Table of Contents Author Guidelines Submit a Manuscript
Journal of Diabetes Research
Volume 2018, Article ID 5804687, 10 pages
https://doi.org/10.1155/2018/5804687
Research Article

A Revised Version of Diabetes Quality of Life Instrument Maintaining Domains for Satisfaction, Impact, and Worry

1Clinical Research Centre, Sarawak General Hospital, Ministry of Health Malaysia, Kuching, Malaysia
2National Clinical Research Centre, Ministry of Health Malaysia, Kuala Lumpur, Malaysia
3Health Clinic Seremban 2, Ministry of Health Malaysia, Seremban Negeri Sembilan, Malaysia

Correspondence should be addressed to Mohamad Adam Bujang; ym.vog.crc@mada

Received 10 April 2018; Accepted 12 June 2018; Published 22 July 2018

Academic Editor: Eusebio Chiefari

Copyright © 2018 Mohamad Adam Bujang et al. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

Abstract

Background. Diabetes quality of life (DQoL) instrument has been widely used to measure quality of life among diabetes patients. This study aimed to develop a revised version of DQoL instrument that incorporated issues of redundancies in the items and strengthen the basis of validity of the instrument. Methods. This was a cross-sectional study where diabetes patients were recruited from December 1, 2014, until end of March 2015 at a public health clinic in Peninsular Malaysia. A questionnaire that included patients’ information and DQoL instrument was distributed to patients. Item selection of DQoL instrument was conducted to screen and finalize the items based on issues of missing values and redundancy. Validity testing was conducted for the revised DQoL instrument based on exploratory factor analysis, confirmatory factor analysis, and Rasch analysis. Results. The pattern structure matrix yielded three domains similar to the original version with 18 items. The minimum factor loading from the structure matrix was 0.358. The item’s and person’s reliability was excellent with 0.92 and 0.84 for “satisfaction” domain, 0.98 and 0.60 for “impact” domain, and 0.99 and 0.57 for “worry” domain, respectively. Confirmatory factor analysis has dropped 5 items and the revised version of DQoL contained 13 items. Composite reliability of the revised version was computed for “satisfaction” domain (0.922; 95% CI: 0.909–0.936), “impact” domain (0.781; 95% CI: 0.745–0.818), and “worry” domain (0.794; 95% CI: 0.755–0.832). Conclusion. A revised version of DQoL that maintains the conceptualization of “satisfaction,” “impact,” and “worry” with 13 items was successfully developed.

1. Introduction

Diabetes quality of life (DQoL) instrument was published in 1988 by the Diabetes Control and Complications Trial (DCCT) Research Group [1]. It was initially developed for a multicenter controlled trial to investigate the effect of two different diabetes treatment interventions on the incidence and progression of early vascular complications. The DQoL instrument which contained 46 items was used to measure health-related quality of life among diabetes patients based on three main domains, namely, “satisfaction,” “impact,” and “worry.” This instrument has been widely used in diabetes research for decades.

As a questionnaire to measure quality of life for patients with diabetes mellitus, DQoL was reported to have very strong reliability and proven to be valid [1, 2]. Its reliability measures were evaluated based on test-retest reliability and internal consistency. Meanwhile, the basis of its validity was supported based on content and concurrent validity. Content validity was conducted among a group of experts while the concurrent validity was determined by the support of other questionnaires such as Symptom Checklist-90-R (SCL), the Bradburn Affect Balance Scale (ABS), and the Psychosocial Adjustment to Illness Scale (PAIS). However, the development of DQoL instrument was not supported by construct validity, although statistical technique such as exploratory factor analysis is a common type of analysis which can be applied in questionnaire development to construct domains for a latent variable [1].

The instrument has been translated and validated into various languages such as Spanish, Turkish, Japanese, Chinese, Taiwanese, Iranian, and Malaysian [39]. Some studies showed improved evidence on the validity of DQoL instrument by testing the association of DQoL instrument and diabetes complications [7]. Proven its stability and validity, DQoL instrument is widely used for diabetes research despite its limitations and is still relevant until now.

Initially, the foundation of DQoL instrument is supported by theory of the conceptual scheme of domains and variables in a quality of life assessment [10]. Perhaps one of the limitations for DQoL instrument is that many items are required to represent the three main domains; nonetheless, it is necessary to ensure that various perspectives of quality of life among diabetes patients are well covered. Other well-known quality of life instruments also have a relatively large number of items with at least 30 items [1113].

