- Email: [email protected]
Somatoform symptoms profiles in relation to psychological disorders – A population classification analysis in a large sample of general adults
Author’s Accepted Manuscript Somatoform Symptoms Profiles in Relation to Psychological Disorders - A Population Classification Analysis in a Large Sample of General Adults Zahra Heidari, Awat Feizi, Hamidreza Roohafza, Ammar Hassanzadeh Keshteli, Payman Adibi www.elsevier.com/locate/psychres
PII: DOI: Reference:
S0165-1781(16)31442-1 http://dx.doi.org/10.1016/j.psychres.2017.04.064 PSY10490
To appear in: Psychiatry Research Received date: 26 August 2016 Revised date: 10 March 2017 Accepted date: 27 April 2017 Cite this article as: Zahra Heidari, Awat Feizi, Hamidreza Roohafza, Ammar Hassanzadeh Keshteli and Payman Adibi, Somatoform Symptoms Profiles in Relation to Psychological Disorders - A Population Classification Analysis in a Large Sample of General Adults, Psychiatry Research, http://dx.doi.org/10.1016/j.psychres.2017.04.064 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting galley proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Somatoform Symptoms Profiles in Relation to Psychological Disorders - A Population Classification Analysis in a Large Sample of General Adults
Zahra Heidaria,b, Awat Feizia,c*, Hamidreza Roohafzad, Ammar Hassanzadeh Keshtelie,f, Payman Adibif,g
a
Department of Biostatistics and Epidemiology, School of Health, Isfahan University of Medical Sciences, Isfahan, Iran b Student Research Center, School of Health, Isfahan University of Medical Sciences, Isfahan, Iran c Psychosomatic Research Center, Isfahan University of Medical Sciences, Isfahan, Iran d Cardiac Rehabilitation Research Center, Cardiovascular Research Institute, Isfahan University of Medical Sciences, Isfahan, Iran e Department of Medicine, University of Alberta, Edmonton, Alberta, Canada f Integrative Functional Gastroenterology Research Center, Isfahan University of Medical Sciences, Isfahan, Iran g Department of Internal Medicine, School of Medicine, Isfahan University of Medical Sciences, Isfahan, Iran *
Corresponding Author: Dr. Awat Feizi. Address: Department of Epidemiology and Biostatistics,
School of Public Health and Psychosomatic Research Center, Isfahan University of Medical Sciences, Hezarjarib St, Isfahan, Iran. Tel.: +983137923250. [email protected]
Abstract In order to identifying somatoform symptoms profiles, classifying study population and evaluating of psychological disorders in extracted classes, we carried out a cross-sectional study on 4762 Iranian adults. Somatoform symptoms were assessed using a comprehensive 30-items 1
questionnaire and psychological disorders were evaluated by 12-item General Health Questionnaire (GHQ-12) and Hospital Anxiety and Depression Scale (HADS) questionnaires. Factor analysis and factor mixture modeling (FMM) were used for data analysis. Four somatoform symptoms profiles were extracted, including ‘psycho-fatigue’, ‘gastrointestinal’, ‘neuro- skeletal’ and ‘pharyngeal-respiratory’. According to FMM results, a two-class fourfactor structure, based somatoform symptoms, was identified in our study population. Two identified classes were labeled as “low psycho-fatigue complaints” and “high psycho-fatigue complaints”. The scores of psychological disorders profile was significantly associated with four somatoform symptoms profiles in both classes; however the stronger relationship was observed in high psycho-fatigue complaints class. The prevalence of all the somatoform symptoms among participants assigned to the "high psycho-fatigue complaints" class was significantly higher than other class. We concluded that somatoform symptoms have a dimensional-categorical structure within our study population. Our study also provided informative pathways on the association of psychological disorders with somatoform symptoms. These findings could be useful for dealing with treatment’s approaches. Keywords: Somatoform symptoms; psychological disorders; depression; anxiety; psychological distress; factor mixture modeling 1. Introduction Somatoform symptoms such as headache, fatigue, dizziness or shortness of breath, are causing significant functional impairments. They are defined as the presence of bodily symptoms as a result of interrelations of mind and body and without physical explanation even after medical evaluation, for that (Manshaee and Hamidi, 2013; Xiong et al., 2015). These symptoms are prevalent not only in patients attending to primary care, secondary care and clinics but also in 2
general populations (Escobar et al., 2010; Novy et al., 2005; Steinbrecher et al., 2011). The higher prevalence of these symptoms puts significant burden on the healthcare delivery system and has considerable impact on quality of life (Gonzalez et al., 2009; Wong et al., 2015). Persons with somatoform symptoms firstly seek help from a physician because of their physical signs. However, after many physical examinations with lacking of results for determining the etiology of them, individuals suffering from these symptoms are referred to a psychiatrist. There are evidences that psychological disorders are risk factors for somatoform symptoms (Sugahara et al., 2004; Wong et al., 2015; Zhu et al., 2012), in which patients with psychological disorders are more likely to have somatoform symptoms than general population (Bener et al., 2013; Gonzalez et al., 2009; Haftgoli et al., 2010; Han et al., 2014; Shidhaye et al., 2013). For instance, high prevalence (73-92%) of somatoform symptoms has been reported in depressed patients (Caballero et al., 2008; Simon et al., 1999; Sugahara et al., 2004). Wong et al. demonstrated that anxiety is a modifiable risk factor for somatoform symptoms in which reducing anxiety could be considered as an effective approach for reducing somatoform symptoms (Wong et al., 2015). Some evidences showed that there is notable heterogeneity in the somatoform symptoms and few studies are available on their classification (Fink et al., 2007; Gara et al., 1998; Kato et al., 2010; Lacourt et al., 2013; Nimnuan et al., 2001). In previous researches, different statistical techniques such as factor analysis, clustering, and latent class analysis (LCA) have been used to address the heterogeneity (Fink et al., 2007; Kato et al., 2010; Lacourt et al., 2013). FMM is a hybrid model that unifies factor analysis and latent class analysis in a single framework and allows the underlying structure to be simultaneously dimensional and categorical (Lubke and Muthén, 2005). This structure is considered categorical because FMM classifies the individuals into sub-groups and it is also considered dimensional because this modeling approach takes into
3
account the heterogeneity within groups using continuous latent variables (Lubke and Muthén, 2005). Therefore, FMM may be superior to other methods both in terms of class detection and class assignment. Some extensions of FMM allow including a set of explanatory variables in the main structure of model. FMM with explanatory variables has been investigated under this assumption that explanatory variables are allowed to affect both latent variables and group membership (Lubke and Muthén, 2005). Given the relations between somatoform symptoms and psychological disorders, the objectives of the present study were identifying profiles of somatoform symptoms (latent factors) and classifying studied population (latent classes) into more homogeneous subgroups based on constructed profiles of somatoform symptoms and evaluating psychological disorders profile in identified classes using FMM. 2. Materials and Methods 2.1 Study Design and Participants This cross-sectional population-based study was conducted in the framework of “Study of the Epidemiology of Psychological, Alimentary Health and Nutrition” (SEPAHAN) project that was performed in 2 phases in a large sample of Iranian adults population in the Isfahan province (Adibi et al., 2012). In the first phase of SEPAHAN project, different questionnaires on demographic information, lifestyle and nutritional factors, were distributed among 10087 invited persons, and 8691 subjects took part (response rate: 86.16%). At the second phase, others questionnaires, which were designed to collect information on gastrointestinal, psychological, and somatoform symptoms, were distributed and 6239 questionnaires were completed (response rate: 64.64%). Then, national identification numbers of the participants used to link the questionnaires from both phases. Finally, after considering missing data, data on 4762 subjects
4
with completed information used in the current analysis. Written informed consent was obtained from all participants. The study was approved by the Bioethics committee of Isfahan University of Medical Sciences, Isfahan, Iran (Project numbers: #189069, #189082, and #189086). More details about SEPAHAN project are presented elsewhere (Adibi et al., 2012). 2.2 Procedures and Assessment of Variables 2.2.1 Assessment of somatoform symptoms No separate specific questionnaire was available to assess somatoform symptoms in the SEPAHAN project; however, we established a validated questionnaire based on 31 items in SEPAHAN’s questionnaires common with “the patient health questionnaire (PHQ)” (Spitzer et al., 1999) and “the 47-items questionnaire used in the Lacourt et al.’s study” (Lacourt et al., 2013), as valid and standard tools for the assessment of somatoform symptoms. We used 31items questionnaire to evaluate the frequency of somatoform symptoms. Participants could indicate how much they had experienced each symptom in the past three months on a four points Likert scale (never, sometimes, often, and always). For one item (i.e. Dry mouth), the rating scale was as: never, low and high. In order to assess the reliability of this instrument, we conducted a separate mini survey of 100 participants selected randomly. There was strong internal reliability, with a Cronbach's alpha score of 0.903 (Heidari et al., 2017b). In the current study, in order to use all study participants in the analysis, we removed women specific “menstrual disorder” item from 31 items; so 30 somatoform symptoms were used in the analysis. 2.2.2 Assessment of Psychological Variables 2.2.2.1 Psychological Distress Iranian validated version of self-report screening instrument of 12-item General Health Questionnaire (GHQ-12) was used to detect psychological distress. The internal consistency was
5
assessed by Cronbach's alpha coefficient and it was found to be 0.87 (Montazeri et al., 2005). The each item of the instrument asks whether the participant has experienced a particular symptom or behavior recently. Each item has a four-point Likert scale (less than usual, no more than usual, rather more than usual, or much more than usual). A respondent’s score could be between 0 and 12 points, and a threshold score of 4 or more was used to identify a respondent with high-distress level (Montazeri et al., 2005). 2.2.2.2 Hospital Anxiety and Depression Scale Self-report screening instrument of Hospital Anxiety and Depression Scale (HADS) was used to assess depression and anxiety of participants. HADS is a self-administered questionnaire that is validated by Montazeri et al. for Iranian populations (Montazeri et al., 2003). It consists of 14 items which 7 of them are allocated to depression. It has a 4-point Likert scale ranging from 0 (not present) to 3 (considerable). The anxiety or depression score of respondent could be between 0 and 21 points (0-7: normal, 8-21: mild, moderate or severe disorder). Internal consistency which is assessed by Cronbach’s alpha has been found to be 0.78 (Montazeri et al., 2003). 2.2.3 Assessment of Other Variables Self-administered standard questionnaires were distributed to gather demographic (age, gender, marital status (single/married), educational level (under diploma, diploma (12 year formal education) and university graduate) etc.) and life styles characteristics (weight (kg), height (m), physical activity (inactive and moderately inactive/moderately active and active) based on General Practice Physical Activity Questionnaire (GPPAQ) (N.C.C., 2008) etc.). 2.3 Statistical Analysis
