Health Policy Research Programme Research Articles

Child maltreatment and living standards in adulthood: findings from a prospective birth cohort

BackgroundChild maltreatment (abuse and neglect) has well-established effects on mental health, but less is known about its influence on adult life economic circumstances. We aimed to establish associations of child maltreatment with such outcomes in mid-life. MethodsWe used data from the 1958 British birth cohort (n=8076) on child neglect and abuse (physical, sexual, psychological, and witnessing abuse) and adult (at age 50 years) long-term sickness absence, not in employment, education, or training (NEET), lacking assets, income-related support, poor educational qualifications, financial insecurity, manual social class, and social mobility. Logistic regression was used to assess associations with maltreatment separately and combined as a score (0, 1, ≥2) with and without adjustment for potential confounding factors (maternal age, birthweight, birth order, social class at birth, parental education, household amenities, crowding, and tenure at age 7 years). Missing values were imputed with multiple imputation chained equations. FindingsAbuse prevalence varied from 1% (n=115) for sexual abuse to 10% (776) for psychological abuse; 1208 (16%) were defined as neglected. 1506 (19%) of the population experienced one type of maltreatment, and 585 (7%) experienced two or more. All maltreatments were associated with most outcomes (eg, for physical abuse, adjusted odds ratio [OR] of long-term sickness absence was 2·33, 95% CI 1·62–3·35). There was a trend across multiple types of maltreatment and unfavourable outcomes—eg, adjusted OR for NEET increased for one and two or more maltreatments compared with no maltreatment (1·30 [1·02–1·65] and 1·72 [1·28–2·33], respectively; ptrend<0·0001). Likewise, adjusted OR for three or more unfavourable adult outcomes increased with more maltreatment compared with none (1·94 [1·59–2·37] for one and 2·23 (1·75–2·84] for two or more; ptrend<0·0001). Children who were sexually abused or neglected were less likely than those not sexually abused or neglected, respectively, to be upwardly mobile from birth to mid-adulthood (unadjusted OR 0·49 [0·30–0·81] and 0·45 [0·39–0·53]). InterpretationChild abuse was reported retrospectively in adulthood, although child neglect and potential confounding factors were ascertained prospectively and we examined a wide range of adult socioeconomic outcomes that are important because of their costs to individuals and society. That childhood maltreatments were associated with most unfavourable outcomes in mid-adulthood, has implications for the broader field of health equity. The increasing risk of poor adult outcome associated with multiple types of maltreatment suggests that they have an accumulating burden. FundingThis work was supported by the Department of Health Policy Research Programme through the Public Health Research Consortium. The views expressed in this abstract are those of the authors and not necessarily those of the Department of Health.

Childhood maltreatment and biomarkers for cardiometabolic disease in mid-adulthood: associations and potential explanations

BackgroundEvidence of the association between childhood maltreatment and risk of adult cardiometabolic disease is sparse. We investigated associations of different forms of child maltreatment with adult cardiometabolic markers and potential explanations. MethodsIn the 1958 British birth cohort, we tested associations of childhood neglect (ascertained at age 7 years and 11 years) and abuse (physical, sexual, psychological, self-reported at 45 years) with adult (45 years) cardiometabolic markers (blood pressure, lipids, glycated haemoglobin [HbA1c]) using linear and logistic regressions. Models were adjusted, first for factors affecting measurements (eg, room temperature, postal delay of blood sample, and for women oral contraception and hormone replacement therapy) and early life factors (eg, birthweight, socioeconomic status) and second for explanatory factors (change in body-mass index from childhood to adulthood, adult socioeconomic status, lifestyles, mental health). We applied multiple imputation to missing data on neglect and covariates, and restricted analyses to individuals with observed cardiometabolic data. FindingsAmong 9349 participants (4650 men, 4699 women), 12% (1143) reported any form of abuse. Prevalence for sexual abuse was 1·6% (149), physical abuse 6·0% (565), and psychological abuse 10·0% (926), and 1627 (17·4%) had two or more indicators of childhood neglect. Childhood neglect was associated in adulthood with raised triglycerides by 3·9% (95% CI 0·4–7·4) and HbA1c by 1·2% (0·4–2·0), and for women lower HDL by 0·05 mmol/L (0·01–0·08), after adjusting for early life covariates. Physical abuse was associated with increased risk of high LDL (odds ratio [OR] 1·24, 95% CI 1·00–1·55) and raised HbA1c in men by 2·4% (0·6–4·2), and lower HDL in women by 0·06 mmol/L (0·01–0·12). Associations for sexual abuse were similar to those for physical abuse but 95% CIs were wide. Psychological abuse was associated with increased risk of high triglycerides (OR 1·23, 1·03–1·46) and low HDL by 0·04 mmol/L (0·01–0·07). Maltreatment was not associated with raised blood pressure. All associations disappeared after further adjustment: adult lifestyle was a key explanatory factor for most associations, adult socioeconomic status was important for associations with neglect but not abuse, body-mass index was important for neglect and physical abuse, and mental health was important for psychological abuse. InterpretationChildhood maltreatments were associated with poor lipid and HbA1c profiles decades later in adulthood in this population cohort. Explanations for associations varied by form of maltreatment. Further work is needed on the role of life-course explanatory factors and on effective strategies to reduce or prevent long-term health consequences of maltreatment. FundingThis work was funded by the Department of Health Policy Research Programme through the Public Health Research Consortium and supported by the National Institute for Health Research Biomedical Research Centre at Great Ormond Street Hospital for Children NHS Foundation Trust and University College London. The views expressed in this abstract are those of the authors and not necessarily those of the Department of Health. Data collection for participants at age 45 years was funded by the Medical Research Council (grant G0000934).

