Simulating the impact of poverty reduction on social inequalities in adolescent mental health in the UK: a causal mediation analysis using the UK Millennium Cohort Study

Background

Adolescent mental health is poor in the UK, with higher prevalence of poor mental health in adolescents living in poverty. However, little experimental evidence exists to understand the potential impact of poverty reduction on inequalities in mental health in the UK population. We aimed to fill this gap by assessing the effect of poverty reduction with a hypothetical intervention on adolescent mental health.

Associations between measures of child poverty,health-care use,and health outcomes: observational study of nationally representative English data

BackgroundThe UK Government has recently proposed the abolition of income-based measures of child poverty in favour of environmental, educational, and employment measures of deprivation. We aimed to study how strongly these proposed measures are associated with health outcomes among children and young people compared with a relative income measure.MethodsWith data from Hospital Episode Statistics for 2013–14 (n=16·4 million), we compared inequalities in inpatient admissions of children and young people aged 0–24 years per thousand (total and chronic conditions) using five deprivation measures (deciles of area-based measures: index of multiple deprivation [IMD], income, index of income deprivation affecting children [IDACI], education, living environment). With data from Health Survey for England for 2014 (n=3085), we compared inequalities in general health and long-standing illness reported by an individual, parent, or carer using equivalised household income quintiles, IMD quintiles, and whether the household reference person was employed (n=2417, 78·3%); not working (397, 12·9%); or retired, long-term sick, or other (274, 8·9%). The ratios of hospital admission rates and weighted prevalence of poor self-reported health were compared between the most and least deprived groups within each measure.FindingsTotal admission rates were higher among the most versus least deprived IMD deciles (ratio 1·60, 95% CI 1·59–1·61). The ratio was greater when income deciles were compared (1·69, 1·68–1·71) and smaller when analysis was by education (1·59, 1·58–1·60), IDACI (1·52, 1·51–1·53), and living environment (1·01, 1·00–1·02). The ratio was lower for admissions for chronic conditions (IMD 1·18, 1·16–1·20). Inequalities were largest when analysis was by income decile (1·25, 1·22–1·27). In the household-level analyses, inequalities in fair or poor, self or parent-reported health were seen when comparing lowest versus highest income quintiles (12·2% vs 3·9; ratio 3·12, 1·99–5·87), not working versus being employed (14·1 vs 6·1; 2·30, 1·68–3·03), and most versus least deprived IMD quintiles (9·8 vs 6·5; 1·49, 1·05–2·21). For long-standing illness, the equivalent data were: income (22·0 vs 11·0; 1·99, 1·50–2·77), employment (23·6 vs 15·8; 1·50, 1·21–1·81), and IMD (18·9 vs 17·2; 1·10, 0·88–1·39).InterpretationAlthough some important aspects of health such as use of primary care and community services are not included, this study shows that child poverty measures differ significantly in their association with key indicators of population health and health-care use. Of the deprivation measures studied here, hospital admissions were most strongly associated with income inequality. Self-reported health outcomes were also more strongly linked to household income than area-level IMD.FundingNone.     

Trends in mental health and wellbeing among children and young people in the UK: a repeated cross-sectional study, 2000–14