6
For following up our main study objective i.e. if participants could be clustered into meaningful subgroups based on their somatoform symptoms using FMM, firstly, we performed factor analysis on the 30 individual somatoform symptoms and resulted four interpretable factors based on orthogonal Varimax rotation procedure. Next, four extracted profiles of somatoform symptoms were labeled based on the loaded items in each factor. Then, LCA was used to determine the appropriate number of latent classes in studied population, models with 2 or 3 latent classes was found based on goodness of fit criteria. After that, we conducted different FMMs with 4 factors and 2 or 3 latent classes. Determination of goodness of fit of models was guided through comparing the Bayesian Information Criterion (BIC) (Schwarz, 1978) and entropy indices across models. Lower BIC and higher entropy values indicate better model fitting and class separation, respectively (Lubke and Muthén, 2007). Finally, we extracted a latent factor from psychological problems (i.e. anxiety, depression and psychological distress) in order to evaluate its association with somatoform symptoms profiles, in identified classes by using FMM. 3. Results Overall, 4762 adults contributed in the study. The mean age was 36.58±0.13 years. They consisted of 2657 (55.8%) females and 3776 (81.2%) married. 2650 (57.2%) of participants had college education. Psychological distress, anxiety, and depression were identified in 23.1%, 5.8% and 10.4% of participants, respectively. About 3.5% of individuals were underweight, 37.1% were overweight and 9.4% were obese. 34.8% of participants had regular physical activity (moderately active and active). 3.1 Extraction of somatoform symptoms profiles using exploratory factor analysis
7
Four extracted profiles based on 30 individual somatoform symptoms were labeled as ‘psychofatigue’, ‘gastrointestinal’, ‘neuro-skeletal’ and ‘pharyngeal-respiratory’; they accounted for 12.4%, 12.3%, 11.4% and 9.3% of total variance, respectively (Table 1) (Heidari et al., 2017b; Shabbeh et al., 2016).
3.2 Correlation between somatoform symptoms profiles and psychological disorders Table 2 shows the correlations between scores of somatoform symptoms profiles and the scores of psychological disorders of studied population. Psychological disorders (i.e. psychological distress, anxiety and depression) are significantly correlated with all somatoform symptoms profiles. Among them, anxiety had stronger associations with somatoform symptoms profiles (Table 2).
3.3 Factor mixture modeling’ results The latent structure or unobserved heterogeneity of studied population in terms of four somatoform symptoms profiles (i.e., ‘psycho-fatigue’, ‘gastrointestinal’, ‘neuro-skeletal’ and ‘pharyngeal-respiratory’) was recognized using FMM. During model fitting, we observed that a two-classes/four-factors model allowing for free intercepts, covariances and means across latent classes had lowest BIC (346926.053) and entropy 0.995. These values indicating that individuals are correctly classified by our fitted model. The two identified classes were labeled as “high psycho-fatigue complaints” and “low psycho-fatigue complaints”; because participants in the first class experienced higher scores of psychological-fatigue somatic symptoms (mean: 0.249 in first class vs. 0 in second class) while experienced lower scores of somatoform symptoms in physical domains (gastrointestinal, neuro-skeletal and pharyngeal-respiratory). There were 4243
8
individuals (89%) in the low psycho-fatigue complaints class and 519 participants (11%) in the high psycho-fatigue complaints class. According to the two-class four-factor FMM solution, approximately, all items are significantly loaded on their respective factor (Heidari et al., 2017a). Table 3 contains regression coefficients on the association of psychological disorders profile with somatoform symptoms profiles in the two extracted classes. In both classes, the profile of psychological disorders is significantly related to the four somatoform symptoms profiles. The regression coefficient of psychological disorders profile was near 0.3 in both classes. There was a significant positive relationship between the profile of psychological disorders and gastrointestinal profile in both classes and its regression coefficient was greater in high psychofatigue complaints class (0.152 and 0.147 in high and low psycho-fatigue complaints classes, respectively; p<0.0001). Furthermore, the regression coefficients for neuro-skeletal profile were 0.219 and 0.210 in high and low psycho-fatigue complaints classes, respectively; (p<0.0001). It can be seen that there is significant positive relationship between psychological disorders profile and pharyngeal –respiratory profile in both classes; however, its regression coefficient on pharyngeal –respiratory profile was greater in high psycho-fatigue complaints class (0.134 and 0.08 in high and low psycho-fatigue complaints classes, respectively; p< 0.0001). (Table 3) The prevalence of the somatoform symptoms in two extracted latent classes is presented in Table 3, too. Although, majority of symptoms had a four-point Likert scale; we only reported the response of participants to ‘often and always’ categories. The prevalence of all 30 somatoform symptoms for participants assigned to the "high psycho-fatigue complaints" class was significantly higher than other class (P<0.0001). In the "often" category, the most common somatoform symptoms reported by participants assigned to "high psycho-fatigue complaints" class were “dry mouth” (45%), “severe fatigue” (39%), followed by “headache”, back pain”,
9
“sleep disorder”, and “feeling low on energy” (all about 24%). In the "always" category, the most frequent somatoform symptoms reported by participants assigned to the "high psychofatigue complaints" class were “severe fatigue” (15%), followed by “feeling low on energy” (13.6%), “pain in joints” (13%), “disturbing thoughts” (12.8%), “back pain” (12.1%) and “dry mouth” (10.5%). (Table 3)
4. Discussion In this cross-sectional population-based study, two classes characterized by high (11%) and low (89%) levels of psychological-fatigue complaints, and four-factors profiles (i.e., ‘psycho-fatigue’, ‘gastrointestinal’, ‘neuro- skeletal’ and ‘pharyngeal-respiratory’) representing underlying structure of the somatoform symptoms were identified from a large sample of Iranian adults using factor mixture modeling (FMM). We observed that the prevalence of all somatoform symptoms for participants assigned to "high psycho-fatigue complaints" class was significantly higher than" low psycho-fatigue complaints" class. Although, fitted FMM led to two classes with mentioned characteristics, however, it should be noted that participants in both classes suffered from other somatoform symptoms profiles with different degrees. We did not find any study such as current study, which stratified a general large population into homogeneous subgroups based on somatoform symptoms using FMM. However, other statistical approaches i.e. clustering, factor analysis and LCA in some studies were used to classify psychosomatic symptoms (Fink et al., 2007; Gara et al., 1998; Kato et al., 2010; Lacourt et al., 2013; Nimnuan et al., 2001). For instance, in the Fink et al.’s study on 978 internal medical, neurological, and primary care patients, a distinct pattern of cardiopulmonary (CP), musculoskeletal/pain (MS), and gastrointestinal (GI) symptom factors as well as three classes of 10
patients including “non-bodily distress”, “modest bodily distress”, and “severe bodily distress” were identified (Fink et al., 2007). The aforesaid study had similarities with our study in terms of gastrointestinal and skeletal profiles. In the Gara et al.’s study 11 clusters of patients with different patterns of medically unexplained symptoms, were identified using hierarchical cluster analysis (Gara et al., 1998). The observed disparities in results of conducted studies in this area of subject can be attributed to geographic, socio-economic status, culture and racial dependency of somatoform symptoms. It is believed that somatoform symptoms are the manifestations of psychological illnesses in the form of physical symptoms (Wong et al., 2015) and there are evidences that psychological disorders are risk factors for somatoform symptoms (Zhu et al., 2012). Majority of previous studies were restricted to the association of psychological disorders with a few somatoform symptoms or with an overall score of somatization (Kinnunen et al., 2010; Wong et al., 2015; Zhu et al., 2012). In current study, three psychological disorders i.e. anxiety, depression and psychological distress were combined (as a latent factor) and its collective association with somatoform symptoms profiles was examined. We observed that the profile of psychological disorders is positively associated with the four somatoform symptoms profiles with greater coefficient in the high psycho-fatigue complaints class, except for psycho-fatigue profile. Previous studies have emphasized on the strong association of psychological disorders with somatoform symptoms. Wong et al. demonstrated that anxiety is a modifiable risk factor for psychosomatic symptoms in general Chinese populations (Wong et al., 2015). Kinnunen et al.’s study showed that nearly all psychosomatic symptoms are associated with mental health symptoms (Kinnunen et al., 2010). The results of Zhu et al.’s study on 2408 clinical patients, revealed that “depression and anxiety” are main risk factors for high somatic symptoms (Zhu et 11
al., 2012). Koh et al.’s study found that both anxiety and depression have direct effects on somatic symptoms in patients suffering from these disorders (Koh et al., 2008). Dales et al. indicated that psychological problems are important determinants of respiratory symptoms (such as cough, wheeze and dyspnea) (Dales et al., 1989). The influence of psychological disorders, such as anxiety and depression, on somatoform symptoms can be explained from biological perspectives (Wong et al., 2015). Psychological disorders have major role in initiation and development of gastrointestinal symptoms potentially via mechanisms involving immune modulation and alteration brain processing of incoming sensory signals (Wouters and Boeckxstaens, 2016). Also, regarding to the association of psychological disorders profile with neuro-skeletal profile of somatic symptoms, the possible role of neurotransmitters and cytokine receptors (Trivedi, 2004; Vargas-Prada and Coggon, 2015; Walker et al., 2014) could be mentioned. It is important to recognize some strengths and limitations of the present study. A major strength of our large population based study is the application of factor mixture model for identifying profiles of somatoform symptoms, and classifying study population, simultaneously, instead of dealing with them, separately. Furthermore, psychological disorders profile was evaluated in identified classes through FMM. However, due to the cross-sectional nature of SEPAHAN design, we could not infer cause–effect relationships from our findings. All used information in the present analysis was collected by self-administered questionnaires that might lead to misclassifying the participants. Finally, because SEPAHAN study’s participants were health centers staffs, thus, generalization of the present findings to the Iranian general population must be done with caution.
