Showing papers by "George Davey Smith published in 1997"

Bias in meta-analysis detected by a simple, graphical test

Abstract: Objective: Funnel plots (plots of effect estimates against sample size) may be useful to detect bias in meta-analyses that were later contradicted by large trials. We examined whether a simple test of asymmetry of funnel plots predicts discordance of results when meta-analyses are compared to large trials, and we assessed the prevalence of bias in published meta-analyses. Design: Medline search to identify pairs consisting of a meta-analysis and a single large trial (concordance of results was assumed if effects were in the same direction and the meta-analytic estimate was within 30% of the trial); analysis of funnel plots from 37 meta-analyses identified from a hand search of four leading general medicine journals 1993-6 and 38 meta-analyses from the second 1996 issue of the Cochrane Database of Systematic Reviews . Main outcome measure: Degree of funnel plot asymmetry as measured by the intercept from regression of standard normal deviates against precision. Results: In the eight pairs of meta-analysis and large trial that were identified (five from cardiovascular medicine, one from diabetic medicine, one from geriatric medicine, one from perinatal medicine) there were four concordant and four discordant pairs. In all cases discordance was due to meta-analyses showing larger effects. Funnel plot asymmetry was present in three out of four discordant pairs but in none of concordant pairs. In 14 (38%) journal meta-analyses and 5 (13%) Cochrane reviews, funnel plot asymmetry indicated that there was bias. Conclusions: A simple analysis of funnel plots provides a useful test for the likely presence of bias in meta-analyses, but as the capacity to detect bias will be limited when meta-analyses are based on a limited number of small trials the results from such analyses should be treated with considerable caution. Key messages Systematic reviews of randomised trials are the best strategy for appraising evidence; however, the findings of some meta-analyses were later contradicted by large trials Funnel plots, plots of the trials9 effect estimates against sample size, are skewed and asymmetrical in the presence of publication bias and other biases Funnel plot asymmetry, measured by regression analysis, predicts discordance of results when meta-analyses are compared with single large trials Funnel plot asymmetry was found in 38% of meta-analyses published in leading general medicine journals and in 13% of reviews from the Cochrane Database of Systematic Reviews Critical examination of systematic reviews for publication and related biases should be considered a routine procedure

Meta-analysis: Principles and procedures

Abstract: Meta-analysis is a statistical procedure that integrates the results of several independent studies considered to be “combinable.”1 Well conducted meta-analyses allow a more objective appraisal of the evidence than traditional narrative reviews, provide a more precise estimate of a treatment effect, and may explain heterogeneity between the results of individual studies.2 Ill conducted meta-analyses, on the other hand, may be biased owing to exclusion of relevant studies or inclusion of inadequate studies.3 Misleading analyses can generally be avoided if a few basic principles are observed. In this article we discuss these principles, along with the practical steps in performing meta-analysis. Meta-analysis should be viewed as an observational study of the evidence. The steps involved are similar to any other research undertaking: formulation of the problem to be addressed, collection and analysis of the data, and reporting of the results. Researchers should write in advance a detailed research protocol that clearly states the objectives, the hypotheses to be tested, the subgroups of interest, and the proposed methods and criteria for identifying and selecting relevant studies and extracting and analysing information. As with criteria for including and excluding patients in clinical studies, eligibility criteria have to be defined for the data to be included. Criteria relate to the quality of trials and to the combinability of treatments, patients, outcomes, and lengths of follow up. Quality and design features of a study can influence the results.4 5 Ideally, researchers should consider including only controlled trials with proper randomisation of patients that report on all initially included patients according to the intention to treat principle and with an objective, preferably blinded, outcome assessment.6 Assessing the quality of a study …

