely the hallmarks of ageing) encompass multiple biological pathways and processes, including genomic instability and epigenetic alterations, modified intracellular and intercellular sensing and communication, and cellular senescence.6–11 At the organismal level, these processes are often mediated by both non-modifiable (eg, highest educational level attained and socioeconomic status),2 12 13 and modifiable (eg, dietary intake and physical activity)14–17 risk factors. While mounting research efforts strive to elucidate the determinants of healthy ageing so as to better inform related public health interventions, they are inherently impeded by current gaps in a common operationalisation and corresponding uniform comprehensive measurement methods for assessing healthy ageing18 based on refined definitions of its outcomes.19 For example, among 50 recently reviewed investigations related to healthy ageing, less than half (19/50) uniformly applied Rowe and Kahn’s three standards for assessing healthy ageing (where successful ageing is defined as a combination of three components, including avoiding disease and disability, high cognitive and physical function, and engagement with life20), half (24/50) employed diverse scales for assessing psychological well-being, and two thirds measured physical capabilities by either activities of daily living (37/50) and/or cognitive function (33/50).21 Furthermore, most recently the WHO proposed in 2015 that healthy ageing ought to be considered as a function of an individual’s functional ability over time (namely the ‘ongoing process of developing and maintaining the functional ability that enables well-being in older age’), rather than the mere presence or absence of disease at a single time point. To this effect, an individual’s intrinsic capacity ought to be considered comprehensively by addressing all physical and mental capacities, including vital, sensory, locomotive, cognitive and psychological capacities without considering interactions with environmental factors (eg, medications, personal support, assistive devices, among others) that may either facilitate or hinder such capacity. Functional ability is in turn determined by the interaction of individuals’ intrinsic capacities and their environment. Hence, there exists an urgent need for deriving a uniform, comprehensive measurement scale for assessing healthy ageing.5

The Ageing Trajectories of Health: Longitudinal Opportunities and Synergies (ATHLOS) project (https://academic.oup.com/ije/article/48/4/1052/5477844) sought to overcome these gaps by developing a single, comprehensive, global health score (namely the ATHLOS Healthy Ageing Scale), using an Item Response Theory (IRT) approach, for assessing healthy ageing across populations.22 The IRT approach attempts to explain the relationship between latent traits (ie, unobservable attributes) and their manifestations (ie, observed outcomes and/or responses), by establishing a link between the properties of items on an instrument, individuals responding to these items and the underlying trait being measured, on an assumed unobservable continuum and/or scale. As of such, the IRT approach focuses on establishing the individual’s position on the continuum scale. Using this approach, the ATHLOS Healthy Ageing Scale rendered excellent reliability (>0.90), establishing its adequacy for conducting group comparisons. Concurrent validity of the scale with sociodemographic, life and health factors, and predictive validity with mortality, have shown that this scale corresponds well with health status and could be potentially useful for conducting international ageing-related investigations. While initial findings of the ATHLOS Healthy Ageing Scale are promising,22 23 the present work aims to evaluate the ATHLOS Scale in an international pooled cohort.

METHODS

As previously detailed,22 the ATHLOS Healthy Ageing Scale was constructed using a two-parameter logistic item response theory model with 41 items related to intrinsic capacity and functional ability, and heterogeneities in the cohort-specific datasets were analysed and addressed to obtain a common scale.2 Specifically, following the harmonisation procedure also detailed previously,24 study-specific variables were transformed in terms of the same definition of variables and format across studies. To this effect, to develop the healthy ageing scale, a comprehensive list of 41 study-specific variables related to items on intrinsic capacity and functional ability were assessed across studies, covering domains such as vitality, sensory skills, locomotion/mobility, cognition and activities and instrumental activities of daily living. Study-specific variables were harmonised into dichotomous items expressing the presence (or absence) of difficulties. Continuous variables were dichotomised in the first quartile indicating the presence of difficulties. Only those items that were successfully harmonised in at least three studies were included in the final model. The harmonisation algorithms of each item per study included can be found at URL: https://github.com/athlosproject/athlos-project.github.io, while the detailed items selected to construct the scale are tabulated in table 1.22

Table 1

List of 41 domains and items related to functioning and health and their parameter estimates from the item response theory (IRT) model

The study sample used to construct the scale included all individuals aged >18 years, who were selected individuals based on enrolment and had non-missing items values in at least one of the items evaluated.24 The study sample is detailed in table 2.

