Abstract
Background: Patients with type 2 diabetes reveal different obstacles in living with the disease. The EGPRN initiated a qualitative research EUROBSTACLE to create a broadly conceptualized diabetes-related quality of life (DR-QoL) instrument. It led to the development of the diabetes obstacle questionnaire (DOQ), a five-point Likert-scaled measure, consisting of 78 items in eight scales.
Objectives: To develop and validate a short, easy-to-use version of the DOQ.
Methods: A cross-sectional study with the DOQ was carried out. Participants answered the DOQ and GPs added some clinical data from their medical records. Data of 853 patients from Belgium, France, Estonia, Serbia, Slovenia, and Turkey were included in the analysis. The selection of items for the short version of the DOQ was achieved with exploratory factor analysis (EFA). Construct validity was proved with EFA and Pearson correlations between the DOQ and the new DOQ-30. Internal reliability was established with Cronbach’s alpha.
Results: DOQ-30 resulted in 30 items in nine subscales. It explained 49.8% of items’ variance. It shows a considerable good internal reliability and construct validity.
Conclusion: The DOQ-30 is a five-point Likert-scaled broadly conceptualized measure of DR-QoL. It addresses a variety of obstacles, such as social, psychological, cognitive and behavioural. The DOQ-30 is ready for implementation in general practice and research in Europe as a valuable instrument to assess DR-QoL.
The DOQ-30 is a new broadly conceptualised DR-QoL instrument addressing a variety of obstacles patients might confront in everyday life.
It is based on qualitative research in six European countries and reveals good internal reliability, and external and construct validity.
KEY MESSAGE
Introduction
Diabetes is a chronic disease of high importance. In the WHO report ‘global status report on non-communicable diseases 2010,’ it is assessed that diabetes will become the seventh leading cause of death by 2030 (Citation1), and diabetic patients are persistently in poor glycaemic control (Citation2). It is known that relevant health outcomes include not only biomedical and functional dimensions of health measures but also subjective considerations such as disease self-management burden, emotional health, and social and physical functioning, the so-called diabetes-related quality of life (DR-QoL) indicators (Citation3). Patients’ self-management behaviours affect DR-QoL, for example, intensification of treatment regimen and patients’ subjective cost-benefit regards can influence patients’ decisions (Citation4). Therefore, the clinical effectiveness assessments should incorporate patient-centred outcome measures.
To attempt a closer understanding of the theme from the point of view of the patient, it is of the utmost importance to research patient-directed experiences and their fears of living with diabetes (Citation5). During the past few decades, considerable effort has been devoted to the study of patients’ self-management behaviours and adherence to the treatment. In DR-QoL research, questionnaires that are in use are listed in Appendix 1 (Citation6–25).
ATT39, PAID, DDS, the QSD-R, DHP, ADS, and ADDQoL are focused only on some aspects such as diabetes-related distress or additionally on the activity or the eating behaviour. EDBS is a measure for elderly people. ASK-20 is not dedicated specifically to patients with diabetes, but also for other chronic diseases like asthma and congestive heart failure.
In their review articles, researchers Watkins and Connell (Citation4) and Achhab et al. (Citation26) were of the same opinion that DCP, DIMS, DQOL, DSQOLS, and the D-39 were broadly conceptualized diabetes-specific QoL questionnaires. They found that DQOL and DSQOLS were mainly dedicated to patients with type 1 diabetes. Patients were not involved in the derivation of items for the ADS, DCP, DIMS and D-39. Factor analysis was not used to support construct validity of the ADS, DQOL and DQLCTQ-R. Almost all instruments were developed and validated in industrialized countries (Citation4,Citation26). Researchers in Bulgaria assessed ADDQoL and DCP as the most promising instruments for measuring the DR-QoL (Citation27). In research, the most often used PAID and ADDQoL are not broadly conceptualized but single-factor measures of diabetes-related distress. These are sometimes used with SF-36 (Citation28,Citation29) . For future research in this area, it is recommended to increase the racial, cultural and ethnic diversity of research participants, and to develop the same questionnaire for low-income countries (Citation4,Citation26).
