Daily Mortality Analysis by Using Different Regression Models in Philadelphia County, 1973–1990

The effects of particulate and ozone pollution on mortality in Moscow, Russia

The objectives of this study were (1) to evaluate how acute mortality responds to changes in particulate and ozone (O(3)) pollution levels, (2) to identify vulnerable population groups by age and cause of death, and (3) to address the problem of interaction between the effects of O(3) and particulate pollution. Time-series of daily mortality counts, air pollution, and air temperature were obtained for the city of Moscow during a 3-year period (2003-2005). To estimate the pollution-mortality relationships, we used a log-linear model that controlled for potential confounding by daily air temperature and longer term trends. The effects of 10 mug/m(3) increases in daily average measures of particulate matter </=10 mum in aerodynamic diameter (PM(10)) and O(3) were, respectively, (1) a 0.33% [95% confidence interval (CI) 0.09-0.57] and 1.09% (95% CI 0.71-1.47) increase in all-cause non-accidental mortality in Moscow; (2) a 0.66% (0.30-1.02) and 1.61% (1.01-2.21) increase in mortality from ischemic heart disease; (3) a 0.48% (0.02-0.94) and 1.28% (0.54-2.02) increase in mortality from cerebrovascular diseases. In the age group >75 years, mortality increments were consistently higher, typically by factor of 1.2 - 1.5, depending upon the cause of death. PM(10)-mortality relationships were significantly modified by O(3) levels. On the days with O(3) concentrations above the 90th percentile, PM(10) risk for all-cause mortality was threefold greater and PM(10) risk for cerebrovascular disease mortality was fourfold greater than the unadjusted risk estimate.

Health effects of particulate air pollution

In the 1980's it was generally felt that particulate air pollution concentrations in the United States were not a hazard to the public health. However, in the early 1990's the application of econometric time-series studies and prospective cohort studies suggested increased mortality associated with acute (daily) and chronic (decades) exposures to particulate air pollution commonly observed in the developed world. The epidemiologic evidence was not supported by evidence of causal associations from other disciplines. Nevertheless, the EPA moved to tighten controls on fine particulate air pollution. The debate over the science was played out in public hearings and the courts. The experience provides lessons on the use of epidemiologic data in setting public policy.

Air pollution and postneonatal mortality in a tropical city: Kaohsiung, Taiwan

With growing evidence of the association between daily mortality and air pollution in adults, it is important to investigate whether infants are also susceptible to the adverse health effects of ambient air pollutants. The purpose of this study is to examine the relationship between air pollution and postneonatal mortality in Kaohsiung, Taiwan, a large industrial city with a tropical climate, during the period 1994-2000, using a case-crossover analysis. Case-crossover analysis provides an alternative to Poisson time-series regression for studying the short-term adverse health effects of air pollution. The air pollutants examined included particulate matter (PM(10)), sulfur dioxide (SO(2)), ozone (O(3)), nitrogen dioxide (NO(2)), and carbon monoxide (CO). The risk of postneonatal deaths was estimated to increase by 4.0% per 67 microg/m(3) (the interquartile range in daily ambient concentration of PM(10)) for PM(10), 1.8% per 17.84 ppb for NO(2), 5.1% per 0.31 ppm for CO, and 4.6% per 19.20 ppb for O(3). Although positive, none of these associations achieved statistical significance. The established link between air pollution levels and infant mortality may not be as strong in cities with tropical climates, although other factors such as differences in pollutant mix or the underlying health of the postneonates may explain the lack of a strong association in this study. Further studies of this type in cities with varying climates and cultures are needed.

Correlation between air pollution and postneonatal mortality in a subtropical city: Taipei, Taiwan

With growing evidence of the association between daily mortality and air pollution exposure in adults, it is important to investigate whether infants are also susceptible. The purpose of this study was to examine the relationship between air pollution exposure and postneonatal, defined as infant of more than 27 d and less than 1 yr old, mortality in Taipei, Taiwan's largest city, which has a subtropical climate, for the period 1994-2000, using a case-crossover analysis. This design is an alternative to Poisson time-series regression for studying the short-term adverse health effects of air pollution. The air pollutants examined included particulate matter (PM10), sulfur dioxide (SO2), ozone (O3), nitrogen dioxide (NO2), and carbon monoxide (CO). The risk of postneonatal deaths was estimated to increase by 3.1% for PM10, 4.1% for SO2, 1.7% for NO2, 3.8% for CO, and 0.1% for O3 for each interquartile range change, respectively. However, the associations were without statistical significance. The established link between air pollution levels and infant mortality may not be as strong in cities with subtropical climates, although other factors such as differences in pollutant component composition or the underlying health of the postneonates may explain the lack of a strong association in this study. Further studies of this type in cities with varying climates and cultures are needed.