Table 2

Descriptive statistics of the study population employed for developing the ATHLOS Healthy Ageing Scale

With a score range between 0 and 100, higher scores on the ATHLOS Healthy Ageing Scale are indicative of more favourable levels of healthy ageing.22 23

For the present work, data from the following 16 studies, with populations from 38 countries in five continents, were considered: the 10/66 Dementia Research Group Population-Based Cohort Study (10/66); the Australian Longitudinal Study of Aging (ALSA); the China Health and Retirement Longitudinal Study (CHARLS); Collaborative Research on Ageing in Europe (COURAGE); the English Longitudinal Study of Ageing (ELSA); Study on Cardiovascular Health, Nutrition and Frailty in Older Adults in Spain (ENRICA); the Health, Alcohol and Psychosocial factors in Eastern Europe Study (HAPIEE); the Health 2000/2011 Survey (H2000/11); the Health and Retirement Study (HRS); the Japanese Study of Aging and Retirement (JSTAR), the Korean Longitudinal Study of Ageing (KLOSA); the pilot-study Longitudinal Aging Study in India (LASI); the Mexican Health and Aging Study (MHAS); the Study on Global Ageing and Adult Health (SAGE); the Survey of Health, Ageing and Retirement in Europe (SHARE); and the Irish Longitudinal Study of Ageing (TILDA).

The study sample consisted of 355 314 unique participants enrolled, of which 44.4% (n=153 597) were male, aged (mean±SD) 61.7±11.5 years old, while 55.6% (n=192 335) were female, aged (mean±SD) 61.2±12.8 years old. The majority of participants (n=222 618; 63.9%) were aged ≤65 years old. Approximately, half (52.1%, n=185 362) were from Europe, 21.6% (n=76 864) from North and South America, 22.9% (n=81 201) from Asia, 0.6% (n=2087) from Oceania and 2.8% (n=9800) from Africa. The scale was also evaluated in a subset of cohorts including ≥3 waves of prospective longitudinal assessment (n=181 504). For this purpose, 8 ATHLOS datasets were employed, including the ALSA, ELSA, ENRICA, HRS, JSTAR, KLOSA, MHAS, and SHARE cohorts, corresponding to 51.1% of the original pooled cohort.

Research ethics approval

All participants of the cohorts included in the ATHLOS study had previously provided written informed consent for study participation within the context of the original studies. Additionally, all cohorts included in the ATHLOS study had previously attained study approval from their respective local research ethics committees and/or Institutional Review Boards for secondary data use and analyses. The ATHLOS study protocol was approved under the data access and ethics governance requirements of the studies of origin. A detailed description of the ATHLOS cohort and all the participating studies, including research ethics approval processes, has been detailed previously.2 22

Statistical analyses

The ATHLOS Scale was evaluated according to gender, countries and/or regions of residence, and age group (namely >60 years or >65 years). It was also evaluated in a subset of 8 ATHLOS cohorts which included ≥3 waves of follow-up assessment, as detailed above. Normally distributed continuous variables are presented as mean±SD and categorical variables as relative frequencies. The independent samples t-test and Mann–Whitney test were applied for comparing normally and skewed continuous variables between groups, respectively. All reported p values were based on two-sided tests. R software (version 3.5.3) was used for all calculations.