BackgroundIncreasing concerns have been raised by professionals in education, health, and other sectors that mental health and wellbeing among children and young people in the UK might be deteriorating, but few nationally representative studies have tested this hypothesis. The objective of this study was to investigate trends in mental health and wellbeing among participants aged 4–24 years in UK national health surveys, 2000–14.MethodsWe used data from national health surveys of four UK countries: England (15 surveys, n=67 386, unweighted), Scotland (8, 16 862), Wales (8, 17 677), and Northern Ireland (1, 339). Trends were evaluated for children aged 4–12 years with the Strengths and Difficulties Questionnaires (SDQ) (parent or carer report) and for young people aged 16–24 years with the General Health Questionnaire (GHQ12) and Warwick Edinburgh Mental Health and Wellbeing Scores (WEMWBS) (self-report for both instruments). We assessed changes over time within countries using weighted t tests of the earliest and latest data for each country and linear regression models using all data.FindingsThere were no significant changes in SDQ scores in England, Wales, or Scotland. The proportion of Scottish parents reporting emotional problems was lower in 2014 than in 2003 (weighted proportions 43/908, 4·7% [95% CI 3·3–6·1] vs 175/1819, 9·6 [8·3–11·0]; p<0·001), but no significant overall trend in any country was seen. According to the GHQ scores, prevalence of mental health problems was higher in Scotland in 2014 than in 2003 (103/570, 18·1% [14·9–22·1] vs 117/931, 12·6 [10·4–14·7]; p=0·004), but unchanged in England. Regression analyses showed no significant trend. Wellbeing scores in England measured by WEMWBS were unchanged but were lower in Scotland in 2014 than in 2008 (49·5 [48·9–50·2] vs 50·5 [50·0–51·0], p=0·04). There was no significant overall trend.InterpretationPrevalence of mental health problems among children and young people has been largely stable in England and Wales over the past 14 years. In Scotland, no significant linear trends were identified, but the most recent data show fewer emotional problems in younger children along with more mental health problems, and decreased wellbeing among young adults.FundingNone.     

Change in maternal smoking behaviour between the first two singleton live pregnancies and childhood obesity: analysis of a UK population-based cohort

BackgroundMaternal smoking during pregnancy has been linked to offspring adiposity. We examined interpregnancy changes in maternal smoking behaviour and the risk of age-specific and sex-specific obesity (≥95^th centile) in the second child (C2).MethodsWe used a population-based cohort of antenatal health-care records (August, 2004–August, 2014) at University Hospital Southampton, linked to measured child body-mass index (BMI) at 4–5 years obtained from child health records at two community NHS Trusts (Solent and Southern). We analysed the first two singleton live pregnancies of 6515 women using logistic regression to examine interpregnancy changes in self-reported maternal smoking in relation to C2 obesity (adjusting for maternal age, ethnicity, BMI, educational attainment, employment, folate supplementation, previous losses, infertility treatment, pre-existing and gestational diabetes and hypertension, interpregnancy interval, C2's birthweight, caesarean section delivery, and gestation).FindingsUnadjusted C2 obesity prevalence for children whose mothers never smoked, smoked at the start of both pregnancy 1 (P1) and pregnancy 2 (P2), P2 only, P1 only, and stopped smoking when both pregnancies were confirmed was 6·9% (215/3118 children), 12·5% (87/698 children), 12·4% (22/178 children), 10·3% (28/271 children), and 13·1% (29/222 children), respectively. Compared with women who never smoked, smoking at the start of both pregnancies was associated with higher odds of C2 obesity (adjusted odds ratio [aOR] 2·03, 95% CI 1·49–2.78). Women not smoking in P1 who smoked at the start of P2, and those who stopped when each pregnancy was confirmed had increased C2 obesity odds (aOR 1·80 [95% CI 1·09–3·00] and 1·73 [1·11–2·69], respectively). Smokers maintaining cessation having quit by P1 confirmation, P1 smokers who ceased by P2 confirmation, and those who smoked between pregnancies but who quit before C2 conception, did not have higher odds of C2 obesity (aOR 1·21 [95% CI 0·95–1·55], 1·39 [0·89–2·17], and 1·13 [0·78–1·63], respectively).InterpretationA mother smoking at the start of her first two pregnancies has twice the odds of having an obese second child compared with a non-smoker. Smoking in the second pregnancy only and between pregnancies up to the first trimester of P2 is also associated with childhood obesity. The interpregnancy period is an opportunity to intervene on modifiable risk factors such as smoking.FundingNIHR Southampton Biomedical Research Centre and University of Southampton Primary Care and Population Sciences PhD studentship (to EJT) and an Academy of Medical Sciences and Wellcome Trust grant (grant number AMS_HOP001\1060; to NAA).     

