STATISTICAL METHODOLOGIES FOR EXPOSURE ASSESSMENT
Funded by
National
Institute of Occupational Safety and Health (NIOSH):
Grant R01-OH03628-01A1
(2001-2004) $400,453; R01-OH03628 (2005-2008) $869,409;
Proposed Aims and Scopes
Exposure measurements obtained at the workplace is the key component for the development of appropriate strategies to monitor and improve occupational safety and health; in particular, for assessing health risks, for identifying cost-effective intervention strategies, for developing exposure-response relationships, for evaluating low-cost, easy to use, non-invasive test kits, and a host of other activities that will enhance occupational health. Furthermore, the collection of exposure data involves considerable human and financial resources. Consequently, it is very crucial that the data be used effectively. Needless to say, the use of valid statistical methodology for exposure data analysis is critical to this endeavor.
Unfortunately, in the context of exposure data analysis, many of the problems that come up are quite different from the problems addressed in the mainstream statistics literature, and available in standard software packages. Routine application of standard statistical techniques to these problems, though easy, can be inappropriate and inefficient; this will result in the ineffective use of the data and, more importantly, invalid conclusion. It appears that new methodologies and novel approaches are required
to handle some of these problems. This proposal aims to: (i) develop such methodologies and approaches, (ii) illustrate their relevance and applicability for exposure data analysis, and (iii) address the related computational issues and provide programming codes, so that occupational hygienists and other practitioners can easily apply the proposed procedures.
While carrying out this research, we became more familiar with the statistical models and problems in occupational hygiene. In fact, all of the models and problems that we propose to investigate are either taken directly from the literature on industrial hygiene, or motivated by the nature of some of the exposure data that we looked at. The problems that we describe later in the proposal are essentially on the following topics: the univariate, bivariate and multivariate lognormal distributions, methodology for comparing several test methods or samplers, data analysis based on exposure samples that include values below the limit of detection, statistical techniques for the evaluation of low-cost, easy to use, and non-invasive test methods, and regression models and calibration with special emphasis on biological monitoring.
Researchers from other disciplines, who use statistical methods, typically rely on what are available in the popular software packages. Researchers in industrial hygiene are no exception, as we noticed from the industrial hygiene literature. It appears to us that this is a severe limitation since, as already noted, the questions that come up in the context of exposure data analysis do not always fit into the statistical routines that are widely available. A major thrust of our research is to address this limitation by developing statistical methodologies to answer the ``right questions.'' Along the way, we will be highlighting the limitations, inaccuracies and inappropriateness of some of the statistical procedures that are commonly used for exposure data analysis. The real success of our efforts will depend on the extent to which our procedures will be noticed and used for analyzing exposure data.
Publications Resulted
from the NIOSH Project
Krishnamoorthy, K. and Xu, Z. (2011). Confidence limits for lognormal percentiles and for lognormal mean based on samples with multiple detection limits. To appear in Annals of Occupational Hygiene.
Krishnamoorthy, K., Mallick, A. and Mathew, T.
(2011).
Inference for the Lognormal Mean and Quantiles Based on Samples with
Non-Detects.
Technometrics, 53, 72-83.
PDF
Krishnamoorthy, K. and Mathew, T. (2009).
Inference on the symmetric range
accuracy.
Annals of Occupational Hygiene,
53, 167-171.
Krishnamoorthy, K.
, Mallick, A. and Mathew, T. (2009).
Model
based imputation approach for data analysis in the presence of
non-detects. Annals of Occupational Hygiene,
59, 249-268.
Krishnamoorthy, K.
and Mathew, T. (2008).
Statistical Methods for Establishing Equivalency of Several Sampling
Devices. Journal of
Occupational and Environmental Hygiene, 5, 15-21.
Krishnamoorthy, K. Mathew, T.
and Mukherjee, S. (2008).
Normal based methods for a Gamma distribution: Prediction and Tolerance
Interval and stress-strength
reliability. Technometrics,
50, 69-78.
Krishnamoorthy , K., Mathew, T. and Lu, F.
(2007). A
Parametric Bootstrap Approach for ANOVA with Unequal Variances: Fixed
and Random Models. Computational Statistics and Data Analysis,
51 5731-5742.
Krishnamoorthy, K., Mathew, T. and Ramachandran, G. (2006).
Generalized
p-values and confidence intervals: A novel approach for analyzing
lognormally distributed exposure data. Journal of Occupational
and Environmental Hygiene, 3, 642-650.
Krishnamoorthy, K. and Guo, H. (2005).
Assessing
occupational exposure via the one-way random effects model with
unbalanced data.
Journal of Statistical Planning and Inference, 128, 219-229.
Krishnamoorthy, K. and Mathew, T. (2004).
One-Sided
tolerance limits in balanced and unbalanced one-way random models based
on generalized confidence limits. Technometrics, 46, 44-52.
Krishnamoorthy, K. and Mathew, T. (2003).
Inferences
on the means of lognormal distributions using generalized p-values and
generalized confidence intervals. Journal of Statistical Planning
and Inference, 115, 103 – 121.
[one of the 25 most downloaded articles from JSPI 2002-04]
Krishnamoorthy, K. and Mathew, T. (2002).
Statistical
methods for establishing equivalency of a sampling device to the OSHA
standard. American Industrial Hygiene Association Journal,
63, 567-571.
Krishnamoorthy, K. and Mathew, T. (2002).
Assessing
occupational exposure via the one-way random effects model.
Journal of Agricultural, Biological and Environmental Statistics, 7,
440-451.