4
General Results

This chapter presents a brief introduction to the examination data, beginning with a more detailed discussion of their collection, coding, and organization. As stated in Chapter 2, all examination data were abstracted and coded by trained nosologists using ninth edition International Classification of Diseases, Clinical Modification (ICD-9-CM) codes. Additionally, they recorded the source of each coded medical condition within the exam as one of the following: discharge summary; history; findings; laboratory, x-ray, or other such services; psychiatric or psychological examination; medical consultation; or VA benefits application form. Medical conditions were also characterized as to their time of onset (while a prisoner of war or not, for example) and whether they were currently resolved. Medications being taken for a given condition were also abstracted and keyed, but the complexity and the lack of standardization of these data prevented their analysis and inclusion in this report.

The additional coding of the source and resolution of each medical condition was undertaken to provide information that might help account for some of the variability in the exams themselves. Although the VA circulars provided directives about the general scope of the exam, the conduct of the exam varied from facility to facility largely because of differences among examiners in the performance of the comprehensive medical examination. Moreover, there were some additional administrative differences among examination sites that contributed to examination variability—for example, the number of days it took to conduct an exam. For the most



The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement



Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.

OCR for page 33
The Health of Former Prisoners of War: Results from the Medical Examination Survey of Former POWs of World War II and the Korean Conflict 4 General Results This chapter presents a brief introduction to the examination data, beginning with a more detailed discussion of their collection, coding, and organization. As stated in Chapter 2, all examination data were abstracted and coded by trained nosologists using ninth edition International Classification of Diseases, Clinical Modification (ICD-9-CM) codes. Additionally, they recorded the source of each coded medical condition within the exam as one of the following: discharge summary; history; findings; laboratory, x-ray, or other such services; psychiatric or psychological examination; medical consultation; or VA benefits application form. Medical conditions were also characterized as to their time of onset (while a prisoner of war or not, for example) and whether they were currently resolved. Medications being taken for a given condition were also abstracted and keyed, but the complexity and the lack of standardization of these data prevented their analysis and inclusion in this report. The additional coding of the source and resolution of each medical condition was undertaken to provide information that might help account for some of the variability in the exams themselves. Although the VA circulars provided directives about the general scope of the exam, the conduct of the exam varied from facility to facility largely because of differences among examiners in the performance of the comprehensive medical examination. Moreover, there were some additional administrative differences among examination sites that contributed to examination variability—for example, the number of days it took to conduct an exam. For the most

OCR for page 33
The Health of Former Prisoners of War: Results from the Medical Examination Survey of Former POWs of World War II and the Korean Conflict part, in larger, urban hospitals administrative difficulties in scheduling appointments for crowded clinics often meant that a patient would have to return several times to complete the exam. In contrast, in smaller, rural VA hospitals, driving distances were often so great as practically to require that the exam be done in a single day. Administrative matters such as these were not standardized, and each VA facility was compelled to make its own best local arrangements to accommodate the research study. That such administrative and other site-to-site differences would contribute to exam variability was recognized from the outset of the study. Because it was impossible to eliminate this variability, the study's strategy was to attempt to limit it (through the directives in the VA circulars) and examine it. For this latter purpose, study staff made site visits to more than 30 VA facilities to explain the research to VA field personnel, answer their questions about the conduct of the study, and gain an appreciation of the actual process by which the examination data were being collected. At each visit, study staff met with the POW physician and administrative coordinators, as well as with other available personnel, and attempted to ''walk through'' a typical examination. Appendix B contains a list of site visits. When the study was designed, it was known that a great deal of information would be collected about each participant, given the advancing age of the study groups and especially of the WW II veterans. The volume of data collected was considerable, as Table 4.1 illustrates. The table displays the mean number of diagnoses per examination by study group and indicates, among other things, that examinations of PWK group members each averaged nearly 70 codable medical conditions; the smallest average among the POW groups was nearly 60. Veteran controls averaged markedly fewer TABLE 4.1 Number of Medical Conditionsa Among Former POWs of World War II (WW II) and the Korean Conflict and Veteran Controls (mean ± 2 standard errors, N in parentheses) War Era POWs Veteran Controls WW II, Pacific 62.5 ± 4.2 (250) 45.4 ± 6.2 (54) WW II, Europe 58.4 ± 5.2 (142) 35.2 ± 7.6 (27) WW II, Europe, malnourished 66.9 ± 6.2 (83) N.a. Korean conflict 69.3 ± 3.8 (408) 49.0 ± 5.3 (103) N.a., not applicable a Number of current and past medical conditions coded from the comprehensive medical examination.

OCR for page 33
The Health of Former Prisoners of War: Results from the Medical Examination Survey of Former POWs of World War II and the Korean Conflict diagnoses than their comparable POWs in every case, but among WP and WK groups, the average number of conditions was still from 45 to nearly 50. Every mention of a medical condition is included in the above counts; if a condition is mentioned in more than one section of the exam (see below), it is counted each time. Such crude data are not meant to represent meaningful differences among study groups. Still, they show the magnitude of information collected and hint at the average "disease burden" of study subjects. The great volume of data collected, as reflected in the large average number of medical conditions, made it difficult to distinguish the effects of normal aging from the sequelae of military captivity. It was assumed that the critical differentiation of historical and current medical conditions could be made easily, using one part of the examination, the "summary sheet." The sheet was meant to be filled out by the POW physician coordinator as part of an "exit interview" for each examinee; presumably, it would contain all current medical conditions of interest. Once analysis began, however, it became apparent that about one in seven examinations had no summary section and that the number of medical conditions mentioned on summary sheets constituted only about 15% of the total diagnostic information. This figure was lower than those for the history (21%), findings (24%), and consultation (18%) sections of the exam, but higher than the figures for the sections on psychiatric exams (8%), lab and x-ray (6%), and VA benefits forms (8%). It was not possible, therefore, to derive complete medical data from any one examination section, and some of the analyses in later chapters will be source dependent. For the remainder of this chapter, data on medical conditions were taken from all parts of the exam. The simplest way to display such data is to categorize them into the customary broad diagnostic categories of the ICD-9-CM classification scheme (e.g., category I, infectious diseases, are all codes from 001 to 139.8; category II, neoplasms, are codes from 140 to 208.9; etc.). For each individual, each coded medical condition is assigned to one of these broad ICD rubrics, but multiple medical conditions within a single category are counted only once, so that the final categorization tells simply the presence or absence of a certain type of medical condition in a person. The resulting individual data can then be aggregated to calculate lifetime prevalence rates (Kleinbaum et al., 1982), although strictly speaking, these might better have been called cumulative lifetime prevalence rates because the lives of these subjects are not yet completed. By the customary definition, a lifetime prevalence rate is the probability that a person has ever had some given medical condition; it is calculated simply by dividing the number of subjects in a group who have ever had that condition by the total number of people in the group.

OCR for page 33
The Health of Former Prisoners of War: Results from the Medical Examination Survey of Former POWs of World War II and the Korean Conflict The examination of the descriptive data in Table 4.2, which contains POW and control lifetime prevalence rates, is hampered by lack of knowledge concerning the stability or, conversely, the variability of these rates. Statistical tests are customarily used to compare such rates because a statistical test takes into account not only the magnitude of differences in rates but also the variability of the rates based on the size of the sample. Given the low response rates (discussed in detail in Chapter 3), however, the customary use of statistical tests would be inappropriate here. Yet without statistical tests, there is no convenient yardstick with which to determine whether a given difference in prevalence rates is noteworthy. If it is recognized that a statistical test done in this setting has no valid inferential use but is instead merely a way of marking more "noteworthy" or "appreciable" differences, then the usual statistical machinery can be pressed into service in this unusual setting. Thus, statistical tests—specifically the chi-square test—will be used to single out appreciable differences, even though they cannot be labeled "significant.'' Additionally on the subject of statistical tests, it should be noted that even the limited use of a statistical test as a marker has the same liability that it has when used in appropriate settings; it customarily classifies differences as either noteworthy or not noteworthy without any gradation between the two. Moreover, a statistical test may conclude that a difference is not noteworthy simply because sample sizes are small (even though the samples are representative); the test is then said to have low power. In the specific case of WW II POW-versus-control comparisons, the very small number of control examinations provides little power to detect and label prevalence rate differences as noteworthy. Finally, in presenting the results of significance tests applied to the data in Table 4.2, it should be remembered that more than 90 such tests were done. By the laws of chance, approximately four "noteworthy" differences would be expected when comparing rates that differed only by random sampling error. In Table 4.2, the PWP group displays an appreciably higher rate of other endocrine disease (this category includes current and historical nutritional deficiencies) and nervous system disease, as well as an appreciably higher rate of neuroses than the WP group; however, PWP has an appreciably lower rate of blood diseases. PWE and PWEM both display appreciably higher rates of psychoses, neuroses, and injuries or poisonings (almost exclusively injuries), and PWEM an appreciably higher rate of urogenital disease. The PWK group shows the greatest number of noteworthy differences—compared with the WK control group—with elevated rates of infectious disease, other endocrine disease, psychoses, neuroses, nervous system, acute respiratory, digestive, and musculoskeletal disease, as well as injury and poisoning (again, almost exclusively injuries). Like PWP, they show an appreciably lower rate of blood diseases. It should be noted that the

