Association between glycemic control and patient-reported outcomes in adults with type 1 diabetes in Japan: the SAGE study subanalysis

Abstract

Aims/introduction

Psychosocial aspects and the quality of life (QOL) of individuals with diabetes are important for achieving glycemic control and treatment goals. Here, we describe patient-reported outcomes (PROs) of Japanese adults with type 1 diabetes (T1D) and evaluate the association thereof with glycemic control.

Materials and methods

This subanalysis of a subgroup of 528 Japanese participants in the SAGE study of adults with T1D used data on glycosylated hemoglobin (HbA1c) and PRO scores [Hypoglycemia Fear Survey-II (HFS-II), Problem Areas In Diabetes (PAID), Insulin Treatment Satisfaction Questionnaire (ITSQ), and Audit of Diabetes-Dependent QOL (ADDQoL)] and summarized the score by the predefined age groups (26–44-years: n = 208, 45–64-years: n = 217, and ≥ 65-years: n = 103). The association between PROs, achieving HbA1c < 7.0%, and individualized targets was explored using multivariate logistic regression analysis.

Results

The HFS-II and PAID scores were lower, and the ITSQ score was higher in the ≥ 65-years group than in the younger groups with a linear trend of better scores with increasing age (P for trend < 0.05). ADDQoL scores were similar across the age groups, and present QOL (ADDQoL subscale) tended to improve with age (P for trend < 0.05). Achieving HbA1c < 7.0% and individualized targets were associated with satisfaction with insulin treatment regarding glycemic control.

Conclusion

In Japanese adults with T1D, the impact on psychosocial aspects and QOL varied across age groups, with a trend of improving scores with age, potentially in relation to the less stringent glycemic control targets adopted in older individuals. Glycemic control was significantly associated with treatment satisfaction.

Supplementary Information

The online version contains supplementary material available at 10.1007/s13340-023-00668-4.

Introduction

Type 1 diabetes mellitus (T1D) poses clinical and psychosocial burdens that affect the health-related quality of life (QOL). Diabetes treatment goals are to “prevent the development and progression of diabetic complications, and maintain QOL and life expectancy at a level comparable to those in individuals without diabetes” [1]. To prevent complications, glycemic control using insulin is vital in individuals with T1D. However, complex, lifelong, and day-to-day self-management challenges individuals with T1D in various aspects of their lives, placing a heavy burden on them. Since self-management is expected to be influenced by individuals’ perceptions, not only laboratory values, such as glycosylated hemoglobin (HbA1c), but also their perceptions, that is, patient-reported outcomes (PRO), play an important role. Emotional distress [2] and/or dissatisfaction with diabetes management [3] may affect self-management behaviors, possibly leading to a poor glycemic control. Excessive fear of hypoglycemia has also been reported to compromise self-management behaviors [4] and constitute a barrier to good glycemic control. The association between glycemic control and QOL has been previously reported [5–7], while others have reported the absence of such an association [8, 9]. However, to date, few studies evaluated QOL and its associations, covering a breadth of PROs, particularly in adults with T1D in Japan.

To describe the glycemic control and PROs in individuals with T1D across adulthood, the Study of Adults’ Glycemia in T1D (SAGE) was conducted in adults with T1D in 17 countries in 2018 [10]. Results of glycemic control and treatment management in a global cohort from 17 countries have been previously reported [10]. Additionally, we conducted a country-level subanalysis and reported the real-world glycemic control status and treatment management in Japan [11]. Our report indicated that overall glycemic control was suboptimal (25.8% of the participants achieved HbA1c < 7.0%, and 27.3% achieved individualized targets) but with a low incidence of severe hypoglycemia (5.5%), suggesting careful management implemented in Japanese daily practice. The use of new devices, including insulin pumps and continuous glucose monitoring (CGM), was less common during the original data collection period (23.5% and 33.9% of the participants used insulin pumps and CGM, respectively), which reflects the treatment practice before use of these devices became widespread.

