|Year : 2013 | Volume
| Issue : 2 | Page : 109-113
Internal consistency and factor structure of 12-item general health questionnaire in visually impaired students
Ajay Kumar Bakhla1, Vijay Verma2, Mahesh Hembram2, Samir Kumar Praharaj3, Vinod Kumar Sinha4
1 Department of Psychiatry, Rajendra Institute of Medical Sciences, Ranchi, Jharkhand, India
2 Department of Psychiatry, Ranchi Institute of Neuropsychiatry and Allied Sciences, Kanke, Ranchi, Jharkhand, India
3 Department of Psychiatry, Kasturba Medical College, Manipal, Karnataka, India
4 Professor of Psychiatry, Central Institute of Psychiatry, Ranchi, Jharkhand, India
|Date of Web Publication||21-May-2014|
Ajay Kumar Bakhla
Assistant Professor of Psychiatry, Rajendra Institute of Medical Sciences, Ranchi - 834 009, Jharkhand
Source of Support: None, Conflict of Interest: None
| Abstract|| |
Background: As there are no instruments to measure psychological wellness or distress in visually impaired students, we studied internal consistency and factor structure of GHQ-12 in visually impaired children. Materials and Methods: Internal consistency analysis (Cronbach's alpha and item total correlation) and exploratory factor analysis (principal component analysis) were carried out to identify factor structure of 12-item general health questionnaire (GHQ-12). Results: All items of GHQ-12 were significantly associated with each other and the Cronbach's alpha coefficient for the scale was 0.7. On analysis of principal component, three-factor solution was found that accounted for 47.92% of the total variance. The factors included, 'general well-being', 'depression' and 'cognitive', with Cronbach's alpha coefficients being 0.70, 0.59, and 0.34, respectively. Conclusion: Our study findings suggest GHQ-12 is a reliable with adequate internal consistency scale and multidimensional factor structure in visually impaired students.
Keywords: Factor analysis, GHQ-12, principal component analysis, visual impairment
|How to cite this article:|
Bakhla AK, Verma V, Hembram M, Praharaj SK, Sinha VK. Internal consistency and factor structure of 12-item general health questionnaire in visually impaired students. Ind Psychiatry J 2013;22:109-13
|How to cite this URL:|
Bakhla AK, Verma V, Hembram M, Praharaj SK, Sinha VK. Internal consistency and factor structure of 12-item general health questionnaire in visually impaired students. Ind Psychiatry J [serial online] 2013 [cited 2022 Jul 4];22:109-13. Available from: https://www.industrialpsychiatry.org/text.asp?2013/22/2/109/132918
There is lack of any extensively researched and well-validated instrument for the identification and measurement of psychological problems in visually impaired, as it is available for normally sighted people.
There are certain instruments to measure "vision-specific QoL".  Quality of life (QoL) tools ascertain social, emotional, and participation aspects of daily living, whereas general psychological health and well-being may be influenced by specific disability but not a synonymous for each other. For a epidemiological study,  reported elsewhere, we used versions of general health questionnaire (GHQ) in visually impaired students, which are one of the most validated instruments being used in general population for this purpose.
Several factor analytic studies have been conducted on GHQ-12 to identify the underlying dimensions.  Exploratory factor analysis using either principal component analysis or factor analysis has identified two or three factor solution. In these studies, the variance explained by the factors ranged from 46%-64%. Confirmatory factor analyses of the GHQ-12 have been carried out which further supports a three-factor model. , Most of these studies were conducted on general population. There is a lack of literature on quality of life/general health of visually impaired children and young people, and this lack is more attributable to lack of measures, instruments or scales for this population. We attempted to see if a validated general population scale may also express similar underlying dimensions or not in visually impaired population.
| Materials and methods|| |
This was a cross-sectional school-based study conducted at Central Institute of Psychiatry, Ranchi, India. The study was approved by the Institutional Ethical Committee and the Central Institute of Psychiatry and a written permission and informed consent were obtained from the school authorities, participants, and their guardians. All the four residential schools for visually impaired children in Ranchi, capital of Jharkhand state of India were surveyed. Study sample consisted of all students of both sexes and having a vision less than 3/60. Those with co-morbid deafness, dumbness or other physical disability were excluded. A total of 110 students was assessed initially, of which 92 (59 males and 33 females) fulfilled the inclusion criteria.
Socio-demographic and clinical data were collected using a specially designed pro-forma. The Hindi version of 12-item GHQ was administered on all the participants; there are many scoring methods, the two popular methods are bimodal (0-0-1-1) and Likert scoring (0-1-2-3). Likert scoring is known to produce a less skewed and less kurtosis of scores, therefore, each item was rated on a 5-point Likert scale. This tool is reliable and has been validated in Indian population. 
