Effects of the choice of age-adjustment method on maps of death rates

Trends in cardiovascular risk factors control among US adults by glycemic statuses, 2007-2018

AIMS

Understanding the national trends in cardiovascular risk factors control of individuals with prediabetes and diabetes is critical for diabetes prevention and management. Our study aims to estimate how cardiovascular risk factors changed in US adults with different glycemic statuses between 2007-2008 and 2017-2018.

METHODS AND RESULTS

This was a serial cross-sectional study based on the National Health and Nutrition Examination Surveys (between 2007-2008 and 2017-2018 cycle). Non-pregnant American participants aged 20 years or older were included. Cardiovascular risk factors including weight, blood pressure, plasma cholesterol, and smoking by glycemic statuses were estimated. A total of 33 040 American adults were included. From 2007-2008 to 2017-2018, the age-adjusted proportions of individuals who reached weight control (body mass index <30 kg/m2) of both normoglycemia group and prediabetes group had a significant decrease over the study period, while the trend in participants with diabetes was not significant (mean difference: -5.34%, 95% confidence interval: -15.28%, 4.59%; P for trend = 0.298). The age-adjusted means of total cholesterol of all three groups decreased during the study decade (P for trend < 0.010), with participants with diabetes maintaining the lowest level. Individuals with high total cholesterol were more likely to receive statin therapy in the diabetes group. Notably, prediabetes participants had the highest level of total cholesterol and low-density lipoprotein cholesterol and were less likely to achieve lipid control with statin therapy. Sensitivity analysis with the second definition of prediabetes and diabetes resulted in a consistent trend.

CONCLUSIONS

In this nationally representative cross-sectional study, we systematically estimated the cardiovascular risk factors control in American adults and found poor weight control in the normoglycemia and prediabetes group. Despite the significant decrease trend of plasma total cholesterol in all groups, the high cholesterol level in the prediabetes group deserves special concern.

Indirect estimation of the need for palliative care during the COVID-19 pandemic: A descriptive cross-sectional study using mortality data in the Biobío Region, Chile

BACKGROUND

People with chronic diseases in their advanced phase require palliative care. This is essential to ensure their quality of life as it ends. However, a very low percentage of patients receive the necessary palliative care. The COVID-19 pandemic has adversely affected the planning and provision of palliative care. Despite this, in Chile, palliative care coverage was extended by law to cover nononcological chronic diseases. Implementation of this law is expected to be a significant challenge in terms of material resources, as well as the need for the formation of specialized palliative care teams. Therefore, it is essential to estimate the need for palliative care for all chronic diseases to generate useful input for planning and decision-making in public health.

OBJECTIVES

To indirectly estimate the need for palliative care among people with Chronic Oncological Diseases (COD) and Chronic Non-Oncological Diseases (CNOD) during the prepandemic and pandemic context due to COVID-19 in the Biobío Region in Chile.

METHODS

Cross-sectional study based on mortality data from chronic oncological and nononcological diseases during the prepandemic (2010-2018) and pandemic (2020-2021) contexts due to COVID-19 in a Region of Chile through indirect estimation using minimal estimate, standardized mortality rates and geographically weighted regression.

RESULTS

It was estimated that 76.25% of deaths from chronic diseases in the Biobío Region would have required palliative care, which represents 77,618 people who should have been included in these health benefits. The pandemic had a significant effect on the average number of deaths from CNOD. People belonging to this group were more likely to die from COVID-19 than from their baseline disease, unlike the deaths of people from COD, where no significant changes were observed.

CONCLUSION

These estimates highlight the potential size of the population requiring palliative care and emphasize the importance of recognizing the rights of individuals with COD and CNOD conditions. It is evident that there is a significant demand for palliative care services, as well as a pressing need for adequate resources, effective management, and strategic planning to cater to the needs of this population. This is particularly crucial in the heavily impacted areas and communes of the Biobío Region, Chile.

Assessing the Distribution of Air Pollution Health Risks within Cities: A Neighborhood-Scale Analysis Leveraging High-Resolution Data Sets in the Bay Area, California

BACKGROUND

Air pollution-attributable disease burdens reported at global, country, state, or county levels mask potential smaller-scale geographic heterogeneity driven by variation in pollution levels and disease rates. Capturing within-city variation in air pollution health impacts is now possible with high-resolution pollutant concentrations.

OBJECTIVES

We quantified neighborhood-level variation in air pollution health risks, comparing results from highly spatially resolved pollutant and disease rate data sets available for the Bay Area, California.

METHODS

We estimated mortality and morbidity attributable to nitrogen dioxide (NO2), black carbon (BC), and fine particulate matter [PM ≤2.5μm in aerodynamic diameter (PM2.5)] using epidemiologically derived health impact functions. We compared geographic distributions of pollution-attributable risk estimates using concentrations from a) mobile monitoring of NO2 and BC; and b) models predicting annual NO2, BC and PM2.5 concentrations from land-use variables and satellite observations. We also compared results using county vs. census block group (CBG) disease rates.

