Development of a New Measure for Assessing Insulin Delivery Device Satisfaction in Patients with Type 1 and Type 2 Diabetes

Abstract

Background:

Although many different types of insulin delivery devices are currently available, there is no well-accepted, validated method to assess patient satisfaction with these devices and their impact on quality of life and other patient-reported outcomes. To address this problem, we developed the Insulin Device Satisfaction Survey (IDSS) and herein describe its construction and validation. We then examine how key patient factors are associated with device satisfaction.

Materials and Methods:

Items were developed from interviews with adults with type 1 diabetes (T1D) (n=10) and type 2 diabetes (T2D) using insulin (n=10), as well as eight healthcare professionals, leading to an initial pool of 32 items. Separate exploratory factor analyses (EFAs) were conducted with T1D subjects (n=279) and insulin-using T2D subjects (n=209). Construct validity was established with overall well-being (World Health Organization-5), diabetes distress (Diabetes Distress Scale), diabetes self-efficacy (Self-Efficacy for Diabetes Management Scale), and subscales from the Insulin Delivery System Rating Questionnaire. Regression analyses examined associations between total scale satisfaction and demographics, insulin adherence, clinical indicators, and device type (pump vs. nonpump users).

Results:

The two EFAs resulted in a 14-item scale for T1D subjects and a 12-item scale for T2D subjects, with eight items common across both samples. The EFAs yielded three coherent, meaningful factors in each sample, accounting for 55.6% (T1D sample) and 64.1% (T2D sample) of the variance. Validity was established by significant correlations with all criterion variables. For both samples, higher IDSS scores were significantly associated with better glycemic control and greater insulin adherence and pump use. For T2D subjects only, IDDS scores were significantly linked to fewer long-term complications, fewer low blood glucose readings, and older age.

Conclusions:

The IDSS is a reliable, valid measure of insulin device satisfaction in both its T1D form and T2 form. It provides a comprehensive profile of sources of device satisfaction for use in clinical care and research.

Introduction

Over the past decade, a range of new insulin delivery devices and new insulin device features has been introduced (e.g., tubeless insulin pumps, patch pumps, insulin pens with digital readouts), and more are soon to follow (e.g., dual-chamber pumps, insulin pens that send usage data to the cloud, insulin delivery devices integrated with blood glucose monitoring data). But how well are these new devices and device features accepted and/or valued by patients? To what degree, if any, do these devices positively influence quality of life (QOL) or support self-care attitudes and behavior? Unfortunately, these have been difficult questions to address because there is no well-accepted, validated method to assess satisfaction with insulin devices or the impact of such devices on key patient-reported outcomes.

Studies on these topics have been limited. Evaluations of insulin devices and device features have examined patient preferences, broad treatment satisfaction, and general impact on overall QOL,¹ and there have also been several early efforts to develop more comprehensive insulin device-focused assessment measures.

One is the Insulin Delivery System Rating Questionnaire (IDSRQ), which assesses multiple dimensions of treatment consequences, satisfaction, and preference.² The IDSRQ includes nine subscales, with many (although not all) focusing on issues specific to the respondent's insulin delivery system. These include treatment convenience, treatment problems, interference with daily activities, perceived efficacy of the device, and overall preference.

Another recent development is the Inhaled Insulin Treatment Questionnaire, which includes subscales assessing perceptions of current insulin therapy (e.g., convenience, comfort, ease of adherence, perceived efficacy), treatment satisfaction, treatment preference, and device performance.³

The Insulin Treatment Satisfaction Questionnaire focuses primarily on patient responses to the type of insulin used, not on the device itself.⁴

The Insulin Delivery Acceptability Questionnaire captures patient responses to potentially key device features (e.g., ease of storage, ease of setting an appropriate dose, ease of deducing when the cartridge is empty, overall effect on lifestyle), but the measure has never been subject to psychometric evaluation and has apparently never been used after its initial publication.⁵

Finally, the Treatment Related Impact Measure, Diabetes Device, which examines the larger impact of injectable delivery devices (not necessarily limited to insulin), assesses two key features—device function and device bother.⁶ Significant differences between type 1 diabetes (T1D) and type 2 diabetes (T2D) subjects on device bother have been reported, pointing to possible qualitative—not just quantitative—differences in how these two groups perceive their devices; this may suggest the need for specific scales that can delineate these potentially critical differences.⁶

Overall, it is unlikely that any of these existing scales will be applicable across the range of current and future insulin-related devices.

