ONS Longitudinal Study Research Articles

Does the health of local populations modify occupational differences in employment rates of older workers? Findings from the ONS Longitudinal Study 2001–2011

Poor health is a key reason for early exit from the labour market. Few studies have explored how the health of local populations is related to occupational differences in employment outcomes among older people. Our study used data for England and Wales from the ONS Longitudinal Study linked with 2001 Census measures of the health of the older working age population at local authority level. We included 128,710 people aged 40-64y in 2001 who were in paid work in the previous five years. We investigated the associations of both occupation and area level with two employment outcomes ten years later (in 2011): i) in paid work or not; ii) economic activity (employed (reference), unemployed, retired, sick/disabled, other). People in elementary occupations were more likely to not be in paid work in 2011 compared to those in managerial occupations (RRR 1·55 [95%CI 1·47,1·64]). Compared to the healthiest third of local authority areas, being resident in the unhealthiest third was associated with greater likelihood of not being in paid employment ten years later (RRR 1·25 [95% CI 1·18,1·33]). While area level health was associated with employment outcomes for all major occupation groups, the gap between the healthiest and unhealthiest areas was most marked for skilled trades; process, plant and machine operatives; and elementary occupations. Occupational differences for the economic activity outcome were most marked for the sick/disabled category. Policies to improve the health of local populations may support retention and reduce occupational inequalities in employment rates of older workers.

Is foster caring associated with an earlier transition to adulthood for caregivers’ own children? ONS Longitudinal Study

This study investigates whether existing children in a fostering household differ from young people in non-caregiving households in the timing of their transitions to key adult roles, known to affect later health and life chances. Using data from the ONS Longitudinal Study, we pooled records from census years 1971–2001 and linked them to follow-up records from 1981–2011. We identified 2,656 children living with a foster child and compared their profiles on the ‘big five’ transitions to roles of adulthood – finishing school, leaving home, finding work and becoming financially independent, getting married and having children – with those of other children without a foster child in the household (N = 209,453). We fitted logistic and multinomial models that controlled for childhood socioeconomic and demographic confounders to estimate the proportion achieving the five roles in early adulthood. When compared to those without a foster child in the household, a modest but reliably higher proportion of caregivers’ children achieved the transition to adulthood. There was some evidence that caregivers’ children might cope better with the transition to adulthood if they were older than the foster child or were female. The findings suggest that supporting foster parents with delaying their children’s transition to adulthood could become part of the role of supervising social workers.

Social care in childhood and adult outcomes: double whammy for minority children?

Background Children in social care report poor outcomes in many aspects of their later lives. Less is known about differences by ethnicity. Objective We examined the health, socio-economic, family and living arrangements across the first three decades of adult life by the intersection of ethnicity (White, Black, South Asian) with social care. Participants and setting Linked census and life events data for a 1% sample of the population of England and Wales in the ONS Longitudinal Study. Participants were dependent children in 1971–2001 (analytic sample n = 669,474). Methods Categorical regression models compared health, socio-economic circumstances, living arrangements and relationships, controlling for country of birth, childhood census year, childhood and adult age in years, gender, and head of household social class, qualifications, employment status and marital status. Results Adverse adult outcomes following social care in childhood were conditional on the interaction of social care with ethnicity, mainly in the socio-economic domain. For some outcomes the White group had the poorest outcomes: for example, 15% lower probability of being employed than other White people (65% versus 80%). Black adults with a history of social care did not differ from other Black adults, except for the lowest probability of acquiring their own home, while care-experienced South Asian adults did not differ from other South Asian adults. Conclusion Minority ethnicity moderated the social care to adult outcomes relationship in both positive and negative ways. Overall, there was little evidence of intersectionality for Black children in social care affecting their life chances.

