Global sensitivity analysis of randomized trials with nonmonotone missing binary outcomes: Application to studies of substance use disorders.

In this paper, we present a method for conducting global sensitivity analysis of randomized trials in which binary outcomes are scheduled to be collected on participants at prespecified points in time after randomization and these outcomes may be missing in a nonmonotone fashion. We introduce a class of missing data assumptions, indexed by sensitivity parameters, which are anchored around the missing not at random assumption introduced by Robins (Statistics in Medicine, 1997). For each assumption in the class, we establish that the joint distribution of the outcomes is identifiable from the distribution of the observed data. Our estimation procedure uses the plug-in principle, where the distribution of the observed data is estimated using random forests. We establish n$\sqrt {n}$ asymptotic properties for our estimation procedure. We illustrate our methodology in the context of a randomized trial designed to evaluate a new approach to reducing substance use, assessed by testing urine samples twice weekly, among patients entering outpatient addiction treatment. We evaluate the finite sample properties of our method in a realistic simulation study. Our methods have been implemented in an R package entitled slabm.

Global identifiability of latent class models with applications to diagnostic test accuracy studies: A Gröbner basis approach.

Identifiability of statistical models is a fundamental regularity condition that is required for valid statistical inference. Investigation of model identifiability is mathematically challenging for complex models such as latent class models. Jones et al. used Goodman's technique to investigate the identifiability of latent class models with applications to diagnostic tests in the absence of a gold standard test. The tool they used was based on examining the singularity of the Jacobian or the Fisher information matrix, in order to obtain insights into local identifiability (ie, there exists a neighborhood of a parameter such that no other parameter in the neighborhood leads to the same probability distribution as the parameter). In this paper, we investigate a stronger condition: global identifiability (ie, no two parameters in the parameter space give rise to the same probability distribution), by introducing a powerful mathematical tool from computational algebra: the Gröbner basis. With several existing well-known examples, we argue that the Gröbner basis method is easy to implement and powerful to study global identifiability of latent class models, and is an attractive alternative to the information matrix analysis by Rothenberg and the Jacobian analysis by Goodman and Jones et al.

Global sensitivity analysis of clinical trials with missing patient-reported outcomes.

Randomized trials with patient-reported outcomes are commonly plagued by missing data. The analysis of such trials relies on untestable assumptions about the missing data mechanism. To address this issue, it has been recommended that the sensitivity of the trial results to assumptions should be a mandatory reporting requirement. In this paper, we discuss a recently developed methodology (Scharfstein et al., Biometrics, 2018) for conducting sensitivity analysis of randomized trials in which outcomes are scheduled to be measured at fixed points in time after randomization and some subjects prematurely withdraw from study participation. The methodology is explicated in the context of a placebo-controlled randomized trial designed to evaluate a treatment for bipolar disorder. We present a comprehensive data analysis and a simulation study to evaluate the performance of the method. A software package entitled SAMON (R and SAS versions) that implements our methods is available at www.missingdatamatters.org .

Global sensitivity analysis for repeated measures studies with informative drop-out: A semi-parametric approach.

In practice, both testable and untestable assumptions are generally required to draw inference about the mean outcome measured at the final scheduled visit in a repeated measures study with drop-out. Scharfstein et al. (2014) proposed a sensitivity analysis methodology to determine the robustness of conclusions within a class of untestable assumptions. In their approach, the untestable and testable assumptions were guaranteed to be compatible; their testable assumptions were based on a fully parametric model for the distribution of the observable data. While convenient, these parametric assumptions have proven especially restrictive in empirical research. Here, we relax their distributional assumptions and provide a more flexible, semi-parametric approach. We illustrate our proposal in the context of a randomized trial for evaluating a treatment of schizoaffective disorder.

Antihypertensive medications and risk of death and hospitalizations in US hemodialysis patients: Evidence from a cohort study to inform hypertension treatment practices.

