Bayesian modeling of spatial ordinal data from health surveys.

Health surveys allow exploring health indicators that are of great value from a public health point of view and that cannot normally be studied from regular health registries. These indicators are usually coded as ordinal variables and may depend on covariates associated with individuals. In this article, we propose a Bayesian individual-level model for small-area estimation of survey-based health indicators. A categorical likelihood is used at the first level of the model hierarchy to describe the ordinal data, and spatial dependence among small areas is taken into account by using a conditional autoregressive distribution. Post-stratification of the results of the proposed individual-level model allows extrapolating the results to any administrative areal division, even for small areas. We apply this methodology to describe the geographical distribution of a self-perceived health indicator from the Health Survey of the Region of Valencia (Spain) for the year 2016.

Socioeconomic inequalities in COVID-19 incidence during the first six waves in Barcelona.

BACKGROUND: The emergence of SARS-CoV-2 affected urban areas. In Barcelona, six waves of COVID-19 hit the city between March 2020 and March 2022. Inequalities in the incidence of COVID-19 have been described. However, no studies have examined the daily trends of socioeconomic inequalities and how they changed during the different phases of the pandemic. The aim of this study is to analyse the dynamic socioeconomic inequalities in the incidence of COVID-19 during the six waves in Barcelona. METHODS: We examined the proportion of daily cases observed in the census tracts in the lower income tercile compared with the proportion of daily cases observed in the sum of the lower and higher income terciles. Daily differences in these proportions were assessed as a function of the epidemic waves, sex, age group, daily incidence and daily change in the incidence. A logistic regression model with an autoregressive term was used for statistical analysis. RESULTS: A time-dynamic effect was found for socioeconomic inequalities in the incidence of COVID-19. In fact, belonging to a lower-income area changed from being a risk factor (Waves 1, 2, 4 and 5) to being a protective factor in the sixth wave of the pandemic. Age also had a significant effect on incidence, which also changed over the different waves of the pandemic. Finally, the lower-income areas showed a comparatively lower incidence during the ascending phase of the epidemic waves. CONCLUSION: Socioeconomic inequalities in COVID-19 changed by wave, age group and wave phase.

Beyond standardized mortality ratios; some uses of smoothed age-specific mortality rates on small areas studies.

BACKGROUND: Most epidemiological risk indicators strongly depend on the age composition of populations, which makes the direct comparison of raw (unstandardized) indicators misleading because of the different age structures of the spatial units of study. Age-standardized rates (ASR) are a common solution for overcoming this confusing effect. The main drawback of ASRs is that they depend on age-specific rates which, when working with small areas, are often based on very few, or no, observed cases for most age groups. A similar effect occurs with life expectancy at birth and many more epidemiological indicators, which makes standardized mortality ratios (SMR) the omnipresent risk indicator for small areas epidemiologic studies. METHODS: To deal with this issue, a multivariate smoothing model, the M-model, is proposed in order to fit the age-specific probabilities of death (PoDs) for each spatial unit, which assumes dependence between closer age groups and spatial units. This age-space dependence structure enables information to be transferred between neighboring consecutive age groups and neighboring areas, at the same time, providing more reliable age-specific PoDs estimates. RESULTS: Three case studies are presented to illustrate the wide range of applications that smoothed age specific PoDs have in practice . The first case study shows the application of the model to a geographical study of lung cancer mortality in women. This study illustrates the convenience of considering age-space interactions in geographical studies and to explore the different spatial risk patterns shown by the different age groups. Second, the model is also applied to the study of ischaemic heart disease mortality in women in two cities at the census tract level. Smoothed age-standardized rates are derived and compared for the census tracts of both cities, illustrating some advantages of this mortality indicator over traditional SMRs. In the latest case study, the model is applied to estimate smoothed life expectancy (LE), which is the most widely used synthetic indicator for characterizing overall mortality differences when (not so small) spatial units are considered. CONCLUSION: Our age-space model is an appropriate and flexible proposal that provides more reliable estimates of the probabilities of death, which allow the calculation of enhanced epidemiological indicators (smoothed ASR, smoothed LE), thus providing alternatives to traditional SMR-based studies of small areas.

Spatio-temporal impact of self-financed rotavirus vaccination on rotavirus and acute gastroenteritis hospitalisations in the Valencia region, Spain.