12
In summary, our study’s findings, in the context of an observational study, suggested that somatoform symptoms had a dimensional-categorical structure within our population that could be useful for dealing with treatment’s approaches. In addition, we showed that the profile of psychological disorders is significantly associated with different somatoform symptoms profiles. Conflicts of interest None Acknowledgements The present article was extracted from a Biostatistics PhD thesis at the School of Health, Isfahan University of Medical Sciences, with project number 394832. SEPAHAN was financially supported by a grant from the Vice Chancellery for Research and Technology, Isfahan University of Medical Sciences (IUMS). We are grateful to thank all staff of Isfahan University of Medical Sciences (MUI) who kindly participated in our study and staff of Public Relations Unit, and other authorities of IUMS for their excellent cooperation.
13
References Adibi, P., Keshteli, A.H., Esmaillzadeh, A., Afshar, H., Roohafza, H., Bagherian-Sararoudi, R., Daghaghzadeh, H., Soltanian, N., Feinle-Bisset, C., Boyce, P., 2012. The study on the epidemiology of psychological, alimentary health and nutrition (SEPAHAN): overview of methodology. J. Res. Med. Sci. 17. Bener, A., Dafeeah, E.E., Chaturvedi, S.K., Bhugra, D., 2013. Somatic symptoms in primary care and psychological comorbidities in Qatar: neglected burden of disease. Int. Rev. Psychiatry 25, 100–106. Caballero, L., Aragonès, E., García-Campayo, J., Rodríguez-Artalejo, F., Ayuso-Mateos, J.L., Polavieja, P., Gómez-Utrero, E., Romera, I., Gilaberte, I., 2008. Prevalence, characteristics, and attribution of somatic symptoms in Spanish patients with major depressive disorder seeking primary health care. Psychosomatics 49, 520–529. Dales, R.E., Spitzer, W.O., Schechter, M.T., Suissa, S., 1989. The influence of psychological status on respiratory symptom reporting. Am. Rev. Respir. Dis. 139, 1459–1463. doi:10.1164/ajrccm/139.6.1459 Escobar, J.I., Cook, B., Chen, C.-N., Gara, M.A., Alegría, M., Interian, A., Diaz, E., 2010. Whether medically unexplained or not, three or more concurrent somatic symptoms predict psychopathology and service use in community populations. J. Psychosom. Res. 69, 1–8. Fink, P., Toft, T., Hansen, M.S., Ørnbøl, E., Olesen, F., 2007. Symptoms and syndromes of bodily distress: an exploratory study of 978 internal medical, neurological, and primary care patients. Psychosom. Med. 69, 30–39. Gara, M.A., Silver, R.C., Escobar, J.I., Holman, A., Waitzkin, H., 1998. A hierarchical classes analysis (HICLAS) of primary care patients with medically unexplained somatic symptoms.
14
Psychiatry Res. 81, 77–86. Gonzalez, D.S., Rodríguez, M., García, C., Prieto, R., Saiz-Ruiz, J., 2009. Gender differences in major depressive disorder: somatic symptoms and quality of life. Rev. Psiquiatr. y salud Ment. 2, 119–127. doi:10.1016/S1888-9891(09)72402-4 Haftgoli, N., Favrat, B., Verdon, F., Vaucher, P., Bischoff, T., Burnand, B., Herzig, L., 2010. Patients presenting with somatic complaints in general practice: depression, anxiety and somatoform disorders are frequent and associated with psychosocial stressors. BMC Fam. Pract. 11, 1-8. Han, H., Wang, S.-M., Han, C., Lee, S.-J., Pae, C.-U., 2014. The relationship between somatic symptoms and depression. Neuro Endocrinol. Lett. 35, 463–469. Heidari, Z., Feizi, A., Roohafza, H., Hassanzadeh Keshteli, A., Shiravi, F.Z., Adibi, P., 2017a. Demographic and life styles determinants of somatic complaints’ structures: a crosssectional study on a large sample of Iranian adults using factor mixture model. Int. J. Prev. Med. 8, 8. doi: 10.4103/2008-7802.200526 Heidari, Z., Keshteli, A.H., Feizi, A., Afshar, H., Adibi, P., 2017b. Somatic complaints are significantly associated with chronic uninvestigated dyspepsia and its symptoms: a large cross-sectional population based study. J. Neurogastroenterol. Motil. 23, 80–91. doi:10.5056/jnm16020 Kato, K., Sullivan, P.F., Pedersen, N.L., 2010. Latent class analysis of functional somatic symptoms in a population-based sample of twins. J. Psychosom. Res. 68, 447–453. Kinnunen, P., Laukkanen, E., Kylmä, J., 2010. Associations between psychosomatic symptoms in adolescence and mental health symptoms in early adulthood. Int. J. Nurs. Pract. 16, 43– 50.