RESULTS

Estimated pollution-attributable deaths per 100,000 people at the 100-m grid-cell level ranged across the Bay Area by a factor of 38, 4, and 5 for NO2 [mean=30 (95% CI: 9, 50)], BC [mean=2 (95% CI: 1, 2)], and PM2.5, [mean=49 (95% CI: 33, 64)]. Applying concentrations from mobile monitoring and land-use regression (LUR) models in Oakland neighborhoods yielded similar spatial patterns of estimated grid-cell-level NO2-attributable mortality rates. Mobile monitoring concentrations captured more heterogeneity [mobile monitoring mean=64 (95% CI: 19, 107) deaths per 100,000 people; LUR mean=101 (95% CI: 30, 167)]. Using CBG-level disease rates instead of county-level disease rates resulted in 15% larger attributable mortality rates for both NO2 and PM2.5, with more spatial heterogeneity at the grid-cell-level [NO2 CBG mean=41 deaths per 100,000 people (95% CI: 12, 68); NO2 county mean=38 (95% CI: 11, 64); PM2.5 CBG mean=59 (95% CI: 40, 77); and PM2.5 county mean=55 (95% CI: 37, 71)].

DISCUSSION

Air pollutant-attributable health burdens varied substantially between neighborhoods, driven by spatial variation in pollutant concentrations and disease rates. https://doi.org/10.1289/EHP7679.

Revealing the Unequal Burden of COVID-19 by Income, Race/Ethnicity, and Household Crowding: US County Versus Zip Code Analyses

OBJECTIVE

To overcome the absence of national, state, and local public health data on the unequal economic and social burden of COVID-19 in the United States.

DESIGN

We analyze US county COVID-19 deaths and confirmed COVID-19 cases and positive COVID-19 tests in Illinois and New York City zip codes by area percent poverty, percent crowding, percent population of color, and the Index of Concentration at the Extremes.

SETTING

US counties and zip codes in Illinois and New York City, as of May 5, 2020.

MAIN OUTCOME MEASURES

Rates, rate differences, and rate ratios of COVID-19 mortality, confirmed cases, and positive tests by category of county and zip code-level area-based socioeconomic measures.

RESULTS

As of May 5, 2020, the COVID-19 death rate per 100 000 person-years equaled the following: 143.2 (95% confidence interval [CI]: 140.9, 145.5) vs 83.3 (95% CI: 78.3, 88.4) in high versus low poverty counties (≥20% vs <5% of persons below poverty); 124.4 (95% CI: 122.7, 126.0) versus 48.2 (95% CI: 47.2, 49.2) in counties in the top versus bottom quintile for household crowding; and 127.7 (95% CI: 126.0, 129.4) versus 25.9 (95% CI: 25.1, 26.6) for counties in the top versus bottom quintile for the percentage of persons who are people of color. Socioeconomic gradients in Illinois confirmed cases and New York City positive tests by zip code-level area-based socioeconomic measures were also observed.

CONCLUSIONS

Stark social inequities exist in the United States for COVID-19 outcomes. We recommend that public health departments use these straightforward cost-effective methods to report on social inequities in COVID-19 outcomes to provide an evidence base for policy and resource allocation.

Risk of Cardiovascular Disease and Mortality in Young Adults With End-stage Renal Disease: An Analysis of the US Renal Data System

Importance

Cardiovascular disease (CVD) is a leading cause of death among patients with end-stage renal disease (ESRD). Young adult (ages 22-29 years) have risks for ESRD-associated CVD that may vary from other ages.

Objective

To test the hypothesis that young adult-onset ESRD is associated with higher cardiovascular (CV) hospitalizations and mortality with different characteristics than childhood-onset disease.

Design, Setting, and Participants

This population-based cohort study used the US Renal Data System to categorize patients who initiated ESRD care between 2003 and 2013 by age at ESRD onset (1-11, 12-21, and 22-29 years). Cardiovascular hospitalizations were identified via International Classification of Diseases, Ninth Revision discharge codes and CV mortality from the Centers for Medicare & Medicaid ESRD Death Notification Form. Patients were censored at death from non-CVD events, loss to follow-up, recovery, or survival to December 31, 2014. Adjusted proportional hazard models (95% CI) were fit to determine risk of CV hospitalization and mortality by age group. Data analysis occurred from May 2016 and December 2017.

Exposures

Onset of ESRD.

Main Outcomes and Measures

Cardiovascular mortality and hospitalization.

