Alastair H Leyland

Multilevel models: applications to health data.

This paper presents an introductory account of multilevel models, highlighting the potential benefits that may be gained by the use of these methods. It draws on recent applications in health services research that have appeared in the literature. Methodological advances in these statistical techniques have taken place in the field of education, where empirical studies have mainly been concerned with comparing pupil achievement across different schools by exploring the relationship between individual and institutional factors. Although recent widespread availability of suitable software packages has enabled other disciplines to adopt these methods, to date they have received little attention in the health services research literature (the investigation of effects of geographical areas on health being a possible exception) despite their obvious application in many areas of current interest. Key areas that could benefit greatly from these techniques include the exploration of variations in clinical practice, comparisons of institutional performance and resource allocation.

Effects of fully-established Sure Start Local Programmes on 3-year-old children and their families living in England: a quasi-experimental observational study

BACKGROUND Sure Start Local Programmes (SSLPs) are area-based interventions to improve services for young children and their families in deprived communities, promote health and development, and reduce inequalities. We therefore investigated whether SSLPs affect the wellbeing of 3-year-old children and their families. METHODS In a quasi-experimental observational study, we compared 5883 3-year-old children and their families from 93 disadvantaged SSLP areas with 1879 3-year-old children and their families from 72 similarly deprived areas in England who took part in the Millennium Cohort Study. We studied 14 outcomes-childrens immunisations, accidents, language development, positive and negative social behaviours, and independence; parenting risk; home-learning environment; fathers involvement; maternal smoking, body-mass index, and life satisfaction; familys service use; and mothers rating of area. FINDINGS After we controlled for background factors, we noted beneficial effects associated with the programmes for five of 14 outcomes. Children in the SSLP areas showed better social development than those in the non-SSLP areas, with more positive social behaviour (mean difference 0.45, 95% CI 0.09 to 0.80, p=0.01) and greater independence (0.32, 0.18 to 0.47, p<0.0001). Families in SSLP areas showed less negative parenting (-0.90, -1.11 to -0.69, p<0.0001) and provided a better home-learning environment (1.30, 0.75 to 1.86, p<0.0001). These families used more services for supporting child and family development than those not living in SSLP areas (0.98, 0.86 to 1.09, p<0.0001). Effects of SSLPs seemed to apply to all subpopulations and SSLP areas. INTERPRETATION Children and their families benefited from living in SSLP areas. The contrast between these and previous findings on the effect of SSLPs might indicate increased exposure to programmes that have become more effective. Early interventions can improve the life chances of young children living in deprived areas.

Multilevel modelling and public health policy

Background: Multilevel modelling is a statistical technique that extends ordinary regression analysis to the situation where the data are hierarchical. Such data form an increasingly common evidence base for public health policy, and as such it is important that policy makers should be aware of this methodology. Method: This paper therefore lays out the a basic description of multilevel modelling, discusses the problems of alternative approaches, and details the relevance for public health policy before describing which levels are relevant and illustrating the different kinds of hypotheses that can be tested using multilevel modelling. A series of examples is used throughout the paper. These relate to regional variations in the incidence of heart disease, the allocation of health resources, the relationship between neighbourhood disorder and mental health, the demand-control model in occupational health, and a school intervention to prevent cardiovascular disease.

An international study of hospital readmissions and related utilization in Europe and the USA.

This study concerns a comparative analysis of hospital readmission rates and related utilization in six areas, including three European countries (Finland, Scotland and the Netherlands) and three states in the USA (New York, California, Washington State). It includes a data analysis on six major causes of hospitalization across these areas. Its main focus is on two questions. (1) Do hospital readmission rates vary among the causes of hospitalization and the study populations? (2) Are hospital inpatient lengths of stay inversely related to readmissions rates? The study demonstrated that diagnoses such as chronic obstructive pulmonary disease (COPD) and congestive heart failure (CHF) were the major causes of hospital readmission rates. The data showed that (initial) hospital stays were generally longer for patients who were readmitted than for those who were not. As a result, short stays were not associated with a higher risk of readmission, meaning that hospital readmissions were not produced by premature hospital discharges in the study population. Furthermore, the spatial variation in readmission rates within 7 versus 8-30 days showed to be identical. Finally, it was found that countries or states with relatively shorter stays showed higher readmission rates and vice versa. Since patients with readmissions in all of the areas had on average longer initial stays, this finding at country level does illustrate that there seems to be a country specific trade off between length of stay and rate of readmission. An explanation should be sought in differences in health care arrangements per area, including factors that determine length of stay levels and readmission rates in individual countries (e.g. managed care penetration, after care by GPs or home care).

