Diabetes Ther (2018) 9:297–307

https://doi.org/10.1007/s13300-017-0365-1

ORIGINAL RESEARCH

Patient Perceptions and Preferences for a Mealtime

Insulin Delivery Patch
Mark Peyrot . Darlene Dreon . Vivien Zraick . Brett Cross .
Meng H. Tan

Received: November 17, 2017 / Published online: January 11, 2018

Ó The Author(s) 2018. This article is an open access publication

ABSTRACT medication) patch use over a 3-day period.

Participants evaluated mealtime insulin-deliv-
Introduction: A basal-bolus insulin regimen is ery systems using eight measures from five
needed to achieve glycated hemoglobin A1c domains (convenience, interference with daily
(HbA1c) below 7.0% in people with type 1 activities, diabetes-related worry, psychological
(T1D) or type 2 (T2D) diabetes who have sig- well-being, and overall satisfaction/preference)
nificant loss of beta-cell function. Nonadher- on the self-administered Insulin Delivery Sys-
ence to therapy is common and negatively tem Rating Questionnaire. User ratings of their
affects the ability to reach treatment goals. We current insulin-delivery systems (syringe, pen,
examined patient assessment of a new, wearable pump) were compared with those for the patch
mealtime insulin-delivery system (patch) rela- by repeated measure analysis of variance and
tive to their current mealtime insulin-delivery one-sample t tests.
system (syringe, pen, or pump). The patch is Results: Participants had significant (p\0.05)
designed to deliver only boluses of fast-acting preference for patch over syringe in all eight
insulin (no basal insulin), mechanically con- comparisons, and over pen in five out of eight
trolled by the patient. comparisons, with no significant preference for
Methods: Adults (n = 101) with T1D or T2D pen. Although there was a preference for patch
assessed their current mealtime insulin-delivery over pump in six out of eight comparisons, only
system and then assessed simulated (no active one showed a significant preference for patch,
and one for pump. Significantly more partici-
Enhanced content To view enhanced content for this pants reported that they would like to switch to
article go to http://www.medengine.com/Redeem/ the patch than continue using a syringe (78% vs
5E1DF06017C067E3. 22%) or pen (76% vs 24%) but this difference
was not significant for the group using a pump
M. Peyrot (&) (52% vs 48%).
Department of Sociology, Loyola University
Maryland, Baltimore, MD, USA
Conclusions: Participants preferred using the
e-mail: mark.peyrot@gmail.com patch over pens and syringes. Its ease of use and
discreet method of insulin delivery may con-
D. Dreon  V. Zraick  B. Cross
tribute to improved patient adherence to
Calibra Medical, Wayne, PA, USA
mealtime insulin regimens among people cur-
M. H. Tan rently using injection devices.
Division of Metabolism, Endocrinology and Funding: Calibra Medical
Diabetes, Department of Internal Medicine,
University of Michigan, Ann Arbor, MI, USA
298 Diabetes Ther (2018) 9:297–307

