Changes in Treatment Satisfaction Over 3 Years in Patients With Type 2 Diabetes After Initiating Second-line Treatment.

CONTEXT: J-DISCOVER is a prospective observational cohort study aiming to understand the current management of patients with early-stage type 2 diabetes mellitus (T2DM) in Japan, enrolling patients initiating second-line treatment.
OBJECTIVE: The current analysis examined the change in treatment satisfaction during the study period and factors affecting this change among patients in J-DISCOVER.
METHODS: We used data from the J-DISCOVER study, in which 1798 patients with T2DM aged ≥ 20 years were enrolled from 142 sites across Japan. Treatment satisfaction was assessed using the Diabetes Treatment Satisfaction Questionnaire (DTSQ).
RESULTS: The mean DTSQ treatment satisfaction score increased from 25.9 points at baseline to 27.3 points at 6 months, which was maintained through 36 months. Among the baseline characteristics examined, higher baseline DTSQ treatment satisfaction scores (P < 0.0001), older age (≥ 75 vs < 65 years, P = 0.0096), living alone (P = 0.0356), and type of facility (clinics vs hospitals, P = 0.0044) had a significantly negative impact on the changes in DTSQ treatment satisfaction scores. Improvement in mean glycated hemoglobin (HbA1c) from baseline (7.7%) to 36 months (7.1%) was associated with positive changes in the DTSQ treatment satisfaction score (P = 0.0003).
CONCLUSION: Changes in DTSQ treatment satisfaction scores were related to HbA1c improvement, suggesting that the management strategy was appropriately planned for each patient. The results also suggest that the availability of social support for patients with T2DM who are elderly or living alone may be an important factor affecting treatment satisfaction, adherence, and clinical outcomes.

The number of people with diabetes mellitus (DM) in Japan is growing. The prevalence has increased from 7.2 million patients in 2015 to 11 million patients in 2021 (1, 2). The treatment goals for diabetes are the prevention of the onset or exacerbation of comorbidities associated with DM, including micro- and macrovascular complications, as well as improvement in quality of life (QoL) of patients (3). Therefore, management of risk factors, including hyperglycemia, at each clinical visit is essential.

Since type 2 diabetes mellitus (T2DM) is a chronic disease, improving the QoL and preventing comorbidities associated with T2DM are important for the patient to lead a healthy life. The Diabetes Treatment Satisfaction Questionnaire (DTSQ) was developed to assess and measure the quality of diabetes-specific treatment satisfaction with prescribed antidiabetic regimens (4). It is recommended by both the World Health Organization and International Diabetes Federation and is currently used worldwide (5-10). In Japan, a translated version of the DTSQ has been prepared and validated (11). Saisho highlighted that DTSQ can be used to compare treatment satisfaction and assess the quality of diabetes care in clinical settings, which will ultimately support a better understanding of patient needs and improve diabetes treatment (6).

The DISCOVERing Treatment Reality of Type 2 Diabetes in Real World Settings (DISCOVER) study program (ClinicalTrials.gov identifiers: NCT02322762 and NCT02226822 [DISCOVERing Treatment Reality of Type 2 Diabetes in Real World Setting in Japan; J-DISCOVER]) (12, 13) was set up to address the knowledge gaps in the management and clinical outcomes of patients with T2DM. The overarching aim of the program was to describe patient characteristics, evolution of disease management patterns, clinical outcomes, and patient-reported outcomes over 3 years in patients with T2DM who were initiating a second-line glucose-lowering therapy in a real-world setting. These patients were chosen as the focus of the study because of the diversity of treatment options recommended at this stage of the disease. A better understanding of practice variations across and within different countries, their determinants, and associated patient outcomes is key for providing effective treatment decisions.

Here we report the results of the DTSQ analysis from the J-DISCOVER study, which examined changes in treatment satisfaction over 3 years in 1798 Japanese patients with T2DM who had initiated a second-line glucose-lowering treatment (defined as an add-on option to first-line treatment of a glucose-lowering drug or a switch between treatments). The analysis aimed to identify factors related with treatment satisfaction after initiation of second-line treatment in patients who have a relatively short duration of the disease.