Lifetime socioeconomic position and mortality: prospective observational study

Abstract: OBJECTIVES: To assess the influence of socioeconomic position over a lifetime on risk factors for cardiovascular disease, on morbidity, and on mortality from various causes. DESIGN: Prospective observational study with 21 years of follow up. Social class was determined as manual or non-manual at three stages of participants' lives: from the social class of their father's job, the social class of their first job, and the social class of their job at the time of screening. A cumulative social class indicator was constructed, ranging from non-manual social class at all three stages of life to manual social class at all three stages. SETTING: 27 workplaces in the west of Scotland. PARTICIPANTS: 5766 men aged 35-64 at the time of examination. MAIN OUTCOME MEASURES: Prevalence and level of risk factors for cardiovascular disease; morbidity; and mortality from broad causes of death. RESULTS: From non-manual social class locations at all three life stages to manual at all stages there were strong positive trends for blood pressure, body mass index, current cigarette smoking, angina, and bronchitis. Inverse trends were seen for height, cholesterol concentration, lung function, and being an ex-smoker. 1580 men died during follow up. Age adjusted relative death rates in comparison with the men of non-manual social class locations at all three stages of life were 1.29 (95% confidence interval 1.08 to 1.56) in men of two non-manual and one manual social class; 1.45 (1.21 to 1.73) in men of two manual and one non-manual social class; and 1.71 (1.46 to 2.01) in men of manual social class at all three stages. Mortality from cardiovascular disease showed a similar graded association with cumulative social class. Mortality from cancer was mainly raised among men of manual social class at all three stages. Adjustment for a wide range of risk factors caused little attenuation in the association of cumulative social class with mortality from all causes and from cardiovascular disease; greater attenuation was seen in the association with mortality from non-cardiovascular, non-cancer disease. Fathers having a manual [corrected] occupation was strongly associated with mortality from cardiovascular disease: relative rate 1.41 (1.15 to 1.72). Participants' social class at the time of screening was more strongly associated than the other social class indicators with mortality from cancer and from non-cardiovascular, non-cancer causes. CONCLUSIONS: Socioeconomic factors acting over the lifetime affect health and risk of premature death. The relative importance of influences at different stages varies for the cause of death. Studies with data on socioeconomic circumstances at only one stage of life are inadequate for fully elucidating the contribution of socioeconomic factors to health and mortality risk.

Meta-analysis: Potentials and promise

Abstract: The number of papers published on meta-analyses in medical research has increased sharply in the past 10 years (fig 1). The merits and perils of the somewhat mysterious procedure of meta-analysis, however, continue to be debated in the medical community.1 23 What, then, is meta-analysis? A useful definition was given by Huque: “A statistical analysis that combines or integrates the results of several independent clinical trials considered by the analyst to be ‘combinable.’” 4 The terminology, however, is still debated, and expressions used concurrently include “overview,” “pooling,” and “quantitative synthesis.” We believe that the term meta-analysis should be used to describe the statistical integration of separate studies, whereas “systematic review” is most appropriate for denoting any review of a body of data that uses clearly defined methods and criteria (box). Systematic reviews can include meta-analyses, appraisals of single trials, and other sources of evidence.6 In this article we examine the potentials and promise of meta-analysis of randomised controlled trials. In later articles of this series we will consider the practical steps involved in meta-analysis,7 examine various extensions beyond the calculation of a combined estimate,8 address potential biases and discuss strategies to detect and minimise the influence of these in meta-analysis of randomised trials9 and of observational studies.10 We will conclude with a discussion of unresolved issues and future developments.11 Details of relevant software will appear on the BMJ's website at the end of the series. Fig 1 Number of publications about meta-analysis, 1987–96 (results from Medline search using text word and medical subject heading “meta-analysis”) #### What's in a name? The case for “meta-analysis” The term meta-analysis for statistically combining and analysing data from separate studies is appropriate because:

Systematic review of randomised controlled trials of multiple risk factor interventions for preventing coronary heart disease.