Ozone exposure and mortality: an empiric bayes metaregression analysis

BACKGROUND

Results from time-series epidemiologic studies evaluating the relationship between ambient ozone concentrations and premature mortality vary in their conclusions about the magnitude of this relationship, if any, making it difficult to estimate public health benefits of air pollution control measures. We conducted an empiric Bayes metaregression to estimate the ozone effect on mortality, and to assess whether this effect varies as a function of hypothesized confounders or effect modifiers.

METHODS

We gathered 71 time-series studies relating ozone to all-cause mortality, and we selected 48 estimates from 28 studies for the metaregression. Metaregression covariates included the relationship between ozone concentrations and concentrations of other air pollutants, proxies for personal exposure-ambient concentration relationships, and the statistical methods used in the studies. For our metaregression, we applied a hierarchical linear model with known level-1 variances.

RESULTS

We estimated a grand mean of a 0.21% increase (95% confidence interval = 0.16-0.26%) in mortality per 10-microg/m increase of 1-hour maximum ozone (0.41% increase per 10 ppb) without controlling for other air pollutants. In the metaregression, air-conditioning prevalence and lag time were the strongest predictors of between-study variability. Air pollution covariates yielded inconsistent findings in regression models, although correlation analyses indicated a potential influence of summertime PM2.5.

CONCLUSIONS

These findings, coupled with a greater relative risk of ozone in the summer versus the winter, demonstrate that geographic and seasonal heterogeneity in ozone relative risk should be anticipated, but that the observed relationship between ozone and mortality should be considered for future regulatory impact analyses.

A review and critique of the EPA’s rationale for a fine particle standard

I review the rationale for the Environmental Protection Agency's 1996 fine particle standard, which was based almost entirely on the epidemiological data with neither support from Toxicology nor understanding of mechanism. While many epidemiological papers available in 1996 reported associations between ambient particles and adverse effects on human health, many others did not and the evidence fell far short of supporting a causal association between particle mass concentration and human health. The literature appearing after 1996 further complicates the picture. The large studies that have appeared after 1996, such as National Mortality Morbidity and Air Pollution Study, and the reanalyses of the American Cancer Society II study, report risks that are substantially smaller than the risks reported in the 1996 Criteria Document and Staff Paper. Moreover, concerns about confounding by weather, temporal trends and co-pollutants remain unresolved. Other issues having to do with model choice have resurfaced as a result of reanalyses of critical data to address a glitch in a widely used software package for time-series epidemiology studies of air pollution. Finally, contemporary examples show that the results of observational epidemiology studies can be seriously biased, particularly when estimated risks are small, as is the case with studies of air pollution. The Agency has largely ignored these issues. I conclude that a particle mass standard is not defensible on the basis of a causal association between ambient particle mass and adverse effects on human health. Such a standard may be justifiable on the basis of the precautionary principle, however. The Agency could argue that the Science raises concerns about current levels of air pollution, and that reduction of ambient fine particulate matter mass, if it could be achieved without an increase in the level of the ultrafines, could have positive effects on human health. If the Agency justifies a particulate matter mass standard on these grounds then the debate over the form and level of the standard will, for all practical purposes, belong strictly in the Policy arena.

Social determinants of health: implications for environmental health promotion

In this article, the authors draw on the disciplines of sociology and environmental and social epidemiology to further understanding of mechanisms through which social factors contribute to disparate environmental exposures and health inequalities. They propose a conceptual framework for environmental health promotion that considers dynamic social processes through which social and environmental inequalities--and associated health disparities--are produced, reproduced, and potentially transformed. Using empirical evidence from the published literature, as well as their own practical experiences in conducting community-based participatory research in Detroit and Harlem, the authors examine health promotion interventions at various levels (community-wide, regional, and national) that aim to improve population health by addressing various aspects of social processes and/or physical environments. Finally, they recommend moving beyond environmental remediation strategies toward environmental health promotion efforts that are sustainable and explicitly designed to reduce social, environmental, and health inequalities.