Results

A total of 33 156 patients aged 1 to 29 years were included in the study population. Young adults (aged 22-29 years) had a 1-year CV hospitalization rate of 138 (95% CI, 121-159) per 1000 patient-years. Young adults had a higher risk for CV hospitalization than children (aged 1-11 years; hazard ratio [HR], 0.41 [95% CI, 0.26-0.64]) and adolescents (aged 12-21 years; HR, 0.86 [95% CI, 0.77-0.97]). Of 4038 deaths in young adults, 1577 (39.1%) were owing to CVD. Five-year cumulative incidence of mortality in this group (7.3%) was higher than in younger patients (adolescents, 4.0%; children, 1.7%). Adjusted HRs for CV mortality were higher for young adults with all causes of ESRD than children (cystic, hereditary, and congenital conditions: HR, 0.22 [95% CI, 0.11-0.46]; P < .001; glomerulonephritis: HR, 0.21 [95% CI, 0.10-0.44]; P < .001; other conditions: HR, 0.33 [95% CI, 0.23-0.49]; P < .001). Adolescents had a lower risk for CV mortality than young adults for all causes of ESRD except glomerulonephritis (cystic, hereditary, and congenital conditions: HR, 0.45 [95% CI, 0.27-0.74]; glomerulonephritis: HR, 0.99 [95% CI, 0.76-1.11]; other: HR, 0.47 [95% CI, 0.40-0.57]). Higher risks for CV hospitalization and mortality were associated with lack of preemptive transplant compared with hemodialysis (hospital: HR, 14.24 [95% CI, 5.92-34.28]; mortality: HR, 13.64 [95% CI, 8.79-21.14]) and peritoneal dialysis [hospital: HR, 8.47 [95% CI, 3.50-20.53]; mortality: HR, 7.86 [95% CI, 4.96-12.45]). Nephrology care before ESRD was associated with lower risk for CV mortality (HR, 0.77 [95% CI, 0.70-0.85]).

Conclusions and Relevance

Cardiovascular disease accounted for nearly 40% of deaths in young adults with incident ESRD in this cohort. Identified risk factors may inform development of age-appropriate ESRD strategies to improve the CV health of this population.

Changing livestock vaccination policy alters the epidemiology of human anthrax, Georgia, 2000-2013

Anthrax is a widely spread zoonotic disease found on nearly every continent. To control the disease in humans and animals, annual livestock vaccination is recommended. However, in 2007, the country of Georgia ended its policy of compulsory annual livestock anthrax vaccination. Our objective was to assess how the epidemiology of human anthrax has evolved from 2000-2013 in Georgia, in the wake of this cessation. We used passive surveillance data on epidemiological surveys of human anthrax case patients. Risk factors and rates of self-reported sources of infection were compared, before and after the change in livestock vaccination policy. We mapped ethnicity-adjusted incidence during the two periods and assessed changes in the spatial pattern of risk. The overall risk of human anthrax increased >5-fold, from 0.7 cases per 100,000 in 2000 to 3.7 cases per 100,000 by 2013. Ethnic disparities in risk became pronounced; from 2000 to 2013, incidence increased >60-fold in Azerbaijanis from 0.35 to 21.1 cases/100,000 Azerbaijanis compared to 0.61 to 1.9 cases/100,000 among ethnic Georgians. Food-borne exposures from purchasing meat increased from 11% in 2000-2006 to 21% in 2007-2013. Spatial analyses revealed a shift from a random pattern of reporting pre-policy change to clustering among district municipalities following the change in policy. Our findings indicate there were unintended human health consequences associated with changing livestock vaccination policy. Following a reduction in the immunizations administered, there was a major shift in the epidemiology of human anthrax in Georgia. Current infection risk is now highest among ethnic minorities. Increased reporting among individuals uncharacteristically at risk for anthrax from foodborne exposures suggests spillover from modes of agricultural production. Given the importance of human-livestock health linkages, careful evaluations of policy need to be undertaken before changes to animal vaccination are made.

Spatial Pattern Detection of Tuberculosis: A Case Study of Si Sa Ket Province, Thailand

This retrospective population-based study was conducted to analyze spatial patterns of tuberculosis (TB) incidence in Si Sa Ket province, Thailand. TB notification data from 2004 to 2008 collected from TB clinics throughout the province was used along with population data to reveal a descriptive epidemiology of TB incidences. Global clustering patterns of the occurrence were assessed by using global spatial autocorrelation techniques. Additionally, local spatial pattern detection was performed by using local spatial autocorrelation and spatial scan statistic methods. The findings indicated clusters of the disease occurred in the study area. More specifically, significantly high-rate clusters were mostly detected in Mueang Si Sa Ket and Khukhan districts, which are located in the northwestern part of the province, while significantly low-rate clusters were persistent in Kantharalak and Benchalak districts, which are located at the southeastern area.

Parkinson's Disease and Pesticides Exposure: New Findings From a Comprehensive Study in Nebraska, USA

BACKGROUND

The association between exposure to agricultural pesticides and Parkinson's Disease (PD) has long been a topic of study in the field of environmental health. This research takes advantage of the unique Nebraska PD registry and state-level crop classification data to investigate the PD-pesticides exposure relationship.

METHODS

First, Geographic Information System and satellite remote sensing data were adopted to calculate exposure to different pesticides for Nebraska residents. An integrated spatial exploratory framework was then adopted to explore the association between PD incidence and exposure to specific pesticide ingredients at the county level.

RESULTS

Our results reveal similarities in geographic patterns of pesticide exposure and PD incidence. The regression analyses indicate that, for most Nebraska counties, PD incidence was significantly associated with exposure to certain pesticide ingredients such as alachlor and broxomy. However, the results also suggest that factors other than pesticide exposure may help further explain the risk of PD at the county level.

CONCLUSIONS

We found significant associations between PD incidence and exposure to different pesticide ingredients. These results have useful implications for PD prevention in Nebraska and other agricultural states in the United States.

