Participants

Table 1 summarizes the sociodemographic characteristics for each participant group. Overall, 411 physicians completed the survey: 259 diabetologists (63.0%) and 152 general internists (37.0%), with a mean of 20 and 25 years of clinical experience, respectively. Of the 330 PWD who completed the survey, 108 (32.7%) had T1D, 108 (32.7%) had T2D and were treated with insulin, and 114 (34.5%) had T2D and were insulin-naive. Mean (standard deviation [SD]) age across all three subgroups was 61.0 (9.5) years. Overall, 328 caregivers completed the survey, comprising 107 family members (32.6%), 117 nurses (35.7%), and 104 care workers (31.7%). Participant flow is presented in Fig. S1.

Table 1 Summary of sociodemographic characteristics of participating physicians, PWD, and caregiversCauses of Resistance to Insulin Initiation Among PWD: Physicians and Insulin-Naive People with T2D

Figure 1 shows the proportions of physicians and insulin-naive people with T2D who agreed with each statement in the negative image of injections, social effects/interpersonal effects, feelings of guilt regarding diabetes self-management, and negative image towards insulin therapy categories. Overall, physicians perceived the most common reasons for PWD resistance to insulin initiation to be “I don’t want to inject myself every day” (89.3%), “I don’t want to be bothered with doing injections” (86.9%), and “I don’t want to inject myself multiple times a day” (86.4%). Similarly, the most common reason for resistance reported by insulin-naive people with T2D was “I don’t want to be bothered with doing injections” (92.1%) (Fig. 1).

Fig. 1

Causes of resistance to insulin initiation among PWD based on responses from physicians and insulin-naive people with T2D. The figure shows the percentage of physicians who responded “a major cause of resistance” or “a little cause of resistance,” and of PWD who responded “completely agree,” “mostly agree,” or “slightly agree” to each statement. *Statistically significant differences (p < 0.05 [χ2]) between physician and PWD responses. PWD people with diabetes, T2D type 2 diabetes

Regarding the proposed statements describing potential causes of resistance to insulin initiation, the beliefs of physicians and those of insulin-naive people with T2D differed statistically significantly (p < 0.05) for 11 statements (Fig. 1). The statements with the largest discrepancies were “my family will be worried” (24.9%-point difference), “I’m on insulin because I didn’t do what I was supposed to” (24.4%-point difference), and “people would think it is my fault that I need to take insulin” (21.4%-point difference).

Causes of Resistance to Insulin Initiation Among PWD: Comparison Between PWD Subgroups

Causes of resistance to insulin initiation among PWD were analyzed by diabetes subgroup to identify any differences in opinions between people with T1D, insulin-naive people with T2D, or insulin-treated people with T2D (Fig. 2). In the negative image of injections category, compared with insulin-treated people with T2D, a statistically significantly higher proportion of insulin-naive people with T2D agreed with each of the proposed statements describing potential barriers to insulin initiation (Fig. 2). The largest differences were for the statements “using injectors is difficult” (47.5%-point difference), “injections are scary” (34.6%-point difference), and “I don’t want to inject myself every day” (28.4%-point difference). Statistically significant differences were also observed between people with T1D and insulin-naive people with T2D for five items in this category (Fig. 2).

Fig. 2

Current feelings towards insulin therapy based on responses from people with T1D, insulin-treated people with T2D, and insulin-naive people with T2D. The figure shows the percentage of PWD who responded “completely agree,” “mostly agree,” or “slightly agree” to each statement. Asterisks highlight statistically significant differences (*p < 0.05 [χ2]) between subgroups. N/A not applicable, PWD people with diabetes, T1D type 1 diabetes, T2D type 2 diabetes

In the social/interpersonal effects category, the proportion of respondents who agreed that “people would think it is my fault that I need to take insulin” was statistically significantly higher (p < 0.05) for insulin-treated and insulin-naive people with T2D (67.6% and 66.7%, respectively) than for people with T1D (50.9%). The level of agreement for the statement “injecting insulin in the presence of others is embarrassing” was consistent in people with T1D and insulin-treated people with T2D (64.8% for both).

In the feelings of guilt regarding diabetes self-management category, a statistically significant difference was observed between the proportion of insulin-naive people with T2D and insulin-treated people with T2D who agreed that “using insulin means my disease is worsening” (71.1% vs. 57.4%, respectively; p < 0.05). Agreement with the statement “I’m on insulin because I didn’t do what I was supposed to do” statistically significantly differed between people with T1D and those with T2D (insulin-naive or insulin-treated) (Fig. 2).