Tuberculosis in migrants moving from high-incidence to low-incidence countries: a population-based cohort study of 519 955 migrants screened before entry to England, Wales, and Northern Ireland

SummaryBackgroundTuberculosis elimination in countries with a low incidence of the disease necessitates multiple interventions, including innovations in migrant screening. We examined a cohort of migrants screened for tuberculosis before entry to England, Wales, and Northern Ireland and tracked the development of disease in this group after arrival.MethodsAs part of a pilot pre-entry screening programme for tuberculosis in 15 countries with a high incidence of the disease, the International Organization for Migration screened all applicants for UK visas aged 11 years or older who intended to stay for more than 6 months. Applicants underwent a chest radiograph, and any with results suggestive of tuberculosis underwent sputum testing and culture testing (when available). We tracked the development of tuberculosis in those who tested negative for the disease and subsequently migrated to England, Wales, and Northern Ireland with the Enhanced Tuberculosis Surveillance system. Primary outcomes were cases of all forms of tuberculosis (including clinically diagnosed cases), and bacteriologically confirmed pulmonary tuberculosis.FindingsOur study cohort was 519 955 migrants who were screened for tuberculosis before entry to the UK between Jan 1, 2006, and Dec 31, 2012. Cases notified on the Enhanced Tuberculosis Surveillance system between Jan 1, 2006, and Dec 31, 2013, were included. 1873 incident cases of all forms of tuberculosis were identified, and, on the basis of data for England, Wales, and Northern Ireland, the estimated incidence of all forms of tuberculosis in migrants screened before entry was 147 per 100 000 person-years (95% CI 140–154). The estimated incidence of bacteriologically confirmed pulmonary tuberculosis in migrants screened before entry was 49 per 100 000 person-years (95% CI 45–53). Migrants whose chest radiographs were compatible with active tuberculosis but with negative pre-entry microbiological results were at increased risk of tuberculosis compared with those with no radiographic abnormalities (incidence rate ratio 3·2, 95% CI 2·8–3·7; p<0·0001). Incidence of tuberculosis after migration increased significantly with increasing WHO-estimated prevalence of tuberculosis in migrants' countries of origin. 35 of 318 983 pre-entry screened migrants included in a secondary analysis with typing data were assumed index cases. Estimates of the rate of assumed reactivation tuberculosis ranged from 46 (95% CI 42–52) to 91 (82–102) per 100 000 population.InterpretationMigrants from countries with a high incidence of tuberculosis screened before being granted entry to low-incidence countries pose a negligible risk of onward transmission but are at increased risk of tuberculosis, which could potentially be prevented through identification and treatment of latent infection in close collaboration with a pre-entry screening programme.FundingWellcome Trust, UK National Institute for Health Research, UK Medical Research Council, Public Health England, and Department of Health Policy Research Programme.

Clinical workload in UK primary care: a retrospective analysis of 100 million consultations in England, 2007–14

SummaryBackgroundPrimary care is the main source of health care in many health systems, including the UK National Health Service (NHS), but few objective data exist for the volume and nature of primary care activity. With rising concerns that NHS primary care workload has increased substantially, we aimed to assess the direct clinical workload of general practitioners (GPs) and practice nurses in primary care in the UK.MethodsWe did a retrospective analysis of GP and nurse consultations of non-temporary patients registered at 398 English general practices between April, 2007, and March, 2014. We used data from electronic health records routinely entered in the Clinical Practice Research Datalink, and linked CPRD data to national datasets. Trends in age-standardised and sex-standardised consultation rates were modelled with joinpoint regression analysis.FindingsThe dataset comprised 101 818 352 consultations and 20 626 297 person-years of observation. The crude annual consultation rate per person increased by 10·51%, from 4·67 in 2007–08, to 5·16 in 2013–14. Consultation rates were highest in infants (age 0–4 years) and elderly people (≥85 years), and were higher for female patients than for male patients of all ages. The greatest increases in age-standardised and sex-standardised rates were in GPs, with a rise of 12·36% per 10 000 person-years, compared with 0·9% for practice nurses. GP telephone consultation rates doubled, compared with a 5·20% rise in surgery consultations, which accounted for 90% of all consultations. The mean duration of GP surgery consultations increased by 6·7%, from 8·65 min (95% CI 8·64–8·65) to 9·22 min (9·22–9·23), and overall workload increased by 16%.InterpretationOur findings show a substantial increase in practice consultation rates, average consultation duration, and total patient-facing clinical workload in English general practice. These results suggest that English primary care as currently delivered could be reaching saturation point. Notably, our data only explore direct clinical workload and not indirect activities and professional duties, which have probably also increased. This and additional research questions, including the outcomes of workload changes on other sectors of health care, need urgent answers for primary care provision internationally.FundingDepartment of Health Policy Research Programme.

The Stroke Prevention Programme: a programme of research to inform optimal stroke prevention in primary care

BackgroundThe management of cardiovascular (CV) risk factors in community populations is suboptimal. The aim of this programme was to explore the role of three approaches [use of a ‘polypill’; self-management of hypertension; and more intensive targets for blood pressure (BP) lowering after stroke] to improve prevention of CV disease (CVD) in the community.Research questions(1) Is it more cost-effective to titrate treatments to target levels of cholesterol and BP or to use fixed doses of statins and BP-lowering agents (polypill strategy)? (2) Will telemonitoring and self-management improve BP control in people on treatment for hypertension or with a history of stroke/transient ischaemic attack (TIA) in primary care and are they cost-effective? (3) In people with a history of stroke/TIA, can intensive BP-lowering targets be achieved in a primary care setting and what impact will this have on health outcomes and cost-effectiveness?DesignMixed methods, comprising three randomised controlled trials (RCTs); five cost-effectiveness analyses; qualitative studies; analysis of electronic general practice data; a screening study; a systematic review; and a questionnaire study.SettingUK general practices, predominantly from the West Midlands and the east of England.ParticipantsAdults registered with participating general practices. Inclusion criteria varied from study to study.InterventionsA polypill – a fixed-dose combination pill containing three antihypertensive medicines and simvastatin – compared with current practice and with optimal implementation of national guidelines; self-monitoring of BP with self-titration of medication, compared with usual care; and an intensive target for systolic BP of &lt; 130 mmHg or a 10 mmHg reduction if baseline BP is &lt; 140 mmHg, compared with a target of &lt; 140 mmHg.ResultsFor patients known to be at high risk of CVD, treatment as per guidelines was the most cost-effective strategy. For people with unknown CV risk aged ≥ 50 years, offering a polypill is cost-effective [incremental cost-effectiveness ratio (ICER) of £8115 per quality-adjusted life-year (QALY)] compared with a strategy of screening and treating according to national guidelines. Both results were sensitive to the cost of the polypill. Self-management in people with uncontrolled hypertension led to a 5.4 mmHg [95% confidence interval (CI) 2.4 to 8.5 mmHg] reduction in systolic BP at 1 year, compared with usual care. It was cost-effective for men (ICER of £1624 per QALY) and women (ICER of £4923 per QALY). In people with stroke and other high-risk groups, self-management led to a 9.2 mmHg (95% CI 5.7 to 12.7 mmHg) reduction in systolic BP at 1 year compared with usual care and dominated (lower cost and better outcome) usual care. Aiming for the more intensive BP target after stroke led to a 2.9 mmHg (95% CI 0.2 to 5.7 mmHg) greater reduction in BP and dominated the 140 mmHg target.ConclusionsPotential for a polypill needs to be further explored in RCTs. Self-management should be offered to people with poorly controlled BP. Management of BP in the post-stroke population should focus on achieving a &lt; 140 mmHg target.Trial registrationCurrent Controlled Trials ISRCTN17585681, ISRCTN87171227 and ISRCTN29062286.FundingThe National Institute for Health Research (NIHR) Programme Grants for Applied Research programme. Additional funding was provided by the NIHR National School for Primary Care Research, the NIHR Career Development Fellowship and the Department of Health Policy Research Programme.

