FormalPara Key Summary Points

Why carry out this study?

Glucose monitoring-related problems affect the social and psychological distress of patients with diabetes, and this distress leads to low compliance; consequently, it is important to be able to comprehensively assess distress due to glucose monitoring.

To measure social and psychological distress from glucose monitoring, we developed and validated the distress of self-glucose monitoring (DSGM) scale in patients aged > 10 years, including children, adolescents, and adults with diabetes.

What was learned from this study?

The DSGM tool has adequate validity and reliability as instrument for measuring the social and psychological distress of patients with diabetes, including children, adolescents, and adults.

The DSGM tool is recommended to healthcare professionals as a means to communicate with patients over the monitoring and management of glucose levels in relation to their DSGM score.

Introduction

Diabetes mellitus is a chronic condition that affects approximately 422 million people worldwide, which means that one out of 11 people worldwide have this condition [1,2,3]. When inadequately managed, it can lead to serious complications, including nephropathy, retinopathy, neuropathy, cardiovascular disease, and death [4, 5]. Self-monitoring of blood glucose (SMBG) is a key aspect of diabetes self-management associated with improved glycemic control. SMBG consists of monitoring blood glucose levels at home using a small carrying machine called a “glucose meter.” Most glucose meters require a drop of blood from a pin-prick on the patient’s finger [6], and patients with diabetes who use insulin generally need to perform four to ten glucose monitoring sessions per day [7]. However, many patients experience difficulty in following the recommended method of monitoring blood glucose levels [8], and SMBG compliance rates as low as 44% and 24% for persons with type 1 and 2 diabetes, respectively, have been reported [9].

Low compliance can be associated with a high burden and stress due to the pain, stigma, time and equipment required to perform SMBG [10]. To resolve these difficulties associated with glucose monitoring, continuous glucose monitoring (CGM) technology has recently evolved as an approach for both clinicians and patients [11]. Although CGM is associated with a relatively higher patient compliance than SMBG, the problems and complaints reported by patients using CGM have been higher than expected, primarily due to patients experiencing problems with skin irritation and difficulties in keeping the sensor and transmitter attached for successful CGM use [12]. In a study of 30 adolescents with type 1 diabetes, the mean CGM usage time decreased from 149 to 134 h per week from the first to third month after the study [9]. Since glucose monitoring-related problems affect the social and psychological distress experience by the patient and this distress leads to low compliance [13], it is important to comprehensively assess distress due to glucose monitoring.

Based on a literature review, we identified four instruments that have been used to measure distress due to glucose monitoring in patients with diabetes. However, these tools can not comprehensively cover distress due to glucose monitoring [14, 15] and can not be used in patients who did not use CGM [16, 17]. Furthermore, previous tools were developed only for adults. As children and adolescents are relatively more comprehensively affected by distress due to glucose monitoring [18], a tool available to children and adolescents is required. Therefore, the aim of this study was to develop and validate the distress of self-glucose monitoring (DSGM) scale as an instrument to comprehensively assess distress from glucose monitoring in patients aged > 10 years, including children, adolescents, and adults with diabetes.

Methods

Instrument Development

Prior to developing a questionnaire to assess the DSGM scale, an extensive literature review was performed by an expert group of two nurses, two pediatricians, three behavioral scientists, and one librarian. This expert group identified 14 studies involving distress measurement tools for use in patients with diabetes who performed SMBG and CGM to measure blood glucose levels. A total of 334 items were identified in the 14 studies, of which 32 were extracted for the present study. The extracted items were classified into general (5 items), physical (5 items), physical society (10 items), technique (3 items), process (8 items), and finance (1 item) domains. To increase content validity, the expert group reviewed the items and subsequently deleted the general and financial domains in the tool.

Following this qualitative study, 27 items were selected for assessing the DSGM scale, including the physical, psychosocial, and process domains. We then translated the English version into Korean. Respondents were instructed to indicate their responses on a 5-point Likert scale (0 = not at all, 1 = a little, 2 = slightly, 3 = quite a bit, 4 = very much). We then created sentences for the questions and instructions. Since the DSGM should cover both children and adults, we attempted to make all sentences simple and easy to understand.