15
Koh, K.B., Kim, D.K., Kim, S.Y., Park, J.K., Han, M., 2008. The relation between anger management style, mood and somatic symptoms in anxiety disorders and somatoform disorders. Psychiatry Res. 160, 372–379. Lacourt, T., Houtveen, J., van Doornen, L., 2013. “Functional somatic syndromes, one or many?”: An answer by cluster analysis. J. Psychosom. Res. 74, 6–11. Lubke, G., Muthén, B.O., 2007. Performance of factor mixture models as a function of model size, covariate effects, and class-specific parameters. Struct. Equ. Model. A Multidiscip. J. 14, 26–47. doi:10.1207/s15328007sem1401_2 Lubke, G.H., Muthén, B., 2005. Investigating population heterogeneity with factor mixture models. Psychol. Methods 10, 21. Manshaee, G.R., Hamidi, E., 2013. Prevalence of psychosomatic symptoms among adolescent’s computer users. Procedia-Social Behav. Sci. 84, 1326–1332. Montazeri, A., Harirchi, A.M., Shariati, M., Garmaroudi, G., Ebadi, M., Fateh, A., Toscani, F., Borreani, C., Boeri, P., Miccinesi, G., 2005. The 12-item General Health Questionnaire (GHQ-12): translation and validation study of the Iranian version. Health Qual. Life Outcomes 1, 66. Montazeri, A., Vahdaninia, M., Ebrahimi, M., Jarvandi, S., 2003. The Hospital Anxiety and Depression Scale (HADS): translation and validation study of the Iranian version. Health Qual. Life Outcomes 1, 14. doi:10.1186/1477-7525-1-14 National Collaborating Centre (N.C.C.) for Nursing and Supportive Care (UK), 2008. Irritable Bowel Syndrome in Adults: Diagnosis and Management of Irritable Bowel Syndrome in Primary Care [Internet]. London: Royal College of Nursing (UK); 2008 Feb. Appendix J, The General Practice Physical Activity Questionnaire (GPPAQ). Available from:
16
https://www.ncbi.nlm.nih.gov/ books/NBK51962/. [Last accessed on 2006 Oct]. Nimnuan, C., Rabe-Hesketh, S., Wessely, S., Hotopf, M., 2001. How many functional somatic syndromes? J. Psychosom. Res. 51, 549–557. Novy, D., Berry, M.P., Palmer, J.L., Mensing, C., Willey, J., Bruera, E., 2005. Somatic symptoms in patients with chronic non-cancer-related and cancer-related pain. J. Pain Symptom Manage. 29, 603–612. Schwarz, G., 1978. Estimating the dimension of a model. Ann. Stat. 6, 461–464. doi:10.1214/aos/1176344136 Shabbeh, Z., Feizi, A., Afshar, H., Hassanzade Kashtali, A., Adibi, P., 2016. Identifying the profiles of psychosomatic disorders in an iranian adult population and their relation to psychological problems. J. Maz. Univ. Med. Sci. 26, 82–94 [In Persian]. Shidhaye, R., Mendenhall, E., Sumathipala, K., Sumathipala, A., Patel, V., 2013. Association of somatoform disorders with anxiety and depression in women in low and middle income countries: A systematic review. Int. Rev. Psychiatry 25, 65–76. Simon, G.E., VonKorff, M., Piccinelli, M., Fullerton, C., Ormel, J., 1999. An international study of the relation between somatic symptoms and depression. N. Engl. J. Med. 341, 1329– 1335. Spitzer, R.L., Kroenke, K., Williams, J.B.W., Group, P.H.Q.P.C.S., 1999. Validation and utility of a self-report version of PRIME-MD: the PHQ primary care study. Jama 282, 1737–1744. Steinbrecher, N., Koerber, S., Frieser, D., Hiller, W., 2011. The prevalence of medically unexplained symptoms in primary care. Psychosomatics 52, 263–271. Sugahara, H., Akamine, M., Kondo, T., Fujisawa, K., Yoshimasu, K., Tokunaga, S., Kubo, C., 2004. Somatic symptoms most often associated with depression in an urban hospital
17
medical setting in Japan. Psychiatry Res. 126, 151–158. Trivedi, M.H., 2004. The link between depression and physical symptoms. Prim Care Companion J Clin Psychiatry 6, 12–16. Vargas-Prada, S., Coggon, D., 2015. Psychological and psychosocial determinants of musculoskeletal pain and associated disability. Best Pract. Res. Clin. Rheumatol. 29, 374– 390. Walker, A.K., Kavelaars, A., Heijnen, C.J., Dantzer, R., 2014. Neuroinflammation and comorbidity of pain and depression. Pharmacol. Rev. 66, 80–101. Wong, J.Y.-H., Fong, D.Y.-T., Chan, K.K.-W., 2015. Anxiety and insomnia as modifiable risk factors for somatic symptoms in Chinese: a general population-based study. Qual. Life Res. 24, 2493–2498. Wouters, M.M., Boeckxstaens, G.E., 2016. Is there a causal link between psychological disorders and functional gastrointestinal disorders? Expert Rev. Gastroenterol. Hepatol. 10, 5–8. Xiong, N., Fritzsche, K., Wei, J., Hong, X., Leonhart, R., Zhao, X., Zhang, L., Zhu, L., Tian, G., Nolte, S., 2015. Validation of patient health questionnaire (PHQ) for major depression in Chinese outpatients with multiple somatic symptoms: A multicenter cross-sectional study. J. Affect. Disord. 174, 636–643. Zhu, C., Ou, L., Geng, Q., Zhang, M., Ye, R., Chen, J., Jiang, W., 2012. Association of somatic symptoms with depression and anxiety in clinical patients of general hospitals in Guangzhou, China. Gen. Hosp. Psychiatry 34, 113–120. doi:10.1016/j.genhosppsych.2011.09.005
18
Table 1. Factor loadings for the four extracted somatoform symptoms profiles from 30 somatoform symptoms Factor Loadings a Somatoform Symptoms Psycho-Fatigue Sleep Disorder
0.46
Pounding heart
0.41
Feeling low on energy
0.69
Feeling like ‘butterflies’
0.78
Difficulty concentrating
0.64
Disturbing thoughts
0.80
Gastrointestinal
NeuroSkeletal
Pharyngeal Respiratory
0.41
Chest pain
0.52
Feeling of fullness
0.69
Nausea
0.50
Gastroesophageal reflux
0.54
Pain or discomfort in the abdomen
0.71
Constipation
0.49
Diarrhea
0.36
Bloating or swelling of the abdomen
0.67
Anal pain
0.48
Headache
0.57
Back pain
0.66
Pain in joints
0.64
Eyesore
0.50
Severe fatigue
0.61
Dizziness and confusion
0.51
Chills and extreme cold
0.42
Hot flashes
0.38
Dry mouth
0.31
19
Neck pain
0.56
Globus sensation
0.55
Having trouble swallowing
0.61
Shortness of breath
0.46
Hoarseness
0.61
Wheezing (asthma)
0.52
Variance explained (%)
12.4
12.3
11.4
9.3
Cumulative variance
12.4
24.7
36.1
45.4
a
Factor loadings<0.3 are not shown for simplicity.
20
Table 2. Correlation between the scores of somatoform symptoms profiles and the scores of psychological disorders
Somatoform Symptoms Profiles Psycho-Fatigue
Gastrointestinal
Neuro-Skeletal
Pharyngeal Respiratory
Depression
0.799
0.415
0.541
0.385
Anxiety Psychological Distress
0.907
0.504
0.612
0.461
0.650
0.373
0.472
0.344
- All presented correlations are significant at the 0.01 level
21
Table 3. Comparison of psychological disorders profile and the prevalence of individual somatoform symptoms in two extracted classes.
Psycho-Fatigue Profile Sleep Disorder Pounding heart Feeling low on energy Feeling like ‘butterflies’ Difficulty concentrating Disturbing thoughts Gastrointestinal Profile Chest pain Feeling of fullness Nausea Gastroesophageal reflux Pain or discomfort in the abdomen Constipation Diarrhea Bloating or swelling of the abdomen Anal pain Neuro-Skeletal Profile
High Psycho-Fatigue Complaints class (n=519) Estimate Often b Always b (SE)a Number Number 0.282 (%) (%) * (0.011) 120 33 (6.7) (24.3) 74 (14.7) 22 (4.4) 121 68 (13.6) (24.2) 36 (7.2) 28 (5.6) 68 (13.5) 40 (8.0) 95 (19.0) 64 (12.8) 0.152 (0.014)*
74 (14.7) 13 (2.6) 91 (18.4)
33 (6.6) 4 (0.8) 44 (8.9)
35 (7.3)
11 (2.3)
Pain in joints Eyesore Severe fatigue
341 (8.2) 56 (1.3) 398 (9.9)
73 (1.8) 6 (0.1) 123 (3.1)
76 (1.9)
9 (0.2)
210 (5.0) 602 (14.5) 524 (12.7) 407 (9.8)
45 (1.1) 106 (2.6)
233 (5.6) 933 (22.6) 215 (5.2) 80 (1.9) 168 (4.0)
54 (1.3) 206 (5.0)
1170 (28.2) 98 (2.4) 52 (1.2) 0 139 (3.3) 47 (1.1) 53 (1.3)
116 (2.8)
0.210 (0.011)* 74 (14.7) 124 (24.8) 124 (24.7) 104 (20.5) 76 (15.0) 197 (39.0) 79 (15.7) 41 (8.1) 59 (11.7)
Back pain
Neck pain Globus sensation Having trouble swallowing Shortness of breath Hoarseness Wheezing (asthma)
13 (2.6) 24 (4.7) 16 (3.2) 11 (2.3) 22 (4.4)
0.219 (0.016)*
Pounding heart Headache
Dizziness and confusion Chills and extreme cold Hot flashes Pharyngeal –Respiratory Profile Dry mouth
77 (15.2) 90 (17.7) 40 (8.1) 30 (6.1) 82 (16.2)
Low Psycho-Fatigue Complaints class (n=4243) Estimate Often Always (SE)a Number Number 0.300 (%) (%) (0.011)* 411 99 (2.4) (10.1) 210 (5.0) 45 (1.1) 574 230 (5.6) (13.9) 148 (3.6) 58 (1.4) 200 (4.8) 78 (1.9) 421 202 (4.9) (10.2) 0.147 (0.008)* 146 (3.5) 16 (0.4) 230 (5.5) 40 (1.0) 78 (1.9) 22 (0.5) 50 (1.2) 16 (0.4) 276 (6.7) 44 (1.1)
22 (4.4) 45 (9.0) 61 (12.1) 66 (13.0) 23 (4.6) 76 (15.0) 29 (5.8) 9 (1.8) 15 (3.0)
0.134 (0.015)*
179 (4.3) 173 (4.2)
34 (0.8) 13 (0.3) 35 (0.8)
0.081 (0.011)* 227 (45.0) 65 (12.9) 61 (12.2) 51 (10.1) 68 (13.5) 26 (5.2) 26 (4.9)
22
53 (10.5) 13 (2.6) 18 (3.6) 5 (1.0) 31 (6.1) 5 (1.0) 6 (1.2)
17 (0.4) 10 (0.2) 0 40 (1.0) 11 (0.3) 17 (0.4)
a
Regression coefficients for the association of psychological disorders profile with somatoform symptoms profiles; *PValue< 0.0001 b The prevalence of all somatoform symptoms was significantly different between two classes (P <0.001).