Antihypertensive medications are commonly prescribed to hemodialysis patients but the optimal regimens to prevent morbidity and mortality are unknown. The goal of our study was to compare the association of routinely prescribed antihypertensive regimens with outcomes in US hemodialysis patients.We used 2 datasets for our analysis. Our primary cohort (US Renal Data System [USRDS]) included adult patients initiating in-center hemodialysis from July 1, 2006 to June 30, 2008 (n = 33,005) with follow-up through December 31, 2009. Our secondary cohort included adult patients from Dialysis Clinic, Inc. (DCI), a national not-for-profit dialysis provider, initiating in-center hemodialysis from January 1, 2003 to June 30, 2008 (n = 11,291) with follow-up through December 31, 2008. We linked the USRDS cohort with Medicare part D prescriptions-fill data and the DCI cohort with USRDS data. Unique aspect of USRDS cohort was pharmacy prescription-fill data and for DCI cohort was detailed clinical data, including blood pressure, weight, and ultrafiltration. We classified prescribed antihypertensives into the following mutually exclusive regimens: ß-blockers, renin-angiotensin system blocking drugs-containing regimens without a ß-blocker (RAS), ß-blocker + RAS, and others. We used marginal structural models accounting for time-updated comorbidities to quantify each regimen's association with mortality (both cohorts) and cardiovascular hospitalization (DCI-Medicare Subcohort).In the USRDS and DCI cohorts there were 9655 (29%) and 3200 (28%) deaths, respectively. In both cohorts, RAS compared to ß-blockers regimens were associated with lower risk of death; (hazard ratio [HR]) (95% confidence interval [CI]) for all-cause mortality, (0.90 [0.82-0.97] in USRDS and 0.87 [0.76-0.98] in DCI) and cardiovascular mortality (0.84 [0.75-0.95] in USRDS and 0.88 [0.71-1.07] in DCI). There was no association between antihypertensive regimens and the risk of cardiovascular hospitalizations.In hemodialysis patients undergoing routine care, renin-angiotensin system blocking drugs-containing regimens were associated with a lower risk of death compared with ß-blockers-containing regimens but there was no association with cardiovascular hospitalizations. Pragmatic clinical trials are needed to specifically examine the effectiveness of these commonly used antihypertensive regimens in dialysis patients.

Effect of intravenous iron use on hospitalizations in patients undergoing hemodialysis: a comparative effectiveness analysis from the DEcIDE-ESRD study.

BACKGROUND: Intravenous iron use in hemodialysis patients has greatly increased over the last decade, despite limited studies on the safety of iron. METHODS: We studied the association of receipt of intravenous iron with hospitalizations in an incident cohort of hemodialysis patients. We examined 9544 patients from Dialysis Clinic, Inc. (DCI). We ascertained intravenous iron use from DCI electronic medical record and USRDS data files, and hospitalizations through Medicare claims. We examined the association between iron exposure accumulated over 1-, 3- or 6-month time windows and incident hospitalizations in the follow-up period using marginal structural models accounting for time-dependent confounders. We performed sensitivity analyses including recurrent events models for multiple hospitalizations and models for combined outcome of hospitalization and death. RESULTS: There were 22 347 hospitalizations during a median follow-up of 23 months. Higher cumulative dose of intravenous iron was not associated with all-cause, cardiovascular or infectious hospitalizations [HR 0.97 (95% CI: 0.77-1.22) for all-cause hospitalizations comparing >2100 mg versus 0-900 mg of iron over 6 months]. Findings were similar in models examining the risk of hospitalizations in 1- and 3-month windows [HR 0.88 (95% CI: 0.79-0.99) and HR 0.88 (95% CI: 0.74-1.03), respectively] or the risk of combined outcome of hospitalization and death in the 6-month window [HR 0.98 (95% CI: 0.78-1.23)]. CONCLUSIONS: Higher cumulative dose of intravenous iron may not be associated with increased risk of hospitalizations in hemodialysis patients. While clinical trials are needed, employing higher iron doses to reduce erythropoiesis-stimulating agents does not appear to increase morbidity in routine clinical care.

Impact of Subsidized Health Insurance Coverage on Emergency Department Utilization by Low-income Adults in Massachusetts.