Keywords: Diabetes mellitus; Insulin delivery; evening, is less of the barrier to insulin therapy
Patient preference; Quality of life than MDI mealtime insulin [10]. Barriers to MDI
adherence include, but are not limited to, the
need for multiple injections daily, difficulties
INTRODUCTION associated with insulin administration, lack of
discreetness, fear of injection-associated pain,
Current clinical practice guidelines by the and availability of insulin when needed [11–13].
American Diabetes Association, the European In the past decade, approximately 60% of
Association for the Study of Diabetes, and people with diabetes taking insulin did not
Canadian Diabetes Association recommend a achieve the proposed glycemic goals [14]. In
goal of glycated hemoglobin A1c (HbA1c) of less people with T2D, the prandial insulin response
than 7.0% for most people with diabetes [1–3]. is delayed and blunted, which contributes to
This goal is based on the findings of the Dia- postprandial hyperglycemia [15–18]. The closer
betes Control and Complications Trial in type 1 people with well-controlled T2D are to achiev-
diabetes (T1D) [4] and the UK Prospective Dia- ing the HbA1c goal of 7.0%, the more important
betes Study in type 2 diabetes (T2D) [5], which it becomes to control prandial glucose excur-
report that better glycemic control leads to a sions [19].
lower risk of onset or progression of both micro- However, initiating mealtime insulin and the
and macrovascular complications. The curvi- anticipation of frequent insulin dosing can feel
linear relationship between the microvascular overwhelming to patients. Therefore, there is a
complication rate and HbA1c was almost ‘‘flat’’ need for discreet and easy-to-use insulin delivery
in individuals with HbA1c close to 7.0% [5]. systems. Advances have been made in the
In people without diabetes, glucose home- development of a variety of tools to help patients
ostasis is attained and maintained by several overcome barriers and simplify insulin delivery,
physiological mechanisms, including continu- ranging from smaller-gauge needles to minimize
ous basal insulin secretion complemented by injection pain to pre-filled and pre-mixed insulin
prandial insulin peaks when glucose rises after a pens, insulin jet injectors using pressurized air
meal. To achieve glycemic goal, people with instead of needles to penetrate the skin, insulin
T1D need a basal-mealtime insulin regimen, pumps to continuously deliver insulin through-
which can be implemented via multiple daily out the day, insulin patch-pumps worn attached
injections (MDI) or continuous subcutaneous to the skin for subcutaneous insulin delivery, and
insulin infusion (CSII). A basal-mealtime insu- a wearable insulin delivery device for on-demand
lin regimen is also often needed to achieve delivery of prandial insulin [20].
glycemic goal in people with T2D when they Here we discuss a wearable mealtime insulin-
develop significant insulin deficiency with pro- delivery system (Calibra Medical, Inc., Wayne,
gressive loss of beta-cell function. PA, USA; in this report referred to as ‘‘the patch’’)
An estimated 30% of people with T2D use that provides a simple, reliable, and discreet
insulin, either as monotherapy or in combina- option for mealtime insulin delivery. It can be
tion with oral antihyperglycemic medications worn on the body for up to 3 days and offers on-
to manage their glycemic control [6]. Studies demand mealtime insulin for people with dia-
show that insulin therapy is initiated in less betes. Insulin is delivered via a subcutaneous
than 50% of people with T2D for whom it is cannula that is inserted as the device is applied,
recommended. If insulin treatment is initiated, which means that the skin is penetrated only
one-third of people with T2D do not continue once every 3 days rather than multiple times per
with it long term [7, 8], and even when treat- day, as with mealtime insulin administration via
ment is continued, more than half of people traditional syringe or pen MDI delivery methods.
with diabetes miss injections, with 20% admit- Similar wearable devices, such as the Omni-
ting to regular omissions of insulin doses [9]. podÒ Insulin Management System (Insulet
Basal insulin, often taken as a once-a-day Corp, Bedford, MA, USA), V-GoTM (Valeritas,
injection in the privacy of home during the Inc., Bridgewater, NJ, USA), or PAQ (CeQurTM,
Diabetes Ther (2018) 9:297–307 299

Montreux, Switzerland), were designed to deli- of either T1D or T2D. Furthermore, they were
ver both basal and bolus insulin doses. Unlike required to already use a basal-bolus insulin
the Calibra patch for bolus insulin delivery regimen with a minimum of one basal insulin
(which is completely mechanical and does not injection and two mealtime insulin injections
require batteries) these other patch-pumps are per day using syringe or pen, or were on a CSII
managed by external controllers or predeter- regimen (insulin pump). Participants were
mined settings for basal insulin delivery [21]. excluded if they were currently employed by a
The aim of this study was to assess user per- research firm or were working in the medical
ception, satisfaction, and preferences for their industry.
current method of mealtime insulin delivery
(syringe, pen, or pump) compared with simu- Study Device
lated use (no active medication) of the patch
during a 3-day period, using a validated self-re- The mealtime insulin-delivery system is a small
port questionnaire. patch (65 9 35 9 8 mm) that can be worn on
the body for up to 3 days (Fig. 1). It is classified
METHODS as a new US Food and Drug Administration
product code OPP, under 21 Code of Federal
Study Design Regulations (CFR) 880.5725; approved for use
with rapid-acting insulins HumalogÒ (Eli Lilly
and Co., Indianapolis, IN, USA) and NovoLogÒ
The study was conducted from 2007 through
(Novo Nordisk Inc., Plainsboro, NJ, USA). The
2010 at three research centers (Redwood City,
patch holds up to 200 units of rapid-acting
CA, USA; Seattle, WA, USA; Orlando, FL, USA).
insulin (at 100 units/mL) that are manually
Each participant evaluated the patch after a
administered in increments of 2 units by press-
simulated-use trial period of 3 days at home.
ing buttons located on each side of the device.
They used the patch with sterile diluent only, and
The buttons can be operated through clothing.
in parallel with their usual device for actual
Mealtime insulin is delivered subcutaneously.
mealtime insulin delivery. The Insulin Delivery
After the two buttons are simultaneously pres-
System Rating Questionnaire (IDSRQ) [22] was
sed, a mechanically timed sequence delivers
administered prior to and after simulated patch
insulin from the reservoir through a soft poly-
use, with the initial administration used to rate
mer cannula into the body (Fig. 2). A built-in
the participants’ current mealtime insulin-de-
alarm (locked buttons) will alert the user if no
livery system and the latter used to rate the patch.
insulin has been delivered.
The study was approved by the BioMedical
Research Institute of America Institutional
Review Board (San Diego, CA, USA). All proce- Participant Training for Use of the Study
dures followed were in accordance with the Device
ethical standards of the appropriate committees
on human experimentation (institutional and Prior to the simulated-use period, participants
national) and with the Helsinki Declaration of received training by a human-factors engineer
1964. Informed consent was obtained from all (‘‘the trainer’’), who followed a training protocol
patients for inclusion in the study. All partici- developed with a diabetes educator. Patients
pants provided written informed consent prior were trained following the step-by-step
to study participation. instructions outlined in the standard mealtime
insulin-delivery system quick-start guide that is
Participants provided with the product. After this demon-
stration, participants completed a comprehen-
sion test of the instructions for the tasks of
Participants eligible for the study were 18–-
opening the packaging, filling the device,
85 years of age and had a pre-existing diagnosis
priming the device, applying the device, dosing,
300 Diabetes Ther (2018) 9:297–307

