Study setting and participants

The study population consisted of consecutive primary care attenders, aged 18–45 years. The age range was deliberately restricted in an attempt to reduce the likelihood of misdiagnosing medically explained fatigue as unexplained because medical disorders that cause fatigue, for example anaemia, diabetes mellitus and hypothyroidism, are much more prevalent in older age groups. ^{Reference Sharpe25} We deliberately chose to oversample Brazilian patients compared with UK ones in a ratio of 2:1. There are already several prevalence studies on unexplained chronic fatigue/chronic fatigue syndrome in British primary care, ^{Reference Wessely, Chalder, Hirsch, Wallace and Wright23,Reference McDonald, David, Pelosi and Mann26,Reference David, Pelosi, McDonald, Stephens, Ledger, Rathbone and Mann27} but none in Brazil. Therefore, a reasonably large sample size was required in Brazil to provide more accurate data on the prevalence of both. Eleven general practices in São Paulo and five in London were selected.

In the UK, primary care is generally the first port of call for the general population when they have a health problem, but Brazilian primary care is more complex. According to the 2000 national census, 24.5% of the Brazilian population are covered by at least one type of health insurance and the rest dependent on public healthcare. Hence, in order to obtain a study population constituted by different socio-economic groups, reasonably representative of the healthcare seeking population, ten public clinics and one private clinic were selected across São Paulo, and 25.4% of the Brazilian sample was recruited from the private clinic. In London, four practices were selected from inner-city areas and one from a suburban area. Additionally, to maximise the comparability between the two countries, only the public clinics with the Family Health Program, an official programme which resembles the UK system, were recruited in Brazil.

Assessments and procedures

Fatigue status including severity and duration was assessed with the Chalder Fatigue Questionnaire (CFQ), an 11-item questionnaire widely used to measure physical and mental fatigue. ^{Reference Chalder, Berelowitz, Pawlikowska, Watts, Wessely, Wright and Wallace28} Prior to this study, we had conducted a rigorous process of translation, back-translation, cross-cultural adaptation and validation of the CFQ in Brazil, the details of which are provided elsewhere. ^{Reference Cho, Costa, Menezes, Chalder, Bhugra and Wessely29} Based on the validation studies, we used a cut-off of 3/4 by bimodal scoring for substantial fatigue in both countries. ^{Reference Chalder, Berelowitz, Pawlikowska, Watts, Wessely, Wright and Wallace28,Reference Cho, Costa, Menezes, Chalder, Bhugra and Wessely29} Chronic fatigue was defined as a score of four or more on the CFQ with a reported duration of 6 months or greater. Unexplained chronic fatigue was defined as medically unexplained substantial fatigue lasting 6 months or more according to the CFQ and the medical examination. Chronic fatigue syndrome, characterised by severe physical and mental fatigue and other accompanying symptoms which cannot be explained by any other medical condition and have persisted for at least 6 months, was assessed using the Centers for Disease Control (CDC) 1994 case definition. ^{Reference Fukuda, Straus, Hickie, Sharpe, Dobbins and Komaroff30} We also determined whether those participants who fulfilled criteria for chronic fatigue had been diagnosed as a case of chronic fatigue syndrome by their GPs. For this purpose, we reviewed their medical records in search for the following diagnostic labels: ‘chronic fatigue’, ‘chronic fatigue syndrome’, ‘myalgic encephalomyelitis’ and ‘post-viral fatigue syndrome’. Finally, the 12-item General Health Questionnaire (GHQ–12) ^{Reference Goldberg and Williams31} was used with the conventional cut-off of 3/4 by bimodal scoring to determine probable common mental disorder, as validated in both British ^{Reference Goldberg and Williams31} and Brazilian primary care. ^{Reference Mari and Williams32} Ethical approval was obtained from the research ethics committees of King's College Hospital, Institute of Psychiatry, London, UK, and the Municipal Department of Health of São Paulo and University of São Paulo Medical School, Brazil.