Visualizing Statistical Mix Effects and Simpson's Paradox

We discuss how "mix effects" can surprise users of visualizations and potentially lead them to incorrect conclusions. This statistical issue (also known as "omitted variable bias" or, in extreme cases, as "Simpson's paradox") is widespread and can affect any visualization in which the quantity of interest is an aggregated value such as a weighted sum or average. Our first contribution is to document how mix effects can be a serious issue for visualizations, and we analyze how mix effects can cause problems in a variety of popular visualization techniques, from bar charts to treemaps. Our second contribution is a new technique, the "comet chart," that is meant to ameliorate some of these issues.

The impact of data suppression on local mortality rates: the case of CDC WONDER

CDC WONDER (Centers for Disease Control and Prevention Wide-Ranging Online Data for Epidemiologic Research) is the nation's primary data repository for health statistics. Before WONDER data are released to the public, data cells with fewer than 10 case counts are suppressed. We showed that maps produced from suppressed data have predictable geographic biases that can be removed by applying population data in the system and an algorithm that uses regional rates to estimate missing data. By using CDC WONDER heart disease mortality data, we demonstrated that effects of suppression could be largely overcome.

A nonparametric approach for determining significance of county cancer rates compared to the overall state rate: illustrated with Minnesota data

BACKGROUND

The study of the geographical distribution of disease has expanded greatly with GIS technology and its application to increasingly available public health data. The emergence of this technology has increased the challenges for public health practitioners to provide meaningful interpretations for county-based state cancer maps.

METHODS

One of these challenges-spurious inferences about the significance of differences between county and overall state cancer rates-can be addressed through a nonparametric statistical method. The Wilcoxon's signed rank test (WSRT) has a practical application for determining the significance of county cancer rates compared to the statewide rate. This extension of the WSRT, developed by John Tukey, forms the basis for constructing a single confidence interval for all differences in county and state directly age-adjusted cancer rates. Empirical evaluation of this WSRT application was conducted using Minnesota cancer incidence data.

RESULTS

The WSRT procedure reduced the impact of statistical artifacts that are frequently encountered with standard normal significance testing of the difference between directly age-adjusted county and the overall state cancer rates.

CONCLUSION

Although further assessment of its performance is required, the WSRT procedure appears to be a useful complement for mapping directly age-adjusted state cancer rates by county.

Mapping suicide in London: a brief methodological case study on the application of the smoothing technique

BACKGROUND

When one intends to globally smooth unstable rates, e.g., suicide rates in a region, one needs to consider whether it is better to smooth the rates toward the global mean of the country or toward the global mean of the same region.

AIMS

The present study aims to provide a methodological framework to answer this question by smoothing suicide rates within London boroughs.

METHODS

Based on the results of the spatial autocorrelation statistics, the noniterative empirical Bayes method of moments was chosen to globally smooth the suicide rate of each borough, first toward the global mean of England and Wales, and second toward the mean of the London region.

RESULTS

The results revealed that smoothing the suicide rates of the boroughs toward the global mean of England and Wales had a stronger influence in reducing the variability of suicide rates than smoothing toward the global mean of the London region.

CONCLUSIONS

Smoothing the rates toward the mean of a region within a country acts somewhat between global and local smoothing.

A New Class of Minimum Power Divergence Estimators with Applications to Cancer Surveillance

The Annual Percent Change (APC) has been adopted as a useful measure for analyzing the changing trends of cancer mortality and incidence rates by the NCI SEER program. Difficulties, however, arise when comparing the sample APCs between two overlapping regions because of the induced dependence (e.g., comparing the cancer mortality change rate of California with the national level). This paper deals with a new perspective of understanding the sample distribution of the test statistics for comparing the APCs between overlapping regions. Our proposal allows for computational readiness and easy interpretability. We further propose a more general family of estimators, namely, the so-called minimum power divergence estimators, including the maximum likelihood estimators as a special case. Our simulation experiments support the superiority of the proposed estimator to the conventional maximum likelihood estimator. The proposed method is illustrated by the analysis of the SEER cancer mortality rates observed from 1991 to 2006.

[Assessment of applicability of standardized rates for health state comparison among areas: 2008 Community Health Survey]

OBJECTIVES

This study shows the issues that should be considered when applying standardized rates using Community Health Survey(CHS) data.

METHODS

We analyzed 2008 CHS data. In order to obtain the reliability of standardized rates, we calculated z-score and rank correlation coefficients between direct standardized rate and indirect standardized rate for 31 major indices. Especially, we assessed the change of correlations according to population composition (age and sex), and characteristics of the index. We used Mantel-Haenszel chi-square to quantify the difference of population composition.

RESULTS

Among 31 major indices, 29 indices' z-score and rank correlation coefficients were over 0.9. However, regions with larger differences in population composition showed lower reliability. Low reliability was also observed for the indices specific to subgroups with small denominator such as 'permanent lesion from stroke', and the index with large regional variations in age-related differences such as 'obtaining health examinations'.

CONCLUSIONS

Standardized rates may have low reliability, if comparison is made between areas with extremely large differences in population composition, or for indicies with large regional variations in age-related differences. Therefore, the special features of standardized rates should be considered when health state are compared among areas.