Fig. 1 The wearable on-demand mealtime insulin-delivery b through clothing by actuating the buttons on both sides
system (patch). a The mealtime insulin-delivery patch of the patch, c which measures no more than
(Calibra Medical, Inc., Wayne, PA, USA) can be worn for 65 9 35 9 8 mm
up to 3 days. Mealtime insulin can be dosed discreetly

Fig. 2 Schematic of the patch. The mechanically timed the valve button first returns into the ‘ out’’ position,
sequence ensures that the reservoir feature is never directly closing the pathway to the cannula and reconnecting the
connected to the user. Sequence: (1) Both the pump and insulin reservoir. (5) Finally, the pump button releases and
valve button are ‘ out’’. (2) Pressing the valve button simultaneously draws the next dose from the insulin
disconnects the insulin reservoir and opens the pathway to reservoir. This sequence is repeated for each subsequent
the cannula. (3) Consequently, the pump button depresses, dosing
delivering a 2-unit dose. (4) When releasing the buttons,

and removing the device in an open-book test of the trainer and were asked a series of ques-
format. Participants then conducted a practice tions to ensure their understanding and safe use
setup and use demonstration under supervision of the patch at home. In addition, all
Diabetes Ther (2018) 9:297–307 301

participants received the standard mealtime item (switch) asked if the participant would like
insulin-delivery system user guide (provided to switch to the patch with four response options:
with the product) containing complete 100, ‘‘definitely yes’’; 66.67, ‘‘probably yes’’;
instructions, including all warnings, frequently 33.33, ‘‘probably not’’; 0, ‘‘definitely not.’’
asked questions, and troubleshooting tips. Par-
ticipants were advised to carefully read the user Statistical Analysis
guide before beginning to use the patch.
Results are reported only for participants who
Analytical Measures completed evaluations for both their existing
and the new device. Descriptive data are repor-
The IDSRQ [22] was used to assess participant ted as percentage for categorical measures or
self-reported satisfaction with the insulin-deliv- mean [standard deviation (SD)] and range
ery system before and after wearing the patch. (min–max scores) for continuous measures.
Participants rated the systems on eight measures Scores on the six IDSRQ measures of conve-
in five domains: convenience, interference of nience, interference of treatment with daily
treatment with daily activities, diabetes-related activities, diabetes-related worry, positive well-
worries, psychological well-being (positive and being, negative well-being, and overall treat-
negative), and treatment system satisfaction/ ment-system satisfaction were analyzed by
preference (overall satisfaction, compare, repeated measures ANOVA comparing ratings of
switch). The treatment-efficacy subscale was not the patch with the ratings of respondents’ cur-
assessed as respondents did not use the study rent mealtime treatment system for those who
device to administer insulin to control blood used either (1) syringe, (2) pen, or (3) pump.
glucose. For the same reason, items related to Pooled ratings of the patch are reported for all
worry about the efficacy of the insulin-delivery participants; scores for those using each of the
systems were excluded from the measure of dia- current devices are reported as mean [standard
betes-related worries. Finally, the item regarding error (SE)] of the deviation from the score for
cost of the insulin-delivery system was excluded the patch (derived separately for respondents
from the convenience measure as there was no who used each particular current device); scores
cost for simulated use of the study device. less than zero indicate that the current meal-
The IDSRQ response options were ordinal time insulin-delivery device received a lower
scales from 1 to 4 or 1 to 5 and summary mea- score. Effect size was used to show the
sures were calculated as the mean of the items quantitative strength of each comparison
completed for four multi-item subscales (con- and was calculated as the difference between
venience, interference of treatment with daily the means of patch and current device
activities, positive well-being, negative well-be- divided by the SD for the rating of the patch
ing). All IDSRQ measures were converted to by all participants (Cohen’s d;
scores of 0–100 using standard arithmetic pro- d ¼ ðmean scoreðpatchÞmean scoreðcurrent deviceÞÞ
SDðpatchÞ ) [23]; a
cedure and the small number of missing values
positive value for d indicates preference for the
was handled as described by Peyrot and Rubin
patch over the current device. Cohen classified
to minimize loss of cases [22].
descriptors for the magnitude of d, where the
Two IDSRQ items asked respondents to rate
effect size of d = 0.2 is considered small, d = 0.5 is
the patch relative to their current treatment sys-
considered a medium effect, and d = 0.8 is con-
tem, and therefore were asked only once, after
sidered a large effect [23]. Thus, if two means
the participant had completed the simulated use
differ by no more than 0.2 SD, in the absence of
of the patch. The first item (compare) was an
objective criteria for assessing the difference it
overall comparison with five response options:
can be concluded that the difference is trivial
100, ‘‘new much better’’; 75, ‘‘new a bit better’’;
even if it is statistically significant. If the differ-
50, ‘‘both about the same’’; 25, ‘‘current a bit
ence is greater than 0.2 SD, it can be considered a
better’’; 0, ‘‘current much better.’’ The second
nontrivial change even if it is not statistically
302 Diabetes Ther (2018) 9:297–307

