Estimating rates of chronic fatigue syndrome from a ... - Springer Link

American Journal of Community Psychology, Vol. 23, No. 4, 1995

Estimating Rates of Chronic Fatigue Syndrome from a Community-Based Sample: A Pilot Study Leonard A. Jason, 2 Renee Taylor, Lynne Wagner, Jay Holden, and Joseph R. Ferrari DePaul University

Audrius V. Plioplys and Sigita Plioplys Mercy Hospital and Medical Center

David Lipkin Northwestern Medical School

Morris Papernik Rush Presbyterian-St. Luke's Medical Center

Most of the Chronic Fatigue Syndrome (CFS) epidemiological studies have relied on physicians who refer patients having at least 6 months of chronic fatigue and other symptoms. However, there are a number of potential problems when using this method to derive prevalence statistics. For example, some individuals with CFS might not have the economic resources to access medical care. Other individuals with CFS might be reluctant to use medical personnel, particularly if they have encountered physicians skeptical of the authenticity of 1The authors appreciate the generous financial support of the CFIDS Association and Minnan, Inc. We also are grateful for many helpful suggestions provided by Judith A. Richman, William McCready, Wendell Richmond, and Stephen E. Goldston. Finally, we are very thankful to the many undergraduate volunteers who helped us complete this study, and they include Cheryl Stenzel, Georgina DeLaTorre, Vickie Chrisos, Don Banik, Hillary Loeb, Leslie Ramesack, Ellen Diamond, Rob Bleeker, Gib Garza III, Mindy Silverstein, Kadip Sen, Meredith Lombrazo, Brian Miller, Caroline Marsden, Lorraine Whitmore, Stan Gayot, Karen Costakis, and Dave Sobotka. 2All correspondence should be sent to Leonard A. Jason, Department of Psychology, DePaul University, 2219 N. Kenmore Avenue, Chicago, Illinois 60614. 557 0091-0562/95/0800-05575ff'/.50/09 1995PlenumPublishingCorporation

558

Jason et al.

their illness. In addition, physicians that are skeptical of the existence of CFS might not identify cases. In the present pilot study, a random community sample (N = 1,031) was interviewed by telephone in order to identify and comprehensively evaluate individuals with symptoms of CFS and those who self-report having CFS. Different definitions of CFS were employed, and higher rates (0.2%) of CFS were found than in previous studies. Methodological benefits in using more rigorous epidemiological methods when estimating CFS prevalence rates are discussed. KEY WORDS: ChronicFatigue Syndrome;epidemiology;communitysample; immunological disorders.

Chronic Fatigue Syndrome (CFS) appears to be a heterogeneous disease syndrome that may be triggered by a variety of factors (Jason et al., 1995). The most prominent symptom is persistent fatigue for at least 6 months. Other symptoms commonly reported include mild fever or chills, sore throats, painful lymph nodes, unexplained generalized muscle weakness, muscle discomfort, prolonged generalized fatigue after levels of exercise, generalized headaches, joint swelling, neuropsychiatric complaints, and sleep disturbances (Bell, 1991). Relatively few patients are cured, with most continuing to experience decreases in exercise tolerance, and mild symptoms related to their CFS (Wilson et al., 1994). There have been many uncontrolled reports of the prevalence of CFS (Shafran, 1991). To date, most CFS epidemiological studies have been based on physician referrals from hospital and community-based clinics. Using this approach, Lloyd, Hickie, Boughton, Spencer, and Wakefield (1990) found the CFS prevalence rate to be 39.6 cases per 100,000; whereas the Centers for Disease Control (CDC) study by Gunn, ConneU, and Randall (1993) found the prevalence rates of CFS to range from 2.0 to 7.3 individuals per 100,000 cases. There are several problems with CFS epidemiologic approaches that involve physician referrals (Richman, Flaherty, & Rospenda, 1994). For example, medical sociological studies have indicated that many low-income individuals do not have access to the health care system (Mechanic, 1983); consequently, it would be inappropriate to estimate prevalence estimates solely from treatment facilities that represent a biased sample of the population (Dutton, 1986). In addition, the negative stigma of this illness, perceived to be a psychiatric disorder by some, probably inhibits some afflicted individuals from seeking help (Goffman, 1961). Of most importance, because many physicians do not believe that CFS exists (Jason et al., 1995), these physicians do not make referrals to these studies, and consequently this results in an underrepresentation of the CFS prevalence rates.