Recently, a secondary analysis of the SAGE global population to evaluate the relationship between glycemic control and PROs was reported [12]. The results demonstrated that among adults with T1D, the impact of diabetes on fear of hypoglycemia, emotional distress, and QOL was relatively low, and they were relatively or highly satisfied with insulin treatment. Better glycemic control was associated with higher treatment satisfaction and lower diabetes-related emotional distress.

PROs closely reflects cultural and societal norms as well as the environment in which an individual lives. QOL is defined as the perception of people of their lives in relation to their culture and value systems ones lives in [13]. Therefore, PROs require evaluation in a specific cultural, social, and clinical context. Assessing PROs and analyzing the associated factors using the data obtained under the current Japanese daily practice would add valuable insights into the needs of individuals with T1D and how to approach them to improve diabetes management. To this end, we conducted a subanalysis of the SAGE study participants in Japan, with a particular focus on PRO data. In this study, we aimed to 1) describe the psychosocial aspects and QOL of Japanese adults with T1D using PRO measures; 2) evaluate the association between PROs, glycemic control, and severe hypoglycemia; and 3) identify factors associated with PROs.

Materials and methods

Study design

The SAGE, a multinational, cross-sectional, observational study, was conducted in 17 non-United States countries from January to December, 2018 [10]. The details of the methods and results of glycemic status and PROs based on all participating countries [10, 12] and a subgroup analysis of glycemic control in Japanese participants have been reported previously [11].

Briefly, we consecutively included individuals who (1) were aged ≥ 26 years, (2) had T1D for ≥ 1 year, (3) were treated with insulin, (4) had available HbA1c values, and (5) provided written informed consent. Data were collected during a single visit, during which the participants responded to the PRO questionnaires described in the following section. Demographic and clinical data, including HbA1c (obtained within 30 days preceding the study visit or within 7 days after the study visit), hypoglycemic events, treatments, and complications, were collected from patient medical records and interviews and entered into an electronic case report form. This manuscript focused on the PRO data of the subgroup of 528 Japanese participants from 21 participating centers (26–44-year-old group: n = 208, 45‒64-year-old group: n = 217, and ≥ 65-year-old group: n = 103).

The study was conducted in accordance with local regulatory requirements, the Declaration of Helsinki, and the guidelines for Good Pharmacoepidemiological Practice.

Outcomes

The primary outcome of the SAGE study was the percentage of participants achieving the general glycemic goal of HbA1c < 7.0%. Secondary outcomes of interest included the achievement of individualized HbA1c targets assigned by treating physicians in daily practice, frequency of severe hypoglycemia during the past 6 months, and PROs. PROs were assessed using validated Japanese versions (or, if unavailable, the Japanese translation) of the Hypoglycemia Fear Survey-II (HFS-II) [14], Problem Areas In Diabetes (PAID) [15], Insulin Treatment Satisfaction Questionnaire (ITSQ) [16], and Audit of Diabetes-Dependent QOL (ADDQoL) [17].

PRO measures

The HFS-II assesses respondents’ fear of hypoglycemia in terms of 15 items on behaviors taken to avoid hypoglycemia (behavior subscale; HFS-B) and 18 items on worries and concerns about hypoglycemia (worry subscale; HFS-W) [14]. Each item is rated on a 5-point scale (0 = never to 4 = almost always). The HFS-II total scores sum the scores of the HFS-B (score 0–60, the higher the score, the higher the likelihood of taking action to avoid hypoglycemia) and HFS-W (score 0–72, the higher the score, the higher the worry about hypoglycemia), yielding a range of 0–132, with a higher score indicating a higher degree of fear of hypoglycemia.

The PAID assesses emotional distress related to diabetes, with 20 items describing the psychosocial problems that individuals with diabetes may face [18]. Respondents rate the extent to which each item affected them on a 5-point Likert scale (0 = not a problem to 4 = serious problem). The sum of each item was multiplied by 1.25 to yield a total score ranging from 0 to 100; a higher score indicates a higher level of emotional distress.