Analysis

Stata Version 10 for Windows ^{Reference Gallagher, Thomas, Hamilton and White33} was employed for all statistical analyses and the significance level was set at P=0.05. Participant characteristics were compared between the two countries by two-tailed chi-squared or t-tests. The prevalence of unexplained chronic fatigue and chronic fatigue syndrome in each country was estimated with the appropriate 95% confidence interval (CI). The aforementioned diagnostic labels were counted during the medical records review, and the percentage frequency in each country with 95% CI was calculated as a proxy for the physician recognition rate of the syndrome. The prevalence of each fatigue category – unexplained chronic fatigue, chronic fatigue syndrome and diagnostic labels – was compared between Brazil and the UK using a chi-squared test. Furthermore, in order to take into account the confounding effect of socio-demographic characteristics and common mental disorder, multivariable logistic regression was conducted with country membership (Brazil v. UK) as the exposure variable and each fatigue category as the outcome variable.

Prevalence of unexplained chronic fatigue and chronic fatigue syndrome

Our study used a design similar to those of previous British primary care studies, and the prevalence of unexplained chronic fatigue (12.2% in Brazil and 10.3% in Britain) was roughly similar to previous estimates (11.2% by McDonald et al ^{Reference McDonald, David, Pelosi and Mann26} and 11.3% by Wessely et al ^{Reference Wessely, Chalder, Hirsch, Wallace and Wright23} ). Chronic fatigue syndrome was also comparably prevalent (1.6% in Brazil and 2.1% in Britain) to previous UK and USA estimates (2.6% by Wessely et al ^{Reference Wessely, Chalder, Hirsch, Wallace and Wright23} and 2.5% by Reeves et al ^{Reference Reeves, Jones, Maloney, Heim, Hoaglin, Boneva, Morrissey and Devlin20} ). In the same way as fatigue is a universal symptom occurring across regions and cultures, chronic fatigue syndrome as defined by the current international consensus was also similar between Brazil and the UK. Furthermore, unexplained chronic fatigue as a less severe sub-syndromal counterpart of chronic fatigue syndrome was similarly common across the two countries, consistent with the notion that fatigue is distributed as a continuous variable in the general population. ^{Reference Pawlikowska, Chalder, Hirsch, Wallace, Wright and Wessely2,Reference Chalder, Berelowitz, Pawlikowska, Watts, Wessely, Wright and Wallace28}

Recognition and labelling of fatigue

Terms such as chronic fatigue syndrome, myalgic encephalomyelitis and post-viral fatigue syndrome are routinely made diagnoses in British primary care as shown by the analysis of the UK General Practice Research Database. ^{Reference Gallagher, Thomas, Hamilton and White33} However, two vignette studies from Brazil showed that even university-based tertiary care doctors are unfamiliar with the construct of chronic fatigue syndrome and rarely diagnose it in their medical practice. ^{Reference Cho, Menezes, Bhugra and Wessely12,Reference Nacul, Perreira, Jacques, Ramos, Rodrigues, Oliva, Gonçalves and Almeida34} Our study confirmed this observation using actual case notes. Although the prevalence of the syndrome in Brazil as assessed using a standardised procedure is comparable to that in the UK, it is not a diagnostic concept currently used in Brazilian medical practice and is neither recognised nor diagnosed.

Fatigue as a symptom elicited by a questionnaire or recorded by an interviewer should mostly reflect the presence or absence of the symptom in an individual. However, fatigue as a diagnosis made by the physician depends upon factors other than the simple presence or absence of the symptom. Much needs to happen for GPs to recognise fatigue cases and label them with such diagnoses as chronic fatigue syndrome, myalgic encephalomyelitis and post-viral fatigue syndrome. First of all, the doctor needs to attribute some importance to the symptom of fatigue. It has been shown that doctors accord fatigue only a minor importance in comparison with patients – doctors rate it as far less important than patients, probably because of its ubiquity and lack of diagnostic specificity. ^{Reference Dohrenwend and Crandell35} Furthermore, the practitioner needs to be aware of these labels and to have at least some knowledge of the diagnostic concept. Even then, a practitioner may be well aware of the concepts and/or labels, but feel that these are not valid or useful, and hence not use them. Studies of doctors' knowledge and attitudes towards chronic fatigue syndrome/myalgic encephalomyelitis have indicated that opinions vary widely about the existence and utility of these labels. ^{Reference Bowen, Pheby, Charlett and McNulty36,Reference Finestone37} There are also changes in physicians' diagnostic patterns for fatigue over time. ^{Reference Gallagher, Thomas, Hamilton and White33}