Neighbourhood poverty and psychological distress in adolescents aged 11–15 in China: a multilevel study of 507 local communities

Background

Neuropsychiatric conditions have become the leading cause of disability in adolescents aged 11–24 globally. Higher neighbourhood economic status was found to be associated with lower incidence of neuropsychiatric conditions in children and adolescents in developed countries, but there is a lack of evidence in China. This study aims to examine the associations between neighbourhood poverty and psychological distress among adolescents in China.

What predicts food insecurity? An online survey

BackgroundFood poverty (inability to afford or access a healthy diet) is a public health emergency, manifesting as a short-term dilemma of accessing food alongside longer-term effects of relying on poor nutritional quality foods to satiate hunger and worrying about food running out because of lack of money. The study aimed to predict what factors might predispose people to this condition by enquiring about respondents’ food poverty experiences.MethodsWe did an online survey (from September to November, 2018) to measure food poverty experiences alongside demographics and self-reported health evaluation in the UK (n=944 respondents). We used a snowball sampling technique, and the researchers and stakeholders promoted the survey by acting as gatekeepers on social media. Complementary paper surveys provided remote access to the research. Inclusion criteria required respondents be the primary householder or main earner (and be aged 18 years or older). We did a binary logistic regression analysis to test if age (18–25, 26–35, 36–45, 46–55, 56–65, ≥66 years old), gender (male, female, other), location (urban, rural), household size, number of children (<18 years old), income (<£10 000, £10 000–19 999,….£100 000–150 000, ≥£150 000), home ownership status (owned, rented), employment (unemployed, employed), education, and self-reported health status predicted worry about running out of food in the past 12 months (Food and Agriculture Organization's Food Insecurity Experience Scale). Ethics approval was granted, and participants gave implied informed consent.FindingsUnweighted case summaries indicated slightly more male respondents (51·8%), and one in five (21·3%) respondents were economically inactive. 8% had a total annual household income of less than £10 000. One in 14 (7·4%) self-reported poor health status. Two-fifths (41·9%) had children living at home. Importantly, 24% of the sample worried about food running out. The model (χ²_{(10, N=944)}=155·158, p<0·0001) significantly distinguished between households that were and were not food poor. Respondents who self-reported poorer health status were more likely to be in food-poor households (odds ratio [OR]=2·10, 95% CI 1·61–2·72; p<0·0001) as were those who had more children (1·52, 1·20–1·93; p=0·001); but food poverty was less likely to be reported in respondents who were older (0·73, 0·58–0·92; p=0·007), in those who owned their property (0·45, 0·25–0·81; p=0·008), and in those who had higher household incomes (0·74, 0·62–0·88; p=0·001). The associations between food poverty and the remaining variables were not significant.InterpretationPersonal and household circumstances can predict food poverty. Given future public health consequences, these predictors could usefully inform targeted interventions for risk profiling vulnerable citizens.FundingUlster University Civic Impact Research Fund.     

Interplay of takeaway food outlet exposure and income on diet and obesity: a cross-sectional study in UK Biobank