BACKGROUND: In 2006, Massachusetts expanded insurance coverage to many low-income individuals. OBJECTIVES: This study aimed to estimate the change in emergency department (ED) utilization per individual among a cohort who qualified for subsidized health insurance following the Massachusetts health care reform. RESEARCH DESIGN: We obtained Massachusetts public health insurance enrollment data for the fiscal years 2004-2008 and identified 353,515 adults who enrolled in Commonwealth Care, a program that subsidizes insurance for low-income adults. We merged the enrollment data with statewide ED visit claims and created a longitudinal file that indicated each enrollee's ED visits and insurance status each month during the preenrollment and postenrollment periods. MEASURES: We estimated the ratio in an individual's odds of an ED visit during the postperiod versus preperiod by conditional logistic regression. RESULTS: Among the 112,146 CommCare enrollees who made at least 1 ED visit during the study period, an individual's odds of an ED visit decreased 4% [odds ratio (OR)=0.96; 95% confidence interval (CI), 0.94, 0.98] postenrollment. However, it varied significantly depending on preenrollment insurance status. A person's odds of an ED visit was 12% higher in the postperiod among enrollees not publicly insured prior (OR=1.12; 95% CI, 1.10, 1.25), but was 18% lower among enrollees who transitioned from the Health Safety Net, a program that pays for limited services for low-income individuals (OR=0.82; 95% CI, 0.78, 0.85). CONCLUSIONS: Expanding subsidized health insurance did not uniformly change ED utilization for all newly insured low-income adults in Massachusetts.

Intravenous iron exposure and mortality in patients on hemodialysis.

BACKGROUND AND OBJECTIVES: Clinical trials assessing effects of larger cumulative iron exposure with outcomes are lacking, and observational studies have been limited by assessment of short-term exposure only and/or failure to assess cause-specific mortality. The associations between short- and long-term iron exposure on all-cause and cause-specific mortality were examined. DESIGN, SETTING, PARTICIPANTS, & MEASUREMENTS: The study included 14,078 United States patients on dialysis initiating dialysis between 2003 and 2008. Intravenous iron dose accumulations over 1-, 3-, and 6-month rolling windows were related to all-cause, cardiovascular, and infection-related mortality in Cox proportional hazards models that used marginal structural modeling to control for time-dependent confounding. RESULTS: Patients in the 1-month model cohort (n=14,078) were followed a median of 19 months, during which there were 27.6% all-cause deaths, 13.5% cardiovascular deaths, and 3% infection-related deaths. A reduced risk of all-cause mortality with receipt of >150-350 (hazard ratio, 0.78; 95% confidence interval, 0.64 to 0.95) or >350 mg (hazard ratio, 0.79; 95% confidence interval, 0.62 to 0.99) intravenous iron compared with >0-150 mg over 1 month was observed. There was no relation of 1-month intravenous iron dose with cardiovascular or infection-related mortality and no relation of 3- or 6-month cumulative intravenous iron dose with all-cause or cardiovascular mortality. There was a nonstatistically significant increase in infection-related mortality with receipt of >1050 mg intravenous iron in 3 months (hazard ratio, 1.69; 95% confidence interval, 0.87 to 3.28) and >2100 mg in 6 months (hazard ratio, 1.59; 95% confidence interval, 0.73 to 3.46). CONCLUSIONS: Among patients on incident dialysis, receipt of ≤ 1050 mg intravenous iron in 3 months or 2100 mg in 6 months was not associated with all-cause, cardiovascular, or infection-related mortality. However, nonstatistically significant findings suggested the possibility of infection-related mortality with receipt of >1050 mg in 3 months or >2100 mg in 6 months. Randomized clinical trials are needed to assess the safety of exposure to greater cumulative intravenous iron doses.

Predialysis systolic BP variability and outcomes in hemodialysis patients.

BP variability (BPV) is an important predictor of outcomes in the general population, but its association with clinical outcomes in hemodialysis patients is not clear. We identified 11,291 patients starting dialysis in 2003-2008 and followed them through December 31, 2008 (median=22 months). Predialysis systolic BPV was assessed over monthly intervals. Outcomes included factors associated with BPV, mortality (all-cause and cardiovascular), and first cardiovascular event (cardiovascular death or hospitalization). Patients' mean age was 62 years, 55% of patients were men, and 58% of patients were white. Modifiable factors associated with higher BPV included obesity, higher calcium-phosphate product levels, and lower hemoglobin concentration; factors associated with lower BPV included greater fluid removal, achievement of prescribed dry weight during dialysis, higher hemoglobin concentration, and antihypertensive regimens without ß-blockers or renin-angiotensin system blocking agents. In total, 3200 deaths occurred, including 1592 cardiovascular deaths. After adjustment for demographics, comorbidities, and clinical factors, higher predialysis BPV was associated with increased risk of all-cause mortality (hazard ratio [HR], 1.18; 95% confidence interval [95% CI] per 1 SD increase in BPV, 1.13 to 1.22), cardiovascular mortality (HR, 1.18; 95% CI, 1.12 to 1.24), and first cardiovascular event (HR, 1.11; 95% CI, 1.07 to 1.15). Results were similar when BPV was categorized in tertiles and patients were stratified by baseline systolic BP. In summary, predialysis systolic BPV is an important, potentially modifiable risk factor for death and cardiovascular outcomes in incident hemodialysis patients. Studies of BP management in dialysis patients should focus on both absolute BP and BPV.