After explaining the nature of study, each student was examined using semi-structured interview for gathering socio-demographic and clinical details. Thereafter, GHQ-12 was administered on all the participants. As they were unable to read printed paper all the questions were asked verbally along with options and answer sheet was marked accordingly to their reply.
Statistical analysis was done on 92 subjects using Statistical Package for Social Sciences (SPSS, Inc., Chicago, Illinois) version 10.0 for Windows® . Frequency analysis for Socio-demographic characteristics and item analysis was done to know the internal consistency of GHQ-12 by calculating Cronbach's alpha coefficient; exploratory factor analysis (principal component analysis) was carried out to identify factor structure of GHQ-12. To retain the number of factors, both Kaiser's criteria of eigenvalues greater than unity as well as scree plot inspection was done. Promax rotation was carried out along with Kaiser's standardization and a cutoff of 0.5 in factor loading was considered significant.
| Results|| |
Socio-demographic characteristics of the sample are summarized in [Table 1]. There were 59 males and 33 females. The mean age was 12.71 (SD 3.53) years. Among them, 63 were congenitally blind, whereas 29 had acquired impairment. The mean duration of impairment was 10.34 (SD 3.73) years and mean age of onset of acquired impairment was 3.34 (SD 2.16) years.
The GHQ-12 scores are summarized in [Table 2]. The mean GHQ score was 8.5 (SD 3.3). The means of the positively worded items (i.e. items 1, 3, 4, 7, 8, and 12) was 4.84 ± 1.89 ranged from 0.67 to 0.93, (Cronbach's alpha = 0.61); whereas mean for negatively worded items (i.e. items 2, 5, 6, 9, 10, and 11) was 3.65 ± 2.12 ranged from 0.42 to 0.80 (Cronbach's alpha = 0.62), which suggest that most of the respondents considered themselves healthy.
On analysis of Pearson Item-total correlations [Table 2] all items were significantly associated with each other. Scale mean if item deleted was measured for all 12 items, which ranged 7.57 to 8.08 and Cronbach's alpha if item deleted for all twelve items ranged from. 646 to 712 (Cronbach's coefficients alpha for complete scale was 0.70).
|Table 2: Score on GHQ‑12, Pearson item‑total correlations (rtt), Cronbach's alpha|
Click here to view
Kaiser-Meyer-Olkin measure of sampling adequacy was 0.629, which indicates adequate sample size for the factor analysis. Bartlett's test of sphericity was significant (χ2 = 216.05, df = 66, P < 0.001). Principal components analysis showed four components with an eigenvalue greater than 1 which accounted for 59.09% of total variance. The first two components had at least three items, whereas third and fourth component had two items each with loading more than 0.5. Scree plot revealed point of inflexion after third component [Figure 1]. Therefore, a three-factor solution was considered appropriate that accounted for 47.92% of the total variance. Following promax rotation, the pattern matrix was found to produce clinically satisfactory solution with high loadings in the obtained factors [Table 3].
|Table 3: Factor structure of GHQ‑12 items (principal component analysis with promax rotation and Kaiser normalization) showing factor loadings >0.5|
Click here to view
The first factor had eigenvalue of 3.01 and explained 25.11% of variance. The item, 'could not overcome difficulties', had the highest loading (0.856). This factor included three positively worded items, 'feeling reasonably happy', 'enjoy normal activities' and 'playing useful part', and two negatively worded items, 'could not overcome difficulties' and 'under stress'. The factor seems to represent and hence named as 'general well-being' factor. The second factor had eigenvalue of 1.70 and explained 14.13% of variance. It had high loading from items related to depression, i.e. 'feeling unhappy and depressed', 'losing confidence' and 'thinking of self as worthless'. The item, 'feeling unhappy and depressed' had the highest loading (0.750) which emerged as central tendency of the factor. This factor seemed to represent ' depression'. The third factor had eigenvalue of 1.28 and explained 10.63% of variance. It had high loadings from two items, 'able to concentrate' (0.738) and 'face up to problems' (0.735), and seemed to represent 'cognitive' symptoms. The first and third factor apparently has unimodal distribution, whereas a bimodal distribution was noted in depression, the second factor. A internal consistency analysis of all items of GHQ-12 showed Cronbach's alpha coefficient  to be 0.70, Cronbach's alpha for the first factor was also 0.70, whereas for second and third factor it was 0.59 and 0.34, respectively.
| Discussion|| |
The GHQ-12 is used for measuring minor psychological distress, which is well validated and widely used in diverse settings, but has not been studied in visually impaired subjects. Our study has addressed the factor structure and internal consistency of GHQ-12 among visually impaired students. The sample size of 92 is slightly small for the factor analytic studies, but the total students studying in all existing school for visually impaired in the city was 110; out of which, 18 were excluded for comorbid disabilities and problems of consent.
