Air Pollution and Mortality: Issues and Uncertainties

Abstract

Results from 31 epidemiology studies linking air pollution with premature mortality are compared and synthesized. Consistent positive associations between mortality and various measures of air pollution have been shown within each of two fundamentally different types of regression studies and in many variations within these basic types; this is extremely unlikely to have occurred by chance. In this paper, the measure of risk used is the elasticity, which is a dimensionless regression coefficient defined as the percentage change in the dependent variable associated with a 1% change in an independent variable, evaluated at the means. This metric has the advantage of independence from measurement units and averaging times, and is thus suitable for comparisons within and between studies involving different pollutants. Two basic types of studies are considered: time-series studies involving daily perturbations, and cross-sectional studies involving longer-term spatial gradients. The latter include prospective studies of differences in individual survival rates in different locations and studies of the differences in annual mortality rates for various communities.

For a given data set, time-series regression results will vary according to the seasonal adjustment method used, the covariates included, and the lag structure assumed. The results from both types of cross-sectional regressions are highly dependent on the methods used to control for socioeconomic and personal lifestyle factors and on data quality. Amajor issue for all of these studies is that of partitioning the response among collinear pollution and weather variables. Previous studies showed that the variable with the least exposure measurement error may be favored in multiple regressions; assigning precise numerical results to a single pollutant is not possible under these circumstances. We found that the mean overall elasticity as obtained from timeseries studies for mortality with respect to various air pollutants entered jointly was about 0.048, with a range from 0.01 to 0.12. This implies that about 5% of daily mortality is associated with air pollution, on average. The corresponding values from population-based cross-sectional studies were similar in magnitude, but the results from the three recent prospective studies varied from zero to about five times as much. Long-term responses in excess of short-term responses might be interpreted as showing the existence of chronic effects, but the uncertainties inherent in both types of studies make such an interpretation problematic.     

Mortality and Air Pollution: Revisited

This paper represents an exploratory effort to estimate a physical nonlinear function between excess mortality rates and the SO₂ concentration with both considerations over econometric problems such as multicollinearity and heteroscedasticity (the residuals of regression analysis), and the threshold levels. Through a recursive and stepwise adjustment procedure, the average physical mortality function was generalized with much more complete specifications. That is, the generalized average mortality model includes not only the demographic, socioeconomic, and climatological determinants but also air pollution variable. The average pollution damage function developed in this study with observations from relevant SMSA’s which have pollution concentrations exceeding the threshold level represents an important departure from the prior studies in which sample observations were selected regardless of the SO₂ concentration level.     

Daily Mortality and Air Pollution in the Netherlands

ABSTRACT

We studied the association of daily mortality with short-term variations in the ambient concentrations of major gaseous pollutants and PM in the Netherlands. The magnitude of the association in the four major urban areas was compared with that in the remainder of the country. Daily cause-specific mortality counts, air quality, temperature, relative humidity, and influenza data were obtained from 1986 to 1994. The relationship between daily mortality and air pollution was modeled using Poisson regression analysis. We adjusted for potential confounding due to long-term and seasonal trends, influenza epidemics, ambient temperature and relative humidity, day of the week, and holidays, using generalized additive models.

Influenza episodes were associated with increased mortality up to 3 weeks later. Daily mortality was significantly associated with the concentration of all air pollutants. An increase in the PM₁₀ concentration by 100 u.g/m³ was associated with a relative risk (RR) of 1.02 for total mortality. The largest RRs were found for pneumonia deaths. Ozone had the most consistent, independent association with mortality. Particulate air pollution (e.g., PM₁₀, black smoke [BS]) was not more consistently associated with mortality than were the gaseous pollutants SO₂ and NO₂. Aerosol SO₄ ^-2, NO₃ ^-, and BS were more consistently associated with total mortality than was PM₁₀. The RRs for all pollutants were substantially larger in the summer months than in the winter months. The RR of total mortality for PM₁₀ was 1.10 for the summer and 1.03 for the winter. There was no consistent difference between RRs in the four major urban areas and the more rural areas.     