The particulate air pollution controversy

Scientists, regulators, legislators, and segments of industry and the lay public are attempting to understand and respond to epidemiology findings of associations between measures of modern particulate air pollutants (PM) and adverse health outcomes in urban dwellers. The associations have been interpreted to imply that tens of thousands of Americans are killed annually by small daily increments in PM. These epidemiology studies and their interpretations have been challenged, although it is accepted that high concentrations of air pollutants have claimed many lives in the past. Although reproducible and statistically significant, the relative risks associated with modern PM are very small and confounded by many factors. Neither toxicology studies nor human clinical investigations have identified the components and/or characteristics of PM that might be causing the health-effect associations. Currently, a massive worldwide research effort is under way in an attempt to identify whom might be harmed and by what substances and mechanisms. Finding the answers is important, because control measures have the potential not only to be costly but also to limit the availability of goods and services that are important to public health.

Relationship between air pollution and daily mortality in a subtropical city: Taipei, Taiwan

Air pollution has been associated with daily mortality in numerous studies over the past decade. However most of these studies were conducted in the United States and Europe with relatively few done in Asia. In the current study, the association between ambient air pollution and daily mortality in Taipei, Taiwan's largest city which has a subtropical climate was undertaken, for the period 1994-1998 using a case-crossover analysis. This design is an alternative to Poisson time series regression for studying the short-term adverse health effects of air pollution. The air pollutants examined included particulate matter (PM(10)), sulfur dioxide (SO(2)), ozone (O(3)), nitrogen dioxide (NO(2)), and carbon monoxide (CO). The largest observed effect, which was without statistical significance, was seen for NO(2) and CO levels on deaths due to respiratory diseases (ORs=1.013 and 1.014, respectively). The well established link between air pollution levels and daily mortality may not be as strong in cities in subtropical areas, although other factors such as differences in pollutant mix or the underlying health of the population may explain the lack of a strong association in this study. Further studies of this type in cities with varying climates and cultures are needed.

Comparison of time series and case-crossover analyses of air pollution and hospital admission data

BACKGROUND

Time series analysis is the most commonly used technique for assessing the association between counts of health events over time and exposure to ambient air pollution. Recently, case-crossover analysis has been proposed as an alternative analytical approach. While each technique has its own advantages and disadvantages, there remains considerable uncertainty as to which statistical methodology is preferable for evaluating data of this type.

METHODS

The objective of this paper is to evaluate the performance of different variations of these two procedures using computer simulation. Hospital admission data were generated under realistic models with known parameters permitting estimates based on time series and case-crossover analyses to be compared with these known values.

RESULTS

While accurate estimates can be achieved with both methods, both methods require some decisions to be made by the researcher during the course of the analysis. With time series analysis, it is necessary to choose the time span in the LOESS smoothing process, or degrees of freedom when using natural cubic splines. For case-crossover studies using either uni- or bi-directional control selection strategies, the choice of time intervals needs to be made.

CONCLUSIONS

We prefer the times series approach because the best estimates of risk that can be obtained with time series analysis are more precise than the best estimates based on case-crossover analysis.

Association Between Particulate Air Pollution and First Hospital Admission for Childhood Respiratory Illness in Vancouver, Canada

In this study, the authors assessed the impact of particulate air pollution on first respiratory hospitalization. Study subjects were children less than 3 years of age living in Vancouver, British Columbia, who had their first hospitalization as a result of any respiratory disease (ICD-9 codes 460-519) during the period from June 1, 1995, to March 31, 1999. The authors used logistic regression to estimate the associations between ambient concentrations of particulate matter (PM) and first hospitalization. The adjusted odds ratios for first respiratory hospitalization associated with mean and maximal PM10-2.5 with a lag of 3 days were 1.12 (95% confidence interval: 0.98, 1.28) and 1.13 (1.00, 1.27). After adjustment for gaseous pollutants, the corresponding odds ratios were 1.22 (1.02, 1.48) and 1.14 (0.99, 1.32). The data indicated the possibility of harmful effects from coarse PM on first hospitalization for respiratory disease in early childhood.

Relationship between air pollution and daily mortality in a tropical city: Kaohsiung, Taiwan

Air pollution has been associated with daily mortality in numerous studies over the past decade. However, most of these studies were conducted in the United States and Europe, with relatively few done in Asia. In this study, the association between ambient air pollution and daily mortality in Kaohsiung, Taiwan, a large industrial city with a tropical climate, was investigated for the period 1994-2000 using a case-crossover analysis. This design is an alternative to Poisson time-series regression for studying the short-term adverse health effects of air pollution. The air pollutants examined included particulate matter (PM10), sulfur dioxide (SO2), ozone (O3), nitrogen dioxide (NO2), and carbon monoxide (CO). No significant effects were found between PM10 and SO2 exposure levels and respiratory-related mortality. The well-established link between air pollution levels and daily mortality may not be as strong in cities in tropical areas, although other factors such as differences in pollutant mixtures or underlying health of the population may explain the lack of a strong association in this study. Further studies of this type in cities with varying climates and cultures are needed.

