- Email: [email protected]
Factor structure and reliability of the 30-item Sinhala version of General Health Questionnaire (GHQ)
Asian Journal of Psychiatry 5 (2012) 150–153
Contents lists available at SciVerse ScienceDirect
Asian Journal of Psychiatry journal homepage: www.elsevier.com/locate/ajp
Factor structure and reliability of the 30-item Sinhala version of General Health Questionnaire (GHQ) Chrishantha Abeysena a,*, Pushpa Jayawardana a, Upali Peiris b a b
Department of Public Health, Faculty of Medicine, University of Kelaniya, Ragama, Sri Lanka Department of Psychiatry, Faculty of Medicine, University of Kelaniya, Sri Lanka
A R T I C L E I N F O
A B S T R A C T
Article history: Received 12 September 2011 Received in revised form 21 February 2012 Accepted 23 February 2012
Objective: To determine the factor structure and the reliability of the Sinhala version of the General Health Questionnaire-30 (GHQ-30). Methods: This was a descriptive study including 368 patients with in the age range of 18–75 years, attending the Out Patient Department (OPD) of Colombo North Teaching Hospital, Ragama, Sri Lanka during the period between June 2009 and September 2010. Sinhala version of GHQ-30 was given to be completed by the participants. Each item of the GHQ was rated on a four-point scale (0-1-2-3). Factor analyses were performed by applying Generalized Least Squares method using oblimin rotation. The internal consistency was assessed by calculating Cronbach’s a coefficient. Results: Median age of the study population was 32.5 years (Inter quartile range [IQR] = 21 years) and the median GHQ score was 9 (IQR 7). GHQ-30 produced a five factor solution which accounted for 51.6% of the total variance (TV). Factor I (Depression) accounted for 38% of TV, Factor II (Loss of confidence) 5%, Factor III (Insomnia) 4%, Factor IV (Social dysfunction) 2.7% and Factor V (Anxiety) 2%. The Cronbach’s alpha coefficients of GHQ-30 was 0.94 indicating satisfactory internal consistency. Conclusions: GHQ-30 comprises five factors/subscales and it displayed adequate reliability for assessment of psychiatric disorders among Sinhala speaking primary care attendees in Sri Lanka. ß 2012 Elsevier B.V. All rights reserved.
Keywords: Factor analysis General Health Questionnaire Reliability Psychometric properties
1. Introduction The General Health Questionnaire (GHQ) is a self-administered questionnaire widely used to detect potential non-psychotic psychiatric disorders. The original questionnaire consists of 60items but subsequently 30, 28 and 12 items versions have been derived from it. Validity is the extent to which a test measures what it claims to measure. It is vital for a test to be valid in order for the results to be accurately applied and interpreted. Construct validity is the extent to which a particular measure relates to other measures consistent with theoretically derived hypothesis concerning the concepts or constructs that are being measured (Abramson and Abramson, 1992). This can be assessed by examining psychometric properties and factor structure for GHQ-30. The GHQ has been validated in different languages and cultures (Huppert et al., 1989; Jakob et al., 1997). The studies that assessed the factor structure had reported it to vary between three and eight factors across the studies (Frydecka et al., 2010; Iwata et al., 1994;
* Corresponding author. Tel.: +94 11 2953411; mobile: +94 77 7591715. E-mail addresses: [email protected], [email protected] (C. Abeysena). 1876-2018/$ – see front matter ß 2012 Elsevier B.V. All rights reserved. doi:10.1016/j.ajp.2012.02.025
Ohta et al., 1995; Shek and Tsang, 1995; Shigemi et al., 2000). The Sinhala translation of it is a commonly used instrument in Sri Lanka and even though it has been reported to be validated and hence widely used for research purposes, the psychometric properties of the instrument have not been published. Thus, the objective of this study was to describe the factor structure of the Sinhala version of the GHQ-30 and to assess its reliability. 2. Methods A descriptive cross sectional study was conducted at the Out Patient Department (OPD) of the Colombo North Teaching Hospital (CNTH), Ragama, Sri Lanka between June 2009 and September 2010. The study participants included both males and females aged 18–75 years who were able to read and understand the Sinhala language. The minimal sample size for factor analysis depends on the number of items in the questionnaire and what is recommended is five participants per question item. Therefore, the minimum required for the GHQ-30 was 150 participants, but a sample size of 323 was computed which will include, on average, 10 participants per item which is preferable. All patients registered for OPD visits at the CNTH and who consented to participate in the study and were eligible for inclusion were recruited until the required sample size was
C. Abeysena et al. / Asian Journal of Psychiatry 5 (2012) 150–153
151
Table 1 Factor loadings of the Sinhala version of GHQ-30. Factor F1 Depression 25. Felt that life is entirely hopeless? 24. Been thinking of you as a worthless person? 23. Been loosing confidence in you? 29. Felt that life is not worth living? 26. Been feeling hopeful about your own future? 21. Found everything getting on top of you? 30. Found at times that you could not do anything because your nerves were too bad? 7. (Not) Felt on the whole you were doing things well? 8. (Not) Been satisfied with the way you have carried out your task? 17. (Not) Been able to enjoy your day to day activities? 1. (Not) Been able to concentrate on whatever you are doing? 12. (Not) Felt that you are playing a useful part in things? 2. Lost much sleep over worry? 3. Been having restless disturbed nights? 6. Been managing as well as most people would in your shoes? 10. Been finging it easy to get on with other people? 11. Spent much time chatting with people? 13. Felt capable of making decisions about things? 20. Been able to face up to your problems? 4. Been managing to keep your self busy and occupied? 5. Been getting out of the house as much as usual? 9. Did you develop love and fondness to your dear ones? 18. Been taking things hard? 19. Been getting scared or panicky for no good reason? 22. Been feeling unhappy and depressed? 28. Been feeling nervous and strung up all the time? 14. Felt constantly under strain? 15. Felt you could not overcome your difficulties? 16. Been finding life a struggle all the time? 21. Found everything getting on top of you?
