The study will be conducted in three phases. Phase I will be conducted from September 1, 1997 through January 14, 1999; phase II from January 15, 1999 through January 14, 2000; phase III from January 15, 2000 through September 30, 2000. The study’s original start date was amended to September 1, 1997; its end date is stipulated as September 30, 2000.

Phase I - September 1, 1997 through January 14, 1999

Phase II - January 15, 1999 through January 14, 2000

Phase III - January 15, 2000 through September 30, 2000

Facilities

Non-federal Funding Sources and Facilities

Chronological Schedule of Project Accomplishments and Milestones

Indicate by whom each element of the work plan will be carried out including supporting agencies, consultants and contractors.

Health Effects Study
Staff	Role	Agency
H. William Barkman, M.D.	Co-Prin. Investigator	University of Kansas Medical Center
John S. Neuberger, DrPH	Co-Prin. Investigator	University of Kansas Medical Center
J. Thomas Pierce, Ph.D.	Investigator	University of Kansas Medical Center
Dennis Wallace, Ph.D.	Investigator	University of Kansas Medical Center
William Jewell, M.D.	Consultant	University of Kansas Medical Center
Gregory A. Reed, Ph.D.	Quality Assurance Mgr.	University of Kansas Medical Center
Mary G. Walker, B.A.	Proj. Coordinator	University of Kansas Medical Center
Tom A. Anderson, M.S.	Driver/Engineer	University of Kansas Medical Center

Environmental Monitoring Study
Staff	Role	Agency
H. William Barkman, M.D.	Co-Prin. Investigator	University of Kansas
Glen A. Marotz, Ph.D.	Investigator	University of Kansas
Richard W. Baldauf, M.S.	Investigator	University of Kansas
Ray E. Carter, Jr., M.S.	Investigator	University of Kansas

top of page

Phase I - September 1, 1997 through January 14, 1999

Prepare Final Study Protocol
Drs. Barkman, Lane, Marotz, Neuberger, Wallace, Messrs. Baldauf and Carter, and Ms. Walker.
Establish Advisory Committee
Drs. Barkman, Neuberger, Pierce, Lane, and Marotz.
Contact and Meet with Leaders of Sedan
Drs. Barkman, Neuberger, Lane, Pierce, and Wallace; Messrs. Baldauf and Carter, and Ms. Walker.
Coordinate and Conduct Stakeholders and Community Information Meetings
Drs. Barkman, Brothers, Neuberger, Lane, Pierce, Messrs. Baldauf and Carter, Ms. Cote, and Ms. Walker.
Develop Health Questionnaire
Drs. Barkman and Wallace.
Review Existing Air Sampling Data
Drs. Lane and Marotz, and Messrs. Richard Baldauf and Ray Carter.
Select Study and Control Communities and Counties for Study’s Cancer Component
Drs. Neuberger and Brothers
Conduct Meetings with Health Officials and Community Residents for Study’s Cancer Incidence Component
Drs. Neuberger and Brothers.
Conduct Literature Review for Study’s Cancer Component
Dr. Brothers.
Obtain Observed Number of Cancer Cases in Exposed Areas for Study’s Cancer Component
Drs. Neuberger and Brothers.
Finalize Statistical Design for Study’s Environmental and Health Components
Drs. Barkman, Wallace, Lane, and Marotz.
Complete and Disseminate Study’s Final Protocol
Drs. Barkman, Neuberger, Lane, Wallace, Mr. Baldauf, and Ms. Walker.

top of page

Phase II - January 15, 1999 through January 14, 2000

Collect and Analyze Air Samples
Drs. Lane, Marotz, and Messrs. Baldauf and Carter.
Collect and Analyze Existing Hospital Data on Visits for Acute Respiratory Illness
Drs. Barkman and Wallace, Ms. Lava, and Mr. Brewer.
Administer Health Questionnaire
Questionnaire will be administered by mail. This task will be organized and executed by Ms. Walker and Mr. Brewer.
Review Health Questionnaires
Drs. Barkman and Wallace.
Select and Notify Study Participants
Dr. Barkman, Mr. Brewer, and Ms. Walker.
Take Health Histories and Conduct Physical Examinations and Pulmonary Function Tests
Dr. Barkman, nurse practitioner and a clerical person.
Collect and Analyze Data on Hospital Visits for Acute Respiratory Illness
Drs. Barkman and Wallace, Ms. Lava, and Mr. Brewer.
Complete Case Findings for Study’s Cancer Component
Drs. Neuberger and Brothers.
Complete Literature Reviews for Study’s Cancer Component
Drs. Neuberger and Brothers.
Complete Statistical Analysis for Study’s Cancer Component
Drs. Neuberger, Wallace, and Ms. Lava.