RESULTS

Descriptive properties of the ATHLOS Healthy Ageing Scale by gender, age and region

The descriptive statistics of the Healthy Ageing Scale in the overall harmonised ATHLOS dataset, by gender, age and region of residence are presented in table 3. In the overall harmonised ATHLOS dataset, the scale had a left skewed distribution, with mean (±SD) value of 50.22±10.0. Males had higher mean scores (mean±SD score in males vs females: 51.6±9.7 vs 49.1±10.1; p<0.01), as well as median (median±SE score in males vs females: 53.0±0.03 vs 49.4±0.02; p<0.01). Both the mean and median values of the ATHLOS Scale were lower among individuals aged >60 years old (mean±SD scores in individuals aged >60 years vs ≤60 years: 47.8±10.1 vs 52.6±9.3; exact p<0.001), as well as those aged >65 years old (mean ± SD scores in individuals aged >65 years vs ≤65 years: 46.7±10.2 vs 52.2±9.3; exact p<0.01).

Table 3

Descriptive statistics of the ATHLOS Healthy Ageing Scale, by gender, age and region of residence

The ATHLOS Healthy Ageing Scale was also assessed separately in each of the geographic regions and countries of participants’ residence (table 3). The countries with the highest ranking scores were primarily from Europe (namely Switzerland and Denmark with mean (±SD) scores of 56.2±7.5 and 55.4±8.7, respectively) and East Asia (namely Japan, 55.7±7.3). Furthermore, the countries with the lowest (mean±SD) score ranking included Ghana (41.8±12.0), India (42.9±10.4) and Russia (43.0±12.2).

ATHLOS Health Ageing Scale in a subset of studies including ≥3 waves of follow-up assessment

In the ATHLOS subset of studies including ≥3 waves of follow-up assessment, the ATHLOS Healthy Ageing Scale had mean and median values of 51.6 and 52.6, respectively (range: 12.49–67.01). Comparisons of the scale score distributions in the original sample and cohort subset are presented in figure 1. The mean value of the Healthy Ageing Scale in the cohort subset was higher than that observed in the original harmonised ATHLOS database (mean±SD score in ≥3 study waves subset vs original pooled cohort: 50.22±10.0 vs 51.62±9.4; Wilcoxon p<0.001). Moreover, comparisons of the distributions of the ATHLOS Healthy Ageing Scale by country of residence in the original pooled cohort as compared to the cohort subset with ≥3 study waves are illustrated in figure 2.

Figure 1

Distributions of the ATHLOS Healthy Ageing Scale in the overall ATHLOS pooled cohort and subset consisting of prospective longitudinal studies with ≥3 study waves. Panel A depicts the distribution of the ATHLOS Healthy Ageing Scale in the original harmonised pooled cohort (n=355 314), while Panel B illustrates the corresponding distribution in the subset of prospective cohorts with ≥3 study waves (n=181 504). ATHLOS, Ageing Trajectories of Health: Longitudinal Opportunities and Synergies.

Figure 2

Distributions of the ATHLOS Healthy Ageing Scale in the overall ATHLOS pooled cohort and subset consisting of prospective longitudinal studies with ≥3 study waves according to participants’ country of residence. Panel A depicts the distribution of the scale in the original pooled analysis (n=355 314), while Panel B illustrates the corresponding distribution in the subset of prospective cohorts with ≥3 study waves (n=181 504). ATHLOS, Ageing Trajectories of Health: Longitudinal Opportunities and Synergies.