An important consideration among diabetes patients is that there are substantial number of elderly patients with varying severity of diabetes-related complications. A questionnaire with substantial items will require more time to be completed. Invalid responses may occur due to quick responses without proper thinking and evaluation on every item by the respondents. A high number of missing values are likely to happen, consequently leading to frustration to the researcher as items were not responded properly. Hence, developing a shorter version of DQoL instrument is crucial.

Some efforts have been made to develop a shorter version of DQoL instrument, and it was considered as a brief version of DQoL instrument. However, the function of the three main domains was not covered [14]. Therefore, the intention of our study is to develop a newly revised instrument for DQoL while maintaining the assessment of the three main domains, namely, “satisfaction,” “impact,” and “worry.” The initial concept of the domains is in line with the requirement of a quality of life measure [10]. Besides the content, the three main domains need to be validated statistically. Therefore, the aims of this study were to introduce the revised version of DQoL instrument and to support the validity with quantitative measures using exploratory factor analysis, confirmatory factor analysis, and Rasch analysis.

2. Methods

The methods were divided into three sections, namely, study design, item selection, and proof of the validity basis based on statistical measures.

2.1. Study Design

This was a cross-sectional study where type 2 diabetes patients were recruited from December 1, 2014, until end of March 2015 at a government health clinic (Health Clinic Seremban 2) in Malaysia. The validated revised DQoL instruments [9] were distributed to diabetes patients attending their follow-up visit. Besides DQoL instrument, other information such as demographic profiles and clinical variables were also collected and analyzed. Subjects’ clinical information were obtained from their health record. Written consent was obtained from all patients before their participation in the study. Ethical approval was received on July 18, 2014, from the Medical Research and Ethics Committee (MREC) of Ministry of Health Malaysia (NMRR ID: NMRR-14-522-19377).

2.2. Item Selection

The original DQoL instrument has 46 items with 15, 20, and 11 items for each domain of “satisfaction,” “impact,” and “worry.” In order to confirm the construct of the three domains, the revised version of DQoL instrument needs to be tested with construct validity such as exploratory factor analysis and confirmatory factor analysis. This was to confirm the ability of the revised version of DQoL instrument in measuring the construct intended as well as the content validity. Besides that, the construct validity also aimed to prove that the shorter version has an appropriate set of relevant items reflecting the full content of the domains being measured.

The development of the revised version started with selection of items. Items were assessed based on two criteria: missing value of the response less than 10.0% and no duplication in terms of the content. Evaluation of missing values is important to identify and omit irrelevant items. Issue of duplication should also be taken into consideration as duplicate items may affect the construct of the domains, and cross-loading between items may occur. Only one item was retained once duplication was identified, and priority was given based on the suitability of the item reflecting the definition of the domain.

2.3. The Proof of Validity Basis for a Newly Modified Version of DQoL Instrument

Exploratory factor analysis was performed to determine the underlying relationship between the constructs of DQoL domains. The scale of reverse scoring items (I8 and I16) was transformed before analysis was conducted [15]. Moderate correlation was expected between the domains, and hence, extraction method “principle axis factors” and rotation method “promax” were chosen. The approach of analysis was as the following: (1)Exploratory factor analysis was first conducted on individual domain to make sure all the items were stable in their respective domain. If the number of domain produced was more than one, only the domain with largest number of items would be chosen.(2)Subsequently, exploratory factor analysis was tested for all the items (no further process took place because the construct of the domains was successful. All the items fell under their respective domains).

Person’s and item’s reliability was analyzed using Rasch analysis while confirmatory factor analysis was conducted to examine how well the measured variables represent the number of constructs. Rasch analysis, introduced by George Rasch (1960), is a powerful tool to develop and validate an instrument [16]. Classically, internal consistency index like Cronbach’s alpha can be used to assess the performance of a scale. However, the measurement can be misleading, as high correlation among a subset of items may be due to the similarity of wordings rather than the relationship of items with the construct in question [17]. Hence, high alpha values might therefore be indicative of an inferior rather than superior quality of the scale [18]. The Rasch analysis has been used to construct new questionnaires and was extensively used to review and improve existing questionnaires that were constructed using Likert scales. Hence, this study assessed and simplified the DQoL questionnaire using Rasch analysis to optimize item that fits to the construct, and the psychometric properties were measured with fit statistics and separation indices. Reliability and empirical validity of the simplified/revised version were analyzed further and compared with those of the full/original version.