BACKGROUND: Several studies have shown a substantial impact of Rotavirus (RV) vaccination on the burden of RV and all-cause acute gastroenteritis (AGE). However, the results of most impact studies could be confused by a dynamic and complex space-time process. Therefore, there is a need to analyse the impact of RV vaccination on RV and AGE hospitalisations in a space-time framework to detect geographical-time patterns while avoiding the potential confusion caused by population inequalities in the impact estimations. METHODS: A retrospective population-based study using real-world data from the Valencia Region was performed among children aged less than 3 years old in the period 2005-2016. A Bayesian spatio-temporal model was constructed to analyse RV and AGE hospitalisations and to estimate the vaccination impact measured in averted hospitalisations. RESULTS: We found important spatio-temporal patterns in RV and AGE hospitalisations, RV vaccination coverage and in their associated adverted hospitalisations. Overall, ~ 1866 hospital admissions for RV were averted by RV vaccination during 2007-2016. Despite the low-medium vaccine coverage (~ 50%) in 2015-2016, relevant 36 and 20% reductions were estimated in RV and AGE hospitalisations respectively. CONCLUSIONS: The introduction of the RV vaccines has substantially reduced the number of RV hospitalisations, averting ~ 1866 admissions during 2007-2016 which were space and time dependent. This study improves the methodologies commonly used to estimate the RV vaccine impact and their interpretation.

Seroprevalence of antibodies against serogroup C meningococci in the region of Valencia, Spain: Impact of meningococcal C conjugate vaccination.

BACKGROUND AND AIMS: Meningococcal C conjugate (MCC) vaccination programs provide direct and indirect protection against meningococcal disease. However, a decrease in the antibodies could affect herd immunity. We conducted a seroprevalence study to assess the immunity in subjects 8-12years after different MCCV vaccination programs were launched and evaluated the impact of vaccination on seroprotection. METHODS: Seroepidemiological study conducted from October 2010 to April 2012 in the region of Valencia, Spain. Sample size was not proportional to the population but to the expected seroprotection by age group. Sera from subjects that were≥3years old were tested using a standardized complement-mediated serum bactericidal antibodies (SBA) assay. Age-stratified proportions of subjects with SBA titers≥8 were considered seroprotected and evaluated. A multivariate logistic regression model was performed to evaluate the impact of vaccination on the seroprotection. RESULTS: Serum samples from 1880 subjects were collected. In total, 523 (27.8%) of the 1880 subjects and 446 (31.2%) of the 1430 subjects<30years (targeted to any vaccination campaign) showed protective SBA titers. The highest percentage of seroprotected subjects (67.8%, 95%CI 56.9-77.4) was observed in those that were vaccinated in a catch-up campaign at 10-13years of age (20-21years old at the time of blood sampling). Those scheduled for immunization in infancy at 2, 4 and 6months of age (7-8years at blood sample) represented the lowest (7.1%, 95% CI 3.3-13.1) number of seroprotected subjects. Having received one vaccine dose after 12months of age was associated with increased seroprotection. The present study revealed a positive correlation between the increasing age at vaccination and longer duration of seroprotection. CONCLUSION: Only one in three subjects who were vaccinated with MCC vaccine was seroprotected after 8-12years. These findings emphasize that seroprevalence studies are essential to identify susceptible cohorts and to inform vaccine policy.

A Multilevel Regression Model for Geographical Studies in Sets of Non-Adjacent Cities.

In recent years, small-area-based ecological regression analyses have been published that study the association between a health outcome and a covariate in several cities. These analyses have usually been performed independently for each city and have therefore yielded unrelated estimates for the cities considered, even though the same process has been studied in all of them. In this study, we propose a joint ecological regression model for multiple cities that accounts for spatial structure both within and between cities and explore the advantages of this model. The proposed model merges both disease mapping and geostatistical ideas. Our proposal is compared with two alternatives, one that models the association for each city as fixed effects and another that treats them as independent and identically distributed random effects. The proposed model allows us to estimate the association (and assess its significance) at locations with no available data. Our proposal is illustrated by an example of the association between unemployment (as a deprivation surrogate) and lung cancer mortality among men in 31 Spanish cities. In this example, the associations found were far more accurate for the proposed model than those from the fixed effects model. Our main conclusion is that ecological regression analyses can be markedly improved by performing joint analyses at several locations that share information among them. This finding should be taken into consideration in the design of future epidemiological studies.