In the negative image towards insulin therapy category, the proportion of insulin-naive people with T2D was statistically significantly higher (all p < 0.05) than that of insulin-treated people with T2D for the statements “I don’t understand why insulin is necessary for me” (47.4% vs. 14.8%, respectively), “I’m afraid of side effects” (57.9% vs. 38.9%, respectively), and “insulin will limit my daily life activities” (68.4% vs. 42.6%, respectively). A statistically significant difference was also observed between the proportion of people with T1D and insulin-treated people with T2D for the statement “insulin will limit my daily life and activities” (61.1% vs. 42.6%, respectively; p < 0.05). The proportion of people with T1D who agreed with the statement “insulin therapy is financially burdensome” was statistically significantly higher than that of people with T2D, regardless of insulin treatment (p < 0.05; Fig. 2). Agreement with the statement “I’m afraid of hypoglycemia” was statistically significantly different between people with T1D and insulin-naive people with T2D. Agreement was consistently high across PWD groups (84.3–91.2%) for the statement “I don’t want to do insulin injections if there is another way” (Fig. 2).

Physician and Family Member Barriers to Insulin Initiation

To explore physician barriers to insulin initiation, physicians were asked about their feelings towards insulin therapy (Fig. S2a). Statements were divided into issues with doctor’s experience, burden related to explanations, considerations of burden on PWD, concerns regarding insulin therapy, and concerns regarding hypoglycemia. Physicians most frequently felt that “PWD would resist insulin therapy” (83.5%), “PWD would have to pay more for treatment” (76.6%), and that “there is a higher risk of hypoglycemia with insulin therapy than with other therapies” (70.8%). Compared with the other categories, relatively few physicians agreed with the statements in the issues with doctor’s experience category.

Family members were also asked about their feelings and possible barriers towards assisting insulin injection. Statements were divided into impact on daily life, perception regarding care, support from family, and impact on health status (Fig. S2b). Family members most frequently agreed with the statements “I am afraid of hypoglycemia” (48.6%), “assisting with insulin self-injection is important to me” (47.7%), and “it is difficult to get support from other family members for assistance with insulin self-injection” (32.7%). Agreement with all other statements was low (≤ 25.0%; Fig. S2b).

Reasons and Criteria for Initiating Insulin Therapy

When considering the initiation of insulin therapy, physicians placed the most importance on “person’s ability to inject (in terms of vision/grip strength)” (75.7%), “person’s cognitive function” (72.3%), and “HbA1c” (72.3%) (Fig. S3). Responses were broadly similar for diabetologists and general internists, with diabetologists also emphasizing “insulin secretion capacity” (75.3% vs. 53.9%, respectively) and “person’s financial status” (41.3% vs. 24.3%, respectively) (Fig. S3). Overall, the most common PWD-reported reasons for starting insulin therapy were “recommendation from a doctor” (70.6%), “aiming for better HbA1c and blood glucose levels” (26.4%), and “worried about progression of diabetes caused by complications” (20.6%) (Fig. S4). Similar trends were observed across PWD subgroups; however, 32.5% of insulin-naive people with T2D also answered “other”.

Ideal and Actual HbA1c Levels at Insulin Initiation

Physicians were asked at which HbA1c level they would ideally initiate insulin if they themselves had T2D. These results were compared with the actual HbA1c levels reported by insulin-treated people with T2D when they started insulin treatment (Fig. 3). More than 70% of physicians (diabetologists or general internists) considered an HbA1c level of under 9.0% (75 mmol/mol) to be ideal for insulin initiation; however, only 33.4% of insulin-treated individuals with T2D reported starting insulin treatment with an HbA1c level of under 9.0% (75 mmol/mol) (Fig. 3). The most popular HbA1c range to start insulin as imagined by physicians if they themselves had T2D was 8.0 to < 8.5% (64 to < 69 mmol/mol) (Fig. 3). The HbA1c level at which physicians would recommend insulin therapy to insulin-naive people with T2D treated with oral glucose-lowering medication varied according to the individual’s age, cognitive function, and activities of daily living, with a trend for a higher ideal HbA1c threshold as clinical background deteriorated (Fig. S5).

Fig. 3

Physicians’ views on the HbA1c level at which they would consider initiating insulin if they themselves had T2D, and actual HbA1c levels at the time of insulin initiation reported by insulin-treated people with T2D. HbA1c glycated hemoglobin, T2D type 2 diabetes

To explore how HbA1c targets are defined in PWD, respondents who reported having a defined HbA1c target were asked how their target was decided (Fig. S6). Overall, 51.5% of these PWD were involved in determining their HbA1c target to some degree, whereas 43.9% were given a target by their physicians without consultation. This trend was observed across PWD subgroups (Fig. S6b). In the overall PWD group, 78.9% of people had an HbA1c target of under 7.0% (53 mmol/mol) and 39.1% had an HbA1c target of under 6.5% (48 mmol/mol) (Fig. S6c).

Difficulties with Using Insulin Injections

People with T1D or insulin-treated people with T2D were asked about difficulties in using insulin injections based on their experience. Overall, the most difficult attribute associated with insulin injections was administering “injections while away from home, traveling or on business,” with 54.6% of people with T1D and 57.4% of insulin-treated people with T2D agreeing this was “very difficult” or “somewhat difficult.” This was followed by “injecting during busy times (hours)” (44.4% and 48.1% of respondents, respectively) (Fig. 4). The proportion of respondents who agreed that “adjusting insulin doses” and “handling hypoglycemia” were difficult was statistically significantly higher for people with T1D (21.3% and 46.3%, respectively) than for insulin-treated people with T2D (11.1% and 25.9%, respectively; both p < 0.05) (Fig. 4).