Another fundamental element of measurement: unidimensionality was used to examine whether the items formed one common underlying latent variable. Unidimensionality was assessed using the fit statistics and principal component analysis (PCA) of the residuals. Both infit and outfit statistics, measured as mean square standardized residuals (MNSQ), indicated how well items fit into the underlying construct. The cut-off range for MNSQ from 0.5 to 1.5 could be considered productive for measurement [19]. Unidimensionality was further assessed by principal component analysis (PCA) of the residuals, using two criteria. Firstly, the variance explained by the measurement dimension is large, at least 40% [20], and the empirical calculation should be similar to the model [21]. Secondly, the first contrasts in the residuals reporting the unexplained variance by the principal component should be small, that is, not more than 15% [22]. Therefore, to strengthen the evidence of the construct, Rasch analysis was applied to evaluate the new construct and compare it with the original construct.

The model was further tested using confirmatory factor analysis and model fit depending on absolute fit using standardized root mean square residual (SRMR), parsimony correction fit index using root mean square of approximation (RMSEA), comparative fit indices using comparative fit index (CFI), and Tucker-Lewis index (TLI). If the model did not fit the data, necessary modifications to improve model fit were done by removing items with low factor loading, high standardized residuals, and high modification index. Modifications were done until the model was reasonably fit as well as theoretically sound. Convergent validity (reliability) of the measurement model was checked using composite reliability by Raykov’s procedure. This method would yield more accurate reliability estimation where correlation term on error was added.

All analyses were conducted using SPSS (IBM Corp. Released 2012. IBM SPSS Statistics for Windows, Version 21.0. Armonk, NY: IBM Corp.), Mplus (Muthén, L. K., & Muthén, B. O., 1998–2011), and Winstep (Linacre, J. M. (2016). Winsteps® Rasch measurement computer program. Beaverton, Oregon: Winsteps.com).

3. Results

3.1. Demographic and Clinical Profile of Patients

A total of 536 patients have participated in the study. Majority were females (53.7%), of Malay race (55.8%) and married (84.3%). The mean (SD) age was 56.7 (11.2) years. About 68.1% of the participants have hypertension and 45.7% of them have dyslipidemia. Some of the participants have experienced diabetes complications such as retinopathy (5.2%), nephropathy (3.4%), neuropathy (1.1%), and cardiovascular (4.7%).

3.2. Item Selection

For “satisfaction” domain, eight items were dropped out of 15 items. Two items were dropped due to missing values more than 10.0%. For example, satisfaction item 10 (S10), concerning satisfaction with sexual life, it was removed as sexual issues is still considered sensitive in Malaysia and many people are uncomfortable to answer such question thus could lead to missing responses. Six items were dropped due to issue of redundancy with items under “impact” domain such as items concerning food intake (S5 and I9) and burden to family members (S6 and I5). Hence, the items retained under “satisfaction” domain were S1, S2, S3, S4, S7, S12, and S15 (Table 1).

Table 1: The evaluation of the original DQoL items based on content validity, missing values and issue of redundancy.

For “impact” domain, seven items were dropped out of 20 items. Items I10, I11, and I20 were dropped due to high percentage of missing values (>10.0%). Item I3 (has low blood sugar) was omitted. Item I1 was suitable to represent item I3 if the patients felt sick or had pain due to diabetes treatment. I14 was similar to I16 and both were dropped as they had less impact in terms of physical and psychological. Items I9 and I19 were comparable, and thus, I9 was chosen since it was simpler and more straightforward. Finally, items I1, I2, I4, I5, I6, I7, I8, I9, I12, I13, I15, I17, and I18 were selected for further evaluation (Table 1).

For “worry” domain, eight items were dropped out of 11 items. All the eight items were dropped due to missing values more than 10.0%. Hence, only W8, W9, and W10 were remained for further evaluation (Table 1).

3.3. Validity of a Newly Revised Version of DQoL Instrument

Results of exploratory factor analysis by each domain showed that only “impact” domain yielded two factors, whereas the rest only yielded one factor solution. Item 8 in “impact” domain was dropped since its communality value was less than 0.2 and produced factor loading with a negative value. Exploratory analysis was conducted again for 12 items (I1, I2, I4, I5, I6, I7, I9, I12, I13, I15, I17, and I18). Two factors were constructed where factor one had eight items and factor two had four items. Thus, the factor with more items (I1, I2, I4, I5, I6, I7, I9, and I18) was selected for further evaluation.