Spatial modeling in environmental and public health research

THIS PAPER HAS TWO AIMS: (1) to summarize various geographic information science methods; and (2) to provide a review of studies that have employed such methods. Though not meant to be a comprehensive review, this paper explains when certain methods are useful in epidemiological studies and also serves as an overview of the growing field of spatial epidemiology.

Spatio-temporal trends of mortality in small areas of Southern Spain

Background

Most mortality atlases show static maps from count data aggregated over time. This procedure has several methodological problems and serious limitations for decision making in Public Health. The evaluation of health outcomes, including mortality, should be approached from a dynamic time perspective that is specific for each gender and age group. At the moment, researches in Spain do not provide a dynamic image of the population's mortality status from a spatio-temporal point of view. The aim of this paper is to describe the spatial distribution of mortality from all causes in small areas of Andalusia (Southern Spain) and evolution over time from 1981 to 2006.

Methods

A small-area ecological study was devised using the municipality as the unit for analysis. Two spatio-temporal hierarchical Bayesian models were estimated for each age group and gender. One of these was used to estimate the specific mortality rate, together with its time trends, and the other to estimate the specific rate ratio for each municipality compared with Spain as a whole.

Results

More than 97% of the municipalities showed a diminishing or flat mortality trend in all gender and age groups. In 2006, over 95% of municipalities showed male and female mortality specific rates similar or significantly lower than Spanish rates for all age groups below 65. Systematically, municipalities in Western Andalusia showed significant male and female mortality excess from 1981 to 2006 only in age groups over 65.

Conclusions

The study shows a dynamic geographical distribution of mortality, with a different pattern for each year, gender and age group. This information will contribute towards a reflection on the past, present and future of mortality in Andalusia.

Racial disparities in lung cancer mortality in U.S. congressional districts, 1990-2001

The objective of this study was to detect statistically significant racial disparities in lung cancer mortality at the U.S. congressional district level. We applied absolute disparity statistics to mortality data from the National Center for Health Statistics (NCHS) for 1990-2001, mapped significant lung cancer mortality disparities by race and gender within U.S. congressional districts, and uncovered previously unreported disparities. The disparity statistics comparing black and white females revealed higher mortality rates for black females in the Midwestern U.S., and higher mortality rates for white females in the South-eastern U.S. Our methodology provides a spatial tool for guiding public health cancer control practices to monitor, target and reduce disparities.

Small-area racial disparity in stroke mortality: an application of bayesian spatial hierarchical modeling

BACKGROUND

In the United States, excess burden of stroke mortality has persisted among African Americans compared with whites despite declines in stroke mortality for both groups. New insights may be gleaned by examining local, small-area patterns in racial disparities in stroke.

METHODS

The study population includes all non-Hispanic African Americans and non-Hispanic whites aged 35 to 64 in the southeastern United States during 1999 to 2002. We assessed county-level numbers of stroke deaths and population estimates in a Bayesian spatial hierarchical modeling framework allowing for inclusion of potential covariates (poverty and rurality), and generating county-specific model-based estimates of both absolute and relative racial disparity. The resulting estimates of race-specific stroke death rates, relative racial disparity, and absolute racial disparity were expressed in maps.

RESULTS

After adjustment for age, poverty, and rurality, county-level estimates of relative racial disparity ranged from 2.3 to 3.3 and estimates of absolute racial disparity ranged from 19 to 45 excess deaths per 100,000. For both racial groups, stroke death rates were higher in rural areas and with increasing poverty. High relative racial disparity was concentrated primarily in the eastern portion of the region and large absolute racial disparity was concentrated primarily in the western portion.

CONCLUSIONS

The results highlight the pervasiveness and magnitude of substantial local racial disparities in stroke mortality in the southeast.

Methodologic implications of social inequalities for analyzing health disparities in large spatiotemporal data sets: an example using breast cancer incidence data (Northern and Southern California, 1988--2002)

Efforts to monitor, investigate, and ultimately eliminate health disparities across racial/ethnic and socioeconomic groups can benefit greatly from spatiotemporal models that enable exploration of spatial and temporal variation in health. Hierarchical Bayes methods are well-established tools in the statistical literature for fitting such models, as they permit smoothing of unstable small-area rates. However, issues presented by 'real-life' surveillance data can be a barrier to routine use of these models by epidemiologists. These include (1) shifting of regional boundaries over time, (2) social inequalities in racial/ethnic residential segregation, which imply differential spatial structuring across different racial/ethnic groups, and (3) heavy computational burdens for large spatiotemporal data sets. Using data from a study of changing socioeconomic gradients in female breast cancer incidence in two population-based cancer registries covering the San Francisco Bay Area and Los Angeles County, CA (1988--2002), we illustrate a two-stage approach to modeling health disparities and census tract (CT) variation in incidence over time. In the first stage, we fit race- and year-specific spatial models using CT boundaries normalized to the U.S. Census 2000. In stage 2, temporal patterns in the race- and year-specific estimates of racial/ethnic and socioeconomic effects are explored using a variety of methods. Our approach provides a straightforward means of fitting spatiotemporal models in large data sets, while highlighting differences in spatial patterning across racial/ethnic population and across time.

