ABSTRACT
Objective:
The aim of this study was to evaluate the validity and reliability of the Turkish translation of the Parent Diabetes Distress Scale (PDDS).
Methods:
The PDDS is a 5-point Likert-type scale with 20 items. After obtaining permission from the scale developers, the study commenced. First, a systematic adaptation of the scale into the Turkish language was performed including translation, expert panel review, back translation, and pilot study. Test-retest was applied to 35 participants. After these procedures, data collection was undertaken using the adapted PDDS and a demographic data collection form. The collected data were analyzed for reliability, including stability of the scale with test-retest and internal consistency of the scale (Cronbach’s α), and validity including construct validity of the scale with confirmatory factor analysis (CFA).
Conclusion:
The Turkish version of the PDDS fulfilled the validity and reliability tests at an acceptable level.
Results:
The parents of 210 teenagers, aged >11 and <18 years, who had been diagnosed with type 1 diabetes mellitus for at least one year were included. Of these parents, 71.9% (n=151) were mothers and 53.3% (n=112) of the children were girls. The Cronbach’s α value was 0.906. The results of the CFA were χ2/df=4.406, p<0.001, comparative fit test 0.704, and goodness of fit tests 0.749. The mean total PDDS score was 2.2±0.7. These results indicate that scores of 1.6 points or less was evaluated as “little or no distress” 1.7-2.4 as “moderate distress,” and >2.4 points as “high distress”. This showed that the majority of the parents in the study experienced moderate or severe diabetes-related distress.
What is already known on this topic?
Families of adolescents with type 1 diabetes mellitus (T1DM) experience many difficulties in different areas related to diabetes.
What this study adds?
The adaptation and validation of the Turkish version of the Parent Diabetes Distress Scale (PDDS) is reliable and valid. The Turkish PDDS may be used to assess parental distress when their adolescent children are diagnosed or living with T1DM.
Introduction
Type 1 diabetes mellitus (T1DM) is a common chronic disease caused by pancreatic β-cell damage in children and adolescents (1,2). The 10th edition of the International Diabetes Federation Atlas estimated that 1,211,900 children and adolescents under 20 years of age have T1DM worldwide (3). The incidence of childhood-onset T1DM is reported to be increasing in many countries. Although the rate of increase varies geographically, the overall global annual increase is estimated to be around 3% (4).
T1DM burdens the whole family of affected children in many ways and is thus a cause for clinical concern (5). Although it is generally accepted that childhood diabetes affects all family members, few studies have focused specifically on parents (5,6,7). Furthermore, relatively few studies have examined the relationships between parental stress and family demographic factors. The young age of the sick adolescent, long-term illness, low socioeconomic level, and being a single parent cause higher levels of stress in parents (5,8,9). It has been reported that in adolescents with poor glycemic control and diabetes self-management, family conflict and parental distress tend to be at a higher level (8). Increased parental emotional distress has been associated with more parental depressive symptoms, lower quality of life, and more family stress (9,10,11).
A number of different scales related to the stress caused by type 1 and type 2 diabetes have been published (12,13). However, some of these scales only measure the distress caused by diabetes (14). There are few scales evaluating diabetes-related stress in parents of children with T1DM (12,14,15). The scale developed by Katz et al. (12), on the other hand, measures the familial effect of T1DM in children. Adolescence is a particularly risky period for families in terms of diabetes management because during adolescence family conflicts tend to be increased (16). Adolescents may experience more difficulties in self-management of their chronic diabetes (5). Therefore, in 2016, Hessler et al. (17) developed the Parent Diabetes Distress Scale (PDDS), a 5-point Likert-type scale with 20 items. This scale measures the effect of adolescent T1DM on the family and has been widely used. To the best of our knowledge, however, there is no Turkish version of the PDDS. Therefore, the aim of this study was to test the factor structure, reliability and validity using the PDDS translated into Turkish.
Methods
Design and Setting
This study used a methodological design. The survey was conducted between October and December 2021.