23
Highlights
The aims of the current study were to classify studied population based on psychosomatic complaints profiles and evaluate the profile of psychological disorders in extracted classes. Factor mixture modeling was used with data from a sample of 4762 Iranian adults. A two-class, four-factor structure was identified for the psychosomatic complaints. The profile of psychological disorders was significantly related to the psychosomatic complaints profiles.
24
PII: DOI: Reference:
S0165-1781(16)31442-1 http://dx.doi.org/10.1016/j.psychres.2017.04.064 PSY10490
To appear in: Psychiatry Research Received date: 26 August 2016 Revised date: 10 March 2017 Accepted date: 27 April 2017 Cite this article as: Zahra Heidari, Awat Feizi, Hamidreza Roohafza, Ammar Hassanzadeh Keshteli and Payman Adibi, Somatoform Symptoms Profiles in Relation to Psychological Disorders - A Population Classification Analysis in a Large Sample of General Adults, Psychiatry Research, http://dx.doi.org/10.1016/j.psychres.2017.04.064 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting galley proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Somatoform Symptoms Profiles in Relation to Psychological Disorders - A Population Classification Analysis in a Large Sample of General Adults
Zahra Heidaria,b, Awat Feizia,c*, Hamidreza Roohafzad, Ammar Hassanzadeh Keshtelie,f, Payman Adibif,g
a
Department of Biostatistics and Epidemiology, School of Health, Isfahan University of Medical Sciences, Isfahan, Iran b Student Research Center, School of Health, Isfahan University of Medical Sciences, Isfahan, Iran c Psychosomatic Research Center, Isfahan University of Medical Sciences, Isfahan, Iran d Cardiac Rehabilitation Research Center, Cardiovascular Research Institute, Isfahan University of Medical Sciences, Isfahan, Iran e Department of Medicine, University of Alberta, Edmonton, Alberta, Canada f Integrative Functional Gastroenterology Research Center, Isfahan University of Medical Sciences, Isfahan, Iran g Department of Internal Medicine, School of Medicine, Isfahan University of Medical Sciences, Isfahan, Iran *
Corresponding Author: Dr. Awat Feizi. Address: Department of Epidemiology and Biostatistics,
School of Public Health and Psychosomatic Research Center, Isfahan University of Medical Sciences, Hezarjarib St, Isfahan, Iran. Tel.: +983137923250. [email protected]
Abstract In order to identifying somatoform symptoms profiles, classifying study population and evaluating of psychological disorders in extracted classes, we carried out a cross-sectional study on 4762 Iranian adults. Somatoform symptoms were assessed using a comprehensive 30-items 1
questionnaire and psychological disorders were evaluated by 12-item General Health Questionnaire (GHQ-12) and Hospital Anxiety and Depression Scale (HADS) questionnaires. Factor analysis and factor mixture modeling (FMM) were used for data analysis. Four somatoform symptoms profiles were extracted, including ‘psycho-fatigue’, ‘gastrointestinal’, ‘neuro- skeletal’ and ‘pharyngeal-respiratory’. According to FMM results, a two-class fourfactor structure, based somatoform symptoms, was identified in our study population. Two identified classes were labeled as “low psycho-fatigue complaints” and “high psycho-fatigue complaints”. The scores of psychological disorders profile was significantly associated with four somatoform symptoms profiles in both classes; however the stronger relationship was observed in high psycho-fatigue complaints class. The prevalence of all the somatoform symptoms among participants assigned to the "high psycho-fatigue complaints" class was significantly higher than other class. We concluded that somatoform symptoms have a dimensional-categorical structure within our study population. Our study also provided informative pathways on the association of psychological disorders with somatoform symptoms. These findings could be useful for dealing with treatment’s approaches. Keywords: Somatoform symptoms; psychological disorders; depression; anxiety; psychological distress; factor mixture modeling 1. Introduction Somatoform symptoms such as headache, fatigue, dizziness or shortness of breath, are causing significant functional impairments. They are defined as the presence of bodily symptoms as a result of interrelations of mind and body and without physical explanation even after medical evaluation, for that (Manshaee and Hamidi, 2013; Xiong et al., 2015). These symptoms are prevalent not only in patients attending to primary care, secondary care and clinics but also in 2
general populations (Escobar et al., 2010; Novy et al., 2005; Steinbrecher et al., 2011). The higher prevalence of these symptoms puts significant burden on the healthcare delivery system and has considerable impact on quality of life (Gonzalez et al., 2009; Wong et al., 2015). Persons with somatoform symptoms firstly seek help from a physician because of their physical signs. However, after many physical examinations with lacking of results for determining the etiology of them, individuals suffering from these symptoms are referred to a psychiatrist. There are evidences that psychological disorders are risk factors for somatoform symptoms (Sugahara et al., 2004; Wong et al., 2015; Zhu et al., 2012), in which patients with psychological disorders are more likely to have somatoform symptoms than general population (Bener et al., 2013; Gonzalez et al., 2009; Haftgoli et al., 2010; Han et al., 2014; Shidhaye et al., 2013). For instance, high prevalence (73-92%) of somatoform symptoms has been reported in depressed patients (Caballero et al., 2008; Simon et al., 1999; Sugahara et al., 2004). Wong et al. demonstrated that anxiety is a modifiable risk factor for somatoform symptoms in which reducing anxiety could be considered as an effective approach for reducing somatoform symptoms (Wong et al., 2015). Some evidences showed that there is notable heterogeneity in the somatoform symptoms and few studies are available on their classification (Fink et al., 2007; Gara et al., 1998; Kato et al., 2010; Lacourt et al., 2013; Nimnuan et al., 2001). In previous researches, different statistical techniques such as factor analysis, clustering, and latent class analysis (LCA) have been used to address the heterogeneity (Fink et al., 2007; Kato et al., 2010; Lacourt et al., 2013). FMM is a hybrid model that unifies factor analysis and latent class analysis in a single framework and allows the underlying structure to be simultaneously dimensional and categorical (Lubke and Muthén, 2005). This structure is considered categorical because FMM classifies the individuals into sub-groups and it is also considered dimensional because this modeling approach takes into
3
account the heterogeneity within groups using continuous latent variables (Lubke and Muthén, 2005). Therefore, FMM may be superior to other methods both in terms of class detection and class assignment. Some extensions of FMM allow including a set of explanatory variables in the main structure of model. FMM with explanatory variables has been investigated under this assumption that explanatory variables are allowed to affect both latent variables and group membership (Lubke and Muthén, 2005). Given the relations between somatoform symptoms and psychological disorders, the objectives of the present study were identifying profiles of somatoform symptoms (latent factors) and classifying studied population (latent classes) into more homogeneous subgroups based on constructed profiles of somatoform symptoms and evaluating psychological disorders profile in identified classes using FMM. 2. Materials and Methods 2.1 Study Design and Participants This cross-sectional population-based study was conducted in the framework of “Study of the Epidemiology of Psychological, Alimentary Health and Nutrition” (SEPAHAN) project that was performed in 2 phases in a large sample of Iranian adults population in the Isfahan province (Adibi et al., 2012). In the first phase of SEPAHAN project, different questionnaires on demographic information, lifestyle and nutritional factors, were distributed among 10087 invited persons, and 8691 subjects took part (response rate: 86.16%). At the second phase, others questionnaires, which were designed to collect information on gastrointestinal, psychological, and somatoform symptoms, were distributed and 6239 questionnaires were completed (response rate: 64.64%). Then, national identification numbers of the participants used to link the questionnaires from both phases. Finally, after considering missing data, data on 4762 subjects
4
with completed information used in the current analysis. Written informed consent was obtained from all participants. The study was approved by the Bioethics committee of Isfahan University of Medical Sciences, Isfahan, Iran (Project numbers: #189069, #189082, and #189086). More details about SEPAHAN project are presented elsewhere (Adibi et al., 2012). 2.2 Procedures and Assessment of Variables 2.2.1 Assessment of somatoform symptoms No separate specific questionnaire was available to assess somatoform symptoms in the SEPAHAN project; however, we established a validated questionnaire based on 31 items in SEPAHAN’s questionnaires common with “the patient health questionnaire (PHQ)” (Spitzer et al., 1999) and “the 47-items questionnaire used in the Lacourt et al.’s study” (Lacourt et al., 2013), as valid and standard tools for the assessment of somatoform symptoms. We used 31items questionnaire to evaluate the frequency of somatoform symptoms. Participants could indicate how much they had experienced each symptom in the past three months on a four points Likert scale (never, sometimes, often, and always). For one item (i.e. Dry mouth), the rating scale was as: never, low and high. In order to assess the reliability of this instrument, we conducted a separate mini survey of 100 participants selected randomly. There was strong internal reliability, with a Cronbach's alpha score of 0.903 (Heidari et al., 2017b). In the current study, in order to use all study participants in the analysis, we removed women specific “menstrual disorder” item from 31 items; so 30 somatoform symptoms were used in the analysis. 2.2.2 Assessment of Psychological Variables 2.2.2.1 Psychological Distress Iranian validated version of self-report screening instrument of 12-item General Health Questionnaire (GHQ-12) was used to detect psychological distress. The internal consistency was
5
assessed by Cronbach's alpha coefficient and it was found to be 0.87 (Montazeri et al., 2005). The each item of the instrument asks whether the participant has experienced a particular symptom or behavior recently. Each item has a four-point Likert scale (less than usual, no more than usual, rather more than usual, or much more than usual). A respondent’s score could be between 0 and 12 points, and a threshold score of 4 or more was used to identify a respondent with high-distress level (Montazeri et al., 2005). 2.2.2.2 Hospital Anxiety and Depression Scale Self-report screening instrument of Hospital Anxiety and Depression Scale (HADS) was used to assess depression and anxiety of participants. HADS is a self-administered questionnaire that is validated by Montazeri et al. for Iranian populations (Montazeri et al., 2003). It consists of 14 items which 7 of them are allocated to depression. It has a 4-point Likert scale ranging from 0 (not present) to 3 (considerable). The anxiety or depression score of respondent could be between 0 and 21 points (0-7: normal, 8-21: mild, moderate or severe disorder). Internal consistency which is assessed by Cronbach’s alpha has been found to be 0.78 (Montazeri et al., 2003). 2.2.3 Assessment of Other Variables Self-administered standard questionnaires were distributed to gather demographic (age, gender, marital status (single/married), educational level (under diploma, diploma (12 year formal education) and university graduate) etc.) and life styles characteristics (weight (kg), height (m), physical activity (inactive and moderately inactive/moderately active and active) based on General Practice Physical Activity Questionnaire (GPPAQ) (N.C.C., 2008) etc.). 2.3 Statistical Analysis