In this study, three of the 27 items (“how satisfied are you with the type of glucose monitoring?,” “how much do you think glucose monitoring interferes with your daily life?,” and “how stressful is glucose monitoring for you?”) covered overall distress instead of a single domain, and those submitted to content experts in the first round were discarded. The resulting set of 21 items in three domains (physical [n = 1], psychosocial [n = 7], process [n = 9]) was administered to the study participants.

Psychometric Validation

Study Participants

We conducted a cross-sectional study in patients with insulin-treated diabetes aged 10–40 years who visited the Department of Pediatrics, Samsung Medical Center, Seoul, Korea, from April 2021 to September 2021. Participants were eligible to participate in this study if they had a confirmed diagnosis of diabetes, had performed glucose monitoring for at least 3 months, and had no evidence of psychological problems at the time of the survey. This study was approved by the Institutional Review Board (IRB) of Samsung Medical Center, Seoul, Republic of Korea, and informed consent was obtained from all study participants (IRB number:2020-12-156-009 SMC). This study was conducted by single center and was performed in accordance with the Declaration of Helsinki 1964 and its later amendments.

Measures

We used the 20 items of the draft DSGM. In addition to the DSGM scale, we assessed patients’ diabetes-related emotional distress using the Korean version of the Problem Areas in Diabetes (PAID-K) instrument to evaluate concurrent and discriminant validities [19, 20]. The PAID-K scale is a 20-item representative self-reported instrument used to measure diabetes-related emotional distress and covers a range of negative emotional problems in patients with diabetes [21]. This instrument was originally developed in the USA for use in patients with diabetes. The validity and reliability of PAID-K have been well established in the Korean language [22].

We asked the participants about their sociodemographic characteristics, including highest educational level, marital status, monthly family income, and current working status. Clinical information, including years since diagnosis, type of diabetes, and type of glucose monitor, was obtained from the electronic medical records.

After conducting the survey, the research nurses asked the participants if there were any items in the DSGM that were difficult to respond to.

Statistical Analyses

Descriptive statistics were used to report the characteristics of the participants and the mean and standard deviation (SD) of each item on the DSGM. Before the exploratory factor analysis (EFA), we performed Bartlett’s test for sphericity and Kaiser–Meyer–Olkin (KMO) test for sampling adequacy. A significant statistical test in Bartlett’s test of sphericity showed that the correlation matrix was not an identity matrix (rejection of the null hypothesis). The KMO test was conducted to examine the strength of partial correlation between the variables; KMO values closer to 1.0 are considered ideal, whereas values < 0.5 are unacceptable. To evaluate the structural validity, which is part of the construct validity, we performed EFA to determine the underlying structure of the DSGM instrument [23]. A common factor model with alpha factor extraction was used. Alpha extraction generates factors by finding item groupings with maximum internal consistency, which makes this method an appropriate choice for instrument development [24].

After extracting the factor structure, we performed a confirmatory factor analysis (CFA) using the maximum likelihood with missing values to test whether our factor structure fitted the data. Several goodness-of-fit indices were used to evaluate the model fit, including the comparative fit index (CFI), and root mean square error of approximation (RMSEA). A CFI > 0.9 and RMSEA < 0.08 indicate a good fit to the data [25, 26]. Factor loadings in CFA were categorized as low (< 0.30), mid-range (0.30–0.59), and high (≥ 0.60) [27].

To test the internal consistency of the DSGM instrument, we calculated Cronbach’s alpha for both derivation and validation data. We expected a value > 0.70, which is the standard for defining acceptable reliability of an instrument.

Regarding hypothesis analysis to confirm construct and criterion validity, we assumed the DSGM domain and total scores to be at least moderately correlated with the PAID-K total score, life interference, and stress due to injection (0.30 ≤ r ≥ 0.70). On the other hand, we assumed that the physical domain would have a weak correlation with PAID-K total score, life interference, and stress due to injection.