Safety of live attenuated influenza vaccine in young people with egg allergy: multicentre prospective cohort study

Study question How safe is live attenuated influenza vaccine (LAIV), which contains egg protein, in young people with egg allergy?Methods In this open label, phase IV intervention study, 779 young...

Screen detection of ductal carcinoma in situ and subsequent incidence of invasive interval breast cancers: a retrospective population-based study

SummaryBackgroundThe value of screen detection and treatment of ductal carcinoma in situ (DCIS) is a matter of controversy. At present, the extent to which the diagnosis and treatment of DCIS could prevent the occurrence of invasive breast cancer in the future is not clear. We sought to estimate the association between detection of DCIS at screening and invasive interval cancers subsequent to the relevant screen.MethodsWe obtained aggregate data for screen-detected cancers from 84 local screening units within 11 regional Quality Assurance Reference Centres in England, Wales, and Northern Ireland from the National Health Service Breast Screening Programme. Data for DCIS diagnoses were obtained for women aged 50–64 years who were invited to and attended mammographic breast screening from April 1, 2003, to March 31, 2007 (4 screening years). Patient-level data for interval cancer arising in the 36 months after each of these were analysed by Poisson regression with invasive interval cancer screen detection rate as the outcome variable; DCIS detection frequencies were fitted first as a continuous and then as a categorical variable. We repeated this analysis after adjustment with both small size and high-grade invasive screen-detected cancers.FindingsWe analysed data for 5 243 658 women and on interval cancers occurring in the 36 months after the relevant screen. The average frequency of DCIS detected at screening was 1·60 per 1000 women screened (median 1·50 [unit range 1·54–3·56] per 1000 women). There was a significant negative association of screen-detected DCIS cases with the rate of invasive interval cancers (Poisson regression coefficient −0·084 [95% CI −0·13 to −0·03]; p=0·002). 90% of units had a DCIS detection frequency within the range of 1·00 to 2·22 per 1000 women; in these units, for every three screen-detected cases of DCIS, there was one fewer invasive interval cancer in the next 3 years. This association remained after adjustment for numbers of small screen-detected invasive cancers and for numbers of grade 3 invasive screen-detected cancers.InterpretationThe association between screen-detected DCIS and subsequent invasive interval cancers suggests that detection and treatment of DCIS is worthwhile in prevention of future invasive disease.FundingUK Department of Health Policy Research Programme and NHS Cancer Screening Programmes.

Time trends in hospital admissions for violent, self-inflicted, and drug-related or alcohol-related injury for adolescents in England and Scotland, 2005–11: observational population-based study

BackgroundIn England, a third of emergency admissions of adolescents for injury are adversity related (violent, self-inflicted, drug-related, or alcohol-related). A comparison of time trends of the incidence of admissions for violent injury between England and Scotland revealed steeper declines in 2005–11 in Scotland. We aimed to determine whether incidence of admissions for any adversity-related injury varied substantially between the two countries. MethodsWe conducted time-series analyses of emergency admissions between 2005 and 2011 for adversity-related injury (defined by the 10th revision of the International Classification of Diseases) to National Health Service hospitals in England (Hospital Episode Statistics) and in Scotland (Scottish Morbidity Records) in 10–18 year olds. Analyses were stratified in groups by sex and age (10–12 years, 13–15, 16–18) and were adjusted for background trends in admissions for any injury. FindingsIn 2005, rates of admissions per 100 000 for adversity-related injury ranged from 48·9 for girls aged 10–12 years in Scotland (95% CI 0–98·9) to 978·2 for boys aged 16–18 years in Scotland (764·0–1184·3). Rates for 10–12-year-old girls and boys, respectively, were similar between the two countries, but were higher in Scotland for 13–15 year olds and 16–18 year olds. From 2005 to 2011, rates decreased in both countries for all groups by −1·5% per year (95% CI −3·2 to 0·21) to −10·0% per year (–15·2 to −4·4), except for 16–18-year-old girls and boys in England, where rates increased by 0·25% per year (0·09–0·41) to 2·5% per year (1·2–3·7). Decreases in all groups were greater in Scotland than in England after adjustment for trends in admissions for any injury. By 2011, although incidences of admissions for adversity-related injury in adolescents aged 13–15 and 16–18 years remained higher in Scotland, differences between England and Scotland were smaller than in 2005. InterpretationOur finding that rates of admissions for adversity-related injury decreased more steeply in Scotland than in England raises questions about the factors driving these discrepancies. Several initiatives within each country might have been influential. For example, the English government attempted to tackle incidence of violence and gangs by targeting high crime areas with higher levels of policing. The Scottish government set up contracts with local gangs to exchange a so-called clean slate for psychosocial support. Further research into potentially successful practices in Scotland could be used to develop future initiatives to reduce harm in adolescents in both countries and further afield. More research is needed, especially into why the incidence of admissions for adversity-related injury increased for older adolescents in England. FundingAH was supported by the Policy Research Unit in the Health of Children, Young People and Families, which is funded by the Department of Health Policy Research Programme (grant reference number 109/0001). AH is also supported by the University College London Impact studentship. The study sponsors played no part in the design, data analysis, and interpretation of this study; the writing of the abstract; or the decision to submit the abstract for publication.