Chronic Fatigue Syndrome Epidemiology

559

Several other CFS epidemiologic studies have also been conducted, but each has significant methodological problems that compromise there ability to estimate prevalence rates. Shortcomings in these studies include estimating rates from tertiary care centers where patients are more severely ill or are more highly motivated in seeking help (Bates et al., 1993), using nonstandard definitions of CFS (Price, North, Wessely, & Fraser, 1992), sampling only individuals with access to health care (Buchwald et al., 1995), and using selfreport data without a needed medical evaluation (Pawlikowska et al., 1994). (See Jason et al., 1995, for a more comprehensive review.) By 1990, a consensus had emerged that CFS was a relatively rare disorder, affecting at most about 10 to 20 thousand individuals in this country (Price et al., 1992). When a new disease syndrome emerges, such as CFS, studies on prevalence can shape public policy, for if few individuals are affected by the syndrome, less federal and state resources might be required for research, prevention, and intervention efforts. Community psychologists have key roles to play in epidemiologic research with new disease syndromes. We have particular training and perspectives on how scientific activities can be negatively influenced by biases and unexamined assumptions. For example, if CFS epidemiologic studies are methodologically flawed because of their reliance on physician referrals, findings that underestimate the prevalence of CFS could minimize the seriousness of this illness. In such a situation, we could participate in assembling multidisciplinary ~teams to conduct more rigorously designed research that better estimates the prevalence of CFS. As a consequence, we could help set the agenda for how CFS epidemio!ogic research is conceptualized and researched, and such activities could have influential roles in helping shape public policy. In 1991, a team of researchers in Chicago, representing diverse disciplines (including epidemiology, psychiatry, medicine, immunology, sociology, biostatistics, and c o m m u n i t y psychology), began conducting preliminary studies, validating measurement approaches, and writing a grant proposal to conduct a CFS community-based epidemiologic study. A central value in community psychology is collaboration and team building. By bringing together scientists from different disciplines, we believed that better research methods could be developed that would allow us to more accurately and sensitively estimate prevalence rates for CFS. Our group initially investigated CFS in a high risk group of nurses (Jason, Taylor, et al., 1993). Out of a sample of 3400, 202 nurses (6%) indicated that they had experienced debilitating fatigue for 6 months or longer, and 23 nurses indicated that they had been diagnosed with CFS (or 680 per 100,000). In our next study (Jason, Fitzgibbon, et al., 1993), we found that it was possible to reach the majority of a sample of individuals with CFS through the use of telephone calls. Despite the fact that

560

Jason et al.

a high percentage of these individuals have answering machines that often screen their calls, it was possible to reach 93% of the sample within seven attempts. The study by Jason, Fitzgibbon, et al. (1993) indicated that repeated telephone calls can be used effectively to contact the majority of CFS cases. A community-based epidemiological study using telephone interviews may clarify the discrepant CFS prevalence rates reported in the literature. Quite possibly the sole reliance on physician referrals has led to an underrepresentation of cases. The present pilot study, then, is the first exploratory community-based epidemiological CFS study that has been conducted.

METHOD

Survey Questionnaire and Training A three-page screening questionnaire was developed and used to interview 1,031 people across five months. The names were obtained from Survey Sample Incorporated. This company generated randomly selected Chicago telephone numbers, using both listed and unlisted numbers. In addition, telephone numbers sampled from three Chicago zip code areas were selected to reflect the ethnic and racial distribution of the city. Thus, we contacted by telephone 344, 344, and 343 adults from three different geographical locales within the Chicagoland area. Every adult, 18 years of age or older, who lived at the residence and was at the house when the initial telephone call was made was included in the sample. There were an average of 1.34, 1.42, and 1.38 adult household members who were interviewed in the homes of interviewees within the three areas, respectively. Demographic data were obtained for the entire sample (e.g., age, ethnicity, socioeconomic status, marital status). Those respondents that indicated that they had experienced 6 months or more of significant debilitating fatigue then were interviewed further to assess their symptoms in order to determine whether or not they might have CFS. Ten interviewers with past experience in surveying were recruited. One Spanish-speaking interviewer was recruited. Training of these individuals took part in three phases, the first involved an orientation to the study. In this phase, all interviewers were familiarized with the survey questionnaire, phone logs, random phone-number lists, and survey procedures. The second phase of the training involved a n intensive role-playing session in which each interviewer participated in mock interviews, received feedback concerning his\her performance, and then role-played other interviews with criticisms and suggestions in mind. The final training phase involved indi-


