Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
APPENDIX C INTERPRETABILITY OF FOLLOW-UP QUESTIONNAIRE DATA The follow-up questionnaire was part of a concerted attempt to determine the nature, extent, and severity of protracted problems associated with exposure of volunteers to a variety of chemicals at the Edgewood test site. Considerable discussion and controversy attended the design and analysis problems. Of particular concern were the use of a specially constructed but untested questionnaire, the relatively small groups of men exposed to some chemicals, the sensitivity of the questionnaire for detecting the problems most probably associated with exposure, and the potential for causing extraordinary concern among the volunteer soldiers. Many special- ists in questionnaire development were consulted regarding the types of information that might be elicited and the specific wording of questions. The resulting questionnaire was a compromise agreed on by the five panels and the National Research Council Committee on Use of Human Subjects. Several subjects of concern were not in- cluded in the final questionnaire, such as probes for specific symptoms, suicide attempts, diseases, treatments, behaviors, de- tailed history of later job-related exposure, accidents, and spon- taneous abortions. An issue of great concern was the relatively small groups of men exposed to the psychochemicals and their effects on interpretability. Briefly stated, it was felt at the outset by the panel reviewing psychochemicals that data obtainable from a sur- vey might add little to our understanding of the long-term health effects of chemicals tested. POPULATION The population to be followed was not contacted regularly after discharge from the Army and had not consented to or expected a follow-up attempt. It was therefore difficult to secure the cooperation and sample sizes desired to make the total response statistically useful. It is assumed that 6,395 of 6,720 soldiers were alive at the time of follow-up. Of those, 4,085 (64% of total or 82% of those located) responded to the questionnaire. Those who did not respond might not constitute a random sample of the entire population. It is possible that many of the nonrespondents failed to respond because they had nothing important to say. If that were true, it would strengthen our belief that long-lasting effects were generally not present. However, it is also possible that the nonrespondents had other reasons for failing to respond, such as -83-
very low economic status, incarceration, and long-term hospitali- zation. The importance of obtaining complete accounting of all those followed is shown by considering the numbers of subjects exposed to each drug. Large samples were exposed to the anticholinesterases, the anticholinergics, and the irritants and vesicants, but fewer than 100 were exposed to the psychochemical Sernyl. The loss of respondents in the smaller groups makes the conclusions for Sernyl more tentative, because of the statistical properties (i.e., power) of the comparison tests. QUESTIONNAIRE DESIGN Most of the problems associated with the questionnaire were due to a lack of explicit hypotheses as to potential long-term effects of the drugs studied. Questions of general interest were included, but specific hypotheses were not assessed with questions. Further- more, the survey instrument was a questionnaire, not an interview. The questions therefore had to be simple, easily understood, and able to be answered quickly. Such constraints limit the specificity and detail of the information to be collected. Even given those characteristics of the survey, some of the questions that were included could have been reworded or recon- sidered. For example, the section on employment might have been longer and more complex than necessary; and the question as to children born to the subject could have indicated more clearly that unmarried subjects should report the numbers and sexes of their children. OTHER CONSIDERATIONS The climate in which the questionnaire was used might raise concern about interpretation of the results. For example, subjects who wished (for whatever reason) to misrepresent the nature and severity of their problems could distort the results (and hence their interpretation), especially if their chemical-test group was fairly small. In addition, as in any cohort study, there was no control of the subjects' environments after discharge from the service, so even legitimate complaints associated with exposure to toxic substances could have been due to occupational or accidental exposure to chemical agents, rather than to exposure at Edgewood. SUMMARY Some caution must be exercised in viewing the data obtained from the questionnaire. There were constraints on the information that -84-
could have been gained on long-term effects of chemical exposure, even if follow-up had been conducted by personal interview with questions designed to test specific hypotheses. There were differ- ences among the subjects exposed, including possible misrepresenta- tion by subjects, variability in their lives after discharge, and an inherent difficulty of finding a representative follow-up sample after 10-30 years. Beyond these constraints, interpretation of the data that were collected entails additional problems. The data on general health, family, and work status are interpretable and appear to show good adjustment to civilian life by most of the men sampled. No major identifiable effects are observable in these data. However, the limited information available from the follow-up on these soldiers does not permit definitive conclusions regarding the nature and extent of possible long-term problems resulting from chemical exposure at Edgewood. -85-