Enabling the examination of long-term mortality trends by educational level for England and Wales in a time-consistent and internationally comparable manner

BackgroundStudying long-term trends in educational inequalities in health is important for monitoring and policy evaluation. Data issues regarding the allocation of people to educational groups hamper the study and international comparison of educational inequalities in mortality. For the UK, this has been acknowledged, but no satisfactory solution has been proposed.ObjectiveTo enable the examination of long-term mortality trends by educational level for England and Wales (E&W) in a time-consistent and internationally comparable manner, we propose and implement an approach to deal with the data issues regarding mortality data by educational level.MethodsWe employed 10-year follow-ups of individuals aged 20+ from the Office for National Statistics Longitudinal Study (ONS-LS), which include education information from each decennial census (1971–2011) linked to individual death records, for a 1% representative sample of the E&W population. We assigned the individual cohort data to single ages and calendar years, and subsequently obtained aggregate all-cause mortality data by education, sex, age (30+), and year (1972–2017). Our data adjustment approach optimised the available education information at the individual level, and adjusts—at the aggregate level—for trend discontinuities related to the identified data issues, and for differences with country-level mortality data for the total population.ResultsThe approach resulted in (1) a time-consistent and internationally comparable categorisation of educational attainment into the low, middle, and high educated; (2) the adjustment of identified data-quality related discontinuities in the trends over time in the share of personyears and deaths by educational level, and in the crude and the age-standardised death rate by and across educational levels; (3) complete mortality data by education for ONS-LS members aged 30+ in 1972–2017 which aligns with country-level mortality data for the total population; and (4) the estimation of inequality measures using established methods. For those aged 30+ , both absolute and relative educational inequalities in mortality first increased and subsequently decreased.ConclusionWe obtained additional insights into long-term trends in educational inequalities in mortality in E&W, and illustrated the potential effects of different data issues. We recommend the use of (part of) the proposed approach in other contexts.

Levelling Up for health in towns? Development of a new deprivation index: the ‘Stronger Towns Index’ and its association with self-rated health and migration in England, between 2001 and 2011

AimTo develop the ‘Stronger Towns Index': a deprivation index that took into account characteristics of areas encompassing towns that may be eligible for redevelopment funding and explore how this index was associated with self-rated health and migration within England between 2001 and 2011.Subject and methodsAll members of the ONS Longitudinal Study in England aged 16 and over in 2001 whose records included a self-rated health response and a valid local authority code.Local authorities in England were ranked using a composite index developed using the five metrics set out in the Stronger Towns Funding: productivity, income, skills, deprivation measures, and the proportion of people living in towns.The index was split into deciles, and logistic regression carried out on the association between decile and self-rated health in 2001 in the main sample (n = 407,878) and decile change and self-rated health in 2011 in a subsample also present in 2011, with migration information (n = 299,008).ResultsThere were areas in the lowest deciles of Town Strength who did not receive funding. After multiple adjustment, LS members living in areas with higher deciles were significantly more likely (7% to 38%) to report good health than those in the lowest decile in 2001. Remaining in the same decile between 2001 and 2011 was associated with 7% lower odds of good self-rated health in 2011.ConclusionIt is important to consider health in towns when allocating funding. Areas in the Midlands may have missed out on funding which might help mitigate poor health.

Maternal educational inequalities in child’s birth weight: ONS Longitudinal Study

Abstract Background Studies found that maternal education is the most frequently assessed predictor of child’s birth weight. Lower level of education was consistently found as indicator of lower child’s birthweight. In turn, inequalities in child’s birthweight have been repeatedly shown to be closely related to inequalities in later adult health. The aim of this project is to evaluate the association between the highest achieved level of maternal education and birthweight of single born babies in large English and Welsh population sample, taking into account factors such as child’s gender, parity, maternal age, partnership status, ethnicity, and household socioeconomic characteristics. Methods Using Office for National Statistics Longitudinal Study (ONS LS) based on English and Welsh Census data, information from almost quarter of a million children born since 1981 to ONS LS sample mothers were used. Maternal education was categorised in 3 categories (below secondary, completed secondary education, degree and higher), and its association with child’s birth weight was analysed by logistic regression accounting for range of available covariates. Results Significant association between the level of education and birth weight was found in crude analysis (p < 0.001). When adjusted, the magnitude of the association with education gradient declined but remained highly significant and was found to considerably increase over the years. The birth weight difference between those born to mothers with below secondary education and those born to mothers with degree increased by more than 60 grams (p for change <0.001) between 1981 and 2016. Conclusions These findings support previous evidence based on different population samples. According to our results, children of mothers with below secondary education tend to have lower birth weight. Our results suggest that the inequalities in birth weight by the highest education level achieved by mother significantly increased since 1981. Key messages • Lower levels of maternal education predict low birth weight in children. • The differences in birth weight by maternal educational demonstrate increase in inequalities over the years.