The development and validation of a comprehensive scale to assess insulin-related devices, one that could examine psychological impact, changes in QOL, and treatment satisfaction across the range of current and future devices, will better enable assessment of patient response to such devices, compare patient responses across devices, and document differences in responses based on device features, patient use factors, and patient demographics. In turn, such scales will aid in the development of better devices and better device training. Therefore, we developed the Insulin Device Satisfaction Survey (IDSS), which builds on the limitations of previously developed instruments. This report describes the construction, evaluation, and validation of the IDSS and examines how patient characteristics are associated with insulin device satisfaction.

Research Design and Methods

With content guided by the current literature, structured interviews were conducted with 10 T1D and 10 insulin-using T2D adults and with eight diabetes healthcare professionals. Verbal descriptions of respondents' personal experiences with insulin delivery devices and the different features of each device were documented, focusing on both the positive and negative attributes of each device, how each device affected respondents' feelings about diabetes and ability to manage diabetes, and the impact of the device on respondents' overall health and QOL. Respondents' comments were reviewed for duplication and were converted into 32 survey items. Patients and healthcare professionals then reviewed the items for completeness and clarity. The draft scale was formatted such that respondents could rate each item on a 5-point scale: 1=“strongly disagree,” 2=“disagree,” 3=“neutral,” 4=“agree,” and 5=“strongly agree.” The 32 survey items were included as part of an online assessment that included other instruments to be used for documenting the construct validity of the new scale.

Samples of insulin-using T1D and T2D adults were recruited from the Taking Control of Your Diabetes (TCOYD) Research Registry, which is a Health Insurance Portability and Accountability Act–protected online platform that includes adults with diabetes recruited primarily from TCOYD's 1-day diabetes education events conducted in multiple cities across the United States. When joining the Registry, patients agree to be contacted about participating in online survey research. For the current study, participants were required to be ≥19 years old, diagnosed with T1D or T2D, and, for those using insulin, doing so ≥1 year. Participants also had to be using their current insulin delivery device for >6 months. After receiving the initial study announcement via e-mail and successfully completing a brief online eligibility questionnaire, participants completed informed consent and the survey online. Each received a $10 electronic gift card for participation. The research protocol was approved by Ethical and Independent Review Services, a community-based, institutional review board.

Measures

In addition to the new 32-item scale, three groups of variables and scales were included to describe the sample and examine correlates of device satisfaction. Demographic indicators included age, gender, ethnicity (white vs. nonwhite), education (years), type of diabetes, number of years since diagnosis, number of years using insulin, and body mass index (calculated from self-reported weight and height). Measures of diabetes status included self-reported glycated hemoglobin (A1C), number of hypoglycemic readings (blood glucose values ≤70 mg/dL) in the past week, current insulin device (pump vs. no pump [i.e., pen or vial and syringe]), number of injections per day (for nonpump users only), and number of diabetes complications (from a list of eight). Finally, a modified version of the Hill-Bone Medication Adherence Scale⁷ was used to assess adherence to insulin recommendations, with subjects indicating how often they missed or forgot insulin injections because of a variety of different circumstances (ranging from 1=“none of the time” to 4=“all of the time”) (α=0.78). Reverse-coding of items indicated that higher scores signified greater adherence.