Phase III - January 15, 2000 through September 30, 2000

Take Health Histories and Conduct Physical Examinations and Pulmonary Functions Tests
Dr. Barkman, nurse practitioner and the clerical person.
Correlate Environmental and Health Data
Drs. Barkman, Wallace, Neuberger, Lane, Marotz and Mr. Baldauf.
Complete Statistical Analysis of Study’s Environmental and Respiratory Health Components
Drs. Barkman, Wallace, Lane, Marotz, Mr. Baldauf and Ms. Lava.
Draft and Review Final Report
Drs. Barkman, Wallace, Neuberger, Pierce, Lane, Marotz, Mr. Baldauf, Ms. Lava, Ms. Walker.
Complete Final Report
Drs. Barkman, Wallace, Neuberger, Pierce, Lane, Marotz, Mr. Baldauf, Ms. Lava, and Ms. Walker.
Hold Stakeholders and Communities Information Meetings on Study Results
Drs. Barkman, Neuberger, Pierce, Lane, Marotz, and EPA.

top of page

Describe sampling and data collection procedures, analytical methods, and methods for evaluating the results and successes of the project.

Environmental Data Collection, Analysis, and Results Evaluation

The sampling frame design for air quality data will permit periodic and discrete sample collection at fixed and designated points within the study and control communities. All air samples will be collected by the study’s environmental scientists using standard protocols. Analytical methods will include gas chromatographic, mass spectrometric, gravimetric, atomic absorption spectrophotometric and inductively-coupled plasma spectroscopic analysis. Inferential parametrical and nonparametrical procedures will be used to test relevant hypotheses. The sampling frame will also be used to collect meteorological information relevant to the selected air quality species at designated locations. The Environmental Monitoring Assessment Quality Assurance Project Plan, outlining the methodology for all air sampling collection and analysis, is attached as Appendix A to this document.

Epidemiological Data Collection, Analysis, and Results Evaluation of the Study’s Cancer Component

Cancer incidence and mortality rates will be developed from the Kansas Cancer Registry and the Kansas Department of Health and Environment. In addition, cancer incidence data will be used from the National Cancer Institute’s Surveillance, Epidemiology and End Result (SEER) program. Data on expected cancers will be compared with the study and control counties. Additional data will be obtained for 1997 from area physicians and hospitals. Additional data will be obtained on cancer incidence and prevalence among the population sample used for the respiratory health questionnaire.

Data Collection, Analysis, and Results Evaluation of the Study’s Respiratory Health
Component

Data Collection and Management

The respiratory health component of this study involves the collection of four primary data streams–questionnaire data received from the household survey, medical data on acute respiratory events obtained from hospital medical records review, data on inhaler use collected from schools in the participating cities, and medical data obtained from two medical examinations on each of the 50 subjects in each community participating in the one-year respiratory health medical evaluation. All data from each stream will be collected using an appropriate data collection form and entered into an electronic data base.

The data entry and data management system will be designed using Microsoft Access ®. For each data stream, 15% of the records will be double entered. If greater than 0.1 percent errors in coding are found in any form, all forms will be double entered and compared with corrections made before the data are considered ready for analysis. After data entry and review are completed, the data will be converted to the SAS data system for analysis. A set of consistency and range checks will be run and all data reviewed before analysis is initiated.

top of page

Statistical Analyses

The primary analytic objective of the respiratory health questionnaire is to obtain estimates of the prevalence of major respiratory illnesses in each of the participating communities and to compare prevalence in each of the combustor communities to that in the control community. Because the data will be collected using a population-based survey instrument, point and interval estimates of population prevalence for each primary respiratory condition will be developed using appropriate sample weights. Weighted least squares techniques that are analogous to ANOVA techniques for continuous outcomes will be used to test for differences in prevalence among the study communities.

Data on use of inhalers in the schools and acute respiratory events based on hospital emergency room use will be collected on a daily basis throughout the year. These data will be used to examine the relationship between air quality and respiratory outcomes in each of the study communities and to assess whether those relationships vary by community.

As part of the initial descriptive analyses, we will examine the year-long temporal patterns in respiratory events and air quality measurements at the different sites in the five communities to identify patterns that suggest possible relationships. As a part of these initial analyses, we will examine patterns of events as a function of both season and day of the week to assess whether those variables should play a role in subsequent analyses. Also these initial analyses will address whether data provide sufficient stability to examine events on a daily basis or whether aggregation across full weeks appears more reasonable.

After the descriptive analyses are completed, formal model based analyses using extensions of generalized linear models (McCullagh and Nelder, 1989; and Diggle, Liang, and Zeger, 1994) that account for the correlation imbedded in the repeated observations from the same communities will be used to assess the relationship of events to air quality. Again, depending on the numbers of outcomes for any particular unit of analysis, we will use either an identity link (if data appear to be normally distributed) or log link (if data behave as Poisson counts). Statistical inferences about the effect of air quality on respiratory events will be based on Wald-type tests using the robust variance estimator described by Liang and Zeger, 1986.

The primary outcome from the respiratory health medical evaluations will be the change in lung function over the course of the year. Level of lung function will be determined by predicted values based on age, gender, height, and race. The initial individual health evaluation will serve as the baseline for the follow-up evaluation that will occur at the end of the study’s data collection phase. Thus, the individual will serve as his/her own control over the sampling period. Depending on the distributional properties of these change scores, either ANOVA models or the non-parametric Kruskal-Wallis analogue will be used to assess differences in change scores in the five communities.

Protocol I	Protocol II
Protocol III	Protocol IV
Protocol V	Protocol VI
Appendix

top of page