Out-of-home care in childhood and socio-economic functioning in adulthood: ONS Longitudinal study 1971–2011

BackgroundChildren who spent time in non-parental care report poor outcomes in many aspects of their later lives on average, but less is known about differences by type of care. We examined whether socioeconomic, family, and living arrangements of adults who had been in non-parental care across the first three decades of adult life varied by type of care (residential, non-relative and relative). MethodsWe used longitudinal data from the Office for National Statistics Longitudinal Study (LS). Participants were aged <18 years and had never been married at baseline of each census year from 1971 to 2001 (n = 242,843). Separately for each adult follow-up age (20–29; 30–39; 40–49), multi-level logistic regression models were used to compare socioeconomic, family, and living arrangements by different out-of-home care (OHC) experiences. ResultsAny OHC increased the likelihood of poorer functioning in the three domains of socioeconomic circumstances, family formation and relationships, and living arrangements. This was evident in participants’ 20 s, 30 s and 40 s; the most adverse outcomes were observed for those with a history of residential care, followed by non-relative OHC, and the least adverse outcomes for relative OHC. Moderation by childhood census year and gender altered the relationship between OHC and some, but not all, adult outcomes. The strongest, most consistent, evidence was for widening of inequalities in outcomes across childhood census years. ConclusionEnduring inequalities in social and economic functioning for OHC-experienced adults were found. The evidence supports the policy to place children in relative care whenever possible, with residential care the least favoured option.

1014Maternal educational qualification and child’s birth weight: Findings from Office for National Statistics Longitudinal Study

Abstract Background Lower maternal education was found to be associated with lower child’s birthweight which in turn was a possible risk factor for later poor health. Presented research aims to assess the association between maternal education and singleton’s birthweight in large UK data, accounting for characteristics such child’s gender, parity, maternal age, partnership status, ethnicity, and household socioeconomic characteristics. Methods Using England and Wales Office for National Statistics Longitudinal Study (ONS-LS), data from over 240,000 children born since 1981 to ONS-LS sample mothers were used. Maternal education was derived into 3 categories (below secondary, complete secondary education, degree and higher). Results Crude analysis confirmed significant association between the level of education and birthweight in each Census cohort (p &amp;lt; 0.001). In adjusted models, the education gradient was partly explained but remained strongly significant, and substantially increased over the years: for example, the birthweight difference between those with below secondary educated mothers and those with degree increased from 29 to 92 grams (p for change&amp;lt;0.001). Conclusions Our findings support previous evidence using different, usually smaller, population samples. Children of mothers with no or low qualification are more prone to be born with lower birthweight leading potentially into health disadvantages in their later life. Our results suggest that the inequalities in birth weight increased over the last 35 years. Key messages Low levels of maternal education predicts low birthweight in children.

Describing socio-economic variation in life expectancy according to an individual's education, occupation and wage in England and Wales: An analysis of the ONS Longitudinal Study.