The ITSQ assesses satisfaction with the insulin treatment [16]. The ITSQ includes 22 questions that consist of five subscales (inconvenience of regimen, lifestyle flexibility, glycemic control, hypoglycemic control, and insulin delivery device satisfaction) rated on a 7-point Likert scale (1–7). Scores were transformed to a 0–100 scale (100 × [(7-scale mean)/6]), with a higher score indicating higher satisfaction with the insulin treatment.

The ADDQoL assesses the diabetes-specific QOL [17]. ADDQoL consists of two overview questions assessing the present QOL on a 7-point Likert scale (− 3: extremely bad to 3: excellent) (overview item 1) and potential QOL given the respondent did not have diabetes on a 5-point Likert scale (− 3: very much better to 1: worse) (overview item 2) and a 19-item domain of life assessing the impact of diabetes on a specific domain of life on a 5-point Likert scale (− 3 to + 1) and importance of each domain on a 4-point Likert scale (0 to + 3). For each domain, a weighted impact score (− 9 to 3) was calculated, by multiplying the impact score by the importance score. For the total ADDQoL score, the average weighted impact score was calculated as the sum of the weighted impact scores of each domain divided by the number of applicable domains, yielding a score ranging from − 9 (maximum negative impact of diabetes) to 3 (maximum positive impact of diabetes).

Factors explored in relationship with PRO

We explored the following potential factors that may be associated with PRO: duration of diabetes (< 10/ ≥ 10 years), at least one microvascular diabetes complication (yes/no), at least one macrovascular comorbidity (yes/no), concomitant ongoing therapies for diabetes other than insulin therapy (yes/no), frequency of titration of short-acting insulin per month (< 4/ ≥ 4 times, the threshold by which participants were divided approximately into half), use of insulin pump (yes/no), use of CGM (yes/no), use of digital applications to manage their diabetes (yes/no), compliance with the recommended diet (yes/no), frequency of ≥ 30-min physical activity per week (no [0 days]/ limited [1–3 days]/ sufficient recommended activity [4 days]), and home health aid (yes/no).

Statistical analysis

PRO scores were summarized as mean (standard deviation [SD]) by age group, and p-values for linear trends across age groups were also calculated. The association between PRO and glycemic control was explored using a multivariable logistic regression model with achievement of the general glycemic goal (HbA1c < 7.0%), individualized target, and incidence of severe hypoglycemia as dependent variables and each PRO score as independent variables, first adjusted by the predefined age group, followed by the interaction between each PRO score and age group. We calculated the odds ratio (OR) and its 95% confidence interval (CI) for a 10-point increase in the HFS-W, PAID, and ITSQ scores and 1-point increase in the ADDQoL score. The model was subsequently adjusted for possible confounding factors. Possible confounders were identified from variables related to participants’ demographics, diabetes complications and comorbidities, diabetes management, and the structure and process of medical care using independent multivariate analyses. The association of these candidate factors with the achievement of the general or individualized goal or the incidence of severe hypoglycemia was assessed using chi-square test. Factors with a p-value < 0.2 were subsequently explored for the relationship between each PRO score by analysis of variance, and factors found to be statistically significant (p-value < 0.05) were identified as potential confounders.

The factors associated with each PRO were explored using a simple regression model that included each factor considered clinically relevant to T1D treatment, as described in the previous section. In addition, stepwise regression analysis was performed to identify factors associated with PRO scores, using an entry criterion of 0.20 and a stay level of 0.05.

A p-value < 0.05 (two-tailed) was considered statistically significant. All analyses used descriptive statistics.

Analyses were performed using SAS version 9.4 (SAS Institute Inc., Cary, NC, USA).

Results

Participants’ characteristics

The characteristics of the 528 participants included in this subanalysis have been previously presented [11]. Briefly, mean age (SD) was 50.3 (13.7) years, and the majority were women (64.4%) in the overall analysis and every age group. The mean body mass index (SD) was 23.1 (3.9) kg/m², and three-quarters of the participants were not obese (< 25 kg/m²). The majority of the participants had T1D for ≥ 10 years (65.1%), with a mean duration (SD) of 16.2 (11.2) years, relatively longer in the ≥ 65-year-old group (20.4 [12.2] years) than in the younger groups (26–44-year-old group: 14.3 [9.0] years and 45‒64-year-old group: 15.9 [12.1] years).

