Identifying extreme COVID-19 mortality risks in English small areas: a disease cluster approach.

The COVID-19 pandemic is having a huge impact worldwide and has highlighted the extent of health inequalities between countries but also in small areas within a country. Identifying areas with high mortality is important both of public health mitigation in COVID-19 outbreaks, and of longer term efforts to tackle social inequalities in health. In this paper we consider different statistical models and an extension of a recent method to analyze COVID-19 related mortality in English small areas during the first wave of the epidemic in the first half of 2020. We seek to identify hotspots, and where they are most geographically concentrated, taking account of observed area factors as well as spatial correlation and clustering in regression residuals, while also allowing for spatial discontinuities. Results show an excess of COVID-19 mortality cases in small areas surrounding London and in other small areas in North-East and and North-West of England. Models alleviating spatial confounding show ethnic isolation, air quality and area morbidity covariates having a significant and broadly similar impact on COVID-19 mortality, whereas nursing home location seems to be slightly less important.

Does depression diagnosis and antidepressant prescribing vary by location? Analysis of ethnic density associations using a large primary-care dataset.

BACKGROUND: Studies have linked ethnic differences in depression rates with neighbourhood ethnic density although results have not been conclusive. We looked at this using a novel approach analysing whole population data covering just over one million GP patients in four London boroughs. METHOD: Using a dataset of GP records for all patients registered in Lambeth, Hackney, Tower Hamlets and Newham in 2013 we investigated new diagnoses of depression and antidepressant use for: Indian, Pakistani, Bangladeshi, black Caribbean and black African patients. Neighbourhood effects were assessed independently of GP practice using a cross-classified multilevel model. RESULTS: Black and minority ethnic groups are up to four times less likely to be newly diagnosed with depression or prescribed antidepressants compared to white British patients. We found an inverse relationship between neighbourhood ethnic density and new depression diagnosis for some groups, where an increase of 10% own-ethnic density was associated with a statistically significant (p < 0.05) reduced odds of depression for Pakistani [odds ratio (OR) 0.81, 95% confidence interval (CI) 0.70-0.93], Indian (OR 0.88, CI 0.81-0.95), African (OR 0.88, CI 0.78-0.99) and Bangladeshi (OR 0.94, CI 0.90-0.99) patients. Black Caribbean patients, however, showed the opposite effect (OR 1.26, CI 1.09-1.46). The results for antidepressant use were very similar although the corresponding effect for black Caribbeans was no longer statistically significant (p = 0.07). CONCLUSION: New depression diagnosis and antidepressant use was shown to be less likely in areas of higher own-ethnic density for some, but not all, ethnic groups.

Assessing intervention effects in a community-based trial to reduce self-harm: a methodological case study.

This paper considers the assessment of the impact of a community-based randomized controlled trial to reduce repeat deliberate self-harm. It considers the drawbacks in simplistic applications of conventional significance testing procedures, as well as possible failures regarding the statistical assumptions underlying such tests. Instead, the paper considers how relevant prior information might be incorporated within a fully Bayesian-model-based assessment procedure. The model includes a latent trait approach to patient morbidity; controlling for morbidity and other patient characteristics enhances the impact of the intervention (measured by a hazard rate ratio). If allowance is made for external information (e.g. ethical approval of the treatment), the weight of evidence shifts towards a positive intervention effect.

The geographical distribution of specialists in public health in the United Kingdom: is capacity related to need?

OBJECTIVES: Recent organizational changes reflect the need to be more responsive to local populations and have included fostering a closer structural relationship between primary care and public health. In light of this, we explore the distribution of the specialist public health workforce and the relationship with population deprivation and need. STUDY DESIGN: Questionnaire survey to all directors of public health working in primary care trusts (PCTs) and strategic health authorities (SHAs) in England to determine the number of specialists in public health working in either PCTs or SHAs. All identified specialists were given the opportunity to self-define in a further questionnaire survey. Whole-time-equivalent staffing, per head of population, was analysed against socio-economic deprivation, measured by the DETR 2000 Index of Multiple Deprivation. The analysis was conducted at the SHA level. RESULTS: The survey was undertaken whilst public health in the UK was undergoing immense change. This presented specific challenges in identifying specialists in public health working within PCTs and SHAs. Seven hundred and eighty-three specialists working in PCTs and SHAs were identified. On average, in England, there are 1.69 specialists in public health per 100,000 population, with some variability at SHA level (range = 0.8-2.89). Findings indicate an overall positive association between capacity at SHA level and socio-economic need, although some discrepancies between need and provision are apparent. CONCLUSIONS: The general positive association between capacity and deprivation should offer some reassurance to policy makers, researchers and patients alike. However, further efforts are needed to redistribute specialists in some areas to address organizational capacity and equity issues.