DISCUSSION

The novel ATHLOS Healthy Ageing Scale was recently refined and evaluated for its within the context of theATHLOS project (EU HORIZON2020–PHC-635316, http://athlosproject.eu/) harmonised data set.23 The ATHLOS Scale was evaluated in a harmonised pooled cohort, including 16 international longitudinal studies and 355 314 participants. In the pooled cohort, the scale had a mean (±SD) value of 50.2±10.0, with males exhibiting higher scores. As expected, mean scores were lower among individuals aged older than 60 or 65 years. As of such, it is inferred that the ATHLOS Scale scores is inversely associated with participants’ age. With regard to the distribution of the ATHLOS Healthy Ageing Scale scores according to participants’ continent of residence, highest mean scores were detected in European residents, while the lowest was in Africa. Specifically, the highest ranking of the scale scores was among residents of Switzerland, Japan and Denmark, while in contrast those with the lowest mean scores were of Ghana, India and Russia. Finally, when the scale was further evaluated in a subset of cohorts with ≥3 waves of prospective longitudinal assessment, the mean ATHLOS scale scores were higher than that observed in the original cohort, potentially due to cohort effects. Therefore, the ATHLOS Healthy Ageing Scale may be adequately and readily applied for assessing healthy ageing both across populations and over time trajectories.

Healthy ageing entails a continuous phenomenon, including various trajectories regarding inherent health attributes and abilities, which may be readily modified by health behaviours and/or individuals’ surrounding environment.17 23 25 26 As of such, it is essential that a metric score of healthy ageing must be comprehensive although refined in nature,19 encompassing the aforementioned domains, as well as their interactive effects over time. It is by this manner that a developed health metric score is anticipated to serve as a sound methodological tool for assessing healthy ageing across different populations and over time trajectories. However, until most recently, such a comprehensive measurement tool remained to be developed and applied at the international level.5 Hence, the generalisability, as well as public health applications, of research efforts dedicated to healthy ageing were hindered. Specifically, as previously reviewed18, healthy ageing often remains without uniform definition and measurement methods, severely deterring the translational capacity of research investigations into public health practice.

The ATHLOS Healthy Ageing Scale seeks to overcome the current gaps in uniformly assessing healthy ageing.22 For the development of the ATHLOS Scale, health was operationalised according to the WHO’s notion of ‘health state’, including (a) an intrinsic attribute of an individual that can be aggregated to the population level and (b) comprising domains of human functioning that describe the actual impact of health conditions on people’s lives.27 This approach seeks to go further than previous ageing scores, such as the Frailty Index, which primarily focuses on ageing and vulnerability to poor outcomes.28 The domains included for assessing health are based on highest methodological quality studies, entailing multiple methodologies and multi-method studies.29 In particular, the ATHLOS Scale was developed based on self-reported health-related questions and measurement tests measured within the context of one of the largest population-based cohort studies, with its performance subsequently assessed in the largest related cohort to date including longitudinal cohorts from 16 European and international countries. In particular, a total of 41 items, including self-reported health questions and measured tests, assessed at baseline, were employed to develop a global health metric score. The IRT approach employed allowed for the inclusion of both anchor items (items that are available in all waves) and items that varied across time trajectories and/or study waves. Thus, the ATHLOS Healthy Ageing Scale assesses health status conceptualised as a vector of functioning in several differing domains (eg, vision, walking, kneeling, Activities of Daily Living, and Instrumental Activities of Daily Living).

Previous studies have demonstrated that robust health metric scores may be constructed based on individual data.30 31 In addition, similar approaches as above have been previously employed to construct a health metric operationalised as domains of functioning, although limited to solely two populations and without encompassing longitudinal assessment.32 In contrast, the ATHLOS Scale employs a comprehensive assessment of healthy ageing, entailing disparate factors related to health including physical functioning and mobility, cognitive functions, sight, Instrumental Activities of Daily Living, among others. Through its development in the largest related international pooled cohort established to date, the procedure employed not only ensures the comprehensive assessment of healthy ageing based on refined outcome measures,19 but the present findings additionally reveal that the ATHLOS Healthy Ageing Scale may be applied across populations. Moreover, as observed in the present findings, the ATHLOS Healthy Ageing Scale may be additionally applied with reasonable robustness longitudinally over time trajectories for assessing healthy ageing across populations over time.