0.886 0.816 0.675 0.642 0.563 0.486 0.434 0.054 0.029 0.029 0.032 0.265 0.002 0.118 0.067 0.040 0.044 0.055 0.249 0.122 0.056 0.112 0.007 0.143 0.284 0.183 0.097 0.193 0.268 0.168
F2 Loss of confidence 0.011 0.003 0.158 0.111 0.087 0.102 0.042 0.911 0.749 0.454 0.367 0.362 0.078 0.014 0.244 0.000 0.070 0.197 0.074 0.218 0.065 0.149 0.058 0.055 0.077 0.092 0.014 0.092 0.027 0.130
F3 Insomnia 0.080 0.009 0.052 0.125 0.055 0.087 0.034 0.063 0.051 0.079 0.165 0.009 0.755 0.695 0.269 0.089 0.091 0.147 0.098 0.159 0.097 0.099 0.162 0.016 0.186 0.136 0.221 0.082 0.111 0.101
F4 Social dysfunction 0.066 0.057 0.183 0.190 0.142 0.012 0.278 0.133 0.023 0.271 0.110 0.198 0.022 0.000 0.256 0.598 0.536 0.501 0.487 0.365 0.332 0.227 0.003 0.031 0.022 0.116 0.230 0.183 0.139 0.264
F5 Anxiety 0.008 0.098 0.266 0.009 0.032 0.348 0.185 0.061 0.003 0.223 0.184 0.177 0.066 0.036 0.007 0.038 0.037 0.097 0.067 0.136 0.133 0.003 0.682 0.545 0.488 0.468 0.448 0.406 0.388 0.304
Extraction method: Generalized Least Squares. Rotation method: Oblimin with Kaiser Normalization. Factor loading >0.3 were bolded.
reached using convenience sampling. The number of participants recruited per day was between 10 and 15. The main study instruments used were a questionnaire on general information to determine relevant socio demographic data and the Sinhala version of the GHQ-30, which is a selfadministered questionnaire. It was completed during the OPD visit. Each item of the GHQ was rated on a four-point Likert scale (0-1-2-3) instead of the conventional bi-modal scale (0-0-1-1) designed for the GHQ, to enable the statistical procedure in relation to factor analysis. Data was entered to EPI info and then converted to SPSS file and analysed. Factorability of the GHQ-30 was assessed with the Bartlett’s test of sphericity (<0.001) and Kaiser–Meyer–Olkin (0.95) measure of sampling adequacy. We applied Generalized Least Squares method with Oblimin rotation, because it produced five distinct interpretable factors. The internal consistency was assessed by calculating Cronbach’s a coefficient of the instrument for the overall sample as well as for selected socio-demographic variables and individual factors/subscales. Informed consent was obtained from all the research participants and confidentiality of data obtained were assured by concealing the identity using a coding system. Ethical clearance was granted by the Ethics Review Committee, Faculty of Medicine, University of Kelaniya.
3. Results Total number of OPD patients who were invited to the study during the period between May 2009 and January 2011 was 432. Of them 22 patients were not willing to participate which gave a nonparticipant rate of 5.0%. Therefore, the total patients recruited to the study was 410. Of them 42 (10%) had missing data for one or
more items of the GHQ-30, which left a study sample of 368 for the final analysis. Median age of the study population was 32.5 years (inter quartile range [IQR] = 21 years). The highest proportion (28.5%; n = 117) of them was in the age group of 21–30 years. The number of males among the study participants was 133 (35%). Fifty nine percent (n = 242) of the study population was Buddhists and 26.6% (n = 110) Catholics or Christians. Two hundred and twenty two (54%) had studied up to General Certificate Examination (Ordinary Level) [GCEOL]. 3.1. Factor structure of the GHQ-30 The median score (based on the Likert’s scale) of the GHQ-30 was 22 (IQR 15) for the total sample and 21 for both males (IQR 21) and females (IQR 13) when analyzed separately. For the age group <32 years, it was 20 (IQR 14) and 32 years, 23 (IQR 16). For the low (up to GCEOL) and high educational categories it was 23 (IQR 16.5) and 19 (IQR 13.5) respectively. Factor analysis produced a five factor solution (Table 1). The five factors together accounted for 51.6% of the total variance (TV). Considering the contribution of individual factors, Factor I (Depression) accounted for 38% of the TV, Factor II (Loss of confidence) 5%, Factor III (Insomnia) 4%, Factor IV (Social dysfunction) 2.7% and Factor V (Anxiety) 2%. Apart from item 21 (Found everything getting on top of you), there was a clear distinction between the items that loaded on the five factors. The item 21 loaded on both depression and anxiety (Table 2). 3.2. Internal consistency of GHQ-30 The Cronbach’s a coefficient of GHQ-30 for the overall sample was found to be 0.94 indicating satisfactory internal consistency of