BackgroundThe proliferation of takeaway food outlets has been called a public health problem by Public Health England and the Chief Medical Officer for England. Evidence indicates that the density of takeaway outlets in neighbourhoods can promote unhealthy eating and obesity. But takeaway outlets tend to cluster in deprived areas, and vulnerability to unhealthy food environments might vary across social groups. This study examined the interplay between exposure to takeaway outlets and individual-level socioeconomic position in relation to diet and obesity.MethodsThis cross-sectional study included UK Biobank participants based in Greater London (n=51 361) aged 38–72 years with valid height and weight data, food frequency data, and food environment exposures. Food environment exposures were estimated using geographic information systems and food outlet data from UKMap, as counts of outlets by type, within one-mile radius Euclidean (circular) buffers centred on each participant's home address. Our primary exposure was the number of takeaway outlets as a proportion of all food outlets. Participants were jointly classified on the basis of household income (four groups) and quartile of neighbourhood takeaway outlet proportion to create 16 exposure groups. Regression models estimated the odds of frequent processed meat consumption and obesity. We calculated relative excess risk due to interaction (RERI), relative to a single reference group. Participant data were collected during 2005–13 and analysed in 2016.FindingsBoth income and takeaway exposure were systematically associated with frequent consumption of processed meat and obesity. In mutually adjusted models, lowest-income participants were more likely than highest-income participants to be obese (odds ratio 1·53, 95% CI 1·38–1·69); and relative to those least exposed, highest takeaway exposure was also associated with obesity (1·76, 1·61–1·91). In additive models, the combination of lowest income and highest takeaway exposure was associated with substantially increased odds of obesity (2·75, 2·33–3·24), with the RERI indicating significant interaction. Results were similar in diet analyses.InterpretationWe have confirmed the findings of earlier work showing independent associations between income, neighbourhood takeaway outlet exposure, and diet and adiposity. Moreover, we provide evidence of the double burden of low income and an unhealthy neighbourhood food environment, resulting in higher rates of unhealthy diet and obesity.FundingBritish Heart Foundation, Cancer Research UK, Economic and Social Research Council, Medical Research Council, National Institute for Health Research, and the Wellcome Trust, under the auspices of the UK Clinical Research Collaboration.     

Understanding social inequalities in childhood hospital admissions: longitudinal analysis of the UK Millennium Cohort Study

BackgroundChildhood hospital admission rates are increasing annually and are socially patterned. To inform policies to reduce hospital admissions in children, we aimed to assess risk factors for hospital admissions and the extent to which any socioeconomic inequality in admissions could be attenuated after accounting for risk factors.MethodsThis analysis used a sample of 11 085 children, across five sweeps from ages 9 months to 11 years, from the UK Millennium Cohort Study. The Millennium Cohort Study oversamples children living in disadvantaged areas and ethnic minority groups by means of a stratified clustered sampling design, and includes sampling weights to address attrition. The primary outcome for the present analysis was parental report of a child “ever being admitted to a hospital ward” (ever admitted) by age 11 years. Parents' educational attainment at the child's birth was the main exposure. Primary outcome data were analysed with Poisson regression, according to parental education, adjusted for covariates including perinatal risk factors, maternal and child health status, and environmental risk factors.Findings4615 parents (45·6%) reported that their child had been admitted to hospital by age 11 years. Children of parents with no educational qualifications were more likely than those of parents with degree level or higher qualifications to have been admitted (risk ratio 1·49, 95% CI 1·20–1·85), and there was evidence of a dose-response association. Controlling for risk factors attenuated the increased risk of admissions in children from households with the lowest educational attainment (adjusted risk ratio 1·09, 95% CI 0·93–1·29). In the fully adjusted analysis, low birthweight (1·24, 1·06–1·45) and longstanding illnesses (1·21, 1·18–1·24) were associated with an increased risk of being ever admitted, and female sex was protective (0·81, 0·75–0·88). Repeating the analysis with the main exposure as income produced similar results.InterpretationIn a contemporary representative sample of children in the UK, nearly half were admitted to hospital by age 11 years. Children of parents with no qualifications had around 50% greater risk of admission, largely explained by differences in childhood longstanding illness and low birthweight. Policies to reduce inequalities in these factors might also reduce inequalities in hospital admissions. Study limitations include the potential for differential self-reporting of hospital admission from households with different educational attainment, and not being able to explore service model differences.FundingNone.     