Specialist public health capacity in England: working in the new primary care organizations.

OBJECTIVES: To determine the capacity and development needs, in relation to key areas of competency and skills, of the specialist public health workforce based in primary care organizations following the 2001 restructuring of the UK National Health Service. STUDY DESIGN: Questionnaire survey to all consultants and specialists in public health (including directors of public health) based in primary care trusts (PCTs) and strategic health authorities (SHAs) in England. RESULTS: Participants reported a high degree of competency. However, skill gaps were evident in some areas of public health practice, most notably "developing quality and risk management" and in relation to media communication, computing, management and leadership. In general, medically qualified individuals were weaker on community development than non-medically qualified specialists, and non-medically qualified specialists were less able to perform tasks that require epidemiological or clinical expertise than medically qualified specialists. Less than 50% of specialists felt that their links to external organizations, including public health networks, were strong. Twenty-nine percent of respondents felt professionally isolated and 22% reported inadequate team working within their PCT or SHA. Approximately 21% of respondents expressed concerns that they did not have access to enough expertise to fulfil their tasks and that their skills were not being adequately utilized. CONCLUSIONS: Some important skill gaps are evident among the specialist public health workforce although, in general, a high degree of competency was reported. This suggests that the capacity deficit is a problem of numbers of specialists rather than an overall lack of appropriate skills. Professional isolation must be addressed by encouraging greater partnership working across teams.

Small area variations in infant mortality in England and Wales in the inter-war period and their link with socio-economic factors.

This paper considers the changing spatial pattern of infant mortality during the inter-war years (1921-1936) at Local Government District level in England and Wales. As well as vital statistics data for these areas, also available are 1921 and 1931 Census data on unemployment, housing, occupational structure and population density, with additional data on unemployment from 1927 to 1936 for the 62 counties in which Districts are located. The statistical analysis relating changing District mortality to these characteristics derives mortality gradients across categorical forms of the social indicators, and assesses impacts of model re-specification to allow for spatio-temporal correlation. Time varying gradients across unemployment, housing and policy status are then investigated with a view to assessing whether there was a widening in mortality inequalities in the regionalized economic depression of the 1930s.

Area effects on health variation over the life-course: analysis of the longitudinal study sample in England using new data on area of residence in childhood.

There is a growing literature which demonstrates that (a) conditions throughout the life-course are important for health outcomes in older people and (b) "contextual" conditions in the place of residence, as well as individual characteristics influence health variations. This paper contributes to this debate by presenting results of an analysis of data from the Office for National Statistics Longitudinal Study (LS) for England and Wales. The analysis makes use of a new set of variables, which have been added to the LS, describing the social and economic conditions in the 1930s in residential areas where members of the LS sample were registered as living in 1939. The analysis focuses on people who were aged 0-16 in 1939. The health outcomes considered are death between 1981 and 1991, and for those still living, whether long-term illness was reported in the 1991 census. Regression analysis is used to examine the effects of residence in 1981, and data on the registered place of residence in 1939. The analysis shows that individual characteristics and type of area of residence in 1981 were associated with health outcomes. In addition, some variables describing socio-economic conditions in the 1930s contribute independently to the regression models (notably measures of economic deprivation and unemployment). The results suggest that conditions in residential area in early life may help to explain relatively poor health experience of populations in some parts of Britain today.

The development of gravity models for hospital patient flows under system change: a Bayesian modelling approach.

This paper discusses models for the impact on patient referral flows from homes to hospitals of reconfigurations of emergency hospital services. Such system change might involve new hospital sites, or expanded bed numbers at some sites, together with possible closure of emergency units at other hospitals. In terms of a gravity model for the flows of patients, this corresponds to removing a destination or adding a new one. While retaining a gravity type approach to this problem, the relevance of supply weighting by distance within the broader accessibility concept is stressed since the independence from irrelevant alternatives property is generally inapplicable. This accessibility based approach may be implemented as a general linear model, with a Poisson outcome, and a Bayesian estimation and predictive approach is adopted. In the context of patient hospitalisations, this facilitates prediction of new patient flows following hospital reconfiguration. A UK based case study of small residential areas and hospitals in North East London and Essex is presented within the context of a review of emergency hospital siting in London.