C. Abeysena et al. / Asian Journal of Psychiatry 5 (2012) 150–153
152 Table 2 Inter-factor correlations.
Factor 1 Depression Factor 2 Loss of confidence Factor 3 Insomnia Factor 4 Social dysfunction Factor 5 Anxiety
Factor 1 Depression
Factor 2 Loss of confidence
Factor 3 Insomnia
Factor 4 Social dysfunction
Factor 5 Anxiety
1.00
0.45
0.29
0.48
0.56
0.45
1.00
0.23
0.53
0.38
0.29
0.23
1.00
0.32
0.39
0.48
0.53
0.32
1.00
0.39
0.56
0.38
0.39
0.39
1.00
Table 3 Item-total correlation coefficients of GHQ-30 for the whole sample and the subgroups.
Item Item Item Item Item Item Item Item Item Item Item Item Item Item Item Item Item Item Item Item Item Item Item Item Item Item Item Item Item Item
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30
Total sample
Male
Female
Age <32
Age 32
Low education
High education
0.53 0.50 0.49 0.49 0.39 0.54 0.54 0.57 0.30 0.53 0.37 0.45 0.53 0.69 0.61 0.65 0.60 0.63 0.58 0.59 0.67 0.76 0.74 0.68 0.70 0.66 0.69 0.71 0.63 0.73
0.56 0.57 0.53 0.56 0.44 0.58 0.59 0.68 0.42 0.56 0.46 0.53 0.58 0.78 0.60 0.75 0.61 0.71 0.64 0.64 0.73 0.78 0.79 0.68 0.67 0.74 0.77 0.73 0.68 0.80
0.53 0.52 0.48 0.50 0.35 0.52 0.49 0.49 0.23 0.54 0.32 0.38 0.48 0.64 0.62 0.62 0.60 0.62 0.56 0.57 0.64 0.76 0.73 0.71 0.74 0.61 0.66 0.69 0.63 0.69
0.43 0.52 0.50 0.44 0.36 0.51 0.41 0.56 0.24 0.51 0.37 0.49 0.55 0.65 0.58 0.63 0.58 0.58 0.61 0.58 0.66 0.73 0.72 0.70 0.65 0.66 0.69 0.67 0.58 0.69
0.62 0.48 0.48 0.54 0.45 0.58 0.69 0.59 0.37 0.55 0.38 0.41 0.50 0.71 0.64 0.66 0.63 0.67 0.54 0.60 0.67 0.79 0.76 0.68 0.74 0.66 0.69 0.73 0.67 0.76
0.46 0.47 0.50 0.51 0.44 0.53 0.49 0.51 0.35 0.51 0.37 0.46 0.50 0.69 0.58 0.64 0.58 0.62 0.56 0.62 0.67 0.75 0.75 0.69 0.73 0.64 0.66 0.70 0.71 0.74
0.64 0.53 0.47 0.49 0.38 0.55 0.59 0.69 0.19 0.56 0.35 0.41 0.58 0.68 0.64 0.65 0.63 0.64 0.62 0.57 0.67 0.77 0.74 0.69 0.65 0.68 0.71 0.69 0.52 0.70
the scale. It was 0.96 for males, 0.94 for females, 0.95 for the age group 32 years and 0.94 for age group <32 years as well as for both educational categories. The corrected item-total correlation coefficients were above 0.40 for the overall sample as well as for the sub-samples (Table 3). The Cronbach’s a coefficients by individual factors/subscales were 0.91 for ‘‘Depression’’, 0.79 for ‘‘Loss of confidence’’, 0.75 for ‘‘Insomnia’’, 0.74 for ‘‘Social dysfunction’’ and 0.90 for ‘‘Anxiety’’. 4. Discussion We found that GHQ-30 consisted of five factors which were named as Depression, Loss of confidence, Insomnia, Social dysfunction and Anxiety which together accounted for 51.6% of the total variance. The factor one was Depression alone accounted for 38% of the total explained variance. All factors were moderately inter-correlated. Item 21 of the GHQ-30 (Found everything getting on top of you?) being common to both, factors on ‘Depression’ and ‘Anxiety’, and demonstrates it’s inability to discriminate between the two constructs. Items 6 (Been managing as well as most people would in your shoes?) and 9 (Did you develop love and fondness to
your dear ones?) were not loaded to any of the factors extracted which suggests its non contribution. The factor structure of the GHQ-30 reported had shown variations from culture to culture. However, Huppert et al. (1989) reported a five factor structure namely Anxiety, Feelings of incompetence, Depression, Difficulty in coping and Social dysfunction which was more or less similar to that of our study. The subscales reported in other studies varied from three (Frydecka et al., 2010), four (Shek, and Tsang, 1995), six and seven (Iwata et al., 1994; Shigemi et al., 2000) factor structures. It is observed that Depression and Anxiety are common subscales to all the cultures detection of which is one of the primary objectives with the use of this instrument. Shek and Tsang (1995) from China, reported that it had a four factor structure namely Anxiety, Depression, Interpersonal dysfuctioning and Adequate coping. Two studies (Iwata et al., 1994; Shigemi et al., 2000) from Japan reported to have six and seven factors respectively. Of the two, in the study reported by Iwata et al. (1994) the six factors extracted were Anxiety, Depressive thoughts, Inhibition of thinking, Social dysfunction, Loss of positive attitudes and Hyperaction. The other study reported by Shigemi et al. (2000) described six factors namely Anxiety and depression, Severe
C. Abeysena et al. / Asian Journal of Psychiatry 5 (2012) 150–153