Trends in socioeconomic inequalities in preventable mortality in urban areas of 33 Spanish cities, 1996-2007 (MEDEA project).

BACKGROUND: Preventable mortality is a good indicator of possible problems to be investigated in the primary prevention chain, making it also a useful tool with which to evaluate health policies particularly public health policies. This study describes inequalities in preventable avoidable mortality in relation to socioeconomic status in small urban areas of thirty three Spanish cities, and analyses their evolution over the course of the periods 1996-2001 and 2002-2007. METHODS: We analysed census tracts and all deaths occurring in the population residing in these cities from 1996 to 2007 were taken into account. The causes included in the study were lung cancer, cirrhosis, AIDS/HIV, motor vehicle traffic accidents injuries, suicide and homicide. The census tracts were classified into three groups, according their socioeconomic level. To analyse inequalities in mortality risks between the highest and lowest socioeconomic levels and over different periods, for each city and separating by sex, Poisson regression were used. RESULTS: Preventable avoidable mortality made a significant contribution to general mortality (around 7.5%, higher among men), having decreased over time in men (12.7 in 1996-2001 and 10.9 in 2002-2007), though not so clearly among women (3.3% in 1996-2001 and 2.9% in 2002-2007). It has been observed in men that the risks of death are higher in areas of greater deprivation, and that these excesses have not modified over time. The result in women is different and differences in mortality risks by socioeconomic level could not be established in many cities. CONCLUSIONS: Preventable mortality decreased between the 1996-2001 and 2002-2007 periods, more markedly in men than in women. There were socioeconomic inequalities in mortality in most cities analysed, associating a higher risk of death with higher levels of deprivation. Inequalities have remained over the two periods analysed. This study makes it possible to identify those areas where excess preventable mortality was associated with more deprived zones. It is in these deprived zones where actions to reduce and monitor health inequalities should be put into place. Primary healthcare may play an important role in this process.

Vaccine coverage estimation using a computerized vaccination registry with potential underreporting and a seroprevalence study.

OBJECTIVE: To develop a method to estimate vaccination coverage using both a computerized vaccine registry with an unknown underreporting rate and a seroprevalence study. A real example of a meningococcal C conjugate vaccine (MCCV) coverage estimation is studied to illustrate the proposed methodology. METHODS: We reviewed the Vaccine Information System of Valencia (Sistema de Información Vacunal, SIV) for the MCCV status of 1430 subjects aged 3-29 years as part of a seroprevalence study. When MCCV was not registered in SIV, subjects were classified into three groups (MCCV non-registered, no vaccination records and missing information) depending on the registry of other vaccines. A Bayesian model was developed to ascertain the percentage of MCCV-vaccinated subjects based on the meningococcal C seroprotection levels from the seroprevalence study. RESULTS: The seroprotection levels in subjects with no MCCV registered in SIV (358) were similar to those in subjects with MCCV registered (1072). This indicated a large proportion of vaccinated subjects with no MCCV registered. The estimated vaccine coverage was over 80% in all age groups, except >22 years, where it was 67.6% (95% CI: [54.0-80.4]), which corresponded to those aged over 13 years at the time of the catch-up campaign. An underreporting rate of 23.5-73.4%, depending on the age group, was estimated in those vaccinated in the 2002 catch-up campaign. CONCLUSION: The Bayesian model allowed for a more realistic estimation of MCCV uptake. In this example, we quantified the underreporting of a vaccine registry, especially occurring during a catch-up campaign that occurred at the establishment of the registry.

Trends in socioeconomic inequalities in amenable mortality in urban areas of Spanish cities, 1996-2007.