In 2000, EGPRN researchers initiated the international research project EUROBSTACLE to fulfil the gap of broadly conceptualized DR-QoL instruments that take in obstacles of patients from different ethnic, cultural and healthcare systems.
The first phase, qualitative research using focus groups on 246 patients was carried out in six European countries: Croatia, Estonia, France, the Netherlands, Slovenia and the UK (Citation30). It resulted in the creation of the diabetes obstacle questionnaire (DOQ). The DOQ is a five-point Likert-scaled instrument, consisting of 78 items grouped in eight scales. The Warwick diabetes care research user group in the UK gave feedback to the research team on the questionnaire design and content (Citation31). In the second phase, a cross-sectional study using the DOQ was conducted in May to November 2009 on patients with type 2 diabetes (T2DM) in seven countries: Belgium, France, Estonia, Serbia, Slovenia, the UK and Turkey. The DOQ was further validated in the UK, Belgium and Estonia (Citation31–33). The Pearson correlation coefficient was calculated between the DOQ and PAID, ADDQoL, and HbA1c in the UK and the Belgian sample to determine construct validity (Citation31,Citation32). A broad variety of obstacles was demonstrated by Estonian sample (Citation33). An important disadvantage of the DOQ is that the questionnaire is rather time-consuming for patients to answer and for healthcare providers to administer. In this study, we intended to develop and validate a short, easy-to-use version of the DOQ.
Methods
Development of the dataset
The design of the cross-sectional study with the DOQ was agreed by researchers from Belgium, France, Estonia, Serbia, Slovenia and Turkey: GPs enrolled at least three consecutive outpatient diabetes patients into the study. Researchers gave a study kit comprising an information leaflet, a questionnaire, extract from their medical records with their latest clinical results, and prepaid, self-addressed envelopes to participants. The method is described elsewhere (Citation31–33). The questionnaire was translated into native languages and back into English in all six countries. Altogether, 860 participants were enrolled in the study. Missing data analyses revealed that 441 respondents answered all items. In 66 cases, one answer was lacking and in 60 cases two answers. Seven cases with more than 25% missing data were eliminated. Consequently, the data for 853 respondents were included in statistical analyses. The number of missing values for items increased from 1.3% to 20.9%, with a mean of 4.4%.
Multiple imputations were used to handle missing data. All the results from six countries were included in the dataset to perform statistical analyses. Descriptive characteristics concerning gender, age, diabetes duration, type of diabetic treatment and complications for each included sample were computed.
Exploratory factor analysis for the selection of items
All 78 items were inserted into exploratory factor analyses (EFA) to identify new subscales, to select the items of greater relevance, and to support construct validity for the short version of the DOQ. We used principal axis factoring (PAF) as the extraction method and Varimax with Kaiser Normalization as the rotation method. PAF is a correlation-focused approach and preferred for causal modelling (Citation33,Citation34). It is shown that the factor pattern obtained by Varimax gives a clearer separation of the factors (Citation35,Citation36). Factor loadings of > 0.5 are considered good, 0.3–0.5 moderate and loadings of <0.3 are considered weak (Citation37,Citation38). The scree plot, eigenvalues and content analysis were examined to determine meaningful factors. Each factor represented as a new subscale. We included 4–2 items with factor loadings ≥0.5 into each subscale for a new short easy-to-use version of the DOQ.
Cronbach’s alpha for internal reliability
Cronbach’s alpha as an internal reliability index was calculated separately for all new subscales.
Pearson correlation coefficient to determine construct validity
To confirm construct validity of the new questionnaire the Pearson correlation coefficients were calculated between the scales of the DOQ and the subscales of the new shortened version of the DOQ. We summed raw scores of the scales of the DOQ and of all items loading on the subscales of the shortened DOQ. This method preserved the variation in the original data (Citation39). Statistical analyses were conducted using the IBM SPSS Statistics Version 20 and R (version 3.1, Lavaan 0.5–17).
Results
Descriptive statistics
In the study group of 853 participants, the age of participants ranged from 27–89 years, the mean age was 64 (SD 10.5) years and the mean duration of T2DM was 7.3 (SD 6.7) years. Among the participants, 49.6% were male. The descriptive characteristics of the whole sample are presented in .
Table 1. Baseline statistics of the study sample.