Analysis in term of Cronbach's alpha and item total correlation suggested that the items of the GHQ-12 were reliable as having adequate internal consistency. Found Cronbach's alpha value of 0.70 is not excellent but considered as the minimum acceptable criterion of instrument's internal consistency as per Kline's criterion.  However, Cronbach's coefficients for the GHQ-12 in other studies have been found to be in much higher range of 0.88 to 0.93, but it has been seen to vary with different scoring methods.  The use of Likert-type scoring of GHQ-12 in our study has advantages over other methods, which have been criticized for under-identification of respondents with existing psychological problems. 
Findings from internal consistency analysis in terms of Pearson Item-Total Correlations suggested that all the 12 items have correlation value more than 0.33 (ranging 0.33 to 0.68) and all were significantly correlated statistically, as shown in [Table 2]. Therefore, the finding suggests that the all the items in the GHQ-12 were correlated among themselves and it is a reliable instrument that could be used in the future to detect distressed visually impaired students.
Results from principal component analysis showed three-factor solutions for GHQ-12 in this stated population that accounted for 47.92% of the variance. Our findings were similar to that of previous factor analytic studies of GHQ-12 ,,, that have found three-factor solution. In accordance, a study by Martin and Newell  in their confirmatory factor analysis of three-factor solution of Graetz  using Likert scoring similar to our study, revealed three factors: Anxiety, social dysfunction, and loss of confidence.
Our first factor appears as a 'general well-being' factor along with two items, 'could not overcome difficulties' and 'under stress', both of which appear in the first factor of Worsley and Gribbin,  Graetz,  Politi et al.,  and Daradkeh et al.  The internal consistency of this factor was found to be 0.70, which is considered as adequate as per Kline criteria.  In our study, the second factor emerged as 'depression factor,' which has three items that are related to depression. These three items loaded in the first factor of Politi et al., and Daradkeh et al., among other items. Thus, depression appears to be a distinct factor in these populations, which is in contrast to few studies in which the depression items loaded on different factors. , Furthermore, depression factor had bimodal distribution in our sample which suggests presence of depressive symptoms in a subset of these students whereas the other group was relatively free of these symptoms. The third factor represents 'cognitive' dimension, includes two items 'able to concentrate' and 'face up to problems'. In other studies, these items loaded on the second factor along with other items, ,, may be explained as depression and cognition usually represents each other.
The variation in factor composition of GHQ-12 in visually impaired students has a number of reasons and implications. Firstly, the source of psychological distress among visually impaired may have different construct then sighted population, as it has been reported to vary across cultures and populations.  Secondly, it has been found that variation of factors may exists between different places and across different time periods, but validity of the scale remains intact.  Thirdly, several other factors such as translation of scale, verbal application of scale on child and adolescent population may affect the study findings.
Our study implicates that among visually impaired students the general psychological well-being as measured by GHQ-12 is not unitary construct, but either presence or absence of depression and cognitive construct has bearing on overall psychological well-being. It is interesting to note that the mean GHQ score for the positively worded items were higher then the negatively worded items, suggesting that most of the participants considered themselves healthy. Although, psychological distress is expected with physical disability such as visual impairment, but the study population might not represent all visually impaired children and adolescents as they belonged to a privileged minority in India, who are getting education in special schools. 
Though, we need better and specific measures to detect and assess quality of life/general health of visually impaired children and young people further studies in these populations are required to re-evaluate the reliability and validity of GHQ-12 and future development of specific instruments is required. This lack of measuring instruments or scales and literature on quality of life/general health of visually impaired children and young people prompted us to attempt this study, and we found that in the cohort of visually impaired students the underlying factor structure of the GHQ-12 is multidimensional, though the underlying structure of dimensions are slightly differing. The validity and internal consistency of the GHQ-12 is not as par with the previous studies on general population, but may be used for studies with its limitations in visually impaired population.