The following scales were included to assess the construct validity of the IDSS. The World Health Organization-5 is a 5-item scale that assesses QOL⁸ (α=0.86). The Diabetes Distress Scale (DDS) assesses worries and concerns specifically related to diabetes and its management; it has been shown to be a good marker of diabetes-related QOL⁹ (α=0.93). T2D participants completed the original, 17-item DDS, whereas T1D subjects completed the new DDS version for T1D, the 28-item T1-DDS¹⁰ (α=0.91). The Self-Efficacy for Diabetes Management Scale is a 10-item scale that assesses confidence in managing the critical diabetes self-care tasks¹¹ (α=0.91). Finally, the IDSRQ assesses multiple dimensions of treatment satisfaction with insulin delivery systems²; subjects completed three of the nine subscales: interference (α=0.91), helpfulness (α=0.89), and satisfaction (α=0.91).

Data analysis

Descriptive statistics examined item and scale distributions for each sample. The χ² and t tests, as appropriate, tested differences in patient characteristics and outcome variables across the two samples. Separate exploratory factor analyses (EFAs) were conducted with promax rotation for the T1D and T2D samples to determine whether the device satisfaction items could be reduced and grouped into coherent subscales and to explore whether differences in response patterns might point to the need to construct separate instruments for each sample. Internal consistency of the subscales and the total scales was determined by Cronbach's α statistic. Construct validity was examined by Pearson correlations between each subscale and total scale scores with QOL, diabetes distress, self-efficacy, and IDSRQ subscales. Multiple regression analyses examined associations of total IDSS satisfaction with the three groups of patient variables: demographics, diabetes status, and insulin adherence. Interactions between insulin device type (pump vs. nonpump) and key variables on total IDSS also were explored. Of note, analyses were initially undertaken to compare all three device types (pump vs. vial/syringe vs. pen), but no IDSS differences between vial/syringe and pen were apparent; therefore, to be parsimonious, we present pump versus nonpump findings only.

Results

Informed consent was completed by 504 patients, with 488 (97%) completing the entire survey. Table 1 gives a description of each of the two samples (T1D, n=279; T2D, n=209). As expected, the T1D group was significantly younger, had a lower body mass index, longer duration of diabetes, and more years using insulin than the T2D group. The T1D sample included significantly more non-Hispanic whites, insulin pump users, and real-time continuous glucose monitoring users than the T2D sample. In addition, T1D subjects reported a significantly lower A1C, fewer long-term complications, higher QOL, and more injections/day than T2D subjects.

Table 1.

Sample Description by Type of Diabetes

	Type 1 diabetes (n=279)	Type 2 diabetes (n=209)	P value ^a
Age (years)	47.1 (15.1)	59.0 (11.6)	<0.001
Gender (female)	156 (55.9%)	129 (61.7%)	0.20
Education level (years)	14.7 (4.3)	14.6 (4.2)	0.39
Ethnicity			<0.001
Non-Hispanic white	242 (86.7%)	146 (69.9%)
African American	3 (1.1%)	9 (4.3%)
Hispanic	5 (1.8%)	11 (5.3%)
Asian or Pacific Islander	3 (1.1%)	17 (8.2%)
Native American	3 (1.1%)	3 (1.4%)
Mixed race	8 (2.9%)	14 (6.7%)
Other	15 (5.4%)	9 (4.3%)
Years since diagnosis	24.6 (15.6)	15.1 (8.7)	<0.001
Number of complications	2.5 (2.6)	4.3 (2.2)	<0.001
Years using insulin	24.3 (15.9)	8.1 (7.2)	<0.001
Insulin delivery system			<0.001
Pump users	207 (74.2%)	19 (9.1%)
Nonpump users
Pen	44 (15.9%)	141 (67.5%)
Vial and syringe	28 (10.0%)	49 (23.4%)
Number of hypoglycemic readings (blood glucose values ≤70 mg/dL) in the past week	4.1 (2.7)	1.4 (2.0)	<0.001
Number of injections/day^b	4.5 (1.3)	3.1 (2.2)	<0.001
BMI (kg/m²)	26.3 (5.8)	33.4 (8.3)	<0.001
A1C (%)	7.1 (1.0)	7.6 (1.3)	<0.001
Insulin adherence	3.9 (0.2)	3.7 (0.4)	<0.001
Quality of life (WHO-5)	2.9 (1.1)	2.6 (1.1)	0.01
Diabetes distress (DDS-T1)	2.1 (0.8)	—
Diabetes distress (DDS)	—	2.5 (1.1)
Management self-efficacy (SEDM)	3.6 (0.8)	3.2 (0.9)	<0.001
IDSRQ subscale
Satisfaction	4.0 (0.9)	3.7 (0.9)	<0.001
Interference	3.2 (0.6)	3.3 (0.7)	0.29
Helpfulness	3.3 (0.8)	3.1 (0.8)	0.01

Data are mean (SD) values for continuous variables.

