Diabetes, which affects many systems as well as mental health [1], is a common disease [2] and when considering the mortality and higher health care costs due to its management and treatment become important. Patients with diabetes are diagnosed with depression by 2 to 4 times more frequent compared to the non-diabetics [3-6]. Recurrent physical and emotional burdens that often accompany diabetic self-management have been proposed to play an important role in the increased risk for depression [7]. Because persons do not consider short- and long-term complications of the disease, they may have difficulty in acceptance of lifestyle changes, and encounter with depression, anxiety and similar psychiatric problems [8]. Depression and anxiety disorders in diabetic persons are characterized by depressed mood, decreased energy and attention, sense of guilt, poor concentration, weight loss, decrease or increase in appetite, sleep disorders, thoughts of death and suicidal ideation and dysfunction [9].

Treatment process of diabetes is often long and challenging. As in almost all chronic diseases, attitudes and approaches of person toward diabetes and its treatment may affect the prognosis [10]. For example, it is known that acceptance of the disease and the difficulties it brings is challenging [11]. Compliance to treatment in diabetes has been shown to increase glycemic control [12]. High prevalence of negative psychological processes and the difficulties due to medical therapy process given raise to the need for development of psychotherapeutic interventions. Accordingly, possible effect of the cognitive-behavioral therapies (CBT) both on anxiety and depression, and on improvement of glycemic control have been studied in recent years. Reviews and meta-analyses including CBT studies on this issue indicate that CBT provided significant decreases in anxiety and depression levels, but did not show a substantial effect on glycemic control [13,14]. However, Grey et al. [15] achieved significant results such as lower glycosylated hemoglobin, higher diabetic and medical self-sufficiency, and decreased effect of diabetes on quality of life. Their group therapy protocol included 6 weekly sessions and monthly visits over the 12 months of follow up was applied to 77 patients with diabetes in 12-20 age group.

Some coping strategies have been found to have negative effects in stress reduction and providing ideal compliance to treatment during the treatment process. For example, a coping method like avoidance has been shown as a factor preventing compliance to treatment in diabetes [16]. Supporting this result, non-acceptance of diabetes by the person has been found to be significantly associated with low active coping, decreased self-care, high HbA1c levels, elevated diabetes related stress, and increased depressive symptoms [17]. Again, high capacity of coping with the disease [18] and low HbA1c levels [19] have been shown in persons with high acceptance. In this direction, it has been found that acceptance-based new generation cognitive-behavioral therapies may be effective in diabetes management and associated mental problems [20,21]. Unlike the traditional CBT approaches, Acceptance and Commitment Therapy (ACT) targets the development of acceptance and mindfulness strategies instead of control and elimination strategies toward negative experiences, skills of ‘defusion’ from verbal processes and exhibition of commitment in line with the values selected by the person [22]. Within this context, goal of ACT is to develop the wellness model which is defined as psychological flexibility, in persons. Psychological flexibility can be defined as “the ability to contact the present moment fully to make contact with internal experiences-without judgement or defense against them-and ability to take action in line with the values” [23]. Psychological flexibility model has been associated with many clinical conditions as well as diabetes related chronic fatigue and quality of life [24]. There are studies showing positive effect of ACT interventions, based on psychological flexibility, on stress accompanying to diabetes in children and young persons [25], depression, psychological well-being and feeling of guilt [26].

Given the above mentioned data, importance of acceptance based process in coping with diabetes and associated psychological problems becomes apparent. Education and therapy programs aimed to improve psychological flexibility related skills can be considered as effective application candidates supporting the treatment. As the first example of such intervention; in a randomized controlled trial by Gregg et al. [27] on patients with type-2 diabetes, ACT+education group was compared with education alone group in terms of self-care, HbA1c levels, diabetes focused self-management, and psychological flexibility skills. Acceptance and Action Diabetes Questionnaire was developed in this study in order to evaluate psychological flexibility model that founds significant iimprovements in ACT+educational group in terms of self-care and HbA1c levels. High internal consistency (Cronbach’s alpha 0.94) was found in the analysis carried out. However, as far as we know any statistical study for construct validity of the questionnaire has not yet been conducted until the time of the present study.

The aim of this study is to assess the validity and reliability of the Acceptance and Action Diabetes Questionnaire (AADQ), and to investigate whether this scale is an appropriate measurement tool for evaluation of psychological flexibility levels in a sample of patients with diabetes in Turkey. By this way, a valid evaluation tool can be provided for acceptance and psychological flexibility based interventions and studies to be conducted in patients with diabetes in Turkey.

Ethics committee approval for the study was received from Ankara Numune Training & Research Hospital (E-14-163/2015). Permission for translation of the scale to Turkish was obtained from the authors who have developed the original questionnaire, and the translation process was carried out by three researchers who have a good command of English language and have been specialized in the relevant field. Translation-retranslation process was performed by a professional translator who had no interest in psychopathology related issues and knew two languages. After the re-translation, original and re-translated versions of AADQ were compared and the final version was created.