Exploratory factor analysis was further conducted for the 18 items where seven items were under “satisfaction” domain, eight items under “impact” domain, and three items under “worry” domain. Results showed that the construct produced three domains, and all the items fell under the respective domains. The minimum coefficient of factor loading from the structure matrix was 0.358 (Table 2). The item’s and person’s reliability was excellent with 0.92 and 0.84, respectively, for “satisfaction” domain, 0.98 and 0.60, respectively, for “impact” domain, and 0.99 and 0.57, respectively, for “worry” domain (Table 3).

Table 2: Evaluation of the revised version of DQoL: results from exploratory factor analysis and internal consistency.
Table 3: Summary of statistics for item and person parameters and unidimensionality using the principal component analysis (PCA) of the DQoL by Rasch analysis.
3.4. Evaluation Based on Rasch Analysis

We subjected the domains and five-point Likert scale to Rasch analysis in a similar manner to that of the full version of the DQoL instrument. The overall performance of both versions for “satisfaction” domain was acceptable with satisfactory fit of high internal consistency to the Rasch model (full version: item separation = 6.01 and reliability = 0.97 and person separation = 2.91 and reliability = 0.89, Cronbach’s α = 0.95; simplified version: item separation = 3.44 and reliability = 0.92 and person separation = 2.27 and reliability 0.84, Cronbach’s α = 0.93) (Table 3). Similar to the full version, the person separation and reliabilities were poor for both the domains: “impact” and “worry.” The person separation and reliability were low (<2.0 and <0.8, resp.), with a relevant person sample, indicating the poor discriminatory ability of the simplified/revised DQoL instrument. This implies that the domains may be less sensitive in distinguishing between those perceiving high and low degree of frequency, and that more items may be needed, although a longer questionnaire would be less desirable.

Table 4 shows that all items fit a single overall construct for the simplified/revised version (infit range 0.78–1.27; outfit range 0.70–1.25). In addition to item fit statistics, unidimensionality was assessed further using principal component analysis (PCA) of the residuals. In the simplified/revised DQoL instrument, the variance explained by the measures for the empirical calculation was almost identical to the model (“satisfaction”: 63.3% and 63.5%, respectively; “impact”: 51.9% and 52.4%; “worry”: 62.6% and 62.4%). The first residual factor in the “satisfaction” domain had an eigenvalue of 1.6, representing 9.8% of the unexplained variance, “impact” domain with 1.6 eigenvalue units (19.0%), and “worry” domain with 1.7 eigenvalue units (20.6%). These results suggested unidimensionality of the scale (Table 3).

Table 4: Item fit statistics of the simplified/revised DQoL from Rasch analysis.
3.5. Evaluation Based on Confirmatory Factor Analysis

Confirmatory factor analysis was performed, and only items with factor loading 0.6 or higher were recruited to ensure all the items recruited explained the domains sufficiently. Item S12 (0.523), I2 (0.526), I6 (0.584), I9 (0.561), and I18 (0.466) were removed. Composite reliability for each domain was computed as well; “satisfaction” domain showed highest composite reliability of 0.922, followed by “worry” domain (0.794) and “impact” domain (0.781). The measurement model for DQoL questionnaire was finalized with 13 items, with six items for “satisfaction” domain, four items for “impact” domain, and three items for “worry” domain, respectively (Table 5).

Table 5: Evaluation of the revised version of DQoL; result from confirmatory factor analysis and internal consistency.

Analysis was further conducted to examine model fit for both original (46 items) and the revised version (13 items). Both measurement models were compared based on the indices. The original version with 46 items did not fit the cut-off value for all the indices while the revised version with 13 items showed good fit for the measurement model. The value CF fit for the revised version with 13 items was 0.528, indicating no difference between observed data and the specified model (Table 6). The flow of data analysis is summarized in Figure 1.

Table 6: Comparison of fit indices for measurement model with 46 items and 15 items.
Figure 1: Flow of data analysis.

4. Discussion

The revision of the original version of DQoL instrument is crucial especially when there are claims that the instrument is less sensitive to measure quality of life among diabetes patients [14, 23]. The claim is due to the findings from the DCCT group when they first used the instrument and concluded that the DQoL instrument failed to detect any significant differences regarding the association of quality of life towards various clinical outcomes such as status of glycemic control, status of severe hypoglycemia, and status of progression of late complications [2]. This might be due to the fact that quality of life is also influenced by various factors such as marital status, social relationships, the existence of other health problems, patient knowledge, treatment satisfaction, and perceived ability to control one’s disease [2427].