Contribution of smoking-related and alcohol-related deaths to the gender gap in mortality: evidence from 30 European countries

Background Women now outlive men throughout the globe, a mortality advantage that is very established in developed European countries. Debate continues about the causes of the gender gap, although smoking is known to have been a major contributor to the difference in the past. Objectives To compare the magnitude of the gender gap in all-cause mortality in 30 European countries and assess the contribution of smoking-related and alcohol-related deaths. Methods Data on all-cause mortality, smoking-related mortality and alcohol-related mortality for 30 European countries were extracted from the World Health Organization Health for All database for the year closest to 2005. Rates were standardised by the direct method using the European population standard and were for all age groups. The proportion of the gender gap in all-cause mortality attributable to smoking-related and alcohol-related deaths was then calculated. Results There was considerable variation in the magnitude of the male ‘excess’ of all-cause mortality across Europe, ranging from 188 per 100 000 per year in Iceland to 942 per 100 000 per year in Ukraine. Smoking-related deaths accounted for around 40% to 60% of the gender gap, while alcohol-related mortality typically accounted for 20% to 30% of the gender gap in Eastern Europe and 10% to 20% elsewhere in Europe. Conclusions Smoking continues to be the most important cause of gender differences in mortality across Europe, but its importance as an explanation for this difference is often overshadowed by presumptions about other explanations. Changes in smoking patterns by gender suggest that the gender gap in mortality will diminish in the coming decades.

Air pollution, social deprivation, and mortality: a multilevel cohort study.

Background: It is becoming increasingly evident that exposure to air pollution and its adverse effects are not equitably distributed. Our goal was to investigate the role of social deprivation in explaining the effect of neighborhood differences in level of air pollution fine particulates (PM2.5) on mortality when the indicators of social deprivation are measured at both individual level and at neighborhood level. Methods: All inhabitants registered in Oslo, Norway on 1 January 1992 in the age group 50–74 years (n = 105,359) constitute the study base. We used an air dispersion model (AirQUIS) to estimate levels of exposure in the period 1992–1995 in all 470 administrative neighborhoods. These data were linked to Census, educational, and death registries. Deaths were recorded in the period 1992–1998. Main Results: PM2.5 was associated with most neighborhood-level indicators of deprivation, as was most clearly seen for type of dwelling and ownership of dwelling. The effect of PM2.5 on mortality was to some extent explained by these indicators independently of the corresponding individual-level indicators. Conclusions: Findings from this study suggest that socially deprived neighborhoods have higher exposure to air pollution. Deprivation at both the individual and neighborhood level is associated with air pollution, accounting for some of the excess mortality associated with air pollution in these neighborhoods.

Cause-specific inequalities in mortality in Scotland: two decades of change. A population-based study.

BackgroundSocioeconomic inequalities in mortality have increased in recent years in many countries. We examined age-, sex-, and cause-specific mortality rates for social groups in and regions of Scotland to understand the patterning of inequalities and the causes contributing to these inequalities.MethodsWe used death records for 1980–82, 1991–92 and 2000–02 together with mid-year population estimates for 1981, 1991 and 2001 covering the whole of Scotland to calculate directly standardised mortality rates. Deaths and populations were coded to small areas (postcode sectors and data zones), and deprivation was assessed using area based measures (Carstairs scores and the Scottish Index of Multiple Deprivation). We measured inequalities using rate ratios and the Slope Index of Inequality (SII).ResultsSubstantial overall decreases in mortality rates disguised increases for men aged 15–44 and little change for women at the same ages. The pattern at these ages was mostly attributable to increases in suicides and deaths related to the use of alcohol and drugs. Under 65 a 49% fall in the mortality of men in the least deprived areas contrasted with a fall of just 2% in the most deprived. There were substantial increases in the social gradients for most causes of death. Excess male mortality in the Clydeside region was largely confined to more deprived areas, whilst for women in the region mortality was in line with the Scottish experience. Relative inequalities for men and women were greatest between the ages of 30 and 49.ConclusionGeneral reductions in mortality in the major causes of death (ischaemic heart disease, malignant neoplasms) are encouraging; however, such reductions were socially patterned. Relative inequalities in mortality have increased and are greatest among younger adults where deaths related to unfavourable lifestyles call for direct social policies to address poverty.