Canadian Forest Fires and the Effects of Long-Range Transboundary Air Pollution on Hospitalizations among the Elderly.

In July 2002, lightning strikes ignited over 250 fires in Quebec, Canada, destroying over one million hectares of forest. The smoke plume generated from the fires had a major impact on air quality across the east coast of the U.S. Using data from the Medicare National Claims History File and the U.S. Environmental Protection Agency (EPA) National air pollution monitoring network, we evaluated the health impact of smoke exposure on 5.9 million elderly people (ages 65+) in the Medicare population in 81 counties in 11 northeastern and Mid-Atlantic States of the US. We estimated differences in the exposure to ambient PM2.5-airborne particulate matter with aerodynamic diameter of ≤2.5 µm-concentrations and hospitalizations for cardiovascular, pulmonary and injury outcomes, before and during the smoke episode. We found that there was an associated 49.6% (95% confidence interval (CI), 29.8, 72.3) and 64.9% (95% CI, 44.3-88.5) increase rate of hospitalization for respiratory and cardiovascular diagnoses, respectively, when the smoke plume was present compared to before the smoke plume had arrived. Our study suggests that rapid increases in PM2.5 concentrations resulting from wildfire smoke can impact the health of elderly populations thousands of kilometers removed from the fires.

Patterns in blood pressure medication use in US incident dialysis patients over the first 6 months.

BACKGROUND: Several observational studies have evaluated the effect of a single exposure window with blood pressure (BP) medications on outcomes in incident dialysis patients, but whether BP medication prescription patterns remain stable or a single exposure window design is adequate to evaluate effect on outcomes is unclear. METHODS: We described patterns of BP medication prescription over 6 months after dialysis initiation in hemodialysis and peritoneal dialysis patients, stratified by cardiovascular comorbidity, diabetes, and other patient characteristics. The cohort included 13,072 adult patients (12,159 hemodialysis, 913 peritoneal dialysis) who initiated dialysis in Dialysis Clinic, Inc., facilities January 1, 2003-June 30, 2008, and remained on the original modality for at least 6 months. We evaluated monthly patterns in BP medication prescription over 6 months and at 12 and 24 months after initiation. RESULTS: Prescription patterns varied by dialysis modality over the first 6 months; substantial proportions of patients with prescriptions for beta-blockers, renin angiotensin system agents, and dihydropyridine calcium channel blockers in month 6 no longer had prescriptions for these medications by month 24. Prescription of specific medication classes varied by comorbidity, race/ethnicity, and age, but little by sex. The mean number of medications was 2.5 at month 6 in hemodialysis and peritoneal dialysis cohorts. CONCLUSIONS: This study evaluates BP medication patterns in both hemodialysis and peritoneal dialysis patients over the first 6 months of dialysis. Our findings highlight the challenges of assessing comparative effectiveness of a single BP medication class in dialysis patients. Longitudinal designs should be used to account for changes in BP medication management over time, and designs that incorporate common combinations should be considered.

The impact of the demographic transition on dengue in Thailand: insights from a statistical analysis and mathematical modeling.