Health status and healthy life measures for population health need assessment: modelling variability and uncertainty.

It is increasingly recognised that population health need assessments based on the comparison of clinical or demographic end points (e.g. area mortality rates) neglect population variation in broader aspects of health status and health-related quality of life. Similarly, outcome measures which neglect impacts on health-related quality of life may be an inadequate basis for assessing the effectiveness of health interventions. This paper reviews issues in assessing needs and outcomes at population level based on health status valuations. It considers especially the modelling of sources of uncertainty in measures of health status by using Bayesian sampling estimation methods which produce a distribution of summary outcome measures. The modelling issues are illustrated in models for individual level health status from survey responses and their incorporation in area life tables to derive total and healthy life expectancies. In particular, a health status index derived from Short Form 36 profile responses in a health and lifestyle survey in a London health authority provides a case study of community health needs assessment.

Assessing health status and outcomes in a geriatric day hospital.

The study objective was to assess the feasibility and usefulness of recommended outcome measures in older people attending a geriatric day hospital for multidisciplinary assessment and rehabilitation. We used the 'Short Form 36' (SF36) questionnaire which had been proposed as a suitable outcome tool for the elderly, as well as standard assessment scales (eg Barthel index). These were administered by interviewers at the start of day hospital attendance and repeated by postal survey three and six months later. Change in overall health status was rated by the clinical team. The study took place in a geriatric day unit based in a support hospital, specialising in assessment and rehabilitation of older people. Participants were older people referred directly from the community, or following an inpatient day, whose assessment indicated a need for multidisciplinary rehabilitation. Stroke and musculo-skeletal disorders were the commonest underlying conditions. There was a high incidence of non-completion on SF36 questions relating to physical and mental function. Subsequent interviews showed that patients found some questions irrelevant. Floor effects were common. In contrast, the standard scales were invariably fully completed. Compared with local population survey data, respondents had low baseline scores on all SF36 dimensions. Differences over time were probably explained by varying methods of administration. In spite of a clinical perception of improved health status during day hospital attendance, both standard and SF36 scores showed overall deterioration. Two conclusions could be drawn from this study. 1. Measures of physical and mental disability and quality of life gave lower results than expected and continued declining over a six month period, even when the clinical team felt that the patient had improved. 2. Administration of SF36 by an interviewer is essential to obtain meaningful results in older people with poor physical health, which should be interpreted with caution. Goal-specific measures may be more useful in this group of patients.

Primary care needs assessment and resourcing: complementary practice and geographic perspectives.

This paper considers the use of small area census data in deriving socio-economic health need indices for primary care practices and assesses how far such indices are related to practice activity. A particular goal is the derivation of weightings for resourcing formulae which incorporate the effect of socio-economic deprivation on activity as well as the more usual allowance for patient age alone, so ensuring a closer match between needs and resources across both practices and localities. A case study is undertaken of a health authority in East London with a population of 380 thousand and 97 primary practices. Total rates of inpatient and outpatient referral and of prescribing by practices all show an influence of practice deprivation after allowing for age, and this effect is enhanced for subcategories such as emergency referrals to A&E departments. After taking account of practice age structure and deprivation, practice characteristics themselves (e.g. number of GPs in relation to practice population) are relatively unimportant influences on activity. The paper discusses how variation in need may be incorporated into resourcing via differential weighting.

A multi-level perspective on small-area health and mortality: a case study of England and Wales.

"This paper considers morbidity and mortality variations at the small-area level in England and Wales and their relation to socio-economic factors. Separate affluence and deprivation effects on illness and mortality are identified from a single-level analysis, both exceeding the influence of social class, urbanity or ethnicity. A multi-level perspective is then adopted to explore ecological associations operating both at micro-area and higher level spatial scales. Contextual effects (higher level variability in the impacts of ward level variables) are identified in the effects of small area deprivation on mortality and illness, as well as cross-level interactions; thus the impact of small area social structure on health is partly defined by the wider regional and district setting."

Modelling inequality in reported long term illness in the UK: combining individual and area characteristics.