The participants were asked to fill out Turkish forms of the Beck Depression Inventory (BDI), Problem Areas in Diabetes Questionnaire (PAID), State-Trait Anxiety Inventory (STAI-I and STAI-II), Audit of Diabetes-Dependent Quality of Life (ADDQoL) and Turkish version of Acceptance and Action Diabetes Questionnaire (TAADQ). The TAADQ were applied to nineteen participants again two weeks after the first measurement for test retest analysis (Supplementary Materials in the online-only Data Supplement).

AADQ which is a scale enabling evaluation of psychological flexibility levels of diabetic patients, allows assessment of the attitudes of diabetics toward the disease and associated problems both in clinical practice and studies. In this study, we investigated whether the Turkish form of scale is a valid and reliable assessment tool in Turkish population.

When calculating sample size for the analysis of construct validity of the scale, it is generally recommended that the number of persons should be higher than the number of variables, and should be at least 5 to 10 for each variable [42]. For this reason, at the beginning of the study we targeted to include 110 persons based on the 11-item AADQ, ultimately reached to 105 persons, thought that the above mentioned condition was met. In the confirmatory factor analysis carried out in order to test TAADQ has not a single-factor construct as its original version; it was found that model fit values of the 11-item version were not at an acceptable level, following respective removal of the 1st and 5th items from the scale, 9-item 3rd method obtained was found to have acceptable values.

CFA was then applied to test whether AADQ has a single-factor construct. CFA also allows to detect measurement errors. Errors showing high correlations may include methodological reasons, use of similar words, overlapping of the contents, and difficulty in understanding, etc [37]. In our study, according to the results of CFA, TAADQ was found to have a single-factor construct like its original version. A high correlation was found between the 8th and 9th items in terms of the measurement errors. The use of similar words such as “I avoid thinking of...” and emphasis on avoidance attitude in both items may explain the high correlation level determined and effect of the method.

Reliability of a scale can be tested through test retest analysis, internal consistency coefficient, and item total score correlation analysis [43]. Analyses (test retest correlation, item-total correlation with 2-week interval, cronbach alpha coefficient, and item-total scale point analyses) were performed on 9-item version. Looking test retest correlations of AADQ, correlation coefficients of each item were found to be between 0.47 and 0.86, which indicated that the scale items did not show significant change over time. Cronbach alpha coefficient, and itemtotal scale point analyses were used to assess the internal consistency [42]. In our study, alpha coefficient of the AADQ was found as 0.83, and Schmitt et al. [17] who studied the original scale’s reliability and validity found similarly alpha value as 0.82. Accordingly, it can be said that the internal consistency is high both in original and Turkish forms of the scale. In item-total scale point analysis which is performed to test reliability of AADQ, it is reported that the coefficients should be higher than 0.30, and the items that do not comply with this rule may be removed from the scale [44]. Looking to our results; the coefficients of item-total scale point analysis varied between 0.27 and 0.75, and we found unnecessary to remove any item.

Concurrent validity is used to evaluate whether the measurement tool measure the expected results. In the present study, significant correlations between severity of depressive symptoms, quality of life levels, and diabetes related emotional stress and AADQ, support concurrent validity of the scale. Because there are studies reporting that avoidance from negative inner life events causes more stress [45], and is associated with higher depression levels and lower quality of life [46,47]. In order to test convergent validity of the scale; correlation between AADQ and STAI-II scale which was developed to measure the levels of trait anxiety and worry was analyzed, and a significant correlation was observed. The correlation between these two scales may not expected to be much strong because ACT’s psychological inflexibility model [23] included other psychopathological processes (eg. experiential avoidance, cognitive fusion, poor self-perception, decreased contact with the values etc.) in addition to anxiety. This findings of our study may be considered as a result which could be an evidence for convergent validity of AADQ.

For divergent validity of AADQ; its correlation with STAI-I scale, developed to assess state anxiety, was assessed. Because AADQ was designed to measure avoidance attitudes rather than situtational signs, it is not expected to be affected by cross-sectional symptoms. As an evidence for the Divergent validity of AADQ, we did not find a significant correlation between the two scales.

Our study has some limitations. Relatively small sample size may have an impact on the study results. Also we had no sufficient rates of the patients with type-1 and type-2 diabetes for creating two separate groups, and we can not assess possible differences in psychological flexibility between two diabetes types. The complications of diabetes may lead to more severe depressive symptoms and anxiety and less psychological flexibility but in this study the patients with diabetes were not classified according to complications of diabetes; this can be another limitation. Further comprehensive studies with a larger sample size or two type-1 and type-2 diabetes group or classified according to complications of diabetes should be conducted for this purpose.

In conclusion; our study demonstrates that the TAADQ has sufficient psychometric properties. Measurement of diabetes related psychological flexibility/inflexibility may be useful in development and implementation of interventions for increasing quality of life in diabetic persons in Turkey. So TAADQ will be a powerfull tool in assessing psychological flexibility in diabetes patients.