Somatisation Disorder and Its Associated Factors in Multiethnic Primary Care Clinic Attenders

Transcription

1 Int.J. Behav. Med. (2012) 19: DOI /s Somatisation Disorder and Its Associated Factors in Multiethnic Primary Care Clinic Attenders E. M. Khoo & N. J. Mathers & S. A. McCarthy & W. Y. Low Published online: 12 May 2011 # International Society of Behavioral Medicine 2011 Abstract Background Somatisation disorder (SD) has been reported as common in all ethnic groups, but the estimates of its prevalence have varied and the evidence for its associated factors has been inconsistent. Purpose This study seeks to determine the prevalence of SD and its associated factors in multiethnic primary care clinic attenders. Methods This cross-sectional study was on clinic attenders aged 18 years and above at three urban primary care clinics in Malaysia. The operational definition of SD was based on ICD- 10 criteria for SD for research, frequent attendance, and excluded moderate to severe anxiety and depression. The instruments used were the ICD-10 symptom list, the Hospital Anxiety and Depression Scale, a semi-structured questionnaire, and SF-36. Results We recruited 1,763 patients (response rate 63.8%). The mean age of respondents was 44.7±15.8 years, 807 (45.8%) were male; there were 35.3% Malay, 30.1% E. M. Khoo (*) Department of Primary Care Medicine, Faculty of Medicine, University of Malaya, Kuala Lumpur, Malaysia khooem@um.edu.my N. J. Mathers Academic Unit of Primary Medical Care, University of Sheffield, Sheffield, UK S. A. McCarthy Hospis Malaysia, Kuala Lumpur, Malaysia W. Y. Low Medical Education and Research Development Unit, Faculty of Medicine, University of Malaya, Kuala Lumpur, Malaysia Chinese and 34.6% Indian. SD prevalence was 3.7%; the prevalence in Malay was 5.8%, Indian 3.0% and Chinese 2.1%. Significant associations were found between SD prevalence and ethnicity, family history of alcoholism, blue-collar workers and the physical component summary (PCS) score of SF-36. Multivariate analysis showed that SD predictors were Malay ethnicity (OR 2.7, 95% CI 1.6, 4.6), blue-collar worker (OR 2.0, 95% CI 1.2, 3.5) and impaired PCS score of SF-36 (OR 0.92, 95% CI 0.90, 0.95). Conclusion The prevalence of SD was relatively uncommon with the stringent operational criteria used. SD preponderance in blue-collar workers may be attributable to secondary gain from getting sickness certificates and being paid for time off work. Keywords Somatisation disorder. Associated factors. Primary care. Ethnic groups. Prevalence Introduction Patients attend primary care clinics for many reasons. The symptoms that they present to their doctors may or may not be due solely to physical disease and family practitioners working in a wide variety of health and welfare systems will be familiar with those patients for whom there is no apparent diagnosis or specific treatment. A large number of labels have been given to such patients by family practitioners that include medically unexplained physical symptoms [1], heartsink patients [2] and somatisation disorder [3]. Somatisation disorder (SD) is characterised as a chronic syndrome of multiple physical complaints that are not explicable medically. It has been defined and conceptualised in a number of different ways and research findings of its prevalence have differed considerably. The

2 166 Int.J. Behav. Med. (2012) 19: prevalence has been estimated to be between 0.03% and 35% of all primary care attenders [4 10]. Part of the reason for this observed variation in prevalence include using the different settings of primary care and the community, ethnic groups surveyed, instrument variation, as well as different operational definitions and criteria for diagnosis. SD has been reported to be relatively common in all ethnocultural groups and societies [11]; however, the prevalence of SD has been shown to vary in different ethnic groups [7]. The reasons for these variations remain unclear. It has also been associated with female gender, lower level of education, lower socioeconomic status, older chronological age [10], occupational and social disability [7, 9 15] and impairment in health-related quality of life [16 20]. However, some of these associations have not been consistently observed. Malaysia is a multiethnic society comprising of the three main ethnic groups Malay, Chinese and Indian. Knowledge of the prevalence of SD in these ethnic groups is lacking, and its associated factors are unknown. Primary care clinics in Malaysia, therefore, offer opportunities to study SD in these diverse ethnicities within a common health and welfare system. We aimed to determine the prevalence of SD in these multiethnic clinic attenders and to examine possible associations with sociodemography and attendance variables such as the need for medical certification and quality of life. In Malaysia, medical certification is needed to allow for paid time off work. The study was designed to help identify the characteristics of patients with SD in primary care. We hypothesised that the prevalence of SD is relatively common and has a similar prevalence in the different ethnic groups and there is no association between the prevalence of SD and the sociodemographic factors, attendance variables and quality of life measures. available resource constraints. Pregnant women and patients who left the clinics prior to the invitation were excluded. Recruitment The data were collected for between 6 and 8 days from each clinic [See Fig. 1]. There were 3,312 consecutive clinic attenders aged 18 years and above during the study period, and 550 attenders had attended their consultations and left the clinics prior to invitation. The remaining 2,762 attenders, who constituted the study sampling frame, were invited to participate the study by one of the 16 trained multilingual enumerators. Each patient was given a patient information sheet and was informed that the purpose of the study was to look at how people from different cultures present their physical symptoms to the doctor. Written consent was sought. As part of the normal registration process when attending primary care clinics in Malaysia, participants were asked to self declare their ethnicity and each individual patient was matched with enumerators with the appropriate language skills. The majority of patients who refused to participate in the study gave the reason of a general lack of time.otherreasons for refusal included a need to see the doctor when their consultation was due, a desire for family consent before participation, anticipation that the time involved in participation was too long or a general reluctance to participate in any form of research eligible attenders aged 18 years and above were approached at registration Methods Design, Setting and Study Sample Invitation to participate Explanation of study Patient information sheet given 999 declined We conducted a cross-sectional study of primary care clinic attenders from three community-based primary care clinics in the Selangor state of Malaysia. Selangor is one of the 13 states in Malaysia situated in the west coast of Malaysia in proximity to Kuala Lumpur, the capital of Malaysia. It covers an area of 7,956 km 2 and has an estimated population of five million. We used purposive sampling to select these clinics (Anika, Bukit Kuda and Kelana Jaya clinics) on the basis of their ethnic distribution of attenders from the past statistics: Anika Clinic has a higher proportion of Indian attenders [21] while Bukit Kuda and Kelana Jaya Clinic have more Malay and Chinese patient attenders respectively [22]. All patients aged 18 years and above, attending the three primary care clinics during the study period, were invited to participate within the 1763 consented Anika Clinic Bukit Kuda Clinic Kelana Jaya Clinic 522 (29.6%) recruited 649(36.8%) recruited 592(33.6%) recruited in in 6 days in 7 days 8 days One-to-one interview prior to consultation Fig. 1 Results: recruitment and response rates

3 Int.J. Behav. Med. (2012) 19: Each participant received a face-to-face interview with an enumerator during which time questionnaires were completed prior to the patient s consultation with the family practitioner. The first questionnaire collected sociodemographic data, selfreported family histories of similar problems (medically unexplained symptoms), alcoholism and psychiatric disorder and recorded the frequency of their self-reported consultations with health care practitioners (including complementary practitioners) and the number of days off work in the preceding 3 months. The frequency of consultations in the preceding 3 months at the primary care clinic was validated from the available medical records. Other questionnaires used were: the ICD-10 symptom list [3], the Hospital Anxiety and Depression Scale (HADS) [23] and SF-36 for quality of life [24]. For the scoring of the 14 symptom list, the participants were askedtoreportanysymptomstheyhadinthepast2yearsfor which their doctors had been unable to find a cause. Definition of SD The operational definition of SD used in this study was a positive score on the 14 symptom list of the ICD- 10 criteria for diagnosis of the condition for research purposes [3], excluding moderate to severe anxiety and depression as measured by the HADS [23]. In addition, our definition also included three or more visits in the preceding 3 months to health care practitioners and/or complementary practitioners, with a reported refusal to accept that there is no physical disease as a basis for the presenting symptoms [See Table 1]. We used published definitions of frequent attendance which has been defined as eight or more visits in a year [25] asit has been shown that patients with SD have increased utilization of health care [4]. We chose a period of 3 months to reduce participants recall bias for previous visits. This is a non-standard definition from the ICD-10 that defines patient seeking three or more consultations for their distressing medically unexplained symptoms. Our operational definition of including more frequent visits over a shorter period of time is stricter than the ICD-10, but was chosen for pragmatic reasons to reduce the risk of false positive diagnoses. The inclusion of visits to complementary practitioners reflects not only the ICD-10 criteria for SD, but also a commonly held view that many patients attending Malaysian primary care clinics also visit complementary practitioners as part of their help-seeking behaviour. The patients with anxiety and depression as measured by the HADs were included in the denominator population. Translation Issues All questionnaires were translated into Bahasa Malaysia (Malay language), Mandarin Chinese and Tamil using forward and backward translations by two independent healthcare professionals. Analysis The extended version of the Statistical Package for Social Sciences (SPSS 15.0) was used for the data analysis. Associations were tested using Chi-square and Students t test Table 1 Operational definition of somatisation disorder (SD) The criteria for SD required all three of the following criteria: 1. A positive score on the 14 symptom list of the ICD-10 criteria for diagnosis of SD [3]: i.e. a total of >6 medically unexplained symptoms from the following list, in at least 2 of 4 designated organ systems for at least the last 2 years. Gastrointestinal symptoms: Abdominal pain Nausea Feeling bloated Bad taste in mouth, or excessively coated tongue Complaints of vomiting or regurgitation of food Complaints of frequent and loose bowel motions or discharge of fluids from anus Cardiovascular symptoms: Breathlessness without exertion Chest pains Genitourinary symptoms: Dysuria or complaints of frequency of micturation Unpleasant sensations in or around the genitals Complaints of unusual or copious vaginal discharge Skin and pain symptoms: Blotchiness or discoloration of skin Pain in the limbs, extremities, or joints Unpleasant numbness or tingling sensations 2. Three or more consultations in the preceding 3 months to health care practitioners and/or complementary practitioners 3. Excluding moderate and severe anxiety and depression as measured by the HADS [23] with a significance level set at p<0.05.categorical data with cells less than five were tested with Fisher s exact test. Multivariate analysis was performed to examine the predictors of SD. The study was approved by the ethics committee of the Ministry of Health, Malaysia. Results Response Rate We approached 2,762 patients (Fig. 1), 1,763 consented to participate, giving a response rate of 63.8% of whom 35.3% were Malay, 30.1% were Chinese and 34.6% were Indian patients. The mean age of respondents was 44.7±15.8 years and 807 (45.8%) were male and 956 (54.2%) were female [see Table 2]. Non-responders were significantly older (48.0±15.4 years, t=5.3, df=2,735, p<0.001, 95% CI 2.1, 4.5), but there was no significant difference in the gender distribution (45.0% male and 55.0% female, χ 2 =0.1, df=1, p=0.71). There

4 168 Int.J. Behav. Med. (2012) 19: Table 2 Sociodemographic characteristics and features of participants with SD, non-sd and the total population SD somatisation disorder, Non- SD non-somatisation disorder a Chi-square test b Independent sample t test c Statistically significant SD n(%) Non-SD n(%) Total n(%) p value Mean age (sd) in years b 42.1 (14.6) 44.8 (15.9 ) 44.7 (15.8) 0.17 (n=1,752) Range Gender a (n=1,763) Male 30 (46.2) 777 (45.8) 807 (45.8) 0.95 Female 35 (53.8) 921 (54.2) 956 (54.2) Ethnic groups a (n=1,763) Malay 36 (55.4) 586 (34.5) 622 (35.3) c Chinese 11 (16.9) 520 (30.6) 531 (30.1) Indian 18 (27.7) 592 (34.9) 610 (34.6) Marital status a (n=1,754) Single 11 (16.9) 325 (19.2) 336 (19.2) 0.31 Married 54 (83.1) 1,313 (77.7) 1,367 (77.9) Separated/divorced/widowed 0 51 (3.0) 51 (2.9) Education levels a (n=1,751) None 3 (4.6) 122 (7.2) 125 (7.1) 0.40 Primary (age 7 12 years) 19 (29.2) 444 (26.3) 463 (26.4) Secondary (age years) 38 (58.5) 862 (51.1) 900 (51.4) Tertiary (college and university) 5 (7.7) 244 (14.5) 249 (14.2) Religious school 0 14 (0.8) 14 (0.8) Occupation a (n=1,722) Professionals 3 (4.7) 89 (5.4) 92 (5.3) 0.46 Associate professionals 5 (7.8) 188 (11.3) 193 (11.2) Clerical workers 2 (3.1) 78 (4.7) 80 (4.6) Service and sales workers 8 (12.5) 103 (6.2) 111 (6.4) Skilled and craft workers 7 (10.9) 112 (6.8) 119 (6.9) Machine operators 9 (14.1) 145 (8.7) 154 (8.9) Elementary workers 3 (4.7) 113 (6.8) 116 (6.7) Housewives 13 (20.3) 359 (21.7) 372 (21.6) Retirees 8 (12.5) 260 (15.7) 268 (15.6) Students 3 (4.7) 95 (5.7) 98 (5.7) Unemployed 3 (4.7) 116 (7.0) 119 (6.9) Employment status a (n=1,722) Employed 37 (57.8) 828 (49.9) 865 (50.2) 0.22 Others 27 (42.2) 830 (50.1) 857 (49.8) Occupation categories a (n=1,763) Blue-collar 27 (41.5) 473 (27.9) 500 (28.4) 0.05 White-collar 10 (15.4) 355 (20.9) 365 (20.7) Non-working 28 (43.1) 870 (51.2) 898 (50.9) Days off work (n=865) b Mean±SD 3.0± ± ± were more Malay (40.2%) and Indian (37.6%) in the nonresponders (χ 2 =20.9, df=2, p<0.001). Prevalence of SD There were 65/1,763 patients (3.7%) who met the operational criteria for a diagnosis of SD. The prevalence of SD was 5.8% in the Malay patients, 2.1% in the Chinese patients and 3.0% in the Indian patients this difference in representation between the ethnic groups was statistically significant (χ 2 =12.56, df=2, p=0.002) [see Table 2]. The prevalence of SD using ICD-10 symptom counts alone for diagnosis was 8.8%. Prevalence of Anxiety and Depression Using the HADS, 28.1% of these primary care attenders had anxiety and 22.4% had depression. A total of 10.1%

5 Int.J. Behav. Med. (2012) 19: and 6.5% of the patients had moderate to severe anxiety and depression, respectively. Prevalence of SD and Other Associated Factors Family History There was a significant association between SD prevalence and a self-reported family history of alcoholism [5 (7.7%) SD patients vs 50 (3.0%) non-sd patients, χ 2 =4.63, df=1, p=0.03]. However, no significant associations were found between SD prevalence and a family history of similar problems [2 (3.1%) SD vs 58 (3.4%) non-sd, Fisher s exact test p=1.00] or a family history of psychiatric disorders [3(4.6%) SD vs 36(2.1%) non-sd, Fisher s exact test p=0.17]. Sociodemography There was no significant association found between SD prevalence and gender (3.7% for both men and women, χ 2 = 0.004, df=1, p=0.95), age (SD 42.1 years vs non-sd 44.8 years, t=1.4, df=1,750, p=0.17), marital status (χ 2 = 2.4, df=2, p=0.31), education level (χ 2 =4.0, df=4, p=0.40) and occupation (χ 2 =9.8, df=10, p=0.46). When occupation was regrouped, a weak association was found between SD prevalence and occupation categories (χ 2 =5.9, df=2, p= 0.05) [see Table 2]. Days Off Work In the total working population (n=865), there were no statistically significant associations found between the prevalence of SD and the number of days taken off work in the preceding 3 months (SD 2.97±4.77 days vs non- SD 1.64±6.74 days, t=1.19, df=863, p=0.23, 95% CI 0.86, 3.54). However, a significantly greater number of days were taken off work by men in the preceding 3 months compared to women (mean 2.09±7.93 days in men vs 1.14±4.27 days in women, t=2.29, df=808, p= 0.02, 95% CI 0.14, 1.78), although the numbers of bodily symptoms were similar in men and women (men 2.20± 2.34 vs women 2.35±2.51, t= 0.93, df=863, p=0.35, 95% CI 0.48, 0.17). In patients with SD (n=65), there were more men working than women (χ 2 =4.64, df=1, p=0.03). However, no significant difference was found in the number of days taken off work in the preceding 3 months (men 3.48±4.34, women 2.31±5.35, t=0.73, df=35, p=0.47, 95% CI 2.07, 4.40) nor a difference in the number of bodily symptoms inmenandwomenwithsd(men7.10±1.55vswomen 7.13±1.46, t=.059, df=35, p=0.95, 95% CI 1.05, 0.99) [see Table 3]. Table 3 Comparison of profiles of men and women in patients with SD Quality of Life SF-36 Using independent sample t tests, patients with SD had significantly more impairment in the following dimensions of SF-36 compared with non-sd: role physical (p<0.001), bodily pain (p<0.001), general health (p<0.001), vitality (p=0.008), role emotional (p=0.002) and the physical component summary (PCS) score (p<0.001). There was no significant difference found in the physical functioning, social functioning, mental health and Mental Component Summary (MCS) score between the two groups. Multiple logistic regression analysis showed that Malay, blue-collar worker and PCS scores were significant predictors of SD [see Table 4]. Discussion n (%) Male (n=30) Female (n=35) p value Proportion of SD 30 (46.2) 35 (53.8) (%) [n=65] In employment a (%) Yes [n=37] 21 (56.8) 16 (43.2) 0.03 c No [n=27] 8 (29.6) 19 (70.4) Mean no. of days off 3.48± ± work b [n=37] No. of bodily symptoms b [n=37] 7.10± ± a Chi-square b Independent sample t tests c Statistically significant The proportion of primary care clinic patients who satisfied the operational definition of SD used in this study was 3.7%. This is consistent with studies on prevalence using the ICD-10, DSM-IV or DSM-III-R criteria of % [6 10, 12, 26 28], but at the lower end of the published prevalence figures of between 0.03% and 35% for primary care clinic attenders and community settings [4 6]. Many definitions and conceptualisations have been used in a wide range of settings in both primary care and community settings; our operational criteria were stringent and probably accounts in part for the observed lower prevalence of SD in our study. However, we included frequent visits in our operational criteria as it has been shown that patients with SD have an increased utilisation of health care [4]. This is a pragmatic approach to operationalisation which reflects the realities of family practice with open access. In Malaysia, a visit to the doctor is necessary for sickness

6 170 Int.J. Behav. Med. (2012) 19: Table 4 Multivariate analysis of SD Predictors B S.E. Wald df Sig. Exp (B) 95% CI for Exp (B) Lower Upper a Statistically significant Malay a Blue-collar worker a Physical component summary a certification. Frequent clinic attendance, therefore, for such certificates may also have contributed to the observed prevalence of SD. In addition, the inclusion of visits to the complementary practitioners could have influenced the observed prevalence of SD too, but it is also part of the criteria in the ICD-10 for SD. It is clear, however, that researchers in this area face considerable methodological difficulties from a lack of agreed operational definitions for example, the ICD-10 [3], the abridged somatisation disorder [5] and other primary care definitions [19] for SD are all different from one another to some extent and will influence measures of prevalence. Other contributory factors that may affect the observed prevalence of SD, apart from the setting, the instruments and the operational definition used, are inherent in the nature of the health and welfare system and the cultural drivers which prompt attendance at primary care clinics for example, the use by patients of physical symptoms as admission tickets to health care may be associated not only with a reluctance to admit to psychological or mental health symptoms, but also with the provision of a socially acceptable reason for medical certification and time off work. Indeed, the observed prevalence of SD within a particular health and welfare system may in part be due to the group s relative familiarity with the health care system and pathways to care [11]. We have found a statistically significant association in the prevalence of SD and ethnicity; the Malay is 2.7 times more likely than other ethnic groups to have SD. To the authors knowledge, there were no previous data available on the prevalence of SD in the Malay, but the observed prevalence of SD in the Chinese and Indian population is consistent with the findings in China and India [7]. Somatising behaviour has been reported as common in all ethno-cultural groups although significant differences in prevalence have been observed in the presence of different health and welfare systems [11]. It was interesting to note that despite living in the same health and welfare system, the prevalence of SD in the Malay was higher than in the other ethnic groups suggesting that there may be other reasons that may have influenced the observed prevalence of SD. This study was not able to explain further any social and cultural factors that could lead to this difference. However, the difference in the prevalence of SD could be a result of the interplay of several factors such as the differences in the idioms of distress [7, 11, 29], differences in psycholinguistic expression of emotional states [30], the cultural beliefs of a society, as well as familiarity with pathways to health care [10, 11, 29, 31]. It is also possible that the higher prevalence in the Malay population could be due to a true difference in the SD prevalence or a difference in instrument translation as has been suggested in other studies involving disparate ethnic groups [13, 32]. Certain ethnic groups have been shown, for example, to have a higher prevalence of somatising behaviour: Vietnamese men and women reported more multiple unexplained physical symptoms than other ethnic groups [11]; and the Latin American compared with ethnic groups from 14 countries worldwide studied using ICD-10 criteria [7]. In addition, Beiser and Fleming have suggested that South East Asians are more likely to consider somatic rather than psychological symptoms as legitimate reasons for consulting physicians [33]. Parker reported that the Chinese in Malaysia were more likely to present somatic symptoms compared to Australian Caucasian depressed patients [34], while Tan found that physical symptoms were the most prominent symptoms among Chinese and Indian patients with anxiety in West Malaysia [35]. Therefore, somatising behaviour, albeit common in all ethno-cultural groups, differs among the groups in both prevalence and presentation. We did not observe a difference in the prevalence of SD between the genders. This was consistent with findings from studies using mainly the abridged somatisation disorder (ASD) criteria [7, 36], although many studies which have used the DSM criteria have found SD to be more prevalent in women [4, 7, 12]. Gureje showed in his study that eight out of 15 primary care centres had a higher prevalence of SD in men than women when ASD criteria were used [7]. Kirmayer has also reported that somatisers in primary care are more likely to be men [37], and that Vietnamese men in particular report multiple unexplained physical symptoms significantly more often than Vietnamese women [11]. This finding may demonstrate a greater resistance by men to acknowledge emotional distress, particularly depression or to label themselves as in need of mental health care [38]. In addition, it has been suggested that symptom perception may differ with gender which may affect symptom reporting, and if symptom perception variables are controlled, men may report more symptoms [39].

7 Int.J. Behav. Med. (2012) 19: In the whole working population of our patients, we observed a greater number of days off work taken by men (mean 2.09 days in men vs days in women) in the preceding 3 months, although the numbers of bodily symptoms reported by men and women were equivalent. However, in the patients with SD, there were more men working than women, and in these working groups, we did not observe any difference in the number of days taken off work by men and women or a difference in the numbers of bodily symptoms. These findings suggest that men in employment have a greater need to justify sickness certification and time off work within the Malaysian health and welfare system and may account for the similarity in the prevalence of SD observed between men and women, i.e. it is plausible in our study that there was a greater need by men than women attending the clinic to justify seeking medical help by increasing the number of symptoms they presented [12]. This would have made it more likely that they would satisfy the operational definition of SD used in this study. It is also possible that traditional notions of masculinity in Malay society among the males in the SD category encouraged a presentation of more somatic symptoms than symptoms of psychological distress that may be perceived as weak and powerless culturally [34]. We found that the blue-collar workers were twice as likely as white-collar workers and the non-working group to be characterised as having SD. It is possible that the nature of an occupation has an influence on somatising behaviour. Blue-collar workers generally have a more physically demanding jobs compared to white-collar workers and the non-working group. Somatisation, therefore, allows people to legitimately use somatic symptoms to adopt the sick role in a society that views mental illness as a stigma. Many studies have found SD to be associated with high indices of occupational and social disability [7, 9, 11 13] and the financial rewards of the paid time off work could reinforce somatising behaviour. In Malaysia, there is no specific provision for unemployment benefit, but sickness certification enables paid time off work. Nevertheless, the mean number of days taken off work, although greater in the patients with SD compared to non-sd, was not significantly different (3.0±4.8 vs 1.6±6.7 days). It is possible that our observation that more blue-collar workers have SD was due to a secondary gain from receiving a sickness certificate and being paid for time off work, but this needs to be further evaluated in a future study. We found both PCS and MCS scores of the SF-36 of patients with SD were lower than the US population norm of 50 which suggested impairment in quality of life. However, there was no significant difference in the MCS score between the SD and the non-sd groups, and multivariate analysis showed that PCS has a significant net effect on SD. This indicates that patients with SD had significant impairment in the physical component of their quality of life, which was consistent with the findings from other studies [16 20]. It also reflects the extent to which SD patients see themselves as physically and not emotionally ill. Strengths and Limitations of the Study This is the first study that has been conducted to determine the prevalence of SD in primary care settings in Malaysia and one of the few studies that examine the prevalence of SD in the three ethnic groups. We believe our operational definition of including frequent attendance is a pragmatic approach to the operationalisation of the definition which reflects the realities of family practice with open access and gives a more accurate reflection of doctors experience of SD in primary care settings. The exclusion of moderate to severe anxiety and depression in the definition of SD allows a better estimate of the prevalence of SD without much overlap with clinically significant psychiatric disorders such as the somatic presentation of anxiety and depression as has often been shown in other studies. The study of three ethnic groups in the same welfare and health care system allows comparisons to be made, avoiding the potential bias due to disparate systems as is often found in other studies that were conducted cross nationally. In addition, multilingual enumerators were trained and care was taken to match the patients preferred language of communication with enumerators language proficiency. No patients refused to participate due to a language barrier which has often been observed in studies involving multiethnic communities. However, our study sample is limited, of course, by the attending clinic population which may not be representative of the whole country s population. The sampling of the clinics was purposive for pragmatic reasons and to ensure an even distribution of the three ethnic groups. Care, therefore, needs to be exercised when generalising the results of the study, as these may not adequately reflect the population of the rest of Malaysia. Although we were able to validate the medical certification histories from the clinical records, the numbers of somatic symptoms reported by patients were not able to be validated from these records since these were not consistently recorded by the practitioners. The response rate of this study was 63.8% which, in our view, was acceptable. There was no significant difference in gender between the responders and non-responders, although the non-responders were older and had a higher proportion of Malay and Indian patients. As most studies have suggested association of SD with older age group, the estimates of prevalence in this study may thus be an underestimate. However, the prevalence of SD in the Indian and the Chinese

8 172 Int.J. Behav. Med. (2012) 19: was found to be consistent with that shown in the same ethnic group, respectively, in India and in China [7]. The big sample size of this study and the good response rate would have reduced the potential sampling error. Implications for Future Research or Clinical Practice The reasons for varied SD prevalence amongst different ethnic groups despite living in the same health and welfare system is intriguing and worthy of further investigation. The associated factors of SD such as the Malay ethnic group, blue-collar workers and physical impairment in health-related quality of life could be attributed by the cultural idiom of distress, the ethnobelief system of a society, the differing psycholinguistic expression of emotional states, the familiarity with the pathways to health care and the possible secondary gain of being paid for time off work. It is clear that the study has to be replicated within the different ethnic groups and pragmatic interventions need to be developed to treat patients with SD as they may be considered a burden on primary health care. Conclusion The prevalence of SD is affected by the definition used, the health and welfare system and the setting as well as the cultural drivers which inform health beliefs and the interpretation of symptoms which prompt attendance at primary care clinics. The current definitions and criteria for SD are complex and not sufficiently pragmatic for use in primary care settings. A review of SD definitions for use in primary care is urgently needed as the lack of a standard definition can cloud its diagnosis. Associated factors of SD require further investigation for confirmation. Acknowledgements This paper is based on a thesis for a doctorate degree of medicine (E.M. Khoo). We would like to acknowledge Professor Shyamala Nagaraj and Dr. Kam Cheong Wong for their statistical help. We would like to thank all the clinic staff, the enumerators, and the patients who had given so generously their time to take part in this study. Disclosure of interests References The authors declared no conflicts of interest. 1. Burton C. Beyond somatisation: a review of the understanding and treatment of medically unexplained physical symptoms (MUPS). Br J Gen Pract. 2003;53: Mathers NJ, Jones N, Hannay DR. Heartsink patients: a study of their general practitioners. Br J Gen Pract. 1995;45: World Health Organization (1993) Somatisation disorder. In: The ICD-10 classification of mental and behavioural disorders. Diagnostic criteria for research. Geneva: World Health Organization, pp Kirmayer LJ, Robbins JM. Three forms of somatization in primary care: prevalence, co-occurrence, and sociodemographic characteristics. J Nerv Ment Dis. 1991;179(11): Escobar JI, Waitzkin H, Cohen Silver R, Gara M, Holman A. Abridged somatization: a study in primary care. Psychosom Med. 1998;60: Peveler R, Kilkenny L, Kinmonth A. Medically unexplained physical symptoms in primary care: a comparison of self report screening questionnaires and clinical opinion. J Psychosom Res. 1997;42 (3): Gureje O, Simon GE, Ustun TB, Goldberg DP. Somatization in cross-cultural perspective: a World Health Organization Study in Primary Care. Am J Psychiatry. 1997;154(7): Escobar JI, Gara M, Silver RC, Waitzkin H, Holman A, Compton W. Somatisation disorder in primary care. Br J Psychiatry. 1998;173(9): Escobar JI, Golding JM, Hough RL, Karno M, Burnam MA, Wells KB. Somatisation in the community: relationship to disability and use of services. Am J Public Health. 