BACKGROUND: While research continues into indicators such as preventable and amenable mortality in order to evaluate quality, access, and equity in the healthcare, it is also necessary to continue identifying the areas of greatest risk owing to these causes of death in urban areas of large cities, where a large part of the population is concentrated, in order to carry out specific actions and reduce inequalities in mortality. This study describes inequalities in amenable mortality in relation to socioeconomic status in small urban areas, and analyses their evolution over the course of the periods 1996-99, 2000-2003 and 2004-2007 in three major cities in the Spanish Mediterranean coast (Alicante, Castellón, and Valencia). METHODS: All deaths attributed to amenable causes were analysed among non-institutionalised residents in the three cities studied over the course of the study periods. Census tracts for the cities were grouped into 3 socioeconomic status levels, from higher to lower levels of deprivation, using 5 indicators obtained from the 2001 Spanish Population Census. For each city, the relative risks of death were estimated between socioeconomic status levels using Poisson's Regression models, adjusted for age and study period, and distinguishing between genders. RESULTS: Amenable mortality contributes significantly to general mortality (around 10%, higher among men), having decreased over time in the three cities studied for men and women. In the three cities studied, with a high degree of consistency, it has been seen that the risks of mortality are greater in areas of higher deprivation, and that these excesses have not significantly modified over time. CONCLUSIONS: Although amenable mortality decreases over the time period studied, the socioeconomic inequalities observed are maintained in the three cities. Areas have been identified that display excesses in amenable mortality, potentially attributable to differences in the healthcare system, associated with areas of greater deprivation. Action must be taken in these areas of greater inequality in order to reduce the health inequalities detected. The causes behind socioeconomic inequalities in amenable mortality must be studied in depth.

Socioeconomic inequalities in injury mortality in small areas of 15 European cities.

This study analysed socioeconomic inequalities in mortality due to injuries in small areas of 15 European cities, by sex, at the beginning of this century. A cross-sectional ecological study with units of analysis being small areas within 15 European cities was conducted. Relative risks of injury mortality associated with the socioeconomic deprivation index were estimated using hierarchical Bayesian model. The number of small areas varies from 17 in Bratislava to 2666 in Turin. The median population per small area varies by city (e.g. Turin had 274 inhabitants per area while Budapest had 76,970). Socioeconomic inequalities in all injury mortality are observed in the majority of cities and are more pronounced in men. In the cities of northern and western Europe, socioeconomic inequalities in injury mortality are found for most types of injuries. These inequalities are not significant in the majority of cities in southern Europe among women and in the majority of central eastern European cities for both sexes. The results confirm the existence of socioeconomic inequalities in injury related mortality and reveal variations in their magnitude between different European cities.

Socio-economic inequalities in mortality due to injuries in small areas of ten cities in Spain (MEDEA Project).

OBJECTIVES: To analyse socio-economic inequalities in mortality due to injuries among census tracts of ten Spanish cities by sex and age in the period 1996-2003. METHODS: This is a cross-sectional ecological study where the units of analysis are census tracts. The study population consisted of people residing in the cities during the period 1996-2003. For each census tract we obtained an index of socio-economic deprivation, and estimated standardized mortality ratios using hierarchical Bayesian models which take into account the spatial structure of the data. RESULTS: In the majority of the cities, the geographical pattern of total mortality from injuries is similar to that of the socio-economic deprivation index. There is an association between mortality due to injuries and the deprivation index in the majority of the cities which is more important among men and among those younger than 45 years. In these groups, traffic injuries and overdoses are the causes most often associated with deprivation in the cities. The percentage of excess mortality from injuries related to socio-economic deprivation is higher than 20% in the majority of the cities, the cause with the highest percentage being drug overdose. CONCLUSIONS: In most cities, there are socio-economic inequalities in mortality due to overdose and traffic injuries. In contrast, few cities have found association between suicide mortality and deprivation. Finally, no association was found between deprivation and deaths due to falls. Inequalities are higher in men and those under 45 years of age. These results highlight the importance of intra-urban inequalities in mortality due to injuries.

Bayesian factor analysis to calculate a deprivation index and its uncertainty.