Participants
The sample of the study consisted of the parents of adolescents over the age of 11 and under the age of 18 who attended Sivas Cumhuriyet University Faculty of Medicine, Pediatric Endocrinology outpatient clinic and had been diagnosed with T1DM for at least one year. As the scale contained 20 items and it is recommended that a sample of 5-10 times the number of items should be reached in cross-cultural scale adaptation (18), the required sample size was calculated as at least 200 participants (20 items x 10). The targeted sample was achieved by administering the research questionnaire to 210 people. It was administered by one of the researchers using the face-to-face interview method. The participants were informed about the study beforehand, and their consent was obtained. Either the mother or father was included in the study. Other relatives were excluded. Parents who were illiterate were also excluded.
Process
The participants answer the items in the PDDS with responses ranging from Not at all to A great deal, according to how they felt about the scale items in the last month. There are no negatively scored items in the scale. To score the scale, the participants’ responses to the items are summed and divided by the number of items in the scale. The scores that can be obtained from the scale range from 0 to 5. High scores are associated with increased stress levels. Hessler et al. (17) defined four sub-dimensions of the scale (personal distress, teen management distress, parent/teen relationship distress, and healthcare team distress). In the original study of Hessler et al. (17), the Cronbach’s a value for the whole scale was 0.94.
In this study, the Turkish version of the PDDS and a data form created by the researchers were administered to the participants. The data form gathered information about the sociodemographic characteristics of the parents, the demographic characteristics of the teenagers, and the characteristics of their diabetes. Permission to use the PDDS was obtained by mail ([email protected]) from Dr. Polonsky, one of the original authors. The adaptation stages of the scale were conducted in line with World Health Organization intercultural adaptation guidelines, as well as recommendations in the literature review (#*#fig1|#*#). The scale was translated into Turkish by two translators. Then the authors committee decided on a common translation. The translated scale was sent to experts and assessed and the content validity index (CVI) was calculated. The Turkish scale was back-translated into English, and after translation was completed, a second confirmation was received from Dr. Polonsky by e-mail. The Turkish version of the scale was administered to 10 people as a pilot study. Then 30 people were tested-retested on the final version of the scale. Correlation analysis between the test-retest was examined. Afterwards, the data collection form and PDDS were administered to 210 participants for the main study. #*#fig1|#*# shows the intercultural adaptation stages applied.
Statistical Analysis
Statistical Package for the Social Sciences (SPSS) for Windows, version 25 and IBM SPSS Amos 20 were used for statistical analysis (IBM Inc., Armonk, NY, USA). Normality analysis was performed with the Shapiro-Wilk test for numerical values. Descriptive statistical analyses were calculated for sociodemographic data and items of the scale. The Student’s t-test was used to compare normally distributed numerical data between two categorical variables. The factors affecting the PDDS score were evaluated with a multiple linear regression model analysis (enter method). The presence of multicollinearity among the independent variables was tested with the variance inflation factor (VIF) value. Within the scope of the reliability analysis of the scale, the stability of the scale was determined by test-retest, and the internal consistency of the scale was evaluated with Cronbach’s a. Content validity was evaluated with the Davis technique to test the validity of the scale (19). Construct validity was tested with confirmatory factor analysis (CFA). The suitability of the scale for factor analysis was evaluated with the sphericity method of Kaiser-Meyer-Olkin (KMO) and Bartlett’s test. CFA results were reported with total variance values and factor loads, χ2/df, comparative fit test (CFI), goodness-of-fit test (GFI), and root mean square error (RMSEA) values of approximate approximation. A p value of less than 0.05 was considered to indicate statistical significance, with a 95% confidence interval.
Results
Descriptive Characteristics of the Participants
The parents of 210 teenagers were included in the study. The mean age of their children was 14.0±2.1 years and the mean duration of diabetes diagnosis was 56.7±37.8 months. Eighty (38.1%) of the parents had completed primary school. The demographic characteristics of the participants are shown in Table 1. The children of 84 (40.0%) of the participants experienced severe hypoglycemia (<50 mg/dL) every month. The median monthly frequency of those who had hypoglycemia was 2 (minimum: 1 - maximum: 6; interquartile range: 2.0).
