Mogens Trab Damsgaard

The mediation proportion: a structural equation approach for estimating the proportion of exposure effect on outcome explained by an intermediate variable.

It is often of interest to assess how much of the effect of an exposure on a response is mediated through an intermediate variable. However, systematic approaches are lacking, other than assessment of a surrogate marker for the endpoint of a clinical trial. We review a measure of “proportion explained” in the context of observational epidemiologic studies. The measure has been much debated; we show how several of the drawbacks are alleviated when exposures, mediators, and responses are continuous and are embedded in a structural equation framework. These conditions also allow for consideration of several intermediate variables. Binary or categorical variables can be included directly through threshold models. We call this measure the mediation proportion, that is, the part of an exposure effect on outcome explained by a third, intermediate variable. Two examples illustrate the approach. The first example is a randomized clinical trial of the effects of interferon-α on visual acuity in patients with age-related macular degeneration. In this example, the exposure, mediator and response are all binary. The second example is a common problem in social epidemiology—to find the proportion of a social class effect on a health outcome that is mediated by psychologic variables. Both the mediator and the response are composed of several ordered categorical variables, with confounders present. Finally, we extend the example to more than one mediator.

Pathways and mechanisms in adolescence contribute to adult health inequalities

Aims: This paper presents a model that encompasses pathways and mechanisms working over adolescence that contribute to adult health inequalities. We review evidence on the four mechanisms: socially differential exposure, tracking, socially differential tracking, and socially differential vulnerability. Methods: We conducted literature searches in English-language peer-reviewed journals using PubMed (from 1966 to May 2009) and PsycINFO, and combined these with hand-searches of reference lists, journals, and authors of particular relevance. Results: Most health indicators are socially patterned in adolescence and track into adulthood, with higher risks of adverse outcomes among individuals from lower socioeconomic positions. Adolescent health behaviours track into adulthood. Smoking, physical activity, and especially fruit and vegetable intake are socially patterned, while evidence for social patterning of alcohol use is less consistent. Relational dimensions like lone parenthood and bullying are socially patterned and track over time, and there are indications of a socially differential vulnerability to the effects of these types of relational strain. Very little research has investigated the social patterning of the above indicators over time or studied social vulnerability of these indicators from adolescence to adulthood. However, all four mechanisms seem to be active in establishing social differences in adult educational attainment. Conclusions: We find the Adolescent Pathway Model useful for providing an overview of what elements and mechanisms in adolescence may be of special importance for adult health inequalities. There is a lack of knowledge of how social patterns of health, health behaviours, and social relations in adolescence transfer into adulthood and to what extent they reflect themselves in adult health.

Social position and health in old age: the relevance of different indicators of social position

Aims: An analysis was undertaken to investigate social inequalities in health among old men and women in relation to five indicators of social position. Methods: The study is based on a population-based cross-sectional survey among 748 75-year-old men and women, which was performed as clinical examinations and interviews in 1989 in Glostrup, a suburban area west of Copenhagen. Social position was measured by vocational education, occupation, social class, income, and housing tenure. Health was measured by number of chronic diseases, tiredness in relation to mobility, need of help in relation to mobility, oral health (number of teeth), and well-being (the CES-D Scale). The statistical analysis included bivariate contingency tables and logistic regression analyses. Results: Two material wealth variables (income and tenure) were consistently related to nearly all health measures while the relationships between the other social position variables and health showed no consistent patterns. Multiple logistic regression analyses with tenure and income as independent variables and each of the health variables as dependent variables and control for education and occupation showed different patterns for men and women. In men the odds ratios of housing tenure on four health variables were strong and unaffected by education and occupation while in women the odds ratios of income on three health variables were strong and unaffected by education and occupation. Conclusion: This study demonstrates strong, consistent associations between variables of material wealth indicators and various measures of health among 75-year-old men and women.

Hospital admissions before and after shipyard closure.

To determine the effect of job loss on health an investigation was made of admissions to hospitals in 887 men five years before and three years after the closure of a Danish shipyard. The control group comprised 441 men from another shipyard. The information on hospital admissions was obtained from the Danish national register of patients. The relative risk of admission in the control group dropped significantly in terms of the number of men admitted from the study group from 1.29 four to five years before closure to 0.74 in the three years after closure. This was especially true of admissions due to accidents (1.33 to 0.46) and diseases of the digestive system (4.53 to 1.03). For diseases of the circulatory system, particularly cardiovascular diseases, the relative risk increased from 0.8 to 1.60, and from 1.0 to 2.6 respectively. These changes in risk of illness after redundancy are probably a consequence of a change from the effects of a high risk work environment to the effects of psychosocial stresses such as job insecurity and unemployment.

