Gender differences and individual, household, and workplace characteristics: Regional geographies of extended working lives.

Increasing labour market participation among older workers is embedded in government policy in the United Kingdom and many other industrialised countries with rises in the state pension age in response to increasing life expectancy. Despite this, many workers stop working before state pension age with around a 20% reduction in the proportion of adults in work between ages 50 and 60 in 2011 in England and Wales. This paper considers the risk of remaining in work by region and gender between 2001 and 2011 for adults aged 40-49 in 2001. Men had significantly higher risk of extended working in the East Midlands (1.4×) East of England (1.5×), South East (1.6×), and South West (1.6×) compared with the North East. Women in all regions apart from London and Wales had significantly higher risk of extended working compared with the North East: ranging from 1.15 times in the North West and West Midlands to 1.6 times in the South West. Adjustment for nonemployment-related socio-economic status, housing tenure, qualifications, and car ownership, and employment status in 2001 attenuated all significant regional differences in extended working in men and in women in most regions. Workplace characteristics attenuated most of the remaining regional differences in women: women working in larger employers in 2001 or working at distances of 200 km or more, abroad or from home, had lower risk of remaining in work, whereas access to a car and higher working hours increased risk. Policies to increase qualifications and skills among older adults are recommended.

GUILD: GUidance for Information about Linking Data sets.

Record linkage of administrative and survey data is increasingly used to generate evidence to inform policy and services. Although a powerful and efficient way of generating new information from existing data sets, errors related to data processing before, during and after linkage can bias results. However, researchers and users of linked data rarely have access to information that can be used to assess these biases or take them into account in analyses. As linked administrative data are increasingly used to provide evidence to guide policy and services, linkage error, which disproportionately affects disadvantaged groups, can undermine evidence for public health. We convened a group of researchers and experts from government data providers to develop guidance about the information that needs to be made available about the data linkage process, by data providers, data linkers, analysts and the researchers who write reports. The guidance goes beyond recommendations for information to be included in research reports. Our aim is to raise awareness of information that may be required at each step of the linkage pathway to improve the transparency, reproducibility, and accuracy of linkage processes, and the validity of analyses and interpretation of results.

Utilising identifier error variation in linkage of large administrative data sources.

BACKGROUND: Linkage of administrative data sources often relies on probabilistic methods using a set of common identifiers (e.g. sex, date of birth, postcode). Variation in data quality on an individual or organisational level (e.g. by hospital) can result in clustering of identifier errors, violating the assumption of independence between identifiers required for traditional probabilistic match weight estimation. This potentially introduces selection bias to the resulting linked dataset. We aimed to measure variation in identifier error rates in a large English administrative data source (Hospital Episode Statistics; HES) and to incorporate this information into match weight calculation. METHODS: We used 30,000 randomly selected HES hospital admissions records of patients aged 0-1, 5-6 and 18-19 years, for 2011/2012, linked via NHS number with data from the Personal Demographic Service (PDS; our gold-standard). We calculated identifier error rates for sex, date of birth and postcode and used multi-level logistic regression to investigate associations with individual-level attributes (age, ethnicity, and gender) and organisational variation. We then derived: i) weights incorporating dependence between identifiers; ii) attribute-specific weights (varying by age, ethnicity and gender); and iii) organisation-specific weights (by hospital). Results were compared with traditional match weights using a simulation study. RESULTS: Identifier errors (where values disagreed in linked HES-PDS records) or missing values were found in 0.11% of records for sex and date of birth and in 53% of records for postcode. Identifier error rates differed significantly by age, ethnicity and sex (p < 0.0005). Errors were less frequent in males, in 5-6 year olds and 18-19 year olds compared with infants, and were lowest for the Asian ethic group. A simulation study demonstrated that substantial bias was introduced into estimated readmission rates in the presence of identifier errors. Attribute- and organisational-specific weights reduced this bias compared with weights estimated using traditional probabilistic matching algorithms. CONCLUSIONS: We provide empirical evidence on variation in rates of identifier error in a widely-used administrative data source and propose a new method for deriving match weights that incorporates additional data attributes. Our results demonstrate that incorporating information on variation by individual-level characteristics can help to reduce bias due to linkage error.

Association between midlife health behaviours and transitions out of employment from midlife to early old age: Whitehall II cohort study.

