You are on page 1of 6

Clinical Care/Education/Nutrition/Psychosocial Research

O R I G I N A L A R T I C L E

Correlates of Insulin Injection Omission


MARK PEYROT, PHD1,2 DAVIDA F. KRUGER, MSN, APN-BC, BC-ADM4 the patient’s emotional well-being, and
RICHARD R. RUBIN, PHD, CDE2,3 LUTHER B. TRAVIS, MD, FAAP5 the patient’s perceptions of medication
side effects and medication-related intru-
sions on activities of daily living.
OBJECTIVE — The purpose of this study was to assess factors associated with patient fre- Medication costs can affect adherence.
quency of intentionally skipping insulin injections. In a U.S. survey of adults with type 2 diabe-
tes using glucose-lowering agents, 11% re-
RESEARCH DESIGN AND METHODS — Data were obtained through an Internet
survey of 502 U.S. adults self-identified as taking insulin by injection to treat type 1 or type 2
ported that they had cut back on their
diabetes. Multiple regression analysis assessed independent associations of various demo- medication in the past year (7). Adherence
graphic, disease, and injection-specific factors with insulin omission. rates are also affected by regimen complex-
ity. Rates of adherence for oral diabetes
RESULTS — Intentional insulin omission was reported by more than half of respondents; medication decline with the number of
regular omission was reported by 20%. Significant independent risk factors for insulin omission times each day the medication should be
were younger age, lower income and higher education, type 2 diabetes, not following a healthy
taken (8), but we could find no report on
diet, taking more daily injections, interference of injections with daily activities, and injection
pain and embarrassment. Risk factors differed between type 1 and type 2 diabetic patients, with the association between insulin injection
diet nonadherence more prominent in type 1 diabetes and age, education, income, pain, and frequency and insulin omission. Depres-
embarrassment more prominent in type 2 diabetes. sion also has been associated with diabetes
medication nonadherence (9) and with in-
CONCLUSIONS — Whereas most patients did not report regular intentional omission of sulin omission among adolescent females
insulin injections, a substantial number did. Our findings suggest that it is important to identify (10).
patients who intentionally omit insulin and be aware of the potential risk factors identified here.
For patients who report injection-related problems (interference with daily activities, injection Several researchers have developed
pain, and embarrassment), providers should consider recommending strategies and tools for questionnaires to assess patient percep-
addressing these problems to increase adherence to prescribed insulin regimens. This could tions of insulin therapy that could affect
improve clinical outcomes. regimen adherence (11–16). These ques-
tionnaires assess factors such as 1) inter-
Diabetes Care 33:240–245, 2010 ference with eating, exercise, and
activities of daily living; 2) dissatisfaction
with the amount of time required to admin-

M
ore than 25% of people with dia- sulin use was 77% of prescribed amounts
betes take insulin (1). The Amer- (3). In this population, two-thirds of ister insulin and with injection-related pain,
ican Diabetes Association and the whom were ⱖ65 years of age, age did not bruising, and embarrassment; 3) worries
European Association for the Study of Di- predict insulin regimen adherence, and about insulin-related side effects such as hy-
abetes recently issued a consensus algo- adherence was nearly identical for men poglycemia; and 4) negative affect associ-
rithm for management of type 2 diabetes and women, but non-Hispanic white pa- ated with administering insulin. Un-
identifying insulin as the most effective tients were more adherent than patients fortunately, none of these studies formally
glucose-lowering agent (2). Lower com- who were African American or Hispanic. assessed the association between any of
pliance with insulin regimens is associ- In another study of patients who switched these factors and intentional insulin omis-
ated with higher A1C levels (3,4) and from using a syringe to deliver insulin to sion. However, a recent publication did find
with higher rates of hospital admissions using a pen, 36% of the patients had med- that insulin adherence is lower among
for diabetes-related complications (3). ication possession rates of ⬎80% while young women who are concerned about
Despite the importance of adhering to using a syringe, but this rose to 55% after their weight (17).
prescribed insulin regimens, little is switching to a pen (5). The current study is designed to ad-
known about the degree to which patients A recent review identified factors as- dress questions about intentional insulin
are adherent or about factors associated sociated with adherence to any diabetes omission, including the frequency of this
with adherence. In a study using a De- medication (6). These factors include behavior and factors hypothesized to be
partment of Veterans Affairs database, in- medication costs, regimen complexity, associated with this behavior, in a large
● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● sample of patients weighted to be repre-
From the 1Department of Sociology, Loyola University Maryland, Baltimore, Maryland; the 2Department of sentative of all adult diabetic patients in
Medicine, Johns Hopkins University, Baltimore, Maryland; the 3Department of Pediatrics, Johns Hopkins the U.S. who take insulin. In a previous
University, Baltimore, Maryland; the 4Division of Endocrinology, Diabetes, Bone and Mineral Disorders, report, we found that most insulin-
Henry Ford Health Systems, Detroit, Michigan; and the 5Department of Pediatrics, University of Texas treated patients wanted to reduce the
Medical Branch, Galveston, Texas.
Corresponding author: Mark Peyrot, mpeyrot@loyola.edu.
number of insulin injections they take
Received 22 July 2009 and accepted 19 October 2009. each day, and some reported that injec-
DOI: 10.2337/dc09-1348 tion-related problems affect the number
© 2010 by the American Diabetes Association. Readers may use this article as long as the work is properly of injections they are willing to take (18).
cited, the use is educational and not for profit, and the work is not altered. See http://creativecommons.
org/licenses/by-nc-nd/3.0/ for details.
Here we formally assess the impact of a
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby broad range of factors that might be ex-
marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. pected to influence intentional insulin
See accompanying editorial, p. 450. omission, including demographic and