6
For following up our main study objective i.e. if participants could be clustered into meaningful subgroups based on their somatoform symptoms using FMM, firstly, we performed factor analysis on the 30 individual somatoform symptoms and resulted four interpretable factors based on orthogonal Varimax rotation procedure. Next, four extracted profiles of somatoform symptoms were labeled based on the loaded items in each factor. Then, LCA was used to determine the appropriate number of latent classes in studied population, models with 2 or 3 latent classes was found based on goodness of fit criteria. After that, we conducted different FMMs with 4 factors and 2 or 3 latent classes. Determination of goodness of fit of models was guided through comparing the Bayesian Information Criterion (BIC) (Schwarz, 1978) and entropy indices across models. Lower BIC and higher entropy values indicate better model fitting and class separation, respectively (Lubke and Muthén, 2007). Finally, we extracted a latent factor from psychological problems (i.e. anxiety, depression and psychological distress) in order to evaluate its association with somatoform symptoms profiles, in identified classes by using FMM. 3. Results Overall, 4762 adults contributed in the study. The mean age was 36.58±0.13 years. They consisted of 2657 (55.8%) females and 3776 (81.2%) married. 2650 (57.2%) of participants had college education. Psychological distress, anxiety, and depression were identified in 23.1%, 5.8% and 10.4% of participants, respectively. About 3.5% of individuals were underweight, 37.1% were overweight and 9.4% were obese. 34.8% of participants had regular physical activity (moderately active and active). 3.1 Extraction of somatoform symptoms profiles using exploratory factor analysis
7
Four extracted profiles based on 30 individual somatoform symptoms were labeled as ‘psychofatigue’, ‘gastrointestinal’, ‘neuro-skeletal’ and ‘pharyngeal-respiratory’; they accounted for 12.4%, 12.3%, 11.4% and 9.3% of total variance, respectively (Table 1) (Heidari et al., 2017b; Shabbeh et al., 2016).
3.2 Correlation between somatoform symptoms profiles and psychological disorders Table 2 shows the correlations between scores of somatoform symptoms profiles and the scores of psychological disorders of studied population. Psychological disorders (i.e. psychological distress, anxiety and depression) are significantly correlated with all somatoform symptoms profiles. Among them, anxiety had stronger associations with somatoform symptoms profiles (Table 2).
3.3 Factor mixture modeling’ results The latent structure or unobserved heterogeneity of studied population in terms of four somatoform symptoms profiles (i.e., ‘psycho-fatigue’, ‘gastrointestinal’, ‘neuro-skeletal’ and ‘pharyngeal-respiratory’) was recognized using FMM. During model fitting, we observed that a two-classes/four-factors model allowing for free intercepts, covariances and means across latent classes had lowest BIC (346926.053) and entropy 0.995. These values indicating that individuals are correctly classified by our fitted model. The two identified classes were labeled as “high psycho-fatigue complaints” and “low psycho-fatigue complaints”; because participants in the first class experienced higher scores of psychological-fatigue somatic symptoms (mean: 0.249 in first class vs. 0 in second class) while experienced lower scores of somatoform symptoms in physical domains (gastrointestinal, neuro-skeletal and pharyngeal-respiratory). There were 4243
8
individuals (89%) in the low psycho-fatigue complaints class and 519 participants (11%) in the high psycho-fatigue complaints class. According to the two-class four-factor FMM solution, approximately, all items are significantly loaded on their respective factor (Heidari et al., 2017a). Table 3 contains regression coefficients on the association of psychological disorders profile with somatoform symptoms profiles in the two extracted classes. In both classes, the profile of psychological disorders is significantly related to the four somatoform symptoms profiles. The regression coefficient of psychological disorders profile was near 0.3 in both classes. There was a significant positive relationship between the profile of psychological disorders and gastrointestinal profile in both classes and its regression coefficient was greater in high psychofatigue complaints class (0.152 and 0.147 in high and low psycho-fatigue complaints classes, respectively; p<0.0001). Furthermore, the regression coefficients for neuro-skeletal profile were 0.219 and 0.210 in high and low psycho-fatigue complaints classes, respectively; (p<0.0001). It can be seen that there is significant positive relationship between psychological disorders profile and pharyngeal –respiratory profile in both classes; however, its regression coefficient on pharyngeal –respiratory profile was greater in high psycho-fatigue complaints class (0.134 and 0.08 in high and low psycho-fatigue complaints classes, respectively; p< 0.0001). (Table 3) The prevalence of the somatoform symptoms in two extracted latent classes is presented in Table 3, too. Although, majority of symptoms had a four-point Likert scale; we only reported the response of participants to ‘often and always’ categories. The prevalence of all 30 somatoform symptoms for participants assigned to the "high psycho-fatigue complaints" class was significantly higher than other class (P<0.0001). In the "often" category, the most common somatoform symptoms reported by participants assigned to "high psycho-fatigue complaints" class were “dry mouth” (45%), “severe fatigue” (39%), followed by “headache”, back pain”,
9
“sleep disorder”, and “feeling low on energy” (all about 24%). In the "always" category, the most frequent somatoform symptoms reported by participants assigned to the "high psychofatigue complaints" class were “severe fatigue” (15%), followed by “feeling low on energy” (13.6%), “pain in joints” (13%), “disturbing thoughts” (12.8%), “back pain” (12.1%) and “dry mouth” (10.5%). (Table 3)
4. Discussion In this cross-sectional population-based study, two classes characterized by high (11%) and low (89%) levels of psychological-fatigue complaints, and four-factors profiles (i.e., ‘psycho-fatigue’, ‘gastrointestinal’, ‘neuro- skeletal’ and ‘pharyngeal-respiratory’) representing underlying structure of the somatoform symptoms were identified from a large sample of Iranian adults using factor mixture modeling (FMM). We observed that the prevalence of all somatoform symptoms for participants assigned to "high psycho-fatigue complaints" class was significantly higher than" low psycho-fatigue complaints" class. Although, fitted FMM led to two classes with mentioned characteristics, however, it should be noted that participants in both classes suffered from other somatoform symptoms profiles with different degrees. We did not find any study such as current study, which stratified a general large population into homogeneous subgroups based on somatoform symptoms using FMM. However, other statistical approaches i.e. clustering, factor analysis and LCA in some studies were used to classify psychosomatic symptoms (Fink et al., 2007; Gara et al., 1998; Kato et al., 2010; Lacourt et al., 2013; Nimnuan et al., 2001). For instance, in the Fink et al.’s study on 978 internal medical, neurological, and primary care patients, a distinct pattern of cardiopulmonary (CP), musculoskeletal/pain (MS), and gastrointestinal (GI) symptom factors as well as three classes of 10
patients including “non-bodily distress”, “modest bodily distress”, and “severe bodily distress” were identified (Fink et al., 2007). The aforesaid study had similarities with our study in terms of gastrointestinal and skeletal profiles. In the Gara et al.’s study 11 clusters of patients with different patterns of medically unexplained symptoms, were identified using hierarchical cluster analysis (Gara et al., 1998). The observed disparities in results of conducted studies in this area of subject can be attributed to geographic, socio-economic status, culture and racial dependency of somatoform symptoms. It is believed that somatoform symptoms are the manifestations of psychological illnesses in the form of physical symptoms (Wong et al., 2015) and there are evidences that psychological disorders are risk factors for somatoform symptoms (Zhu et al., 2012). Majority of previous studies were restricted to the association of psychological disorders with a few somatoform symptoms or with an overall score of somatization (Kinnunen et al., 2010; Wong et al., 2015; Zhu et al., 2012). In current study, three psychological disorders i.e. anxiety, depression and psychological distress were combined (as a latent factor) and its collective association with somatoform symptoms profiles was examined. We observed that the profile of psychological disorders is positively associated with the four somatoform symptoms profiles with greater coefficient in the high psycho-fatigue complaints class, except for psycho-fatigue profile. Previous studies have emphasized on the strong association of psychological disorders with somatoform symptoms. Wong et al. demonstrated that anxiety is a modifiable risk factor for psychosomatic symptoms in general Chinese populations (Wong et al., 2015). Kinnunen et al.’s study showed that nearly all psychosomatic symptoms are associated with mental health symptoms (Kinnunen et al., 2010). The results of Zhu et al.’s study on 2408 clinical patients, revealed that “depression and anxiety” are main risk factors for high somatic symptoms (Zhu et 11
al., 2012). Koh et al.’s study found that both anxiety and depression have direct effects on somatic symptoms in patients suffering from these disorders (Koh et al., 2008). Dales et al. indicated that psychological problems are important determinants of respiratory symptoms (such as cough, wheeze and dyspnea) (Dales et al., 1989). The influence of psychological disorders, such as anxiety and depression, on somatoform symptoms can be explained from biological perspectives (Wong et al., 2015). Psychological disorders have major role in initiation and development of gastrointestinal symptoms potentially via mechanisms involving immune modulation and alteration brain processing of incoming sensory signals (Wouters and Boeckxstaens, 2016). Also, regarding to the association of psychological disorders profile with neuro-skeletal profile of somatic symptoms, the possible role of neurotransmitters and cytokine receptors (Trivedi, 2004; Vargas-Prada and Coggon, 2015; Walker et al., 2014) could be mentioned. It is important to recognize some strengths and limitations of the present study. A major strength of our large population based study is the application of factor mixture model for identifying profiles of somatoform symptoms, and classifying study population, simultaneously, instead of dealing with them, separately. Furthermore, psychological disorders profile was evaluated in identified classes through FMM. However, due to the cross-sectional nature of SEPAHAN design, we could not infer cause–effect relationships from our findings. All used information in the present analysis was collected by self-administered questionnaires that might lead to misclassifying the participants. Finally, because SEPAHAN study’s participants were health centers staffs, thus, generalization of the present findings to the Iranian general population must be done with caution.