The significance level was p < 0.05 (two-sided), and all statistical analyses were performed using Stata version 16 (StataCorp LP, College Station, TX, USA) and R 4.1.2 ® Foundation for Statistical Computing, Vienna, Austria).

Results

Study Participants

The mean age (SD) of the participants was 20.9 (± 6.7) years; 47.1% were male. In total, 89.7% had type 1 diabetes, and 60.3% used CGM for glucose monitoring. Patients with type 1 and 2 diabetes were typically diagnosed with diabetes at ages 9.6 (± 4.7) and 13.5 (± 2.1) years, respectively. The mean level of glycated hemoglobin for the last 3 months at the time of conducting this study was 7.08 ± 1.08%, and diabetes-related complications were observed in 42% of the participants (Table 1).

Table 1 Characteristics of study participants (N = 136)
Full size table

All participants responded to all items without exception and reported that there had been no item that was difficult to respond to. Even children aged < 10 years responded to all the item without missing any and reported that they understood the meaning of each item.

The highest distress item was “bothered to carry the materials for glucose monitoring” followed by “have skin-related problems due to glucose monitoring,” and “bothered to prepare for glucose monitoring” (Fig. 1). The CGM group was more likely to experience stress due to “feeling annoyed due to alarm from a glucose monitor” than the non-CGM group.

Fig. 1
figure 1

Confirmatory factor analysis of distress of glucose self-monitoring. CGM Continuous glucose monitoring group, non-CGM group not performing CGM

Full size image

Structural Validity: Exploratory and Confirmatory Factor Analysis

All 24 items satisfied Bartlett’s test for sphericity (p < 0.01) and the KMO test for sampling adequacy (p = 0.91). EFA indicated a three-factor solution with an eigen value > 1.0; thus, it was initially designed for a three-factor solution. In addition, three items (“concerned with cost of paying for glucose monitoring,” “I do not want to monitor glucose due to stress,” and “feeling annoyed due to the difference in glucose levels between the measured and the real one”) were excluded and did not load significantly on an interpretable factor solution. Subsequently, “feeling annoyed due to error from a glucose monitor” (r = 0.32) and “feeling annoyed due to alarm from a glucose monitor” (r = 0.36) were excluded based on relatively low factor loading values; however, we did not exclude these items based on the expert opinion that they were important items to measure process distress due to glucose monitoring (Table 2). Ultimately, 21 items were included in the DSGM instrument. In addition, the items “it is difficult to have physical and outdoor activities due to glucose monitoring,” “it is difficult to participate in all activities during the camp or trip due to glucose monitoring,” “it is difficult to do hobbies in free time due to glucose monitoring,” “feeling uncomfortable because people might stare at me,” “feeling bothered to prepare glucose monitoring,” and “feeling bothered to rotate the monitoring site” showed moderate correlations in the physical, psychosocial, and process domains. We assigned these items to a domain that had a relatively higher correlation than that of the other domains (Table 2).

Table 2 Factor loadings from the exploratory factor analysis and reliability of the distress of glucose self-monitoring dimensions
Full size table

Further examination of the factor structure of the 21-item DSGM scale was evaluated using CFA, which revealed high loadings (0.38–0.89) in general. In the overall group, the fit indices for this model were good: CFI = 0.938 and RMSEA = 0.060 (Fig. 2a). In the CGM group, the fit indices for this model were good: CFI = 0.927 and RMSEA = 0.068 (Fig. 2b). In the non-CGM group, the fit indices for this model were good: CFI = 0.776 and RMSEA = 0.122 (Fig. 2c).

Fig. 2
figure 2

Distribution of distress of glucose self-monitoring

Full size image

Internal Consistency

Cronbach’s alpha coefficients of the total DSGM were 0.92, 0.92, and 0.93, in the overall, CGM, and non-CGM groups, respectively. Cronbach’s alpha coefficients of the three subscales ranged from 0.69 to 0.92, indicating satisfactory internal consistency (Table 2).