Apart from the external factors mentioned earlier, the original DQoL instrument has a major limitation where the construct of this instrument was not supported by factor analysis. There was no study that validated DQoL successfully using exploratory factor analysis including our data so far. This could be due to some of the items that are not be suitable for majority of the diabetic patients. For instance, a question with regard to sexual life is quite sensitive for some countries [28] and thus leads to a more protective response towards the question. In addition, this item may be irrelevant for patients who are not sexually active. Apart from that, there were also redundancies among the items especially between “satisfaction” and “impact” domains. Therefore, the idea to come out with the revised version was to overcome these issues.

Ideally, all the items should be relevant as a proxy to measure quality of life. However, a well-constructed instrument should be supported by the statistical evidence besides the content alone. Some analyses of validation had been conducted for DQoL instrument [19], but it was insufficient since all the effort to prove the validity of current DQoL instrument based on exploratory factor analysis had failed. As exploratory factor analysis is the most common method to validate a quality of life instrument [1113], it is necessary to reevaluate the items and retest the validity of the revised version of DQoL instrument. It is also crucial to strengthen and improve the stability of the DQoL questionnaire to be used for diabetes research.

The process to validate the current DQoL instrument was successful after few items were omitted. Justifications to omit the items were based on the percentage of missing values and issue of duplication. The items with high percentage of missing values might indicate that the items were irrelevant for most of the diabetic patients. Majority of the items with high missing values were from “worry” domain. For example, items such as whether you will marry or get children are irrelevant for patients who are already married and have children. These items will lead to missing values if they are retained, and subsequently, the total score by domain will be difficult to compute.

Some items were found to have similar meaning especially from domains of “satisfaction” and “impact.” For example, items S6 and I5 were redundant in terms of meaning since both represent the extent of burden of diabetes which contributes to the family members. Individual item should only represent a single construct [29]. Hence, priority was given to impact domain when redundancy was detected between these two domains. The decision was done with consideration that the items reflect more in terms of the impact from diabetes.

For the revised version of DQoL instrument, we proposed the score for each domain and the total score to be converted to percentage as presented in Table 7. Any missing value, although it is unlikely to happen, can be substituted with the median score based on the response from the respective domain if the missing value is not more than one. If missing value is more than one in each domain, the domain score and the total score cannot be calculated based on the proposed rule. The original DQoL instrument allows more than one missing value in calculating a score of the domain [15]. However, in order to maintain the validity of the interpretation especially when the number of items in each domain had been reduced more than 50.0%, only one missing value is allowed for each domain.

Table 7: The proposed scoring for each domain and total score for a revised DQoL.

Result from exploratory factor analysis showed that the revised version of DQoL instrument has stable and strong construct. The validation was conducted among more than 500 patients in which the sample size was more than enough to conduct an exploratory factor analysis for 18 or even 46 items [30]. In other words, besides content validity, the conceptualization of the three domains was successfully validated. A sufficient sample size in addition to the fact that sufficient sample is needed to conduct factor analysis and to capture an accurate pattern of the responses from the respondents in the intended population is presented [31, 32]. The findings were also supported by Rasch analysis. The item’s and person’s reliability has improved compared with the original version. In addition, the infit and outfit of each and every item were also within the acceptable range.

The final model for confirmatory factor analysis consisted of 13 items with three factors fit well after respecification. The present study confirmed the validity for the three factors’ structure measurement model based on CF fit, RMSEA, SRMR, and TLI, providing a foundation for future study to be conducted to measure quality of life among diabetes patients in Malaysia. The composite reliability estimates proved the good construct validity of present study. Therefore, the evidence from the confirmatory factor analysis and Rasch analysis helped to strengthen the validity of the revised version of DQoL instrument.

Besides an excellent construct, the major strength with the revised version of DQoL instrument was that it has lesser number of items. This is certainly an advantage since few studies found that lengthy questionnaires are less likely to be completed [33, 34]. More complete responses for the questionnaire can be expected from the respondents using lesser items. The revised version of DQoL instrument has only 13 items, and hence, lesser time is needed to answer all the questions. The idea to produce a revised version is not only for patients’ convenience but most importantly to strengthen the instrument.