Comparing patients diagnosed with type 1 versus type 2 diabetes (by t test or χ² test).

For nonpump users only.

A1C, glycated hemoglobin; BMI, body mass index; DDS, Diabetes Distress Scale; DDS-T1, Diabetes Distress Scale, Type 1; IDSRQ, Insulin Delivery System Rating Questionnaire; SEDM, Self-Efficacy for Diabetes Management Scale; WHO-5, World Health Organization-5.

Factor analysis and construct validity of the IDSS

An EFA of the 32 initial items yielded three-factor solutions (eigenvalues ≥1.00) for each sample that accounted for 52.8% of the common item variance for T1D and 51.4% for T2D. A review of the items in each factor and the scree plot of successive eigenvalues further supported a three-factor solution in each analysis. Poorly loaded items (<0.50) or those that were cross-loaded on multiple factors (i.e., ≥30) were dropped, and the remaining items were subjected to a second EFA for each patient group. This analysis included 14 items for T1D and 12 items for T2D, eight of which were common across the two samples. This EFA yielded three coherent factors in each sample, accounting for 55.6% of the variance in the T1D sample and 64.1% in the T2D sample (Table 2).

Table 2.

Factor Analysis of Insulin Delivery Satisfaction Items by Type of Diabetes

	Type 1 diabetes			Type 2 diabetes
Item	Effective	Burdensome	Inconvenient	Effective	Burdensome	Freeing
Helps me to feel more in control of my diabetes.	0.921	0.083	−0.002
Works well when I need it.	0.806	−0.067	0.148	0.819	−0.078	0.004
Doesn't really benefit me much. (R)	–0.671	0.278	−0.192	–0.786	0.227	0.262
Helps me feel more positive about the future.	0.668	0.093	−0.246	0.274	−0.082	0.598
Helps me keep my blood glucose levels stable.				0.690	0.116	0.220
Helps me to have good blood glucose control.	0.753	0.183	−0.076	0.780	0.127	0.152
Is too complicated.	−0.010	0.689	0.017	−0.082	0.800	0.135
Is too much of a hassle to use.	−0.163	0.581	0.091	0.012	0.753	−0.180
Has too many pieces and parts to manage.	0.112	0.674	0.210	−0.178	0.790	0.054
Takes too much time to use.	−0.038	0.739	−0.017	0.176	0.849	−0.132
Is too expensive.	0.120	0.749	−0.218
Is often embarrassing to use when I am in public.	−0.041	0.021	0.735
Makes it difficult to be as spontaneous as I'd like to be.	−0.133	0.221	0.525
Is inconvenient to use when I am away from home.	0.088	−0.222	0.826
Is a hassle to carry around.	0.047	0.279	0.579
Helps me to avoid missing or forgetting doses.				−0.091	−0.137	0.707
Helps me to be more spontaneous in my life.				−0.093	−0.124	0.810
Helps me feel less restricted by diabetes.				0.147	0.217	0.659

(R) indicates the item is reverse-scored before averaging for subscale score. In addition, items that load onto the Emotional Burden and Behavioral burden subscales are reverse-scored before averaging across all items to obtain the total scale score.

Bolded items are those that load most highly on the individual factor.