significant (i.e., because of a lack of statistical Perceptions of Devices

power). An effect size of d = 0.5 has been pro-
posed to represent the minimum detectable dif- The four multi-item composite IDSRQ measures
ference (MDD), a change that is considered all exhibited adequate reliability ([0.70) for the
recognizable to a careful observer; this concept ratings of both the current treatment system and
parallels that of a clinically meaningful differ- the patch: convenience (Cronbach’s a = 0.90/
ence (CMD) and can be used when there are no 0.94), interference of treatment with daily activi-
objective criteria for assessing CMD [24]. ties (a = 0.74/0.77), positive well-being (a = 0.77/
One-sample t tests were used to analyze the 0.81), and negative well-being (a = 0.78/0.77).
two IDSRQ items that asked respondents to rate Table 2 reports the results for comparisons of
the patch relative to their current treatment the patch with participants’ current mealtime
system; deviation of the mean from the mid- insulin-delivery system. Participants signifi-
point of the item scores was assessed for each of cantly favored the patch over syringe on all eight
the different current devices (a value of 50 rep- comparisons (p\0.05), with effect size ranging
resenting equivalence of the patch with the from 1.49 to 0.41 (median d = 0.70). Participants
current device). significantly favored patch over pen for five of
Two sets of ancillary analyses were performed eight comparisons (convenience, positive and
to identify potential correlates to overall satis- negative well-being, overall comparison and
faction with the patch (the latter measure was preference), with no comparisons favoring the
chosen since it is the summary outcome measure pen (effect size ranging from 0.72 to 0.27, median
that does not require splitting the sample into d = 0.47). Although the patch was favored over
subsets, which reduces the number of cases to the the pump on six of eight comparisons, only one
point at which correlational analysis is inappro- significantly favored the patch (negative well-
priate). The analyses examined associations being) and one significantly favored the pump
between overall satisfaction and (1) all patient (overall comparison); effect size ranged from 0.65
characteristics and (2) all domain-specific IDSRQ to - 0.51 (median d = 0.24).
measures. Preliminary analyses estimated Significantly more participants reported that
bivariate correlations; if multiple significant the patch was better overall vs their current
correlates were observed, multiple regression system for syringe (n = 37 vs n = 8) and pen
analysis was conducted to determine which fac- (n = 13 vs n = 4), but the opposite was true for
tors had significant independent associations the pump (n = 6 vs n = 18). Significantly more
with overall satisfaction with the patch. participants reported that they would like to
Statistical significance for all analyses was switch to the patch than continue using syringe
defined as p B 0.05, two-tailed. (n = 37 vs n = 10) or pen (n = 16 vs n = 5), but
the difference was not significant for the pump
(n = 14 vs n = 13).
RESULTS
Study Population Ancillary Analyses