240 DIABETES CARE, VOLUME 33, NUMBER 2, FEBRUARY 2010 care.diabetesjournals.org


Peyrot and Associates

disease factors, as well as perceived bur- jection time needed, ease of use, pain, total of 77% of the sample said they had
den of insulin therapy (i.e., interference inflammation/bruising, and embarrass- type 2 diabetes, and the rest said they had
with activities of daily living), the injec- ment— each measured by a single item type 1 diabetes; patients reported having
tion experience (e.g., pain and embarrass- (response options: 1 ⫽ very satisfied, 2 ⫽ diabetes for an average of almost 15 years.
ment), and negative emotions (e.g., satisfied, 3 ⫽ somewhat satisfied, 4 ⫽ not A total of 61% of the patient sample
dread) associated with insulin injections. at all satisfied). identified a primary care physician as
Negative affect toward insulin injec- their primary diabetes health care pro-
RESEARCH DESIGN AND tions was measured as the mean of three vider, whereas 28% named an endocri-
METHODS — Data were obtained items: “I dread insulin injections”; “Inject- nologist and 11% named another
through an Internet survey of U.S. adults ing myself with insulin is the hardest part (nonphysician) health care provider. Of
self-identified as taking insulin to treat of managing my diabetes”; “I have to men- the sample, 39% reported engaging in
type 1 or type 2 diabetes; the survey was tally prepare myself before each injection” physical activity and 55% said they fol-
conducted 13 June to 7 July 2008 by Har- (response options: 1 ⫽ strongly disagree, lowed a healthy diet. A total of 70% of
ris Interactive, a contract research organi- 2 ⫽ somewhat disagree, 3 ⫽ somewhat patients surveyed said they took insulin
zation. The sample was drawn from the agree, 4 ⫽ strongly agree). The reliability using a syringe and 30% said they used
Harris Interactive Chronic Illness Panel. of this scale was moderate (␣ ⫽ 0.85). a pen; most (56%) changed their needle
Patients were recruited by email if they Worry about hypoglycemia was mea- with each injection. Patients reported
had diabetes and currently used a syringe sured by a single item (response options: taking an average of 2.7 injections a day
or insulin pen to deliver insulin. The re- 1 ⫽ never, 2 ⫽ rarely, 3 ⫽ sometimes, (maximum of five recorded).
cruitment quota was 500 participants. 4 ⫽ often). A substantial minority of respondents
Institutional review board approval Frequency of intentional insulin omis- (22%) said they planned their daily activ-
for the study protocol was obtained from sion. The dependent variable in this ities around their insulin injections, and
the Human Subject Research Committee study was the response to the question, similar proportions reported that insulin
of Loyola University Maryland. “How often do you skip insulin injections injections interfered with their lives: 23%
that you know you should take?” (re- said insulin injections interfered with
Measures sponse options: 1 ⫽ never, 2 ⫽ rarely, their eating/exercising schedule more
Data collected from participants included 3 ⫽ sometimes, 4 ⫽ often). than a little, and 25% said that insulin
the following: 1) basic demographic infor- injections had a negative effect on one or
mation; 2) disease type, duration, complica- Statistical analysis more activity of daily living. Further, a
tions, and treatment; 3) perceived burden of The sample was weighted to be represen- substantial minority of respondents
insulin injections; 4) the experience of in- tative of the U.S. population of people (22%) reported they had to mentally pre-
jections; 5) negative affect toward insulin in- with diabetes. Multiple regression analy- pare themselves before each injection,
jections; and 6) frequency of skipping sis was used to assess independent rela- and 33% identified they had some level of
insulin injections. Respondents reported tionships with frequency of skipping dread associated with taking their daily
whether they had ever been diagnosed with insulin injections. Control variables (de- injections.
type 1 or type 2 diabetes, depression, obe- mographic and disease characteristics) Attitudinal measures tended to fall
sity, or cardiovascular disease (“high blood were entered first, and then injection- below the halfway point of the response
pressure” or “heart disease”) and whether related experience and attitudes were en- options (i.e., ⬍2.5). