MicSem Publications

Schizophrenia and Chronic Mental Illness in Micronesia: An Epidemiological Survey

by Francis X. Hezel, SJ, with Michael Wylie

1992 Health

A community-based epidemiological survey using key informants and facility records in case finding was undertaken to better understand the occurrence of severe mental illness in Palau, the Federated States of Micronesia, and the Republic of the Marshall Islands. The goal of the survey was to identify all cases of schizophrenia and chronic psychosis, including affective psychosis and paranoid delusional disorder, using community identiflcation of abnormality rather than formal psychiatric diagnosis. The average prevalence rate for schizophrenia and chronic mental illness was 5.4 per 1,000 population, with a range of 3.2 to 16.7/1,000. Although rates varied greatly in different island groups–Palau's rate was almost twice that of Yap (the next highest area) and several times higher than rates in eastern Micronesia–the rates fall within accepted prevalence ranges established across cultures for schizophrenia. In Micronesia, schizophrenia and chronic mental illness were heavily weighted toward males, who constituted 77 percent of the total sample and outnumbered females by a ratio of 3.4/1.

The mentally ill have become a common feature of the social landscape in Micronesia in the last decade. One encounters them on the road wandering from place to place, often garishly dressed (or undressed), stopping people to beg cigarettes and engaging them in bewildering conversation, usually in fluent English. Others, far less visible, skulk in their homes and have dealings with almost no one other than their most intimate relatives. Many, if not most, of these people would be psychiatrically diagnosed as having some form of severe thought disorder or "psychosis." Schizophrenia is a common diagnostic category of such psychosis and typically reflects the presence of hallucinations and delusions in addition to severely disturbed thinking. Psychotics, like the poor, may be with us always, but the impression of many Micronesians is that mental illness is a growing problem in their societies, particularly among the young.

The problem of mental illness has not gone unattended, however. In 1974, the former Trust Territory Headquarters hired a staff psychiatrist and a clinical psychologist to run its flrst mental health program and to assist personnel in the districts. In subsequent years, as the Division of Mental Health took shape, the first serious attempt was made to collect data on mental illness in Micronesia. Drawing on the monthly reports submitted to headquarters for the years 1977-1980, Dr. Paul Dale, a former staff psychiatrist, published an article comparing rates of schizophrenia in different parts of Micronesia (Dale, 1981). Some of the data collected by Dale are presented in Table 1. Dale noted the relatively high rates of schizophrenia in Yap and Palau and the extremely low rates reported for some of the remote atolls. During the same time period, a psychiatric resident from Loma Linda Medical School conducted a 3-month study of schizophrenia in Palau that showed a heavy preponderance of male victims. He and his colleagues attributed the high male rate of illness in Palau to the "relatively recent and rapid disintegration of the traditional culture" (Kauders, MacMurray, & Hammond, 1982, p. 101) and to rampant drug use among men (Hammond, Kauders, & MacMurray, 1 983).

Table 1: Prevalence of Schizophrenia and Chronic Mental Illness in Micronesia as reported by Dale (1981)


(Age 15+)

# Cases

# Cases
Bipolar Psychosis

# Cases
Depressive Psychosis

Specific Rate
(per 1,000)

Specific Rate
& Psychosis 
(per 1,000)
Palau 8,264 63 5 10 7.6 9.4
Yap 6,009 40 1 9 6.7 8.3
Chuuk 18,279 40 1 3 2.2 2.4
Pohnpei 13,206 17 3 5 1.4 1.9
Kosrae 2,562 3 1 0 1.2 1.6
Marshalls 13,481 11 1 0 .8 .9
TOTAL 61,807 174 12 27 2.8 3.4

Note. Data compiled using DSM-II (American Psychiatric Association, 1968) diagnostic criteria and US Trust Territory mental health records. Data in columns 3, 4, and 5 are reprinted with permission from the Journa/ of Psychia~ric Research 16, Paul W. Dale, Prevalence of Schizophrenia in the Pacific Island Populations, 1981, Pergamon Press Ltd.

The epidemiological survey reported here attempts to explore further the patterns of schizophrenia and chronic psychosis using community-based case-finding methods rather than the more limited mental health case records of earlier psychiatric researchers in Micronesia. We present findings on some of the more striking distribution patterns of psychosis in Micronesia and raise pertinent questions about the relationship between these patterns and the features of the sociocultural environment of present-day Micronesia. We should note, however, that the methods of this survey are limited by the lack of psychiatric resources in Micronesia that are necessary to develop thorough and formal diagnostic work-ups of those people identifled as having severe mental illness. Therefore, this report might be best represented as an informal epidemiological survey that emphasizes community perceptions and schizophrenia. We review literature on epidemiological methods used in the study of schizophrenia, which indicates that a variety of diagnostic practices have been used when examining the occurrence of schizophrenia. In the context of this review, our use of community perceptions in the identiflcation of schizophrenia and chronic psychosis represents a "broad" deflnition of schizophrenia.


Researchers in the fleld of epidemiology study the presence of disease by using measures of mortality (death from disease) or measures of morbidity (the occurrence of disease), or both. Ongoing chronic conditions with low mortality rates–like schizophrenia–are usually studied by using measures of morbidity, including incidence, prevalence, and expectancy (Mausner & Kramer, 1985; Westermeyer, 1989). As shown in Table 2, different rates of occurrence of schizophrenia are provided by different measures. Incidence provides a measure of disease onset, or the development of new cases in a previously healthy population during a period of time (e.g., 1 year). Rates may be presented as crude rates–the ratio of new cases over the total population–or as speciflc rates, which allow further examination of disease dynamics within subgroups of a population (Mausner & Kramer, 1985). In the study of schizophrenia, the population at risk is often considered to be those in the age range of 15-54 (Jablensky et al., 1992) and schizophrenia morbidity rates are sometimes expressed as a speciflc rate: the number of identifled cases over the population at risk (age group 15-54). As Table 2 shows, the incidence of schizophrenia is typically less than 1 per 1,000 population, although variation in the rate can be obseNed.

Table 2: Liferafure Review of Incidence, Prevalence, and Expecfancy Rafes of Schizophrenia




.34/1000 .11-.75 Eaton,1985 review of 12 studies in 6 countries; higher rates in US (.30-.70) than elsewhere (.11-.25); crude rates
.44/1000 .11-2.26 Jablensky et al., 1992 review of 12 studies in 10 countries; mixed rates (both crude & specific)
3.3/1000 1.5-4.2 Cooper, 1978 review of 15 studies in 12 countries; crude rates
3.7/1000 .6-8.3 Eaton, 1985 review of 20 studies in 14 countries; point, period, & lifetime rates show only small differences; mixed rates
5.4/1000 1.4-17 Jablensky et al., 1992 review of 16 studies in 8 countries; mixed rates 
5.5/1000 1.9-17.9 Nakane et al., 1992 review of 16 studies in Japan
  .6-7.1 Karno & Norquist, 1989 point prevalence; review of 10 studies
  3.6-7.3 Karno & Norquist, 1989 3-6 month period prevalance; review of 4 studies
  2.7-7.0 Karno & Norquist, 1989 12-month period prevalence; review of 7 studies
  .9-11.0 Karno & Norquist, 1989 lifetime prevalence; review of 21 studies
11.3/1000 3.6-26.8 Jablensky, et al., 1992 review of 13 studies in 6 countries; mixed rates
  7.0-30.0 Cooper, 1978 <10 in non-US countries; > 10 in US with differences likely due to diagnostic criteria and bias

Prevalence rates refer to the presence of an illness in a population at a given time and includes both newly developing cases and ongoing cases of illness. For chronic conditions like schizophrenia, prevalence rates portray a better estimate of disease occurrence because at any given time there are more ongoing cases than newly developing cases. Consequently, as shown in Table 2, prevalence rates of schizophrenia are larger than incidence rates and reflect an occurrence of perhaps 5 cases per 1,000 population, again with rate variation present. It is possible to examine the prevalence of a disease during varying periods of time, and epidemiologists may elect to study point, period, or lifetime prevalence rates. Point prevalence refers to the rate of disease in a population at a particular point in time of short duration (e. g., 1 day or 1 week). Period prevalence reflects the disease rate over a longer period of time, typically 3, 6, or 12 months duration, although it is possible to specify other time periods. Finally, /ifefime preva/ence refers to the proportion of a given population that at a point in time either has an active illness or has had a history of the illness. Eaton (1985), in reviewing prevalence studies of schizophrenia, concludes that these different types of prevalence measures provide similar estimates of occurrence of schizophrenia, and he suggests that this is because of the chronic, but not fatal, nature of schizophrenia. The varying procedures and methods of data collection used across 20 studies did not affect the reviewed results signiflcantly, and Eaton reports that the obtained range of rates corresponded with those of earlier reviews.

Expectancy (also referred to as lifetime risk or morbid risk) is a theoretical statistic calculated from prevalence data (Reid, 1960) or incidence data (Jablensky et al., 1992) that reflects the probability of healthy individuals becoming ill with a specific disease if they survive or live through the relevant period of highest risk (ages 15-54 for schizophrenia). Such lifetime risk for schizophrenia is popularly referred to as Nabout 1 percent," which is statistically equivalent to a rate of 10 per 1,000 population, similar to the rates shown in Table 2. Cooper (1978) notes that expectancy rates for schizophrenia are often higher in the United States than in other countries (also see the Table 2 incidence comments for Eaton, 1985). This reflects a long-held diagnostic bias in American psychiatry toward a broad conceptualization of schizophrenia compared with the more restricted conceptualization found in other countries (Karno & Norquist, 1989). It is likely that much of the variation in the ranges of schizophrenia summarized in Table 2 is due to such diagnostic bias rather than to signiflcantly different patterns of occurrence across countries, cultures, and populations. Evidence for this conclusion comes from two recent, major, mental health epidemiology projects: the US National Institute of Mental Health (NIMH) Epidemiologic Cafchmenf Area Sfudy(Robins & Regier, 1991) and the World Healfh Organizafion (WHO) Ten-counfry sfudy–Schizophrenia: Manifesfafions, Incidence and Course in Differenf Culfures (Jablensky et al., 1992).

In the NIMH Epidemiologic Cafchmenf Area (ECA) Sfudy, coordinated and standardized epidemiologic data were collected on major mental disorders in the early 1980s across flve different sites in the United States (Robins & Regier, 1991). This has been the largest coordinated mental health epidemiological project in the United States to date and is the flrst large-scale project to use DSM-III (American Psychiatric Association, 1980) criteria in establishing the prevalence of psychiatric disorders. The 6-month period prevalence rate for schizophrenia in the ECA project ranged from a low of 4/1,000 (Los Angeles) to a high of 13/1,000 (Piedmont, NC) and averaged 8.8/1,000 (Burnham et al., 1987). These rates are high when compared with the prevalence rates in Table 2. Also, lifetime prevalence rates from the ECA study are higher, averaging 13/1,000 with a range from 6/1,000 to 19/1,000 (as reported in Karno & Norquist, 1989). These data suggest that the United States has a higher prevalence of schizophrenia than elsewhere, or that the ECA study employed more effective case-flnding procedures than other studies have, or that there is a continued diagnostic bias in the classiflcation of schizophrenia by American researchers. A recent World Health Organization study suggests that the last alternative is correct.

In the late 1960s and early 1970s, the World Health Organization undertook the International Pilot Study of Schizophrenia (Leff, Sartorius, Jablensky, Korten, & Ernberg, 1992) and examined patterns and course of schizophrenia in nine countries by following 1,200 patients over a 5-year period. This research established that coordinated multicenter cross-cultural research studies of schizophrenia could occur successfully and that schizophrenics across the cultures studied showed similar behavior patterns and symptoms. Subsequently, in the late 1970s, the World Health Organization undertook an even larger 12-site/10-country investigation of schizophrenia to provide basic epidemiologic information about schizophrenia across cultures using standardized case-flnding and diagnostic tools (Jablensky et al., 1992). The research design included identiflcation of flrst-time cases of schizophrenia and related psychoses during a 1-year period, thereby establishing annual incidence rates across the 10 countries. Potential cases were identifled by establishing extensive community links with traditional and nontraditional social and helping agencies, including practitioners of traditional medicine and religious healers. It was estimated that about 200 cases (out of the total study population of 1,379) would have been missed without the cooperation of traditional practitioners in case flnding. The research attempted to explore diagnostic bias and the effects of broad versus restricted conceptualizations of schizophrenia, and so a two-stage case-deflnition process occurred. First, the individuals included in the study by broad diagnostic criteria were those with "symptoms and signs which most psychiatrists would describe as psychotic or strongly suggestive of psychosis" (Jablensky et al., 1992, p. 84). Then further fllters were employed to narrow and identify those with the symptoms and syndromes that would be regarded as schizophrenic only by very strict criteria. Clear cases of affective disorder and cases with a clear basis for organic brain damage were excluded. A formal diagnosis was made of each case using both broad and narrow deflnitions of schizophrenia.

As reported by Jablensky et al. (1992), results of the research project showed that rates of schizophrenia were variable across countries when broad diagnostic criteria were used and that these rates were consistent with previously reported rates of the magnitudes shown in Table 2. Using broad diagnostic criteria, the annual incidence of schizophrenia across eight sites averaged .25/1,000 with a range of .15/1,000 (Denmark and Honolulu) to .72/1,000 (India). The researchers also calculated the average expectancy using broad criteria at 9.2/1,000 with a range of 5/1,000 (Honolulu) to 17.2/1,000 (India). When such broad diagnostic criteria were used, there were statistically signiflcant differences in the incidence and expectancy of schizophrenia across each study site. However, in a major flnding, there were no differences in the 1 0 rate of schizophrenia across cultures when narrow or restricted diagnostic criteria of schizophrenia were utilized. Using restricted criteria, an annual incidence rate of .10/1,000 with a range of .07/1,000 to .14/1,000 was observed and an expectancy rate of 3.8/1,000 with a range of 2.6/1,000 to 5.4/1,000. These rates are substantially lower than those reported in Table 2. These researchers conclude that there is a common "nuclear" schizophrenia that occurs across cultures, as evidenced by the similar rates observed when restricted diagnostic criteria are used. It seems, then, that diagnostic bias and varying epidemiologic methodologies are responsible for much of the observed variation in rates of schizophrenia, as presented in Table 2. Given the finding of this recent WHO International study, it would be premature at this time to suggest that differences in the rate of schizophrenia across cultures have been demonstrated.

Such recent interest in the occurrence of schizophrenia across cultures provides a timely context for us to examine the epidemiology of severe mental illness in Micronesia. The data on which earlier studies of mental illness in Micronesia drew are now more than 10 years old and were probably inadequate even then. Prevalence rates were almost certainly understated because reports were derived exclusively from patients who sought hospital treatment, and the reliability of even these data depended on the thoroughness of the reporting procedures in each island group. In an effort to provide a more up-to-date and comprehensive body of data on serious mental illness, the Micronesian Seminar undertook its own survey beginning in July 1988. We attempted to draw up a complete list of schizophrenics in the Republic of the Marshall Islands (RMI), Palau, and the Federated States of Micronesia (FSM). (Herein, we shall continue to refer to these three areas collectively as Micronesia.) Our sample population included people identifled in the community as having a severe thought disorder for a period of over 12 months and who, in our opinion, would likely receive a formal psychiatric diagnosis of schizophrenia or a related psychosis based on criteria from the DSM-III (American Psychiatric Association, 1980), if formal psychiatric diagnosis were to occur.

Psychosis has always had its folk explanations, and Micronesia is no different in this respect from any other part of the world. Many people, when asked about the problem of mental illness, offer a standard explanation for what they regard as a sharp increase in mental illness in the last 10 or 15 years. They point to travel abroad and high educational attainments as common characteristics of the psychotic population, implying that either or both may put such pressure on the individual as to cause a breakdown. They also attribute mental illness to the widespread use of alcohol and drugs, particularly marijuana, among the young. Because these factors are singled out so frequently by local people as causes of mental illness, the following survey also examines drug use habits, travel abroad, and level of education of our sample population.

METHOD General Paraprofessional researchers visited seven island centers (Chuuk, Ebeye, Koror, Kosrae, Majuro, Pohnpei, Yap) collecting information on the prevalence of severe mental illness throughout all inhabited islands of Palau, the FSM, and the RMI. Although outer islands surrounding each center were not visited, it was possible to identify individuals from each remote area presently living in the island centers and thereby gather information about cases from remote areas. Twelve-month period prevalence information was gathered at each location. Typically, from 1 to 3 months were spent at each island center during intensive data collection, with periodic follow-up over the following year. Because the Micronesian Seminar was located in Chuuk at the time of the study, particularly thorough community data were collected there over a 12to 24-month period. Information was gathered on those people with established histories of abnormal behavior, even if they were not experiencing an active psychotic episode at the time of the survey. Survey Information Biographical data were collected for each identified subject, including birth date, residence, educational background, religion, travel history, and drug use patterns. To this was added illness information on disease symptomology, date of onset, history and course, community attributions of cause of illness, and treatment history, including traditional versus Western medical approaches. When available, hospital record information, including psychiatric diagnosis, was obtained. Family information was collected on marital status, children, parents, siblings, and extended family. Additional information was obtained on family history of mental illness, early family history, family conflicts, continuing family care and support, and family attributions of illness. Defning Schizophrenia and Chronic Mental Illness Mental illness has various manifestations, and diagnosis is far from a simple matter, as anyone knows who has picked up a copy of DSM-III or DSM ~ R (American Psychiatric Association, 1980, 1987). In this paper we are concerned exclusively with what are generally called psychoses: "mental disorders in which impairment of mental function has developed to a degree that interferes grossly with insight, ability to meet some ordinary demands of life or to maintain adequate contact with reality" (American Psychiatric Association, 1980, p. 410). We have attempted to stake out as our area of research what would roughly correspond to the DSM-III and III-R categories 290-299, with the following exceptions. We excluded from the survey those types of psychoses that were primarily organic in nature: senile dementia, arteriosclerotic dementia, and psychotic conditions attributable exclusively to alcohol or drug abuse. Hence, the organic psychotic conditions described in DSM-III under the classifications 290-294 were excluded. Moreover, we made every effort to screen out the following: individuals suffering from personality disorders; individuals who were retarded from birth and those 1 4 whose mental problems stemmed from trauma or physical disabilities such as epilepsy; persons who suffer from rather severe but temporary problems, such as those who have undergone bouts of depression or those who have experienced somatoform or dissociative disorders, for example, "possession syndrome"; and those who display occasional odd behavior, but in the judgment of the community are capable of functioning rather normally. Additionally, the data in this study do not include borderline cases–that is, those in which the symptoms are not indisputably psychotic, at least by our perception, or in which the symptoms have not persisted for at least 1 year. The data, however, do include six psychotic individuals who died during the 1 8-month period in which this study was conducted and another three who have since died.

Although we had access to the formal psychiatric diagnosis for many of the subjects included in this study, we decided against exclusive reliance on Western medical norms in deflning psychosis. In determining whether a person for whom we had no formal diagnosis should be regarded as a psychotic, we adopted a loose communitybased norm. Our fleld workers asked informants whether the individual had been acting "crazy," at least intermittently, for a period of a year or longer. Through additional questions, we attempted to determine the nature of the symptoms and make a judgment on the severity of the person's mental illness. Case-Finding Mefhods Three methods of case flnding occurred in each island center. Initially, mental health 1 5 programs were contacted and information was gathered from case records and from community contacts provided by mental health staff. Second, extensive church contacts were utilized in flnding cases and developing further community networks. Finally, civil authorities, professionals, and paraprofessionals–school teachers, island magistrates, Peace Corps Volunteers–were contacted. The general model of case flnding was simply to ask repeatedly an expanding network of community members about any continuing (more than 12 months) strange, crazy, or inexplicable actions by people they knew of or knew.

Field researchers, conducting survey interviews in English for the most part, interviewed dozens of informants, often at great length, and laboriously pieced together the case histories of each of the psychotics. Each case was checked with several informants to verify the information obtained from interviews. By the end of the survey, a compact but detailed case flle had been assembled on the great majority of subjects. There remained a relatively small number, about 20 percent, who were established with reasonable certainty to be truly psychotic, but whose case flles had significant gaps. In keeping with the goal of achieving maximum comprehensiveness, the decision was made to include these subjects in the survey.

Calculafion of Prevalence Rafes The population statistics that were used to determine age- and sex-speciflc prevalence rates were based on the most recent census taken in each of the island groups, with 16 flgures representing the de facto indigenous population. Because the censuses for the Marshalls and Chuuk fell within the period of the survey, the sex-age data for the census years were used in unaltered form to calculate prevalence rates (Federated States of Micronesia, 1991, p.16; Republic of Marshall Islands, 1989, pp. 3~34). For Yap, unadjusted flgures from the 1987 census were used in the absence of more recent population projections by sex and age group (Yap State, 1988, Vol. I, p. 27). Unadjusted flgures from 1986 were used for Palau because the island group has shown virtually no population change over the last 20 years (Republic of Palau, 1987, p.18). Population projections for 1990 were derived for Pohnpei and Kosrae from the 1985 and 1986 census data, respectively, (Kosrae State, 1990, p. 104; Pohnpei State Government, 1990, p. 127) in the calculation of age- and sex-speciflc prevalence rates. REsuLTs The survey revealed a total of 445 individuals suffering from psychosis in Micronesia. Twenty-seven percent of the cases had not been receiving care from relevant mental health programs, and many of these had not come to the attention of mental health offlcers. Of those cases that had been treated by mental health staff, nearly three fourths (73 percent) had received a DSM-III diagnosis of schizophrenia in one of its many forms. What was diagnosed as bipolar psychosis (manic-depression) accounted for 13 percent of the cases. Acute paranoia comprised 5 percent of the cases, psychotic depression comprised 2 percent, and assorted other diagnoses comprised 7 percent. In 1 7 addition to the 445 people included in this report, case information was collected on another 136 people; however, their symptoms did not warrant inclusion in the most seriously mentally ill categories of schizophrenia and chronic psychosis.

As shown in Table 3, the overall speciflc rate (adjusted for the period of increased risk, i.e., aged 15 years and older) of schizophrenia and chronic mental illness throughout the islands surveyed is 5.4 psychotics per 1,000 adults. There are differences in prevalence rates of schizophrenia and chronic mental illness from one island group to another. Palau's combined male and female rate of 16.7 is by far the highest in the region. It is nearly twice the rate of Yap (8.4), which has the next highest rate, and more than four times the lowest observed rates of Chuuk and Pohnpei.

Table 3: Prevalence of Schizophrenia and Chronic Mental Illness in Micronesia

Location Males Females Total Males Females Total Males Females Total
Palau 3,960 3,709 7,666 89 39 128 22.5 10.5 16.7
Yap 3,220 3,069 6,289 19 14 53 12.1 4.6 8.4
Chuuk 11,946 12,544 24,490 78 14 92 6.5 1.1 3.8
Pohnpei 8,830 8,644 17,474 42 14 56 4.8 1.6 3.2
Kosrae 2,027 2,094 4,061 26 0 26 12.8 0 6.4
Marshalls 10,819 10,425 21,244 70 20 90 6.5 1.9 4.2
TOTAL 40,802 40,422 81,224 344 101 445 8.4 2.5 5.4

One of the most startling flndings in this survey was the great difference in male and female rates: The overall male/female ratio was a striking 3.4/ 1. For the area as a whole, 77 percent of the subjects in this survey were males, and the overall prevalence of schizophrenia for males was 8.4/1,000. The speciflc rate for males in Palau was 22.5/1,000, meaning that about 2 percent of the adult male population suffers from psychosis, and the male rates in Kosrae and Yap were over 10/1,000. The other three areas–Chuuk, Pohnpei, and the Marshall Islands–had male rates of 4.8-6.5.

Everywhere in Micronesia females seem to be less at risk for schizophrenia than males. The overall female rate in Micronesia was 2.5 per 1,000 adults. Despite the considerable variation in the speciflc rates of psychosis, ranging from zero in Kosrae to 10.5 in Palau, the female rates appear to be proportional to the total measured rates. Thus, Palau, which has the highest combined rate, also shows the highest female rate, while Chuuk and Pohnpei, areas with low combined rates, also have very low female rates. Women comprised 17 to 30 percent of the total psychotic population in each island group except Kosrae, where no females were recorded. (Kosrae, however, has had cases of psychosis among females in the past, including a woman with schizophrenic symptoms who apparently took her own life 10 years ago.)

As Table 4 shows, the mean age of Micronesian men was about 25 when they flrst experienced schizophrenia, but women were closer to 30. Additionally, the average age of females with schizophrenia and chronic mental illness at the time of our survey was signiflcantly higher than that of males throughout Micronesia. There was a difference of between 3 (Yap) and 13 (Chuuk) years in the mean ages of males and females in the various island groups, with an average age gap between genders of slightly more than 7 years. Hence, females with schizophrenia and chronic psychosis throughout Micronesia are not only much rarer than males, but they tend to show symptoms of the disease later than males do.

Table 4: Average Age at Onset of Schizophrenia and Chronic Mental Illness by Sex (and Average Age at Time of Survey)

                                                                       AT ONSET        AT SURVEY

Location Male Female Male Female
Palau 26.1 33.1 38.4 44.6
Yap 28.3 28.8 41.2 44.7
Chuuk 24.1 33.2 36.0 48.7
Pohnpei 25.2 27.0 35.0 43.3
Kosrae 23.8 35.5  
Marshalls 23.5 24.3 33.8 38.6
Overall 25.2 29.7 36.7 43.9

A more precise breakdown of these patterns is shown in Table 5, which shows speciflc rates of schizophrenia and chronic mental illness in 10-year age and gender groups across locations. This method of presentation controls for changing population demographics and allows direct comparison of prevalence across gender, age, and location. As can be seen, the highest rates of schizophrenia for women occur in either the 40-49 age group (Yap, Chuuk) or the 5~59 age group (Palau, Pohnpei, Marshall Islands), while the highest rates for men are in the lower 20-29 (Kosrae), 30-39 (Palau, Chuuk, Pohnpei, Marshall Islands), or 40-49 (Yap) age groups. Across the entire Table 5, there are only three instances of schizophrenia prevalence rates being higher for women than men, and this occurs in Palau (ages 40-49 and 50-59) and in the Marshall Islands (50-59). When examining the overall rate for females across age groups, the prevalence of schizophrenia generally increases with age; for men, prevalence rates decrease in older age. Over 5 percent of Palauan males in the age group of 30-39 suffer from schizophrenia or chronic psychosis (54.4/1,000).

Table 5: Age & Sex Adjusted Prevalence Rates (per 1,000) of Schizophrenia and Chronic Mental Illness in Micronesia

                     Palau               Yap               Chuuk          Pohnpei           Kosrae          Marshalls                Overall

Age M F M F M F M F M F M F Male Female M&F
10-19 0 0 0 0 0 0 .3 0 0 .2 0 .1 0 .05
20-29  18.1 5.2 7.5 1.2 8.0 0.3 5.4 1.5 28.4 0 5.9 1.8 8.7 1.4 5.0
30-39  54.4 16.8 20.6 4.7 11.4 0.7 9.1 0.6 11.4 0 12.5 2.6 16.0 2.8 9.6
40-49  17.8 18.8 24.5 19.2 7.2 4.9 3.7 1.2 3.5 0 8.4 1.7 9.2 5.6 7.4
50-59  21.6 22.8 18.1 11.4 2.9 0.9 3.6 2.7 12.0 0 0 5.9 6.0 5.6 5.8
60+ 15.0 7.0 8.2 2.6 4.1 2.3 3.5 2.1 11.9 0 0 0 5.0 2.3 3.6

The survey flgures on alcohol and drug use also show a strong differentiation along gender lines. As shown in Table 6,88 percent of the males have a history of moderate to heavy drug use, compared with only 36 percent of the females. The most commonly used drug, as might be expected, was alcohol; all but a veN few individuals of both sexes who had a drug history were at least occasional drinkers. The clear pattern of drug use found among the subjects of this suNey is one of accretion rather than substitution. Hence, all but 11 of the males and 3 of the females who smoked marijuana were also regular users of alcohol, and a much smaller number added other more potent drugs to their personal pharmacopoeia. This pattern, which seems to hold true for the Micronesian population at large, was conflrmed in a recent study of alcohol use in Chuuk (Marshall & Marshall, 1990). Only a few individuals in our suNey were known to have used heavier drugs than marijuana and alcohol. Only 44 males and 1 female, or 11 percent of the total sample, have a record of heavy drug use (e.g., heroin, cocaine, LSD, amphetamine). The medical histories of these individuals seem to suggest a relationship between their drug abuse and the onset of their mental illness, but their number is too small to be of signiflcance in this study.

Table 6:  Percentage of People with Schizophrenia and Chronic Mental Illness in Micronesia Who Regularly Use Drugs

Users of: # Males % Males # Females % Females # Both % Both
Any Drug 287 88 31 36 318 77
Alcohol 271 83 28 33 299 73
Cannabis 198 61 12 14 210 51
Other 44 13 1 1 45 11

Note: The total number of cases in Table 6 (400) is less than the total number of identified cases in Table 3 (445). This difference reflects the fact that it was not possible to obtain reliable drug use data for some cases.

Males: n = 326
Females: n = 74
Both: n = 400

Individuals with schizophrenia or chronic mental illness seem as a group to be above average in their educational attainments. In the four areas for which there were adequate census data to compare the mentally ill with the general adult population (the 1989 Chuuk census and the 1985 Pohnpei census do not provide detailed information on education), the level of formal education of our sample is higher than that of the general population. According to census documents, the average numbers of years of educational attainment for the adult populations of Palau, Yap, Kosrae, and the Marshall Islands are 9.7, 7.8, 8.8, and 8.7, respectively. But in our survey, the educational level for people from these areas was 10.0, 10.1, 10.7, and 9.3, respectively. The difference is over 2 years of schooling in Yap and nearly the same in Kosrae, a half year in the Marshalls, and a fraction of a year in Palau. Nearly one fourth (23 percent) of the subjects surveyed attended college for at least a year, and many spent 2 or 3 years

Our survey also suggests that individuals with schizophrenia and chronic mental illness in Micronesia are well traveled; almost half (47 percent) have lived for more than 6 months outside their own state or republic, and most of these have resided on Guam or in the United States. Unfortunately, there is no data for the general population against which this can be measured. DlscussloN This survey showed the average overall prevalence rate of schizophrenia and chronic mental illness in Micronesia to be 5.4/1,000, with a range from 3.2 to 16.7/1,000. The 22 rates of schizophrenia and chronic mental illness obtained in the study fall within the accepted prevalence ranges established across cultures for schizophrenia, as reviewed in Table 2. Rates varied greatly in different island groups, with Palau's rate almost twice that of Yap, the next highest area, and several times higher than rates in most of the eastern parts of Micronesia.

The patterning of mental illness reported in our research is similar to that reported by Dale (1981), as presented in Table 1, although we report higher rates of schizophrenia and psychosis in Kosrae and the Marshalls than Dale did. Additionally, the present prevalence rates are higher throughout Micronesia when compared with those presented by Dale. This may represent an increase in the prevalence of schizophrenia over the last 10 years or this may be an apparent increase only, due to (a) better case-finding methods utilized in our survey; (b) a "broader" deflnition of schizophrenia and chronic mental illness; or, most likely, (c) some combination of these factors. Only with continued epidemiologic research and careful monitoring of mental health patterns can trends in prevalence be revealed. Gender Differences Micronesian psychosis is heavily weighted toward males, who constitute 77 percent of the total sample and outnumber females by a ratio of 3.4/1. Additionally, the average age of illness onset occurs earlier for men than women by approximately 5 years and, within our sample, men with schizophrenia and chronic mental illness are younger than women by approximately 7 years. When examining age- and sex-speciflc prevalence rates, the prevalence of schizophrenia for women generally increases with age; for men, prevalence rates are highest at an earlier age. Finally, as discussed below, marked gender differences appear in drug use patterns as well.

Most of these patterns correspond to those typically seen in recent epidemiologic research on schizophrenia. Over the last 10 years a good deal of research has focused on several different areas in the study of gender and schizophrenia (Goldstein & Tsuang, 1990) and several conclusions seem to be emerging. Prior to becoming ill, men have a poorer history of adjustment and social functioning (Dworkin, 1990; Foerster, Lewis, Owen, & Murray, 1991) and may display either social withdrawal or aggressive behavior, or both. Men have an earlier onset of schizophrenia by about 5 years, with typical illness and flrst hospitalization for schizophrenia beginning at around age 25 (Hafner et al., 1989). Once identifled, and typically hospitalized, men have a poorer response than women to treatment, whether pharmacological (Seeman & Lang, 1990) or psychosocial (Spencer et al., 1988). Hence, men are more likely to spend longer periods in hospitals and are rehospitalized more often than women (Angermeyer, Goldstein, & Kuehn, 1989; Goldstein, 1988).