| References|| |
|1.||Cochrane GM, Marella M, Keeffe JE, Lamoureux EL. The Impact of Vision Impairment for Children (IVI_C): Validation of a vision-specific pediatric quality-of-life questionnaire using Rasch analysis. Invest Ophthalmol Vis Sci 2011;52:1632-40. |
|2.||Bakhla AK, Sinha VK, Verma V, Sarkhel S. Prevalence of psychiatric morbidity in visually impaired children. Indian Pediatr 2011;48:225-7. |
|3.||Cambell A, Walker J, Farrell G. Confirmatory factor analysis of the GHQ-12: Can I see that again? Aust N Z J Psychiatry 2003;37:475-83. |
|4.||Martin AJ. Assessing the multidimensionality of the 12-item General Health Questionnaire. Psychol Rep 1999;84:927-35. |
|5.||Martin CR, Newell RJ. The factor structure of the 12-item General Health Questionnaire in individuals with facial disfigurement. J Psychosom Res 2005;59:193-9. |
|6.||Gautam S, Nijhawan M, Kamal P. Standardization of hindi version of goldberg's general health questionnaire. Indian J Psychiatry 1987;29:63-6. |
|7.||Cronbach LJ. Coefficient alpha and the internal structure of tests. Psychometrika 1951;16:297-334. |
|8.||Kline P. A Psychometrics Primer. London: Free Association Books; 2000. |
|9.||Newman SC, Bland RC, Orn H. A comparison of methods of scoring the General Health Questionnaire. Compr Psychiatry 1988;29:402-8. |
|10.||Worsley A, Gribbin CC. A factor analytic study of the twelve item general health questionnaire. Aust N Z J Psychiatry 1977;11:269-72. |
|11.||Farrell GA. The mental health of hospital nurses in Tasmania as measured by the 12-item General health Questionnaire. J Adv Nurs 1998;28:707-12. |
|12.||Daradkeh TK, Ghubash R, el-Rufaie OE. Reliability, validity and factor structure of the Arabic version of the 12-item General Health Questionnaire. Psychol Rep 2001;89:85-94. |
|13.||Graetz B. Multidimensional properties of the General Health Questionnaire. Soc Psychiatry Psychiatr Epidemiol 1991;26:132-8. |
|14.||Politi PL, Piccinelli M, Wilkinson G. Reliability, validity and factor structure of the 12-item General Health Questionnaire among young males in Italy. Acta Psychiatr Scand 1994;90:432-7. |
|15.||Werneke U, Goldberg DP, Yalcin I, Ustün BT. The stability of the factor structure of the General Health Questionnaire. Psychol Med 2000;30:823-9. |
|16.||Kundu CL. Status of disability in India - 2000. New Delhi: Rehabilitation Council of India, Ministry of Welfare, Government of India; 2000. |
[Table 1], [Table 2], [Table 3]
|This article has been cited by|
||Memantine augmentation of sertraline in the treatment of symptoms and executive function among patients with obsessive-compulsive disorder: A double-blind placebo-controlled, randomized clinical trial
| ||Sanaz Askari, Saba Mokhtari, Seyed Vahid Shariat, Behnam Shariati, Masoomeh Yarahmadi, Mohammadreza Shalbafan |
| ||BMC Psychiatry. 2022; 22(1) |
|[Pubmed] | [DOI]|
||Understanding the relationship between parenting style and chronic pain in adolescents: a structural equation modelling approach
| ||Maryam Shaygan, Pardis Bostanian, Mina Zarmehr, Hamidreza Hassanipour, Maryam Mollaie |
| ||BMC Psychology. 2021; 9(1) |
|[Pubmed] | [DOI]|
||Cardiac autonomic tone, plasma BDNF levels and paroxetine response in newly diagnosed patients of generalised anxiety disorder
| ||Ravi Dutt, Nilima Shankar, Shruti Srivastava, Asha Yadav, Rafat S Ahmed |
| ||International Journal of Psychiatry in Clinical Practice. 2020; 24(2): 135 |
|[Pubmed] | [DOI]|
||The Prediction of Metro Shield Construction Cost Based on a Backpropagation Neural Network Improved by Quantum Particle Swarm Optimization
| ||Lanjun Liu, Denghui Liu, Han Wu, Xinyu Wang, Hugo Rodrigues |
| ||Advances in Civil Engineering. 2020; 2020: 1 |
|[Pubmed] | [DOI]|
||Different psychometric properties of the Emotional Reaction Instrument-English (ERI-E) between hospitalized African American and Caucasian children
| ||Kye-ha Kim,Roxie L. Foster,Jeong-hwan Park |
| ||Journal for Specialists in Pediatric Nursing. 2017; : e12176 |
|[Pubmed] | [DOI]|