Maternal adversity and variation in the rate of children entering local authority care during infancy in England: a longitudinal ecological study

BackgroundEngland has seen unsustainable increases in the number of children entering care in recent years. One in five children entering care are infants, although rate of infant entry to care varies considerably by local authority. Understanding this variation and its drivers could inform upstream strategies to improve outcomes for children at risk of, or receiving, care. Parental ill health or risk behaviour from pregnancy onwards could trigger state involvement in infants’ care. Therefore, we explored whether adversity indicated within women's hospitalisation history, pre-delivery, could help to explain local authority variation in rates of infant entry to care.MethodsWe combined data on hospitalisations (hospital episode statistics) and entries to care (children looked after return) with publicly available data on local authority characteristics (2011 Census, Office for National Statistics Population Estimates, and Public Health England Fingertips) for 131 English local authorities. Our primary outcome was annual (April–March) local authority rate of infant entry to care (per 10 000 infants in the local authority population) between April 1, 2006, and March 31, 2014. We used linear mixed-effects models to analyse the relationship between the outcome and local authority-specific proportion of livebirths with maternal history of adversity-related hospital admissions (ie, substance misuse, mental health problems, or violence-related admissions in the 3 years before delivery), adjusted for seven other predictors for entry to care (including local authority-specific prevalence of maternal deprivation, births to teenage mothers, and community violence).FindingsRate of infant entry to care (mean 85·16 per 10 000 [SD 41·07, range 0·00–318·51) and proportion of livebirths with maternal history of adversity-related hospital admissions (mean 4·62% [2·44, 0·52–16·19]) varied greatly by local authority. Prevalence of maternal adversity accounted for 24% (95% CI 14–35%) of variation in rate of entry. After adjustment, a percentage point increase in prevalence of maternal adversity (both within and between local authorities) was associated with an additional 2·56 infants (95% CI 1·31–3·82) per 10 000 entering care.InterpretationPrevalence of maternal adversity before birth helps explain differing rates of infant entry to care among English local authorities. This study also highlights the need for linked parent–child data on health and child protection to inform policy interventions to improve maternal wellbeing and potentially reduce entries to care.FundingNational Institute for Health Research—Great Ormond Street Hospital Biomedical Research Centre.     

Maternal depression and risk of injuries in children aged 0–4 years: a population-based cohort study

BackgroundThe association between maternal depression and childhood injuries is underexplored, with existing studies relying on maternal reporting of injury occurrences. We aimed to assess the association between the incidence of three common childhood injuries and exposure to maternal antenatal depression, postnatal depression, or both.MethodsUsing a cohort of 209 418 mother–child pairs who had linked data from the Clinical Practice Research Datalink and Hospital Episode Statistics for 1997–2014, we estimated incidence of poisonings, fractures, and burns per 10 000 person-years from birth to the child's fifth birthday, stratified by whether the mother had antenatal depression, postnatal depression, or both. One child was randomly selected per mother to prevent clustering. Using Poisson regression, we estimated incidence rate ratios (IRRs), adjusted for maternal age at delivery, socioeconomic status, number of older siblings, and number of children aged 0–4 years in the household. The study was approved by the Independent Scientific Advisory Committee for the Medicines and Healthcare products Regulatory Agency in February, 2014.FindingsCrude injury rates were higher for each injury type among children whose mothers had antenatal depression, postnatal depression, or both. Poisoning incidence per 10 000 person-years was 59·6 (95% CI 48·7–72·8) for children of mothers with antenatal depression, 64·4 (57·4–72·2) for children of mothers with postnatal depression, and 74·2 (63·2–87·1) for children of mothers with both, compared with 36·3 (34·8–37·8) for those unexposed to either antenatal or postnatal depression. Similar patterns were seen for fractures and burns. After adjustment, IRRs remained significant for poisonings and burns, with the greatest risk for children whose mothers had both antenatal and postnatal depression (poisonings 1·94, 95% CI 1·63–2·32; burns 1·32, 1·13–1·55). However, there was no significant association between fractures and perinatal depression.InterpretationAntenatal and postnatal depression were associated with a higher risk of child poisonings and burns, suggesting that maternal depression could be a modifiable risk factor for these types of injury. By using routinely collected health data we only identified women who sought medical attention for depression, and were unable to adjust for other variables such as child behaviour. Differences in the ascertainment of injury events between mothers with and without perinatal depression could lead to an overestimation of injury risk. Further work could include assessing certain injuries (eg, long-bone fractures) for which ascertainment is likely to be almost complete.FundingRB is funded by the National Institute for Health Research (NIHR) School for Primary Care Research and the University of Nottingham.     