Fig. 4

Views on the difficulties in the use of insulin injections as reported by people with T1D and insulin-treated people with T2D. The bar chart shows the proportion of PWD who responded “very difficult” or “somewhat difficult” for each item. aFor “taking insulin with meals,” n = 91 for people with T1D and n = 55 for insulin-treated people with T2D. This item was not applicable to people who used basal insulin only. *Statistically significant differences (p < 0.05 [χ2]) between PWD subgroups. PWD people with diabetes, T1D type 1 diabetes, T2D type 2 diabetes

Degree of Burden on Caregivers

Caregivers (family members, nurses, and care workers) were asked about the burden of assisting or preparing for self-injections based on their experience (Fig. 5). In general, a smaller proportion of nurses than family members and care workers considered assisting or preparing self-injections to be burdensome. Nurses considered “instruct/reinstruct PWD and family members on how to perform injections” to be the highest burden, with 30.0% of respondents agreeing this was “somewhat burdensome” or “very burdensome.” More than one-third (37.3%) of family members reported feeling some degree of burden when assisting or preparing for insulin self-injection (Fig. S7). The highest burden for family members was associated with “check for insulin balls” (42.1%), “handling hypoglycemia (talking to PWD, wiping perspiration, taking temperature etc.)” (38.9%), and “giving assistance when injecting insulin” (35.0%) (Fig. 5). Care workers considered “handling hypoglycemia” to be the highest burden (43.6%). Statistically significant differences (all p < 0.05) were found across ten items, including “calling out (encourage insulin injections before meals etc.),” “take out the necessary tools for self-injection (injectors, alcohol swabs etc.),” and “attaching and removing needles,” between nurses and both family members and care workers (Fig. 5).

Fig. 5

Caregivers’ views on the degree of the burden on them to assist or prepare for insulin self-injections. The bar chart shows the proportions of caregivers who responded “very burdensome” or “somewhat burdensome” for each item. *Statistically significant differences (p < 0.05 [Fisher’s exact test]) between caregiver subgroups. The numbers of respondents varied between question items in each subgroup because selection of multiple items was permitted, and not all respondents in each subgroup selected the same number of items. N/A not applicable, PWD people with diabetes

The median (minimum–maximum) time spent assisting or preparing for self-injections was 30 (0–600) min/week for family members, 30 (1–120) min/week for nurses, and 15 (0–350) min/week for care workers. Family members were also asked about their reactions when the person with diabetes they care for has a sick day: 49.5% agreed with the statement “if the person is very sick and cannot check blood glucose levels, seek medical assistance,” 27.1% agreed with “measure blood glucose more often (every 2–4 h),” 20.6% agreed with “stop taking all diabetes medications and/or using insulin,” 19.6% agreed with “try to exercise as much as possible to lower blood glucose levels,” and 16.8% agreed with “if blood glucose exceeds 270 mg/dL, take lots of drinks with no added sugar.”

Expectations for Future Insulin Therapy

Physicians, PWD, and caregivers were asked for their views on expected attributes for new insulin therapies, specifically when initiating a new insulin therapy in insulin-naive PWD or when switching from an existing insulin. Figures 6, 7, and 8 show the most frequently selected attributes by respondent subgroup. Across subgroups, regardless of whether the respondent was considering insulin initiation or switching from an existing insulin therapy, “less frequent injections (e.g., daily becomes weekly),” “simpler treatment (e.g., fewer diabetes medications),” “further improved HbA1c and blood glucose levels,” “lower risk of hypoglycemia,” and “simpler preparation for injections” were selected as the most important attributes expected from a future insulin therapy. In the nurse subgroup, “further improved HbA1c and blood glucose levels” was not in the top five; instead, “counter measures can be taken in case of a forgotten injection or when it is not possible to inject at a prescribed time” was selected (Fig. 8). However, nurses selected “simpler preparation for injections” more often than family members (30.9%-point difference between subgroups) and care workers (21.9%-point difference between subgroups) when considering switching from existing insulin to a new insulin.

Fig. 6

Treatment characteristics expected from a new insulin therapy when initiating insulin treatment or switching from an existing insulin to a new insulin, as reported by physicians. It was mandatory for respondents to select at least one characteristic; respondents were able to select up to five characteristics in total. HbA1c glycated hemoglobin

Fig. 7

Treatment characteristics expected from a new insulin therapy when initiating insulin treatment or switching from an existing insulin to a new insulin, as reported by PWD. It was mandatory for respondents to select at least one characteristic; respondents were able to select up to five characteristics in total. HbA1c glycated hemoglobin, PWD people with diabetes, T1D type 1 diabetes, T2D type 2 diabetes

Fig. 8

Treatment characteristics expected from a new insulin therapy when initiating insulin treatment or switching from an existing insulin to a new insulin, as reported by caregivers. It was mandatory for respondents to select at least one characteristic; respondents were able to select up to five characteristics in total. HbA1c glycated hemoglobin