Screening was conducted to reevaluate the content so that it can be more representative of the domains that DQoL instrument aimed to measure. This is a common process in developing a questionnaire where item development and selection are one of the critical phases to produce a valid instrument with support from stronger statistical evidence [3537]. Reassessment has been made to all the DQoL items. There were 18 items selected for further evaluation using confirmatory factor analysis. Results from the analysis suggested that there were 13 items for the revised DQoL instrument.

Besides a shorter version, most importantly, the revised version of DQoL will be more suitable to reflect quality of life among patients with diabetes. Although validation of the revised instrument was conducted for the Malay version, the findings were supported with excellent statistical evidences. Therefore, we hope similar findings can be derived when the revised version of DQoL instrument is validated in other languages.

This study has some limitations. First of all, test-retest reliability of this questionnaire was not conducted. Future studies may consider to test the concurrent validity of the revised version of DQoL instrument using other questionnaires. Besides that, researchers are unsure to what extent the shorter version of DQoL instrument is sensitive in predicting future outcomes such as hospitalization due to diabetes complications and this too warrants future studies to be conducted.

5. Conclusion

In conclusion, a revised version of DQoL instrument that maintains the conceptualization of satisfaction, impact, and worry domains with only 13 items was successfully developed. Future studies should be conducted to validate the revised version of DQoL instrument in other languages.

Abbreviations

DQoL:Diabetes quality of life
DCCT:Diabetes Control and Complications Trial.

Data Availability

The data used to support the findings of this study are available from the corresponding author upon request.

Ethical Approval

An ethical approval was received on July 18, 2014, by the Medical Research and Ethics Committee (MREC) and Ministry of Health Malaysia with the NMRR ID: NMRR-14-522-19377.

Consent

A written consent was taken from all patients before they participated in the study.

Conflicts of Interest

The authors declare that they have no conflicts of interests.

Authors’ Contributions

Mohamad Adam Bujang formulated the research question, study design, interpretation, writing, and final proof of the manuscript. Tassha Hilda Adnan and Chien Joo Lim performed the statistical analysis, interpretation, writing, and final proof of the manuscript. Nur Khairul Bariyyah did the study design, data collection, writing, and final proof of the manuscript. Mastura Ismail executed the study design, writing, and final proof of the manuscript.

Acknowledgments

The authors would like to extend their appreciation to the Director General, Ministry of Health Malaysia, for giving them permission to publish this study. In addition, the authors would like to thank the National Institute of Health, Ministry of Health Malaysia, for sponsoring a grant for this study and Ms. Shirin Tan Hui for proofreading the manuscript.

Supplementary Materials

Supplementary material contains the Revised Version of DQOL (RV-DQOL13) questionnaire with 13 items. (Supplementary Materials)