Emergency admissions across the transition from paediatric to adult care: cross-sectional analysis of English hospital data

BackgroundTransition from paediatric to adult health services is often sub-optimally achieved for children with long-term conditions. We investigated trends in emergency admissions across the transition from paediatric to adult care. MethodsUsing Hospital Episodes Statistics data, we analysed 1 186 392 emergency admissions between April 1, 2009, and March 31, 2012, in children and young people aged 10–24 years. We calculated emergency admission rates for this cohort before (10–15 years), during (16–18), and after transition (19–24) using mid-year population estimates as denominators. We stratified analyses by underlying long-term conditions, which we defined by validated International Classification of Diseases 10th revision codes. We excluded injury-related and maternity-related admissions. We used negative binomial regression to determine incidence rate ratios (IRRs). FindingsPrevalence rates for emergency admissions increased 64% for girls (29/1000 before transition to 48/1000 after transition) and 17% for boys (26/1000 to 31/1000). This increase overwhelmingly occurred during transition (16–18 years), when rates increased 4·0/1000 per year for girls and 2·3/1000 for boys. At the same time, average length of hospital stay increased from 2·1 to 2·9 days for girls, and 1·8 to 5·0 days for boys (increases of 38% and 186% for girls and boys, respectively). Disparities in admission rates increased during transition, with children and young people from the most deprived quintile more likely than those from the least deprived quintile to have an emergency admission (IRR 1·63, 95% CI 1·60–1·66), and more so than before transition (1·51, 1·45–1·56). The increase in admission rates was driven by underlying long-term conditions: overall, rates increased 14% (IRR 1·14, 95% CI 1·13–1·14), but for children and young people with long-term conditions rates increased 46% (1·46, 1·46–1·47). Rates increased most rapidly for mental health problems, metabolic and endocrine disorders, multiple long-term conditions (both sexes), and respiratory disorders (girls only). InterpretationOur results show that emergency admission rates, duration, and inequalities increase across transition, especially in children and young people with underlying long-term conditions. The extent to which this increase is due to transition within the health service, rather than other developments, requires further research. Our results are in keeping with the widely recognised difficulties that young people face when they transition to adult health services. FundingLPMMW and JW were supported by funding from the Department of Health Policy Research Programme through funding to the Policy Research Unit in the Health of Children, Young People and Families (grant reference 109/0001). RG is supported by awards establishing the Farr Institute of Health Informatics Research at University College London Partners from the Medical Research Council and a consortium of funders (MR/K006584/1).

Effectiveness of a nurse-led intensive home-visitation programme for first-time teenage mothers (Building Blocks): a pragmatic randomised controlled trial

SummaryBackgroundMany countries now offer support to teenage mothers to help them to achieve long-term socioeconomic stability and to give a successful start to their children. The Family Nurse Partnership (FNP) is a licensed intensive home-visiting intervention developed in the USA and introduced into practice in England that involves up to 64 structured home visits from early pregnancy until the child's second birthday by specially recruited and trained family nurses. We aimed to assess the effectiveness of giving the programme to teenage first-time mothers on infant and maternal outcomes up to 24 months after birth.MethodsWe did a pragmatic, non-blinded, randomised controlled, parallel-group trial in community midwifery settings at 18 partnerships between local authorities and primary and secondary care organisations in England. Eligible participants were nulliparous and aged 19 years or younger, and were recruited at less than 25 weeks' gestation. Field-based researchers randomly allocated mothers (1:1) via remote randomisation (telephone and web) to FNP plus usual care (publicly funded health and social care) or to usual care alone. Allocation was stratified by site and minimised by gestation (<16 weeks vs ≥16 weeks), smoking status (yes vs no), and preferred language of data collection (English vs non-English). Mothers and assessors (local researchers at baseline and 24 months' follow-up) were not masked to group allocation, but telephone interviewers were blinded. Primary endpoints were biomarker-calibrated self-reported tobacco use by the mother at late pregnancy, birthweight of the baby, the proportion of women with a second pregnancy within 24 months post-partum, and emergency attendances and hospital admissions for the child within 24 months post-partum. Analyses were by intention to treat. This trial is registered with ISRCTN, number ISRCTN23019866.FindingsBetween June 16, 2009, and July 28, 2010, we screened 3251 women. After enrolment, 823 women were randomly assigned to receive FNP and 822 to usual care. All follow-up data were retrieved by April 25, 2014. 304 (56%) of 547 women assigned to FNP and 306 (56%) of 545 assigned to usual care smoked at late pregnancy (adjusted odds ratio [AOR] 0·90, 97·5% CI 0·64–1·28). Mean birthweight of 742 babies with mothers assigned to FNP was 3217·4 g (SD 618·0), whereas birthweight of 768 babies assigned to usual care was 3197·5 g (SD 581·5; adjusted mean difference 20·75 g, 97·5% CI −47·73 to 89·23. 587 (81%) of 725 assessed children with mothers assigned to FNP and 577 (77%) of 753 assessed children assigned to usual care attended an emergency department or were admitted to hospital at least once before their second birthday (AOR 1·32, 97·5% CI 0·99–1·76). 426 (66%) of 643 assessed women assigned to FNP and 427 (66%) 646 assigned to usual care had a second pregnancy within 2 years (AOR 1·01, 0·77–1·33). At least one serious adverse event (mainly clinical events associated with pregnancy and infancy period) was reported for 310 (38%) of 808 participants (mother–child) in the usual care group and 357 (44%) of 810 in the FNP group, none of which were considered related to the intervention.InterpretationAdding FNP to the usually provided health and social care provided no additional short-term benefit to our primary outcomes. Programme continuation is not justified on the basis of available evidence, but could be reconsidered should supportive longer-term evidence emerge.FundingDepartment of Health Policy Research Programme.

Characteristics of trafficked adults and children with severe mental illness: a historical cohort study

Evidence regarding the mental health needs of trafficked people is limited; however, prevalence of depression and post-traumatic stress disorder is high among trafficked people who are in contact with shelter services. We aimed to investigate the sociodemographic and clinical characteristics of trafficked people with severe mental illness. We did a historical cohort study of trafficked people in contact with secondary mental health services in South London, UK, between Jan 1, 2006, and July 31, 2012. We searched and retrieved comprehensive clinical electronic health records for over 200 000 patients from the Case Register Interactive Search database to identify trafficked patients. A matched cohort of non-trafficked adults was generated by simple computer-generated random selection of potential controls for each case within the parameters of matching criteria. We extracted data on sociodemographic and clinical characteristics and abuse history, and used multiple imputation to deal with missing data. We fitted logistic regression models to compare trafficked and non-trafficked patients. We identified 133 trafficked patients, including 37 children. 78 (81%) of 96 adults and 25 (68%) of 37 children were female. 19 (51%) of 37 children were trafficked for sexual exploitation. Among both adults and children, the most commonly recorded diagnoses were post-traumatic stress disorder, severe stress, or adjustment disorder (27 adults [28%] and ten children [27%]) and affective disorders (33 adults [34%] and ten children [27%]). Records documented childhood physical or sexual abuse among trafficked adults (41 [43%]) and children (28 [76%]), and adulthood abuse among trafficked adults (58 [60%]). Trafficked adults were more likely to be compulsorily admitted as a psychiatric inpatient than non-trafficked adults (adjusted odds ratio 7·61, 95% CI 2·18-26·60; p=0·002) and had longer admissions (1·48, 1·01-2·15; p=0·045). No association was found between trafficking status and either adverse pathway into care (adjusted odds ratio 0·91, 95% CI 0·40-2·05; p=0·82) or substance misuse problems (0·55, 0·27-1·17; p=0·12). Severe mental illness in trafficked people is associated with longer admissions and high levels of abuse before and after trafficking. Evidence is needed on the effectiveness of interventions to promote recovery for this vulnerable group. Department of Health Policy Research Programme.