BACKGROUND: It is important to determine whether unhealthy behaviours might influence transitions out of employment from midlife to old age, given the anticipated need for adults to work for longer. Our aim was to determine the association between repeated assessments of cigarette smoking, heavy/problem alcohol drinking, low physical activity and poor diet at midlife, in relation to work exit from midlife to old age. METHODS: Data from 7704 participants (5392 men) from the Whitehall II cohort study in employment at midlife were used to evaluate the association between unhealthy behaviours and a subsequent transition out of work during 22 years follow-up, using logistic regression models. RESULTS: Men who smoked cigarettes, consistently drank alcohol heavily, or reported problem drinking, were more likely to leave employment over follow-up. Women with a consistently poor diet were more likely to leave employment. Associations were stronger when the reason for leaving was health grounds, and stronger among those with persistently unhealthy behaviours over follow-up. The size of the effects were broadly equivalent to one advancing year of age on employment. Physical health functioning over follow-up only partly accounted for the associations with work exit, whereas physical and mental functioning accounted for most of the associations with work exit on health grounds. CONCLUSIONS: Unhealthy behaviours in midlife are associated with transitions out of employment into old age. Promoting healthy behaviours at midlife might support current policy initiatives aimed at extending working life. Future research should consider possible mechanisms that link behaviours to transitions out of employment, and consider sex differences in larger cohorts.

Sexual orientation and symptoms of common mental disorder or low wellbeing: combined meta-analysis of 12 UK population health surveys.

BACKGROUND: Previous studies have indicated increased risk of mental disorder symptoms, suicide and substance misuse in lesbian, gay and bisexual (LGB) adults, compared to heterosexual adults. Our aims were to determine an estimate of the association between sexual orientation identity and poor mental health and wellbeing among adults from 12 population surveys in the UK, and to consider whether effects differed for specific subgroups of the population. METHODS: Individual data were pooled from the British Cohort Study 2012, Health Survey for England 2011, 2012 and 2013, Scottish Health Survey 2008 to 2013, Longitudinal Study of Young People in England 2009/10 and Understanding Society 2011/12. Individual participant meta-analysis was used to pool estimates from each study, allowing for between-study variation. RESULTS: Of 94,818 participants, 1.1 % identified as lesbian/gay, 0.9 % as bisexual, 0.8 % as 'other' and 97.2 % as heterosexual. Adjusting for a range of covariates, adults who identified as lesbian/gay had higher prevalence of common mental disorder when compared to heterosexuals, but the association was different in different age groups: apparent for those under 35 (OR = 1.78, 95 % CI 1.40, 2.26), weaker at age 35-54.9 (OR = 1.42, 95 % CI 1.10, 1.84), but strongest at age 55+ (OR = 2.06, 95 % CI 1.29, 3.31). These effects were stronger for bisexual adults, similar for those identifying as 'other', and similar for 'low wellbeing'. CONCLUSIONS: In the UK, LGB adults have higher prevalence of poor mental health and low wellbeing when compared to heterosexuals, particularly younger and older LGB adults. Sexual orientation identity should be measured routinely in all health studies and in administrative data in the UK in order to influence national and local policy development and service delivery. These results reiterate the need for local government, NHS providers and public health policy makers to consider how to address inequalities in mental health among these minority groups.

Local area unemployment, individual health and workforce exit: ONS Longitudinal Study.

BACKGROUND: In many developed countries, associations have been documented between higher levels of area unemployment and workforce exit, mainly for disability pension receipt. Health of individuals is assumed to be the primary driver of this relationship, but no study has examined whether health explains or modifies this relationship. METHODS: We used data from 98 756 Office for National Statistics Longitudinal Study members who were aged 40-69 and working in 2001, to assess whether their odds of identifying as sick/disabled or retired in 2011 differed by local authority area unemployment in 2001, change in local area unemployment from 2001 to 2011 and individual reported health in 2001 (self-rated and limiting long-term illness). RESULTS: Higher local area unemployment and worse self-rated health measures in 2001 were independently related to likelihood of identifying as sick-disabled or retired, compared to being in work, 10 years later, after adjusting for socio-demographic covariates. Associations for local area unemployment were stronger for likelihood of identification as sick/disabled compared to retired in 2011. Associations for changes in local area unemployment from 2001 to 2011 were only apparent for likelihood of identifying as retired. For respondents that identified as sick/disabled in 2011, effects of local area unemployment in 2001 were stronger for respondents who had better self-rated health in 2001. CONCLUSIONS: Strategies to retain older workers may be most effective if targeted toward areas of high unemployment. For persons in ill health, local area unemployment interventions alone will not be as efficient in reducing their exit from the workforce.