Measurements of Morbidity and Mortality Related to Air Pollution

Examination of total deaths in New York City by day of occurrence shows periodic peaks in mortality which are associated with periods of high air pollution. These peaks are usually associated with periods of low wind speed and temperature inversion conditions which permit air pollution to build up to high levels. Unlike the experience of London, fog is not a necessary part of this picture, and therefore the presence of these episodes is often not apparent at the time to most inhabitants. A characteristic feature of these episodes is the immediate rise in mortality occurring on the same day as the peaks of pollution. A second characteristic is their frequent influence on death rates in the 45–64 year age group as well as in those over 65. These characteristics suggest that if these excess deaths are related to air pollution (as we believe to be highly likely) the mechanism is probably protean and pervasive affecting the course of a variety of different diseases through a basic physiologic effect. The relationship of morbidity in a normal urban population to air pollution is also demonstrated by time series analysis of correlation coefficients. The relation of two symptoms (cough and eye irritation) in a group of 1090 persons observed for three years is compared with two measures of air pollution (SO₂ and particulate density) to which they were exposed. Time lags of up to 28 days were introduced and a constant relationship between air pollution levels and those symptoms was demonstrated.     

State-Space Modeling of the Relationship between Air Quality and Mortality

ABSTRACT

A portion of a population is assumed to be at risk, with the mortality hazard varying with atmospheric conditions including total suspended particulates (TSP). This at-risk population is not observed and the hazard function is unknown; we wish to estimate these from mortality count and atmospheric variables. Consideration of population dynamics leads to a state-space representation, allowing the Kalman Filter (KF) to be used for estimation. A harvesting effect is thus implied; high mortality is followed by lower mortality until the population is replenished by new arrivals.

The model is applied to daily data for Philadelphia, PA, 1973-1990. The estimated hazard function rises with the level of TSP and at extremes of temperature and also reflects a positive interaction between TSP and temperature. The estimated at-risk population averages about 480 and varies seasonally. We find that lags of TSP are statistically significant, but the presence of negative coefficients suggests their role may be partially statistical rather than biological. In the population dynamics framework, the natural metric for health damage from air pollution is its impact on life expectancy. The range of hazard rates over the sample period is 0.07 to 0.085, corresponding to life expectancies of 14.3 and 11.8 days, respectively.     

Modeling Traffic-Related Air Pollution in Street Canyons of Beijing

ABSTRACT

It is important to develop a general model to accurately simulate the air pollution in urban street areas. In this paper, the Operational Street Pollution Model (OSPM) initially developed in Denmark is tested with measured data from a relatively wide and open street in Beijing. Major factors influencing the dispersion, such as emission factors, stationary source emissions, and solar radiation, are analyzed. Results show that the model can reflect the basic dispersion pattern in the street but gives systematically higher concentrations. After modifications to estimate street-level wind speed in the model, performance is obviously improved.     

A Movable Laboratory for Controlled Clinical Studies of Air Pollution Exposure

The cause of the Yokkaichi asthma episode (1960-1969) has been analyzed. It Is concluded that the respiratory diseases were due not to sulfur dioxide but to concentrated sulfuric acid mists emitted from stacks of calciners of a titanium oxide manufacturing plant located windward of the residential area.     

Spatial Modeling for Air Pollution Monitoring Network Design: Example of Residential Woodsmoke

Abstract

The purpose of this paper is to demonstrate how to develop an air pollution monitoring network to characterize small-area spatial contrasts in ambient air pollution concentrations. Using residential woodburning emissions as our case study, this paper reports on the first three stages of a four-stage protocol to measure, estimate, and validate ambient residential woodsmoke emissions in Vancouver, British Columbia. The first step is to develop an initial winter nighttime woodsmoke emissions surface using inverse-distance weighting of emissions information from consumer woodburning surveys and property assessment data. Second, fireplace density and a compound topo-graphic index based on hydrological flow regimes are used to enhance the emissions surface. Third, the spatial variation of the surface is used in a location-allocation algorithm to design a network of samplers for the woodsmoke tracer compound levoglucosan and fine particulate matter. Measurements at these network sites are then used in the fourth stage of the protocol (not presented here): a mobile sampling campaign aimed at developing a high-resolution surface of woodsmoke concentrations for exposure assignment in health effects studies. Overall the results show that relatively simple data inputs and spatial analysis can be effective in capturing the spatial variability of ambient air pollution emissions and concentrations.     

Measurement of Photochemical Air Pollution with a Sensitive Monitoring Plant

With newly developed monitoring techniques, investigators used the sensitive tobacco variety, Bel-W3, as a monitor for photochemical air pollution at different locations within 75 miles of Cincinnati, Ohio. Observations during the summers of 1966 through 1968 revealed almost daily injury to monitoring plants, fairly uniform total seasonal injury at all locations, and a marked variation in both oxidant and injury from one season to another within the city. The relationship between oxidant level and injury was not found to be consistent.     