OCR for page 33
The Health of Former Prisoners of War: Results from the Medical Examination Survey of Former POWs of World War II and the Korean Conflict TABLE 4.2 Lifetime Prevalence Rates (percentages) of Medical Conditions Ascertained Through Examination, by ICD Rubrica and Study Group ICD Rubric PWP PWE PWEM PWK WP WE WK Infectious diseases 96.4 96.5 98.8 98.0 100.0 92.6 94.2b Malignant neoplasms 20.0 19.7 16.9 9.1 13.0 22.2 12.6 Benign neoplasms 15.6 15.5 19.3 17.7 25.9 14.8 13.6 Diabetes 14.8 13.4 12.1 12.8 18.5 11.1 10.7 Other endocrine diseases 85.6b 68.3 81.9 70.8b 66.7b 70.4 50.5b Blood diseases 14.0b 11.3 7.2 12.0b 29.6b 14.8 22.3b Psychoses 19.6 15.5b 24.1b 23.0b 9.3 0.0b 10.7b Neuroses 91.2b 88.7b 90.4b 93.1b 79.6b 63.0b 83.5b Nervous system 36.4b 18.3 22.9 25.3b 18.5b 11.1 11.7b Sense organs 86.8 85.9 86.8 88.2 90.7 77.8 83.5 Heart disease 62.0 60.6 59.0 51.2 66.7 63.0 54.4 Cerebrovascular disease 8.4 9.2 12.1 5.2 1.9 0.0 8.7 Hypertension and other circulatory disease 78.4 71.8 79.5 68.6 66.7 66.7 68.9 Acute respiratory 36.4 26.7 31.3 35.8b 25.9 33.3 24.3b Chronic respiratory 68.4 65.5 67.5 61.0 64.8 59.3 56.3 Digestive 94.4 90.1 91.6 94.4b 87.0 81.5 83.5b Urogenital 60.4 54.2b 49.4 47.1 66.7 33.3b 44.7 Skin 75.2 71.1 79.5 70.3 72.2 63.0 66.0 Musculoskeletal 93.6 86.6 89.2 92.4b 88.9 81.5 85.4b Congenital conditions 8.0 4.9 7.2 5.6 3.7 3.7 3.9 Symptoms and ill defined conditions 98.8 97.9 98.8 99.3b 98.2 100.0 94.2b Injury and poisoning 85.6 84.5b 91.6b 91.4 81.5 66.7b 94.2 V-codesc 52.4 54.9 57.8 50.5 46.3 55.6 54.4 N 250 142 83 408 54 27 103 PWP, prisoner of war, Pacific theater, WW II; PWE, prisoner of war, European theater, WW II; PWEM, prisoner of war, malnourished, European theater, WW II; PWK, prisoner of war, Korean conflict; WP, war veteran, Pacific theater, WW II; WE, war veteran, European theater, WW II; WK, wounded war veteran, Korean conflict. a Coded in ninth revision, International Classification of Disease, Clinical Modification. b Noteworthy (see text) difference between POW and comparable control group. c Factors influencing health status and contact with health services. PWK and WK groups contain the most examinations, which could well affect the larger number of noteworthy differences (see the discussion of statistical power above). In addition, the PWP group, like the PWK, has somewhat higher, but not noteworthy, rates of digestive disease. When faced with so many results, one needs to look for consistency