BACKGROUND: Procedures for calculating deprivation indices in epidemiologic studies often show some common problems because the spatial dependence between units of analysis and uncertainty of the estimates is not usually accounted for. This work highlights these problems and illustrates how spatial factor Bayesian modeling could alleviate them. METHODS: This study applies a cross-sectional ecological design to analyze the census tracts of 3 Spanish cities. To calculate the deprivation index, we used 5 socioeconomic indicators that comprise the deprivation index calculated in the MEDEA project. The deprivation index was estimated by a Bayesian factor analysis using hierarchical models, which takes the spatial dependence of the study units into account. We studied the relationship between this index and the one obtained using principal component analysis. Various analyses were carried out to assess the uncertainty obtained in the index. RESULTS: A high correlation was observed between the index obtained and the non-Bayesian index, but this relationship is not linear and there is disagreement between the methods when the areas are grouped according to quantiles. When the deprivation index is calculated using summary statistics based on the posterior distributions, the uncertainty of the index in each census tract is not taken into account. Failure to take this uncertainty into account may result in misclassification bias in the census tracts when these are grouped according to quantiles of the deprivation index. CONCLUSIONS: Not taking uncertainty into account may result in misclassification bias in the census tracts. This bias could interfere in subsequent analyses that include the deprivation index. Our proposal provides another tool for identifying groups with greater deprivation and for improving decision-making for public policy planning.

FluDetWeb: an interactive web-based system for the early detection of the onset of influenza epidemics.

BACKGROUND: The early identification of influenza outbreaks has became a priority in public health practice. A large variety of statistical algorithms for the automated monitoring of influenza surveillance have been proposed, but most of them require not only a lot of computational effort but also operation of sometimes not-so-friendly software. RESULTS: In this paper, we introduce FluDetWeb, an implementation of a prospective influenza surveillance methodology based on a client-server architecture with a thin (web-based) client application design. Users can introduce and edit their own data consisting of a series of weekly influenza incidence rates. The system returns the probability of being in an epidemic phase (via e-mail if desired). When the probability is greater than 0.5, it also returns the probability of an increase in the incidence rate during the following week. The system also provides two complementary graphs. This system has been implemented using statistical free-software (R and WinBUGS), a web server environment for Java code (Tomcat) and a software module created by us (Rdp) responsible for managing internal tasks; the software package MySQL has been used to construct the database management system. The implementation is available on-line from: http://www.geeitema.org/meviepi/fludetweb/. CONCLUSION: The ease of use of FluDetWeb and its on-line availability can make it a valuable tool for public health practitioners who want to obtain information about the probability that their system is in an epidemic phase. Moreover, the architecture described can also be useful for developers of systems based on computationally intensive methods.

Improving multilevel analyses: the integrated epidemiologic design.

Multilevel analysis has been widely used to allow the simultaneous examination of the effects of individual- and group-level variables on individual health outcomes. In spite of its utility, multilevel design can have some drawbacks in the estimation of risk factor effects when the within-group variation of variables of interest is small relative to between-group variation. An extreme case of this is a group-level risk factor, which by definition has no within-group variation. To improve the estimation of group-level and individual-level risk factor effects, we consider an integrated epidemiologic design using a population-based estimating equation approach that can be considered a further extension of the multilevel design. Although the integrated design uses the same individual-level and group-level data as the multilevel design, it includes aggregated health outcome data in each group as additional information. This paper explains differences between the 2 designs, describing advantages and disadvantages of the integrated design over the multilevel design. The 2 designs are applied to a real example of mortality following chronic kidney disease, illustrating differences that might be encountered in practice.

[Methods to smooth mortality indicators: application to analysis of inequalities in mortality in Spanish cities [the MEDEA Project]]. / Métodos para la suavización de indicadores de mortalidad: aplicación al análisis de desigualdades en mortalidad en ciudades del Estado español (Proyecto MEDEA).

Although there is some experience in the study of mortality inequalities in Spanish cities, there are large urban centers that have not yet been investigated using the census tract as the unit of territorial analysis. The coordinated project <> was designed to fill this gap, with the participation of 10 groups of researchers in Andalusia, Aragon, Catalonia, Galicia, Madrid, Valencia, and the Basque Country. The MEDEA project has four distinguishing features: a) the census tract is used as the basic geographical area; b) statistical methods that include the geographical structure of the region under study are employed for risk estimation; c) data are drawn from three complementary data sources (information on air pollution, information on industrial pollution, and the records of mortality registrars), and d) a coordinated, large-scale analysis, favored by the implantation of coordinated research networks, is carried out. The main objective of the present study was to explain the methods for smoothing mortality indicators in the context of the MEDEA project. This study focusses on the methodology and the results of the Besag, York and Mollié model (BYM) in disease mapping. In the MEDEA project, standardized mortality ratios (SMR), corresponding to 17 large groups of causes of death and 28 specific causes, were smoothed by means of the BYM model; however, in the present study this methodology was applied to mortality due to cancer of the trachea, bronchi and lung in men and women in the city of Barcelona from 1996 to 2003. As a result of smoothing, a different geographical pattern for SMR in both genders was observed. In men, a SMR higher than unity was found in highly deprived areas. In contrast, in women, this pattern was observed in more affluent areas.