References

  1. The DCCT Research Group, “Reliability and validity of a diabetes quality-of-life measure for the diabetes control and complications trial (DCCT),” Diabetes Care, vol. 11, no. 9, pp. 725–732, 1988. View at Publisher · View at Google Scholar · View at Scopus
  2. The Diabetes Control and Complications Trial Research Group, “Influence of intensive diabetes treatment on quality-of-life outcomes in the diabetes control and complications trial,” Diabetes Care, vol. 19, no. 3, pp. 195–203, 1996. View at Publisher · View at Google Scholar · View at Scopus
  3. J. Reviriego, M. D. Millan, and M. Millan, “Evaluation of the diabetes quality-of-life questionnaire in a Spanish population. An experience of translation and reliability,” PharmacoEconomics, vol. 10, no. 6, pp. 614–622, 1996. View at Publisher · View at Google Scholar · View at Scopus
  4. A. Yildirim, F. Akinci, H. Gozu, H. Sargin, E. Orbay, and M. Sargin, “Translation, cultural adaptation, cross-validation of the Turkish diabetes quality-of-life (DQOL) measure,” Quality of Life Research, vol. 16, no. 5, pp. 873–879, 2007. View at Publisher · View at Google Scholar · View at Scopus
  5. F. Sato, T. Mita, R. Yamamoto et al., “Reliability and validity of the Japanese version of the diabetes quality-of-life questionnaire for Japanese patients with type 2 diabetes mellitus,” Diabetology international, vol. 5, no. 1, pp. 21–29, 2014. View at Publisher · View at Google Scholar · View at Scopus
  6. A. Y. Cheng, E. Y. Tsui, A. J. Hanley, and B. Zinman, “Developing a quality of life measure for Chinese patients with diabetes,” Diabetes Research and Clinical Practice, vol. 46, no. 3, pp. 259–267, 1999. View at Publisher · View at Google Scholar · View at Scopus
  7. I. C. Huang, J. H. Liu, A. W. Wu, M. Y. Wu, W. Leite, and C. C. Hwang, “Evaluating the reliability, validity and minimally important difference of the Taiwanese version of the diabetes quality of life (DQOL) measurement,” Health and Quality of Life Outcomes, vol. 6, no. 1, pp. 87–90, 2008. View at Publisher · View at Google Scholar · View at Scopus
  8. A. H. Pakpour, M. Saffari, and A. Burri, “Translation and validation of an Iranian version of the diabetes quality of life measure,” Journal of diabetes investigation, vol. 3, no. 5, pp. 471–478, 2012. View at Publisher · View at Google Scholar · View at Scopus
  9. M. A. Bujang, M. Ismail, N. K. B. H. Hatta, S. H. Othman, N. Baharum, and S. S. M. Lazim, “Validation of the Malay version of diabetes quality of life (DQoL) questionnaire for adult population with type 2 diabetes mellitus,” Malaysian Journal of Medical Sciences, vol. 24, no. 4, pp. 86–96, 2017. View at Publisher · View at Google Scholar · View at Scopus
  10. A. T. Marcia and C. S. Donald, “Assessment of quality-of-life outcomes,” The New England Journal of Medicine, vol. 334, pp. 835–840, 1996. View at Publisher · View at Google Scholar · View at Scopus
  11. J. E. Ware, M. A. Kosinski, and S. D. Keller, SF-36 Physical and Mental Health Summary Scales: a User’s Manual, The health institute, New England Medical Center, Boston, MA, USA, 1994.
  12. L. Vileikyte, M. Peyrot, C. Bundy et al., “The development and validation of a neuropathy- and foot ulcer-specific quality of life instrument,” Diabetes Care, vol. 26, no. 9, pp. 2549–2555, 2003. View at Publisher · View at Google Scholar · View at Scopus
  13. E. Eypasch, J. I. Williams, S. Wood-Dauphinee et al., “Gastrointestinal quality of life index: development, validation and application of a new instrument,” The British Journal of Surgery, vol. 82, no. 2, pp. 216–222, 1995. View at Publisher · View at Google Scholar · View at Scopus
  14. E. B. Thomas, D. Radhika, M. W. Brian, G. Debra, and M. G. Janet, “Development and validation of the diabetes quality of life brief clinical inventory,” Diabetes Spectrum, vol. 17, no. 1, pp. 41–49, 2004. View at Publisher · View at Google Scholar
  15. C. Bradley, “Chapter 5: the diabetes quality of life measure,” in Handbook of Psychology and Diabetes, p. 67, Psychology Press, New York, NY, USA, 2003. View at Google Scholar
  16. G. Rasch, Probabilistic Models for Some Intelligence and Attainment Tests, MESA Press, Chicago, IL, USA, 1960.
  17. L. Steinberg and T. David, “Uses of item response theory and the testlet concept in the measurement of psychopathology,” Psychological Methods, vol. 1, no. 1, pp. 81–97, 1996. View at Publisher · View at Google Scholar · View at Scopus
  18. A. M. Smith, “Some problems when adopting Churchill’s paradigm for the development of service quality measurement scales,” Journal of Business Research, vol. 46, no. 2, pp. 109–120, 1999. View at Publisher · View at Google Scholar · View at Scopus