Female streetworking prostitution and HIV infection in Glasgow.

OBJECTIVES--To identify the extent of HIV infection and injecting drug use among female streetworking prostitutes in Glasgow; to estimate the size of the female streetworking prostitute population in the city; and to estimate the number of HIV positive women working as prostitutes on the streets in Glasgow. DESIGN--Observation and interviewing of female prostitutes over seven months in red light district; analysis of saliva samples for presence of antibodies to HIV; capture-recapture approach to estimating the size of the female streetworking prostitute population. SETTING--Glasgow. SUBJECTS--206 female streetworking prostitutes. MAIN OUTCOME MEASURES--Number of women with antibodies to HIV, self reported use of injecting drugs, history of contact with 206 women. RESULTS--Saliva samples were requested from 197 women; 159 (81%) provided samples. Four (2.5%, 95% confidence interval 0.7%-6.3%) of the samples were positive for HIV, all of which had been provided by women who injected drugs. Of the 206 streetworking women contacted 147 (71%) were injecting drug users. About 1150 women are estimated to work on the streets in Glasgow over a 12 month period. CONCLUSIONS--HIV is not as widespread among female prostitutes as many reports in the tabloid press suggest. A greater proportion of female streetworking prostitutes in Glasgow are injecting drugs than has been reported for other British cities.

Cohort Profile: The Scottish Health Surveys Cohort: linkage of study participants to routinely collected records for mortality, hospital discharge, cancer and offspring birth characteristics in three nationwide studies

Background Although life expectancy is increasing in Scotland, the nation still has the highest rates of coronary heart disease (CHD) and selected malignancies in the UK and higher rates than most countries in Western Europe. The Scottish Health Surveys (SHeSs)—conducted in 1995, 1998 and 2003––were established to provide detailed, contemporary health information on a large, representative sample of the Scottish population. By capturing a range of behavioural, biological, psychological and social characteristics, their purpose was to monitor health in order to assist in policy formulation and the development of new health initiatives across the whole of Scotland.

Spatial clustering of mental disorders and associated characteristics of the neighbourhood context in Malmo, Sweden, in 2001

Study objective: Previous research provides preliminary evidence of spatial variations of mental disorders and associations between neighbourhood social context and mental health. This study expands past literature by (1) using spatial techniques, rather than multilevel models, to compare the spatial distributions of two groups of mental disorders (that is, disorders due to psychoactive substance use, and neurotic, stress related, and somatoform disorders); and (2) investigating the independent impact of contextual deprivation and neighbourhood social disorganisation on mental health, while assessing both the magnitude and the spatial scale of these effects. Design: Using different spatial techniques, the study investigated mental disorders due to psychoactive substance use, and neurotic disorders. Participants: All 89 285 persons aged 40–69 years residing in Malmö, Sweden, in 2001, geolocated to their place of residence. Main results: The spatial scan statistic identified a large cluster of increased prevalence in a similar location for the two mental disorders in the northern part of Malmö. However, hierarchical geostatistical models showed that the two groups of disorders exhibited a different spatial distribution, in terms of both magnitude and spatial scale. Mental disorders due to substance consumption showed larger neighbourhood variations, and varied in space on a larger scale, than neurotic disorders. After adjustment for individual factors, the risk of substance related disorders increased with neighbourhood deprivation and neighbourhood social disorganisation. The risk of neurotic disorders only increased with contextual deprivation. Measuring contextual factors across continuous space, it was found that these associations operated on a local scale. Conclusions: Taking space into account in the analyses permitted deeper insight into the contextual determinants of mental disorders.