Abstract: Objective: To assess the effectiveness of multiple risk factor intervention in reducing cardiovascular risk factors, total mortality, and mortality from coronary heart disease among adults. Design: Systematic review and meta-analysis of randomised controlled trials in workforces and in primary care in which subjects were randomly allocated to more than one of six interventions (stopping smoking, exercise, dietary advice, weight control, antihypertensive drugs, and cholesterol lowering drugs) and followed up for at least six months. Subjects: Adults aged 17-73 years. 903 000 person years of observation were included in nine trials with clinical event outcomes and 303 000 person years in five trials with risk factor outcomes alone. Main outcome measures: Changes in systolic and diastolic blood pressure, smoking rates, blood cholesterol concentrations, total mortality, and mortality from coronary heart disease. Results: Net decreases in systolic and diastolic blood pressure, smoking prevalence, and blood cholesterol were 4.2 mm Hg (SE 0.19 mm Hg), 2.7 mm Hg (0.09 mm Hg), 4.2% (0.3%), and 0.14 mmol/l (0.01 mmol/l) respectively. In the nine trials with clinical event end points the pooled odds ratios for total and coronary heart disease mortality were 0.97 (95% confidence interval 0.92 to 1.02) and 0.96 (0.88 to 1.04) respectively. Statistical heterogeneity between the studies with respect to changes in mortality and risk factors was due to trials focusing on hypertensive participants and those using considerable amounts of drug treatment, with only these trials showing significant reductions in mortality. Conclusions: The pooled effects of multiple risk factor intervention on mortality were insignificant and a small, but potentially important, benefit of treatment (about a 10% reduction in mortality) may have been missed. Changes in risk factors were modest, were related to the amount of pharmacological treatment used, and in some cases may have been overestimated because of regression to the mean, lack of intention to treat analyses, habituation to blood pressure measurement, and use of self reports of smoking. Interventions using personal or family counselling and education with or without pharmacological treatments seem to be more effective at reducing risk factors and therefore mortality in high risk hypertensive populations. The evidence suggests that such interventions implemented through standard health education methods have limited use in the general population. Health protection through fiscal and legislative measures may be more effective. Key messages The effectiveness of health education approaches modifying lifestyle to prevent coronary heart disease is in doubt Health promotion interventions result in only small changes in risk factors and mortality in the general population In people with hypertension and in other high risk groups risk factor interventions have beneficial effects Health protection by fiscal and legislative means deserves a higher priority

Socio-economic differentials in health: the role of nutrition

Abstract: This paper examines the role of nutrition in generating inequalities in health. While the precise role is difficult to delineate, in several cases, nutritional deficiencies or excesses which will generate health problems are socially patterned in such a way to further disadvantage the poor. This paper discusses differentials in mortality and morbidity, and their origins; social class, central obesity, and metabolism; social patterning of diet and nutrition; and life-course influences on health inequalities.

Meta-analysis: Beyond the grand mean?

Abstract: In the previous two articles1 2 we outlined the potentials and principles of meta-analysis and the practical steps in performing a meta-analysis. Now we will examine how to use meta-analysis to do more than simply combine the results from all the individual trials into a single effect estimate. Firstly, we discuss the advantages and disadvantages of performing subgroup analyses. Secondly, we consider the situation in which the differences in effects between individual trials are related in a graded way to an underlying phenomenon, such as the degree of mortality risk of the trial participants. #### Summary points Meta-analysis can be used to examine differences in treatment effects across trials; however, the fact that randomised trials are included in meta-analyses does not mean that comparisons between trials are also randomised comparisons Meta-analytic subgroup analyses, like subgroup analyses within trials, are prone to bias and need to be interpreted with caution A more reliable way of assessing differences in treatment effects is to relate outcome to some underlying patient characteristic on a continuous, or ordered, scale The underlying level of risk is a key variable which is often related to a given treatment effect, with patients at higher risk receiving more benefit then low risk patients Individual patient data, rather than published summary statistics, are often required for meaningful subgroup analyses The main aim of a meta-analysis is to produce an estimate of the average effect seen in trials of a particular treatment. The direction and magnitude of this average effect is intended to guide decisions about clinical practice for a wide range of patients. Clinicians are thus being asked to treat their patients as though each one is well represented by the patients in the clinical trials included in the meta-analysis. This runs against doctors' concerns to use the specific characteristics of a …

