No abstract

Recent concern has centered on "sick buildings" in which there has been an unusually high percentage of health complaints by the building's occupants. Typically, these gymptoms are thought to be tied to indoor air quality characteristics, such as high levels of respirable particles or volatiles, thermal conditions, etc. In addition, recent studies have drawn connections between SBS symptoms and non-environmental variables, i.e., personal and occupational factors.We review Hedge, et al. (1995) and perform additional analyses of their data. In a study of 27 air-conditioned office buildings, they measured nine indoor environmental conditions at various locations within each building and concurrently questioned workers on sixteen SBS symptoms and a number of other personal factors. The analyses we perform are among the first to attempt to draw formal statistical connections between SBS symptoms and both personal worker characteristics and indoor air pollutants simultaneously. The analyses are based on severity scales for each symptom which include information not only on the frequency with which an indi\idual experienced a symptom, but also on how much the symptom disrupted the individual's work. Results from sixteen linear mixed effects models indicate that significant predictors are primarily personal and occupational (rather than environmental) in nature.

Our analyses indicate that the risk of CNS disease associated with Kaposi's sarcoma depends strongly on the reference or control group chosen. When compared to individuals with other non-Kaposi's sarcoma AIDS-defining diseases, Kaposi's sarcoma is associated with a lower risk of CNS disease in HIV-1 positive individuals. However, when compared to individuals with no AIDS-defining disease or with a similarly mild AIDS-defining disease such as invasive candidiasis, Kaposi's sarcoma is associated with an equivalent risk of CNS disease.

The two-sample Student's t-interval can be derived by parametric bootstrap coverage calibration of the naive, plug-in interval that fails to account for the estimation of the common variance parameter. The same technique results in an essentially exact con dence interval for the ratio of two means based on independent samples from gamma distributions. The gamma problem is more di cult computationally because an exact pivot is not available. However, the computational burden of bootstrap calibration can be reduced to a routine level using the fact that the ratio of sample means is proportional to an exact Fvariate that is independent of the conditional maximum likelihood estimator of the shape parameter. The density of the conditional estimator can be approximated with extreme accuracy using the p formula, and a rejection sampler can be used to simulate from this approximate density. This greatly simpli es bootstrap calibration in this context. An alternative interval is obtained by inverting an exact test described by Jensen (1986). Simulation studies suggest that the two intervals have similar performance in terms of length and coverage. In practice, we recommend the bootstrap calibration method because it is much simpler to implement and is more versatile.