People who live in more deprived areas have poorer health outcomes, and this inequality is a major driver of health and social policy. Many interventions targeting these disparities implicitly assume that poorer health is predominantly associated with area-level factors, and that these inequalities are the same for men and women. However, health differentials due to individual socio-economic status (SES) of men and women are less well documented. We used census data linked to the ONS Longitudinal Study to derive individual-level SES in terms of occupation, education and estimated wage, and examined differences in adult mortality and life expectancy. We modelled age-, sex- and SES-specific mortality using Poisson regression, and summarised mortality differences using life expectancy at age 20. We compared the results to those calculated using area-level deprivation metrics. Wide inequalities in life expectancy between SES groups were observed, although differences across SES groups were smaller for women than for men. The widest inequalities were found across men's education (7.2-year (95% CI: 3.0–10.1) difference in life expectancy between groups) and wage (7.0-year (95% CI: 3.5–9.8) difference), and women's education (5.4-year (95% CI: 2.2–8.1) difference). Men with no qualifications had the lowest life expectancy of all groups. In terms of the number of years' difference in life expectancy, the inequalities measured here with individual-level data were of a similar magnitude to inequalities identified previously using area-level deprivation metrics. These data show that health inequalities are as strongly related to individual SES as to area-level deprivation, highlighting the complementary usefulness of these different metrics. Indeed, poor outcomes are likely to be a product of both community and individual influences. Current policy which bases health spending decisions on evidence of inequalities between geographical areas may overlook individual-level SES inequalities for those living in affluent areas, as well as missing important sex differences.

Is studying medicine good for your health? Long-term health outcomes of a cohort of clinical medicine graduates in England and Wales in the ONS Longitudinal Study

ObjectiveTo quantify the potential protective effect on health associated with study of a clinical medicine degree.DesignProspective population-based cohort data collected at census and linked over time: cohort born before 1976...

Non-parental care in childhood and health up to 30 years later: ONS Longitudinal Study 1971-2011.

Children who spend time in non-parental care report worse health later in life on average, but less is known about differences by type of care. We examined whether self-rated health of adults who had been in non-parental care up to 30 years later varied by type of care. We used longitudinal data from the office for National Statistics Longitudinal Study. Participants were aged <18 and never-married at baseline of each census year from 1971 to 2001. Separately for each follow-up period (10, 20 and 30 years later), multi-level logistic regression was used to compare self-rated health outcomes by different care types. For combined census years, sample sizes were 157 896 dependent children with 10 years of follow-up, 166 844 with 20 years of follow-up and 173 801 with 30 years of follow-up. For all follow-up cohorts, longitudinal study members who had been in care in childhood, had higher odds of rating their health as 'not good' vs. 'good'; with highest odds for residential care. For example, 10-year follow-up odds ratios were 3.5 (95% confidence interval: 2.2-5.6) for residential care, 2.1 (1.7-2.5) for relative households and 2.6 (2.1-3.3) for non-relative households, compared with parental households after adjustment for childhood demographics. Associations were weakest for 10-year, and strongest for 20-year, follow-up. Additional adjustment for childhood social circumstances reduced, but did not eliminate, associations. Decades after children and young people are placed in care, they are still more likely to report worse health than children who grew up in a parental household.

Are ‘healthy cohorts’ real-world relevant? Comparing the National Child Development Study (NCDS) with the ONS Longitudinal Study (LS)

Comparisons between cohort studies and nationally representative ‘real-world’ samples are limited. The NCDS (1958 British birth cohort) follows those born in Britain in a single week in March 1958 (n=18,558); and the ONS Longitudinal Study (LS) contains linked census data and life events for a 1% sample of the population of England and Wales (&gt; 1 million records; allowing for sub-samples by age, ethnicity, or other socio-demographic factors). Common country-and age-matched socio-demographic variables were extracted from the closest corresponding time-points, NCDS 55-year survey in 2013 (n=8107) and LS respondents aged 55 in 2011 (n=7052). Longitudinal associations between socio-demographic exposures (from the NCDS 46-survey in 2004 and LS respondents aged 45 in 2001) and long-term limiting illness (from NCDS 2013 and LS respondents 2011, aged 55) were assessed using logistic regression. The NCDS 55-year sample had similar characteristics to LS respondents aged 55 for sex and marital status, but the NCDS sample had lower levels of long-term limiting illness (19.7% vs 22.8%), non-white ethnicity (2.1% vs 11.7%) and living in South England (46.9% vs 50.1%), and higher levels of full-time employment (61.2% vs 55.2%), working in professional/higher managerial occupations (35.7% vs 29.2%), and living with a spouse (69.1% vs 64.9%), all p&lt;0.001. Nevertheless, longitudinal associations between socio-demographic exposures and long-term limiting illness were similar in the NCDS and LS samples (all tests of between-study heterogeneity in mutually adjusted models p&gt;0.09) suggesting these NCDS findings are largely generalisable to the population of England and Wales.