12
In summary, our study’s findings, in the context of an observational study, suggested that somatoform symptoms had a dimensional-categorical structure within our population that could be useful for dealing with treatment’s approaches. In addition, we showed that the profile of psychological disorders is significantly associated with different somatoform symptoms profiles. Conflicts of interest None Acknowledgements The present article was extracted from a Biostatistics PhD thesis at the School of Health, Isfahan University of Medical Sciences, with project number 394832. SEPAHAN was financially supported by a grant from the Vice Chancellery for Research and Technology, Isfahan University of Medical Sciences (IUMS). We are grateful to thank all staff of Isfahan University of Medical Sciences (MUI) who kindly participated in our study and staff of Public Relations Unit, and other authorities of IUMS for their excellent cooperation.
13
References Adibi, P., Keshteli, A.H., Esmaillzadeh, A., Afshar, H., Roohafza, H., Bagherian-Sararoudi, R., Daghaghzadeh, H., Soltanian, N., Feinle-Bisset, C., Boyce, P., 2012. The study on the epidemiology of psychological, alimentary health and nutrition (SEPAHAN): overview of methodology. J. Res. Med. Sci. 17. Bener, A., Dafeeah, E.E., Chaturvedi, S.K., Bhugra, D., 2013. Somatic symptoms in primary care and psychological comorbidities in Qatar: neglected burden of disease. Int. Rev. Psychiatry 25, 100–106. Caballero, L., Aragonès, E., García-Campayo, J., Rodríguez-Artalejo, F., Ayuso-Mateos, J.L., Polavieja, P., Gómez-Utrero, E., Romera, I., Gilaberte, I., 2008. Prevalence, characteristics, and attribution of somatic symptoms in Spanish patients with major depressive disorder seeking primary health care. Psychosomatics 49, 520–529. Dales, R.E., Spitzer, W.O., Schechter, M.T., Suissa, S., 1989. The influence of psychological status on respiratory symptom reporting. Am. Rev. Respir. Dis. 139, 1459–1463. doi:10.1164/ajrccm/139.6.1459 Escobar, J.I., Cook, B., Chen, C.-N., Gara, M.A., Alegría, M., Interian, A., Diaz, E., 2010. Whether medically unexplained or not, three or more concurrent somatic symptoms predict psychopathology and service use in community populations. J. Psychosom. Res. 69, 1–8. Fink, P., Toft, T., Hansen, M.S., Ørnbøl, E., Olesen, F., 2007. Symptoms and syndromes of bodily distress: an exploratory study of 978 internal medical, neurological, and primary care patients. Psychosom. Med. 69, 30–39. Gara, M.A., Silver, R.C., Escobar, J.I., Holman, A., Waitzkin, H., 1998. A hierarchical classes analysis (HICLAS) of primary care patients with medically unexplained somatic symptoms.
14
Psychiatry Res. 81, 77–86. Gonzalez, D.S., Rodríguez, M., García, C., Prieto, R., Saiz-Ruiz, J., 2009. Gender differences in major depressive disorder: somatic symptoms and quality of life. Rev. Psiquiatr. y salud Ment. 2, 119–127. doi:10.1016/S1888-9891(09)72402-4 Haftgoli, N., Favrat, B., Verdon, F., Vaucher, P., Bischoff, T., Burnand, B., Herzig, L., 2010. Patients presenting with somatic complaints in general practice: depression, anxiety and somatoform disorders are frequent and associated with psychosocial stressors. BMC Fam. Pract. 11, 1-8. Han, H., Wang, S.-M., Han, C., Lee, S.-J., Pae, C.-U., 2014. The relationship between somatic symptoms and depression. Neuro Endocrinol. Lett. 35, 463–469. Heidari, Z., Feizi, A., Roohafza, H., Hassanzadeh Keshteli, A., Shiravi, F.Z., Adibi, P., 2017a. Demographic and life styles determinants of somatic complaints’ structures: a crosssectional study on a large sample of Iranian adults using factor mixture model. Int. J. Prev. Med. 8, 8. doi: 10.4103/2008-7802.200526 Heidari, Z., Keshteli, A.H., Feizi, A., Afshar, H., Adibi, P., 2017b. Somatic complaints are significantly associated with chronic uninvestigated dyspepsia and its symptoms: a large cross-sectional population based study. J. Neurogastroenterol. Motil. 23, 80–91. doi:10.5056/jnm16020 Kato, K., Sullivan, P.F., Pedersen, N.L., 2010. Latent class analysis of functional somatic symptoms in a population-based sample of twins. J. Psychosom. Res. 68, 447–453. Kinnunen, P., Laukkanen, E., Kylmä, J., 2010. Associations between psychosomatic symptoms in adolescence and mental health symptoms in early adulthood. Int. J. Nurs. Pract. 16, 43– 50.
