Factor analysis of the Schizophrenia Quality of Life Scale Revision 4 (SQLS-R4) Chinese version and related factors

Abstract

Objectives. To validate the Schizophrenia Quality of Life Scale Revision 4 (SQLS-R4) Chinese version with factor analysis. We also investigated factors influencing quality of life results. Method. One hundred patients with schizophrenia participated in this study. Experienced psychiatrists evaluated each patient's current psychological condition to assure intact judgment. Participants then filled out the quality of life questionnaire, SQLS-R4. We used principal component analysis to analyze SQLS-R4 with oblique rotation. We also collected demographic data, another subjective and objective quality of life questionnaire, and information about the functional status of the participants at the same time. Multiple linear regression was used to analyze collected factors which may influence the quality of life of schizophrenic patients. Results. We extracted seven factors that explained a 68.1% variance in the patients' quality of life. The factors were relationships with others, loneliness, exhaustion, depressed thinking, somatic concerns, vitality, and worry. Conclusion. We found that the SQLS-R4 factor analysis provided insights into the structure of the quality of life of patients with schizophrenia. We also found that subjective social participation, subjective health, and depression were quality-of-life-related factors. Whether other factors influence the quality of life in patients with schizophrenia requires further investigation.

The Construct of the Schizophrenia Quality of Life Scale Revision 4 for the Population of Taiwan

“…However, as a newly developed QoL measure [ 8 , 11 ], the construct of SQLS-R4 still needs to be confirmed. Although the SQLS-R4 has good psychometric properties [ 6 , 10 , 13 ], to the best of our knowledge, only one UK study [ 7 ] has used confirmatory factor analysis (CFA) to examine the construct of SQLS-R4. The advantage of using CFA is being able to compare several proposed models and to clarify the construct of the tested instruments [ 14 ].…”

“…Therefore, the CFA is suitable for understanding the constructs of measurements that are under development, such as the SQLS-R4. Although the SQLS-R4 was suggested to be two-dimensional [ 6 , 7 , 12 , 13 ], Martin and Allan [ 7 ] indicated that some CFA fit indices are not acceptable (root mean square error of approximation [RMSEA] = 0.11 and weighted root mean square residual [WRMR] = 0.94). In addition, one Singaporean study [ 11 ] mentions that the cultural difference between Asia and the West may influence the psychometric properties of the SQLS-R4.…”

“…The former version of the SQLS-R4, SQLS, is a 30-item self-reported measure compromising three domains (psychosocial, motivation and energy, and symptoms and side effects) [ 8 ]. Several revisions were carried out for the SQLS to improve its psychometric properties, and finally the SQLS-R4 with two factors was developed with different language versions [ 7 , 9 ], including Taiwanese version [ 6 , 13 ]. However, when Kuo et al [ 6 ] examined the psychometric properties of the Taiwan version of SQLS-R4, they found that four items had weak corrected item-total correlation with their domains.…”

Minimum Sample Size Requirements for a Validation Study of The Schizophrenia Quality of Life Scale-Revision 4 (SQLS-R4)

“…We would advocate, that whenever possible, the complete data set should be used within the analysis, with the smallest number of missing data points (<5%). It must be acknowledged however, that the 'rule of thumb' sample size established by Martin and Allan (6) and utilised by other studies (3,8,25) used an estimated sample size that was much too small, thus lacking in power. This underlines the importance of conducting simulations such as the Monte Carlo simulation to determine the sample size required to offer a good fit to data for an adequately powered study.…”

“…To date most translation/adaptation/validation studies have used just 100 patients, presumably based on the 'rule of thumb' sample size established by Martin and Allan (6) see for example Su, Yang (3), Kuo, Chen-Sea (8) Kuo, Ma (25). However, selection of such a modest sample size may be problematic particularly in terms of the relatively large number of items (N=33) within the measure which would violate other minimum sample size criteria such as number of participants to number of items ratio (26).…”