Socioeconomic status, non-communicable disease risk factors, and walking speed in older adults: multi-cohort population based study

Abstract Objective To assess the association of low socioeconomic status and risk factors for non-communicable diseases (diabetes, high alcohol intake, high blood pressure, obesity, physical inactivity, smoking) with loss of physical functioning at older ages. Design Multi-cohort population based study. Setting 37 cohort studies from 24 countries in Europe, the United States, Latin America, Africa, and Asia, 1990-2017. Participants 109 107 men and women aged 45-90 years. Main outcome measure Physical functioning assessed using the walking speed test, a valid index of overall functional capacity. Years of functioning lost was computed as a metric to quantify the difference in walking speed between those exposed and unexposed to low socioeconomic status and risk factors. Results According to mixed model estimations, men aged 60 and of low socioeconomic status had the same walking speed as men aged 66.6 of high socioeconomic status (years of functioning lost 6.6 years, 95% confidence interval 5.0 to 9.4). The years of functioning lost for women were 4.6 (3.6 to 6.2). In men and women, respectively, 5.7 (4.4 to 8.1) and 5.4 (4.3 to 7.3) years of functioning were lost by age 60 due to insufficient physical activity, 5.1 (3.9 to 7.0) and 7.5 (6.1 to 9.5) due to obesity, 2.3 (1.6 to 3.4) and 3.0 (2.3 to 4.0) due to hypertension, 5.6 (4.2 to 8.0) and 6.3 (4.9 to 8.4) due to diabetes, and 3.0 (2.2 to 4.3) and 0.7 (0.1 to 1.5) due to tobacco use. In analyses restricted to high income countries, the number of years of functioning lost attributable to low socioeconomic status by age 60 was 8.0 (5.7 to 13.1) for men and 5.4 (4.0 to 8.0) for women, whereas in low and middle income countries it was 2.6 (0.2 to 6.8) for men and 2.7 (1.0 to 5.5) for women. Within high income countries, the number of years of functioning lost attributable to low socioeconomic status by age 60 was greater in the United States than in Europe. Physical functioning continued to decline as a function of unfavourable risk factors between ages 60 and 85. Years of functioning lost were greater than years of life lost due to low socioeconomic status and non-communicable disease risk factors. Conclusions The independent association between socioeconomic status and physical functioning in old age is comparable in strength and consistency with those for established non-communicable disease risk factors. The results of this study suggest that tackling all these risk factors might substantially increase life years spent in good physical functioning.

completed using a computer-assisted personal interview (CAPI) and the other half using paper and pencil. SAGE Mexico used CAPI throughout and the other four countries used paper and pencil format for all interviews. Multistage cluster sampling strategies were used in all countries where, except for Mexico, households were classified into one of two mutually exclusive categories: (1) all persons aged years and older were selected from households classified as + households ; and one person aged 18-years was selected from a household classified as an -household . The arrangement in Mexico was similar, but included supplementary and replacement samples to account for losses to follow up in selected sampling units since Wave 0 (more sampling details provided when accessing the data through http://apps.who.int/healthinfo/systems/surveydata/index.php/catalog). The sample in India is also representative at the sub-national and sub-state levels for the selected states. Household enumerations were carried out for the final sampling units. One household questionnaire was completed per household-where a household informant and individual respondent need not be the same individual. One individual was selected from 18-49 households, whereas for 50+ households all individuals aged 50+ were invited to complete the individual interview. Proxy respondents were identified for selected individuals who were unable to complete the interview. Household-level analysis weights and person-level analysis weights were calculated for each country, which included sample selection and a post-stratification factor. Post-stratification correction techniques used the most recent population estimates provided by the national statistical offices. The pooled Wave 1 six-country totals for individual respondents included 34,124 respondents aged 50+ and 8,340 aged 18-49. Taiwan, 2000 and2006, provides information regarding the health and well-being of older persons in Taiwan. Taiwan has undergone rapid demographic, social, and economic changes, becoming a highly urbanized and industrial society with a growing population of persons age 65 or older. SEBAS explores the relationship between life challenges and mental and physical health, the impact of social environment on the health and well-being of the elderly, as well as biological markers of health and stress. The study collected self-reports of physical, psychological, and social well-being, plus extensive clinical data based on medical examinations and laboratory analyses. Examination of health outcomes included chronic illnesses, functional status, psychological well-being, and cognitive function. Questions regarding life challenges focused on perceived stress, economic difficulties, security and safety, and the consequences of a major earthquake. Biological markers were used to identify cardiovascular risk factors, metabolic process measures, immune-system activity, the hypothalamic-pituitary adrenal axis, and sympathetic nervous system activity. Two rounds of biomarker data collected in 2000 and 2006 were complemented by face-to-face interviews with the participants. Demographic and background variables included age, sex, education, ethnicity, occupation, and residency. Additional information about the Social Environment and Biomarkers of Aging Study can be found at the Georgetown University Center for Populations and Health Web site.