  19. J. M. Linacre, “Size vs. significance: standardized chi-square fit statistic,” Rasch Measurement Transactions, vol. 17, p. 918, 2003. View at Google Scholar
  20. J. M. Linacre, “Data variance explained by measures,” Rasch Measurement Transactions, vol. 20, pp. 1045–1047, 2006. View at Google Scholar
  21. T. G. Bond and C. M. Fox, Applying the Rasch Model: Fundamental Measurement in the Human Sciences, Lawrence Erlbaum Associates, Inc, Mahwah, NJ, USA, 2001.
  22. W. P. Fisher, “Rating scale instrument quality criteria,” Rasch Measurement Transaction, vol. 21, article 1095, 2007. View at Google Scholar
  23. U. Bott, I. Muhlhauser, H. Overmann, and M. Berger, “Validation of a diabetes-specific quality-of-life scale for patients with type 1 diabetes,” Diabetes Care, vol. 21, no. 5, pp. 757–769, 1998. View at Publisher · View at Google Scholar · View at Scopus
  24. R. R. Rubin and M. Peyrot, “Quality of life and diabetes,” Diabetes/Metabolism Research and Reviews, vol. 15, no. 3, pp. 205–218, 1999. View at Publisher · View at Google Scholar
  25. G. R. J. Parkerson, R. T. Connis, W. E. Broadhead, D. L. Patrick, T. R. Taylor, and C. K. Tse, “Disease-specific versus generic measurement of health-related quality of life in insulin-dependent diabetic patients,” Medical Care, vol. 31, no. 7, pp. 629–639, 1993. View at Publisher · View at Google Scholar · View at Scopus
  26. A. M. Jacobson, J. A. Samson, M. De Groot, and J. A. Samson, “The evaluation of two measures of quality of life in patients with type I and type II diabetes,” Diabetes Care, vol. 17, no. 4, pp. 267–274, 1994. View at Publisher · View at Google Scholar · View at Scopus
  27. G. C. Williams and A. Zeldman, “Patient-centered diabetes self-management education,” Current Diabetes Reports, vol. 2, no. 2, pp. 145–152, 2002. View at Publisher · View at Google Scholar · View at Scopus
  28. S. Hisasue, Y. Kumamoto, Y. Sato et al., “Prevalence of female sexual dysfunction symptoms and its relationship to quality of life: a Japanese female cohort study,” Urology, vol. 65, no. 1, pp. 143–148, 2005. View at Publisher · View at Google Scholar · View at Scopus
  29. K. E. A. Burns, M. Duffett, M. E. Kho et al., “A guide for the design and conduct of self-administered surveys of clinicians,” CMAJ, vol. 179, no. 3, pp. 245–252, 2008. View at Publisher · View at Google Scholar · View at Scopus
  30. M. Bujang, P. A. Ghani, S. A. Soelar, and N. A. Zulkifli, “Sample size guideline for exploratory factor analysis when using small sample: taking into considerations of different measurement scales,” in 2012 International Conference on Statistics in Science, Business and Engineering (ICSSBE), pp. 1–5, Langkawi, Malaysia, September 2012. View at Publisher · View at Google Scholar · View at Scopus
  31. M. A. Bujang, P. A. Ghani, N. A. Zolkepali, S. Selvarajah, and J. Haniff, “A comparison between convenience sampling versus systematic sampling in getting the true parameter in a population: explore from a clinical database: the Audit Diabetes Control Management (ADCM) registry in 2009,” in 2012 International Conference on Statistics in Science, Business and Engineering (ICSSBE), pp. 1–5, Langkawi, Malaysia, September 2012. View at Publisher · View at Google Scholar · View at Scopus
  32. M. A. Bujang, N. Sa’at, R. A. Joys, and M. M. Ali, “An audit of the statistics and the comparison with the parameter in the population,” AIP Conference Proceedings, vol. 1682, article 050019, 2015. View at Publisher · View at Google Scholar · View at Scopus
  33. J. Fox, “Designing research: basics of survey construction,” Minimally Invasive Surgical Nursing, vol. 8, no. 2, pp. 77–79, 1994. View at Google Scholar
  34. D. A. Dillman, “Chapter 10 - mail and other self-administered questionnaires,” in Handbook of Survey Research, P. H. Rossi, J. D. Wright, and A. B. Anderson, Eds., pp. 359–376, Academic Press, San Diego, CA, USA, 1983. View at Publisher · View at Google Scholar
  35. E. F. Juniper, G. H. Guyatt, R. S. Epstein, P. J. Ferrie, R. Jaeschke, and T. K. Hiller, “Evaluation of impairment of health related quality of life in asthma: development of a questionnaire for use in clinical trials,” Thorax, vol. 47, no. 2, pp. 76–83, 1992. View at Publisher · View at Google Scholar · View at Scopus
  36. P. L. Hudak, P. C. Amadio, and C. Bombardier, “Development of an upper extremity outcome measure: the DASH (disabilities of the arm, shoulder, and head),” American Journal of Industrial Medicine, vol. 29, no. 6, pp. 602–608, 1996. View at Publisher · View at Google Scholar
  37. E. F. Juniper, P. M. Obyrne, G. H. Guyatt, P. J. Ferrie, and D. R. King, “Development and validation of a questionnaire to measure asthma control,” European Respiratory Journal, vol. 14, no. 4, pp. 902–907, 1999. View at Publisher · View at Google Scholar · View at Scopus