Descriptive statistics and occurrence of complication showed a considerable range between countries. The data for diabetes complications were missing in the Turkish and Belgian samples; these were marked as NR.
Exploratory factor analyses
EFA was computed on the whole set of 78 items. It resulted in item-to-factor loadings from 0.42 to 0.85 and 18 factors with an eigenvalue >1 explaining 51.5% of items’ variance. A close examination of the scree plot, eigenvalues, and content analysis indicated nine meaningful subscales with four items in the first six and two in the last three—altogether 30 items. The new scale explained 49.8% of the items’ variance. Compared to the 18-factor structure that explains 51.5% of items’ variance, we have lost only 0.7% of items’ variance by nine-factor construction. The Kaiser–Meyer–Olkin measure of sampling adequacy was 0.92, and Bartlett’s test of sphericity was significant (P <0.001), which indicated that the nine-factor solution was appropriate. We named the short easy-to-use version of the DOQ the DOQ-30. The results are shown in (the DOQ-30 questionnaire can be found in the Supplementary material online).
Table 2. Factor structure in EFA of DOQ-30 of the dataset of six European countries; Cronbach’s alpha of the subscales.
Cronbach’s alpha ranged from 0.52–0.89
Cronbach’s alpha was 0.52 for ‘uncertainty about a consultation’ containing two items drawn from different scales of the DOQ.
Pearson correlations between the DOQ and the DOQ-30
Pearson correlation between scale scores of the DOQ and the DOQ-30 showed high coefficients from 0.58–0.99. The results are shown in .
Table 3. Pearson correlation between scale scores of the DOQ and the DOQ-30 of the dataset of six European countries. All coefficients are significant (P <0.001).
The subscale ‘uncertainty about a consultation’ was extracted from ‘obstacles at diagnosis scale’ r = 0.71 and ‘obstacles in relationships with healthcare professionals scale’ r = 0.49. The original scale ‘medication obstacles scale’ was cleaved into two subscales: ‘medication’ r = 0.58 and ‘Insulin-use’ r = 0.70. No item was presented from ‘obstacles to coping with diabetes scale.’ Conversely, it revealed high correlations with three new subscales: ‘lifestyle changes’ r = 0.58, ‘support from friends and family’ r = 0.55, and ‘uncertainty about a consultation’ r = 0.43. The contents of the DOQ-30 in six countries that emerged were comparable to the original scales of the DOQ validated in the UK, Belgium, and Estonia.
Discussion
Main findings
At the beginning of this millennium, there were several DR-QoL measures available that did not meet requirements of a broadly conceptualized easy-to-use DR-QoL instrument. In 2000, the EGPN initiated the project EUROBSTACLE for the development of a DR-QoL instrument meeting these demands. Patients from countries with different ethnic, cultural and healthcare systems had to be involved in the development of the tool. The goal succeeded in the creation of the DOQ-30.
The development of the DOQ-30 began with the qualitative research project EUROBSTACLE on 246 patients in six European countries (Citation30). Previously, the development of DM-QoL instrument had been based on different generic and disease-specific measures of subjective health status, followed by consultations of diabetes healthcare professionals and conversations or small, unstructured interviews with patients. For ADDQoL, there were 12 interviews (Citation19), for D-39 an unknown number of interviews with patients (Citation7) and for DIMS and DCP no conversations with a patient were mentioned (Citation14,Citation21). We did not find any qualitative research with international design in the development of a DR-QoL instrument. The EUROBSTACLE project resulted in the creation of the diabetes obstacle questionnaire (DOQ) in 78 items. We intended to create an ideal research tool of a reasonable volume by means of EFA and validate its construct with EFA and Pearson correlations between the original DOQ and the new shortened one. We formed a database for factor analysis from a cross-sectional study with the DOQ in six countries. The descriptive characteristics concerning age, gender, duration of the disease and type of treatment showed considerable differences between data of included samples (). The EFA resulted in a nine-factor structure, which explains 49.8% of the items’ variance. So, the DOQ-30 in nine subscales with 30 items was created (). We decided to preserve the two-item subscale ‘uncertainty with the consultation’ because of the very important theme. The subscale showed a moderate internal homogeneity with a Cronbach’s alpha of 0.52. It is shown that Cronbach’s alpha is influenced too much by the amount of the items included in a scale. The lower the number of items, the lower the value for Cronbach’s alpha (Citation40).