SHARE.
The Survey of Health, Ageing and Retirement in Europe (SHARE) is a multidisciplinary and cross-national panel database of micro data on health, socio-economic status and social and family networks of approximately 123,000 individuals aged 50 or older (more than 293,000 interviews). SHARE covers 27 European countries and Israel. To date, SHARE has collected more than 293,000 interviews in four panel waves on current living circumstances and one wave on retrospective life histories (SHARELIFE). The first wave was collected in 2004/2005, the second in 2006/2007, SHARELIFE in 2008/2009, the fourth wave mainly in 2011 and the fifth wave in 2013. Up to the present 20 countries participated in SHARE. However, not all countries were part of each wave and also the timing of data collection differs between countries. Table S1 gives information on the participation of countries and time of data collection for Wave 1 to Wave 5.

NSHAP.
The National Social Life, Health, and Aging Project (NSHAP) is a longitudinal, populationbased study of health and social factors, aiming to understand the well-being of older, communitydwelling Americans by examining the interactions among physical health and illness, medication use, cognitive function, emotional health, sensory function, health behaviors, social connectedness, sexuality, and relationship quality. NSHAP provides policy makers, health providers, and individuals with useful information and insights into these factors, particularly on social and intimate relationships. The study contributes to finding new ways to improve health as people age. To date there are two waves of NSHAP data available to researchers, and funding has now been awarded for a third wave of data collection. In 2005 and 2006, NORC and Principal Investigators at the University of Chicago conducted the first wave of NSHAP, completing more than 3,000 interviews with a nationally representative sample of adults born between 1920 and 1947 (aged 57 to 85 at the time of Wave 1 interview). In 2010 and 2011, nearly 3,400 interviews were completed for Wave 2 with these Wave 1 Respondents, Wave 1 Non-Interviewed Respondents, and their spouses or cohabiting romantic partners. The second wave of NSHAP is essential to understanding how social and biological characteristics change. By eliciting a variety of information from respondents over time, NSHAP provides data that will allow researchers in a number of fields to examine how specific factors may or may not affect each other across the life course.
TILDA. The Irish Longitudinal Study on Ageing (TILDA) is a large prospective cohort study examining the social, economic, and health circumstances of 8,175 community-dwelling adults aged 50 years and older resident in the Republic of Ireland. The sample was generated using a three-stage selection process and the Irish Geodirectory as the sampling frame. The Irish Geodirectory is a comprehensive listing of all addresses in the Republic of Ireland which is compiled by the national post service and ordnance survey Ireland. Subdivisions of district electoral divisions pre-stratified by SES, age, and geographical location served as the primary sampling units. The second stage involved the selection of a random sample of 40 addresses from within each PSU resulting in an initial sample of 25,600 addresses. The third stage involved the recruitment of all members of the household aged 50 years and over. Consequently, the response rate was defined as the proportion of households including an eligible participant from whom an interview was successfully obtained. A response rate of 62.0% was achieved at the household level. There were three components to the survey. Respondents completed a computer-assisted personal interview (CAPI) and a separate self-completion paper and pencil module which collected information that was considered sensitive. All participants were invited to undergo a separate health assessment at one of two national centers using trained nursing staff. At the initial interview, respondents were invited to undergo a detailed health assessment at one of two national centers in Dublin and Cork using trained nursing staff. If a respondent could not attend the health center but was agreeable to completing a health assessment, a trained nurse administered a subset of the tests in the respondent s home. The Trinity College Dublin Research Ethics Committee granted ethical approval for the study.
EPIPORTO. The EPIPorto study was initiated in 1999 and recruited 2,485 adult dwellers aged 18 years or more in the city of Porto, northwest of Portugal. Briefly, simple random digit dialing of landline telephones was used to select households. The vast majority of houses (>95%) had a landline telephone at the time of this procedure. We used a table of random numbers to define the last four digits that are specific to individual houses, assuming the local prefix codes to limit the universe to the city of Porto. Non-existing numbers, those corresponding to fax numbers or telephone numbers of non-individual subscribers were ignored. The household was considered unreachable after at least four dialing attempts at different hours and including week and weekend days. Within each household, a permanent resident aged 18 years or more was selected using simple random sampling. The proportion of participation was 70%. A follow-up evaluation was conducted from 2005 to 2008 (participation rate=68% of the baseline sample), by trained interviewers, using structured questionnaires and forms, following the same protocol for data collection as at baseline. In both evaluations, participants were invited to visit our Department at Medical School for an interview, which included a questionnaire on social, demographic, behavioural and clinical data, and a physical examination including blood collection.