A total of 101 participants completed the IDSRQ The analysis examining associations between
prior to and after simulated use of the patch. overall satisfaction and all patient characteris-
Baseline characteristics are summarized in tics showed that none were significantly related
Table 1. The study population primarily com- to overall satisfaction with the patch. The sec-
prised patients with T1D (71%) and individuals ond ancillary analysis showed that four of the
of middle age (mean age 46 years). The mean five domain-specific IDSRQ measures (all except
duration of insulin use was 20 years. Most worry) were related (p\0.01) with the expected
patients used a syringe (n = 49), with the sign to overall satisfaction with the patch; the
remainder divided between pens (n = 22) and absolute value of the five correlations ranged
CSII (n = 28). from 0.17 to 0.59, with a median of 0.27. When
Diabetes Ther (2018) 9:297–307 303

Table 1 Baseline characteristics showing a significant advantage for their cur-

rent insulin treatment system, specifically the
Characteristics Participants
pump, possibly because of the participants’
(N 5 101)
established active approaches to their diabetes
Gender, n (%) self-care [25]. After 3 days of simulated patch
Female 51 (50) use, 58% of syringe or pen users reported that,
in terms of overall satisfaction, the patch was
Male 50 (50) better than their current injection system and
Age, years 46 ± 17 (18–85) 12% indicated it was equal to their current
injection system. By contrast, 64% of pump
Type of diabetes, n (%)a users reported that the pump was better overall
T1D 72 (71) and 14% indicated it was equal to their current
system. Most participants (71%) reported that
T2D 29 (28)
they would like to switch to the patch, primarily
Age at diagnosis, years 25 ± 17 (1–73) in preference to syringe and pen.
Patch use resulted in significantly more
Duration of diabetes, years 21 ± 15 (1–67)
favorable participant responses for most sum-
Glucose monitoring (frequency per day), n (%)b mary scales when compared with the user’s
\4 times 22 (22) current injection device and was similar to the
current insulin-delivery system for all but one of
4–6 times 30 (30) the other comparisons. This reflects findings of
[6 times 30 (30) a real-life study into the use of this device in
people with T1D and T2D [11]. Participants
Insulin use, years 20 ± 15 (1–67) currently using an insulin pen or injecting
Current device, n (%)b insulin via a syringe had a more favorable
assessment of the patch, indicating a possible
Syringe 49 (49)
preference for an alternative system that does
Insulin pen 22 (22) not require multiple daily needle insertions for
mealtime insulin.
Insulin pump 28 (28)
These favorable responses to the patch may
Duration of use of current device, 11 ± 11 (1–56) have substantial implications for patient treat-
years ment adherence. For people with diabetes, bar-
riers to insulin therapy adherence include the
Values are shown as mean ± SD (range), unless indicated
perception that insulin therapy is complicated,
otherwise
fear of hypoglycemia, weight gain, and multiple
SD standard deviation, T1D type 1 diabetes, T2D type 2
injections, and the perception that insulin
diabetes
a
May not sum to 100% because of rounding therapy significantly restricts a normal lifestyle
b
May not sum to 100% because of missing values [9, 26]. However, the use of a device that would
facilitate discreet insulin delivery and reduce
the number of needle–skin penetrations may
all variables were entered into a stepwise offer a solution to some of these barriers
regression analysis, only one (convenience) had [27–31]. Compared with traditional MDI insulin
an independent association (r-squared = 0.34, regimens, an individual (eating three meals a
p\0.001). day) using the patch can potentially decrease
the number of mealtime device needle–skin
penetrations over 3 days from nine (or more) to
DISCUSSION one.
This study did not assess objective glucose
The patch was well received by the study par-
control or perceived clinical efficacy; however,
ticipants, with only one of 24 comparisons
greater treatment satisfaction (as reported in
304 Diabetes Ther (2018) 9:297–307