Cynical hostility, socioeconomic position, health behaviors, and symptom load: a cross-sectional analysis in a Danish population-based study.

Objective: To analyze the cross-sectional association between cynical hostility and high symptom load in a Danish population-based study. Furthermore, the aim was to investigate to what extent health risk behaviors mediated this association. Methods: Data were based on a postal questionnaire in a Danish random sample of 3426 men and 3699 women aged 40 or 50 years. Cynical hostility was measured by the 8-item Cynical Distrust Scale. High symptom load was assessed by physiological and mental symptoms experienced within the last 4 weeks. Confounders were age and socioeconomic position, while potential mediators were alcohol consumption, smoking, physical activity, and BMI. Results: Higher cynical hostility was associated with self-reported symptom load. Health behaviors did not seem to mediate this effect. Socioeconomic position was a strong confounder for the effect on both health and health behaviors. After adjustment the effects of hostility on health remained with odds ratios of 2.1 (1.7–2.6) for women and 2.3 (1.8–2.8) for men. Conclusion: After adjustment for socioeconomic position, cynical hostility has an effect on self-reported high symptom load, and this effect is not mediated by health behaviors.

Socioeconomic position and smoking behaviour in Danish adults

Aims: The associations between smoking and various socioeconomic indicators may have different implications and causes, which may also vary according to sex and birth cohort. This study analyses how two dimensions of socioeconomic position, an individual ( education) and a structural ( occupation) indicator, are associated with ever, current and ex-smoking. Methods: Data on smoking behaviour were collected in five cross-sectional surveys of random samples of the general Danish population aged 20 years or more at intervals between 1982 and 1994. In total, 8,054 men and 8,281 women participated. Logistic regression was used to analyse the infl uence of education and occupation on smoking behaviour controlling for sex and birth cohort. Results: In cohorts born after 1930 ever and current smoking were related to years of school education and current occupation. The prevalences of ever and current smoking were highest among the least educated, unskilled workers, unemployed persons and persons who received welfare benefits. A significant interaction between birth cohort and education indicated that the educational difference in ever and current smoking increased significantly with increasing year of birth. In multivariate analysis controlling for sex and birth cohort, ex-smoking seemed to be more strongly associated with education than occupation .Those with 12 or more years of school education had twice as high a chance of being ex-smokers as those with 7 years of school or less. Conclusion: Smoking behaviour is strongly associated with both individual and structural indicators of socioeconomic position in Danish adults in all cohorts except for those born before 1930.

Postthrombotic syndrome and quality of life in patients with iliofemoral venous thrombosis treated with catheter-directed thrombolysis

BACKGROUNDnPostthrombotic syndrome (PTS) is a common complication after iliofemoral venous thrombosis, often resulting in poor quality of life (QOL) among the affected patients. This study assessed development of PTS and its effect on QOL among patients treated for iliofemoral venous thrombosis by catheter-directed thrombolysis.nnnMETHODSnPatients admitted with an iliofemoral venous thrombosis and treated with catheter-directed thrombolysis at Gentofte University Hospital from 1999 to 2008 were invited to participate. Duplex ultrasound imaging was used to assess venous patency and valve function. Each patient completed the generic Short-Form 36-item (SF-36) health survey assessment, producing physical component (PCS) and mental component summary (MCS) scores, and the disease-specific Venous Insufficiency Epidemiological and Economic Study (VEINES)-Quality of Life (QOL)/Symptoms (Sym), questionnaires to assess QOL. PTS was assessed using the Villalta scale.nnnRESULTSnThe study included 109 patients. Median follow-up was 71 months. PTS developed in 18 patients (16.5%) and of those, initial thrombolysis was successful in 13. Patients with PTS had significantly worse mean ± standard deviation scores than patients without PTS on VEINES-QOL (34.2 ± 9.6 vs 53.1 ± 6.6; P < .0001), VEINES-Sym (34.0 ± 8.8 vs 53.2 ± 6.6; P < .0001), SF-36 MCS (44.2 ± 15.5 vs 52.3 ± 11.0; P = .005), and SF-36 PCS (42.3 ± 9.1 vs 53.5 ± 7.8; P < .0001) subscales. Patients with reflux or chronic occlusions, or both, had significantly lower mean ± SD scores than patients with patent veins without reflux on VEINES-QOL (43.5 ± 14.3 vs 51.0 ± 8.8; P = .044) and SF-36 PCS (47.2 ± 10.9 vs 52.4 ± 8.5; P = .049) scales.nnnCONCLUSIONnPTS was associated with worse QOL, although only a few patients developed PTS after catheter-directed thrombolysis of iliofemoral venous thrombosis. Patients with patent veins and sufficient valves have higher QOL scores than patients with reflux and occluded veins.