Methods

J-DISCOVER Study Design

The study rationale and design, baseline characteristics, and results of the J-DISCOVER study have been published elsewhere (13, 14). Full explanations of the definitions and diagnostic criteria for T2DM, as well as full inclusion and exclusion criteria and treatment targets, have also been published (15). The J-DISCOVER study was a 3-year, multicenter, prospective, observational, longitudinal study conducted at 142 sites across Japan, which enrolled patients with T2DM whose diabetes was inadequately controlled with a first-line treatment and who had initiated a second-line treatment. Patients were recruited from hospitals and clinics, including both diabetes specialist and nonspecialist care settings. Men and women aged ≥ 20 years were eligible for enrollment if they were diagnosed with T2DM and their treatment pattern involved an add-on option or switching to a second oral or parenteral antidiabetic medication after a first-line oral monotherapy. All procedures were in accordance with the International Conference on Harmonization of Good Clinical Practice, the Ethical Guidelines for Epidemiological Research of Japan, the ethical standards of the responsible committee on human experimentation (institutional and national), and the Helsinki Declaration of 1964, as revised in 2013. Informed consent was obtained from all patients included in the study. This study was approved by the relevant ethical committees of the participating sites and informed consent was obtained from all patients at each site. All patients participated voluntarily and were free to discontinue their participation at any point.

Full explanations of the primary objective and key secondary objectives are provided in separate manuscripts (12, 15). One of the key secondary objectives was patient-reported outcomes for the overall population and each second-line treatment. After obtaining informed consent from patients, baseline data were collected and follow-up data were obtained from patient records at 6, 12, 24, and 36 months, with a ± 2-month buffer period. Patients were followed up for information on their diabetes treatment during the study period, and relevant data were collected at the study sites until the completion of the study. This manuscript adheres to the applicable Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guidelines (16), and the STROBE checklist is provided as Supplementary Information (17).

DTSQ Assessment

Patient-reported outcomes were measured using the Japanese version of the DTSQ developed specifically for assessing treatment satisfaction in T2DM (11). The questionnaire was completed by patients during their allocated visits at 6, 12, 24, and 36 months. The DTSQ comprises 8 questions, of which 6 (questions 1 and 4-8) are used to assess different domains of treatment satisfaction: (a) overall satisfaction, (b) convenience, (c) flexibility, (d) understanding of diabetes, (e) willingness to recommend current treatment to others, and (f) willingness to continue the current treatment (4), each of which is scored on a 7-point Likert scale from 0 (very dissatisfied) to 6 (very satisfied). Thus, the DTSQ treatment satisfaction score can range from 0 to 36 and represents the treatment satisfaction of the patient with their diabetes treatment (5, 6). Two additional items, perceived hyperglycemia and perceived hypoglycemia, are assessed through 2 questions (questions 2 and 3) on the frequency of hyperglycemia and hypoglycemia, using a 7-point scale from 0 (none of the time) to 6 (most of the time).

Questionnaires were administered at the initiation of the second-line treatment and at 6, 12, 24, and 36 months of treatment and were completed by the patients before or after the examination at the site visit. Responses were also allowed via telephonic interviews during the follow-up period (at 6, 12, 24, and 36 months).

Statistical Analysis

The analysis population consisted of all patients (n = 1798) enrolled in the study. Categorical variables are shown as frequencies and percentages. For continuous measures, means (SDs) were calculated. Changes in DTSQ scores at 6, 12, 24, and 36 months from baseline were analyzed using a paired t test with the Holm method. The relation between changes in DTSQ treatment satisfaction score and patient characteristics was analyzed using a linear mixed-effects model. Explanatory variables included age (< 65 years vs ≥ 65 years and < 75 years vs ≥ 75 years), sex, baseline glycated hemoglobin A1c (HbA1c; continuous variable, per 1% increment), body mass index (BMI; < 22 kg/m2 and ≥ 25 kg/m2 vs ≥ 22 kg/m2 and < 25 kg/m2), renal function (estimated glomerular filtration rate [eGFR] < 60 mL/min/1.73 m2 vs ≥ 60 mL/min/1.73 m2), history of cardiovascular disease, and first-line antihyperglycemic medication. Individual differences in DTSQ treatment satisfaction scores were set as a random effect. To assess the relation between changes in DTSQ treatment satisfaction score and changes in HbA1c, a linear mixed-effects model was performed using the following variables: changes in DTSQ treatment satisfaction score from baseline to 36 months as the objective variable and changes in HbA1c from baseline to 36 months, DTSQ treatment satisfaction score at baseline, and HbA1c at baseline as the explanatory variables. A P value of < 0.05 was considered statistically significant.