The Office for National Statistics Longitudinal Study

Background and Rationale The ONS Longitudinal Study (‘the LS’) covers England and Wales and includes individual data from the 1971 – 2011 decennial censuses and linked information on births, deaths and cancer registrations. It is representative of the population of England and Wales.&#x0D; Aim This presentation describes the LS and the linked administrative data, and showcases recent/ prominent examples of research.&#x0D; Methods and Approach The LS is built around samples drawn from decennial censuses, with its initial sample drawn from the 1971 Census. It also contains information about other people living in a sample-member’s household. Substantial emphasis is placed on security of access to the data and its responsible use. All research outputs are checked and are only released to users once disclosure control requirements are met. Linkage of study members from one census to another and vital events is carried out by ONS.&#x0D; Results The LS has been used for a variety of research. Using linked census and death records occupational differences in mortality rates have been researched. Individual records from all five censuses have been used to contribute to research social mobility, and research has also investigated the effects of long-term exposure to air pollution. Research has provided evidence of impact for social policy issues, e.g. health inequalities and the State Pension Age Review.&#x0D; Discussion The main strength of the LS is its large sample size (&gt;1 million), making it the largest nationally representative longitudinal dataset in the UK. This allows analysis of small areas and specific population groups. Sampling bias is almost nil, and response rates are very high relative to other cohort and panel studies.&#x0D; Conclusion The ONS Longitudinal Study is a vital UK research asset, providing access to a large sample of census data linked across five censuses. It is strengthened through linkage to events data.

Inequalities in time from stopping paid work to death: findings from the ONS Longitudinal Study, 2001–2011

BackgroundUK state pension eligibility ages are linked to average life expectancy, which ignores wide socioeconomic disparities in both healthy and overall life expectancy.ObjectivesInvestigate whether there are occupational social class differences...

Religious Intermarriage in England and Wales: Differences in Individual and Area Characteristics of Endogamous and Exogamous Couples.

Religious intermarriage in England and Wales is an under-researched area, compared with ethnically divergent and immigrant/host intermarriage. More specifically, the aim of the study is to identify whether there are any statistically significant differences in personal or spatial characteristics between religiously endogamous and exogamous couples, and if so, to provide some explanation as to why. Use is made of the responses to the religion question in the 2001 and 2011 censuses and the Office for National Statistics (ONS) longitudinal study, so as to avoid bias inherent in trying to establish random survey samples of minor-religion members and their partners. Although the paper considers religious intermarriage generally, there is a focus on Jews as the only group which demonstrates a high level of stability of religion reporting and material levels of exogamy. Logistic regression indicates that Jewish population density, start decade of the partnership and whether the individuals are in work and live in an area with a large synagogue are statistically significant differentiators that have a material impact on levels of exogamy and endogamy. A strong link between exogamy and cohabitation (and delayed childbearing) is also established for this group. Some comparisons between Jews and religiously intermarried groups in general are also made, and the utility of the ONS longitudinal study for examination of small groups is established.

The parental co-immunization hypothesis: An observational competing risks analysis

The main interest of this study is the hypothesis that contact with small children may be beneficial for the parents’ later health and mortality (because of changes in their immune system). For this purpose, we document the relationship of a set of individual characteristics—including parenthood and marital state—and socioeconomic status with an individual’s cause of death. Using a novel and rich data set made available by the Office for National Statistics Longitudinal Study (ONS-LS), which follows 1% of the population of England and Wales along five census waves 1971, 1981, 1991, 2001, and 2011, our competing risks analysis yields several striking results: (1) Females with children have a 72.5% reduced risk of dying of cancer compared to childless females (for childless females of age 70, this corresponds to a risk of dying of cancer of 1.3% compared to a risk of about 1.3 × 0.275 = 0.4% for females with children). (2) Males have a 171% increased chance of dying of cancer when they are married (e.g., a baseline probability of 1.2% when 75 year old) compared to unmarried males. (3) Females with children have only a 34% risk of dying of heart disease (corresponding to a conditional probability of 0.3% when aged 65) relative to females without children and (4) a 53% chance of dying of infections (i.e., 0.1% at 65 years of age) compared to the risk for females without children. (5) At the same age, married men have an increased expectation of 123% of dying of heart disease (corresponding to an expected death probability of 0.7%) compared to unmarried men. (6) High income and house ownership is always associated with higher survival but less so than having children. While these results document a relationship between the presence of children and mortality, the specific transmission mechanisms remain unclear and we cannot make causality assertions.