STUDY OBJECTIVE: To assess the nature of the relation between health and social factors at both the aggregated scale of geographical areas and the individual scale. DESIGN AND SETTING: The individual data are derived from the sample of anonymised records (SAR) from the census of 1991 in Great Britain, and are combined with area data from this census. The ecological setting (context) was defined using multivariate methods to classify the 278 districts of residence identifiable in the SAR. The outcome health variable is the 1991 census long-term limiting illness question. Health variations were analysed by multilevel logistic regression to examine the compositional variation (at the level of the individual) and the contextual variation (variability operating at the level of districts) in reported illness. PARTICIPANTS: 10 per cent randomised subsample of the SAR who are aged 16+ and are resident in households. MAIN RESULTS: The multi-level modelling revealed that area factors have a significant association with individual health outcome but their effect is smaller than that of individual attributes. The results show evidence for both compositional and contextual effects in the pattern of variation in propensity to report illness. CONCLUSIONS: The results suggest generally higher levels of ill health for individuals who are older, not married, in a semi/unskilled manual social class, and socioeconomically deprived (as measured by a composite deprivation score). All individuals living in areas with high levels of illness (which tend to be more deprived areas) show greater morbidity, even after allowing for their individual characteristics. However, within affluent areas, where morbidity was generally lower, the health inequality (health gradient) between rich and poor individuals was particularly strong. We consider the implications of these findings for health and resource allocation policy.

Life table analysis for areas using vital register data.

"This study considers the utility of parameterised life tables derived by survival analysis for comparing mortality between areas, using death registration records and accompanying information on the social characteristics for each individual deceased. Such methods enable a comparison of summary measures of mortality experience such as life expectancy and median age at death before and after adjustment for socio-economic variables. In the absence of comparable information on the survivor population an approximate life table method is investigated as a means of comparing mortality profiles and the effects of social factors. Such factors may pertain both to the individuals (e.g. their birthplace) or to their small area of residence (e.g. measures of area deprivation). These methods also permit a comparison of the impact of socio-economic factors on different causes of death. The application is to mortality in London over the period 1990-92 and to its constituent boroughs and electoral wards."

Modelling frailty in area mortality.

This paper investigates the impact on area life tables of the specification of unobserved frailty. Frailty specification may affect both the regression effects of area and individual level covariates, and lead to changes in the value of summary mortality parameters, such as life expectancy. The paper also investigates how frailty affects life tables for specific causes of death, especially lung cancer and heart disease. Implications for choice between different model specifications, both in terms of age dependence and frailty, are discussed. The focus is on registered deaths recorded by age, birthplace and by small area within a borough of Greater London.

Approaches to modelling overdispersion in the analysis of migration.

"In this paper the modelling of overdispersion in generalised Poisson and multinomial models of migration flows and rates is considered, and its importance within the wider question of substantive model specification is shown. It is argued that substantive specification and the modelling of overdispersion are closely interrelated. Simplified ways of estimating the form of overdispersion--moments methods and pseudo-likelihood--are considered wherever possible. Overdispersion is set within the broader context of correlation effects which relate to migration--correlation across different destinations, between spatially adjacent origins and destinations, and between different time periods. The data used to illustrate the modelling development [concern England and Wales and] are drawn from the Office of Population Censuses and Surveys Longitudinal Study, the decennial Census, and the National Health Service Central Register."

Statistical graduation in local demographic analysis and projection.

"This paper considers parametric graduation for mortality, fertility and migration with particular reference to the development of parameterized local and regional demographic projections. Parametric graduations facilitate comparisons of demographic schedules across many areas and across time points--a feature which can be used to advantage in making forecasts of the three demographic components and thus in setting the assumptions for projections. Particular methodological issues raised are the questions of parsimony in fit and...of overdispersion in relation to binomial or Poisson assumptions. The analysis is illustrated with cross-sectional material for the 32 London boroughs and with time series at the level of Greater London."

Aspects of general linear modelling of migration.

"This paper investigates the application of general linear modelling principles to analysing migration flows between areas. Particular attention is paid to specifying the form of the regression and error components, and the nature of departures from Poisson randomness. Extensions to take account of spatial and temporal correlation are discussed as well as constrained estimation. The issue of specification bears on the testing of migration theories, and assessing the role migration plays in job and housing markets: the direction and significance of the effects of economic variates on migration depends on the specification of the statistical model. The application is in the context of migration in London and South East England in the 1970s and 1980s."