Trends in overweight prevalence among 11-, 13- and 15-year-olds in 25 countries in Europe, Canada and USA from 2002 to 2010

BACKGROUNDnThe purpose of this study was to assess recent changes in the prevalence of overweight (including obesity) among 11-, 13- and 15-year-olds in 33 countries from 2002 to 2010.nnnMETHODSnData from 25 countries from three consecutive survey cycles (2002, 2006 and 2010) that had at least 80% response rate for self-reported height, weight and age were analysed using logistic regression analysis.nnnRESULTSnOverweight prevalence increased among boys in 13 countries and among girls in 12 countries; in 10 countries, predominantly in Eastern Europe, an increase was observed for both boys and girls. Stabilization in overweight rates was noted in the remaining countries; none of the countries exhibited a decrease over the 8-year period examined. In the majority of countries (20/25) there were no age differences in trends in overweight prevalence.nnnCONCLUSIONnIn over half of the countries examined overweight prevalence did not change during 2002-2010. However, increasing overweight prevalence was noted in many Eastern European countries over this time period. Overweight prevalence remained high in several countries in Europe and North America. These patterns call for continued research in youth overweight and highlight the need to understand cross-national differences by examining macro-level indicators. Such research should feed into developing sound translations and practices to prevent and reduce overweight in youth.

Adolescent alcohol use: a reflection of national drinking patterns and policy?

AIMSnTo analyse how adolescent drunkenness and frequency of drinking were associated with adult drinking patterns and alcohol control policies.nnnDESIGN, SETTING AND PARTICIPANTSnCross-sectional survey data on 13- and 15-year-olds in 37 countries who participated in the Health Behaviour in School-Aged Children (HBSC) Study in 2010 (nu2009=u2009144u2009788) were linked to national-level indicators on alcohol control policies and adult drinking patterns.nnnMEASUREMENTSnOutcome measures were self-reported weekly drinking and life-time drunkenness (drunk once or more). Data were analysed using multi-level logistic regression models.nnnFINDINGSnIn the mutually adjusted models, adolescent drunkenness was associated significantly with high adult alcohol consumption [odds ratio (OR)u2009=u20093.15 among boys, 95% confidence interval (CI)u2009=u20092.13-4.64, OR girlsu2009=u20092.44, CIu2009=u20091.57-3.80] and risky drinking patterns in the adult population (OR boysu2009=u20092.02, CIu2009=u20091.33-3.05, OR girlsu2009=u20091.61, CIu2009=u20091.18-2.18). The level of abstainers in the adult population was also associated significantly with girls drunkenness; a 10% increase in the number of abstainers in a country reduced the odds of drunkenness with 21% (ORu2009=u20090.79, CIu2009=u20090.68-0.90). Weekly drinking was associated significantly with weak restrictions on availability (OR boysu2009=u20092.82, CIu2009=u20091.74-4.54, OR girlsu2009=u20092.00, CIu2009=u20091.15-3.46) and advertising (OR boysu2009=u20091.56, CIu2009=u20091.02-2.40, OR girlsu2009=u20091.79, CIu2009=u20091.10-2.94).nnnCONCLUSIONSnComparing data cross-nationally, high levels of adult alcohol consumption and limited alcohol control policies are associated with high levels of alcohol use among adolescents.

European study of frequency of participation of adolescents with and without cerebral palsy

Children with cerebral palsy participate less in everyday activities than children in the general populations. During adolescence, rapid physical and psychological changes occur which may be more difficult for adolescents with impairments. Within the European SPARCLE project we measured frequency of participation of adolescents with cerebral palsy by administering the Questionnaire of Young Peoples Participation to 667 adolescents with cerebral palsy or their parents from nine European regions and to 4666 adolescents from the corresponding general populations. Domains and single items were analysed using respectively linear and logistic regression. Adolescents with cerebral palsy spent less time with friends and had less autonomy in their daily life than adolescents in the general populations. Adolescents with cerebral palsy participated much less in sport but played electronic games at least as often as adolescents in the general populations. Severity of motor and intellectual impairment had a significant impact on frequency of participation, the more severely impaired being more disadvantaged. Adolescents with an only slight impairment participated in some domains as often as adolescents in the general populations. Regional variation existed. For example adolescents with cerebral palsy in central Italy were most disadvantaged according to decisional autonomy, while adolescents with cerebral palsy in east Denmark and northern England played sports as often as their general populations. Participation is an important health outcome. Personal and environmental predictors of participation of adolescents with cerebral palsy need to be identified in order to design interventions directed to such predictors; and in order to inform the content of services.