The three subscales for the T1D sample were labeled as follows: “Effective” highlighted the perception that the device is useful and valuable (e.g., “Helps me to feel more in control of my diabetes”) (five items; α=0.83). “Burdensome” focused on frustrations and concerns regarding the operation of the device (e.g., “Takes too much time to use”) (five items; α=0.72). “Inconvenient” related to the specific hassles and discomfort of using the device while trying to live one's life (e.g., “Is often embarrassing to use when I am in public”) (four items; α=0.70). For the T2D sample, the three subscales were as follows: “Useful” centered on the perceived value of the device (e.g., “Helps me to keep my blood glucose levels stable”) (four items; α=0.80). “Difficult” was concerned with the obstacles associated with device usage (e.g., “Has too many pieces and parts to manage”) (four items; α=0.84). “Freeing” related to the QOL benefits resulting from the device (e.g., “Helps me feel less restricted by diabetes”) (four items; α=0.73). Overall, both IDSS versions—the 14-item scale for T1D subjects and the 12-item scale for T2D subjects—demonstrated good internal consistency for the total scale (α=0.85 and 0.86, respectively) and for each subscale (α range, 0.70–0.84).

Moderate intercorrelations among subscales (r=± 0.45 to r=± 0.58 for T1D; r=± 0.40 to r=± 0.53 for T2D), suggesting interrelated, but separate, dimensions of insulin device satisfaction. Mean scores for the T1D subscales were as follows: Effective, 4.1 (±0.7); Burdensome, 2.2 (±0.7); and Inconvenient, 2.3 (±0.9). Mean scores for the T2D subscales were as follows: Useful, 3.8 (±0.8); Difficult, 2.1 (±0.9); and Freeing, 3.2 (±0.8). The average total IDSS satisfaction score was 3.9 (±0.6) for T1D and 3.7 (±0.7) for T2D.

Construct validity

Significant associations (all P<0.05) were found between the total IDSS scale score (and each subscale score) in both samples with each of the construct validity variables (Table 3): greater satisfaction with one's insulin delivery device was associated with better QOL, less distress, and greater self-efficacy. Thus, the findings support the construct validity of the IDSS for both T1D and T2D patients across several interrelated constructs.

Table 3.

Construct Validity Associations

			IDSRQ
	WHO-5	Diabetes Distress	Satisfaction	Noninterference	Helpfulness	SEDM
Type 1 diabetes
Effective	0.34^a	−0.38^a	0.66^a	0.25^a	0.61^a	0.48^a
Burdensome	−0.26^a	0.42^a	−0.50^a	−0.34^a	−0.42^a	−0.39^a
Inconvenient	−0.26^a	0.34^a	−0.62^a	−0.37^a	−0.42^a	−0.38^a
Total score	0.35^a	−0.47^a	0.73^a	0.39^a	0.59^a	0.51^a
Type 2 diabetes
Useful	0.27^a	−0.42^a	0.54^a	0.18^a	0.53^a	0.47^a
Difficult	−0.24^a	0.57^a	−0.58^a	−0.46^a	−0.38^a	−0.29^a
Freeing	0.38^a	−0.34^a	0.64^a	0.07	0.52^a	0.49^a
Total score	0.37^a	−0.56^a	0.73^a	0.31^a	0.59^a	0.50^a

P<0.001.

IDSRQ, Insulin Delivery System Rating Questionnaire; SEDM, Self-Efficacy for Diabetes Management Scale; WHO-5, World Health Organization-5.

Patient characteristics and insulin device satisfaction

Four multiple regression equations were run separately for each sample to determine how each of four groups of patient characteristics was associated with total insulin device satisfaction: demographics (age, gender, education, ethnicity), insulin adherence, clinical indicators (A1C, number of complications, number of hypoglycemic episodes in past week), and pump versus nonpump use. All analyses included controls for age, gender, ethnicity, and education (Table 4).

Table 4.

Predictors of Insulin Device Satisfaction

	IDSS total satisfaction
	Type 1 diabetes		Type 2 diabetes
Predictor	β	P	β	P
Demographics
Age	0.09	0.19	0.19	0.01
Gender (female)	−0.01	0.83	−0.03	0.69
Education	0.07	0.29	0.05	0.52
Ethnicity (white)	0.01	0.88	−0.01	0.98
Insulin adherence	0.23	<0.001	0.36	<0.001
Clinical indicators
A1C	−0.23	<0.001	−0.35	<0.001
Number of complications	−0.03	0.64	−0.13	0.05
Number of hypoglycemic readings in the past week	−0.07	0.24	−0.14	0.04
Device type (pump vs. nonpump use)	0.35	<0.001	0.24	0.001

Results are adjusted for age, gender, ethnicity, and education.