Burden of recurrent emergency hospital admissions in children and young people in England: a cohort study

BackgroundIn England, the number of emergency admissions in children and young people has increased by 28% since 1999. Evidence is lacking about the contribution of recurrent admissions to the high rate of emergency admissions among children and young people. We quantified the contribution of recurrent admissions to the total burden of admissions in England. MethodsWe analysed all hospital admissions to the National Health Service in England using Hospital Episode Statistics from 2009 to 2011 for children and young people aged 0–24 years. We followed up children and young people for 2 years from their first emergency admission (index admission) in 2009. We determined the number of subsequent admissions, time to next admission, length of stay, and the proportion of admissions for injury and of children affected by a chronic condition measured by diagnostic codes in all admissions during the 2 years. Findings869 895 children had an index admission in 2009, resulting in a further 939 710 admissions (of which 600 322 [64%] were emergency admissions) over the next 2 years. We excluded 4371 children (0·5%) with inconsistent records (eg, multiple birth admissions). After discharge from the index admission, 274 986 children (32%) had a recurrent emergency admission, accounting for 41% (n=1 470 107) of all emergency admissions in the 2-year cohort. A few children and young people (37 311, 4%) had four or more emergency admissions. The proportion of patients with recurrent admissions was similar across all age groups. 73 830 first recurrent admissions (26%) occurred within the first month after discharge from the index admission. 360 633 index admissions (41%), in contrast with 459 167 recurrent emergency admissions (76%), were in patients with a chronic condition. InterpretationThis snapshot of hospital flow during 2 years shows that recurrent emergency admissions account for a substantial minority of all emergency admissions and predominantly affect children and young people with chronic conditions. Since we only included inpatient hospital data, our analysis provides a limited overview of health-care use. Interventions to reduce re-admissions should consider discharge planning for support by community services for parents caring for children with chronic conditions. FundingLW was supported by funding from the Department of Health Policy Research Programme through funding to the Policy Research Unit in the Health of Children, Young People and Families. This is an independent report commissioned and funded by the Department of Health.

Risk of future harm in adolescents admitted to hospitals in England for injury related to victimisation, self-harm, or drug or alcohol misuse: a retrospective cohort study

BackgroundInjury related to adversity (ie, maltreatment, violence, self-harm, or drug or alcohol misuse) accounts for most mortality in adolescents. Targeting reduction of future harm in adolescents admitted to hospital with adversity-related injuries is of great public health interest in the UK. Evidence is lacking on the risks of emergency re-admission and death after an adolescent admission for adversity-related injury. We aimed to quantify these risks and to establish whether they are higher than those in adolescents with accident-related injury only. MethodsWe used linked Hospital Episode Statistics and Office for National Statistics mortality data for 10–19-year-old adolescents, who had at least one emergency admission for injury (the index admission; for multiple emergency admissions for injury, we randomly selected one) between April 1, 1997, and March 31, 2012. We conducted time-to-event analyses for emergency re-admission (non-injury or injury for any reason) or death before age 31 years, following the index. FindingsOf 1 080 368 adolescents who had emergency admissions for injury, 335 103 (31·0%) had adversity-related injury at the index admission and 649 774 (60·1%) had accident-related injury. The cumulative incidence of an emergency re-admission 1 year from adversity-related injury was greater than that from an accident-related injury (girls 18·5% [95% CI 18·2–18·7%] vs 8·3 [8·2–8·5], boys 11·2 [11·0–11·3] vs 7·0 [6·9–7·0]), and remained about twice as high 5 and 10 years later. Rates of death within 1 year were greater after adversity-related injury (girls 114/100 000 [99–130] vs 52 [42–64], boys 201 [180–226] vs 56 [49–63]) and was twice to four times as high after 10 years. Relative hazards of re-admissions and death from adversity-related injury for girls and boys were higher than from accident-related injury by 1·65–2·16 times, adjusting for multiple re-admissions, age, chronic conditions, ethnicity, and socioeconomic status. InterpretationThe risk of emergency re-admission or death from an adolescent admission for adversity-related injury is twice as high as that from accident-related injury. Since International Classification of Diseases 10 codes used were moderately sensitive for identifying adversity, the true relative risk could be higher than our estimates. Development of hospital-based interventions in adolescents with adversity-related injury, such as motivational interviewing, which has shown promise for reducing alcohol-related injury in the USA, could reduce this risk. FundingAH was supported by the Policy Research Unit in the Health of Children, Young People and Families, which is funded by the Department of Health Policy Research Programme (grant reference number 109/0001). AH is also supported by the University College London IMPACT studentship. The study funders played no part in the design, data analysis, and interpretation of this study, the writing of the manuscript, or the decision to submit the paper for publication. The authors' work was independent of their funders.

Early life predictors of adult leisure-time physical inactivity: prospective findings from the 1958 British Birth Cohort