depression, Lack of interpersonal network or social support, Poor social activity, Insomnia and Loss of positive attitudes. The items related to Insomnia in the later were similar to that of our study. Most of the items related to ‘Social dysfunction’ and ‘Loss of confidence’ in the present study were categorized into factors on ‘Inhibition of thinking’ and ‘Social dysfunction’ in the study reported by Iwata et al. (1994) and ‘Lack of interpersonal network or social support’ and ‘Poor social activity’ in the study reported by Shigemi et al. (2000). Similar to our study, item 6 was not loaded on to any of the factors extracted by Shigemi et al. (2000). We could not perform factor analysis by sex or age groups because of the inadequacy of subgroup sample sizes. The internal consistency of the GHQ-30 in our total sample was 0.94, which was higher than in the previous studies published (Iwata et al., 1994; Shigemi et al., 2000). It is clearly demonstrated from the above that direct comparison between studies can be made only with regard to the number of factors derived, but not the structure within the derived factors, if all relevant details are not published. The nomenclature of the factors derived are arbitrarily chosen, based on the factor structure, which has the tendency to vary from culture to culture thus making it a futile exercise to describe the differences between the structure of individual factors. Therefore, for comparison purposes, what matters is the psychometric properties of the summated score of all the items included in the GHQ, which has been shown to have consistency across studies in relation to adequate validity and reliability. We recruited the study participants from the OPD patients without applying any probability sampling method, which can be interpreted as a limitation to the study as this affects the genralizability of the results to all OPD attendees of this tertiary care institution. Despite above the minimal non-response rate observed (5%; n = 22), it would be described as a strength, which is likely to mitigate the selection bias which would have otherwise exerted a serious impact on the external validity of the study. In conclusion, according to the results of this study, the GHQ-30 comprises five factors and it displayed adequate reliability for assessment of psychiatric disorders among Sinhala speaking primary care attendees in Sri Lanka. It also indicates that the factor structure of the Sinhala version is similar to that of the GHQ30 in other cultural settings. Therefore, the use of the Sinhala version of GHQ-30 may be recommended to screen common psychiatric disorders in primary care attendees.
153
Role of funding source Funding for this study was provided by the University of Kelaniya; the university had no further role in study design; in the collection, analysis and interpretation of data; in the writing of the report; and in the decision to submit the paper for publication. Contributors All authors designed the study and wrote the protocol. CA managed the literature searches and analyses and undertook the statistical analysis. CA wrote the first draft of the manuscript. All authors contributed to and have approved the final manuscript. Conflict of interest The authors declare that they have no conflicts of interest. Acknowledgements We are grateful to two research assistants who assisted for data collection. References Abramson, J.H., Abramson, Z.H., 1992. Survey Methods in Community Medicine. Churchil Livingston, London. Frydecka, D., Małyszczak, K., Chachaj, A., Kiejna, A., 2010. Factorial structure of the general health questionnaire (GHQ-30). Psychiatr. Pol. 44 (3), 341–359. Huppert, F.A., Walters, D.E., Day, N.E., Elliott, B.J., 1989. The factor structure of the General Health Questionnaire (GHQ-30). A reliability study on 6317 community residents. Br. J. Psychiatry 155, 178–185. Iwata, N., Uno, B., Suzuki, T., 1994. Psychometric properties of the 30-item version general Health Questionnaire in Japanese. Jpn. J. Psychiatry Neurol. 48 (3), 547–556. Jakob, K.S., Bhugra, D., Mann, H., 1997. The validation of the 12-item General Health Questionnaire among ethnic Indian women living in the United Kingdom. Psychol. Med. 27, 1215–1217. Ohta, Y., Kawasaki, N., Araki, K., Mine, M., Honda, S., 1995. The factor structure of the general health questionnaire (GHQ-30) in Japanese middle-aged and elderly residents. Int. J. Soc. Psychiatry 41 (4), 268–275. Shek, D.T., Tsang, S.K., 1995. Reliability and factor structure of the Chinese GHQ-30 for parents with preschool mentally handicapped children. J. Clin. Psychol. 51 (2), 227. Shigemi, J., Mino, Y., Tsuda, T., 2000. Stability of factor structure and correlation with perceived job stress in General Health Questionnaire: a three-wave survey over one year in Japanese workers. J. Occup. Health 42, 284–291.