The associations between social factors and suboptimal use of paediatric dental services (Tiny Teeth): a mixed method study

BackgroundDental decay is the leading reason for hospital admission of children aged 5–9 years. Prevalence of dental decay is strongly associated with deprivation, decreased maternal education, and ethnicity. This study aimed to investigate social and parental factors that prevent timely and regular use of dental services in Haringey and Enfield (London, UK), and to address the suitability of the emergency department (ED) as a site of opportunistic oral health education, using a dental intervention (DI).MethodsThe study was undertaken at the North Middlesex University Hospital's paediatric ED in North London, UK. It was a mixed method study with cross-sectional surveys, a brief DI, and qualitative interviews. Parents of children aged 10 years and younger in the ED were invited to participate in the study. Informed written consent was obtained from 101 participants. Data collection was from Jan 31, 2019, to March 13, 2019. Questionnaires collected background social and medical data, and assessed each caregiver's oral health literacy. The Big Bites and Pearly Whites DI provided parents with up-to-date information about teeth cleaning practices, dentistry, and diet. A post-intervention questionnaire assessed changes in parental knowledge. Adjusted logistic regressions explored associations between social factors and child dental registration. Seven semistructured interviews were conducted, collecting data on barriers faced to the use of paediatric dental services.FindingsBlack British, Caribbean, and African children had an odds ratio of 0·158 (95% CI 0·037–0·679) for dental registration compared with their white counterparts, after adjusting for age of child and sibling number. The DI led to a significant difference between pre and post intervention median knowledge scores (Wilcoxon signed rank test; p=0·009) with a small effect size (r=0·2). Qualitative interviews exposed that current oral health advice, from dentists alone, is inconsistent and not universally delivered, resulting in late, suboptimal, and problem-based attendance.InterpretationTo maximise and reinforce early and universal parental receipt of oral health messages, collaboration and coordination is needed between educational, dental, and general health providers. Disparities in health-seeking behaviour due to ethnicity and knowledge should be accounted for when planning service engagement programmes. The study supports the ED as a site to opportunistically make every contact count by reinforcing preventive oral health messages. The major limitations of the study are its size and cross-sectional design.FundingNone.     

Domestic violence and abuse in gay and bisexual men attending a sexual health clinic: pilot testing of an educational support intervention for sexual health practitioners

BackgroundDomestic violence and abuse is recognised globally as a pervasive public health issue, but there is less research in relation to gay and bisexual men. This study aimed to measure the occurrence of negative and potentially abusive behaviour and associations with health problems in gay and bisexual men; and pilot test an educational intervention for practitioners to increase awareness of domestic violence and abuse and provide skills in inquiry and support.MethodsBetween Sept 21, 2010, and May 25, 2011, an anonymous Health and Relationships survey was administered to English-literate men (aged 18 years) attending a sexual health clinic in London, UK. Negative behaviour included: feeling frightened of the behaviour of a partner; having to ask permission (beyond being considerate to and checking with a partner); being slapped, hit, kicked, or physically hurt; and forced sex or sexual activity. Health problems included mental health, sexual health, and alcohol and illicit drug use. Sexual health practitioners received training on the educational intervention HERMES (HEalthcare Responding to MEn for Safety). 6 weeks after training, 21 semi-structured interviews were conducted with the practitioners to assess acceptability of HERMES, changes in awareness, and confidence in identifying and responding to gay and bisexual men who were experiencing domestic violence and abuse.FindingsOf 2657 men who attended the clinic, 1132 (42·4%) completed the survey. 532 (47·2%) self-identified as gay or bisexual. 33·9% (95% CI 24·9–37·9) of gay and bisexual men reported ever experiencing a negative and potentially abusive behaviour from a partner. Ever being frightened by a partner and having to ask a partner's permission were associated with increased odds of being anxious (odds ratio [OR] 2·5, 95% CI 2·0–3·1 and 2·7, 1·6–4·7, respectively). Being frightened of a partner, being physically hurt, and physically hurting a partner were associated with increased odds of using a class A drug in the previous 12 months (OR 2·2, 1·5–3·2; 2·3, 1·4–3·8; and 3·1, 2·3–4·2, respectively). Practitioners welcomed HERMES and felt that it raised their awareness and confidence in dealing with domestic violence and abuse among gay and bisexual men.InterpretationIn this clinical setting the occurrence of potentially abusive behaviours in gay and bisexual men was high. HERMES increased sexual health practitioners’ awareness of domestic violence and abuse and confidence in asking about abuse. The results can only be interpreted within the context of the sexual health clinic. Because of the cross-sectional design, it is not possible to determine the temporal direction of the associations between negative behaviour and health problems.FundingThis report presents independent research commissioned by the National Institute for Health Research (NIHR) under its Programme Grants for Applied Research scheme (RP-PG-0108-10084).     