BackgroundLeisure-time physical inactivity has a high prevalence and associated disease burden. Adult inactivity research ignores earlier life factors from which later life influences can originate. We aimed to establish whether early life factors influence adult inactivity. MethodsThe 1958 British Birth Cohort is a nationwide follow-up study of all births during 1 week in March, 1958. The outcome of the present study was leisure-time inactivity, defined as activity frequency of less than once a week, assessed at ages 33 years, 42 years, and 50 years (n=12 776). Early life factors (birth to 16 years) were categorised into three domains (physical, social, behavioural). We assessed stability of inactivity from 33 years to 50 years and associations with adult inactivity using logistic regression of: factors within domains, the three domains combined (ie, multivariable associations), and allowing for adult factors. Missing values were imputed with multiple imputation chained equations. FindingsAt each adult age, about 32% of participants were inactive (31% at 33 years, 34% at 42 years, and 30% at 50 years). 1189 (9%) were inactive at all three ages. In analysis of the three domains simultaneously, factors related to adult inactivity were: short prepubertal stature, poor hand control or physical coordination, and poor cognition (physical); low class at birth, minimal parental education, poor household amenities, parental divorce, and institutional care (social); and inactivity, average or lower sports aptitude, smoking, and externalising and unsociable behaviours (behavioural). Odds ratios for inactivity at age 33 years ranged from 0·86 per SD increase in cognition (95% CI 0·82–0·91) to 1·41 (1·23–1·61) for average or lower sports aptitude. Associations weakened slightly but were mostly maintained after adjustment for adult covariates. After allowing for adult covariates participants with unskilled manual backgrounds had 23% higher odds of inactivity at 50 years than those from professional or managerial backgrounds. InterpretationOur study, based on self-report, focuses only on leisure-time inactivity. However, to have repeat, prospective data on inactivity spanning several decades in adulthood is rare, and leisure-time inactivity is likely to be amenable to modification. Adult inactivity is only moderately stable, providing opportunities for behaviour change. Factors from early life are associated with adult inactivity, allowing for early identification of groups vulnerable to later inactivity. FundingThis work was supported by the Department of Health Policy Research Programme through the Public Health Research Consortium. The Great Ormond Street and University College London Institute of Child Health was supported in part by the Department of Health's National Institute for Health Biomedical Research Centre.

Assessment of the implementation of the National Cold Weather Plan for England

BackgroundTo address the large wintertime health burden in England, the Department of Health introduced a Cold Weather Plan (CWP) in 2011, which provides advice to relevant organisations and the general public on actions that should be taken to protect vulnerable individuals during cold weather. The Department of Health commissioned the Policy Innovation Research Unit at the London School of Hygiene and Tropical Medicine to undertake an assessment of the implementation of the CWP over winter 2012–13 and to begin to assess its potential health benefits and costs. MethodsThe research consisted of time-series regression and episode analysis of historic records of daily mortality, 1993–2006 (from the Office for National Statistics) and emergency hospital admissions and accident and emergency department visits, 1997–2010 (both from the Health and Social Care Information Centre), linked to weather data for each region of England. Models were adjusted for seasonal factors unrelated to temperature and possible day-of-week effects. Semistructured telephone interviews (n=52) with staff involved in the implementation of the CWP were done in ten local authorities selected from within regions of the country with different winter weather patterns to provide a spread of urban versus rural areas and levels of socioeconomic deprivation. Interviews focused on the operational policy changes and health-care or social-care services' response to the introduction of the CWP for comparison with the expectations contained in the plan. 25 local authority managers, 13 National Health Service (NHS) commissioners, four NHS acute or community trust managers, two NHS ambulance trust managers, six care home managers, one family doctor, and one voluntary sector manager were interviewed, dependent on the local allocation of responsibility for CWP implementation between agencies. Analysis of CWP-related plans and related documents in the ten local authority areas was done. A national survey was undertaken of 8500 community nurse members of the Royal College of Nursing, of whom 437 (5%) responded. Semi-structured telephone interviews (n=35) with individuals at high risk from cold weather exposure living in their own homes in two local authorities were also done. People were interviewed shortly after a level 3 cold weather alert had been issued, which is issued when severe winter weather is occurring. FindingsCold-related effects were noted for both mortality and hospital use but the biggest cold burdens occurred at temperatures outside the alert thresholds used in the CWP. Effects could be delayed by up to 4 weeks after initial exposure. The CWP and the cold weather alerts were viewed as a helpful exposition of good practice by local health-care and social-care managers. The CWP formalised existing activity and helped to coordinate the actions of a wide range of staff in different agencies. There were examples of innovations that local staff attributed directly to the CWP (eg, a push to raise influenza immunisation levels of front-line staff). Analysis of the interviews with people living in the community suggested that interventions aimed at people who are vulnerable to cold should focus on people who live in cold homes and are socially isolated. InterpretationThe CWP is broadly welcomed and has the potential to reduce cold-related public health effects and health services usage, but the alerts may need revision. The present study is limited by the fact that the most robust quantitative evidence on the effectiveness of the CWP will only come once there are sufficient years after intervention to assess any changes in weather–health relations. FundingDepartment of Health Policy Research Programme.

Did legislation to regulate school meals in England widen dietary inequalities in children aged 4–7 years? A repeat cross-sectional study

BackgroundThere is an urgent need for policies to address obesity and its health consequences. Children's diets lay the foundations for healthy nutrition across the life course. School meals can contribute substantially to total diet. From 2008, primary schools in England were required to comply with legislation specifying the frequency with which certain foods can and cannot be served and maximum and minimum levels of nutrients (food and nutrient based standards) in school meals. The effect of this legislation on dietary inequalities is unknown. We aimed to assess the effect of the legislation on children's mean nutrient intake at lunch and in their total diet. MethodsWe undertook cross-sectional surveys in 2003–04 (before) and 2008–09 (after legislation) in 12 primary schools in northeast England. Schools represented the socioeconomic spectrum. All children aged 4–7 years were eligible. Ethics approval and informed, written parental consent was obtained. We used a validated, prospective 4-day food diary to record children's dietary intake (n=368 in 2003–04; n=624 in 2008–09). The index of multiple deprivation 2007 was used to categorise socioeconomic status into quintiles. A linear mixed-effect model explored the effect of year, lunch type, socioeconomic status, and interactions between these factors on nutrient intake. FindingsSchool lunch uptake decreased significantly in the most deprived quintile after compared with before legislation (74% to 60%, χ2 test p=0·019). After legislation, children's mean intake from a school lunch had a lower percentage of energy from fat (difference in mean change −8·4%, 95% CI −9·9 to −6·8) and saturated fat (−4·5%, −5·4 to −3·6) and higher non-starch polysaccharides (NSPs; 1·6 g, 1·3 to 1·9) and iron (0·5 mg, 0·3 to 0·7) than those consuming a home-packed lunch. After legislation, children in the least deprived quintile had higher mean nutrient intakes from their lunch than children in the most deprived quintile (differences in mean change: NSP 0·8 g, 95% CI 0·4–1·2; iron 0·4 mg, 0·2–0·6; zinc 0·3 mg, 0·1–0·5). After legislation, children's mean micronutrient intakes in their total diet increased in the least deprived quintile and fell in the most deprived quintile (differences in mean change: NSP 1·2 g, 95% CI 0·3–2·1; iron 0·9 mg, 0·3–1·5; zinc 0·64 mg 0·3–1·0). For children consuming a school lunch, mean total dietary intake of vitamin C increased across all quintiles but increased more for children in the most deprived quintile; for children consuming a home-packed lunch, mean vitamin C intake fell in the most deprived quintile and increased in the least deprived quintile (p=0·035). A key limitation was the use of repeated cross-sectional surveys and lack of control group, which limits the extent to which changes in nutrient intakes can be fully attributed to the implementation of the school food legislation. InterpretationThis legislation has led to children who consume school lunches having a healthier mean nutrient profile than those who consume home-packed lunches. Nutrient intakes have improved overall for the most deprived children. However, inequalities in micronutrient intakes persist at lunchtime and in total diet regardless of lunch type. Additional interventions are needed: for the most deprived families, uptake of school lunch needs to be encouraged; for all children, guidance is needed on how to make the healthiest choices from the foods offered. FundingThis work was undertaken as part of the research programme of the Public Health Research Consortium, which is funded by the Department of Health Policy Research Programme. The views expressed in this publication are those of the authors. MW and AJA are partly funded as members of Fuse, the Centre for Translational Research in Public Health, a UK Clinical Research Collaboration (UKCRC) Public Health Research Centre of Excellence. We thank the British Heart Foundation, Cancer Research UK, Economic and Social Research Council, MRC, and the NIHR, under the auspices of the UKCRC for funding. AJA is funded by a NIHR Professorship.