Contents lists available at SciVerse ScienceDirect
Asian Journal of Psychiatry journal homepage: www.elsevier.com/locate/ajp
Factor structure and reliability of the 30-item Sinhala version of General Health Questionnaire (GHQ) Chrishantha Abeysena a,*, Pushpa Jayawardana a, Upali Peiris b a b
Department of Public Health, Faculty of Medicine, University of Kelaniya, Ragama, Sri Lanka Department of Psychiatry, Faculty of Medicine, University of Kelaniya, Sri Lanka
A R T I C L E I N F O
A B S T R A C T
Article history: Received 12 September 2011 Received in revised form 21 February 2012 Accepted 23 February 2012
Objective: To determine the factor structure and the reliability of the Sinhala version of the General Health Questionnaire-30 (GHQ-30). Methods: This was a descriptive study including 368 patients with in the age range of 18–75 years, attending the Out Patient Department (OPD) of Colombo North Teaching Hospital, Ragama, Sri Lanka during the period between June 2009 and September 2010. Sinhala version of GHQ-30 was given to be completed by the participants. Each item of the GHQ was rated on a four-point scale (0-1-2-3). Factor analyses were performed by applying Generalized Least Squares method using oblimin rotation. The internal consistency was assessed by calculating Cronbach’s a coefficient. Results: Median age of the study population was 32.5 years (Inter quartile range [IQR] = 21 years) and the median GHQ score was 9 (IQR 7). GHQ-30 produced a five factor solution which accounted for 51.6% of the total variance (TV). Factor I (Depression) accounted for 38% of TV, Factor II (Loss of confidence) 5%, Factor III (Insomnia) 4%, Factor IV (Social dysfunction) 2.7% and Factor V (Anxiety) 2%. The Cronbach’s alpha coefficients of GHQ-30 was 0.94 indicating satisfactory internal consistency. Conclusions: GHQ-30 comprises five factors/subscales and it displayed adequate reliability for assessment of psychiatric disorders among Sinhala speaking primary care attendees in Sri Lanka. ß 2012 Elsevier B.V. All rights reserved.
Keywords: Factor analysis General Health Questionnaire Reliability Psychometric properties
1. Introduction The General Health Questionnaire (GHQ) is a self-administered questionnaire widely used to detect potential non-psychotic psychiatric disorders. The original questionnaire consists of 60items but subsequently 30, 28 and 12 items versions have been derived from it. Validity is the extent to which a test measures what it claims to measure. It is vital for a test to be valid in order for the results to be accurately applied and interpreted. Construct validity is the extent to which a particular measure relates to other measures consistent with theoretically derived hypothesis concerning the concepts or constructs that are being measured (Abramson and Abramson, 1992). This can be assessed by examining psychometric properties and factor structure for GHQ-30. The GHQ has been validated in different languages and cultures (Huppert et al., 1989; Jakob et al., 1997). The studies that assessed the factor structure had reported it to vary between three and eight factors across the studies (Frydecka et al., 2010; Iwata et al., 1994;
* Corresponding author. Tel.: +94 11 2953411; mobile: +94 77 7591715. E-mail addresses: [email protected], [email protected] (C. Abeysena). 1876-2018/$ – see front matter ß 2012 Elsevier B.V. All rights reserved. doi:10.1016/j.ajp.2012.02.025
Ohta et al., 1995; Shek and Tsang, 1995; Shigemi et al., 2000). The Sinhala translation of it is a commonly used instrument in Sri Lanka and even though it has been reported to be validated and hence widely used for research purposes, the psychometric properties of the instrument have not been published. Thus, the objective of this study was to describe the factor structure of the Sinhala version of the GHQ-30 and to assess its reliability. 2. Methods A descriptive cross sectional study was conducted at the Out Patient Department (OPD) of the Colombo North Teaching Hospital (CNTH), Ragama, Sri Lanka between June 2009 and September 2010. The study participants included both males and females aged 18–75 years who were able to read and understand the Sinhala language. The minimal sample size for factor analysis depends on the number of items in the questionnaire and what is recommended is five participants per question item. Therefore, the minimum required for the GHQ-30 was 150 participants, but a sample size of 323 was computed which will include, on average, 10 participants per item which is preferable. All patients registered for OPD visits at the CNTH and who consented to participate in the study and were eligible for inclusion were recruited until the required sample size was
C. Abeysena et al. / Asian Journal of Psychiatry 5 (2012) 150–153
151
Table 1 Factor loadings of the Sinhala version of GHQ-30. Factor F1 Depression 25. Felt that life is entirely hopeless? 24. Been thinking of you as a worthless person? 23. Been loosing confidence in you? 29. Felt that life is not worth living? 26. Been feeling hopeful about your own future? 21. Found everything getting on top of you? 30. Found at times that you could not do anything because your nerves were too bad? 7. (Not) Felt on the whole you were doing things well? 8. (Not) Been satisfied with the way you have carried out your task? 17. (Not) Been able to enjoy your day to day activities? 1. (Not) Been able to concentrate on whatever you are doing? 12. (Not) Felt that you are playing a useful part in things? 2. Lost much sleep over worry? 3. Been having restless disturbed nights? 6. Been managing as well as most people would in your shoes? 10. Been finging it easy to get on with other people? 11. Spent much time chatting with people? 13. Felt capable of making decisions about things? 20. Been able to face up to your problems? 4. Been managing to keep your self busy and occupied? 5. Been getting out of the house as much as usual? 9. Did you develop love and fondness to your dear ones? 18. Been taking things hard? 19. Been getting scared or panicky for no good reason? 22. Been feeling unhappy and depressed? 28. Been feeling nervous and strung up all the time? 14. Felt constantly under strain? 15. Felt you could not overcome your difficulties? 16. Been finding life a struggle all the time? 21. Found everything getting on top of you?