Validity and Reliability Analyses of the PDDS
Content validity analysis was evaluated with the Davis technique to test the validity of the scale. With this technique a CVI >0.80 will indicate content validity (18). The CVI for all items of the Turkish version of the PDDS was above 0.80 (Table 4).
The Cronbach’s α value calculated within the scope of the internal consistency analysis of the scale was 0.906. The item-total correlation coefficients ranged from 0.208 to 0.700. There was a single value below 0.30, but since there was no change in the Cronbach a value when this item was removed, it was decided to retain the item in the scale. The reliability of the scale was also measured using the test-retest method. The test-retest correlation coefficients of the items of the scale ranged from 0.535 to 0.953. The test-retest correlation coefficient of the total score was calculated as 0.942 (Table 5). The KMO coefficient, which was calculated within the scope of the construct validity of the scale, was 0.842. Bartlett’s sphericity test results were χ2=2003.303, p<0.001. In the CFA performed on the sub-dimensions defined in the scale, the χ2/df value was calculated as 4.406 (p<0.001), GFI 0.749, CFI 0.704 and RMSEA 0.128. Figure 2 shows the CFA model of the PDDS.
Discussion
Families who have teenagers with T1DM experience many difficulties, which may include their own personal distress, difficult relationships with their teenagers, and teenagers’ problems with managing their diabetes. Identifying and supporting families’ diabetes-related distress plays a key role in disease management. Adolescence is already a challenging period for young people, in which many changes occur physically and psychologically. It is more difficult for teenagers to struggle with a chronic illness, such as T1DM, during this period. A happy and psychologically strong parent can better support their child in this process. The current study describes the adaptation of the PDDS for use in Turkey and the subsequent validity testing in order to provide a tool to assess the stress experienced by parents of adolescents with T1DM.
Content validity analysis of the scale was conducted. The CVI value of all items of the PDDS scale were higher than 0.80, which was interpreted as the Turkish version of the PDDS having content validity (20). The Cronbach’s a, item-total correlation, and test-retest correlation analyses were performed within the scope of reliability analyses of the scale. The Cronbach’s α value calculated for the Turkish version of the PDDS was 0.906 which was comparable with 0.940 reported by Hessler et al. (17) following the original development of the scale. In the literature, a value of 0.60 and above is acceptable for the Cronbach’s α value (21) although some researchers suggest that values above 0.70 are reliable (22). Item-total correlation analysis, one of the reliability analyses, explains the relationship between the scores obtained from the test items and the total score of the test. If this value is 0.30 and above, it may be interpreted as the discrimination rate of the items is high (18). Only one of the 20 items of the Turkish version of the PDDS scale had a value below 0.30, but when this item was removed it did not change the Cronbach’s α value so it was decided to retain the item in the scale. The test-retest method is the reliability calculation applied to determine the stability of a scale. In this assessment, the time between testing and retesting is important. Deciding the delay should be done according to the feature measured. Generally, it is stated that a period of 2-3 or 4-6 weeks is sufficient (23). In the present study, the test was administered again to 30 of the original 35 participants three weeks after first completion. The test-retest reliability for the present study was 0.942 and there was a strong and significant positive correlation between the two sessions.
In the present study, KMO and Bartlett tests were applied to determine the adequacy of the sample and whether the data were suitable for factor analysis. The present study found a KMO value of 0.842 while Bartlett’s sphericity test results were significant at χ2=2003.303 and p<0.001. A KMO value greater than 0.60 and a significant Bartlett test indicate that the data are suitable for factor analysis (18). In the CFA performed on the sub-dimensions defined in the scale, χ2/df, GFI, CFI, and RMSEA values were used to evaluate the fit index. A χ2/df, value of ≤3 is a very good indicator of model fit. GFI and CFI values of 0.90 and above indicate adequate good fit, while values between 0.80 and 0.90 indicate that the structure is suitable for a good fit. Regarding the RMSEA value, while some researchers accept <0.06 as being good, others accept 0.07 as the threshold value (24,25,26). In our study, the CFA, performed in accordance with the study of Hessler et al. (17), showed that the fit indexes of the sub-dimensions were low. For this reason, it was concluded that the scale did not show any sub-dimensions in the analyses conducted in our sample and it should therefore be evaluated using the total score.