We computed Pearson correlations to ensure that by reducing the questionnaire we do not lose important information gathered in the qualitative phase of the project. It also showed very high correlation coefficients from 0.58–0.99 (). Conclusively, the results of EFA and Pearson correlation coefficients were of high construct validity.
The DOQ-30 showed comprehensive areas of barriers in everyday life with T2DM. Social aspects are studied in the scale ‘relationships with medical professionals,’ which primarily describes physicians’ skills of communication and elucidates the concerns related to the patient, and in the scale ‘support from friends and family,’ which deals with loneliness and desire for support from family, friends and society. Psychological aspects are studied in the scales ‘lifestyle changes,’ ‘exercising’ and ‘uncertainty about a consultation.’ The scales showed how considerably resentful patients with T2DM may feel if they have to change their habits because of the disease. The scale ‘knowledge of the disease’ reflects cognition and understanding about the disease. The three scales ‘self-monitoring,’ ‘medication,’ and ‘use of insulin’ express attitudes and fears to treating process and behaviour. In comparing the DOQ-30 with other diabetes-related QoL research measures, DOQ-30 is more multidimensional and covers a comprehensive variety of barriers. Almost all diabetes questionnaires that were evaluated as broadly conceptualized exclude statements about themes, like relationships with medical personnel (DIMS, DCP, diabetes-39), physical activity (DIMS, diabetes-39), self-monitoring (DIMS), or concerns about lack of knowledge of T2DM (DIMS). However, all discussed questionnaires deal with patients’ feelings and moods associated with separate aspects of psychosocial distress in diabetes. The use of DOQ-30 may help to stimulate conversation between the caregiver and patient, to explain the obstacles in adhering to their self-management, and targeted to promote patients.
Limitations
Despite the strength of its contributions, the DOQ-30 shows that some limitations have to be acknowledged.
First, the dataset contained 4.4% of missing values. We used multiple imputations to analyse incomplete data. The most optimal solution has been proposed. Second, even though the new subscales of the DOQ-30 cover those of the DOQ, we lost some information. This loss can be justified by reducing duplication of the information and creating a reasonable easy-to-use questionnaire.
As the validation of measures is an ongoing process, future studies should corroborate DOQ-30 and prove its relevance for the diabetic patient. Further research is required to evaluate responsiveness through longitudinal comparisons of instruments within clinical trials.
Conclusion
The DOQ-30 is a five-point Likert-scaled measure of DR-QoL in nine subscales and contains 30 items with good internal reliability, and external and construct validity.
The DOQ-30 addresses a variety of obstacles, which patients might confront, such as social, psychological, cognitive and behavioural. The DOQ-30 could be implemented in general practice and research in Europe as a valuable instrument to assess DR-QoL.
Supplementary_appendix:_DOQ-30
Download PDF (63.6 KB)Acknowledgements
The authors gratefully acknowledge the patients and their GPs who agreed to participate in this study and express their gratitude to the EGPRN group. They should like to thank K. The, A. Kuusk, M.Raag, A. Rozeik and R. Salupere for their time and advice. The Turkish part of the study was supported by the Akdeniz University Project Management Unit. During the design and conduct of the study, the University of Tartu funded L. Pilv as a full-time state-funded Ph.D. student. The data referred to in the current study were gathered with the support of the Estonian Science Foundation Grant no. ETF 7596.
Declaration of interest
The authors report no conflicts of interest. The authors alone are responsible for the content and writing of the paper.
References
- World Health Organization. Global status report on noncommunicable diseases 2010. Geneva: World Health Organization; 2011.
- American Diabetes Association. Standards of medical care in diabetes—2011. Diabetes Care 2011;34:S11–S61.
- Barr JT. The outcomes movement and health status measures. J Allied Health 1995;24:13–28.
- Watkins K, Connell CM. Measurement of health-related QOL in diabetes mellitus. Pharmaco Economics 2004;22:1109–26.