Poor fetal nutrition causes long-term changes in expression of insulin signaling components in adipocytes

Abstract: Insulin action on adipocytes was studied in the offspring of mothers who had been fed either a control (20% protein) or a low (8%)-protein diet during pregnancy and lactation. Adipocytes isolated from low-protein offspring had significantly higher basal and insulin-stimulated glucose uptakes than controls. This may be related to a threefold increase in insulin receptors in low-protein adipocytes. Consistent with these observed changes in glucose transport, adipocytes from low-protein animals had significantly higher basal and insulin-stimulated insulin receptor substrate (IRS)-1-associated phosphatidylinositol 3-kinase (PI 3-kinase) activities. There was also more p85-associated PI 3-kinase activity in these adipocytes. There was no difference in expression in the p85 regulatory subunit or the p110-alpha catalytic subunit of PI 3-kinase. In contrast, there was a sixfold reduction in the p110-beta catalytic subunit of PI 3-kinase in adipocytes from low-protein animals. These results suggest that poor fetal nutrition during pregnancy and lactation can have long-term effects on glucose transport and on the expression of key components of the insulin signaling pathway in adipocytes.

Disease aetiology and materialist explanations of socioeconomic mortality differentials

Abstract: The social structure can affect health by distributing exposure to environmental hazards and by conditioning behaviours which damage or promote health. The present paper concentrates on the former route, which the Black Report described as the structural or materialist type of explanation of health inequalities. The contribution of materialist factors is assessed by examining socioeconomic mortality differentials as a form of relative deprivation and by presenting evidence of the sensitivity of these differentials to multiple indicators of social position. A definition of materialist factors is developed which generates hypotheses about disease causation. The literature on the aetiology of disease is reviewed to estimate the proportion of all deaths attributable to the materialist factors so-defined. An apparent discrepancy is identified between the lesser importance attributed to materialist factors in the aetiological literature and that suggested by the evidence initially presented. This discrepancy, it is argued, may be partly explained by artefactual processes but also reflects the complexity of the mechanisms by which socioeconomic gradients in disease are produced. An understanding of the processes through which materialist factors combine and accumulate over the course of life is an important target for understanding disease aetiology as well as socioeconomic differences in health.

The relationship between socioeconomic status, hostility, and blood pressure reactions to mental stress in men: Data from the Whitehall II study.

Abstract: This study investigated the relationship among blood pressure reactions to mental stress, cynical hostility, and socioeconomic status (SES) in 1,091 male public servants. Occupational grade served to index SES and cynical hostility was assessed using the Cook-Medley scale. (Cook & Medley, 1954). The magnitude of systolic, but not diastolic, blood pressure change scores to stress was positively associated with occupational grade: the higher the grade, the greater the reactions. Mental stress task performance also varied with occupational grade but was unrelated to reactivity. Ratings of task difficulty aid not vary with occupational grade. Cynical hostility-was negatively related to occupational grade, and, contrary to previous findings, negatively related to systolic blood pressure reactivity. Cynical hostility was also negatively related to mental stress task performance but unrelated to ratings of task difficulty.

Socio-economic Differentials in Health

Abstract: The first Whitehall Study of British civil servants demonstrated an inverse gradient in mortality. The lower the grade the higher was the mortality risk. This higher mortality risk applies to most but not all causes of death. The Whitehall II Study, set up to investigate causes of this social gradient shows similarly marked gradients in morbidity. A review of potential causes of the gradient suggests that it is due neither to health selection nor simply to differences in lifestyle, but that relative deprivation-a psychosocial concept-and the accumulation of socially-patterned exposures over the life course, must be important. Whitehall II suggests that the operation of these factors is to be found in the specific circumstances under which people grow, live and work.

