Improved Semiparametric Time Series Models of Air Pollution and Mortality

Dominici, Francesca; McDermott, Aidian; Hastie, Trevor

doi:10.1198/016214504000000656

Cited by 155 publications

(142 citation statements)

References 29 publications

Supporting

Mentioning

140

Contrasting

Unclassified

Order By: Relevance

“…In our analyses, GAM with default convergence criteria overestimated the health effects and underestimated their standard errors. These biases can be removed by incorporating stricter convergence criteria in the backfitting and local scoring algorithms of GAM along with implementing recent methods to more accurately calculate standard errors from generalized additive models (Dominici et al, 2004).…”

Section: Discussionmentioning

confidence: 99%

See 1 more Smart Citation

Association between particulate matter and emergency room visits, hospital admissions and mortality in Spokane, Washington

Slaughter

Kim

Sheppard

et al. 2004

J Expo Sci Environ Epidemiol

111

View full text Add to dashboard Cite

There is conflicting evidence regarding the association between different size fractions of particulate matter (PM) and cardiac and respiratory morbidity and mortality. We investigated the short-term associations of four size fractions of particulate matter (PM 1 , PM 2.5 , PM 10 , and PM 10-2.5 ) and carbon monoxide with hospital admissions and emergency room (ER) visits for respiratory and cardiac conditions and mortality in Spokane, Washington. We used a log-linear generalized linear model to compare daily averages of PM and carbon monoxide with daily counts of the morbidity and mortality outcomes from January 1995 to June 2001. We examined pollution lags ranging from 0 to 3 days and compared our results to a similar log-linear generalized additive model. Effect estimates tended to be smaller and have larger standard errors for the generalized linear model. Overall, we saw no association with respiratory ER visits and any size fraction of PM. However, there was a suggestion of greater respiratory effect from fine PM when compared to coarse fraction. Carbon monoxide was associated with both all respiratory ER visits and visits for asthma at the 3-day lag. We feel that carbon monoxide may be serving as a marker for combustion-derived pollutants, which is one large component of the diverse air pollutant mixture. We also found no association with any size fraction of PM or CO with cardiac hospital admissions or mortality at the 0-to 3-day lag. We found no consistent associations between any size fraction of PM and cardiac or respiratory ER visits or hospital admissions.

show abstract

Section: Discussionmentioning

confidence: 99%

“…Computation in the GAM model relies on the backfitting algorithm in addition to iteratively reweighted least squares, while GLM only uses iteratively reweighted least squares. We used standard S-plus convergence criteria for the GAM; these have been shown to be too lax in this setting (Dominici et al, 2002(Dominici et al, , 2004.…”

Section: Methodsmentioning

confidence: 99%

Association between particulate matter and emergency room visits, hospital admissions and mortality in Spokane, Washington

Slaughter

Kim

Sheppard

et al. 2004

J Expo Sci Environ Epidemiol

111

View full text Add to dashboard Cite

show abstract

“…It suggests that one should be concerned only with avoiding the use of too few degrees of freedom when adjusting for confounders. Dominici et al 25 investigated this issue in more detail and proposed a strategy for selecting the number of degrees of freedom used to adjust for confounders.…”

Section: Resultsmentioning

confidence: 99%

“…␣ is a scaling factor that allows the amount of smoothing used for the confounder adjustments to be increased (␣ Ͻ 1) or decreased (␣ Ͼ 1), similar to the approach of Dominici et al 25 Because ␣ ϭ 1 for the generated mortality time series, if Model 7 is fit to the generated mortality time series with ␣ ϭ 1, the confounder adjustments are correctly specified. For ␣ Ͻ 1, the confounder adjustments use too few degrees of freedom, and for ␣ Ͼ 1, the confounder adjustments use too many degrees of freedom.…”

Section: Simulation Study Estimation Model For Each [mentioning

confidence: 99%

An Investigation of Distributed Lag Models in the Context of Air Pollution and Mortality Time Series Analysis