BACKGROUND: An increase in the average age of dengue hemorrhagic fever (DHF) cases has been reported in Thailand. The cause of this increase is not known. Possible explanations include a reduction in transmission due to declining mosquito populations, declining contact between human and mosquito, and changes in reporting. We propose that a demographic shift toward lower birth and death rates has reduced dengue transmission and lengthened the interval between large epidemics. METHODS AND FINDINGS: Using data from each of the 72 provinces of Thailand, we looked for associations between force of infection (a measure of hazard, defined as the rate per capita at which susceptible individuals become infected) and demographic and climactic variables. We estimated the force of infection from the age distribution of cases from 1985 to 2005. We find that the force of infection has declined by 2% each year since a peak in the late 1970s and early 1980s. Contrary to recent findings suggesting that the incidence of DHF has increased in Thailand, we find a small but statistically significant decline in DHF incidence since 1985 in a majority of provinces. The strongest predictor of the change in force of infection and the mean force of infection is the median age of the population. Using mathematical simulations of dengue transmission we show that a reduced birth rate and a shift in the population's age structure can explain the shift in the age distribution of cases, reduction of the force of infection, and increase in the periodicity of multiannual oscillations of DHF incidence in the absence of other changes. CONCLUSIONS: Lower birth and death rates decrease the flow of susceptible individuals into the population and increase the longevity of immune individuals. The increase in the proportion of the population that is immune increases the likelihood that an infectious mosquito will feed on an immune individual, reducing the force of infection. Though the force of infection has decreased by half, we find that the critical vaccination fraction has not changed significantly, declining from an average of 85% to 80%. Clinical guidelines should consider the impact of continued increases in the age of dengue cases in Thailand. Countries in the region lagging behind Thailand in the demographic transition may experience the same increase as their population ages. The impact of demographic changes on the force of infection has been hypothesized for other diseases, but, to our knowledge, this is the first observation of this phenomenon. Please see later in the article for the Editors' Summary.

Emergency admissions for cardiovascular and respiratory diseases and the chemical composition of fine particle air pollution.

BACKGROUND: Population-based studies have estimated health risks of short-term exposure to fine particles using mass of PM(2.5) (particulate matter or= 65 years of age). RESULTS: In multiple-pollutant models that adjust for the levels of other pollutants, an interquartile range (IQR) increase in EC was associated with a 0.80% [95% posterior interval (PI), 0.34-1.27%] increase in risk of same-day cardiovascular admissions, and an IQR increase in OCM was associated with a 1.01% (95% PI, 0.04-1.98%) increase in risk of respiratory admissions on the same day. Other components were not associated with cardiovascular or respiratory hospital admissions in multiple-pollutant models. CONCLUSIONS: Ambient levels of EC and OCM, which are generated primarily from vehicle emissions, diesel, and wood burning, were associated with the largest risks of emergency hospitalization across the major chemical constituents of PM(2.5).

Mortality in the Medicare population and chronic exposure to fine particulate air pollution in urban centers (2000-2005).

BACKGROUND: Prospective cohort studies constitute the major source of evidence about the mortality effects of chronic exposure to particulate air pollution. Additional studies are needed to provide evidence on the health effects of chronic exposure to particulate matter < or = 2.5 microm in aerodynamic diameter (PM(2.5)) because few studies have been carried out and the cohorts have not been representative. OBJECTIVES: This study was designed to estimate the relative risk of death associated with long-term exposure to PM(2.5) by region and age groups in a U.S. population of elderly, for the period 2000-2005. METHODS: By linking PM(2.5) monitoring data to the Medicare billing claims by ZIP code of residence of the enrollees, we have developed a new retrospective cohort study, the Medicare Cohort Air Pollution Study. The study population comprises 13.2 million participants living in 4,568 ZIP codes having centroids within 6 miles of a PM(2.5) monitor. We estimated relative risks adjusted by socioeconomic status and smoking by fitting log-linear regression models. RESULTS: In the eastern and central regions, a 10-microg/m(3) increase in 6-year average of PM(2.5) is associated with 6.8% [95% confidence interval (CI), 4.9-8.7%] and 13.2% (95% CI, 9.5-16.9) increases in mortality, respectively. We found no evidence of an association in the western region or for persons > or = 85 years of age. CONCLUSIONS: We established a cohort of Medicare participants for investigating air pollution and mortality on longer-term time frames. Chronic exposure to PM(2.5) was associated with mortality in the eastern and central regions, but not in the western United States.

Coarse particulate matter air pollution and hospital admissions for cardiovascular and respiratory diseases among Medicare patients.