Coffee consumption and coronary heart disease mortality in Scottish men: a 21 year follow up study.

Abstract: The effects of coffee consumption on coronary heart disease (CHD) mortality remain uncertain. Recent meta-analyses\"2 have produced equivocal findings, with weaker associations being seen in cohort than in case-control studies. Data regarding potential confounding factors have often been limited in these earlier studies. We have examined this issue in a large prospective study with a long follow up from the west of Scotland.

The effects of recent food, alcohol, and tobacco intake and the temporal scheduling of testing on cardiovascular activity at rest and during psychological stress

Abstract: Participants in cardiovascular psychophysiological studies receive various instructions regarding restraints on alcohol, food, and tobacco intake prior to testing. Further, although the hour of testing is frequently a concern, little attention is paid to month of the year. In the present study, we examined cardiovascular activity at rest and in response to a laboratory stress task in 1,272 men in terms of these variables. Alcohol consumption among alcohol users and food intake within 4 hr of testing were associated with higher baseline diastolic blood pressure (BP) but lower baseline heart rate (HR). Tobacco use in smokers within 1 hr of testing was associated with elevated baseline BP and HR. Hour of testing was related to baseline HR, whereas month of the year was associated with baseline BP. In contrast to these baseline effects, cardiovascular reactions to the laboratory stress task were largely immune to these contextual influences.

Geographical and social class differentials in stroke mortality--the influence of early-life factors: comments on papers by Maheswaran and colleagues.

Abstract: Marked and consistent geographical differences in stroke mortality are seen in many countries. In the United States the \"stroke belt\" in the south east has maintained its relative position, of considerably higher rates than the rest of the country, against a background of persistently declining stroke mortality rates.'2 Rural areas within the US have consistently experienced higher stroke mortality rates over the past 60 years.3 Coronary heart disease (CHD) mortality, conversely, was lower in rural areas in the 1930s, with this advantage having been lost in recent years.3 In the UK, the north west and north Wales have continued to experience higher stroke mortality rates than the south east.45 This does not simply reflect differences in socioeconomic conditions and thus in all cause mortality. The geography of stroke is different to that of CHD, respiratory disease, and all cause mortality, but closely similar to that of stomach cancer mortality, which has also shown declining mortality rates this century.4 When stroke mortality is analysed by geographical region and social class, the geographical differences are seen within each social class grouping,6 and vice versa. The stable geographic distribution of CHD mortality in Norway led Forsdahl to speculate that deprivation in early life, which could be indexed by the infant mortality rate earlier this century, increased susceptibility to CHD in adult life.7 This work was replicated and extended to other causes of death, including stroke and respiratory disease, and indicators of influences acting at particular stages of early development, including maternal mortality, neonatal mortality, and postneonatal mortality.8 Stroke mortality was strongly related to maternal and neonatal mortality, suggesting a particular influence of intrauterine development, whereas CHD mortality was related to all three past indicators, suggesting an influence of intrauterine and infant environment. Chronic bronchitis was associated mainly with postneonatal mortality, indicating an important contribution of environmental conditions in early infancy. There is also a strong correlation (r= 0.66) between infant mortality rates in the 1930s and current stroke, across the countries included in the WHO Monica project, with no correlation (r= 0. 10) being seen with regard to CHD mortality.9 The hypothesis that early-life factors importantly determine stroke risk in adulthood is an exciting and attractive one. A problem is clearly introduced by the persistence of socioeconomic disadvantage and other environmental influences, many of which have also remained geographically stable over many years. Areas with high rates of neonatal mortality earlier this century are the same areas with high levels of socioeconomic deprivation today. Thus, the strong correlations between current stroke mortality and infant mortality earlier this century (r=0.68 for women and r=0.77 for men) are greatly attenuated when present-day socioeconomic factors are taken into account (r= 0.24 for women and r= 0.16 for men).'0 Prospective studies of individuals, associating low birth weight with elevated blood pressure and stroke mortality risk,\" 12 offer more robust evidence of direct influences of poor development in early life on stroke risk in adulthood, although they do not entirely escape from possible confounding by environmental factors acting later in life. A particularly striking geographical difference in stroke mortality, which has been taken to provide evidence for the importance of early life influences on stroke risk, is the lower mortality for Greater London than the rest of England and Wales.'3 It has been suggested that young women who migrated to London to work in domestic service from rural areas were generally well nourished and thus had well developed offspring.'3 Maheswaran et al'4 have subjected this interpretation to critical scrutiny. They find that the London advantage has decreased over time and is now only seen for those aged 75 years and older, with younger cohorts experiencing higher stroke mortality in London than the surrounding South East Region. Separating cohort effects from period effects is notoriously difficult,'5 but the crossover in stroke mortality allowed the identification of cohorts born since about 1920 as having higher stroke mortality in London, whereas those born before this time had lower stroke mortality. This suggests some change in an influence favouring cohorts in London born before 1920 and favouring cohorts in the South East Region born after 1920. The early-life influences which have been directly related to blood pressure and stroke occurrence in adult life have been birth weight and head circumference at birth.\" 12 Reliable comparative data for London and the remainder of the south east are not available for the time period under consideration. Therefore Maheswaran and colleagues utilised the two Department of Social Medicine, University of Bristol, Canynge Hall, Whiteladies Road, Bristol BS8 2PR G Davey Smith Y Ben-Shlomo