Investigation of bidirectional associations between depressive symptoms and physical activity from age 11 to 50 years in the 1958 British Birth Cohort

BackgroundUnderstanding the relation between physical activity and depressive symptoms is important because if activity could alleviate or prevent depressive symptoms, it would represent a low-risk, low-cost intervention at the population level with auxiliary health benefits. Alternatively, if depressive symptoms are related to an absence of activity, that would inform understanding of barriers to activity. Yet, the direction of the relation and whether it changes with age is unclear. In the 1958 British Birth Cohort, we investigated whether depressive symptoms are concurrent with activity and assessed the direction of association from age 11 to 50 years. MethodsThe 1958 Cohort is a large nationwide study, with follow-up of all births during 1 week in March, 1958. Activity frequency and depressive symptoms were assessed by questionnaires at 11, 16, 23, 33, 42, and 50 years (N varied from 14 072 at 11 years to 9633 at 50 years). These measures—ie, activity frequency and depressive symptoms at different life stages—were the primary outcomes. Relations were examined for the full depressive symptom spectrum and for those with elevated symptom levels (top 10% at each age, sex specific), referred to as depressed. Cross-sectional associations between activity and depressive symptoms were examined with linear and logistic regression. For longitudinal analyses, we used mixed effects models (to account for repeated measures in the same individuals), providing yearly estimates for activity associations with the depressive symptoms trajectory (in non-lagged and lagged models) and vice versa. We tested whether relations varied with age, using an interaction (age) term. Mixed effects models allow for missing outcome data; however, to minimise data loss, missing covariates were imputed with multiple imputation chained equations. FindingsCross-sectionally, at all ages except 33 years, there was evidence (p<0·05) of fewer depressive symptoms with more frequent activity. In longitudinal analyses, 11–50 years, activity was associated with a lower symptoms trajectory and the magnitude of association did not vary with age; for example, those who were inactive at 23 years and who remained inactive 5 years later had no change in symptom level, whereas those who increased activity to five times per week had lower mean number of symptoms by 0·30 (95% CI 0·25–0·35). This finding is equivalent to an estimated reduction in depression from 10% to 6·5%. The longitudinal relation between symptoms and activity trajectory weakened with age (pinteraction<0·0001): mean increase in activity in those with no symptoms at 23 years and 5 years later was 0·60 times per week (95% CI 0·57–0·64), for those increasing to one symptom it was 0·53 times per week (0·49–0·56), whereas by 48 years symptom level was not associated with activity (p>0·05). In general, associations were similar for the depression trajectory. Results were replicated, though weaker, in lagged models. InterpretationThe relation between activity and depressive symptoms was bidirectional, albeit more persistent over decades of life for activity to depressive symptoms. Repeated, prospective data for activity and depressive symptoms spanning adolescence to mid-adulthood are rare; however, measures are self-reported. Findings emphasise the importance of activity to alleviate depressive symptoms and, in turn, depressive symptoms before mid-life could be a potential barrier to uptake and increase in activity. FundingThis work was supported by the Department of Health Policy Research Programme through the Public Health Research Consortium. The GOSH/UCL Institute of Child Health was supported in part by the Department of Health's NIHR Biomedical Research Centre.

From tobacco control research to policy: the example of smoke-free legislation in England

BackgroundTobacco control policies, such as legislation to make all enclosed public places and workplaces smoke free, are often highly contested. Despite growing evidence that smoke-free laws are effective in reducing secondhand smoke exposure and provide early health benefits, efforts to contest evidence for secondhand smoke and the effects of legislation are widespread. The Tobacco Control Research Group at Bath, UK, has produced a wide range of work assessing various effects of smoke-free legislation introduced on July 1, 2007, in England. We use our assessment of whether the legislation reduced hospital admissions for heart attacks as an example of our research. When this assessment was published in the British Medical Journal in 2010, several studies showing a decrease in admissions for heart attack after the introduction of smoke-free laws in other countries had already been published. In view of efforts to contest these findings by pro-tobacco bloggers, media articles questioning the credibility of tobacco control research, and spurious coverage of our work in national newspapers before it was published, we had two aims: to build on existing research by taking into account long-term trends; and to ensure the findings, when published, were accurately reported in the media and to policy makers. MethodsWe investigated the short-term effect of the English smoke-free legislation with numbers of emergency hospital admissions per week coded with a primary diagnosis of a myocardial infarction from July 1, 2002, to Sept 30, 2008, in all patients aged 18 years and older. The analysis was done with a segmented Poisson regression analysis that adjusted for long-term trends and other potential confounders. Cost savings to the NHS were calculated in collaboration with the London Health Observatory. For the media strategy, we prepared simple but accurate media briefings, which we disseminated to ten key parties that were likely to be contacted by the media. We held a press briefing at the Science Media Centre (London), and worked closely with them, Bath University press office, and the British Medical Journal. FindingsSmoke-free legislation was association with a 2·4% (95% CI −4·06 to −0·66) reduction in the number of emergency admissions for myocardial infarction. This reduction equates to 1200 fewer emergency admissions for myocardial infarction in the first year after legislation. The London Health Observatory estimated that £8·4 million has been saved in emergency hospital care for heart attacks in the first year since smoke-free legislation was introduced. The research received wide reaching, positive coverage in both print and broadcast media, featuring on the BBC, BBC Radio 4's Today programme, BBC Radio 5 Live, and local BBC radio stations. National newspapers that featured the story included The Times, Financial Times, Independent, and Daily Mail. Internationally, the research was reported by media in several countries, including the USA, Canada, and New Zealand. Our efforts to ensure that coverage was accurate were successful, despite later efforts by pro-tobacco bloggers to undermine it. InterpretationThis study adds to the growing evidence that smoke-free legislation leads to reductions in myocardial infarctions and builds on previous work by showing that such decreases are noted even when underlying reductions in admissions and other potential confounders are controlled for. With careful planning and communication, public health research in a contested field can be effectively and accurately disseminated. The research has subsequently contributed to the present government's tobacco control strategy. FundingDepartment of Health's Policy Research Programme.