15
Koh, K.B., Kim, D.K., Kim, S.Y., Park, J.K., Han, M., 2008. The relation between anger management style, mood and somatic symptoms in anxiety disorders and somatoform disorders. Psychiatry Res. 160, 372–379. Lacourt, T., Houtveen, J., van Doornen, L., 2013. “Functional somatic syndromes, one or many?”: An answer by cluster analysis. J. Psychosom. Res. 74, 6–11. Lubke, G., Muthén, B.O., 2007. Performance of factor mixture models as a function of model size, covariate effects, and class-specific parameters. Struct. Equ. Model. A Multidiscip. J. 14, 26–47. doi:10.1207/s15328007sem1401_2 Lubke, G.H., Muthén, B., 2005. Investigating population heterogeneity with factor mixture models. Psychol. Methods 10, 21. Manshaee, G.R., Hamidi, E., 2013. Prevalence of psychosomatic symptoms among adolescent’s computer users. Procedia-Social Behav. Sci. 84, 1326–1332. Montazeri, A., Harirchi, A.M., Shariati, M., Garmaroudi, G., Ebadi, M., Fateh, A., Toscani, F., Borreani, C., Boeri, P., Miccinesi, G., 2005. The 12-item General Health Questionnaire (GHQ-12): translation and validation study of the Iranian version. Health Qual. Life Outcomes 1, 66. Montazeri, A., Vahdaninia, M., Ebrahimi, M., Jarvandi, S., 2003. The Hospital Anxiety and Depression Scale (HADS): translation and validation study of the Iranian version. Health Qual. Life Outcomes 1, 14. doi:10.1186/1477-7525-1-14 National Collaborating Centre (N.C.C.) for Nursing and Supportive Care (UK), 2008. Irritable Bowel Syndrome in Adults: Diagnosis and Management of Irritable Bowel Syndrome in Primary Care [Internet]. London: Royal College of Nursing (UK); 2008 Feb. Appendix J, The General Practice Physical Activity Questionnaire (GPPAQ). Available from:
16
https://www.ncbi.nlm.nih.gov/ books/NBK51962/. [Last accessed on 2006 Oct]. Nimnuan, C., Rabe-Hesketh, S., Wessely, S., Hotopf, M., 2001. How many functional somatic syndromes? J. Psychosom. Res. 51, 549–557. Novy, D., Berry, M.P., Palmer, J.L., Mensing, C., Willey, J., Bruera, E., 2005. Somatic symptoms in patients with chronic non-cancer-related and cancer-related pain. J. Pain Symptom Manage. 29, 603–612. Schwarz, G., 1978. Estimating the dimension of a model. Ann. Stat. 6, 461–464. doi:10.1214/aos/1176344136 Shabbeh, Z., Feizi, A., Afshar, H., Hassanzade Kashtali, A., Adibi, P., 2016. Identifying the profiles of psychosomatic disorders in an iranian adult population and their relation to psychological problems. J. Maz. Univ. Med. Sci. 26, 82–94 [In Persian]. Shidhaye, R., Mendenhall, E., Sumathipala, K., Sumathipala, A., Patel, V., 2013. Association of somatoform disorders with anxiety and depression in women in low and middle income countries: A systematic review. Int. Rev. Psychiatry 25, 65–76. Simon, G.E., VonKorff, M., Piccinelli, M., Fullerton, C., Ormel, J., 1999. An international study of the relation between somatic symptoms and depression. N. Engl. J. Med. 341, 1329– 1335. Spitzer, R.L., Kroenke, K., Williams, J.B.W., Group, P.H.Q.P.C.S., 1999. Validation and utility of a self-report version of PRIME-MD: the PHQ primary care study. Jama 282, 1737–1744. Steinbrecher, N., Koerber, S., Frieser, D., Hiller, W., 2011. The prevalence of medically unexplained symptoms in primary care. Psychosomatics 52, 263–271. Sugahara, H., Akamine, M., Kondo, T., Fujisawa, K., Yoshimasu, K., Tokunaga, S., Kubo, C., 2004. Somatic symptoms most often associated with depression in an urban hospital
17
medical setting in Japan. Psychiatry Res. 126, 151–158. Trivedi, M.H., 2004. The link between depression and physical symptoms. Prim Care Companion J Clin Psychiatry 6, 12–16. Vargas-Prada, S., Coggon, D., 2015. Psychological and psychosocial determinants of musculoskeletal pain and associated disability. Best Pract. Res. Clin. Rheumatol. 29, 374– 390. Walker, A.K., Kavelaars, A., Heijnen, C.J., Dantzer, R., 2014. Neuroinflammation and comorbidity of pain and depression. Pharmacol. Rev. 66, 80–101. Wong, J.Y.-H., Fong, D.Y.-T., Chan, K.K.-W., 2015. Anxiety and insomnia as modifiable risk factors for somatic symptoms in Chinese: a general population-based study. Qual. Life Res. 24, 2493–2498. Wouters, M.M., Boeckxstaens, G.E., 2016. Is there a causal link between psychological disorders and functional gastrointestinal disorders? Expert Rev. Gastroenterol. Hepatol. 10, 5–8. Xiong, N., Fritzsche, K., Wei, J., Hong, X., Leonhart, R., Zhao, X., Zhang, L., Zhu, L., Tian, G., Nolte, S., 2015. Validation of patient health questionnaire (PHQ) for major depression in Chinese outpatients with multiple somatic symptoms: A multicenter cross-sectional study. J. Affect. Disord. 174, 636–643. Zhu, C., Ou, L., Geng, Q., Zhang, M., Ye, R., Chen, J., Jiang, W., 2012. Association of somatic symptoms with depression and anxiety in clinical patients of general hospitals in Guangzhou, China. Gen. Hosp. Psychiatry 34, 113–120. doi:10.1016/j.genhosppsych.2011.09.005
18
Table 1. Factor loadings for the four extracted somatoform symptoms profiles from 30 somatoform symptoms Factor Loadings a Somatoform Symptoms Psycho-Fatigue Sleep Disorder
0.46
Pounding heart
0.41
Feeling low on energy
0.69
Feeling like ‘butterflies’
0.78
Difficulty concentrating
0.64
Disturbing thoughts
0.80
Gastrointestinal
NeuroSkeletal
Pharyngeal Respiratory
0.41
Chest pain
0.52
Feeling of fullness
0.69
Nausea
0.50
Gastroesophageal reflux
0.54
Pain or discomfort in the abdomen
0.71
Constipation
0.49
Diarrhea
0.36
Bloating or swelling of the abdomen
0.67
Anal pain
0.48
Headache
0.57
Back pain
0.66
Pain in joints
0.64
Eyesore
0.50
Severe fatigue
0.61
Dizziness and confusion
0.51
Chills and extreme cold
0.42
Hot flashes
0.38
Dry mouth
0.31
19
Neck pain
0.56
Globus sensation
0.55
Having trouble swallowing
0.61
Shortness of breath
0.46
Hoarseness
0.61
Wheezing (asthma)
0.52
Variance explained (%)
12.4
12.3
11.4
9.3
Cumulative variance
12.4
24.7
36.1
45.4
a
Factor loadings<0.3 are not shown for simplicity.
20
Table 2. Correlation between the scores of somatoform symptoms profiles and the scores of psychological disorders
Somatoform Symptoms Profiles Psycho-Fatigue
Gastrointestinal
Neuro-Skeletal
Pharyngeal Respiratory
Depression
0.799
0.415
0.541
0.385
Anxiety Psychological Distress
0.907
0.504
0.612
0.461
0.650
0.373
0.472
0.344
- All presented correlations are significant at the 0.01 level
21
Table 3. Comparison of psychological disorders profile and the prevalence of individual somatoform symptoms in two extracted classes.
Psycho-Fatigue Profile Sleep Disorder Pounding heart Feeling low on energy Feeling like ‘butterflies’ Difficulty concentrating Disturbing thoughts Gastrointestinal Profile Chest pain Feeling of fullness Nausea Gastroesophageal reflux Pain or discomfort in the abdomen Constipation Diarrhea Bloating or swelling of the abdomen Anal pain Neuro-Skeletal Profile
High Psycho-Fatigue Complaints class (n=519) Estimate Often b Always b (SE)a Number Number 0.282 (%) (%) * (0.011) 120 33 (6.7) (24.3) 74 (14.7) 22 (4.4) 121 68 (13.6) (24.2) 36 (7.2) 28 (5.6) 68 (13.5) 40 (8.0) 95 (19.0) 64 (12.8) 0.152 (0.014)*
74 (14.7) 13 (2.6) 91 (18.4)
33 (6.6) 4 (0.8) 44 (8.9)
35 (7.3)
11 (2.3)
Pain in joints Eyesore Severe fatigue
341 (8.2) 56 (1.3) 398 (9.9)
73 (1.8) 6 (0.1) 123 (3.1)
76 (1.9)
9 (0.2)
210 (5.0) 602 (14.5) 524 (12.7) 407 (9.8)
45 (1.1) 106 (2.6)
233 (5.6) 933 (22.6) 215 (5.2) 80 (1.9) 168 (4.0)
54 (1.3) 206 (5.0)
1170 (28.2) 98 (2.4) 52 (1.2) 0 139 (3.3) 47 (1.1) 53 (1.3)
116 (2.8)
0.210 (0.011)* 74 (14.7) 124 (24.8) 124 (24.7) 104 (20.5) 76 (15.0) 197 (39.0) 79 (15.7) 41 (8.1) 59 (11.7)
Back pain
Neck pain Globus sensation Having trouble swallowing Shortness of breath Hoarseness Wheezing (asthma)
13 (2.6) 24 (4.7) 16 (3.2) 11 (2.3) 22 (4.4)
0.219 (0.016)*
Pounding heart Headache
Dizziness and confusion Chills and extreme cold Hot flashes Pharyngeal –Respiratory Profile Dry mouth
77 (15.2) 90 (17.7) 40 (8.1) 30 (6.1) 82 (16.2)
Low Psycho-Fatigue Complaints class (n=4243) Estimate Often Always (SE)a Number Number 0.300 (%) (%) (0.011)* 411 99 (2.4) (10.1) 210 (5.0) 45 (1.1) 574 230 (5.6) (13.9) 148 (3.6) 58 (1.4) 200 (4.8) 78 (1.9) 421 202 (4.9) (10.2) 0.147 (0.008)* 146 (3.5) 16 (0.4) 230 (5.5) 40 (1.0) 78 (1.9) 22 (0.5) 50 (1.2) 16 (0.4) 276 (6.7) 44 (1.1)
22 (4.4) 45 (9.0) 61 (12.1) 66 (13.0) 23 (4.6) 76 (15.0) 29 (5.8) 9 (1.8) 15 (3.0)
0.134 (0.015)*
179 (4.3) 173 (4.2)
34 (0.8) 13 (0.3) 35 (0.8)
0.081 (0.011)* 227 (45.0) 65 (12.9) 61 (12.2) 51 (10.1) 68 (13.5) 26 (5.2) 26 (4.9)
22
53 (10.5) 13 (2.6) 18 (3.6) 5 (1.0) 31 (6.1) 5 (1.0) 6 (1.2)
17 (0.4) 10 (0.2) 0 40 (1.0) 11 (0.3) 17 (0.4)
a
Regression coefficients for the association of psychological disorders profile with somatoform symptoms profiles; *PValue< 0.0001 b The prevalence of all somatoform symptoms was significantly different between two classes (P <0.001).
23
Highlights
The aims of the current study were to classify studied population based on psychosomatic complaints profiles and evaluate the profile of psychological disorders in extracted classes. Factor mixture modeling was used with data from a sample of 4762 Iranian adults. A two-class, four-factor structure was identified for the psychosomatic complaints. The profile of psychological disorders was significantly related to the psychosomatic complaints profiles.
24