Healthy but unhappy? Cross-national comparative analysis of depressive symptoms in Japanese vs. Korean elders

BackgroundJapan and South Korea, two neighbouring countries in East Asia, enjoy the highest life expectancies in the world, yet suffer paradoxically from high suicide rates.AimWe sought to conduct a cross-national comparative analysis of depressive symptoms among older adults in Japan vs. Korea, focusing particularly on poverty and physical health status.MethodsWe used nationally representative samples aged 65 and over from the Comprehensive Survey of Living Conditions in Japan and the Korean Community Health Survey in South Korea. Multivariate logistic regression models were conducted to examine if equivalized household income, poor self-rated health, disability and comorbidity (number of diseases) were associated with depressive symptoms, adjusting for age, education, marital status, alcohol use, smoking and living alone.ResultsOlder Japanese adults with poor self-rated health and disability were more likely to report depressive symptoms, but income level was not significantly associated with mental distress. By contrast, among older Korean people, depressive symptoms were strongly patterned by household income level, as well as poor self-rated health, disability, and comorbidity.ConclusionPoor physical health status was correlated with depressive symptoms among both Japanese and Korean seniors. However, income level was associated with depressive symptoms among only Korean elders, but not Japanese. Thus, the current generation of older Japanese adults appears to enjoy (relative) financial security, longevity, and mental wellbeing. By contrast, older Koreans experience high levels of mental distress, especially if they are financially insecure.     

Using Scotland-wide record linkage to investigate the educational and health outcomes of children treated for chronic medical conditions: a retrospective population cohort study

BackgroundChronic ill health can affect children's educational and health outcomes. Previous studies have reported poorer school outcomes among children with diabetes, asthma, epilepsy, attention deficit hyperactivity disorder (ADHD), and depression; however, many have limitations, very few have analysed multiple outcomes or population-wide data, and literature in Scotland and the UK is scarce. We aimed to identify children prescribed medication for these chronic medical conditions and investigate their educational and health outcomes compared with their peers.MethodsWe retrospectively linked Scotland-wide education records for children attending primary and secondary schools in Scotland between 2009 and 2013 to prescribing data, acute and psychiatric hospital admissions, maternity records, and deaths. School absenteeism, exclusion, and special educational need were analysed annually using generalised estimating equations. Academic attainment was analysed using generalised ordinal logistic regression and unemployment with binomial logistic regression. All-cause hospitalisation and mortality were analysed using Cox proportional hazards or Poisson piecewise regression. Analyses were restricted to singleton children born in Scotland and adjusted for sociodemographic or maternity factors and comorbid conditionsFindings766 244 schoolchildren were included in the analyses. All conditions were associated with increased school absenteeism (diabetes incidence rate ratio [IRR] 1·34, asthma 1·25, epilepsy 1·50, ADHD 1·18, depression 1·95), special educational need (diabetes odds ratio [OR] 2·45, asthma 1·28, epilepsy 10·11, ADHD 8·77, depression 2·24], and all-cause hospitalisation (diabetes hazard ratio [HR] 3·97, asthma 1·69, epilepsy 3·72, ADHD 1·33, depression 2·56). All, excluding diabetes, were associated with poorer academic attainment (asthma OR 1·11, epilepsy 4·07, ADHD 3·64, depression 3·44] and all, excluding ADHD, were associated with increased all-cause mortality [diabetes HR 3·84, asthma 1·77, epilepsy 24·77, depression 6·21]. ADHD (IRR 5·82) and depression (1·65) were associated with increased exclusion from school, whereas epilepsy (OR 1·99), ADHD (1.42), and depression (1·98) were associated with subsequent unemployment.InterpretationIn addition to poorer health outcomes, schoolchildren with these chronic medical conditions have significant educational disadvantage compared with their peers. Interventions and further understanding of the relationships between health and education among children with these conditions is required.FundingHealth Data Research UK     