Opportunities for longitudinal data linkage in Scotland.

Scotland has existing data resources which are competitive internationally and available to researchers from elsewhere. The Scottish Informatics and Linkage Collaboration (SILC) was recently launched, allowing data sets to be linked within and between sectors (e.g. health to non-health). The purpose of this review article is to introduce and define key terms in data linkage, to describe the emerging data linkage resources available in Scotland and to describe the opportunities available in Scotland to researchers internationally. The review is aimed at researchers internationally who are interested in data linkage using Scottish data resources. The review makes particular reference to longitudinal health data but emphasises that linkage to non-health data allows research questions to be considered that were previously not answerable. The review is focused on longitudinal data resources (e.g. cohort studies and repeated measures designs), since they are usually the focus of data linkage research. The review concludes that any intended data linkage for research should be driven by a clear research question. The infrastructure already available and the launch of SILC will accelerate research in Scotland and generate new research questions that previously could not be considered.

Reciprocal associations between smoking cessation and depression in older smokers: findings from the English Longitudinal Study of Ageing.

BACKGROUND: Depression is a particular problem in older people and it is important to know how it affects and is affected by smoking cessation. AIMS: To identify reciprocal, longitudinal relationships between smoking cessation and depression among older smokers. METHOD: Across four waves, covering six years (2002-2008), changes in smoking status and depression, measured using the 8-item Centre for Epidemiologic Studies Depression Scale, were assessed among recent ex-smokers and smokers (n = 2375) in the English Longitudinal Study of Ageing. RESULTS: In latent growth curve analysis, smoking at baseline predicted depression caseness longitudinally and vice versa. When both processes were modelled concurrently, depression predicted continued smoking longitudinally (B(ß) = 0.21 (0.27); 95% CI = 0.08-0.35) but not the other way round. This was the case irrespective of mental health history and adjusting for a range of covariates. CONCLUSIONS: In older smokers, depression appears to act as an important barrier to quitting, although quitting has no long-term impact on depression.

Association between questionnaire- and accelerometer-assessed physical activity: the role of sociodemographic factors.

The correlation between objective and self-reported measures of physical activity varies between studies. We examined this association and whether it differed by demographic factors or socioeconomic status (SES). Data were from 3,975 Whitehall II (United Kingdom, 2012-2013) participants aged 60-83 years, who completed a physical activity questionnaire and wore an accelerometer on their wrist for 9 days. There was a moderate correlation between questionnaire- and accelerometer-assessed physical activity (Spearman's r = 0.33, 95% confidence interval: 0.30, 0.36). The correlations were higher in high-SES groups than in low-SES groups (P 's = 0.02), as defined by education (r = 0.38 vs. r = 0.30) or occupational position (r = 0.37 vs. r = 0.29), but did not differ by age, sex, or marital status. Of the self-reported physical activity, 68.3% came from mild activities, 25% from moderate activities, and only 6.7% from vigorous activities, but their correlations with accelerometer-assessed total physical activity were comparable (range of r 's, 0.21-0.25). Self-reported physical activity from more energetic activities was more strongly associated with accelerometer data (for sports, r = 0.22; for gardening, r = 0.16; for housework, r = 0.09). High-SES persons reported more energetic activities, producing stronger accelerometer associations in these groups. Future studies should identify the aspects of physical activity that are most critical for health; this involves better understanding of the instruments being used.

Bidirectional association between mental health and physical activity in older adults: Whitehall II prospective cohort study.