Social theories of gender differences in schizophrenia (which usually attempt to explain differences in age at onset and social functioning) tend to focus on sex roles and expectations, opportunities for social functioning, perceptions of illness as they affect 24 men versus women, and tendencies to use drugs and alcohol as exacerbating factors in the development of schizophrenia (Goldstein & Kriesman, 1988; Hafner et al., 1989; Salokangas, 1983). For example, Salokangas suggests that the earlier symptoms and poorer prognosis for men may be related to the generally more stressful conditions in which they live and work, while women, not as often in the work force, may survive in more protected family environments longer. Similarly, within families, Goldstein and Kreisman suggest that social perceptions are different for sons and daughters and that parents may be more protective and tolerant of daughters than sons. They suggest that "parents treat sons and daughters differently, in part influenced by social norms and expectations associated with gender" (p. 871).

Even though research suggests that age at onset is earlier for men, and treatment response seems to be worse for men, it is generally accepted that the cumulative risk of schizophrenia is the same for men and women (Hafner et al., 1989; Jablensky et al., 1992). In other words, over time there is a 1/1 illness ratio seen between men and women. In the World Health Organization's Infernationa/PilotStudy of Schizophrenia and subsequent Ten-Country StudJ/, for example, the male-female ratios of schizophrenic patients were 1/1 and 1.2/1, respectively (Jablensky et al., 1992). Our flndings of a schizophrenia and chronic psychosis male/female ratio of 3.4/1 represents a clear and unique difference from established patterns in the epidemiology of schizophrenia.

Perhaps our relatively broad deflnition of schizophrenia influenced the male/female ratio of cases. This is unlikely, however, because research into diagnostic criteria suggests that male rates increase rather than diminish as stricter diagnostic criteria are used (Lewine, Burbach, & Meltzer, 1984). In the recent World Health Organization study, which speciflcally examined diagnostic bias and the prevalence of schizophrenia (Jablensky et al., 1992), no signiflcant difference in the male/female ratio was found when diagnostic criteria were relaxed.

It is possible that our case-flnding methods selectively identifled male patients based on some easily identiflable community perspective, such as abnormally aggressive behavior, or that our methods selectively overlooked female patients. Extensive case biographies were developed through community interviews following initial case identiflcation, and decisions to include speciflc cases in the survey pool were based on the presence of speciflc psychotic symptomology rather than any other criteria. This suggests the possibility that our case-flnding methods might systematically exclude females, if for some reason they were not being identifled by community agents. There have been suggestions that many female psychotics go unrecognized because their svmDtomoloav. for cultural reasons. is muted. while that of males is exaaaerated. again for cultural reasons (Hezel, 1 987a). Although we cannot entirely discount this possibility, such radical attenuation of symptoms in a cognitive disorder like schizophrenia is highly improbable. It should be noted that despite the exhaustive 26 community suNey made in Chuuk, where each lagoon island was visited and a wide network of community and church leaders was mobilized to scrutinize the villages for unreported cases, the prevalence rate for female schizophrenia was one of the lowest in Micronesia. We must somehow account for what appears to be a real disparity in male and female rates of psychoses, which do not usually differ according to gender.

Because the genetic predisposition for the major psychoses is not sex linked, as far as we know, the preponderance of males suggests that perhaps environmental factors are responsible. These might include the exposure of males to role changes and other stresses, from which females are better protected by the culture. The sociocultural environment of Micronesia, affected as it has been by modernization and its concomitant changes in recent decades, would appear to be more stressful for many than ever before now. It is conceivable that this could have an effect on the overall rates of psychosis in Micronesia and its distribution among particular segments of these societies. The traditional patterns of social organization in Micronesian cultures tend to shelter women, conflning them to the home and keeping them from the public roles that bring men both greater satisfaction and greater stress. Anthropologists who studied Micronesian societies soon

after World War ll noted that the social pressures on men were greater and the supports fewer (Gladwin & Sarason, 1953; Lessa, 1951). This should be all the more true today if, as seems to be the case, men are subject to greater role changes and social dislocation than women (Hezel, 1987b).

Alcohol and Drug Use Drug use, particularly alcohol and marijuana, which is common among young men in Micronesia and infrequent among women, may contribute to the high rate of mental illness among males. Everywhere in Micronesia the use of alcohol and drugs such as marijuana is regarded as a mainly male activity, and drug use has become a culturally sanctioned means by which males seek relief from stress. The strong differentiation along sex lines with respect to drug use is too great to be ignored when we cast about for an explanation of the high rate of male psychosis. A relationship between drugs and psychosis has been suggested before, as when Hammond and colleagues hypothesized that substance abuse results "in increased morbidity and contributes to the observed male-predominance among Palauan schizophrenics" (Hammond et al., 1983, p. 168).

A linkage of drug use to schizophrenia and similar diseases could offer an explanation for the wide variation in the morbidity rates of these diseases by cultural and geographical area. Palau and Yap, which show the highest rates of chronic mental illness, would also appear to be among the islands in which drug use is heaviest. It is no secret that during the 1 980s Palau experienced a drug problem, including the use of heroin and cocaine, that was unparalleled in other parts of Micronesia. Although there are no accurate means of gauging comparative prevalence of drugs in different island groups, the following figures furnish a rough index to the relative amount of alcohol 28 consumed in these places. Trust Territory import flgures for 1977 (the latest year available) indicate that the per capita annual expenditure on alcohol in Yap ($66) was about three times higher than in Pohnpei ($21) and the Marshall Islands ($24), while Palau's per capita expenditure ($55) was twice that of Pohnpei and the Marshall Islands (Hezel, 1981). The flgures for Chuuk and Kosrae were unrecorded.

How does the prevalence of alcohol and drug use in our sample compare with that in the general population? Statistical data needed for such comparison is lacking for most parts of Micronesia, but a large representative sample of the Chuukese people offers some comparative flgures for that island group. In a 1985 survey, 86 percent of the general male population in Chuuk were either former or current users of alcohol (Marshall, 1991, p. 339); this flgure is similar to our flnding that 83 percent of the male population had a history of alcohol consumption. The 61 percent rate of marijuana use among males with mental illness, however, is more than twice the 27 percent use rate recorded for all males over the age of 15 in Chuuk (Marshall, 1991, p. 339). Female rates for alcohol and marijuana use are much higher in our study than in the general Chuukese population, where only 2.3 percent ever drank alcohol and 1.2 percent ever smoked marijuana.