OCR for page 33
The Health of Former Prisoners of War: Results from the Medical Examination Survey of Former POWs of World War II and the Korean Conflict across group comparisons and to draw on external results to assess clinical plausibility. For example, the fact that neuroses and psychoses are elevated in nearly all comparisons, coupled with the knowledge that previous studies have shown that psychological illness is found at higher rates among POWs, argues that the apparently increased rates of psychoses and neuroses are more likely to be a real and important clinical difference. In interpreting these apparently very high rates, it should be noted that the psychoses category of ICD-9-CM includes major depression and drug psychoses, such as drug withdrawal syndrome and pathological drug intoxication. The neuroses category includes anxiety states, neurotic depression, depressive disorder, and posttraumatic stress disorder, all later shown individually to be highly prevalent. Similarly, an elevated rate of other endocrine disease, a category that comprises the nutritional deficiencies, is consistent over several POW groups, as are the elevated rates of injuries among POWs. Other isolated findings, particularly those in the PWK group, are more difficult to interpret, although the somewhat elevated rate of digestive disease in the PWP group and the studies of Australian prisoners of war (which found a higher lifetime rate of ulcer) together argue that this finding has an underlying clinical basis. The most puzzling of the findings common to more than one group is the appreciably lower rate of blood diseases among PWP and PWK. Further discussion of this finding is undertaken in Chapter 6. Given the potential importance of consistency of findings, one can further note that Table 4.2 shows few large differences among disease rates for POWs, despite the somewhat younger age of those in the PWK group compared with the WW II veterans in the PWP and PWE groups. Age differences are not an explicit consideration in the POW-versus-control comparisons because control groups were selected to have the same age distribution as their respective POW samples. Given the importance of external comparisons, this chapter ends by comparing examination-based lifetime prevalence rates for selected conditions with similar rates from Eberly and Engdahl's (1991) examination study of some 426 former POWs seen at the Minneapolis VA Medical Center. These 426 POWs represented some 60% of the 696 POWs known to be living in the region, and participants in the Minneapolis study were comparable to nonparticipants in terms of age, education, marital status, and military service-connected disability status. The Minneapolis examinations were conducted between April 1984 and April 1987, somewhat early in the course of the MFUA study, which limits the overlap of the Minneapolis and MFUA examination data to only 8 POWs. Like the lifetime prevalence estimates from the MFUA study, those from the Eberly and Engdahl study are based on a single, comprehensive examination that collected current and historical information. POWs were

OCR for page 33
The Health of Former Prisoners of War: Results from the Medical Examination Survey of Former POWs of World War II and the Korean Conflict randomly assigned to one of ten examining psychiatrists, all of whom were experienced in examining and diagnosing psychiatric illnesses in POWs and combat veterans. The psychiatrists collected information on the POW's premilitary history, military and POW experiences, postmilitary adjustment, and current psychiatric status. Historical and current diagnoses were based on this information. For roughly two-thirds of the 426 Minneapolis exams, the examining psychiatrist used either the Research Diagnostic Criteria of the SADS-L (Spitzer et al., 1978) or a PTSD checklist. The remaining one-third of the sample was reviewed by Eberly and Engdahl, who found that frequencies of the diagnoses obtained using the SADS-L and PTSD checklist were comparable to frequencies later obtained by record review. A sample of 30 cases was rated using both methods, and there was complete agreement between the two in 23 (77%) of these cases. The lifetime prevalence estimates for somatic disorders were derived after experienced coding personnel reviewed the entire medical exam and coded all disorders using the VA's coding system for medical compensation. Unlike the MFUA study, the Minneapolis study did not include veteran controls. Instead, Eberly and Engdahl compared their findings to lifetime prevalence rates available from a number of studies of selected conditions in the general population (National Institute on Aging, 1986; Weissman and Myers, 1978; Robins et al., 1984; Helzer et al., 1987). For comparison purposes, the following discussion of MFUA examination data will be limited to these same selected medical conditions, and the general population rates cited by Eberly and Engdahl will also be displayed. Table 4.3 compares lifetime prevalence rates of selected medical and psychiatric conditions for POWs and controls in the MFUA study, POWs in the Minneapolis study, and the general population; the selected conditions shown are those chosen by Eberly and Engdahl. Despite the crude nature of these data (both WW II and Korean conflict POWs are combined, for example), there is a remarkable degree of agreement across all groups. Among the medical conditions, with the exception of intermittent claudication (discussed in detail later in this report), the MFUA data for POWs and controls and the general population data are virtually equivalent. Among Minneapolis POWs, only the lifetime prevalence rate of hypertension is virtually identical to those for MFUA POWs and controls and the general population; for diabetes, cerebrovascular accident, and myocardial infarct, the rates for Minneapolis POWs all seem a bit low in comparison. Among psychiatric conditions, with the exception of posttraumatic stress disorder (PTSD) and depressive symptoms in MFUA POWs, there is a similar congruence of lifetime prevalence rates (only schizophrenia in MFUA controls appears exceptionally low)—although sample sizes are quite small. It is reasonable to assume that somewhat different diagnostic schemes ac-