561

vidual supervision of each interviewer during their first 10% of interviews with immediate feedback provided. Procedures adopted by W. McCready (personal communication, Oct. 27, 1992) were used to increase response rates. For example, to decrease the number of people who might immediately hang up, respondents were told that the interviewer was from a university and no products were being sold. In the standard introduction, the respondents were assured confidentiality. The interviewers also explained that the results of the study referred only to the group, not to individual participants. The respondents were told that the study involved the topic of fatigue. Telephone calls were made on Monday-Friday, from 1:00 p.m. to 8:00 p.m., and on Saturday and Sunday from 11:00 a.m. to 8:00 p.m. If the interviewer continued to reach an answering machine after seven attempts, a message was left on the eighth call giving the standard introduction and requesting the person call the survey group in order to be interviewed. The response rate was calculated by dividing the number of completed interviews by the number of adults with whom contact was attempted, either successfully or unsuccessfully (e.g., refusals, failure to return our phone call). Phase 2

Those respondents who either self-reported having CFS (i.e., had no exclusionary medical illnesses and met minimum symptom severity criteria) or had most of the symptoms of CFS were considered for Phase 2 of the study. These individuals were contacted by follow-up telephone call and asked to participate in the second phase of the study. Participants were told that the study would attempt to determine the causes of their fatigue. They also were told that participation would earn them $50 for their involvement. Individuals agreeing to participate first completed a psychiatric survey by telephone, then they were scheduled for a psychiatric assessment, a blood sampling, and finally a medical examination and history. Impaired subjects were offered the option of having a research assistant pick them up at their home and help them get to the study site. A trained advanced clinical psychology graduate student administered the Diagnostic Interview Schedule (DIS) to assess psychiatric comorbidity (Robins & Regier, 1991). Phase 3

After the selected individuals had been interviewed on the DIS, they were psychiatrically evaluated by a psychiatrist. Participants were asked to

562

Jason et ai.

sign a medical release form so that medical data from their primary physician could be reviewed. These individuals were then medically examined by one of two physicians. The laboratory tests were the minimum necessary to rule out other illnesses (Schluederberg et al., 1992). Participants with abnormalities were referred to their primary care physician. If a participant did not have a primary care physician, a referral to a physician was provided, and copies of the test data were forwarded to that physician. A team of two physicians and one psychiatrist were responsible for making final CFS diagnoses. Four definitions of CFS were used, two from the United States (Holmes et al., 1988, and a second with modifications by Schluederberg et al., 1992), one from Great Britain (Sharpe et al., 1991), and one from Australia (Lloyd et al., 1990). The United States definitions require at least 6 months of persisting or relapsing, debilitating fatigue, and eight or more minor symptoms such as sore throats and mild fevers. The British and Australian definitions are less restrictive as they do not require eight or more minor symptoms. RESULTS Only 11% of our sample of 1,031 adults from three Chicago area areas were refusals to be interviewed. Demographics information for the total sample is shown in Ihble I. Among the interviewees within the three areas, there was a significant difference in age, F(2) = 23.00, p < .01; sex, X2(2, N = 868) = 11.05, p < .01; race, X2(8, N = 807) = 479.58, p < .01; education, X2(8, N = 785) = 317.39, p < .01; marital status, Z2(4, N = 798) = 56.23, p < .01; and occupation, X2(8, N = 692) = 167.20, p < .01. A comparison by respondents indicates that one area comprised mostly African Americans (and there were more females in this area, and they had the highest completion rate for agreeing to be interviewed, 99%), another area was predominately Caucasian (respondents in this area had a higher educational level and a higher percentage were unmarried, with a completion rate of 88%), and the third area was a mixture of African American, Caucasian, and Asian Americans (respondents from this locale had more unskilled workers, and they had the lowest completion rate of 78%). Only 5% of the sample, or 53 individuals, indicated that they had unexplained severe fatigue for 6 months or more (see Tables I and 1I). 3 Six 3After the study was completed, a new definition of CFS was published (Fukuda et al., 1994), and this definition is less restrictive than the previous CDC definition, as it requires 4 or more rather than 8 or more symptoms. According to Fukuda (personal communication, August 30, 1995), the first author of the new CFS definition, in CFS epidemiologic research, individuals with psyeholosocial, psychological stress, and many psychiatric reasons for their fatigue should not be excluded. Using these less restrictive criteria, the phase 1 screening questionnaire would have identified 22 of the 53 fatigued individuals as possibly having CFS.