A1C, glycated hemoglobin; IDSS, Insulin Device Satisfaction Survey.

Among the four demographic variables, only age was significantly and independently associated with device satisfaction and only for the T2D sample (β=0.19, P=0.01): older T2D subjects reported greater satisfaction with their current insulin device than younger T2D subjects. No other patient demographic characteristic was significantly associated with device satisfaction in either sample.

In the second analysis, for both T1D subjects and T2D subjects, higher insulin adherence was significantly associated with greater device satisfaction (β=0.23, P<0.001 for T1D subjects; β=0.36, P<.001 for T2D subjects).

In the third analysis, which examined clinical indicators, lower A1C (β=−0.35, P<0.001), less frequent hypoglycemic episodes (β=−0.14, P=0.04), and fewer long-term complications (β=−0.13, P=0.05) were each significantly linked to greater device satisfaction for T2D subjects. However, for T1D subjects, only lower A1C was significantly associated with higher insulin device satisfaction (β=−0.23, P<0.001). Adding insulin adherence to the third analysis, both insulin adherence (β=0.15, P<0.05 for T1D subjects; β=0.19, P<0.05 for T2D subjects) and lower A1C (β=−0.20, P<0.01 for T1D subjects; β=−0.31, P<0.001 for T2D subjects) remained significant and independent predictors of greater device satisfaction for both T1D subjects and T2D subjects.

In the last analysis, for both T1D subjects and T2D subjects, pump users reported significantly greater device satisfaction than nonpump users (β=0.35, P<0.001 for T1D subjects; β=0.24, P=0.001 for T2D subjects). Given this significant effect, we also explored in supplementary analyses the extent to which device satisfaction varied by patient characteristics differentially for pump and nonpump users. Interactions were added to each of the initial equations for T1D subjects and T2D subjects separately. Three significant interactions were found for T1D subjects only: between insulin device type and age (P<0.01), education level (P=0.05), and insulin adherence (P<0.01). Those T1D subjects using pumps displayed no significant differences in device satisfaction based on any demographic characteristic. In contrast, those not using pumps displayed significant positive associations between device satisfaction and age (r=0.27, P=0.01), education (r=0.22, P<0.05), and insulin adherence (r=0.47, P<0.001). Thus, for T1D nonpump users only, device satisfaction was significantly higher among older than younger adults, more highly educated compared with less educated subjects, and more adherent than less adherent subjects.

Given these significant effects of insulin device type on satisfaction, we also asked in supplementary analyses what effect the number of injections/day might have on device satisfaction—among nonpump users only. For both T1D and T2D nonpump users, fewer injections/day was significantly linked to greater device satisfaction (β=−0.34, P=0.004 for T1D subjects; β=−0.30, P<0.001 for T2D subjects). These results add to the findings that device satisfaction is not uniform across patient samples; it is linked to patient characteristics, regimen demands, and patient behavior.

Discussion

To better understand patients' attitudes and perspectives regarding their insulin delivery devices, we developed the IDSS, with moderately different versions for T1D patients (n=14 items) and T2D patients (n=12 items). Both versions contain three clinically meaningful subscales that capture key features contributing to insulin device satisfaction. For the T1D version, the subscales are labeled “Effective,” which represents the perceived value of the device, “Burdensome,” which reflects the day-to-day bothers associated with using the device, and “Inconvenient,” which represents the hassles of using the device while trying to live one's life. For the T2D version, the subscales are labeled “Useful,” which represents the perceived benefits resulting from device usage, “Difficult,” which centers on the obstacles to using the device effectively, and “Freeing,” which reflects the broad positive impact of the device on QOL.