OCR for page 33
The Health of Former Prisoners of War: Results from the Medical Examination Survey of Former POWs of World War II and the Korean Conflict TABLE 4.3 Lifetime Prevalence Rates (percentages) of Medical and Psychiatric Diagnoses Among All POWs and Controls in the MFUA Cohort, Among All POWs in the Minneapolis Study, and in the General Population Diagnosis MFUA POWs Minneapolis POWs MFUA Controls General Population Medical conditions Hypertension 44.6 39.2 40.2 40.4 Diabetes mellitus 13.4 8.9 13.0 13.2 Cerebrovascular accident 6.2 0.9 5.5 6.4 Myocardial infarction 17.9 7.6 19.6 15.5 Intermittent claudication 13.9 1.9 7.6 2.1 Psychiatric conditions Major or minor depressiona 57.6a 23.3 20.7a 16.5 Bipolar I or II disorder 0.5 0.9 0.5 1.2 Alcohol abuse or dependence 23.6 21.1 23.4 18.2 Schizophrenia 2.3 1.9 0.5 1.1 Posttraumatic stress disorder 35.9 70.9 9.2 0.5 Number examined 883 426 184 N.a. N.a, not applicable. a For the MFUA cohort, depression or depressive disorder. SOURCE: Eberly and Engdahl (1991). count for most of these differences. For example, the SADS-L criteria used in the Minneapolis study were probably stricter than the simple diagnosis of depressive symptoms used in the MFUA study—certainly the SADS-L criteria were more uniformly applied—and this would explain the lower rate of depression in the Minneapolis cohort. The difference in PTSD rates is more difficult to explain, although the analyses presented in Chapter 5 show that the rates of PTSD vary appreciably by measurement instrument. There is an important comorbidity of PTSD and depressive symptoms, which can also make such comparisons more difficult. The medical conditions of Table 4.3 are shown in Table 4.4 for all POWs and controls broken down into war period and theater. (This is a reasonable exercise because known differences in age, length of captivity, and harshness of treatment during captivity make it likely that lifetime prevalence rates vary among study groups.) In addition to the conditions listed in Table 4.3, ulcer and general anxiety are included (as they were by Eberly and Engdahl), as well as dysentery, malaria, beriberi, and frozen