PRO

The overall HFS-II total score (mean [SD]) was 26.91 (17.34) (Fig. 1a). By age group, the total score was numerically lower in the ≥ 65-year-old group (24.90 [16.45]) than in the 26–44-year-old and 45‒64-year-old groups (27.73 [17.79] and 27.09 [17.32], respectively), albeit with no statistical significance in a linear trend with age (P for trend = 0.177). The HFS-W score also decreased (p < 0.001), but the HFS-B score slightly increased with increasing age (p = 0.003).

Fig. 1.

PRO scores by age group. a HFS-II. Each score calculated only if more than 75% of the items have responses. Total score: 0 to 132, higher total score reflects greater fear of hypoglycemia. b PAID questionnaire. Total score considered as missing when ≥ 5 items not responded. Total score: 0 to 100, higher score reflects higher emotional distress due to diabetes. c ITSQ. A total score was calculated only when all five subscale scores were not missing; if < 20% of the items in the subscale were missing, the score was calculated by imputing the missing values based on the average of the non-missing items. Otherwise, the subscale was considered missing. Scores were transformed into a 0–100 scale (100 × [(7-scale mean)/6]), with higher scores indicating higher satisfaction with the insulin treatment. d ADDQoL questionnaire. Weighted and total score: − 9 to 3, higher score reflects greater positive impact of diabetes. Overview item 1 score: − 3 (extremely bad) to 3 (excellent). Overview item 2 score: − 3 (very much better) to 1 (worse). HFS-B, Hypoglycemia Fear Survey-II behavior subscale; HFS-W, Hypoglycemia Fear Survey-II worry subscale; PAID, Problem Areas In Diabetes; PRO, patient-reported outcome; SD, standard deviation. P for trend across age groups is presented

The mean total PAID score (mean [SD]) of the participants was 31.57 (20.77) (Fig. 1b). By age group, the total score was slightly lower in the ≥ 65-year-old group (26.48 [22.08]) than in the younger age groups: 26–44-year-old and 45‒64-year-old groups (33.08 [18.92] and 32.55 [21.54], respectively) (p = 0.008).

The overall ITSQ summary score (mean [SD]) was 64.8 (18.4) (Fig. 1c). By age group, the overall summary score was significantly higher in the ≥ 65-year-old group (71.4 [19.3]) than in the 26–44-year-old and 45‒64-year-old groups (62.7 [17.8] and 63.7 [18.0], respectively) with a linear trend of increase with age (p < 0.001) A similar trend of a higher score in the ≥ 65-year-old group than in the other younger groups and a trend of linear increase with age was observed across all the subscales (p < 0.05), except for lifestyle and hypoglycemic control subscales.

The ADDQoL total score (mean [SD]) of the participants was − 2.32 (1.83) (Fig. 1d). For all the age groups, the mean total score was similar (− 2.3) (p = 0.960), and for overview item 1 (present QOL), a trend of linear increase with increased age (p = 0.028) was observed. Overview item 2, reflecting potential QOL without diabetes numerically decreased in the ≥ 65-year-old group (− 1.42 [1.08]) than in the younger age groups: 26–44-year-old and 45‒64-year-old groups (− 1.29 [1.19] and − 1.37 [1.09], respectively), but without statistical significance in a linear trend (p = 0.342).

Association between glycemic control and PRO

Association with the achievement of the general glycemic goal (HbA1c < 7%)

After adjusting for age, a significant association was found between the ITSQ glycemic control subscale and ADDQoL overview item 1 (present QOL) (OR [95% CI] 1.21 [1.09–1.34] and 1.26 [1.05–1.51], respectively) with the achievement of the general glycemic target (Fig. 2a). The interaction between PRO scores and age group was significant only in the HFS-W subscale (p = 0.046). Further, adjusted by age and the potential confounders (Table S1), those two associations still remained significant (OR [95% CI] 1.21 [1.09–1.33] and 1.25 [1.04–1.50], respectively) (Fig. 3a).