Results

Patient baseline data is shown in Table 1. Briefly, the mean age of patients at baseline was 61.6 ± 12.8 years and 61.8% of patients were male. The mean HbA1c value in the overall population was 7.7 ± 1.3%; 49.0% of patients had a history of drinking, 80.4% of patients attended clinics and lived with someone, and 69.2% of the physicians in charge were endocrinologists/diabetologists (Table 1). Additional patient characteristics have been described in previous studies (14, 15). The most prescribed first-line treatment was a dipeptidyl peptidase-4 inhibitor (DPP4i), and the distribution of first-line treatments was comparable across second-line treatment groups, except for DPP4i (14). The mean DTSQ treatment satisfaction score was 25.9 points at baseline and increased to 27.3 points at 6 months (5.4 percentage points increase), which was maintained through 36 months (Fig. 1a). A similar trend was observed for each medication, except for glucagon-like peptide-1 receptor agonists (GLP-1RAs), likely due to the small number of patients in this category. The mean score of patient-perceived hyperglycemia frequency was 2.8 points at baseline, which subsequently decreased to 2.3 points at 6 months and was maintained between 2.2 and 2.3 points through 36 months (Fig. 1b). Similar trends were observed for each treatment, except for GLP-1RAs and insulin, likely due to the small number of patients in those categories. In addition, higher baseline HbA1c (11.7%) in the insulin group may account for the higher score at baseline (14). The mean score of perceived hypoglycemia frequency was approximately 1.0 point throughout the 36-month observation period (Fig. 1c). Similar trends were observed for each treatment. Details of the DTSQ scores, for the overall population and for each second-line treatment, by DTSQ question item are shown in Supplementary Tables 1-9 (17).

Table 1.