A whole-economy model of the health co-benefits of strategies to reduce greenhouse gas emissions in the UK

BackgroundThe UK Government has set specific targets for greenhouse gas emissions to lower the risk of dangerous climate change. Previous research has shown that important health co-benefits could result from strategies targeting the domains of transport, built environment, and agriculture. This study assesses the full general equilibrium economy-wide macroeconomic effects of health co-benefits from three similar UK strategies to meet locally specific 2030 greenhouse gas emission targets. MethodsEconomy-wide effects of health co-benefits were modelled with a dynamic extension of the widely used International Food Policy Research Institute standard computable general equilibrium model for 2011–30. Four forms of economic agents are modelled: firms (who combine resource inputs to maximise profits), consumers (who consume and save to maximise their welfare), government, and foreign agents. The method consists of simulation of three greenhouse gas policy scenarios and a counterfactual do-nothing scenario. Basic health co-benefits (years lived with disability [YLD] and years of life lost [YLL]) were measured for a range of illnesses, on the basis of the comparative risk assessment approach. Combined with incidence numbers and prevalence trends, these basic YLD (morbidity) and YLL (mortality) co-benefits were used to calculate dynamic sequences of demographic and labour market effects on population and productive labour supply, and public budget implications for averted health-care costs and increased social security transfers (including benefit payments for working-age individuals and pension payments for old-age individuals). These economic shocks were subsequently imposed on the computable general equilibrium model and used to measure the combined macroeconomic effect of health co-benefits. The method for measuring averted health-care costs was published in The Lancet in 2012. Three scenarios were modelled: active travel (transport sector; health co-benefits of an assumed transformation of urban transport behaviour to reduce motorised transportation and increase walking [2·5-times] and cycling [8·0-times] in urban England and Wales); healthy diet (food and agriculture sector; health co-benefits of an assumed UK-wide 30% reduction in consumption of dietary saturated fat); and household energy (household energy sector; health co-benefits of an assumed UK-wide improvement in home insulation and ventilation, including reduced household energy use, improved indoor temperature, and associated changes in indoor pollutants). FindingsFor all scenarios, the macroeconomic effects of health co-benefits are positive. Overall, substantial savings on health-care costs represent the main contributing factor. Increased labour supplies also contribute positively, whereas increased social security transfers (due to larger working-age and old-age population segments) detract. The largest potential cumulative gross domestic product gains from health co-benefits are associated with the active travel scenario (around £19 billion), in which increased physical activity averts large-scale and long-term chronic disease burdens and health-system costs. The healthy diet scenario also leads to important potential gains (around £5 billion), whereas the full potential health co-benefits from the household energy scenario will not be realised until beyond 2030. Three economic sensitivity analyses were undertaken to test the sensitivity of results to variations in assumptions concerning: the substitutability of labour for other factors of production; the effectiveness of the interventions; and changes in the discount rate (the present value of the economic effects). Overall, the core results can be considered relatively robust to changes in these three factors. InterpretationStrategies to reduce greenhouse gas emissions and improve health are likely to result in substantial and increasing positive contributions to the economy. This effect might offset some economic costs and thereby allow such strategies to be seen more favourably, especially in times of economic austerity. FundingDepartment of Health Policy Research Programme.

Premature death after self-harm: a multicentre cohort study

People who self-harm have an increased risk of premature death. The aim of this study was to investigate cause-specific premature death in individuals who self-harm, including associations with socioeconomic deprivation. We undertook a cohort study of patients of all ages presenting to emergency departments in Oxford, Manchester, and Derby, UK, after self-poisoning or self-injury between Jan 1, 2000, and Dec 31, 2007. Postcodes of individuals' place of residence were linked to the Index of Multiple Deprivation 2007 in England. Mortality information was supplied by the Medical Research Information Service of the National Health Service. Patients were followed up to the end of 2009. We calculated age-standardised mortality ratios (SMRs) and years of life lost (YLL), and we tested for associations with socioeconomic deprivation. 30 950 individuals presented with self-harm and were followed up for a median of 6·0 years (IQR 3·9-7·9). 1832 (6·1%) patients died before the end of follow-up. Death was more likely in patients than in the general population (SMR 3·6, 95% CI 3·5-3·8), and occurred more in males (4·1, 3·8-4·3) than females (3·2, 2·9-3·4). Deaths due to natural causes were 2-7·5 times more frequent than was expected. For individuals who died of any cause, mean YLL was 31·4 years (95% CI 30·5-32·2) for male patients and 30·7 years (29·5-31·9) for female patients. Mean YLL for natural-cause deaths was 25·9 years (25·7-26·0) for male patients and 25·5 years (25·2-25·8) for female patients, and for external-cause deaths was 40·2 years (40·0-40·3) and 40·0 years (39·7-40·5), respectively. Disease of the circulatory (13·1% in males; 13·0% in females) and digestive (11·7% in males; 17·8% in females) systems were major contributors to YLL from natural causes. All-cause mortality increased with each quartile of socioeconomic deprivation in male patients (χ(2) trend 39·6; p<0·0001), female patients (13·9; p=0·0002), and both sexes combined (55·4; p<0·0001). Socioeconomic deprivation was related to mortality in both sexes combined from natural causes (51·0; p<0·0001) but not from external causes (0·30; p=0·58). Alcohol problems were associated with death from digestive-system disease, drug misuse with mental and behavioural disorders, and physical health problems with circulatory-system disease. Physical health and life expectancy are severely compromised in individuals who self-harm compared with the general population. In the management of self-harm, clinicians assessing patients' psychosocial problems should also consider their physical needs. Department of Health Policy Research Programme.