Métodos para la suavización de indicadores de mortalidad: aplicación al análisis de desigualdades en mortalidad en ciudades del Estado español (Proyecto MEDEA) / Methods to smooth mortality indicators: application to analysis of inequalities in mortality in Spanish cities [the MEDEA Project]

Aunque la experiencia en el estudio de las desigualdadesen la mortalidad en las ciudades españolas es amplia, quedangrandes núcleos urbanos que no han sido investigadosutilizando la sección censal como unidad de análisis territorial.En este contexto se sitúa el proyecto coordinado ®Desigualdadessocioeconómicas y medioambientales en la mortalidaden ciudades de España. Proyecto MEDEA», en el cualparticipan 10 grupos de investigadores de Andalucía, Aragón,Cataluña, Galicia, Madrid, Comunitat Valenciana y PaísVasco. Cabe señalar cuatro particularidades: a) se utiliza comoárea geográfica básica la sección censal; b) se emplean métodosestadísticos que tienen en cuenta la estructura geográficade la región de estudio para la estimación de riesgos; c) seaprovechan las oportunidades que ofrecen 3 fuentes de datoscomplementarias (información sobre contaminación atmosférica,información sobre contaminación industrial y registrosde mortalidad), y d) se emprende un análisis coordinado degran alcance, favorecido por la implantación de la redes temáticasde investigación. El objetivo de este trabajo es explicarlos métodos para la suavización de indicadores de mortalidaden el proyecto MEDEA. El artículo se centra en lametodología y los resultados del modelo de mapa de enfermedadesde Besag, York y Mollié (BYM). Aunque en el proyectose han suavizado, mediante el modelo BYM, las razonesde mortalidad estandarizadas (RME) correspondientesa 17 grandes grupos de causas de defunción y 28 causasespecíficas, aquí se aplica esta metodología a la mortalidadpor cáncer de tráquea, de bronquios y de pulmón en ambossexos en la ciudad de Barcelona durante el período 1996-2003(AU)

Como resultado se aprecia un diferente patrón geográfico enlas RME suavizadas en ambos sexos. En los hombres se observanunas RME mayores que la unidad en los barrios conmayor privación socioeconómica. En las mujeres este patrónse observa en las zonas con un mayor nivel socioeconómico(AU)

Although there is some experience in the study of mortalityinequalities in Spanish cities, there are large urban centersthat have not yet been investigated using the census tract asthe unit of territorial analysis. The coordinated project ®Socioeconomicand environmental inequalities in mortality in Spanishcities. The MEDEA project» was designed to fill this gap,with the participation of 10 groups of researchers in Andalusia,Aragon, Catalonia, Galicia, Madrid, Valencia, and the BasqueCountry. The MEDEA project has four distinguishing features:a) the census tract is used as the basic geographicalarea; b) statistical methods that include the geographical structureof the region under study are employed for risk estimation;c) data are drawn from three complementary data sources(information on air pollution, information on industrialpollution, and the records of mortality registrars), and d) a coordinated,large-scale analysis, favored by the implantation ofcoordinated research networks, is carried out. The main objective of the present study was to explain the methods for smoothingmortality indicators in the context of the MEDEA project.This study focusses on the methodology and the resultsof the Besag, York and Mollié model (BYM) in disease mapping.In the MEDEA project, standardized mortality ratios(SMR), corresponding to 17 large groups of causes of deathand 28 specific causes, were smoothed by means of the BYMmodel; however, in the present study this methodology wasapplied to mortality due to cancer of the trachea, bronchi andlung in men and women in the city of Barcelona from 1996 to2003. As a result of smoothing, a different geographical patternfor SMR in both genders was observed. In men, a SMRhigher than unity was found in highly deprived areas. In contrast,in women, this pattern was observed in more affluentareas(AU)