0.886 0.816 0.675 0.642 0.563 0.486 0.434 0.054 0.029 0.029 0.032 0.265 0.002 0.118 0.067 0.040 0.044 0.055 0.249 0.122 0.056 0.112 0.007 0.143 0.284 0.183 0.097 0.193 0.268 0.168
F2 Loss of confidence 0.011 0.003 0.158 0.111 0.087 0.102 0.042 0.911 0.749 0.454 0.367 0.362 0.078 0.014 0.244 0.000 0.070 0.197 0.074 0.218 0.065 0.149 0.058 0.055 0.077 0.092 0.014 0.092 0.027 0.130
F3 Insomnia 0.080 0.009 0.052 0.125 0.055 0.087 0.034 0.063 0.051 0.079 0.165 0.009 0.755 0.695 0.269 0.089 0.091 0.147 0.098 0.159 0.097 0.099 0.162 0.016 0.186 0.136 0.221 0.082 0.111 0.101
F4 Social dysfunction 0.066 0.057 0.183 0.190 0.142 0.012 0.278 0.133 0.023 0.271 0.110 0.198 0.022 0.000 0.256 0.598 0.536 0.501 0.487 0.365 0.332 0.227 0.003 0.031 0.022 0.116 0.230 0.183 0.139 0.264
F5 Anxiety 0.008 0.098 0.266 0.009 0.032 0.348 0.185 0.061 0.003 0.223 0.184 0.177 0.066 0.036 0.007 0.038 0.037 0.097 0.067 0.136 0.133 0.003 0.682 0.545 0.488 0.468 0.448 0.406 0.388 0.304
Extraction method: Generalized Least Squares. Rotation method: Oblimin with Kaiser Normalization. Factor loading >0.3 were bolded.
reached using convenience sampling. The number of participants recruited per day was between 10 and 15. The main study instruments used were a questionnaire on general information to determine relevant socio demographic data and the Sinhala version of the GHQ-30, which is a selfadministered questionnaire. It was completed during the OPD visit. Each item of the GHQ was rated on a four-point Likert scale (0-1-2-3) instead of the conventional bi-modal scale (0-0-1-1) designed for the GHQ, to enable the statistical procedure in relation to factor analysis. Data was entered to EPI info and then converted to SPSS file and analysed. Factorability of the GHQ-30 was assessed with the Bartlett’s test of sphericity (<0.001) and Kaiser–Meyer–Olkin (0.95) measure of sampling adequacy. We applied Generalized Least Squares method with Oblimin rotation, because it produced five distinct interpretable factors. The internal consistency was assessed by calculating Cronbach’s a coefficient of the instrument for the overall sample as well as for selected socio-demographic variables and individual factors/subscales. Informed consent was obtained from all the research participants and confidentiality of data obtained were assured by concealing the identity using a coding system. Ethical clearance was granted by the Ethics Review Committee, Faculty of Medicine, University of Kelaniya.
3. Results Total number of OPD patients who were invited to the study during the period between May 2009 and January 2011 was 432. Of them 22 patients were not willing to participate which gave a nonparticipant rate of 5.0%. Therefore, the total patients recruited to the study was 410. Of them 42 (10%) had missing data for one or
more items of the GHQ-30, which left a study sample of 368 for the final analysis. Median age of the study population was 32.5 years (inter quartile range [IQR] = 21 years). The highest proportion (28.5%; n = 117) of them was in the age group of 21–30 years. The number of males among the study participants was 133 (35%). Fifty nine percent (n = 242) of the study population was Buddhists and 26.6% (n = 110) Catholics or Christians. Two hundred and twenty two (54%) had studied up to General Certificate Examination (Ordinary Level) [GCEOL]. 3.1. Factor structure of the GHQ-30 The median score (based on the Likert’s scale) of the GHQ-30 was 22 (IQR 15) for the total sample and 21 for both males (IQR 21) and females (IQR 13) when analyzed separately. For the age group <32 years, it was 20 (IQR 14) and 32 years, 23 (IQR 16). For the low (up to GCEOL) and high educational categories it was 23 (IQR 16.5) and 19 (IQR 13.5) respectively. Factor analysis produced a five factor solution (Table 1). The five factors together accounted for 51.6% of the total variance (TV). Considering the contribution of individual factors, Factor I (Depression) accounted for 38% of the TV, Factor II (Loss of confidence) 5%, Factor III (Insomnia) 4%, Factor IV (Social dysfunction) 2.7% and Factor V (Anxiety) 2%. Apart from item 21 (Found everything getting on top of you), there was a clear distinction between the items that loaded on the five factors. The item 21 loaded on both depression and anxiety (Table 2). 3.2. Internal consistency of GHQ-30 The Cronbach’s a coefficient of GHQ-30 for the overall sample was found to be 0.94 indicating satisfactory internal consistency of