Table 2 Summary of results for device comparison by standardized IDSRQ scores

Measure (number of items) Patcha Syringeb Penb Pumpb
Conveniencec (11) 78.0 ± 16.9 - 25.3 ± 4.0*** (1.49) - 11.9 ± 3.3** (0.70) - 4.1 ± 3.6 (0.24)
d
Interference (7) 13.4 ± 15.2 13.4 ± 3.72*** (0.88) 7.3 ± 5.3 (0.48) 9.9 ± 5.0 (0.65)
Worryd (1) 37.6 ± 27.3 18.4 ± 4.1*** (0.67) 10.7 ± 6.8 (0.39) 8.3 ± 5.0 (0.30)
c
Positive well-being (5) 67.2 ± 20.3 - 11.0 ± 2.5*** (0.54) - 8.3 ± 2.0*** (0.41) - 4.8 ± 3.0 (0.24)
Negative well-beingd (4) 36.7 ± 19.5 8.0 ± 2.3*** (0.41) 9.2 ± 4.3* (0.47) 9.8 ± 3.7* (0.50)
c
Overall satisfaction (1) 69.4 ± 23.4 - 21.8 ± 6.6*** (0.93) - 6.3 ± 9.1 (0.27) 11.9 ± 6.2 (- 0.51)
Device comparisonc,e (1) NA - 25.0 ± 4.3*** (0.73) - 20.2 ± 6.4** (0.59) 14.3 ± 6.1* (- 0.42)
c,e
Switching to patch (1) NA - 19.5 ± 4.9*** (0.61) - 23.0 ± 5.9*** (0.72) - 5.6 ± 5.9 (0.18)
All measures were scored on a scale of 0–100 to standardize, as described in the original article [22]
NA no independent value for the patch because ratings compare devices, IDSRQ Insulin Delivery System Rating Ques-
tionnaire, SD standard deviation, SE standard error
* p B 0.05; ** p B 0.01; *** p B 0.001
a
Data for the patch are shown as mean ± SD
b
Data for syringe, pen, and pump users are shown as mean difference ± SE relative to the patch (Cohen’s d, effect size
shown as pooled SD units for the patch)
c
Higher score is better
d
Lower score is better
e
Test assumes value of 50, representing equal scores for both devices; difference represents deviation from 50

this study) has been associated with better gly- conclusions regarding long-term patient-re-
cemic outcomes [32]. In addition, a clinical ported outcomes to be made. In addition, only
study by Bohannon et al. [11] reported that user evaluation data was collected, so it was not
participants using the study device had equiv- possible to clinically determine whether use of
alent mean daily blood glucose and signifi- the patch would impact medication adherence
cantly less glycemic variability when compared or glucose control.
with using their pen or syringe. The safety The study cohort was modest in size, but for
profile of this patch has been shown to be the comparison between the current insulin-
similar to currently available injection devices delivery device and the patch, only one effect
for mealtime insulin [11]. Current literature size larger than the MDD did not reach signifi-
indicates that more user-friendly and safe-to-use cance (d = 0.51, p = ns), and some effect sizes
devices, such as insulin pens or pumps, can lead less than MDD did reach significance (e.g.,
to improved treatment satisfaction and gly- d = 0.51, p[0.001). This suggests that sample
cemic control compared with using a syringe for size was not a problem, as meaningful differ-
insulin delivery [28, 30]. ences in patient-reported outcomes should be
This study had several limitations. The use of recognizable by those making the ratings.
the patch was simulated; this might create bias Patient characteristics were not significantly
toward less favorable patient-reported outcomes correlated with the key rating of the patch,
because participants, in addition to using the suggesting that observed perceptions of the
patch, also had to perform insulin administra- study device are robust and it was unnecessary
tion with their existing mealtime insulin-deliv- for our analysis to control for case-mix vari-
ery system. Furthermore, the short duration ables. However, future research should examine
(3 days) of simulated use did not allow for the degree to which patient characteristics,
Diabetes Ther (2018) 9:297–307 305

including regimen intensity, are related to per- with pen or syringe. Research is needed to
ceptions of the study device in actual clinical examine this possibility.
practice.
It also is worth noting that the device is
currently labeled to be worn on the abdomen. ACKNOWLEDGEMENTS
In addition, although the age distribution of
participants in this study was 18–85 years, the
device is currently labeled for use by patients Funding. The studies were funded by Calibra
over 21 years of age. Medical. The study sponsor, Calibra Medical
Physicians may be reluctant to initiate or (Wayne PA), was involved in all stages of
intensify insulin treatment in people with T2D research and manuscript preparation; all
despite recognizing the need for it (a phe- authors had access to the data, and contributed
nomenon known as ‘‘clinical inertia’’). Physi- to the manuscript development. Calibra Medi-
cians are concerned that patients will be unable cal funded the article processing charges.
to follow the insulin-treatment regimen [33].
There is also a concern that training people with Authorship. All named authors meet the
diabetes to properly calculate insulin doses and International Committee of Medical Journal
administer insulin correctly is too complicated Editors (ICMJE) criteria for authorship for this
and time-consuming [34]. Future studies should manuscript, take responsibility for the integrity
examine physicians’ reactions to the study of the work as a whole, and have given final
device to determine whether the use of a system approval for the version to be published. The
that is discreet to wear and use and requires authors would like to thank Brian L. Levy, MD,
fewer daily needle–skin penetrations might and Ramachandra G. Naik, MD for their critical
have an impact on prescribing behaviors and feedback.
reduce clinical inertia.
Medical Writing Assistance. Editorial and
writing support in the preparation of this
CONCLUSIONS manuscript was provided by Martina Fuchs-
berger, PhD, of Excerpta Medica, funded by
The patch evaluated here is a small and easy-to- Calibra Medical.
use device that allows for discreet delivery of
mealtime insulin. While the system does not Disclosures. Mark Peyrot has received con-
eliminate the need for daily basal-insulin sulting fees from Calibra Medical. Meng H. Tan
injections, it is intended to assist in mealtime has received consulting fees from Calibra Med-
dosing by providing an alternative to the mul- ical. Darlene Dreon is a full-time employee of
tiple daily injections required with syringe or Calibra Medical. Vivien Zraick is a former full-
pen devices. Rather than having to penetrate time employee of Calibra Medical. Brett Cross is
the skin multiple times per day, application of a former full-time employee of Calibra Medical.
the patch requires only one skin penetration
every 3 days. Given that a mealtime dose of Compliance with Ethics Guidelines. All
insulin from this device is not associated with a procedures followed were in accordance with
needle–skin penetration, use of the patch may the ethical standards of the appropriate com-
help increase patient adherence to multiple- mittees on human experimentation (institu-
daily-insulin-injection regimens, not only tional and national) and with the Helsinki
because of user satisfaction with the patch but Declaration of 1964. Informed consent was
also reductions in the patient burden associated obtained from all patients for inclusion in the
with MDI. Use of this patch may lead to better study.
control of prandial hyperglycemic excursions
and improved HbA1c in those people with dia- Data Availability. The datasets generated
betes who miss mealtime insulin injections and/or analyzed during the current study are
306 Diabetes Ther (2018) 9:297–307