Comparing trends in cancer rates across overlapping regions

Monitoring and comparing trends in cancer rates across geographic regions or over different time periods have been major tasks of the National Cancer Institute's (NCI) Surveillance, Epidemiology, and End Results (SEER) Program as it profiles healthcare quality as well as decides healthcare resource allocations within a spatial-temporal framework. A fundamental difficulty, however, arises when such comparisons have to be made for regions or time intervals that overlap, for example, comparing the change in trends of mortality rates in a local area (e.g., the mortality rate of breast cancer in California) with a more global level (i.e., the national mortality rate of breast cancer). In view of sparsity of available methodologies, this article develops a simple corrected Z-test that accounts for such overlapping. The performance of the proposed test over the two-sample "pooled"t-test that assumes independence across comparison groups is assessed via the Pitman asymptotic relative efficiency as well as Monte Carlo simulations and applications to the SEER cancer data. The proposed test will be important for the SEER * STAT software, maintained by the NCI, for the analysis of the SEER data.

Temporal changes in the spatial pattern of disease rates incorporating known risk factors

Examining the geographical pattern of temporal changes in infant mortality rates illustrates the methodological problems of documenting and understanding temporal changes in any spatial pattern of disease. Early research on geographical differences in infant mortality rates showed strong ecological correlations with socio-economic factors such as poverty rates. More recent research established relationships between individual-level socio-economic values and probabilities of death. With geographic information available at the level of individuals, it is possible to estimate the probabilities of death on a person-by-person basis from knowledge of the relationships between individual factors and socio-economic measures. These estimated probabilities provide an expected geographic pattern of deaths. The difference between the observed spatial pattern and the expected pattern is the remaining spatial variation adjusted for this knowledge. For the study area, individual factors and some socio-economic measures were available for each year of the study period. Using data from the Iowa Birth Defects Registry and the Iowa Department of Public Health (USA), I tested the stability and continuity of these cross-sectional relationships and investigated whether any temporal lags in these variables relate to the unexplained spatial variations in infant mortality rates that remain. I accounted for the 'Change of Support Problem' [Gotway C. A. & Young L. J. (2002). Combining incompatible spatial data. Journal of the American Statistical Association, 97458, 632-648] inherent in frame-based geographical analysis. The analysis involved a generalized linear model (GLM) to estimate individual risks and a Monte Carlo simulation model to generate the non-linear probability density functions for disease rates whose densities are theoretically intractable. Results show the temporal changes in the observed spatial pattern and the expected spatial pattern differ by geographic location. In conclusion such differences are the result of a combination of unexplained place-based risk and unmeasured individual risks.

Do hot spots of deprivation predict the rates of suicide within London boroughs?

INTRODUCTION

The ecological associations between suicide rates and different indices of deprivation within London have been investigated at least for half a century. In the present study, the association between rates of suicide with newly developed hot spots of deprivation index within London boroughs have been studied taking into account the results of the spatial dependency between suicide rates in nearby boroughs.

METHODS

Suicide data were provided by the National Confidential Inquiry into Suicide and Homicide by People with Mental Illness. The hot spots index of deprivation and the population counts were provided by the Department of the Environment, Transport and the Region (DETR) and Office for the National Statistics (ONS), respectively.

RESULTS

The results show that there is no strong spatial dependency between suicide rates in the London boroughs, the 'hot spots' index of deprivation predicts the rates of suicide in males 30-49, better than other age and sex groups. The rate of suicide decreases with decreasing deprivation as indicated by the 'hot spots' index.

CONCLUSION

These findings suggest that at the London boroughs the 'hot spots' index of deprivation (together with other socio-economic and social fragmentation indices) should be considered as a potential explanatory variable to explain the effects of age on rates of suicide in men and women.

Geographical epidemiology, spatial analysis and geographical information systems: a multidisciplinary glossary

We provide a relatively non-technical glossary of terms and a description of the tools used in spatial or geographical epidemiology and associated geographical information systems. Statistical topics included cover adjustment and standardisation to allow for demographic and other background differences, data structures, data smoothing, spatial autocorrelation and spatial regression. We also discuss the rationale for geographical epidemiology and specific techniques such as disease clustering, disease mapping, ecological analyses, geographical information systems and global positioning systems.

Bayesian statistics in medicine: a 25 year review

This review examines the state of Bayesian thinking as Statistics in Medicine was launched in 1982, reflecting particularly on its applicability and uses in medical research. It then looks at each subsequent five-year epoch, with a focus on papers appearing in Statistics in Medicine, putting these in the context of major developments in Bayesian thinking and computation with reference to important books, landmark meetings and seminal papers. It charts the growth of Bayesian statistics as it is applied to medicine and makes predictions for the future. From sparse beginnings, where Bayesian statistics was barely mentioned, Bayesian statistics has now permeated all the major areas of medical statistics, including clinical trials, epidemiology, meta-analyses and evidence synthesis, spatial modelling, longitudinal modelling, survival modelling, molecular genetics and decision-making in respect of new technologies.