How much do people spend on physical activity in England: a nationwide cross-sectional survey of expenditure

BackgroundPriced-based policies (eg, the new UK sugar tax) are an important tool for public health. Spending on physical activity might be a barrier for participation. However, little is known about the level of expenditure, its correlates, and potential for health policies. We assessed nationwide expenditure patterns on physical activity in England, the processes in individuals' decision making, and policy implications.MethodsParticipants were from a 2010 telephone survey of a representative sample of adults nationwide (≥16 years old) who did sports and exercise in the Health Survey for England 2008 and agreed to be re-contacted. We invited 3130 adults using stratified two-stage random sampling (households and then individuals) to allow interview of one person per household and control for any household interaction effect. Interviews covered types and frequency of sports and exercise expenditure, frequency and intensity of activity, and individual characteristics. A two-part regression model was fitted to explain the decision to spend on sports and exercise and level of expenditure.FindingsOf 1686 respondents (56% response rate), 1393 (83%) had undertaken sports and exercise via 72 activities. 710 (51%) were women, 821 (59%) were aged 45 years or older, 864 (62%) were employed, and 237 (17%) had an annual household income above £46 800. Popular activities were working out (n=534, 38%), swimming (362, 26%), cycling (344, 25%), running (319, 23%), and walking (231, 17%). 957 (69%) of 1393 sports and exercise participants spent money to do so (mean expenditure per month £176, SD 912). Most frequent expenses were fees for membership, competition, and classes, and this expenditure rose with activity. The decision to spend was related to older age (β coefficient −0·788), high income (0·733), being single (0·428), and not smoking (0·575), whereas level of expenditure was associated with low educational qualification (0·378), high income (0·471), households with at least one child (one child −0·524, two children −0·511, three or more children −0·522), and being resident in south and central England (0·542).InterpretationThis study confirms that whether and how much to spend money on sports and exercise are two separate decisions that are associated with different sets of predictors. Proposals for price-based policies on physical activity in the UK need to consider this subtle difference in individual behaviour, either by targeting specific groups (eg, spenders on sports and exercise) or by implementing wider environmental interventions for non-spenders.FundingDepartment of Health's Policy Research Programme (England).     

Caregiver mental health and HIV-infected child wellness: perspectives from Ugandan caregivers

Prior studies indicate a substantial link between maternal depression and early child health but give limited consideration to the direction of this relationship or the context in which it occurs. We sought to create a contextually informed conceptual framework of this relationship through semi-structured interviews with women that had lived experience of caring for an HIV-infected child while coping with depression and anxiety symptoms. Caregivers explained their role in raising healthy children as complex and complicated by poverty, stigma, and isolation. Caregivers discussed the effects of their own mental health on child well-being as primarily emotional and behavioral, and explained how looking after a child could bring distress, particularly when unable to provide desired care for sick children. Our findings suggest the need for investigation of the reciprocal effects of child sickness on caregiver wellness and for integrated programs that holistically address the needs of HIV-affected families.