Fig. 2.

Relationship between glycemic control and PRO, adjusted by age. a Relationship with achieving the general HbA1c target of < 7.0%. b Relationship with achieving the individualized glycemic target. Relationships among the overall population were adjusted for age. OR (95% CI) for achieving the glycemic target for a 10-point increase in the HFS-W, PAID, and ITSQ scores and for a 1-poinit increase in the ADDQoL score. ADDQoL, Audit of Diabetes Dependent Quality of Life; CI, confidence interval; HbA1c, glycosylated hemoglobin; HFS-W, Hypoglycemia Fear Survey-II worry subscale; ITSQ, Insulin Treatment Satisfaction Questionnaire; OR odds ratio; PAID, Problem Areas In Diabetes; PRO, patient-reported outcome

Fig. 3.

Relationship between glycemic control and PRO, adjusted by confounding factors. a Relationship with achieving the general HbA1c target of < 7.0%. Confounding factors included are summarized in Table S1. b Relationship with achieving the individualized glycemic target. Confounding factors included are summarized in Table S2. OR (95% CI) for achieving the glycemic target for a 10-point increase in the HFS-W, PAID, and ITSQ scores and for a 1-poinit increase in the ADDQoL score. ADDQoL, Audit of Diabetes Dependent Quality of Life; CI, confidence interval; HbA1c, glycosylated hemoglobin; HFS-W, Hypoglycemia Fear Survey-II worry subscale; ITSQ, Insulin Treatment Satisfaction Questionnaire; OR odds ratio; PAID, Problem Areas In Siabetes; PRO, patient-reported outcome

Association with the achievement of individualized HbA1c target

Adjusted by age, a significant association was found for ITSQ glycemic control subscale (OR [95% CI] 1.18 [1.07–1.30]) with the achievement of individualized HbA1c targets (Fig. 2b). The interaction between the PRO scores and age group was significant for the HFS-W and ITSQ hypoglycemic control subscales (p = 0.043 and p = 0.037, respectively). In particular, higher ITSQ hypoglycemic control scores in the 45–64-year-old group were more likely to be associated with target achievement than in other age groups. Furthermore, even after adjusting for age and potential confounders (Table S2), the association remained significant (OR [95% CI] 1.18 [1.07–1.30]) (Fig. 3b).

Association with the incidence of severe hypoglycemia

No significant association between PRO scores and incidence of severe hypoglycemia was observed after adjusting for age (Fig. S1a). The interaction between PRO scores and age group was not significant for any of the PROs. After adjusting for age and potential confounders (Table S3), statistical significance was found (p = 0.048) for the ITSQ hypoglycemic control subscale (OR [95% CI] 0.84 [0.70–1.00]) (Fig. S1b).

Factors associated with PRO

The results of the simple regression analysis are summarized in Table S4. Stepwise regression analysis identified significant associations in the following (Table 1): use of CGM with higher HFS-II score (greater fear for hypoglycemia); a duration of diabetes ≥ 10 years and concomitant ongoing therapies with lower PAID score (less emotional distress) and compliance with diet with higher PAID score (greater emotional distress); a duration of diabetes ≥ 10 years with higher ITSQ score (greater treatment satisfaction); and more frequent titration of short-acting insulin (≥ 4 times per month) and compliance with diet associated with lower ADDQoL score (greater negative impact on QOL).

Table 1.