Mapping and measuring social disparities in premature mortality: the impact of census tract poverty within and across Boston neighborhoods, 1999-2001

The identification and documentation of health disparities are important functions of public health surveillance. These disparities, typically falling along lines defined by gender, race/ethnicity, and social class, are often made visible in urban settings as geographic disparities in health between neighborhoods. Recognizing that premature mortality is a powerful indicator of disparities in both health status and access to health care that can readily be monitored using routinely available public health surveillance data, we undertook a systematic analysis of spatial variation in premature mortality in Boston (1999-2001) across neighborhoods and sub-neighborhoods in relation to census tract (CT) poverty. Using a multilevel model based framework, we estimated that the incidence of premature mortality was 1.39 times higher (95% credible interval 1.09-1.78) among persons living in the most economically deprived CTs (>/=20% below poverty) compared to those in the least impoverished tracts (<5% below poverty). We present maps of model-based standardized mortality ratios that show substantial within-neighborhood variation in premature mortality and a sizeable decrease in spatial variation after adjustment for CT poverty. Additionally, we present maps of model-based direct standardized rates that can more readily be compared to externally published rates and targets, as well as maps of the population attributable fraction that show that in some of Boston's poorest neighborhoods, the proportion of excess deaths associated with CT poverty reaches 25-30%. We recommend that these methods be incorporated into routine analyses of public health surveillance data to highlight continuing social disparities in premature mortality.

U.S. congressional district cancer death rates

BACKGROUND

Geographic patterns of cancer death rates in the U.S. have customarily been presented by county or aggregated into state economic or health service areas. Herein, we present the geographic patterns of cancer death rates in the U.S. by congressional district. Many congressional districts do not follow state or county boundaries. However, counties are the smallest geographical units for which death rates are available. Thus, a method based on the hierarchical relationship of census geographic units was developed to estimate age-adjusted death rates for congressional districts using data obtained at county level. These rates may be useful in communicating to legislators and policy makers about the cancer burden and potential impact of cancer control in their jurisdictions.

RESULTS

Mortality data were obtained from the National Center for Health Statistics (NCHS) for 1990-2001 for 50 states, the District of Columbia, and all counties. We computed annual average age-adjusted death rates for all cancer sites combined, the four major cancers (lung and bronchus, prostate, female breast, and colorectal cancer) and cervical cancer. Cancer death rates varied widely across congressional districts for all cancer sites combined, for the four major cancers, and for cervical cancer. When examined at the national level, broad patterns of mortality by sex, race and region were generally similar with those previously observed based on county and state economic area.

CONCLUSION

We developed a method to generate cancer death rates by congressional district using county-level mortality data. Characterizing the cancer burden by congressional district may be useful in promoting cancer control and prevention programs, and persuading legislators to enact new cancer control programs and/or strengthening existing ones. The method can be applied to state legislative districts and other analyses that involve data aggregation from different geographic units.

Hierarchical modeling and other spatial analyses in prostate cancer incidence data

BACKGROUND

State central cancer registries are often asked to respond to questions about the spatial distribution of cancer cases. Spatial analysis methods and technology are evolving rapidly, and can be a considerable challenge to registries that do not have staff with training in this area. The purpose of this article is to describe a general methodological approach that potentially might be a starting point for many cancer registry spatial analyses at the county level.

METHODS

Prostate cancer incident cases (N=31,159) from the Louisiana Tumor Registry from 1988 to 1999 were used for illustrative purposes. To explore spatio-temporal patterns, analyses focused on four time periods, each 3 years in length: 1998-1990, 1991-1993, 1994-1996, and 1997-1999. For each time period, race-specific (white and black), direct age-adjusted incidence rates and indirect standardized incidence ratios (SIRs) were calculated, smoothed using Bayesian methods, and assessed for evidence of spatial autocorrelation using global and local Moran's I. Hierarchical generalized linear models (HGLM) were fitted to identify significant covariates. Clusters of elevated and lower rates were identified using a spatial scan statistic (SaTScan).

RESULTS

Temporal trends in SIRs in both race groups were consistent with the introduction of prostate specific antigen (PSA) testing in Louisiana during the late 1980s and early 1990s, but possibly with a lag in black males. Clusters of lower than expected values were observed for white males in the central (p=0.001) and southeastern coastal areas (p=0.001), and to a greater extent for black males in the central (p=0.001), southwestern and southeastern coastal parishes (p=0.001).

CONCLUSIONS

Mapping disease occurrence by time period is an effective way to explore spatio-temporal patterns. HGLM models and software are available to control for covariates and for unstructured and spatially structured variability that may confound spatial variability patterns.

A multiscale method for disease mapping in spatial epidemiology

The effects of spatial scale in disease mapping are well-recognized, in that the information conveyed by such maps varies with scale. Here we provide an inferential framework, in the context of tract count data, for describing the distribution of relative risk simultaneously across a hierarchy of multiple scales. In particular, we offer a multiscale extension of the canonical standardized mortality ratio (SMR), consisting of Bayesian posterior-based strategies for both estimation and characterization of uncertainty. As a result, a hierarchy of informative disease and confidence maps can be produced, without the need to first try to identify a single appropriate scale of analysis. We explore the behaviour of the proposed methodology in a small simulation study, and we illustrate its usage through an application to data on gastric cancer in Tuscany.