Health and Socio-economic Position: A Commentary.

Abstract: This Special Issue focuses on health variations contingent on socio-economic position. Among the numerous reasons why health psychologists should attend to such variations are their pervasiveness, their magnitude and their continuation into better-off social groups. The latter raises the possibility that psychosocial factors may be involved. Recent data revealing that socio-economic health inequalities have increased in recent years, in parallel with increasing income inequalities in countries such as the UK and USA, provide further reason. Understanding the underlying physical, behavioural and psychosocial mechanisms is clearly an important research mission. However, effective intervention will require structural approaches that directly counter socio-economic disparity.

An evaluation of the cross-linking model for the interaction of insulin with its receptor.

Abstract: The cross-linking model for insulin receptor interactions, in which a single insulin molecule may form a cross-link between an insulin receptor's alpha-subunits, has been expressed as a formal comp...

The spread of disability living allowance

Abstract: This article examines the spread of Disability Living Allowance across a group of low income households in an industrial town in North West England from the introduction of the new benefits in 1992 until the Spring of 1996. We focus on the changes over time of the incidence of Disability Living Allowance in the low income population and look at the relative spread of different elements of the benefit. We speculate on whether the new benefits have provided the most effective way of fulfilling the aims of the White Paper, The Way Ahead , and discuss the implications for social and medical models of disability.

Atomic Scale Study Of Precipitate / Matrix Interfaces in a Metallic Alloy

Abstract: The study of phase transformations involves making assumptions about interfaces within a material in order to apply mathematical models. An example of this is the Gibbs classical theory of nucleation and growth which is based upon the assumption that the precipitate / matrix interface is sharp. Recent developments in atom probe microscopy have made it possible for the first time to characterize complex three-dimensional internal interfaces within materials to sub-nanometre accuracy. We have used the OPoSAP (Optical Position Sensitive Atom Probe) to characterize the precipitate / matrix interface in the dilute Cu-Co system which is a model alloy for the study of homogeneous nucleation and growth. Interface widths derived from radial composition profiles were measured to be 0.9 ± 0.2 nm in size. The effects of thermal energy, positioning inaccuracies, and statistical limitations on the measurement of interface widths are examined through the use of mathematical models and computer simulations. Similar widths are measured for precipitates in the coarsening regime which shows the interface velocity does not affect the interface width.