Stepwise regression analysis for the factors associated with PRO

Factors and variables		Coefficient	(95% CI)	t value	P value
Association with HFS-II
Use of CGM	Yes (vs. No)	8.609	(5.516 to 11.701)	5.468	< 0.0001
Association with PAID
Duration of diabetes	≥ 10 years (vs. < 10 years)	− 4.559	(− 8.353 to − 0.766)	− 2.361	0.0186
Concomitant ongoing therapies for diabetes other than insulin therapy	Yes (vs. No)	− 6.852	(− 12.836 to − 0.868)	− 2.250	0.0249
Compliance with diet	Yes (vs. No)	4.299	(0.096 to 8.502)	2.009	0.0450
Association with ITSQ
Duration of diabetes	≥ 10 years (vs. < 10 years)	4.267	(0.916 to 7.617)	2.502	0.0127
Association with ADDQoL
Frequency of titration of short acting insulin per month	≥ 4 times (vs. < 4 times)	− 0.344	(− 0.669 to − 0.018)	− 2.076	0.0384
Compliance with diet	Yes (vs. No)	− 0.529	(− 0.908 to − 0.149)	− 2.735	0.0065

The following factors were considered: duration of diabetes, at least one microvascular diabetes complication, at least one macrovascular comorbidity, concomitant ongoing therapies for diabetes other than insulin therapy, frequency of titration of short acting insulin per month, use of insulin pump, use of CGM, use of digital applications to manage their diabetes, compliance with diet, frequency of ≥ 30-min physical activity per week, and home health aid. Only selected factors are presented

ADDQoL, Audit of Diabetes-Dependent Quality of Life; CI, confidence interval; HFS-II, Hypoglycemia Fear Survey-II; ITSQ, Insulin Treatment Satisfaction Questionnaire; PAID, Problem Areas In Diabetes; PRO, patient-reported outcome

Discussion

This subanalysis of the Japanese subgroup of the global SAGE study describes the impact of T1D on the psychosocial aspects and QOL of adults with T1D in Japan. This study was based on a large number of Japanese participants from the SAGE study and encompassed a wider range of age groups. The psychosocial aspects and QOL of Japanese adults with T1D were characterized by a relatively low fear of hypoglycemia and emotional distress, moderate-to-high satisfaction with insulin treatment, and a small degree of negative impact on QOL. The overall trend in the PRO score was generally in line with that observed in the SAGE global cohort [12].

Among the five regions of the SAGE global cohort (Asia, Eastern Europe, Western Europe, Latin America, and the Middle East), fear of hypoglycemia was the lowest in Asia [12] and numerically lower in Japan (global: 38.59, Asia: 30.07, and Japan 26.91). In the ≥ 65-year-old group, the oldest group, the score was even lower than in the same age group of the global cohort (global: 39.74, Japan: 24.90). The previously reported low incidence of severe hypoglycemia in Japan [11] suggests that careful management to prevent severe hypoglycemia in Japan may have translated into a lower fear of hypoglycemia than in the global cohort. Within the Japanese subgroup, the fear of hypoglycemia was numerically lower in the ≥ 65-year-old group than in the younger groups. Particularly for older individuals, for whom less stringent glycemic control targets were adopted than for the younger groups, as recommended [19], such careful management might be more evidently reflected in the lowest fear of hypoglycemia.

The ITSQ-total score was also low in Asia in the SAGE global cohort [12] and in Japan (Global: 69.1, Asia: 66.5, and Japan: 64.8). By age group, treatment satisfaction tended to increase with age and was lowest in the youngest age group within the Japanese subgroup and lower than that of its global counterpart (global: 67.7; Japan: 62.7). As the ITSQ inconvenience subscale score was particularly low in this age group, day-to-day diabetes self-management using insulin while managing work, social, and/or family life may be demanding and challenging, resulting in reduced treatment satisfaction. In particular, some young adults may have been in the midst of, or just around, the transition period from pediatric to adult care, possibly struggling to better adjust to the new clinical care environments. These life transitions may have affected their perceptions of treatment.

In addition to the aforementioned fear of hypoglycemia and insulin treatment satisfaction, the total and/or subscale scores of diabetes-related emotional distress and present QOL generally varied by age group and showed a linear trend of better PRO with increasing age (p < 0.05). These results may substantiate the different needs of different age groups of adults with T1D and underscore the importance of care tailored to each individual.