Roberts

2005

Journal of the Air & Waste Management Association

View full text Add to dashboard Cite

In particulate air pollution mortality time series studies, the particulate air pollution exposure measure used is typically the current day's or the previous day's air pollution concentration or a multi-day moving average air pollution concentration. Distributed lag models (DLMs) that allow for differential air pollution effects that are spread over multiple days are seen as an improvement over using a single-or multi-day moving average air pollution exposure measure. However, at the current time, the statistical properties of DLMs as a measure of air pollution exposure have not been investigated. In this paper, a simulation study is used to investigate the performance of DLMs as a measure of air pollution exposure in comparison with single-and multi-day moving average air pollution exposure measures under various forms for the true effect of air pollution on mortality. The simulation study shows that DLMs offer a more robust measure of the effect of air pollution on mortality and avoid the potential for a large negative bias compared with singleor multi-day moving average air pollution exposure measures. This is important information. In many U.S. cities, particulate air pollution concentrations are observed only once every six days, meaning it is often only possible to use single-day particulate air pollution exposure measures. The results from this paper will help quantify the magnitude of the negative bias that can result from using single-day exposure measures. The implications of this work for future air pollution mortality time series studies are discussed. The data used in this paper are concurrent daily time series of mortality, weather, and particulate air pollution from Cook County, IL, for the period 1987-1994. INTRODUCTIONThere have been numerous community time series studies on the effect of air pollution on mortality. [2][3][4][5][6][7][8][9][10][11][12] These studies typically fit a generalized additive model 13 or generalized linear model 14 to concurrent time series of daily mortality, air pollution, and meteorological covariates. The fitted models then are used to quantify the effect of air pollution on mortality. The measure of air pollution exposure used is typically a single-or multi-day moving average air pollution concentration. Smith et al. 10 investigated the effect of using different exposure measures for air pollution on the overall air pollution mortality effect estimate. They found that the air pollution mortality effect estimate was sensitive to the averaging period used for the exposure measure and suggested that distributed lag models that avoid the problem of selecting an averaging period for the air pollution exposure measure may be the way forward.Distributed lag models (DLMs) allow for differential air pollution effects that are spread over multiple days. In a DLM, the current day's air pollution concentration and previous days' (or lagged) air pollution concentrations are included in the model, each lag having its own coefficient. This allows but does not require th...

show abstract

“…Controlling for confounding bias and exploring effect modification in time series studies of air pollution and mortality is challenging (Samet et al, 2000a,b;Dominici et al, 2004), and requires a systematic assessment of the sensitivity of findings to model specification. For example, temperature is both a potential confounder and effect modifier of the ozonemortality association.…”

Section: Introductionmentioning

confidence: 99%

Bayesian hierarchical distributed lag models for summer ozone exposure and cardio‐respiratory mortality

2005

Self Cite

View full text Add to dashboard Cite

In this paper we develop Bayesian hierarchical distributed lag models for estimating associations between daily variations in summer ozone levels and daily variations in cardiovascular and respiratory (CVDRESP) mortality counts for 19 large U.S. cities included in the National Morbidity Mortality Air Pollution Study (NMMAPS) for the period 1987 -1994.At the first stage, we define a semi-parametric distributed lag Poisson regression model to estimate city-specific relative rates of CVDRESP associated with short-term exposure to summer ozone. At the second stage, we specify a class of distributions for the true cityspecific relative rates to estimate an overall effect by taking into account the variability within and across cities. We perform the calculations with respect to several random effects distributions (normal, t-student, and mixture of normal), thus relaxing the common assumption of a two-stage normal-normal hierarchical model. We assess the sensitivity of the results to: 1) lag structure for ozone exposure; 2) degree of adjustment for long-term trends; 3) inclusion of other pollutants in the model; 4) heat waves; 5) random effects distributions; and 6) prior hyperparameters.On average across cities, we found that a 10ppb increase in summer ozone level for every day in the previous week is associated with 1.25 percent increase in CVDRESP mortality (95% posterior regions: 0.47, 2.03). The relative rate estimates are also positive and statistically significant at lags 0, 1, and 2. We found that associations between summer ozone and CVDRESP mortality are sensitive to the confounding adjustment for P M 10 , but are robust to: 1) the adjustment for long-term trends, other gaseous pollutants (N O 2 , SO 2 , and CO); 2) the distributional assumptions at the second stage of the hierarchical model; and 3) the prior distributions on all unknown parameters.Bayesian hierarchical distributed lag models and their application to the NMMAPS data allow us estimation of an acute health effect associated with exposure to ambient air pollution in the last few days on average across several locations. The application of these methods and the systematic assessment of the sensitivity of findings to model assumptions provide important epidemiological evidence for future air quality regulations.

show abstract

Improved Semiparametric Time Series Models of Air Pollution and Mortality

Cited by 155 publications

References 29 publications

Association between particulate matter and emergency room visits, hospital admissions and mortality in Spokane, Washington

Association between particulate matter and emergency room visits, hospital admissions and mortality in Spokane, Washington

An Investigation of Distributed Lag Models in the Context of Air Pollution and Mortality Time Series Analysis

Bayesian hierarchical distributed lag models for summer ozone exposure and cardio‐respiratory mortality

Contact Info

Product

Resources

About