The impact of youth labour market experiences on later employment opportunities: What roles do ethnicity and gender play?

Youth joblessness often leaves a scar. However, some ethnic groups appear to be more successful in recovering from this than others. Using a unique dataset (ONS Longitudinal Study) linking census records for a 1% sample of the population of England and Wales, we examine the relationship between early labour market experiences and later employment outcomes for men and women from Indian, Pakistani, Bangladeshi and Caribbean origins to those of white British individuals. Our results show that, on average, being unemployed or inactive in youth (vs being employed or in education) reduces employment opportunities later in life. However, this varies greatly by ethnicity and gender: Indian and, especially, Bangladeshi men are substantively less affected by previous non-employment compared with white British men; for women, having an ethnic minority background continues to limit their labour market integration. Addressing gender and ethnic labour market inequalities requires a more nuanced understanding of how these disadvantages unfurl over time for different communities.

Why did fewer people change address in England and Wales in the 2000s than in the 1970s? Evidence from an analysis of the ONS Longitudinal Study

AbstractThe paper explores why internal migration rates fell in England and Wales between 1971 to 1981 and 2001 to 2011. It considers all‐address changing rates, short‐distance migration (address changes involving moves of 10 km or less) and long‐distance migration (moves of 50 km or more). It does so by using the Blinder–Oaxaca method to decompose the differences between the 2 time periods into a compositional component (changes in the aggregate migration rate attributable to increases or decreases in more or less migratory groups of the population) and a rate component—the extent to which these population groups behave differently in the 2 time periods. The analysis finds that changing population structure since the 1970s does not fully account for the observed decreases. Instead, it seems that changed migratory behaviour and in particular, a large general downward effect that cuts across all population groups are the most important component. The greater decrease in address changing observed for all moves and those of less than 10 km (compared with longer distance moves) points the search for further explanations in the direction of factors known to be associated with short‐distance moves such as the housing market, and this is identified as an avenue for more research.

Let the Little Children Come to Me

We document the relationship of a set of individual choices - including parenthood, marital state, and income - with an individual’s cause of death. Using the data set of the Office for National Statistics Longitudinal Study (ONS-LS) which follows one percent of the population of England and Wales along five census waves 1971, 1981, 1991, 2001, and 2011, our competing risks analysis yields several striking results. 1) Females have only a 28% chance to die of cancer when they have children (compared to childless females); 2) males have a 71% increased chance of dying from cancer when they are married (compared to unmarried males); 3) females with children have only a 34% risk to die of heart disease and 4) a 53% chance of dying from infections (compared to females without children); 5) married men have an increased expectation of 23% to die of heart disease (compared to unmarried men); 6) high income and house ownership always is associated with higher survival but less so than having children.

Declining social mobility? Evidence from five linked censuses in England and Wales 1971-2011.

In this paper we add to the existing evidence base on recent trends in inter-generational social mobility in England and Wales. We analyse data from the Office for National Statistics Longitudinal Study (ONS-LS), which links individual records from the five decennial censuses between 1971 and 2011. The ONS-LS is an excellent data resource for the study of social mobility because it has a very large sample size, excellent population coverage and low rates of nonresponse and attrition across waves. Additionally, the structure of the study means that we can observe the occupations of LS-members' parents when they were children and follow their own progress in the labour market at regular intervals into middle age. Counter to widespread prevailing beliefs, our results show evidence of a small but significant increase in social fluidity between 1950s and the 1980s for both men and women.