A population exposure model for particulate matter: case study results for PM(2.5) in Philadelphia, PA

A population exposure model for particulate matter (PM), called the Stochastic Human Exposure and Dose Simulation (SHEDS-PM) model, has been developed and applied in a case study of daily PM(2.5) exposures for the population living in Philadelphia, PA. SHEDS-PM is a probabilistic model that estimates the population distribution of total PM exposures by randomly sampling from various input distributions. A mass balance equation is used to calculate indoor PM concentrations for the residential microenvironment from ambient outdoor PM concentrations and physical factor data (e.g., air exchange, penetration, deposition), as well as emission strengths for indoor PM sources (e.g., smoking, cooking). PM concentrations in nonresidential microenvironments are calculated using equations developed from regression analysis of available indoor and outdoor measurement data for vehicles, offices, schools, stores, and restaurants/bars. Additional model inputs include demographic data for the population being modeled and human activity pattern data from EPA's Consolidated Human Activity Database (CHAD). Model outputs include distributions of daily total PM exposures in various microenvironments (indoors, in vehicles, outdoors), and the contribution from PM of ambient origin to daily total PM exposures in these microenvironments. SHEDS-PM has been applied to the population of Philadelphia using spatially and temporally interpolated ambient PM(2.5) measurements from 1992-1993 and 1990 US Census data for each census tract in Philadelphia. The resulting distributions showed substantial variability in daily total PM(2.5) exposures for the population of Philadelphia (median=20 microg/m(3); 90th percentile=59 microg/m(3)). Variability in human activities, and the presence of indoor-residential sources in particular, contributed to the observed variability in total PM(2.5) exposures. The uncertainty in the estimated population distribution for total PM(2.5) exposures was highest at the upper end of the distribution and revealed the importance of including estimates of input uncertainty in population exposure models. The distributions of daily microenvironmental PM(2.5) exposures (exposures due to time spent in various microenvironments) indicated that indoor-residential PM(2.5) exposures (median=13 microg/m(3)) had the greatest influence on total PM(2.5) exposures compared to the other microenvironments. The distribution of daily exposures to PM(2.5) of ambient origin was less variable across the population than the distribution of daily total PM(2.5) exposures (median=7 microg/m(3); 90th percentile=18 microg/m(3)) and similar to the distribution of ambient outdoor PM(2.5) concentrations. This result suggests that human activity patterns did not have as strong an influence on ambient PM(2.5) exposures as was observed for exposure to other PM(2.5) sources. For most of the simulated population, exposure to PM(2.5) of ambient origin contributed a significant percent of the daily total PM(2.5) exposures (median=37.5%), especially for the segment of the population without exposure to environmental tobacco smoke in the residence (median=46.4%). Development of the SHEDS-PM model using the Philadelphia PM(2.5) case study also provided useful insights into the limitations of currently available data for use in population exposure models. In addition, data needs for improving inputs to the SHEDS-PM model, reducing uncertainty and further refinement of the model structure, were identified.

Assessment of recent ozone short-term epidemiologic studies

The U.S. Environmental Protection Agency (EPA) revised the National Ambient Air Quality Standards (NAAQS) for ozone in 1997 based largely on short-term ozone studies published up to 1995. The U.S. EPA's conclusions must now be updated because (1) the agency did not consider many new studies published since 1995 and (2) the agency did not critically review the studies published before 1995 (i.e., it accepted the stated conclusions). In this article, we examine many recently published short-term ozone studies including 17 hospital admissions studies, 10 mortality studies, and 6 summer-camp studies. Almost all of these studies reported a significant association between ambient levels of ozone and adverse health effects. However, on close examination, it is apparent that there are mixed findings from one study to another and even within the results of a single study. Moreover, questionable statistical analyses and failure to consider confounders make a number of the reported findings doubtful and even negative.