available from the corresponding author on 8. Karter AJ, Subramanian U, Saha C, et al. Barriers to
reasonable request. insulin initiation: the translating research into
action for diabetes insulin starts project. Diabetes
Care. 2010;33:733–5.
Open Access. This article is distributed
under the terms of the Creative Commons 9. Peyrot M, Rubin RR, Kruger DF, Travis LB. Corre-
Attribution-NonCommercial 4.0 International lates of insulin injection omission. Diabetes Care.
2010;33:240–5.
License (http://creativecommons.org/licenses/
by-nc/4.0/), which permits any noncommercial 10. Peyrot M, Barnett AH, Meneghini LF, Schumm-
use, distribution, and reproduction in any Draeger PM. Insulin adherence behaviours and
medium, provided you give appropriate credit to barriers in the multinational Global Attitudes of
Patients and Physicians in Insulin Therapy study.
the original author(s) and the source, provide a
Diabet Med. 2012;29:682–9.
link to the Creative Commons license, and
indicate if changes were made. 11. Bohannon N, Bergenstal R, Cuddihy R, et al.
Comparison of a novel insulin bolus-patch with
pen/syringe injection to deliver mealtime insulin
for efficacy, preference, and quality of life in
adults with diabetes: a randomized, crossover,
REFERENCES multicenter study. Diabetes Technol Ther.
2011;13:1031–7.
1. American Diabetes Association. Standards of medi-
cal care in diabetes. 5. Glycemic targets. Diabetes 12. Rubin RR. Adherence to pharmacologic therapy in
Care. 2016;39(Suppl 1):S39–46. patients with type 2 diabetes mellitus. Am J Med.
2005;118(Suppl 5A):S27–34.
2. Inzucchi S, Bergenstal R, Buse J, et al. Management
of hyperglycaemia in type 2 diabetes, 2015: a 13. Kruger DF, LaRue S, Estepa P. Recognition of and
patient-centred approach. Update to a position steps to mitigate anxiety and fear of pain in
statement of the American Diabetes Association injectable diabetes treatment. Diabetes Metab
and the European Association for the Study of Syndr Obes. 2015;8:49–56.
Diabetes. Diabetologia. 2015;58:429–42.
14. Giugliano D, Maiorino MI, Bellastella G, Chiodini
3. Imran SA, Rabasa-Lhoret R, Ross S. Canadian Dia- P, Ceriello A, Esposito K. Efficacy of insulin analogs
betes Association 2013 clinical practice guidelines in achieving the hemoglobin A1c target of \ 7% in
for the prevention and management of diabetes in type 2 diabetes: meta-analysis of randomized con-
Canada: targets for glycemic control. Can J Dia- trolled trials. Diabetes Care. 2011;34:510–7.
betes. 2013;37(Suppl 1):S31–4.
15. Vinik AI, Kalk WJ, Botha JL, Jackson WP, Blake KC.
4. The Diabetes Control and Complications Trial The inexhaustible beta cell. Diabetes. 1976;25:11–5.
(DCCT) Research Group. The effect of intensive
treatment of diabetes on the development and 16. Service FJ, Hall LD, Westland RE, et al. Effects of
progression of long-term complications in insulin- size, time of day and sequence of meal ingestion on
dependent diabetes mellitus. N Engl J Med. carbohydrate tolerance in normal subjects. Dia-
1993;329:977–86. betologia. 1983;25:316–21.