Baseline patient characteristics

Patient characteristic
First-line treatment
BG, n (%)	384 (21.4)
SU, n (%)	128 (7.1)
α-GI, n (%)	122 (6.8)
TZD, n (%)	62 (3.4)
Glinide, n (%)	64 (3.6)
DPP-4i, n (%)	967 (53.8)
SGLT2i, n (%)	71 (3.9)
Second-line treatment
Add-on, n (%)	1546 (86.0)
Switched, n (%)	252 (14.0)
BG, n (%)	503 (28.0)
SU, n (%)	195 (10.8)
α-GI, n (%)	102 (5.7)
TZD, n (%)	116 (6.5)
Glinide, n (%)	87 (4.8)
DPP-4i, n (%)	557 (31.0)
SGLT2i, n (%)	219 (12.2)
GLP-1RA, n (%)	6 (0.3)
Insulin, n (%)	13 (0.7)
Age, mean (SD), years	61.6 (12.8)
Age < 65 years, n (%)	988 (54.9)
Age ≥ 65 years and < 75 years, n (%)	526 (29.3)
Age ≥ 75 years, n (%)	284 (15.8)
Male, n (%)	1111 (61.8)
BMI, mean (SD), kg/m 2	25.5 (4.6)
BMI ≥ 25 kg/m2, n (%)	885 (49.2)
BMI ≥ 22 and < 25 kg/m2, n (%)	531 (29.5)
BMI < 22 kg/m2, n (%)	371 (20.6)
HbA1c, mean (SD), %	7.7 (1.3)
HbA1c ≥ 7.0%, n (%)	1278 (71.1)
HbA1c ≥ 8.0%, n (%)	518 (28.8)
eGFR, mean (SD), mL/min/1.73 m2	78.9 (20.5)
Chronic kidney disease, n (%)	251 (14.0)
Cardiovascular disease, n (%)	204 (11.3)
Facility
Clinic, n (%)	1445 (80.4)
General/community hospital, n (%)	352 (19.6)
University/teaching hospital, n (%)	1 (0.1)
Other type of center, n (%)	0 (0.0)
Specialty of the physician-in-charge
General practitioner, n (%)	237 (13.2)
Endocrinology/diabetology, n (%)	1245 (69.2)
Cardiology, n (%)	291 (16.2)
Nephrology, n (%)	6 (0.3)
Geriatrics, n (%)	0 (0.0)
Other specialty, n (%)	19 (1.1)
Living style
Living alone, n (%)	263 (14.6)
Living with someone, n (%)	1479 (82.3)
Declined to answer, n (%)	55 (3.1)
Education level
Primary (1-6 years of education), n (%)	17 (0.9)
Secondary (7-13 years of education), n (%)	1138 (63.3)
University/ higher education degree (> 13 years of education), n (%)	454 (25.3)
Declined to answer, n (%)	188 (10.5)
Employment status
Full-time, n (%)	783 (43.6)
Part-time, n (%)	209 (11.6)
Not working, n (%)	405 (22.5)
Retired, n (%)	277 (15.4)
Other, n (%)	63 (3.5)
Declined to answer, n (%)	60 (3.3)
Smoking history
Nonsmoker, n (%)	774 (43.1)
Former smoker, n (%)	506 (28.2)
Current smoker, n (%)	447 (24.9)
Unknown, n (%)	70 (3.9)
Alcohol intake
Lifetime abstainer, n (%)	440 (24.5)
Former drinker, n (%)	379 (21.1)
Drinker, n (%)	881 (49.0)

Abbreviations: α-GI, α-glucosidase inhibitors; BG, biguanides; BMI, body mass index; DPP-4i, dipeptidyl peptidase-4 inhibitors; eGFR, estimated glomerular filtration rate; GLP-1RA, glucagon-like peptide-1 receptor agonists; HbA1c, hemoglobin A1c; SGLT2i, sodium-glucose cotransporter-2 inhibitors; SU, sulfonylureas; TZD, thiazolidinediones.

Figure 1.

(a), Changes in DTSQ treatment satisfaction scores over time by second-line therapy. Data are shown as mean with SD. (b), Changes in the perceived hyperglycemia frequency scores over time by second-line therapy. Data are shown as mean with SD. (c), Changes in the perceived hypoglycemia frequency scores over time by second-line therapy. Data are shown as mean with SD. Abbreviations: α-GI, α-glucosidase inhibitors; BG, biguanides; DPP-4i, dipeptidyl peptidase-4 inhibitors; DTSQ, Diabetes Treatment Satisfaction Questionnaire; GLP-1RA, glucagon-like peptide-1 receptor agonists; SGLT2i, sodium-glucose cotransporter-2 inhibitors; SU, sulfonylureas; TZD, thiazolidinediones.

The DTSQ treatment satisfaction scores of 0-36 points were categorized into 6 groups (0-6, 7-12, 13-18, 19-24, 25-30, and 31-36) to further examine the changes in scores over 36 months (Fig. 2). A decrease in the proportion of patients from baseline to 6 months was observed in the following groups: (a) group of 0-6 points (0.7% to 0.2%), (b) 7-12 points (1.7% to 1.1%), (c) 13-18 points (14.5% to 9.4%), and (d) 19-24 points (24.4% to 23.2%). After 6 months, the proportion of patients in these groups remained at a similar level through 36 months. In contrast, the proportion of patients in 2 groups increased from baseline to 6 months: 25-30 points (31.2% to 34.4%) and 31-36 points (27.5% to 31.8%). After 6 months, the proportion of patients in these 2 groups remained higher compared with baseline through 36 months.

Figure 2.