The influence of substance abuse in the development of schizophrenia is controversial. Medical research indicates that some drugs (such as hallucinogens, amphetamine, and alcohol during the withdrawal stage) can cause psychotic-like reactions in normal people, while other drugs (e.g., cannabis and amphetamine) can produce paranoia (as reviewed by Mueser et al., 1990). These studies indicate that the use of some drugs can produce psychotic-like symptoms, although the drugs in question, with the exception of cannabis, are not those widely used by the psychotic population in Micronesia.

There is a small body of research that suggests that drug abuse may be a contributing factor in the development of schizophrenia. Dixon, Haas, Weiden, Sweeny, and Frances (1990) review a study by Andreasson, Engstrom, Allebeck, and Rydeberg (1987) that suggests that heavy-users of cannabis have a higher risk of developing schizophrenia than nonusers. Other studies have not observed such a relation (Leiberman & Bowers, 1990; Test, Wallisch, Allness, & Ripp, 1989). The US NIMH Epidemiologic Cafchmenf Area Sfudyrevealed similar prevalence rates of schizophrenia among males and females despite the fact that males were observed to have twice the rate of drug abuse and more than four times the rate of alcohol abuse of females (Robins & Regier, 1991). A review of 20 studies of substance abuse and schizophrenia by Mueser et al. (1990) underscores a pattern of rather high drug abuse in schizophrenic patients, but a National Institute of Drug Abuse (1987) household survey in the United States suggests that the rates of abuse among schizophrenics may be no higher than among the general population (as reported by Leiberman & Bowers, 1990).

Whatever the controversies surrounding the attempts to link the incidence of 30 schizophrenia with alcohol and drug use, there is evidence that drug use may affect the course of the disease and its severity. Several studies, as reviewed by Dixon et al. (1990), indicate that drug use may be correlated with more rapid precipitation of schizophrenia and an earlier onset age for the disease. Other research suggests that alcohol and cannabis use can have negative effects on the level at which schizophrenics function and the outcome of the disease. Negrete, Knapp, Douglass, and Smith (1986), for example, reported a correlation among schizophrenic outpatients between the frequency of hallucinations and delusions and the use of drugs. Alcohol use, even in relatively small amounts, may also exacerbate schizophrenic symptomology (Drake, Osher, & Wallach, 1989, as reported by Drake et al., 1990). Thus, recent research suggests that even if drug use cannot be cited as clearly a factor in causing schizophrenia, it may produce psychotic-like symptoms in heavy-users who are not afflicted with the disease and it may aggravate the course of the disease in genuine schizophrenics.

Conclusion This epidemiological survey raises a number of critical questions that warrant further study by epidemiologists, psychological anthropologists, and psychiatric researchers. The striking disparity between male and female prevalence rates of schizophrenia and chronic mental illness requires further and careful research. Also, the signiflcant variation in rates of illness from one island group to another demands explanation. Even if a more thorough epidemiological study should show a much lower and consistent rate of schizophrenia, narrowly deflned, in males and females and throughout island groups, the preponderance of males afflicted by psychotic-like symptomology would remain a mystery. At present it is unclear whether explanations for the present patterns can be explained by the signiflcantly higher rates of drug abuse among males, by greater stress levels among males relative to females in these cultures, by both of these, or by other as yet undetermined factors. Acknowledgments The authors would like to acknowledge the contributions of Otong Emilio, Mariano Marcus, and Marty Doyle, the people who did the fleld research for this survey. Otong Emilio, the principal fleld researcher and major contributor to the data, died after the completion of this project. We wish to dedicate this article to his memory. The authors are also grateful to the Catholic Conference for Justice and Development (CCJD), a program of the Paciflc Catholic Bishops' Conference, for their funding assistance for this project.


American Psychiatric Association. (1968). Diagnosfic and sfafisfical manual of menfal disorders. Washington, DC: American Psychiatric Association.

American Psychiatric Association. (1980). Diagnosfic and sfafisfical manual of menfal disorders (DSM-III) (3rd ed.). Washington, DC: American Psychiatric Association.

American Psychiatric Association. (1987). Diagnosfic and sfafisfical manual of menfal disorders ~DSM ~ R) (3rd ed. rev.). Washington, DC: American Psychiatric Association.

Andreasson, S., Engstrom, A., Allebeck, P., & Rydeberg, U. (1987). Cannabis and schizophrenia: A longitudinal study of Swedish conscripts. Lancef, Il, 1483-1486.

Angermeyer, M. C., Goldstein, J. M., & Kuehn, L. (1989). Gender differences in schizophrenia: Rehospitalization and community survival. PsychologicalMedicine, 19, 365-382.

Burnham, M. A., Hough, R. L., Esocbar, J. I., Karno, M., Timbers, D. M., Telles, C. A., & Locke, B. Z. (1987). Six-month prevalence of specific psychiatric disorders among Mexican Americans and non-Hispanic whites in Los Angeles. Archives of General Psychiafry, 44, 687-694.

Cooper, B. (1978). Epidemiology. In J. K. Wing (Ed.), Schizophrenia: Towards a new synfhesis (pp. 31-51). New York: Grune & Stratton.

Dale, P. W. (1981). Prevalence of schizophrenia in the Paciflc Island populations of Micronesia. Journal of PsychiafricResearch, 16, 103-111.

Dixon, L., Haas, P. W., Weiden, P., Sweeny, J., & Frances, A. (1990). Acute effects of drug abuse in schizophrenic patients: Clinical observations and patients' self-reports. Schizophrenia Bullefin, 16, 69-79.

Drake, R. E., Osher, F. C., Noordsy, D. L., Hurlbut, S. C., Teague, G. B., & Beaudett, M. S. (1990). Diagnosis of alcohol use disorders in schizophrenia. Schizophrenia Bullefin, 16, 56-67.

Drake, R. E., Osher, F. C., & Wallach, M. A. (1989). Alcohol use and abuse in schizophrenia: A prospective community study. Journal of Nervous and Menfal Disease, 177, 408-414.

Dworkin, R. H. (1990). Patterns of sex differences in negative symptoms and social functioning consistent with separate dimensions of schizophrenic psychopathology. American Journal of Psychiafry, 147, 347-349.

Eaton, W. W. (1985). Epidemiology of schizophrenia. EpidemiologicReviews, 7, 1 05-126.

Federated States of Micronesia (FSM). (1991). Chuuk Sfafe 1989 Census Reporf. Division of Statistics, Office of Planning and Statistics, Kolonia, Pohnpei State.

Foerster, A., Lewis, S., Owen, M., & Murray, R. (1991). Pre-morbid adjustment and personality in psychosis: Effects of sex and diagnosis. Brifish Journal of Psychiatry, 158, 171-176.

Gladwin, T., & Sarason, S. (1953). Truk. Man in paradise. Viking Fund Publications in Anthropology, No. 20. New York: Wenner-Gren Foundation.

Goldstein, J. M. (1988). Gender differences in the course of schizophrenia. American Journal of Psychiafry, 145, 684-689.