Achievement of general (HbA1c < 7.0%) and individualized HbA1c targets was significantly associated with higher satisfaction with insulin treatment in terms of glycemic control itself, rather than overall satisfaction or any other PROs. Previous studies have also suggested an association between good glycemic control and treatment satisfaction [20, 21]. In addition to treatment satisfaction, emotional distress was also associated with glycemic control in the global cohort, but not in the Japanese subgroup, albeit the similarly low emotional distress [12]. The difference in these associations and potential confounders between the global cohort [12] and Japanese subgroup may also underline the importance of evaluating PRO in each country or region.

Better present QOL, represented by the ADDQoL overview item 1 score, was also indicated as a factor significantly associated with achieving the general glycemic target. While the OR for achieving the general target varied, that for individualized targets was similar across age groups. Considering that individualized glycemic targets varied by and within age groups [11], the relevance of “achieving HbA1c < 7.0%” may also vary and likely reflect on differences in achievement levels and degrees of association with QOL across age groups.

Notably, no significant association between worry about hypoglycemia and achieving general or individualized targets was found in the overall Japanese subgroup, in contrast to the global cohort. However, a significant impact of age on the relationship was indicated, with a pronounced effect in the oldest group, demonstrating that the likelihood of achieving glycemic targets with the higher HFS-W scores decreased with age. Considering that a greater proportion of the ≥ 65-year-old group in Japan had their individualized target set at ≥ 8.0% than the younger and same age group in the global cohort it reflects the general HbA1c target may be limited as a universal treatment goal and potentially becomes challenging to achieve due to fear of hypoglycemia, particularly for the older individuals.

The actual incidence of severe hypoglycemia was associated only with measures directly related to hypoglycemia. Nevertheless, the number of participants who experienced severe hypoglycemia was limited in the Japanese subgroup (29 [5.5%]) [11]. This subanalysis might not have sufficient sample size to conclude an association between PRO and severe hypoglycemia, warranting further consideration.

Although included as a confounder, the availability of glucagon at home was the only factor confounded by the relationship between all PROs, except for HFS-W, and glycemic control. As glucagon administration requires the understanding and assistance of family members or caregivers, support from people around individuals with T1D may be another important factor related to QOL and glycemic control.

Contrary to expectations based on previous reports on the benefits of CGM use to reduce the number of hypoglycemic episodes [22] and lessen the fear of hypoglycemia [23], the present study showed that the use of CGM was the only factor associated with increasing fear of hypoglycemia. SAGE was conducted immediately after the intermittently scanned CGM device was approved in September, 2017 in Japan; this device was often introduced primarily to those experiencing frequent hypoglycemic episodes, which possibly contributed to the observed results. Exploring this association may yield different trends if data is collected after CGM use becomes more common in clinical practice. Among the SAGE participating countries, CGM usage differed by region, ranging from 2.5% in the Middle East to 46.4% in Western Europe [10]. Although the association between CGM and PRO was not analyzed in the global cohort, this association is expected to differ by the CGM introduction status and other factors including health insurance system in each country or region.

Duration of diabetes was identified as a factor inversely associated with the PAID score and positively with ITSQ total score; individuals with T1D for < 10 years reported greater emotional distress and lower treatment satisfaction than those with T1D for ≥ 10 years. In those with limited experience with the treatment and self-management required for T1D, new adjustments and the resulting struggles may affect their perception of insulin treatment. In contrast, extensive experience in insulin and diabetes management may enable those with a long history of T1D to appreciate the advancement of T1D treatments over the course of time, which may have translated to higher treatment satisfaction.

Compliance with recommended diets was identified as another factor associated with emotional distress, as previously suggested [18, 24]. Additionally, titration frequency of short-acting insulin was associated with QOL (ADDQoL total score). Asian individuals, including the Japanese, generally consume more dietary carbohydrates than Western individuals. Meal-related insulin doses are mainly conditioned on carbohydrate intake; the more frequent the titration required, the more burdensome the individual may feel. Furthermore, coping with dietary restrictions adds another layer of complexity to self-management in individuals with T1D. Treatment strategies may require careful consideration of the perceptions of individuals with T1D and how they cope with the fear, distress, and various burdens encountered.