Changes in the distribution of DTSQ treatment satisfaction score among patients in 6 categorical groups from baseline to 36 months. Patients were categorized into 6 groups based on the DTSQ treatment satisfaction scores at each time point, and the proportions of patients in each category are shown. Abbreviation: DTSQ, Diabetes Treatment Satisfaction Questionnaire.

A linear mixed-effects model was performed to identify the factors related with the changes in DTSQ treatment satisfaction scores. The baseline characteristics used as variables in this analysis were the types of first-line and second-line treatments, age, sex, BMI, HbA1c, eGFR, type of facility, the specialty of the physician-in-charge, living style, education level, employment status, smoking history, and alcohol intake (Table 1). Among these variables, duration of treatment (P < 0.0001), baseline DTSQ treatment satisfaction score (P < 0.0001), age (P = 0.0049), type of facility (P = 0.0044), and living style (P = 0.0356) had a statistically significant impact on the changes in DTSQ treatment satisfaction scores (Table 2).

Table 2.

The linear mixed-effects model for testing the influence of variables on changes in DTSQ treatment satisfaction score

Effect	F valuea	P value
Duration of treatment (6 months, 12 months, 24 months, 36 months)	F (3, 992) = 8.02	< 0.0001
Baseline DTSQ treatment satisfaction score (per 1 score increment)	F (1, 992) = 618.81	< 0.0001
First-line therapy	F (6, 992) = 1.06	0.3833
Second-line therapy	F (8, 992) = 0.86	0.5521
Age (< 65 years, ≥ 65 years to < 75 years, ≥ 75 years)	F (2, 992) = 5.34	0.0049
Sex (male, female)	F (1, 992) = 0.00	0.9582
BMI (< 22 kg/m 2 , ≥ 22 to < 25 kg/m 2 , ≥ 25 kg/m 2 )	F (2, 992) = 2.11	0.1214
HbA1c (per 1% increment)	F (1, 992) = 0.01	0.9275
DTSQ hyperglycemia awareness score	F (1, 992) = 0.48	0.4904
DTSQ hypoglycemia awareness score	F (1, 992) = 0.07	0.7914
Renal function (eGFR) (< 60 mL/min/1.73 m 2 , ≥ 60 mL/min/1.73 m 2 )	F (1, 992) = 2.99	0.0840
History of cardiovascular disease	F (1, 992) = 0.99	0.3200
Type of facility (clinics/ hospitals/ university hospital)	F (1, 992) = 8.14	0.0044
Specialty of the physician-in-charge(DM specialists/endocrinologists/cardiologists/general practitioners/nephrologists/other specialists)	F (1, 992) = 3.77	0.0524
Living style (living alone, living with someone)	F (1, 992) = 4.43	0.0356
Education level (university graduate or above/high school graduate or below)	F (1, 992) = 0.15	0.6965
Employment status (full-time/part-time/not working/ retired)	F (2, 992) = 0.57	0.5641
Smoking history (current smoker/former smoker/nonsmoker)	F (2, 992) = 0.50	0.6041
Alcohol intake (drinker/former drinker/lifetime abstainer)	F (2, 992) = 0.11	0.9002

Abbreviations: BMI, body mass index; DM, diabetes mellitus; DTSQ, Diabetes Treatment Satisfaction Questionnaire; eGFR, estimated glomerular filtration rate; HbA1c, hemoglobin A1c.

aF value (numerator degree of freedom, denominator degree of freedom).

For the variables with a P value lower than 0.1 in Table 2, we further investigated whether each variable had a positive or a negative impact on the changes in DTSQ treatment satisfaction scores. As shown in Table 3, a negative impact on the DTSQ treatment satisfaction score was observed with 24 months treatment duration compared with 6 months treatment duration, higher baseline DTSQ treatment satisfaction scores, patients aged ≥ 75 years compared with patients aged < 65 years, and patients who lived alone compared with those who lived with someone. Type of facility (hospitals/university hospitals compared with clinics) had a positive impact on DTSQ treatment satisfaction scores. Although lacking statistical significance, eGFR less than 60 mL/min/1.73 m2 (vs ≥ 60 mL/min/1.73 m2) and specialty of the physician-in-charge (general practitioner/cardiologist/nephrologist/other specialist vs endocrinologist/DM specialist) had a negative impact on the changes in DTSQ treatment satisfaction scores.