Goldstein, J. M., & Kreisman, D. (1988). Gender, family environment, and schizophrenia. PsychologicalMedicine, 18, 861-872.

Goldstein, J. M., & Tsuang, M. T. (1990). Gender and schizophrenia: An introduction and synthesis of flndings. Schizophrenia Bullefin, 16, 179-183.

Hafner, H., Riecher, A., Maurer, K., Loffler, W., Munk-Jorgensen, P., & Stromgren, E. (1989). How does gender influence age at flrst hospitalization for schizophrenia? A transnational case register study. PsJ~chological Medicine, 19, 903-91 8.

Hammond, K. W., Kauders, F. R., & MacMurray, J. P. (1983). Schizophrenia in Palau: Adescriptive study. InfernafionalJournalofSocialPsychiafry, 24, 161-170.

Hezel, F. X. (1981, November). Youfh drinking in Micronesia. Paper presented at a seminar on alcohol use and abuse among Micronesian youth, Kolonia, Pohnpei.

Hezel, F. X. (1 987a). In search of the social roots of mental health pathology in Micronesia. In A. Robillard & A. Marsella (Eds.), Contemporary issues in mental health research in fhe Pacifc Islands (pp. 12-31). Honolulu: University of Hawai'i Press.

Hezel, F. X. (1987b). The dilemmas of development: The effects of modernization on three areas of island life. In S. Stratigos & P. Hughes (Eds.), The efhics of developmenf: The Pacifcin fhe 21st Century(pp. 60-74). Port Moresby: UPNG Press.

Jablensky, A., Sartorius, N., Ernberg, G., Anker, M., Korten, A., Cooper, J. E., Day, R., & Bertelsen, A. (1992). Schizophrenia: Manifestations, incidence, and course in different cultures. A World Health Organization Ten-Country Study. Psychological Medicine, Monograph Supplement 20, 1-97.

Karno, M., & Norquist, G. S. (1989). Schizophrenia: Epidemiology. In H. Kaplan & B. Sadock (Eds.), Comprehensive texfbook of psychiafrJ/ (5th ed.) (pp. 699-705). Baltimore: Williams and Wilkens.

Kauders, F. R., MacMurray, J. P., & Hammond, K. W. (1982). Male predominance among Palauan schizophrenics. Infernafional Journal of Social Psychology, 28, 97-1 02.

Kosrae State. (1990, March). Kosrae Sfafe sfafisficalJvearbook, 1990. Bureau of Planning and Statistics, Office of Budget and Planning.

Leff, J., Sartorius, N., Jablensky, A., Korten, A., & Ernberg, G. (1992). The international pilot study of schizophrenia: Five-year follow-up flnding. Psychological Medicine, 22, 1 31 -145.

Lessa, W. A. (1951). The ethnography of UlifhiAfoll. Unpublished doctoral dissertation, University of California, Los Angeles.

Lewine, R., Burback, D., & Meltzer, H. (1984). Effect of diagnostic criteria on the ratio of male to female schizophrenic patients. American Journal of Psychiafry, 141, 8~87.

Lieberman, J. A., & Bowers, M. B., Jr. (1990). Substance abuse comorbidity in schizophrenia: Editors' introduction. Schizophrenia Bullefin, 16, 29-30.

Marshall, M. (1991). Beverage alcohol and other psychoactive substance use by young people in Chuuk, Federated States of Micronesia (Eastern Caroline Islands). In ConfemporaryDrugProblems(pp. 331-371). Washington, DC: Federal Legal Publications.

Marshall, M., & Marshall, L. B. (1990). Silenf voices speak. Women andprohibifion in Truk. Belmont, CA: Wadsworth.

Mausner, J. S., & Kramer, S. (1985). Mausner and Bahn. Epidemiology. An infroducfory fexf (2nd ed.). Philadelphia: W. B. Saunders.

Mueser, K. T., Yarnold, P. R., Levinson, H. S., Bellack, A. S., Kee, K., Morrison, R. L., & Yadalam, K. G. (1990). Prevalence of substance abuse in schizophrenia: Demographic and clinical correlates. Schizophrenia Bullefin, 16, 31-56.

Nakane, Y., Ohta, Y., & Radford, M. H. B. (1992). Epidemiological studies of schizophrenia in Japan. Schizophrenia Bullefin, 18, 75-84.

National Institute of Drug Abuse. (1987). Nationalhousehold surveyon drugabuse. Populafion esfimafes 1985. DHHS Pub. No. (ADM) 87-1539. Rockville, MD: The Institute.

Negrete, J. C., Knapp, W. P., Douglas, D. E., & Smith, W. B. (1986). Cannabis affects the severity of schizophrenic symptoms: Results of a clinical suNey. Psychological Medicine, 16, 51 5-520.

Pohnpei State Government. (1990). PohnpeiSfafe sfafisficsyearbook, 1990. Office of Budget, Planning, and Statistics, Kolonia, Pohnpei.

Reid, D. D. (1960). Epidemiological mefhods in fhe sfudy of menfal disorders. Geneva: World Health Organization.

Republic of the Marshall Islands (RMI). (1989). Census of populafion and housing 1988, fnal reporf. Office of Planning and Statistics, Majuro, Marshall Islands.

Republic of Palau. (1987). 1986 census of populafion and housing of fhe Republic of Palau (BelauJ. Office of Planning and Statistics, Koror, Palau.

Robins, L. N., & Regier, D. A. (1991). PsychiafricdisordersinAmerica. The epidemiologic cafchmenf area sfudy. New York: Free Press.

Salokangas, R. K. R. (1983). Prognostic implications of the sex of schizophrenic patients. Brifish JournalofPsychiafry, 142, 145-151.

Seeman, M, V., & Lang, M. (1990). The role of estrogens in schizophrenia gender differences. Schizophrenia Bulletin, 16, 185-194.

Spencer, J. H., Glick, I. D., Haas, G. L., Clarkin, J. F., Lewis, A. B., Peyser, J., DeMane, N., Good-Ellis, M., Harris, E., & Lestelle, V. (1988). A randomized clinical trial of inpatient family intervention, Ill: Effects at 6-month and 18-month follow-ups. American Journal of Psychiatry, 145, 1115-1121.

Test, M. A., Wallisch, L. S., Allness, D. J., & Ripp, K. (1989). Substance use in young adults with schizophrenic disorder. Schizophrenia Bulletin, 15, 465-476.

Yap State. (1988). Reporton the 1987 Yap State census of population (Vols. I ~ ll). Office of Planning and Statistics, Colonia, Yap.

Westermeyer, J. (1989). Psychiatric epidemiology across cultures: Current issues and trends. Transcultural Psychiatric Research Reviews, 26, 5-25.