1987;77: Escobar JI, Rubio-Stipec M, Canino G, Karno M. Somatic Symptom Index (SSI): a new and abridged somatisation construct: prevalence and epidemiological correlates in two large community samples. J Nerv Ment Dis. 1989;177(3): Kirmayer LJ, Young A. Culture and somatization: clinical, epidemiological, and ethnographic perspectives. Psychosom Med. 1998;60(4): Creed F, Barsky A. A systematic review of the epidemiology of somatisation disorder and hypochondriasis. J Psychosom Res. 2004;56: Escobar JI, Burnam MA, Karno M, Forsythe A, Golding JM. Somatization in the community. Arch Gen Psychiatry. 1987;44 (8): Simon GE, Vonkorff M. Somatization and psychiatric disorder in the NIMH Epidemiologic Catchment Area Study. Am J Psychiatry. 1991;148: Zocolillo M, Cloninger CR. Somatization disorder: psychologic symptoms, social disability, and diagnosis. Compr Psychiatry. 1986;27: Aiarzaguena JM, Grandes G, Salazar A, Gaminde I, Sánchez A. The diagnostic challenges presented by patients with medically unexplained symptoms in general practice. Scand J Prim Health Care. 2008;26(2): Dickinson WP, Dickinson LM, degruy FV, Candib LM, Main DS, Libby AM, et al. The somatization in primary care study: a tale of three diagnoses. Gen Hosp Psychiat. 2003;25(1): Duddu V, Husain N, Dickens C. Medically unexplained presentations and quality of life: a study of a predominantly South Asian primary care population in England. J Psychosom Res. 2008;65 (4): Kroenke K, Spitzer RL, degruy FV, Hahn SR, Linzer M, Williams JBW, et al. Multisomatoform disorder: an alternative to undifferentiated somatoform disorder for the somatizing patient in primary care. Arch Gen Psychiatry. 1997;54(4): Löwe B, Spitzer RL, Williams JB, Mussell M, Schellberg D, Kroenke K. Depression, anxiety and somatization in primary care: syndrome overlap and functional impairment. Gen Hosp Psychiatry. 2008;30(3): Anika clinic outpatient attendance record (2002). 22. Ministry of Health Malaysia. (2002). Health Information System, Outpatient attendance. 23. Snaith RP, Zigmond AS. The hospital anxiety and depression scale. England: The NFER-NELSON Publishing Company Ltd; 1994.

9 Int.J. Behav. Med. (2012) 19: Medical Outcomes Trust. SF-36 Health Survey. UK: QualityMetric Incorporated; Jyväsjärvi S, Joukamaa M, Väisänen E, Larivaara P, Kivelä S, Keinänen-Kiukaanniemi S. Somatizing frequent attenders in primary health care. J Psychosom Res. 2001;50: degruy F, Columbia L, Dickinson P. Somatization disorder in a family practice. J Fam Pract. 1987;25: Lobo A, Garcia-Campayo J, Campos R, Marcos G, Perez- Echeverria M. Somatisation in primary care in Spain: I. Estimates of prevalence and clinical characteristics. Br J Psychiatry. 1996;168(3): Smith RC, Gardiner JC, Lyles JS, Sirbu C, Dwamena FC, Hodges A, et al. Exploration of DSM-IV criteria in primary care patients with medically unexplained symptoms. Psychosom Med. 2005;67 (1): Bäärnhielm S, Ekblad S. Introducing a psychological agenda for understanding somatic symptoms an area of conflict for clinicians in relation to patients in a multicultural community. Cult Med Psychiatry. 2008;32(3): Cheng TA. Symptomatology of minor psychiatric morbidity: a cross cultural comparison. Psychol Med. 1989;19: Katon W, Ries RK, Kleinman A. The prevalence of somatization in primary care. Compr Psychiatry. 1984;25: Hsu LKG, Folstein MF. Somatoform disorders in Caucasian and Chinese Americans. J Nerv Ment Dis. 1997;185(6): Beiser M, Fleming JAE. Measuring psychiatric disorder among Southeast Asian refugees. Psychol Med. 1986;16: Parker G, Cheah YC, Roy K. Do the Chinese somatize depression? A cross-cultural study. Soc Psychiatry Psychiatr Epidemiol. 2001;36: Tan ES. The symptomatology of anxiety in West Malaysia. Aust New Zeal J Psychiatr. 1969;3: Piccinelli M, Simon G. Gender and cross-cultural differences in somatic symptoms associated with emotional distress. An international study in primary care. Psychol Med. 1997;27(2): Kirmayer LJ, Robbins JM. Patients who somatize in primary care: a longitudinal study of cognitive and social characteristics. Psychol Med. 1996;26(5): Emslie C, Ridge D, Ziebland S, Hunt K. Men s account of depression: reconstructing or resisting hegemonic masculinity? Soc Sci Med. 2006;62(9): Gijsbers van Wijk CMT, Kolk AM. Sex differences in physical symptoms: the contribution of symptom perception theory. Soc Sci Med. 1997;45(2): Funding body China Medical Board, University of Malaya is the funding body of this study. Ethics approval This study was approved by the Medical Research & Ethics Committee, Ministry of Health Malaysia [reference number KKM/ JEPP/Jld.6 (175)].