Table 3.

Influence of variables (P < 0.1 in Table 2) on changes in DTSQ treatment satisfaction score

Effect	Level	Estimate	Standard error	t value	Pr > | t |
Period	12 months (ref: 6 months)	0.1248	0.1718	t (992) = 0.73	0.4676
	24 months (ref: 6 months)	−0.7763	0.1941	t (992) = −4.00	< 0.0001
	36 months (ref: 6 months)	−0.2375	0.1991	t (992) = −1.19	0.2332
Baseline DTSQ treatment satisfaction score	Continuous value: per 1 score increment	−0.5144	0.0207	t (992) = −24.88	< 0.0001
Age	≥ 65 years, < 75 years (ref: < 65 years)	0.1218	0.3639	t (992) = 0.33	0.7379
	≥ 75 years (ref: < 65 years)	−1.3030	0.5023	t (992) = −2.59	0.0096
Type of facility	Hospital/university hospital (ref: clinic)	1.0179	0.3568	t (992) = 2.85	0.0044
Living style	Living alone(ref: living with someone)	−0.8039	0.3821	t (992) = −2.10	0.0356
Renal function (eGFR)	< 60 mL/min/1.73 m2 (ref: ≥ 60 mL/min/1.73 m2)	−0.7114	0.4112	t (992) = −1.73	0.0840
Specialty of the physician-in-charge	general practitioner/cardiologist/ nephrologist/other specialist (ref: endocrinologist/DM specialist)	−0.5978	0.3078	t (992) = −1.94	0.0524

Abbreviations: DTSQ, Diabetes Treatment Satisfaction Questionnaire; ref, reference; eGFR, estimated glomerular filtration rate.

Finally, we examined whether changes in HbA1c from baseline to 36 months (mean ± SD; 7.7 ± 1.3% at baseline and 7.1 ± 1.0% at 36 months) (15), DTSQ treatment satisfaction score at baseline, and HbA1c at baseline were related with the change in DTSQ treatment satisfaction score from baseline to 36 months and found a statistically significant effect with each variable (Table 4). Specifically, the DTSQ treatment satisfaction score from baseline to 36 months decreased by 0.698 points for every 1.0% increase in HbA1c from baseline to 36 months (parameter estimates −0.698 [95% CI: −1.074, −0.322]). The DTSQ treatment satisfaction score also decreased by 0.569 points for every 1.0-point increase in DTSQ treatment satisfaction score at baseline (parameter estimates −0.569 [95% CI: −0.617, −0.522]) and by 0.376 points for every 1.0% increase in baseline HbA1c (parameter estimates −0.376 [95% CI: −0.742, −0.011]).

Table 4.

Multiple linear regression analysis for testing the influence of variables on changes in DTSQ treatment satisfaction score

Explanatory variable	Parameter estimates	95% CI	P value	Adjusted R2
HbA1c changes: Baseline to 36 months	−0.698	−1.074, −0.323	0.0003	0.3231
Baseline DTSQ treatment satisfaction score	−0.569	−0.617, −0.522	< 0.0001
Baseline HbA1c	−0.376	−0.742, −0.011	0.0437

Explanatory variables were continuous values and the association was estimated per 1% increment in HbA1c and per 1 point increment in DTSQ treatment satisfaction score. A P value < 0.05 was considered significant.

Abbreviations: DTSQ, Diabetes Treatment Satisfaction Questionnaire; HbA1c, hemoglobin A1c.