Combining contingency tables with missing dimensions

We propose a methodology for estimating the cell probabilities in a multiway contingency table by combining partial information from a number of studies when not all of the variables are recorded in all studies. We jointly model the full set of categorical variables recorded in at least one of the studies, and we treat the variables that are not reported as missing dimensions of the study-specific contingency table. For example, we might be interested in combining several cohort studies in which the incidence in the exposed and nonexposed groups is not reported for all risk factors in all studies while the overall numbers of cases and cohort size is always available. To account for study-to-study variability, we adopt a Bayesian hierarchical model. At the first stage of the model, the observation stage, data are modeled by a multinomial distribution with fixed total number of observations. At the second stage, we use the logistic normal (LN) distribution to model variability in the study-specific cells' probabilities. Using this model and data augmentation techniques, we reconstruct the contingency table for each study regardless of which dimensions are missing, and we estimate population parameters of interest. Our hierarchical procedure borrows strength from all the studies and accounts for correlations among the cells' probabilities. The main difficulty in combining studies recording different variables is in maintaining a consistent interpretation of parameters across studies. The approach proposed here overcomes this difficulty and at the same time addresses the uncertainty arising from the missing dimensions. We apply our modeling strategy to analyze data on air pollution and mortality from 1987 to 1994 for six U.S. cities by combining six cross-classifications of low, medium, and high levels of mortality counts, particulate matter, ozone, and carbon monoxide with the complication that four of the six cities do not report all the air pollution variables. Our goals are to investigate the association between air pollution and mortality by reconstructing the tables with missing dimensions, to determine the most harmful pollutant combinations, and to make predictions about these key issues for a city other than the six sampled. We find that, for high levels of ozone and carbon monoxide, the number of cases with a high number of deaths increases as the levels of particulate matter, PM10, increases and that the most harmful combinations corresponds to high levels of PM10, confirming prior findings that levels of PM10 higher than the NAAQS standard are harmful.

Daily mortality in relation to weather and air pollution in Christchurch, New Zealand

OBJECTIVE

To investigate the relationship between the daily number of deaths, weather and ambient air pollution.

METHOD

An ecological study. We assembled daily data for the city of Christchurch, New Zealand (population 300,000) from June 1988 to December 1993. We used Poisson regression models, controlling for season using a parametric method.

RESULTS

Above the third quartile (20.5 degrees C) of maximum temperature, an increase of 1 degree C was associated with a 1% (95% CI: 0.4 to 2.1%) increase in all-cause mortality and a 3% (0.1 to 6.0%) increase in respiratory mortality. An increase in PM10 of 10 micrograms/m3 was associated (after a lag of one day) with a 1% (0.5 to 2.2%) increase in all-cause mortality and a 4% (1.5 to 5.9%) increase in respiratory mortality. We found no evidence of interaction between the effects of temperature and particulate air pollution.

CONCLUSIONS

High temperatures and particulate air pollution are independently associated with increased daily mortality in Christchurch. The fact that these results are consistent with those of similar studies in other countries strengthens the argument that the associations are likely to be causal.

IMPLICATIONS

These findings contribute to evidence of health consequences of fuel combustion, both in the short term (from local air pollution) and in the long term (from the global climatic effects of increased atmospheric CO2).

Mortality and air pollution in Helsinki

In Helsinki, Finland, from 1987 to 1993, the authors studied the associations between daily concentrations of sulphur dioxide, nitrogen dioxide, ozone, total suspended particulates, and particulates with aerodynamic diameters less than 10 microm (PM10), and the daily number of deaths from all causes and from cardiovascular causes. Investigators used Poisson regressions to conduct analyses in two age groups, and they controlled for temperature, relative humidity, day of the week, month, year, long-term trend, holidays, and influenza epidemics. The PM10 levels were associated significantly with all-cause and cardiovascular mortality among persons under the age of 65 y of age. In the less-than-65-y age group, sulfur dioxide and ozone were also associated significantly with cardiovascular mortality. The effect of ozone was independent of the PM10 effect, whereas sulfur dioxide became nonsignificant when modeled with PM10. An increase of 10 microg/m3 in PM10 resulted in increases in total mortality and cardiovascular mortality of 3.5% (95% confidence interval=1.0, 5.8) and 4.1% (95% confidence interval=0.4, 10.3), respectively. A 20 microg/m3 increase in ozone was associated with a 9.9% (95% confidence interval=1.1, 19.5) increase in cardiovascular mortality; however, ozone results were inconsistent. Moreover, in addition to their separate effects, high concentrations of PM10, ozone, and nitrogen dioxide had a further harmful additive effect. Typically, PM10 was a better indicator of particulate pollution than total suspended particulates. The authors' findings suggest that (a) even low levels of particulates are related to an increase in cardiovascular mortality; (b) ozone--even in low concentrations--is associated, independently, with cardiovascular mortality; and (c) PM10, ozone, and nitrogen dioxide--the essential components of summertime pollution--have harmful interactions at high concentrations.