These factors, upon which the recognition and labelling of fatigue depend, may have an important link with the sociocultural setting. The degree of medicalisation of the population and awareness of chronic fatigue syndrome among the population and the medical professionals may be important. The explanatory models held by individuals with chronic fatigue seem to be more biomedically oriented in Western affluent societies compared with non-Western societies. ^{Reference Chalder, Power and Wessely38–Reference Karasz and McKinley40} Further data from the current study, reported elsewhere, ^{Reference Cho, Bhugra and Wessely41} demonstrated that British people with unexplained chronic fatigue were more likely to attribute their fatigue to physical causes than their Brazilian counterparts in line with this proposition. Moreover, while chronic fatigue syndrome is well known and officially endorsed as a medical condition in the UK, ^{Reference Hutchinson42} it is little known in Brazil by either patients or doctors. ^{Reference Cho, Menezes, Bhugra and Wessely12,Reference Nacul, Perreira, Jacques, Ramos, Rodrigues, Oliva, Gonçalves and Almeida34} We believe that these differences in sociocultural context between Brazil and the UK have contributed to the current findings. Conversely, these findings indicate the importance of sociocultural factors not so much in the occurrence and distribution of fatigue but more in the recognition and labelling of fatigue. In Brazil, where unexplained fatigue is not sanctioned as a medical condition worthy of medical treatment, sick leave or sickness benefit, ^{Reference Cho, Menezes, Bhugra and Wessely12,Reference Nacul, Perreira, Jacques, Ramos, Rodrigues, Oliva, Gonçalves and Almeida34} it may be more likely to be considered as part of everyday adversity and less likely to be recognised as a medical disorder. Likewise, individuals with a similar range of symptoms are considerably less likely to receive a label of chronic fatigue syndrome/myalgic encephalomyelitis in Brazil than in the UK.

Limitations and strengths

The study has some limitations. First, the medical records review for the assessment of fatigue diagnosis was incomplete in Brazil – medical records could not be traced for almost 30% of Brazilian participants with chronic fatigue. In addition, although the UK medical records were all electronic and mostly complete, the available Brazilian medical records were all on paper and frequently lacking information on examination results. None the less, the absence of any fatigue diagnosis in a far larger pool of medical records in Brazil is unlikely to be explained by this limitation. In addition, the type of quality problem observed in the Brazilian medical records was not directly related to diagnostic labels and we conducted laboratory tests in Brazil to compensate for this particular deficiency. Furthermore, as mentioned above, there is also evidence that Brazilian doctors are unfamiliar with the construct of chronic fatigue syndrome and rarely use this diagnosis in their practice. ^{Reference Cho, Menezes, Bhugra and Wessely12,Reference Nacul, Perreira, Jacques, Ramos, Rodrigues, Oliva, Gonçalves and Almeida34} Second, general practices were not randomly selected and consequently selection bias was possible. In order to minimise this problem, we employed the sampling strategy of including different social classes in proportion to the national data and selecting general practices with similar characteristics in the two countries.

Despite these limitations, this study also had some strengths as it included large samples with a reasonable number of clinics; made a direct comparison of a poorly explored topic in two culturally and economically distinct settings using the same method; employed cross-culturally validated measures of fatigue and psychological distress; and conducted multivariable analyses.

This is the first study to specifically estimate the prevalence of unexplained chronic fatigue/chronic fatigue syndrome in Brazil. If taken at face value, it raises the question of unexplained chronic fatigue/chronic fatigue syndrome as a hidden public health issue in Brazil given the prevalence of and the disability caused by the conditions. More importantly, this is the first study to examine the epidemiology of chronic fatigue syndrome using the same standardised methodology across two culturally and economically distinct countries, an Anglophone affluent country and a Latin American middle-income country. Despite its limitations, this study provides some evidence about the role of sociocultural factors in the recognition of fatigue and the use of labels such as chronic fatigue syndrome/myalgic encephalomyelitis. In other words, the recognition of this illness rather than the illness itself seems to be culturally induced. The overall conclusion is that unexplained chronic fatigue/chronic fatigue syndrome can be found in Brazil in similar proportions as the UK, if one cares to look. At the moment, it seems that Brazilian society, or more specifically its healthcare system, does not care to look. How appropriate that is, and what the impact is, will remain a matter of speculation. Further research on the pragmatic implications of the current findings may shed more light on the understanding of this controversial, continuing but real health problem.