5. UK Prospective Diabetes Study (UKPDS) Group. 17. Cerasi E, Luft R. The plasma insulin response to
Intensive blood-glucose control with sulphony- glucose infusion in healthy subjects and in diabetes
lureas or insulin compared with conventional mellitus. Acta Endocrinol (Copenh).
treatment and risk of complications in patients 1967;55:278–304.
with type 2 diabetes (UKPDS 33). Lancet.
1998;352:837–53. 18. Del Prato S. Loss of early insulin secretion leads to
postprandial hyperglycaemia. Diabetologia.
6. National Diabetes Statistics Report. 2014. http:// 2003;46(Suppl 1):M2–8.
www.cdc.gov/diabetes/pubs/statsreport14/national-
diabetes-report-web.pdf. Accessed Sept 15, 2016. 19. Monnier L, Lapinski H, Colette C. Contributions of
fasting and postprandial plasma glucose increments
7. Roussel R, Charbonnel B, Behar M, Gourmelen J, to the overall diurnal hyperglycemia of type 2 dia-
Emery C, Detournay B. High rates of early discon- betic patients: variations with increasing levels of
tinuation of insulin therapy in type 2 diabetes: a HbA(1c). Diabetes Care. 2003;26:881–5.
nationwide study. Diabetes. 2015;64(Suppl 1):A252.
Diabetes Ther (2018) 9:297–307 307

20. Varshney HM, Kumar R, Mohan S. Novel approa- 28. Klausmann G, Hramiak I, Qvist M, Mikkelsen KH,
ches for insulin delivery: current status. Int J Ther Guo X. Evaluation of preference for a novel durable
Appl. 2012;7:25–31. insulin pen with memory function among patients
with diabetes and health care professionals. Patient
21. Anhalt H, Bohannon NJV. Insulin patch pumps: Prefer Adherence. 2013;7:285–92.
their development and future in closed-loop sys-
tems. Diabetes Technol Ther. 2010;12(Suppl 29. Mader JK, Lilly LC, Aberer F, et al. A feasibility study
1):S51–8. of a 3-day basal-bolus insulin delivery device in
individuals with type 2 diabetes. Diabetes Care.
22. Peyrot M, Rubin RR. Validity and reliability of an 2014;37:1476–9.
instrument for assessing health-related quality of
life and treatment preferences: the Insulin Delivery 30. Barnard KD, Bromba M, de Lange M, et al. High
System Rating Questionnaire. Diabetes Care. reported treatment satisfaction in people with type
2005;28:53–8. 1 diabetes switching to latest generation insulin
pump regardless of previous therapy. J Diabetes Sci
23. Cohen J. Statistical power analysis for the behav- Technol. 2015;9:231–6.
ioral sciences. 2nd ed. Hillsdale: Lawrence Earl-
baum Associates; 1988. 31. Berard L, Cameron B, Woo V. Pen needle preference
in a population of Canadians with diabetes: results
24. Norman GR, Sloan JA, Wyrwich KW. Interpretation from a recent patient survey. Can J Diabetes.
of changes in health-related quality of life: the 2015;39:206–9.
remarkable universality of half a standard devia-
tion. Med Care. 2003;41:582–92. 32. Hanestad BR, Albreksten G. Quality of life, per-
ceived difficulties in adherence to diabetes regimen,
25. Ritholz MD, Smaldone A, Lee J, Castillo A, Wolpert and blood glucose control. Diabet Med.
H, Weinger K. Perceptions of psychosocial factors 1991;8:759–64.
and the insulin pump. Diabetes Care.
2007;30:549–54. 33. Polinski JM, Smith BF, Curtis BH, et al. Barriers to
insulin progression among patients with type 2
26. Summers KH, Szeinbach SL, Lenox SM. Preference diabetes: a systematic review. Diabetes Educ.
for insulin delivery systems among current insulin 2013;39:53–65.
users and nonusers. Clin Ther. 2004;26:1498–505.
34. Sorli C, Heile MK. Identifying and meeting the
27. Kerr D, Hoogma RP, Buhr A, Petersen B, Storms FE, challenges of insulin therapy in type 2 diabetes.
Study investigators. Multicenter user evaluation of J Multidiscip Healthc. 2014;7:267–82.
ACCU-CHEKÒ Combo, an integrated system for
continuous subcutaneous insulin infusion. J Dia-
betes Sci Technol. 2010;4:1400–7.