- Vermeire E, Hearnshaw H, Van Royen P, Denekens J. Patient adherence to treatment: Three decades of research. A comprehensive review. J Clin Pharm Ther. 2001;26:331–42.
- Dunn SM, Smartt HH, Beeney LJ, Turtle JR. Measurement of emotional adjustment in diabetic patients: Validity and reliability of ATT39. Diabetes Care 1986;9:480–89.
- Boyer JG, Earp JAL. The development of an instrument for assessing the quality of life of people with diabetes: Diabetes-39. Med Care 1997;35:440–53.
- Polonsky WH, Anderson BJ, Lohrer PA, Welch G, Jacobson AM, Aponte JE, et al. Assessment of diabetes-related distress. Diabetes Care 1995;18:754–60.
- Welch GW, Jacobson AM, Polonsky WH. The problem areas in diabetes scale. An evaluation of its clinical utility. Diabetes Care 1997;20:760–66.
- Welch G, Weinger K, Anderson B, Polonsky WH. Responsiveness of the problem areas in diabetes (PAID) questionnaire. Diabet Med. 2003;20:69–72.
- Polonsky WH, Fisher L, Earles J, Dudl RJ, Lees J, Mullan J, et al. Assessing psychosocial distress in diabetes: Development of the diabetes distress scale. Diabetes Care 2005;28:626–31.
- Herschbach P, Duran G, Waadt S, Zettler A, Amm C, Marten-Mittag B. Psychometric properties of the questionnaire on stress in patients with diabetes—revised (QSD-R), Health Psychol. 1997;16:171–74.
- Meadows KA, Abrams C, Sandbaek A. Adaptation of the diabetes health profile (DHP-1) for use with patients with type 2 diabetes mellitus: Psychometric evaluation and cross-cultural comparison. Diabet Med. 2000;17:572–80.
- Hammond GS, Aoki TT. Measurement of health status in diabetic patients: Diabetes impact measurement scales. Diabetes Care 1992;15:469–77.
- Shen W, Kotsanos JG, Huster WJ, Mathias SD, Andrejasich CM, Patrick DL. Development and validation of the diabetes quality of life clinical trial questionnaire. Med Care 1999;37:AS45–66.
- Hahn SR, Park J, Skinner EP, Yu-Isenberg KS, Weaver MB, Crawford, B, et al. Development of the ASK-20 adherence barrier survey. Curr Med Res Opin. 2008;24:2127–38.
- Matza LS, Yu-Isenberg KS, Coyne KS, Park J, Wakefield J, Skinner J, et al. Further testing of the reliability and validity of the ASK-20 adherence barrier questionnaire in a medical center outpatient population. Curr Med Res Opin. 2008;24:3197–206.
- Carey MP, Jorgensen RS, Weinstock RS, Sprafkin RP, Lantinga LJ, Carnrike CLM Jr, et al. Reliability and validity of the appraisal of diabetes scale. J Behav Med. 1991;14:43–50.
- Bradley C, Todd C, Gorton T, Symonds E, Martin A, Plowright R. The development of an individualized questionnaire measure of perceived impact of diabetes on quality of life: The ADDQoL. Qual Life Res 1999;8:79–91.
- Bradley C, Speight J. Patient perceptions of diabetes and diabetes therapy: Assessing quality of life. Diabetes Metab Res Rev. 2002;18:S64–9.
- Fitzgerald JT, Davis WK, Connell CM, Hess GE, Funnell MM, Hiss RG. Development and validation of the diabetes care profile. Eval Health Prof. 1996;19:208–30.
- Fitzgerald JT, Gruppen LD, Anderson RM, Funnell MM, Jacober SJ, Grunberger G, et al. The influence of treatment modality and ethnicity on attitudes in type 2 diabetes. Diabetes Care 2000;23:313–18.
- The DCCT Research Group. Reliability and validity of a diabetes quality-of-life measure for the diabetes control and complications trial (DCCT). Diabetes Care 1988;11:725–32.
- Bott U, Mühlhauser I, Overmann H, Berger M. Validation of a diabetes-specific quality-of-life scale for patients with type 1 diabetes. Diabetes Care 1998;21:757–69.