CONTEXT: Health risks of fine particulate matter of 2.5 microm or less in aerodynamic diameter (PM2.5) have been studied extensively over the last decade. Evidence concerning the health risks of the coarse fraction of greater than 2.5 microm and 10 microm or less in aerodynamic diameter (PM10-2.5) is limited. OBJECTIVE: To estimate risk of hospital admissions for cardiovascular and respiratory diseases associated with PM10-2.5 exposure, controlling for PM2.5. DESIGN, SETTING, AND PARTICIPANTS: Using a database assembled for 108 US counties with daily cardiovascular and respiratory disease admission rates, temperature and dew-point temperature, and PM10-2.5 and PM2.5 concentrations were calculated with monitoring data as an exposure surrogate from January 1, 1999, through December 31, 2005. Admission rates were constructed from the Medicare National Claims History Files, for a study population of approximately 12 million Medicare enrollees living on average 9 miles (14.4 km) from collocated pairs of PM10 and PM2.5 monitors. MAIN OUTCOME MEASURES: Daily counts of county-wide emergency hospital admissions for primary diagnoses of cardiovascular or respiratory disease. RESULTS: There were 3.7 million cardiovascular disease and 1.4 million respiratory disease admissions. A 10-microg/m3 increase in PM10-2.5 was associated with a 0.36% (95% posterior interval [PI], 0.05% to 0.68%) increase in cardiovascular disease admissions on the same day. However, when adjusted for PM2.5, the association was no longer statistically significant (0.25%; 95% PI, -0.11% to 0.60%). A 10-microg/m3 increase in PM10-2.5 was associated with a nonstatistically significant unadjusted 0.33% (95% PI, -0.21% to 0.86%) increase in respiratory disease admissions and with a 0.26% (95% PI, -0.32% to 0.84%) increase in respiratory disease admissions when adjusted for PM2.5. The unadjusted associations of PM2.5 with cardiovascular and respiratory disease admissions were 0.71% (95% PI, 0.45%-0.96%) for same-day exposure and 0.44% (95% PI, 0.06% to 0.82%) for exposure 2 days before hospital admission. CONCLUSION: After adjustment for PM2.5, there were no statistically significant associations between coarse particulates and hospital admissions for cardiovascular and respiratory diseases.

Fine particulate matter and mortality: a comparison of the six cities and American Cancer Society cohorts with a medicare cohort.

BACKGROUND: The American Cancer Society study and the Harvard Six Cities study are 2 landmark cohort studies for estimating the chronic effects of fine particulate air pollution (PM2.5) on mortality. Using Medicare data, we assessed the association of PM2.5 with mortality for the same locations included in these studies. METHODS: We estimated the chronic effects of PM2.5 on mortality for the period 2000-2002 using mortality data for cohorts of Medicare participants and average PM2.5 levels from monitors in the same counties included in the 2 studies. We estimated mortality risk associated with air pollution adjusting for individual-level (age and sex) and area-level covariates (education, income level, poverty, and employment). We controlled for potential confounding by cigarette smoking by including standardized mortality ratios for lung cancer and chronic obstructive pulmonary disease. RESULTS: Using the Medicare data, we estimated that a 10 microg/m increase in the yearly average PM2.5 concentration is associated with 10.9% (95% confidence interval = 9.0-12.8) and with 20.8% (14.8-27.1) increases in all-cause mortality for the American Cancer Society and Harvard Six Cities study counties, respectively. The estimates are somewhat higher than those reported by the original investigators. CONCLUSION: Although Medicare data lack information on some potential confounding factors, we estimated risks similar to those in the previously published reports, which incorporated more extensive information on individual-level confounders. We propose that the Medicare files can be used to construct on-going cohorts for tracking the risk of air pollution over time.

Fine particulate air pollution and hospital admission for cardiovascular and respiratory diseases.

CONTEXT: Evidence on the health risks associated with short-term exposure to fine particles (particulate matter < or =2.5 microm in aerodynamic diameter [PM2.5]) is limited. Results from the new national monitoring network for PM2.5 make possible systematic research on health risks at national and regional scales. OBJECTIVES: To estimate risks of cardiovascular and respiratory hospital admissions associated with short-term exposure to PM2.5 for Medicare enrollees and to explore heterogeneity of the variation of risks across regions. DESIGN, SETTING, AND PARTICIPANTS: A national database comprising daily time-series data daily for 1999 through 2002 on hospital admission rates (constructed from the Medicare National Claims History Files) for cardiovascular and respiratory outcomes and injuries, ambient PM2.5 levels, and temperature and dew-point temperature for 204 US urban counties (population >200,000) with 11.5 million Medicare enrollees (aged >65 years) living an average of 5.9 miles from a PM2.5 monitor. MAIN OUTCOME MEASURES: Daily counts of county-wide hospital admissions for primary diagnosis of cerebrovascular, peripheral, and ischemic heart diseases, heart rhythm, heart failure, chronic obstructive pulmonary disease, and respiratory infection, and injuries as a control outcome. RESULTS: There was a short-term increase in hospital admission rates associated with PM2.5 for all of the health outcomes except injuries. The largest association was for heart failure, which had a 1.28% (95% confidence interval, 0.78%-1.78%) increase in risk per 10-microg/m3 increase in same-day PM2.5. Cardiovascular risks tended to be higher in counties located in the Eastern region of the United States, which included the Northeast, the Southeast, the Midwest, and the South. CONCLUSION: Short-term exposure to PM2.5 increases the risk for hospital admission for cardiovascular and respiratory diseases.