Appendix 2. Details on measures
Smoking was self-reported and was categorized into current smoking for individuals currently smoking, former smoking for individuals not currently smoking but who smoke in the past, and never smoking for individuals who never smoke.
Alcohol consumption was measured in alcohol units. For most cohorts, it was the average number of alcohol units over a week. For ELSA, it was the highest intake over a week. High alcohol intake was defined as a consumption exceeding 21 alcohol units for men and 14 for women. Abstinence was defined as no alcohol consumption, and moderate intake was defined as consumption less or equal 21 alcohol units for men and 14 for women. For CRELES Pre 1945 and CRELES-RC, the strata were determined based on drinking frequency only. In SHARE, heavy alcohol consumption was assessed based on frequency of drinking more than 2 units of beer, wine, and liquor.
Physical inactivity was considered as none/light activity only for ELSA, as less than 1 hour of moderate and vigorous physical activity per week in WHITEHALL II and TILDA, as not practicing sport in GAZEL, EPIPORTO and CRELES Pre 1945. Physical inactivity was determined as no moderate or vigorous physical activity in CRELES-RS, HAALSI, NHANES, and SHARE; no moderate or vigorous physical activity at least once a week in the HRS, MIDUS, WLSG, and WLSS; no moderate or vigorous physical activity and no walking or biking more than 10 minutes continuously in SAGE; no regular exercise in SEBAS; no or less than once a month vigorous physical activity in NSHAP.
Body mass index (BMI) was measured as weight (Kg) divided by the square of height (m 2 ). Obesity is defined as BMI  30; Overweight is defined as BMI  25 and BMI<30; Normal BMI is defined as BMI  18.5 and BMI<25. For GAZEL, HRS, MIDUS, SHARE, WLSG and WLSS the values were self-reported, measured in all other ones. Hypertension was defined as the presence of at least one of the following conditions: systolic blood pressure  140 mmHg, recorded diastolic blood pressure  90 mmHg, current intake of anti-hypertensive medication, self-report. For GAZEL and CRELES-RC only the selfreported information was available. Blood pressure was self-reported in WLSG, and WLSS, and measured as the mean of two or three readings in the remaining studies. Knowledge of antihypertensive treatment was available in HAALSI, CRELES Pre 1945, TILDA, HRS, MIDUS, NHANES1999, NHANES2001, SAGE, and SHARE.
Diabetes was defined as the presence of at least one of the following conditions: fasting glucose  7 mmol/L, 2h postload glucose  11.1 mmol/L, glycated hemoglobin  6.5%, self-report. All had selfreported information, WHITEHALL-II and HAALSI further has both fasting glucose and 2h post load glucose, ELSA and CRELES-RC have glycated hemoglobin, CRELES Pre 1945 fasting glucose and glycated haemoglobin. In the NHANES, diabetes was determined based on fasting glucose and glycated haemoglobin or taking medication. In SEBAS, diabetes was determined based on glycated haemoglobin or taking medication.
Ethnicity was measured as white, Asian, black African or other. Out of 68,112 participants with known ethnicity, 52.9% were white, 22.9% Asian and 16.7% black African.