- Araki A, Ito H. Development of elderly diabetes burden scale for elderly patients with diabetes mellitus. Geriatr Gerontol Int. 2003;3:212–24.
- El Achhab Y, Nejjari C, Chikri M, Lyoussi B. Disease-specific health-related quality of life instruments among adults diabetic: A systematic review. Diabetes Res Clin Pract. 2008;80:171–84.
- Levterova BA, Dimitrova DD, Levterov GE, Dragova EA. Instruments for disease-specific quality-of-life measurement in patients with type 2 diabetes mellitus–a systematic review. Folia Med. (Plovdiv) 2013;55:83–92.
- Aguiar CC, Vieira AP, Carvalho AF, Montenegro-Junior RM. Assessment instruments for a health-related quality of life in diabetes mellitus. Arq Bras Endocrinol Metabol. 2008;52:931–9. Portuguese.
- Woodcock, A J, Julious SA, Kinmonth AL, Campbell MJ; Diabetes Care from Diagnosis Group. Problems with the performance of the SF-36 among people with type 2 diabetes in general practice. Qual Life Res. 2001;10:661–70.
- Vermeire E, Hearnshaw H, Rätsep A, Levasseur G, Petek D, van Dam H, et al. Obstacles to adherence in living with type-2 diabetes: An international qualitative study using meta-ethnography (EUROBSTACLE). Prim Care Diabetes 2007;1:25–33.
- Hearnshaw, H, Wright K, Dale J, Sturt J, Vermeire E, van Royen P. Development and validation of the diabetes obstacles questionnaire (DOQ) to assess obstacles in living with type 2 diabetes. Diabetic Med. 2007;24:878–82.
- Pilv, L, Rätsep A, Oona M, Kalda R. Prevalent obstacles and predictors for people living with type 2 diabetes. Int J Family Med. 2012;2012:842–912.
- Garson GD. Factor analysis. Asheboro, NC: Statistical Associates Publishers; 2013.
- de Vet HCW, Adér HJ, Terwee CB, Pouwer F. Are factor analytical techniques used appropriately in the validation of health status questionnaires? A systematic review on the quality of factor analysis of the SF-36. Qual Life Res. 2005;14:1203–18.
- Pett MA, Lackey NR, Sullivan JJ, editors. Making sense of factor analysis. Thousand Oaks, CA: SAGE Publications, Inc.; 2003. doi: http://dx.doi.org/10.4135/9781412984898.
- Kim J, Mueller CW. Factor analysis. Thousand Oaks, CA: SAGE Publications, Inc.; 1978. doi: http://dx.doi.org/10.4135/9781412984256.
- Gaur AS, Gaur SS. Statistical methods for practice and research: A guide to data analysis using SPSS. 2nd ed. New Delhi: SAGE Publications India Pvt Ltd; 2009. doi: http://dx.doi.org/10.4135/9788132108306.
- Cohen J. Statistical power analysis for the behavioral sciences. 2nd ed. New Jersey: Lawrence Erlbaum Associates; 1988.
- DiStefano, C, Zhu M, Mindrila D. Understanding and using factor scores: Considerations for the applied researcher. Practical Assessment, Research & Evaluation 2009;14:1–11.
- Cortina JM. What is coefficient alpha? An examination of theory and applications. J Appl Psychol. 1993;78:98.
Appendix 1
DR-QOL RESEARCH QUESTIONNAIRES
• ATT39, measurement of emotional adjustment in diabetic patients (1986)
• D-39, diabetes-39 (1997)
• PAID, the problem areas in diabetes (1995)
• DDS, the diabetes distress scale (2005)
• QSD-R, the questionnaire on stress in patients with diabetes (1986)
• DHP-1, the diabetes health profile (2000)
• DIMS, diabetes impact measurement scales (1992)
• DQLCTQ-R, the diabetes quality of life clinical trial questionnaire (1999)
• ASK-20, questionnaire to assess barriers to medication adherence (2008)
• ADS, the appraisal of diabetes scale
• ADDQoL, the audit of diabetes-dependent QoL (1999)
• DCP, the diabetes care profile (1996)
• DQOL, the diabetes quality of life
• DSQOLS, the diabetes-specific quality of life scale
• EDBS the elderly diabetes burden scale