Sequential updating via dynamic hierarchical models: A Bayesian approach to longitudinal data analysis

Most statistical analysis, theory and practice, is concerned with static models; models with a proposed set of parameters whose values are fixed across observational units. Static models implicitly assume that the quantified relationships remain the same across the design space of the data. While this is reasonable under many circumstances this can be a dangerous assumption when dealing with sequentially ordered data. The mere passage of time always brings fresh considerations and the interrelationships among parameters, or subsets of parameters, may need to be continually revised. ^ When data are gathered sequentially dynamic interim monitoring may be useful as new subject-specific parameters are introduced with each new observational unit. Sequential imputation via dynamic hierarchical models is an efficient strategy for handling missing data and analyzing longitudinal studies. Dynamic conditional independence models offers a flexible framework that exploits the Bayesian updating scheme for capturing the evolution of both the population and individual effects over time. While static models often describe aggregate information well they often do not reflect conflicts in the information at the individual level. Dynamic models prove advantageous over static models in capturing both individual and aggregate trends. Computations for such models can be carried out via the Gibbs sampler. An application using a small sample repeated measures normally distributed growth curve data is presented. ^

Optimization of a DNSH passive sampling method to measure airborne carbonyls

Various airborne aldehydes and ketones (i.e., airborne carbonyls) present in outdoor, indoor, and personal air pose a risk to human health at present environmental concentrations. To date, there is no adequate, simple-to-use sampler for monitoring carbonyls at parts per billion concentrations in personal air. The Passive Aldehydes and Ketones Sampler (PAKS) originally developed for this purpose has been found to be unreliable in a number of relatively recent field studies. The PAKS method uses dansylhydrazine, DNSH, as the derivatization agent to produce aldehyde derivatives that are analyzed by HPLC with fluorescence detection. The reasons for the poor performance of the PAKS are not known but it is hypothesized that the chemical derivatization conditions and reaction kinetics combined with a relatively low sampling rate may play a role. This study evaluated the effect of absorption and emission wavelengths, pH of the DNSH coating solution, extraction solvent, and time post-extraction for the yield and stability of formaldehyde, acetaldehyde, and acrolein DNSH derivatives. The results suggest that the optimum conditions for the analysis of DNSHydrazones are the following. The excitation and emission wavelengths for HPLC analysis should be at 250nm and 500nm, respectively. The optimal pH of the coating solution appears to be pH 2 because it improves the formation of di-derivatized acrolein DNSHydrazones without affecting the response of the derivatives of the formaldehyde and acetaldehyde derivatives. Acetonitrile is the preferable extraction solvent while the optimal time to analyze the aldehyde derivatives is 72 hours post-extraction. ^

Characterization of the East Houston particulate matter <2.5 atmosphere; a novel method for desorbing organic constituents from airborne particulates

I have developed a novel approach to test for toxic organic substances adsorbed onto ultra fine particulate particles present in the ambient air in Northeast Houston, Texas. These particles are predominantly carbon soot with an aerodynamic diameter (AD) of <2.5 μm. If present in the ambient air, many of the organic substances will be absorbed to the surface of the particles (which act just like a charcoal air filter), and may be adducted into the respiratory system. Once imbedded into the lungs these particles may release the adsorbed toxic organic substances with serious health consequences. I used a Airmetrics portable Minivol air sampler time drawing the ambient air through collection filters samples from 6 separate sites in Northeast Houston, an area known for high ambient PM 2.5 released from chemical plants and other sources (e.g. vehicle emissions).(1) In practice, the mass of the collected particles were much less than the mass of the filters. My technique was designed to release the adsorbed organic substances on the fine carbon particles by heating the filter samples that included the PM 2.5 particles prior to identification by gas chromatography/mass spectrometry (GCMS). The results showed negligible amounts of target chemicals from the collection filters. However, the filters alone released organic substances and GCMS could not distinguish between the organic substances released from the soot particles from those released from the heated filter fabric. However, an efficacy tests of my method using two wax burning candles that released soot revealed high levels of benzene. This suggests that my method has the potential to reveal the organic substances adsorbed onto the PM 2.5 for analysis. In order to achieve this goal, I must refine the particle collection process which would be independent of the filters; the filters upon heating also release organic substances obscuring the contribution from the soot particles. To obtain pure soot particles I will have to filter more air so that the soot particles can be shaken off the filters and then analyzed by my new technique. ^

Nitrogen dioxide exposures among a selected group of residents of substandard housing in Houston, Texas