C. Abeysena et al. / Asian Journal of Psychiatry 5 (2012) 150–153
152 Table 2 Inter-factor correlations.
Factor 1 Depression Factor 2 Loss of confidence Factor 3 Insomnia Factor 4 Social dysfunction Factor 5 Anxiety
Factor 1 Depression
Factor 2 Loss of confidence
Factor 3 Insomnia
Factor 4 Social dysfunction
Factor 5 Anxiety
1.00
0.45
0.29
0.48
0.56
0.45
1.00
0.23
0.53
0.38
0.29
0.23
1.00
0.32
0.39
0.48
0.53
0.32
1.00
0.39
0.56
0.38
0.39
0.39
1.00
Table 3 Item-total correlation coefficients of GHQ-30 for the whole sample and the subgroups.
Item Item Item Item Item Item Item Item Item Item Item Item Item Item Item Item Item Item Item Item Item Item Item Item Item Item Item Item Item Item
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30
Total sample
Male
Female
Age <32
Age 32
Low education
High education
0.53 0.50 0.49 0.49 0.39 0.54 0.54 0.57 0.30 0.53 0.37 0.45 0.53 0.69 0.61 0.65 0.60 0.63 0.58 0.59 0.67 0.76 0.74 0.68 0.70 0.66 0.69 0.71 0.63 0.73
0.56 0.57 0.53 0.56 0.44 0.58 0.59 0.68 0.42 0.56 0.46 0.53 0.58 0.78 0.60 0.75 0.61 0.71 0.64 0.64 0.73 0.78 0.79 0.68 0.67 0.74 0.77 0.73 0.68 0.80
0.53 0.52 0.48 0.50 0.35 0.52 0.49 0.49 0.23 0.54 0.32 0.38 0.48 0.64 0.62 0.62 0.60 0.62 0.56 0.57 0.64 0.76 0.73 0.71 0.74 0.61 0.66 0.69 0.63 0.69
0.43 0.52 0.50 0.44 0.36 0.51 0.41 0.56 0.24 0.51 0.37 0.49 0.55 0.65 0.58 0.63 0.58 0.58 0.61 0.58 0.66 0.73 0.72 0.70 0.65 0.66 0.69 0.67 0.58 0.69
0.62 0.48 0.48 0.54 0.45 0.58 0.69 0.59 0.37 0.55 0.38 0.41 0.50 0.71 0.64 0.66 0.63 0.67 0.54 0.60 0.67 0.79 0.76 0.68 0.74 0.66 0.69 0.73 0.67 0.76
0.46 0.47 0.50 0.51 0.44 0.53 0.49 0.51 0.35 0.51 0.37 0.46 0.50 0.69 0.58 0.64 0.58 0.62 0.56 0.62 0.67 0.75 0.75 0.69 0.73 0.64 0.66 0.70 0.71 0.74
0.64 0.53 0.47 0.49 0.38 0.55 0.59 0.69 0.19 0.56 0.35 0.41 0.58 0.68 0.64 0.65 0.63 0.64 0.62 0.57 0.67 0.77 0.74 0.69 0.65 0.68 0.71 0.69 0.52 0.70
the scale. It was 0.96 for males, 0.94 for females, 0.95 for the age group 32 years and 0.94 for age group <32 years as well as for both educational categories. The corrected item-total correlation coefficients were above 0.40 for the overall sample as well as for the sub-samples (Table 3). The Cronbach’s a coefficients by individual factors/subscales were 0.91 for ‘‘Depression’’, 0.79 for ‘‘Loss of confidence’’, 0.75 for ‘‘Insomnia’’, 0.74 for ‘‘Social dysfunction’’ and 0.90 for ‘‘Anxiety’’. 4. Discussion We found that GHQ-30 consisted of five factors which were named as Depression, Loss of confidence, Insomnia, Social dysfunction and Anxiety which together accounted for 51.6% of the total variance. The factor one was Depression alone accounted for 38% of the total explained variance. All factors were moderately inter-correlated. Item 21 of the GHQ-30 (Found everything getting on top of you?) being common to both, factors on ‘Depression’ and ‘Anxiety’, and demonstrates it’s inability to discriminate between the two constructs. Items 6 (Been managing as well as most people would in your shoes?) and 9 (Did you develop love and fondness to
your dear ones?) were not loaded to any of the factors extracted which suggests its non contribution. The factor structure of the GHQ-30 reported had shown variations from culture to culture. However, Huppert et al. (1989) reported a five factor structure namely Anxiety, Feelings of incompetence, Depression, Difficulty in coping and Social dysfunction which was more or less similar to that of our study. The subscales reported in other studies varied from three (Frydecka et al., 2010), four (Shek, and Tsang, 1995), six and seven (Iwata et al., 1994; Shigemi et al., 2000) factor structures. It is observed that Depression and Anxiety are common subscales to all the cultures detection of which is one of the primary objectives with the use of this instrument. Shek and Tsang (1995) from China, reported that it had a four factor structure namely Anxiety, Depression, Interpersonal dysfuctioning and Adequate coping. Two studies (Iwata et al., 1994; Shigemi et al., 2000) from Japan reported to have six and seven factors respectively. Of the two, in the study reported by Iwata et al. (1994) the six factors extracted were Anxiety, Depressive thoughts, Inhibition of thinking, Social dysfunction, Loss of positive attitudes and Hyperaction. The other study reported by Shigemi et al. (2000) described six factors namely Anxiety and depression, Severe
C. Abeysena et al. / Asian Journal of Psychiatry 5 (2012) 150–153