Discussion

In the present study, using the DTSQ, we first investigated long-term (3-year) changes in treatment satisfaction after the initiation of second-line treatment in patients with T2DM in Japan. Treatment satisfaction scores improved at 6 months and were maintained at a similar level for the remainder of the follow-up period. Choice of the first-line medication or second-line medication was not related with the change in the DTSQ treatment satisfaction score, and baseline characteristics, such as duration of treatment, baseline DTSQ treatment satisfaction score, age (≥ 75 years vs < 65 years), lifestyle (living alone vs living with someone), and type of facility (hospital/university hospital vs clinic) showed a significant relation with the changes in DTSQ treatment satisfaction scores during the follow-up period. Furthermore, the change in DTSQ treatment satisfaction score from baseline to 36 months was related with HbA1c improvement after the initiation of second-line treatment. To the best of our knowledge, no previous studies have demonstrated the factors affecting long-term changes in DTSQ treatment satisfaction scores and the relation between changes in HbA1c and DTSQ treatment satisfaction scores.

Assessment of the quality of diabetes treatment should include clinical aspects such as glycemic control and prevention of comorbidities and psychological aspects of patients such as treatment satisfaction and mental well-being (5). Diabetes is a progressive chronic disease, and longer treatment durations tend to lead to increases in the number of antidiabetic drugs prescribed (18). As the dosing frequency of diabetes medications increases, the total burden score as measured by the diabetes treatment burden questionnaire tends to increase (19, 20). Additionally, treatment adherence may be compromised when QoL and/or treatment satisfaction is affected by a treatment regimen, such as increased number of medications (20). The target patient population of the J-DISCOVER study was patients with T2DM who were required to initiate a second-line treatment to control their HbA1c, and these patients were followed for 36 months after the initiation of the second-line treatment; 48.2% of these patients changed their medications at least once during the follow-up period (14). In the current analysis, the mean DTSQ treatment satisfaction score showed an increase of 0.8 to 1.4 points from baseline through the 36-month follow-up, and this change was significantly related with changes in HbA1c from baseline to 36 months. These results suggest that improved HbA1c control by add-on therapy may have helped to maintain patient treatment motivation and adherence, and thus may have further influenced their treatment satisfaction in spite of the increase in the drugs prescribed.

Management of diabetes involves self-care by patients, including diet restriction and exercise, in addition to pharmacological therapy (21). If a drug regimen negatively impacts a patient’s QoL, adherence to above interventions may be compromised (20). The importance of treatment satisfaction in diabetes treatment is well recognized, and an understanding of the factors that affect treatment satisfaction of patients with T2DM could help improve treatment adherence and clinical outcomes. Though results vary depending on study population and design, previous studies have reported that both clinical and socio-demographic factors are related with treatment satisfaction and QoL in patients with T2DM (19, 20, 22-26). A linear mixed-effects model in the current study identified older age and living alone as factors related with negative changes in treatment satisfaction. Park et al have shown that medication adherence decreased in elderly patients when multiple prescriptions were given due to deterioration of their recognition and memory, visual acuity, and/or hearing (27). Lower observed medication adherence has generally been shown to negatively impact treatment satisfaction (10, 28). For elderly patients with T2DM, higher adherence is related with a lower probability of hospitalization or emergency department visits, as well as lower healthcare costs (29). Therefore, diabetes medications should be selected for elderly patients with attention to dosing frequency and timing. Another study has also shown that living alone predicted a decline in adherence; however, living alone was not a predictor among patients who had support from family and friends (30). This difference may suggest that the availability of social support is critical in delivering better clinical outcomes and treatment satisfaction for patients living alone. In a large-scale meta-regression analysis, Charvat et al demonstrated that the prevalence of T2DM would likely rise over the next few decades in Japan, mainly because of the aging population (31). The proportion of Japanese adults aged ≥ 65 years has increased over the years and is expected to reach almost 36 million by 2050, almost 35% of the total population (32). Furthermore, the proportion of individuals living alone in Japan is particularly high in the older age group and is expected to increase from 14% in 2015 to 20.8% in 2040 among men aged ≥ 65 years and from 12.8% to 18.4% among those aged ≥ 75 years. The proportion is also expected to rise in women from 21.8% in 2015 to 24.5% in 2040 for those aged ≥ 65 years (33). Therefore, further studies will be of great value for providing insights on whether support by social workers or healthcare multidisciplinary teams could improve treatment satisfaction of elderly patients with T2DM.