Nitrogen dioxide (NO$\sb2)$ levels in sixteen substandard houses located in Houston, Texas were examined. The classification of the houses as substandard was based on an assessment of structural integrity which would affect air exchange rates. In these homes, unvented gas space heaters were operated as the primary source of heat.^ The Ogawa passive sampling device was used to measure NO$\sb2$ concentrations over 24 to 48-hour periods during generally cold weather. A sampler was placed in the kitchen and bedroom of each house. The female head of household was asked to wear a monitor during area monitoring to assess her personal exposure. Outdoor levels of NO$\sb2$ were also measured.^ Mean (standard deviation) levels of kitchen, bedroom and personal exposures were 280 (125) ppb, 256 (155) ppb and 164 (102) ppb, respectively. Additional short-term ($<$24 hours) samples were measured in three houses. The mean level of NO$\sb2$ measured outdoors was 51 ppb over the course of the study.^ The measurements obtained with the Ogawa sampler were compared to those levels obtained using a reference method (chemiluminescence). Outdoor levels measured with the diffusion samplers were 48% higher.^ These results suggest that wintertime NO$\sb2$ levels within substandard houses using gas appliances for heating and cooking are extremely elevated. Further work is needed to investigate the prevalence of possible health effects associated with these exposures. ^

A Bayesian analysis for spatial processes with application to mapping

In geographical epidemiology, maps of disease rates and disease risk provide a spatial perspective for researching disease etiology. For rare diseases or when the population base is small, the rate and risk estimates may be unstable. Empirical Bayesian (EB) methods have been used to spatially smooth the estimates by permitting an area estimate to "borrow strength" from its neighbors. Such EB methods include the use of a Gamma model, of a James-Stein estimator, and of a conditional autoregressive (CAR) process. A fully Bayesian analysis of the CAR process is proposed. One advantage of this fully Bayesian analysis is that it can be implemented simply by using repeated sampling from the posterior densities. Use of a Markov chain Monte Carlo technique such as Gibbs sampler was not necessary. Direct resampling from the posterior densities provides exact small sample inferences instead of the approximate asymptotic analyses of maximum likelihood methods (Clayton & Kaldor, 1987). Further, the proposed CAR model provides for covariates to be included in the model. A simulation demonstrates the effect of sample size on the fully Bayesian analysis of the CAR process. The methods are applied to lip cancer data from Scotland, and the results are compared. ^

TRANSMISSION DYNAMICS OF ENTERIC BACTERIA IN DAY-CARE CENTERS, HOUSTON, TEXAS

An investigation was undertaken to evaluate the role of fomites in the transmission of diarrhea in day-care centers (DCC) and to elucidate the paths by which enteric organisms spread within this setting.^ During a nine-month period (December 1980-August 1981) extensive culturing of inanimate objects, as well as children and staff was done routinely each month and again repeated during diarrhea outbreaks. Air was sampled from the classrooms and toilets using a Single-Stage Sieve Sampler (Ross Industries, Midland, VA.). Stool samples were collected from both ill and well children and staff in the affected rooms only during outbreaks. Environmental samples were processed for Shigella, salmonella and fecal coliforms while stools were screened for miscellaneous enteropathogens.^ A total of 11 outbreaks occurred in the 5 DCC during the study period. Enteric pathogens were recovered in 7 (64%) of the outbreaks. Multiple pathogens were identified in 3 outbreaks. The most frequently identified pathogen in stools was Giardia lamblia which was recovered in 5 (45%) of the outbreaks. Ten of the 11 (91%) outbreaks occurred in children less than 12 months of age.^ Environmental microbiology studies together with epidemiologic information revealed that enteric organisms were transmitted from person-to-person. On routine sampling, fecal coliforms were most frequently isolated from tap handles and diaper change areas. Contamination with fetal coliforms was wide-spread during diarrhea outbreaks. Fecal coliforms were recovered with significantly greater frequency from hands, toys and other classroom objects during outbreaks than during non-outbreak period. Salmonella typhimurium was recovered from a table top during an outbreak of Salmonellosis. There was no association between the level of enteric microbial contamination in the toilet areas and the occurrence of outbreaks. No evidence was found to indicate that enteric organisms were spread by the airborne route via aerosols.^ Toys, other classroom objects and contaminated hands probably play a major role in the transmission of enteropathogens during day-care center outbreaks. The presence of many enteric agents in the environment undoubtedly explains the polymicrobial etiology of the day-care center associated diarrhea outbreaks. ^