563

Table I. Demographics of Participants Total (n = 913) Gender Women Men Age M years

Fatigued subjects (n = 53)

Self-report CFS (n = 13)

n

%

n

%

n

%

501 367

58 42

36 17

68 32

10 3

77 23

42

40

36

Race African American Asian American Caucasian Latino Other

287 51 386 71 12

36 6 48 9 1

17 0 27 4 2

34 0 54 8 4

3 0 7 2 1

23 0 54 15 8

Education Less than high school High school graduate Partial college College degree Graduate degree

137 131 157 209 151

17 17 20 27 19

12 6 15 12 8

23 11 28 23 15

4 1 4 3 1

31 8 31 22 8

Marital status Never married Married Separated Divorced Widowed

328 287 23 63 97

41 36 3 8 12

26 11 2 10 4

49 20 4 19 8

9 2 1 1 0

69 15 8 8 0

Occupation Student Unskilled worker Skilled worker Not employed Retired

48 107 319 102 116

7 15 46 15 17

5 8 19 10 7

10 16 39 20 14

3 2 5 3 0

23 15 39 23 0

interviewees (4 w o m e n , 2 m e n ; M age = 36 years) felt that they h a d or might have h a d CFS in the past. Five were C a u c a s i a n a n d o n e was A f r i c a n A m e r i c a n . T h e o n e m i n o r i t y in this small sample, who c l a i m e d t h a t she h a d CFS in the past, was a physician, a n d we could n o t find any alternative r e a s o n for h e r fatigue. Reviewing the medical records of these six individuals, o n e p e r s o n ' s fatigue was d u e to a n e m i a , a n d a n o t h e r p e r s o n ' s fatigue was d u e to depression. F o r four individuals, we could n o t find a n alternative e x p l a n a t i o n for their fatigue. Because these individuals h a d recovered, we were n o t able to diagnose these individuals. T h e r e were 4 individuals (3 women, 1 man; M age = 32 years) who did n o t indicate that they had CFS, b u t their responses to the initial q u e s t i o n n a i r e

564

Jason et ai.

indicated that they had significant fatigue and many of the other CFS symptoms. These individuals were medically and psychiatrically evaluated. Three were Caucasian and one was African American. Three did not have a physician overseeing their medical condition. Following the medical and psychiatric evaluations, none of these subjects were evaluated as having CFS, one had a medical problem (anemia) and three had psychiatric disorders (two cases of depression, one case of conversion disorder and depression). Eleven interviewees stated that their current fatigue was due to CFS, and two others stated that CFS might be accounting for their fatigue. Data from their surveys are shown in Tables I and II. During inspection of medical records, it was apparent that two of these subjects had exclusionary medical diagnoses that could explain their fatigue (i.e., aneurism, kidney disease). In addition, two people had a psychosocial explanation of their fatigue, and one person indicated that she no longer had fatigue when we recontacted her, and she requested to not participate in the next phase of the study. This means that for the target group of 13 individuals who we were most interested in further examining, there was 1 explicit refusal to participate in Phases 2 and 3 of the study. The remaining eight individuals (7 women, 1 man; M age = 32.6 years) were medically and psychiatrically examined. Within this group, there were two minorities (1 Latino, 1 African American). Of these eight individuals who indicated that they had or might currently have CFS, five were evaluated as not having CFS. One individual had no exclusionary reasons, but her severity of symptoms were not extreme enough to meet the definitions. The others had psychosocial, psychological stress, or other (i.e., substance abuse) exclusionary reasons for their fatigue. Three individuals were judged to have or have had CFS. One individual met all four definitions of CFS. Two other individuals had previously met all four definitions, but their conditions had somewhat improved, and when examined one met the less stringent British and Australian CFS definitions and the other person was sufficiently improved to no longer meet these criteria. Therefore two individuals were diagnosed with current CFS for a point prevalence of 0.2% (2/1031). A Caucasian woman, a 47-year-old former schoolteacher, met all four definitions. Before becoming sick, she had an active lifestyle, but this was curtailed severely after illness onset. She contracted influenza and this was followed by the syndrome onset 14 years ago. Her physician had not diagnosed her with CFS. No psychiatric illnesses were found for this subject. The second person diagnosed with CFS was a 48-year-old Caucasian female, who was a skilled executive, and had never been married. She became sick with mononucleosis 3~2 years earlier and was not rated as having a prior or current psychiatric disorder. She had not been diagnosed by a physician as having CFS. The third individual diagnosed as having had CFS was a


565

Table H. Symptomatology of Fatigued Participants Fatigued subjects (n =

n Length sick (M months) Fatigue present >50% time for 6 months or more Avg. daily activities