For both T1D and T2D samples, greater device satisfaction (including IDSS total and all subscale scores) is associated with less diabetes distress, greater well-being, higher levels of diabetes self-efficacy, and better adherence to insulin regimens, enhancing the construct validity of the IDSS. Also, reliability analyses indicate good internal consistency of each subscale and the total satisfaction scale for each sample.

It is noteworthy that T1D and T2D pump users report significantly greater device satisfaction than nonpump users and that the occurrence of fewer daily injections among nonpump users is associated with greater device satisfaction. This underscores patients' broad dissatisfaction with injections—as number of injections increase, device satisfaction is reduced, and no injections (i.e., pump use) is viewed as more satisfying than any frequency of injections. Indeed, these data are consistent with a previous report from Nicolucci et al.,¹² in which pump users reported significantly greater treatment satisfaction than nonpump users.

Greater device satisfaction is also associated with more successful self-reported behavioral outcomes, such as better insulin adherence, and clinical outcomes, especially better glycemic control and, for T2D subjects only, fewer hypoglycemic episodes and fewer long-term complications. Although causation cannot be determined from these cross-sectional findings, one explanation is that IDSS scores reflect patients' broader satisfaction or dissatisfaction with their current diabetes status and overall diabetes regimen. Alternatively, greater device satisfaction may serve as an instigator to glycemic success, with satisfaction leading to greater insulin adherence, resulting in better clinical outcomes. Of note, our analyses indicate that insulin adherence in the T1D sample is associated with device satisfaction only among nonpump users. This suggests that T1D pump users, who have likely made a greater financial and perhaps emotional commitment to their insulin delivery device than nonpump users, may feel more committed to following insulin regimen recommendations, even when frustrated or dissatisfied with their insulin delivery system.

The current findings suggest that the IDSS is relatively stable across demographic groups, with one important exception—greater device satisfaction is significantly higher among older patients than younger patients. This is consistent with previous studies demonstrating that younger patients, especially in the T2D population, have greater diabetes-related emotional distress, more depressive symptoms, and poorer adherence to self-care behaviors than older patients.¹³

The IDSS provides several advantages over currently available instruments. First, it contains separate scales for T1D and T2D respondents; previous scales were limited to T2D only^4,5 or did not examine differences between T1D and T2D.^3,6 Second, the IDSS delineates coherent, psychometrically sound subscales of device satisfaction, thus allowing for a richer perspective on patients' responses to insulin delivery devices. Although the IDSRQ also described an array of subscales, these were apparently derived a priori and were not subjected to factor analytic procedures.² Finally, the IDSS is one of the few scales devoted entirely to device satisfaction,^2,6 and it is not limited to one type of device.³ Thus, these attributes may be of unique value in future studies of device satisfaction.

Key advantages of the current study are that IDSS items were derived from the reported experiences of patients and healthcare professionals, final scale development was based on the responses of large samples of T1D and insulin-using T2D patients, and the results indicate acceptable reliability and validity. Study limitations must also be noted. First, the study samples were primarily non-Hispanic white and well educated (Table 1), factors that may limit generalizability. The scales should be reevaluated with a broader patient sample. Second, glycemic control (A1C) was assessed via self-report rather than by a laboratory analysis, which could have introduced important bias. However, fewer than 3% of subjects were unable to report a value or reported an improbable value, and in our related work with T1D adults there was a high level of agreement between self-reported A1C and the laboratory result (r=0.84; L.F., 2015, unpublished data).

In conclusion, the IDSS is a new measure of device satisfaction that has been shown to be reliable and valid. Separate forms have been developed for T1D and T2D patients. The IDSS subscales also describe primary, underlying dimensions of device satisfaction, thus providing a more comprehensive picture of a patient's experience with his or her insulin device. The current findings also point to important qualitative differences in insulin device satisfaction between T1D and T2D adult subjects and between pump users and nonpump users. This strongly suggests that device satisfaction is not uniform across populations; therefore, designers of future insulin delivery devices might benefit by targeting specific patient groups with features tailored to address these variations.

Footnotes

Acknowledgments

This investigator-initiated study was supported by Sanofi.