Estimating the Effect of Being Indoors on Total Personal Exposure to Outdoor Air Pollution

A personal air quality model (PAQM) has been developed to estimate the effect of being indoors on total personal exposure to outdoor-generated air pollution. Designed to improve air toxics risk assessment, PAQM accounts for individual hourly activity patterns, indoor-outdoor differences, physical exercise level, and geographic location for up to 56 different population groups. Unique hourly activity profiles are specified for each population group; group members are assigned each hour to one of up to 10 different indoor and outdoor microenvironments. To illustrate PAQM use, we apply it to two example cases: a long-term example representative of situations where pollutant health impact is related to integrated exposure (as in the case of potentially carcinogenic air toxics) and a short-term example representative of situations where health impact is related to acute exposure to peak concentrations (as with ozone).

Case study results illustrate that personal exposure, and thus health risk, attributable to outdoor-generated air pollution is sensitive to indoor-outdoor differences and population mobility. Where health impact is related to long-term integrated exposure (e.g., air toxics), exposure and subsequent risk are likely to be lower than that estimated by previous modeling techniques which do not account for such effects.     

The Crisis in Air Pollution Manpower Development

There is a crisis in air pollution manpower development within the United States today. This is the conclusion drawn from a series of three studies recently conducted by the National Air Pollution Manpower Development Advisory Committee. These studies, designed to define the essential components of a total manpower development program and to evaluate current efforts on specific segments of such a program, showed that the existing Federal manpower program fails to address many of the problem areas needing attention, that greater effort needs to be directed to meeting the air pollution educational requirements of State highway departments and planning agencies, and that the quality of most graduate level university programs in air pollution control is on the decline because of the withdrawal of Federal financial support.     

A Preliminary Study of Modeling the Air Pollution Effects fromTraffic Engineering Alternatives

Three separate mathematical models were combined to calculate the changes in carbon monoxide (CO) concentrations that might result from traffic engineering changes. The three models used were: (1) The Dynamic Highway Transportation model (DHTM) which relates traffic flow patterns to physical parameters and traffic signal characteristics of a network; (2) an emission model that predicts CO emissions from traffic flow parameters such as number of stops, idling time, etc; and (3) the APRAC-1A urban diffusion model which calculates CO concentrations from source distributions and meteorological factors. The composite model was applied to traffic in downtown Chicago for a specific set of meteorological conditions. Results are compared for two traffic signal control schemes. In those blocks where concentrations were highest, the model indicates a 20% reduction is possible through improved traffic signal controls. The model should be useful for testing other traffic control measures.     

The Use of Aircraft in Air Pollution Research

An instrumented single engine aircraft is being used to characterize plumes from large emission sources. By cross sectioning these plumes at a number of points we hope to describe the character and concentration of pollution as it ages and travels downwind under known meteorological conditions. This paper reports experimental work thus far completed on aerosol measurement. Flights have been made through plumes from a large coal-fired power station, forest fires, and a gravel plant. Particle counts in various size classifications were made with a Bausch and Lomb 40–1 Dust Counter which was carried in the aircraft. Selected meteorological parameters were recorded. Results show the aerosol concentrations in various size classifications that are found at successive distances downwind from these sources for distances up to 32 miles. Background aerosol concentrations are also reported, and variations in these can be noted with respect to time, location and altitude.     

The Influence of Annual Meteorological Variations On Regional Air Pollution Modeling: A Case Study of Allegheny County,Pennsylvania

The influence of year-to-year meteorological variations on the prediction of annual average ground-level pollutant concentrations has been examined via case studies of Allegheny County, Pa. Twenty-two stability wind roses representing different averaging intervals of from one to seven years were employed in the Air Quality Display Model to predict annual average SO₂ concentrations in two multiple source sub-basins, and from two single point sources representing industrial and utility boiler stacks. Effects of annual meteorological variations were manifested by changes in the magnitude of peak concentrations, the location of peak concentrations, and the geographic distribution of pollutants. For fixed rates of emission, the peak annual average SO₂ ground-level concentration varied by an average of up to 33% of highest values for point sources and 17% for sub-basin complexes. In both cases, there was relatively little change in the location of peak concentration, though occasional directional shifts were noted. In contrast, marked variations were noted in the geographic area exposed to annual average concentrations in excess of several selected values. To aid in regional planning, several methods were formulated which considerably reduced the uncertainty in predicting peak annual concentration for varying degrees of historical data on regional stability wind rose. These methods are-especially applicable to analysis of control strategies directed at attaining annual ambient air quality standards which nominally must never be exceeded.     

The Law of Federal Air Pollution Control

The Clean Air Act, while recognizing the primary responsibility of states and local governments for the prevention and control of air pollution, provides authority for federal initiative and action to prevent and to abate air pollution which “endangers the health or welfare of any persons.” The applicable statutory procedures and the scope of federal authority are analyzed and discussed.