OBJECTIVE: To investigate longitudinal and bidirectional associations between mental health and physical activity from midlife into old age. METHODS: Analysis was based on data from 6909 participants (aged 45 to 69 in 1997/99) from the Whitehall II cohort in the UK. Latent growth curve analysis examined possible bidirectional associations between the SF-36 Mental Component Summary and weekly physical activity measured at three time-points over ten years. RESULTS: Mental health and physical activity were associated at baseline (ß=0.17, 95% CI 0.13, 0.21) and associations persisted into old age. In the latent growth curve model, both mental health and physical activity increased and their rates of change 'moved together' over time (ß=0.24, 95% CI 0.11, 0.37). Relatively high baseline levels of either variable were associated with slightly slower increases in the other outcome (ß=-0.02, 95% CI -0.03, -0.01; ß=-0.07, 95% CI -0.11, -0.13), which are thought to reflect regression to the mean. However, those who started high on either variable remained the most advantaged at end of follow-up. CONCLUSIONS: From midlife to old age, greater physical activity is associated with better mental health and vice versa. These findings suggest persistent longitudinal and bidirectional associations between physical activity and mental health.

Combined impact of smoking and heavy alcohol use on cognitive decline in early old age: Whitehall II prospective cohort study.

BACKGROUND: Identifying modifiable risk factors for cognitive decline may inform prevention of dementia. AIMS: To examine the combined impact of cigarette smoking and heavy alcohol consumption on cognitive decline from midlife. METHOD: Prospective cohort study (Whitehall II cohort) with three clinical examinations in 1997/99, 2002/04 and 2007/09. Participants were 6473 adults (72% men), mean age 55.76 years (s.d. = 6.02) in 1997/99. Four cognitive tests, assessed three times over 10 years, combined into a global z-score (mean 0, s.d. = 1). RESULTS: Age-related decline in the global cognitive score was faster in individuals who were smoking heavy drinkers than in non-smoking moderate alcohol drinkers (reference group). The interaction term (P = 0.04) suggested that the combined effects of smoking and alcohol consumption were greater than their individual effects. Adjusting for age, gender, education and chronic diseases, 10-year decline in global cognition was -0.42 z-scores (95% CI -0.45 to -0.39) for the reference group. In individuals who were heavy alcohol drinkers who also smoked the decline was -0.57 z-scores (95% CI -0.67 to -0.48); 36% faster than the reference group. CONCLUSIONS: Individuals who were smokers who drank alcohol heavily had a 36% faster cognitive decline, equivalent to an age-effect of 2 extra years over 10-year follow-up, compared with individuals who were non-smoking moderate drinkers.

Trait anxiety levels before and after antidepressant treatment: a 3-wave cohort study.

The aim of this study was to examine change in "trait anxiety" levels assessed repeatedly before and after antidepressant treatment in a large cohort of men and women. A total of 18,732 participants of the Finnish Public Sector Study with no initial record of depression or self-reported doctor diagnosis of depression completed the short form of the Spielberger Trait-Anxiety Inventory in 2000-2002 (T1), 2004-2005 (T2), and 2008-2009 (T3). We used prescription data from the nationwide Drug Prescription Register to identify antidepressant treatment between T1 and T2 (n = 710). Both men (ß = 0.435, P < 0.001) and women (ß = 0.300, P < 0.001) who received antidepressant treatment had higher trait anxiety levels at T1. Mixed models analyses of repeated measures showed a small but statistically significant decrease in trait anxiety scores for the overall sample of men (ß = 0.023, P = 0.033) and women (ß = 0.011, P = 0.031) between T1 and T3. The interaction term between time and antidepressant treatment status suggested a greater decrease in trait anxiety levels among men receiving antidepressant treatment, with an adjusted excess decrease in mean trait anxiety scores of 0.163 (P = 0.012) between T1 and T3. We found some evidence suggesting that this is also the case in women, although the evidence in our data was less consistent for women. This large-scale study provides evidence suggesting that antidepressant treatment is associated with a reduction in trait anxiety levels, particularly in men.

Cigarette smoking and alcohol drinking in a representative sample of English school pupils: cross-sectional and longitudinal associations.

OBJECTIVE: The aim of our study was to examine cross-sectional and longitudinal associations between cigarette smoking and alcohol drinking, in a representative sample of English pupils. METHOD: Data from 13,635 school pupils in the Longitudinal Study of Young People in England (LSYPE) on usage of cigarettes from 2004 (typical age 14) to 2006 (age 16) and alcohol from 2004 to 2007 (age 17), analyzed with latent growth curve models. RESULTS: The weighted percentage of pupils drinking alcohol increased from 26% at age 14 to 71% by age 17, smoking from 12% to 27% by age 16. Pupils with lower socio-economic status were more likely to smoke but less likely to drink alcohol regularly. Both behaviors were positively correlated at age 14, adjusted for several confounding factors. The rate of increase over time was also positively correlated. CONCLUSION: Cigarette smoking and alcohol drinking are already correlated by age 14, are socio-economically patterned, and 'move together' during adolescence. Future studies and interventions should be targeted at a younger age range, to identify early smoking and potentially hazardous alcohol drinking patterns.