Construct Validity and Criterion Validity

Regarding correlations among DSGM, PAID-K, life interference, and stress due to glucose monitoring, the total DSGM was moderately correlated with the PAID-K total score (r = 0.59), life interference (r = 0.70), and stress (r = 0.65) due to glucose monitoring. In addition, the psychosocial domain and process in the DSGM also moderately correlated with the PAID-K total score, life interference due to glucose monitoring, and stress due to glucose monitoring. However, the physical domain of the DSGM was only weakly correlated with the PAID-K total score (Table 3).

Table 3 Concordance validity
Full size table

Discussion

In this study, we found that the DSGM scale was a valid and reliable tool to evaluate distress due to glucose monitoring in adults, adolescents, and children with diabetes. The three factors that emerged in this analysis were reflected in three subdomains: physical, psychosocial, and process. The EFA and CFA confirmed the structural validity of the tool in patients with and without CGM. Furthermore, construct and criterion validity were demonstrated by its varying degrees of correlation with the PAID-K total score, life interference, and stress due to glucose monitoring. Taken together, our results provide strong evidence to support the validity and reliability of DSGM as an instrument to measure distress due to glucose monitoring in patients.

All participants completed all of the questions without missing any item. Considering that more than 50% of the study participants were aged < 18 years, we expect the DSGM scale to be a feasible means to evaluate the distress of glucose monitoring regardless of age and illiteracy rate when used in patients aged > 10 years. The EFA and CFA confirmed our hypothesis regarding the underlying structure of DSGM: physical, psychosocial, and process. The themes of the three domains were consistent with previously identified problems related to the distress of glucose monitoring among children and adolescents.

In this study, the highest distress among items was noted for the item “bothered to carry the materials for glucose monitoring” followed by “have skin-related problems due to glucose monitoring” and “bothered to prepare for glucose monitoring”. However, there are no tools that can measure this content. To date, previous tools related to the distress of blood glucose measurement have not been comprehensive because they were selective according to the blood glucose check method used or limited to individuals by type of diabetes. The DSGM scale is a tool that can comprehensively evaluate distress in patients who need to measure their blood glucose level, regardless of the type of diabetes or blood glucose check method [14,15,16,17, 28,29,30,31,32].

In general, the internal consistency of the DSGM scale was high. In contrast, the items “feeling annoyed due to error from a glucose monitor” and “feeling annoyed due to alarm from a glucose monitor” had relatively weaker correlations with the other items in addition to low factor loading values in the EFA. The low correlation in error and alarm items might be due to these issues being more likely to be machine issues, not patient process issues.

In the construct validity for the DSGM, the DSGM total score and the score of each subdomain moderately correlated with the PAID-K total score, life interference, and stress due to glucose monitoring. However, physical distress in the DSGM scale had a weak correlation or no correlation at all with the PAID-K total score, life interference, and stress due to glucose monitoring. Because the PAID-K instrument evaluates only emotional distress due to diabetes, it cannot address physical distress. However, because physical problems from glucose monitoring also occur frequently, the DSGM scale will be more useful in evaluating these problems [33].

This study has a number of limitations. First, the DSGM scale was developed using Korean patients with diabetes; therefore, additional validation studies are necessary in other patient populations. Nonetheless, previous studies have shown that patients with diabetes experience similar problems related to glucose monitoring [33]. Second, we did not perform a test–retest analysis. Our study design did not include an evaluation to confirm clinical stability and clinical outcomes such as glycemic control and adherence to glucose monitoring. Further study using a prospective cohort design with larger sample size should be necessary to evaluate the association between DSGM and clinical outcomes. Third, in an effort to minimize respondents’ distress, the convergent validity of the physical and process distress subscales of the DSGM was not examined using separate instruments.

Conclusion

In conclusion, our study confirmed that the DSGM tool is a valid and reliable scale for measuring distress of glucose monitoring, including the physical, psychosocial, and process aspects. It can be used to assess the DSGM in patients with diabetes, including children and adolescents aged > 10 years. The DSGM tool is recommended to healthcare professionals as a means to communicate with patients regarding the monitoring and managing of glucose levels in relation to their DSGM scale.