Spatial analysis of disease--applications

The application of spatial statistical analysis to health data has reached adolescence. The theory and the software are both still maturing. We are drawing upon the experiences of the geostatisticians in modeling surfaces and the econometricians in modeling time series. "New and improved" computer algorithms are constantly being provided to implement the evolving theory or to improve the processing in terms of stability, reliability, and efficiency. We will come of age when we have the theory, the software, and the process to reliably produce "generalized spatio-temporal" models suitable for health data. In the meantime, biostatisticians need to acknowledge when their data is not independently distributed and to consider the spatial correlation in their analysis. This chapter provided examples using four available methods. The methods were spatial filtering, identifying clusters using the spatial scan statistic, hierarchical modeling, and conditional autoregression modeling.

Spatial analysis of disease

In this chapter, we have reviewed the history of the spatial analysis of disease and the statistical methods used for the exploratory analysis, testing and modeling of spatial patterns. In the next chapter, the principles described here will be illustrated.

Public health, GIS, and spatial analytic tools

We review literature that uses spatial analytic tools in contexts where Geographic Information Systems (GIS) is the organizing system for health data or where the methods discussed will likely be incorporated in GIS-based analyses in the future. We conclude the review with the point of view that this literature is moving toward the development and use of systems of analysis that integrate the information geo-coding and data base functions of GISystems with the geo-information processing functions of GIScience. The rapidity of this projected development will depend on the perceived needs of the public health community for spatial analysis methods to provide decision support. Recent advances in the analysis of disease maps have been influenced by and benefited from the adoption of new practices for georeferencing health data and new ways of linking such data geographically to potential sources of environmental exposures, the locations of health resources and the geodemographic characteristics of populations. This review focuses on these advances.

Multiple sclerosis distribution in northern Sardinia: spatial cluster analysis of prevalence

BACKGROUND

A heterogeneous geographic distribution of MS has been reported among different ethnic groups, and also within small communities. Epidemiologic studies conducted over the past two decades using repeated assessments clearly show that Sardinia is at high risk for MS, with a prevalence of 150 per 100,000 in 1997.

OBJECTIVE

To present spatial analysis of the disease prevalence to disclose possible "hot" or "cold" spots of disease, further allowing correlations with risk factors.

METHODS

A spatial analysis of the whole province of Sassari, in northern Sardinia, at a microgeographic level (i.e., in the 89 administrative communes and 6 linguistic areas) was conducted. Because of the small number of cases per commune and to overcome random variability, a hierarchical Bayesian approach was adopted. The distribution of prevalent cases by commune of residence on December 31, 1997 and from age 5 to 15 years was analyzed.

RESULTS

A clustering pattern was found in the southwestern communes of the province based on geographic distribution by both prevalence and residence at age 5 to 15 years. A west-to-east gradient also was observed.

CONCLUSIONS

This study highlights a hot spot of MS in the southwestern part of Sassari province, bordering with the commune of Macomer, where MS was once hypothesized as having occurred as an epidemic. Interestingly, these areas of MS clustering comprise the Common Logudorese linguistic domain. The Catalan area, linguistically and genetically distant from the remaining Sardinian domains, does not show such high estimates. Because MS is not a single-source infectious disease, this study may help test the hypothesis that a widely and evenly spread environmental (infectious?) agent may produce disease in subgroups of genetically more susceptible individuals in areas at higher inbreeding rates, wherein a disease mode of inheritance could be better investigated.

Methodological issues in the development of the Canadian Cancer Incidence Atlas

The Canadian Cancer Incidence Atlas is among recent national atlases using incidence rather than mortality data. Methods used to assess the significance and spatial correlation of the age-standardized rates (ASIRs) for the 290 census divisions are described. The expected number of cases by area was used to determine cancer sites with sufficient cases to be mapped. ASIR significance was assessed using a simulation based on a Poisson distribution. The consistency of the observed case distributions with the Poisson distribution was examined. The bootstrap confidence interval (CI) for the ASIR developed by Swift was used in the atlas. Spatial correlation was assessed with Moran's I/I(max) and the significance determined by a simulation in order to allow for area population variation. Data quality indicators typically used for cancer registries were presented, supplemented by a registry questionnaire.

All maps of parameter estimates are misleading

Maps are frequently used to display spatial distributions of parameters of interest, such as cancer rates or average pollutant concentrations by county. It is well known that plotting observed rates can have serious drawbacks when sample sizes vary by area, since very high (and low) observed rates are found disproportionately in poorly-sampled areas. Unfortunately, adjusting the observed rates to account for the effects of small-sample noise can introduce an opposite effect, in which the highest adjusted rates tend to be found disproportionately in well-sampled areas. In either case, the maps can be difficult to interpret because the display of spatial variation in the underlying parameters of interest is confounded with spatial variation in sample sizes. As a result, spatial patterns occur in adjusted rates even if there is no spatial structure in the underlying parameters of interest, and adjusted rates tend to look too uniform in areas with little data. We introduce two models (normal and Poisson) in which parameters of interest have no spatial patterns, and demonstrate the existence of spatial artefacts in inference from these models. We also discuss spatial models and the extent to which they are subject to the same artefacts. We present examples from Bayesian modelling, but, as we explain, the artefacts occur generally.