Using the same population as in a previous report on a Japanese subgroup of SAGE participants, the limitations of the present analysis overlap with those appraised in the previous report [11]. First, causal relationships could not be inferred owing to the nature of a cross-sectional study. Second, as an observational study, selection bias, particularly survivor bias, might exist, that is, the older-age group included those with better conditions than the younger groups because of the healthy survivor effect. Moreover, we recruited participants to maintain the predefined country-level age-group ratio of 40%, 40%, and 20% of the 26–44-year-old, 45‒64-year-old, and ≥ 65-year-old groups, respectively. These ratios, however, were uniformly applied to all the participating countries, and did not necessarily reflect the prevalence ratio in each country. In Japan, the prevalence of older individuals with T1D is projected to increase and may account for > 20% of the T1D population in Japan. Therefore, the generalizability of the present findings to the Japanese individuals with T1D may be limited. Third, the possibility of information bias needs to be considered because of the retrospective nature of data collection. Fourth, we might not have included all potentially relevant confounding factors (e.g., severity of comorbidities) in this secondary analysis.

Conclusion

This sub-analysis of the Japanese subgroup of the global SAGE study suggested that the score for psychosocial aspects and quality of life varied across age groups, with a trend of improving scores with age, in relation to the less strict glycemic control targets adopted in older individuals in Japan. Glycemic control is significantly associated with treatment satisfaction. Considering the variability in individualized glycemic targets, a tailored approach to treatment and diabetes management is essential to maintain the preferred glycemic status in the long term.

Supplementary Information

Below is the link to the electronic supplementary material.

Funding

This work was supported by Sanofi K.K., Tokyo, Japan.

Data availability

The datasets used and analyzed during the current study are available from the corresponding author on reasonable request. Qualified researchers may request access to patient level data and related study documents including the clinical study report, study protocol with any amendments, blank case report form, statistical analysis plan and dataset specifications. Patient level data will be anonymised, and study documents will be redacted to protect the privacy of trial participants. Further details on Sanofi's data sharing criteria, eligible studies and process for requesting access can be found at: https://www.clinicalstudydatarequest.com.

Declarations

Conflict of interest

RN received honoraria from Sanofi, Medtronic Japan, Nippon Boehringer Ingelheim, Takeda Pharmaceutical, KISSEI PHARMACEUTICAL, Novartis Pharma, Eli Lilly Japan, Novo Nordisk Pharma, MSD, Astellas Pharma, and Abbott; and subsidies or donations from Taisho Pharmaceutical, Ono Pharmaceutical, Takeda Pharmaceutical, Nippon Boehringer Ingelheim, and Abbott. AS declares that he has no conflict of interest. NA received honoraria from Novo Nordisk Pharma and Eli Lilly Japan. MM received honoraria from Sanofi, Eli Lilly Japan, Novo Nordisk Pharma, Abbott Japan, Sumitomo Pharma, Kyowa Kirin, Nippon Boehringer Ingelheim, and Orizuru Therapeutics; research funding from Novo Nordisk Pharma; and subsidies or donations from Sysmex and Nissui. YT is an employee of Sanofi. HI received honoraria from Novo Nordisk Pharma, Sanofi, Terumo, and Sumitomo Pharma; and subsidies or donations from LifeScan Japan, Novo Nordisk Pharma, Sumitomo Pharma, and Taisho Pharmaceutical.

Human rights statement

All procedures followed were in accordance with the ethical standards of the responsible committee on human experimentation (institutional and national) and the Helsinki Declaration of 1964 and later versions. This study was approved by the ethics committee of the Nagasaki University Hospital (approval date: December 11, 2018; Approval No. 18052134-2) and each participating site.

Informed consent

Informed consent was obtained from all participants for being included in the study.