The results of the present study in more than 200 parents of Turkish adolescents with T1DM found that diabetes-related parental stress was common, which is similar to previous studies (8,11) and the findings of the original scale. The PDDS scale mean score of the parents was 2.2±0.7 which falls towards the upper end of the moderate distress category. This suggests that most of the study participants experienced moderate or severe DDS. Furthermore, earlier studies reported that mothers experience more stress than fathers when caring for a child with diabetes (27,28,29). This finding was replicated with the Turkish version of the PDDS. Common stressors include food management, family conflict related to diabetes, injecting insulin, and monitoring blood sugar (30).
There may be differences between the sexes in terms of adaptation to chronic disease due to hormonal and metabolic changes during adolescence. While it is more difficult for boys to adapt to diabetes during childhood, it is more difficult for girls to adapt during adolescence. This situation also makes a difference in the impact of the disease on their families. Different results have been found in the literature on this issue (31,32). In the study of Hessler et al. (17), it was reported that parents of sons experienced more diabetes-related stress. In another study, it was found that families of adolescent girls experienced more stress (33). The present study found no significant patient gender effect and the scale score.
In the study of Hessler et al. (17), it was reported that parents of sons experienced more diabetes-related stress. Hessler et al. (17) reported no correlation between the duration of diabetes diagnosis and stress and this was also found in the present study. Many studies have reported a relationship between degree of glycemic control and diabetes-related stress (34,35,36). In the present study, there was no relationship between stress level and glycated hemoglobin (HbA1c) level, but a positive correlation was found with increasing frequency of severe hypoglycemia. Hessler et al. (17) found a positive correlation between both frequency of hypoglycemia and HbA1c level and stress. In a study among Lithuanian youth, parents of young people with good diabetes control had lower stress levels (33).
Parents of children with T1DM experience fear and stress during insulin injection and glucose-testing procedures. Previous studies reported that 13.6% of mothers experienced needle phobia and related stress following the diagnosis of diabetes (37,38). In the present study, higher levels of stress were reported by parents of adolescents who used an insulin pen for injections compared with those who used pumps. This was similar to the original report of Hessler et al. (17) but in a later study by Polonsky et al. (38) research, the opposite was the case.
Studies have shown that the most frequent conflicts between families and young people were around adolescents’ blood sugar control and adherence to their diets (39,40). In the present study, parents of adolescents who had difficulties in adapting to their diabetes regimen had increased stress. Frey et al. (41) found that low income level was associated with difficulties in coping with the disease and also with more stress in mothers. Mothers had higher levels of diabetes-related stress than fathers in the present study, in line with earlier reports.
Approximately 20-30% of parents of children with T1DM experience clinically significant depressive symptoms and anxiety (42). These symptoms are mostly related to parents’ involvement in their child’s diabetes management tasks (43). In addition, parents’ fear of hypoglycemia and adolescents’ poor glycemic control also increased parents’ stress (5,44). In the present study, stress increased in the families of adolescents who had severe hypoglycemia in the month preceding completion of the Turkish version of the PDDS, or if their children had been hospitalized in the previous year, or had an additional chronic disease. These findings are again in line with previously published reports.
Study Limitations
The study data were collected from a single center in the central part of the country. Statistical power was limited. Therefore, the generalization of results by country are likely to be affected. However, the present study also has several strengths. First of all, to the best of our knowledge, this is the first study conducted in Turkey in which this scale was adapted and used. Therefore, this study indicates that the Turkish version of the PDDS may be used in both clinical and research settings, and the validity and reliability of the Turkish version are at a sufficient level. In addition, cross-cultural comparative studies can be carried out using this scale.
Conclusion
Diabetes-related stress is a common and important problem for parents of adolescents with T1DM. This stress is associated with sociodemographic characteristics of families and adolescents’ diabetes management. This study examined the psychometric properties of the Turkish version of the PDDS and the results showed that the Turkish version of the scale was a valid and reliable measurement tool.