depression, Lack of interpersonal network or social support, Poor social activity, Insomnia and Loss of positive attitudes. The items related to Insomnia in the later were similar to that of our study. Most of the items related to ‘Social dysfunction’ and ‘Loss of confidence’ in the present study were categorized into factors on ‘Inhibition of thinking’ and ‘Social dysfunction’ in the study reported by Iwata et al. (1994) and ‘Lack of interpersonal network or social support’ and ‘Poor social activity’ in the study reported by Shigemi et al. (2000). Similar to our study, item 6 was not loaded on to any of the factors extracted by Shigemi et al. (2000). We could not perform factor analysis by sex or age groups because of the inadequacy of subgroup sample sizes. The internal consistency of the GHQ-30 in our total sample was 0.94, which was higher than in the previous studies published (Iwata et al., 1994; Shigemi et al., 2000). It is clearly demonstrated from the above that direct comparison between studies can be made only with regard to the number of factors derived, but not the structure within the derived factors, if all relevant details are not published. The nomenclature of the factors derived are arbitrarily chosen, based on the factor structure, which has the tendency to vary from culture to culture thus making it a futile exercise to describe the differences between the structure of individual factors. Therefore, for comparison purposes, what matters is the psychometric properties of the summated score of all the items included in the GHQ, which has been shown to have consistency across studies in relation to adequate validity and reliability. We recruited the study participants from the OPD patients without applying any probability sampling method, which can be interpreted as a limitation to the study as this affects the genralizability of the results to all OPD attendees of this tertiary care institution. Despite above the minimal non-response rate observed (5%; n = 22), it would be described as a strength, which is likely to mitigate the selection bias which would have otherwise exerted a serious impact on the external validity of the study. In conclusion, according to the results of this study, the GHQ-30 comprises five factors and it displayed adequate reliability for assessment of psychiatric disorders among Sinhala speaking primary care attendees in Sri Lanka. It also indicates that the factor structure of the Sinhala version is similar to that of the GHQ30 in other cultural settings. Therefore, the use of the Sinhala version of GHQ-30 may be recommended to screen common psychiatric disorders in primary care attendees.
153
Role of funding source Funding for this study was provided by the University of Kelaniya; the university had no further role in study design; in the collection, analysis and interpretation of data; in the writing of the report; and in the decision to submit the paper for publication. Contributors All authors designed the study and wrote the protocol. CA managed the literature searches and analyses and undertook the statistical analysis. CA wrote the first draft of the manuscript. All authors contributed to and have approved the final manuscript. Conflict of interest The authors declare that they have no conflicts of interest. Acknowledgements We are grateful to two research assistants who assisted for data collection. References Abramson, J.H., Abramson, Z.H., 1992. Survey Methods in Community Medicine. Churchil Livingston, London. Frydecka, D., Małyszczak, K., Chachaj, A., Kiejna, A., 2010. Factorial structure of the general health questionnaire (GHQ-30). Psychiatr. Pol. 44 (3), 341–359. Huppert, F.A., Walters, D.E., Day, N.E., Elliott, B.J., 1989. The factor structure of the General Health Questionnaire (GHQ-30). A reliability study on 6317 community residents. Br. J. Psychiatry 155, 178–185. Iwata, N., Uno, B., Suzuki, T., 1994. Psychometric properties of the 30-item version general Health Questionnaire in Japanese. Jpn. J. Psychiatry Neurol. 48 (3), 547–556. Jakob, K.S., Bhugra, D., Mann, H., 1997. The validation of the 12-item General Health Questionnaire among ethnic Indian women living in the United Kingdom. Psychol. Med. 27, 1215–1217. Ohta, Y., Kawasaki, N., Araki, K., Mine, M., Honda, S., 1995. The factor structure of the general health questionnaire (GHQ-30) in Japanese middle-aged and elderly residents. Int. J. Soc. Psychiatry 41 (4), 268–275. Shek, D.T., Tsang, S.K., 1995. Reliability and factor structure of the Chinese GHQ-30 for parents with preschool mentally handicapped children. J. Clin. Psychol. 51 (2), 227. Shigemi, J., Mino, Y., Tsuda, T., 2000. Stability of factor structure and correlation with perceived job stress in General Health Questionnaire: a three-wave survey over one year in Japanese workers. J. Occup. Health 42, 284–291.