Chronic inflammation as a determinant of future aging phenotypes.

BACKGROUND: The importance of chronic inflammation as a determinant of aging phenotypes may have been underestimated in previous studies that used a single measurement of inflammatory markers. We assessed inflammatory markers twice over a 5-year exposure period to examine the association between chronic inflammation and future aging phenotypes in a large population of men and women. METHODS: We obtained data for 3044 middle-aged adults (28.2% women) who were participating in the Whitehall II study and had no history of stroke, myocardial infarction or cancer at our study's baseline (1997-1999). Interleukin-6 was measured at baseline and 5 years earlier. Cause-specific mortality, chronic disease and functioning were ascertained from hospital data, register linkage and clinical examinations. We used these data to create 4 aging phenotypes at the 10-year follow-up (2007-2009): successful aging (free of major chronic disease and with optimal physical, mental and cognitive functioning), incident fatal or nonfatal cardiovascular disease, death from noncardiovascular causes and normal aging (all other participants). RESULTS: Of the 3044 participants, 721 (23.7%) met the criteria for successful aging at the 10-year follow-up, 321 (10.6%) had cardiovascular disease events, 147 (4.8%) died from noncardiovascular causes, and the remaining 1855 (60.9%) were included in the normal aging phenotype. After adjustment for potential confounders, having a high interleukin-6 level (> 2.0 ng/L) twice over the 5-year exposure period nearly halved the odds of successful aging at the 10-year follow-up (odds ratio [OR] 0.53, 95% confidence interval [CI] 0.38-0.74) and increased the risk of future cardiovascular events (OR 1.64, 95% CI 1.15-2.33) and noncardiovascular death (OR 2.43, 95% CI 1.58-3.80). INTERPRETATION: Chronic inflammation, as ascertained by repeat measurements, was associated with a range of unhealthy aging phenotypes and a decreased likelihood of successful aging. Our results suggest that assessing long-term chronic inflammation by repeat measurement of interleukin-6 has the potential to guide clinical practice.

Neuroticism and cardiovascular disease mortality: socioeconomic status modifies the risk in women (UK Health and Lifestyle Survey).

OBJECTIVE: The association between personality traits and mortality might differ as a function of socioeconomic status (SES). Our aim was to evaluate the all-cause, cardiovascular disease (CVD), and cancer mortality risk associated with neuroticism or extraversion and their interactions with SES in a representative sample of the UK adult population. METHODS: A total of 5450 participants (2505 men) from the Health and Lifestyle Survey completed the Eysenck Personality Inventory at baseline and were monitored for vital status over 25 years. SES was defined as a latent variable comprising occupational social class, educational attainment, and income. RESULTS: A significant neuroticism-by-SES-by-sex interaction (p = .04) for CVD mortality revealed a neuroticism-by-SES interaction specific to women. Compared to women with average SES, those with both high neuroticism and low SES were at an increased risk for CVD mortality (hazard ratio = 2.02, 95% confidence interval = 1.45-2.80), whereas those with high neuroticism and high SES combined were at a decreased risk for CVD mortality (hazard ratio = 0.61, 95% confidence interval = 0.38-0.97, p for interaction = 0.003). The interaction term was not explained by health behaviors (10% attenuation) and physiological variables (11% attenuation). This interaction was not observed for all-cause and cancer mortality risks or among men for CVD mortality. CONCLUSIONS: High neuroticism is a risk factor for cardiovascular mortality in women with low SES, whereas in women with higher SES, it is protective. Further research is needed to replicate this finding and identify the mechanisms behind the modifying effect of SES on neuroticism.

Influence of individual and combined healthy behaviours on successful aging.