Revised analyses of the National Morbidity, Mortality, and Air Pollution Study: mortality among residents of 90 cities.

This article presents findings from updated analyses of data from 90 U.S. cities assembled for the National Morbidity, Mortality, and Air Pollution Study (NMMAPS). The data were analyzed with a generalized additive model (GAM) using the gamfunction in S-Plus (with default convergence criteria previously used and with more stringent criteria) and with a generalized linear model (GLM) with natural cubic splines. With the original method, the estimated effect of PM(10) (particulate matter 10 microm in mass median aerodynamic diameter) on total mortality from nonexternal causes was a 0.41% increase per 10-microg/m(3) increase in PM(10); with the more stringent criteria, the estimate was 0.27%; and with GLM, the effect was 0.21%. The effect of PM(10) on respiratory and cardiovascular mortality combined was greater, but the pattern across models was similar. The findings of the updated analysis with regard to spatial heterogeneity across the 90 cities were unchanged from the original analyses.

Ozone and short-term mortality in 95 US urban communities, 1987-2000.

CONTEXT: Ozone has been associated with various adverse health effects, including increased rates of hospital admissions and exacerbation of respiratory illnesses. Although numerous time-series studies have estimated associations between day-to-day variation in ozone levels and mortality counts, results have been inconclusive. OBJECTIVE: To investigate whether short-term (daily and weekly) exposure to ambient ozone is associated with mortality in the United States. DESIGN AND SETTING: Using analytical methods and databases developed for the National Morbidity, Mortality, and Air Pollution Study, we estimated a national average relative rate of mortality associated with short-term exposure to ambient ozone for 95 large US urban communities from 1987-2000. We used distributed-lag models for estimating community-specific relative rates of mortality adjusted for time-varying confounders (particulate matter, weather, seasonality, and long-term trends) and hierarchical models for combining relative rates across communities to estimate a national average relative rate, taking into account spatial heterogeneity. MAIN OUTCOME MEASURE: Daily counts of total non-injury-related mortality and cardiovascular and respiratory mortality in 95 large US communities during a 14-year period. RESULTS: A 10-ppb increase in the previous week's ozone was associated with a 0.52% increase in daily mortality (95% posterior interval [PI], 0.27%-0.77%) and a 0.64% increase in cardiovascular and respiratory mortality (95% PI, 0.31%-0.98%). Effect estimates for aggregate ozone during the previous week were larger than for models considering only a single day's exposure. Results were robust to adjustment for particulate matter, weather, seasonality, and long-term trends. CONCLUSIONS: These results indicate a statistically significant association between short-term changes in ozone and mortality on average for 95 large US urban communities, which include about 40% of the total US population. The findings indicate that this widespread pollutant adversely affects public health.

Airborne particulate matter and mortality: timescale effects in four US cities.

While time-series studies have consistently provided evidence for an effect of particulate air pollution on mortality, uncertainty remains as to the extent of the life-shortening implied by those associations. In this paper, the authors estimate the association between air pollution and mortality using different timescales of variation in the air pollution time series to gain further insight into this question. The authors' method is based on a Fourier decomposition of air pollution time series into a set of independent exposure variables, each representing a different timescale. The authors then use this set of variables as predictors in a Poisson regression model to estimate a separate relative rate of mortality for each exposure timescale. The method is applied to a database containing information on daily mortality, particulate air pollution, and weather in four US cities (Pittsburgh, Pennsylvania; Minneapolis, Minnesota; Seattle, Washington; and Chicago, Illinois) from the period 1987-1994. The authors found larger relative rates of mortality associated with particulate air pollution at longer timescale variations (14 days-2 months) than at shorter timescales (1-4 days). These analyses provide additional evidence that associations between particle indexes and mortality do not imply only an advance in the timing of death by a few days for frail individuals.