We have found that hospitals, as a type of facility in which patients were treated, were related with positive changes in treatment satisfaction. Compared with clinics, hospitals typically can offer more diverse support through an interdepartmental collaborative approach of multidisciplinary teams that include dietitians, social workers, and physiotherapists. Our results indicate that the availability of social support from family, social workers, and medical staff has a significant impact on the ability of diabetes patients to achieve clinical treatment goals, as well as to maintain an enhanced QoL.

Although we also found that higher baseline DTSQ treatment satisfaction scores were related with negative changes in treatment satisfaction, this result may be due to the nature of the DTSQ treatment satisfaction score, e.g., regression to the mean effect and ceiling effects.

In this study, the frequency of perceived hyperglycemia decreased from 2.8 to 2.3 points in 6 months and was maintained throughout the 3-year study period. However, the frequency of perceived hypoglycemia slightly increased from 0.9 to 1.0 point and was maintained at the same level throughout the study period. Only 2 cases of severe hypoglycemic events were reported (14). A history of severe and symptomatic hypoglycemia has been related with poor QoL in patients with T2DM (34, 35), since these patients can develop a fear of hypoglycemia when treatment is escalated or switched, which thereby reduces patients’ willingness to take medication as directed (35).

Patients on sulfonylurea or insulin have been reported to have a high incident rate of hypoglycemia (36, 37). However, our study showed no significant increase in perceived hypoglycemia frequency among patients who started sulfonylureas or insulin as a second-line treatment. It is possible that these medications were given for patients with a high baseline HbA1c and those who were less susceptible to hypoglycemic episodes. Overall, our findings suggest that Japanese physicians selected the most suitable diabetes medication for each patient based on individual patient backgrounds and treatment preferences.

Strengths and Limitations

A major strength of this study is that it was designed to analyze the determinants of treatment decisions and relations between treatment patterns and patient-reported outcomes. It is one of the few studies to assess T2DM treatment satisfaction in a large population of patients in Japan by different medical facilities and provide insight into treatment adherence and QoL of patients across different age groups. Patients with T2DM of short disease duration (median 3.1 years) were selected to evaluate treatment patterns after the initiation of second-line therapy regardless of the agent(s) prescribed. The long follow-up period of 3 years is another key strength of this study; this is one of the longest studies of patients with T2DM from Japan.

The study has some limitations. First, the procedure for completing the DTSQ was not standardized. Although the questionnaire was given and completed at the time of the site visit, the timing of entry was not specified (whether the questionnaire was completed before or after the patient was informed of their blood glucose levels or any other test results), which may have impacted treatment satisfaction. Additionally, factors such as waiting time at a clinic/hospital, the change of physician-in-charge, and transfer to another medical facility during the follow-up period may have affected the results. Second, no information on patient disease education was acquired. Patients’ understanding of the disease could be biased because of misrecognized symptoms of hyperglycemia as thirst, polydipsia, polyuria, and/or hypoglycemia. Third, we did not collect information regarding escalation or switching of antidiabetic medications during the follow-up period. Therefore, the relation between these factors and the DTSQ treatment satisfaction score was not considered. Lastly, general limitations of the DTSQ score such as ceiling or floor effects and the effects of regression to the mean should be considered when interpreting the results. In particular, a higher baseline DTSQ treatment satisfaction score was related with negative changes in DTSQ treatment satisfaction scores in our analysis, which indicates a DTSQ ceiling effect where high baseline scores result in little improvement in satisfaction over time (37).

Conclusion

In this study, the DTSQ treatment satisfaction scores improved at 6 months and were maintained at a similar level for the remainder of the 3-year follow-up period. The DTSQ treatment score change was related with HbA1c improvement after the initiation of second-line treatment, which suggests that treatment patterns were appropriately planned for each patient. Age, living style, and type of facility were factors related with a significant change in the DTSQ treatment satisfaction score. Considering the increasing population of the elderly and those living alone in Japan, the availability of social support at clinics or hospitals as well as at home is likely a critical factor to affect treatment satisfaction, adherence, and clinical outcomes of patients with T2DM.