BACKGROUND: Increases in life expectancy make it important to remain healthy for as long as possible. Our objective was to examine the extent to which healthy behaviours in midlife, separately and in combination, predict successful aging. METHODS: We used a prospective cohort design involving 5100 men and women aged 42-63 years. Participants were free of cancer, coronary artery disease and stroke when their health behaviours were assessed in 1991-1994 as part of the Whitehall II study. We defined healthy behaviours as never smoking, moderate alcohol consumption, physical activity (≥ 2.5 h/wk moderate physical activity or ≥ 1 h/wk vigorous physical activity), and eating fruits and vegetables daily. We defined successful aging, measured over a median 16.3-year follow-up, as good cognitive, physical, respiratory and cardiovascular functioning, in addition to the absence of disability, mental health problems and chronic disease (coronary artery disease, stroke, cancer and diabetes). RESULTS: At the end of follow-up, 549 participants had died and 953 qualified as aging successfully. Compared with participants who engaged in no healthy behaviours, participants engaging in all 4 healthy behaviours had 3.3 times greater odds of successful aging (95% confidence interval [CI] 2.1-5.1). The association with successful aging was linear, with the odds ratio (OR) per increment of healthy behaviour being 1.3 (95% CI 1.2-1.4; population-attributable risk for 1-4 v. 0 healthy behaviours 47%). When missing data were considered in the analysis, the results were similar to those of our main analysis. INTERPRETATION: Although individual healthy behaviours are moderately associated with successful aging, their combined impact is substantial. We did not investigate the mechanisms underlying these associations, but we saw clear evidence of the importance of healthy behaviours for successful aging.

Direct and indirect pathways connecting cognitive ability with cardiovascular disease risk: socioeconomic status and multiple health behaviors.

OBJECTIVE: To model and test direct and indirect pathways connecting general cognitive ability (g) with cardiovascular disease risk factors, via socioeconomic status (SES) and multiple health behaviors. METHODS: A sample comprising participants in the Health and Lifestyle Survey, a prospective cohort study of a representative sample of U.K. adults in 1984/5 (n = 4939, 2426 males). RESULTS: Two mediating latent variables were proposed that connected a latent cognitive trait (named g) with a latent trait of cardiovascular disease (CVD) risk: multiple health behaviors (defined by smoking, physical inactivity, and weekly saturated fat intake) and SES (defined by educational attainment, occupational social class, and income). In males and females, SES mediated the association between g and CVD risk, but the mediation was moderated by years of age. A direct effect from g to CVD risk was also significant, but this was restricted to older males. Multiple health behaviors offered no explanatory power, because they were not influenced by g. CONCLUSIONS: SES may connect g with CVD risk in males, but not systematically across the life course. Moderated mediation is a novel way to illustrate that direct and indirect pathways can vary as a function of age. Explanations that emphasize g or SES are not mutually exclusive; there are direct and indirect contributions to CVD risk from each source, and these vary across the life course.

The SF-36 component summary scales and the daytime diurnal cortisol profile.

PURPOSE: Higher daytime cortisol output has been associated with higher levels of perceived stress and worse mental and physical health outcomes. Hypothalamic-pituitary-adrenal (HPA) axis dysregulation, such as elevated secretion of daytime cortisol, occurs in many mental and physical illnesses. However, the nature of the association between functional health status and daytime cortisol production has not been established. METHODS: Healthy adult volunteers (n=68, 45 females) provided saliva samples 3, 6, 9 and 12 h after waking, for two consecutive days, in everyday settings. Bivariate correlations between log salivary cortisol concentrations were calculated, and the SF-36 component summary scores were calculated. Latent growth curve modeling was used to model the daytime profile and adjust for covariates (age, sex and waking time). RESULTS: Higher PCS scores were not associated with cortisol three hours after waking (cortisol intercept), or the diurnal decline (cortisol slope). Higher MCS scores were correlated with faster cortisol decline across the day (r=-.31, P<.01) but not with cortisol intercepts. In a latent growth curve model adjusting for age, sex and waking time, the association was no longer statistically significant. CONCLUSIONS: Large scale epidemiological studies involving salivary cortisol would benefit from measuring SF-36 component summary scores. Cortisol intercepts and slopes may be differentially related to the PCS and MCS, although greater statistical power is needed to test this hypothesis more fully. Associations between daytime cortisol and the PCS or MCS could reflect the regulatory competence of bodily systems, common causes or unmeasured confounding factors.