Author Disclosure Statement

W.H.P. has worked as a consultant for Sanofi Diabetes Care, Dexcom, Roche Diabetes Care, Abbott Diabetes Care, and Johnson & Johnson. L.F. has worked as a consultant for Roche Diabetes Care, Eli Lilly, Abbott Diabetes Care, and Sanofi Diabetes Care. D.H. reports no competing interest. S.V.E. has worked as a consultant for Sanofi Diabetes Care, Dexcom, Bayer Diabetes Care, Abbott Diabetes Care, and Johnson & Johnson.

W.H.P. and L.F. were responsible for study conception/protocol. D.H. was responsible for statistical analysis. W.H.P., L.F., D.H., and S.V.E. were responsible for manuscript development. All authors read and approved the final manuscript.

References

Molife

, Lee

, Shi

, Sawhney

, Lenox

: Assessment of patient-reported outcomes of insulin pen devices versus conventional vial and syringe. Diabetes Technol Ther, 2009; 11:529–538.

Peyrot

, Rubin

: Validation and reliability of an instrument for assessing health-related quality of life: the Insulin Delivery System Rating Questionnaire. Diabetes Care, 2005; 28:53–58.

Rubin

, Peyrot

: Psychometric properties of an instrument for assessing the experience of patients treated with inhaled insulin: the Inhaled Insulin Treatment Questionnaire. Health Qual Life Outcomes, 2010; 8:32.

Anderson

, Skovlund

, Marrero

, Levine

, Meadows

, Brod

, Balkrishnan

: Development and validation of the Insulin Treatment Satisfaction Questionnaire. Clin Ther, 2004; 26:565–578.

Barnett

, Bowen

, Burden

: Multicentre study to assess quality of life and glycaemic control of type 2 diabetic patients treated with insulin compared with oral hypoglycaemic agents. Pract Diabetes Int, 1996; 13:179–183.

Brod

, Hammer

, Christensen

, Lessard

, Bushnell

: Understanding and assessing the impact of treatment in diabetes: the Treatment-Related Impact Measures for Diabetes and Devices (TRIM-Diabetes and TRIM-Diabetes Device). Health Qual Life Outcomes, 2009; 7:83.

Kim

, Hill

, Bone

, Levine

: Development and testing of the Hill-Bone Compliance to High Blood Pressure Therapy Scale. Prog Cardiovasc Nurs, 2000; 15:90–96.

Hajos

, Pouwer

, Skovlund

, Den Oudsten

, Geelhoed-Duijvestijn

, Tack

, Snoek

: Psychometric and screening properties of the WHO-5 well-being index in adult outpatients with Type 1 or Type 2 diabetes mellitus. Diabet Med, 2013; 30:e63–e69.

Polonsky

, Fisher

, Earles

, Dudl

, Lees

, Mullan

, Jackson

: Assessing psychosocial distress in diabetes: development of the Diabetes Distress Scale. Diabetes Care, 2005; 28:626–631.

10.

Fisher

, Polonsky

, Hessler

, Mashrani

, Blumer

, Peters

, Strycker

, Bowyer

: Understanding the sources of diabetes distress in adults with type 1 diabetes. J Diabetes Complications, 2015; 29:572–577.

11.

Iannotti

, Schneider

, Nansel

, Haynie

, Plotnick

, Clark

, Sobel

, Simons-Morton

: Self-efficacy, outcome expectations, and diabetes self-management in adolescents with type 1 diabetes. J Dev Behav Pediatr, 2006; 27:98–105.

12.

Nicolucci

, Maione

, Franciosi

, Amoretti

, Busetto

, Capani

, Bruttomesso

, Di Bartolo

, Girelli

, Leonetti

, Morviducci

, Ponzi

, Vitacolonna

: Quality of life and treatment satisfaction in adults with Type 1 diabetes: a comparison between continuous subcutaneous insulin infusion and multiple daily injections. Diabet Med, 2008; 25:213–220.

13.

Hessler

, Fisher

, Mullan

, Glasgow

, Mashrani

: Patient age: a neglected factor when considering disease management in adults with type 2 diabetes. Patient Educ Couns, 2011; 85:154–159.