Power calculations for generalized linear models in observational longitudinal studies: a simulation approach in SAS

Repeated measurements arising from longitudinal studies occur frequently in applied research. Methods to calculate power in the context of repeated measures are available for experimental settings where the covariate of interest is a discrete treatment indicator. However, no closed form expression exists to calculate power for generalized linear models with non-zero within-cluster correlation that are common in epidemiological and observational studies in which the covariate of interest varies over time and is often measured on a continuous scale, and where the researchers control for several potential confounders. We describe a Monte Carlo simulation approach conducted to calculate power, and illustrate its application in two models frequently encountered in practice, the normal linear mixed model, and the logistic regression model, both with repeated measurements and non-zero within-cluster correlation. This approach can be used to calculate the effect on power of changing various simulation conditions controlled by the researcher, such as sample size, within-cluster correlation structure, smallest meaningful difference to detect, and distributional assumptions.

Association of FEV1 in asthmatic children with personal and microenvironmental exposure to airborne particulate matter

Exposure to particulate matter (PM) air pollution has been shown to exacerbate children's asthma, but the exposure sources and temporal characteristics are still under study. Children's exposure to PM is likely to involve both combustion-related ambient PM and PM related to a child's activity in various indoor and outdoor microenvironments. Among 19 children with asthma, 9-17 years of age, we examined the relationship of temporal changes in percent predicted forced expiratory volume in 1 sec (FEV1) to personal continuous PM exposure and to 24-hr average gravimetric PM mass measured at home and central sites. Subjects were followed for 2 weeks during either the fall of 1999 or the spring of 2000, in a southern California region affected by transported air pollution. FEV(subscript)1(/subscript) was measured by subjects in the morning, afternoon, and evening. Exposure measurements included continuous PM using a passive nephelometer carried by subjects; indoor, outdoor home, and central-site 24-hr gravimetric PM2.5 (PM of aerodynamic diameter < 2.5 microm) and PM10; and central-site hourly PM10, nitrogen dioxide, and ozone. Data were analyzed with linear mixed models controlling for within-subject autocorrelation, FEV1 maneuver time, and exposure period. We found inverse associations of FEV1 with increasing PM exposure during the 24 hr before the FEV1 maneuver and with increasing multiday PM averages. Deficits in percent predicted FEV1 (95% confidence interval) for given PM interquartile ranges measured during the preceding 24-hr were as follows: 128 microg/m3 1-hr maximum personal PM, -6.0% (-10.5 to -1.4); 30 microg/m3 24-hr average personal PM, -5.9% (-10.8 to -1.0); 6.7 microg/m3 indoor home PM2.5, -1.6% (-2.8 to -0.4); 16 microg/m3 indoor home PM10, -2.1% (-3.7 to -0.4); 7.1 microg/m3 outdoor home PM2.5, -1.1% (-2.4 to 0.1); and 7.5 microg/m3 central-site PM2.5, -0.7% (-1.9 to 0.4). Stronger associations were found for multiday moving averages of PM for both personal and stationary-site PM. Stronger associations with personal PM were found in boys allergic to indoor allergens. FEV1 was weakly associated with NO2 but not with O3. Results suggest mixed respiratory effects of PM in asthmatic children from both ambient background exposures and personal exposures in various microenvironments.

Conception and pregnancy during the Persian Gulf War: the risk to women veterans

PURPOSE

To enumerate Gulf-War (GW) exposed conceptions and to compare reproductive outcomes of GW-exposed pregnancies with postwar conceptions of women Gulf War veterans (GWV) and women nondeployed veterans (NDV).

METHODS

Deployment data and inpatient records from 153 military hospitals identified servicewomen who were pregnant between August 1990 and May 1992 and belonged to military units that were deployed to the Gulf War. Postal surveys were used in 1997 and 1998 to elicit reproductive history; responses were validated against military hospitalization records. Reproductive outcomes of GW-exposed pregnancies were compared with postwar conceptions of women GWVs and NDVs.

RESULTS

3285 women had a pregnancy-related admission; of these, 1558 completed the questionnaire. Self-reported reproductive outcomes and dates, gestational data, and individual deployment dates identified 415 Gulf War-exposed pregnancies, 298 GWV postwar conceptions, and 427 NDV conceptions. Compared with NDV conceptions, adverse reproductive outcomes were similar among GW-exposed pregnancies. However, spontaneous abortions [NDV: 9.1%, GWV postwar: 22.8%, adjusted odds ratio (OR)=2.92, 95% confidence interval (CI): 1.9, 4.6] and ectopic pregnancies (NDV: 1.4%, GWV postwar: 10.7%, adjusted OR=7.70, 95% CI, 3.0, 20) were elevated for GWV postwar conceptions.

CONCLUSION

GW-exposed conceptions and nondeployed conceptions had similar outcomes. However, GWV postwar conceptions were at increased risk for ectopic pregnancies and spontaneous abortions.

Self-reported symptoms and medical conditions among 11,868 Gulf War-era veterans: the Seabee Health Study

US Navy Seabees have been among the most symptomatic Gulf War veterans. Beginning in May 1997, the authors mailed Gulf War-era Seabees a health survey in serial mailings. As of July 1, 1999, 68.6% of 17,559 Seabees contacted had returned the questionnaire. Compared with other Seabees, Gulf War Seabees reported poorer general health, a higher prevalence of all 33 medical problems assessed, more cognition difficulties, and a higher prevalence of four physician-diagnosed multisymptom conditions: chronic fatigue syndrome, posttraumatic stress disorder, multiple chemical sensitivity, and irritable bowel syndrome. Because the four multisymptom conditions were highly associated with one another, the authors aggregated them into a working case definition of Gulf War illness. Among the 3,831 (22% cases) Gulf War Seabee participants, multivariable modeling revealed that female, Reserve, and enlisted personnel and participants belonging to either of two particular Seabee units were most likely to meet the case definition. Twelve of 34 self-reported Gulf War exposures were mildly associated with meeting the definition of Gulf War illness, with exposure to fumes from munitions having the highest odds ratio (odds ratio = 1.9, 95% confidence interval: 1.5, 2.4). While these data do not implicate a specific etiologic exposure, they demonstrate a strong association and a high prevalence of self-reported multisymptom conditions in a large group of symptomatic Gulf War veterans.