The ATHLOS Healthy Ageing Scale has been previously demonstrated to be inversely associated with well-established health determinants,23 including gender, socioeconomic status, and educational level.33 In particular, based on mixed-effects multilevel regression models, individuals with non-modifiable risk factors for adverse ageing, including lower educational and socioeconomic status, have been shown to have worse performance on the ATHLOS Scale.23 Additionally, those with modifiable risk factors for adverse ageing,34 including smoking and/or physical inactivity, are inversely associated with the health metric score.3 Additionally, within the context of the present work it is demonstrated that with the exception of Greece, most Mediterranean countries had intermediate and/or low (ie, Portugal) score rankings. Thus, it is hypothesised that the ATHLOS Scale may also reflect inequities and barriers to health between population groups. The detection of such barriers to healthcare access may serve as pivotal public health policy targets aimed at enhancing healthy ageing.35 However, it cannot be precluded that the observed differences may be also driven by differences in culture, lifestyle, and other health-related factors, as well as wider societal factors which are known drivers of health inequalities and outcomes, between countries and regions.

Strengths and limitations

The study strengths include the vast number of longitudinal studies, including 16 prospective cohorts arising from five continents, applied for the development of the ATHLOS Healthy Ageing Scale. The harmonisation approach and tools used were both systematic and rigorous to ensure that harmonised variables are comparable. The study limitations include that since harmonisation is a retrospective process, studies were not initially designed to be harmonised. As of such, the heterogeneity in study design, instruments and data collection limits the amount and quality of information pooled and applied for the development of the final health metric.

CONCLUSIONS

ATHLOS Healthy Ageing Scale may be adequately applied for assessing healthy ageing across populations. The further application of the ATHLOS Healthy Ageing Scale in other research settings is anticipated to facilitate concerted future international and longitudinal investigations which may optimally inform and guide future related public health interventions worldwide.35

What is already known

  • Uniform international measurement tools for assessing healthy ageing are currently lacking.

  • Most recently, the Horizons 2020 Ageing Trajectories of Health: Longitudinal Opportunities and Synergies (ATHLOS) project sought to overcome these gaps by developing a single, comprehensive, global health score (namely the ATHLOS Healthy Ageing Scale) using an Item Response Theory (IRT) approach, for assessing healthy ageing across populations worldwide.

  • The ATHLOS Healthy Ageing Scale remains to be evaluated in international settings.

What this study adds

  • The ATHLOS Healthy Ageing Scale (including 41 items related to intrinsic capacity and functional ability) was evaluated in a pooled international cohort (n=355 314 from 16 studies) according to gender, country of residence, and age group, while it was also assessed in a subset of 8 cohorts with ≥3 waves of follow-up assessment.

  • Higher mean scores of the ATHLOS Scale (range: 12.49–68.84) are indicative of more favourable healthy ageing. Highest mean scores were detected in populations from Switzerland, Japan, and Denmark, while lowest were among those from Ghana, India and Russia.

  • The ATHLOS Healthy Ageing Scale may be adequately applied for assessing healthy ageing across populations.

REFERENCES

Footnotes

  • Contributors Conception and design: JMH and DP. Acquisition of data: EC, JLAM, FFC, JDLF, JMH and DP. Analysis of data: EC, DP, ASN, JMH, DP. Interpretation of data: EC, DP, ASN, IG-V, JLAM, FFC, JDLF, JMH, DP. Drafting of manuscript: EC, DP, DP. Critical revision of manuscript: ASN, IG-V, JLAM, FFC, JDLF, JMH. Final approval: All authors.

  • Funding This work was conducted within the context of the Ageing Trajectories of Health: Longitudinal Opportunities and Synergies (ATHLOS) project (European Union’s Horizon 2020 research and innovation programme, grant agreement no. 635316).

  • Competing interests None declared.

  • Patient consent for publication Not required.

  • Provenance and peer review Not commissioned; externally peer reviewed.

  • Data availability statement Data may be obtained from a third party and are not publicly available.