OCR for page 33
The Health of Former Prisoners of War: Results from the Medical Examination Survey of Former POWs of World War II and the Korean Conflict TABLE 4.4 Lifetime Prevalence Rates (percentages) of Medical and Psychiatric Diagnoses Among POWs and Controls in the MFUA Cohort and Among POWs in the Minneapolis Study, by Study Group   MFUA POWs MFUA Controls Minneapolis POWs Diagnosis PWP PWE PWEM PWK WP WE WK PWP PWE PWK Medical conditions Hypertension 52.4a 42.3 51.8 39.2 37.0 40.7 41.8 42 39 39 Diabetes mellitus 14.8 13.4 12.1 12.8 18.5 11.1 10.7 N.a. N.a. N.a. Cerebrovascular accident 7.6 5.6 9.6 4.9 1.9 0.0 8.7 0 1 0 Myocardial infarction 20.0 19.7 24.1 13.5 24.1 11.1 19.4 19 21 22 Intermittent claudication 17.6 9.2 14.5 13.2 11.1 3.7 6.8 3 2 0 Dysentery 47.2a 25.4a 37.4a 42.9a 7.4 3.7 5.8 20 3 11 Malaria 60.8 2.8 1.2 20.6 50.0 7.4 22.3 54 3 11 Beriberi 61.2a 0.0 6.2 25.7a 0.0 0.0 1.0 48 <1 17 Frozen feet 10.0 16.2 38.6a 30.6a 1.9 14.8 8.7 5 17 11 Ulcer 19.6 20.4 24.1 26.0a 20.4 14.8 15.5 N.a N.a. N.a. Psychiatric conditions Depressive disorderb 52.4a 35.9a 53.0a 56.4a 16.7 0.0 28.2 31 22 28 Bipolar I or II disorder 0.8 0.0 0.0 0.5 1.9 0.0 0.0 3 <1 0 Alcohol abuse or dependence 19.6 18.3 21.7 28.2 14.8 7.4 32.0 22 22 11 Schizophrenia 1.2 0.7 2.4 3.4 0.0 0.0 1.0 3 2 6 Posttraumatic stress disorder 40.8a 23.2 27.7 39.0a 3.7 11.1 11.7 79 70 53 Generalized anxiety 44.0a 46.5a 62.7a 58.6a 13.0 7.4 25.2 58 47 44 Number examined 250 142 83 54 27 103 408 65 343 18 PWP, prisoner of war, Pacific theater, WW II; PWE, prisoner of war, European theater, WW II; PWEM, prisoner of war, malnourished, European theater, WW II; PWK, prisoner of war, Korean conflict; WP, war veteran, Pacific theater, WW II; WE, war veteran, European theater, WW II; WK, wounded war veteran, Korean conflict. N.a., not applicable. a Noteworthy (see text) difference between POWs and comparable control group. b For Minneapolis POWs, major or minor depression.

OCR for page 33
The Health of Former Prisoners of War: Results from the Medical Examination Survey of Former POWs of World War II and the Korean Conflict feet, conditions that are all known to be linked to harshness of treatment during captivity. Table 4.4 shows only two medical conditions not associated with prison camp treatment that have appreciably higher prevalences among POWs than controls in the MFUA study: hypertension (in the PWP group) and ulcer (in the PWK). Dysentery is the only prison camp symptom appreciably elevated in all four MFUA POW groups, with beriberi and frozen feet markedly elevated in two of the four groups; there were no noteworthy differences between POWs and controls for lifetime malaria prevalence. The prevalence of cerebrovascular accident appears high in the WK group, especially compared with WP and WE, but it is based on a small sample and is therefore not appreciably higher than the PWK rate (or the PWP, PWE, or PWEM rates). Given the general agreement among all the latter rates, it may be that WP and WE cerebrovascular prevalence rates are unusually low—these are both based on very small samples—rather than the WK rates being unusually high. Psychiatric conditions are a different matter, with depressive disorder and generalized anxiety showing appreciably higher prevalence rates across all four MFUA POW groups; the rate of 0% among WE is probably attributable to small (N = 27) sample size. PTSD is markedly elevated in PWP and PWK and is somewhat higher, although not appreciably so, in PWEM. None of the other psychiatric disorders showed an appreciably higher lifetime prevalence among POWs. Chapters 5 and 6 present a more detailed analysis of the examination data, as well as of the structured psychiatric interview and psychological data. Further discussion of the Eberly and Engdahl data can be found in Chapter 7. REFERENCES Eberly RE, Engdahl BE. 1991. Prevalence of somatic and psychiatric disorders among former prisoners of war. Hosp. Commun. Psychiatry 42:807–813. Helzer JE, Robins LN, McEvoy L. 1987. Post-traumatic stress disorder in the general population. N. Engl. J. Med. 317:1630–1634. Kleinbaum DG, Kupper LL, Morgenstern H. 1982. Epidemiologic Research: Principles and Quantitative Methods, Van Nostrand Reinhold, New York, p. 118. National Institute on Aging. 1986. Established Populations for Epidemiological Studies of the Elderly. NIH publication 86–2443. Washington, D.C., U.S. Government Printing Office. Robins LN, Heizer JE, Weissman MM, et al. 1984. Lifetime prevalence of specific psychiatric disorders at three sites. Arch. Gen. Psychiatry 41:949–958. Spitzer RL, Endicott J, Robin E. 1978. Research Diagnostic Criteria: rationale and reliability. Arch. Gen. Psychiatry 35:773–782. Weissman MM, Myers JK. 1978. Affective disorders in a U.S. urban community: the use of the Research Diagnostic Criteria in an epidemiologic survey. Arch. Gen. Psychiatry 35:1304–1311.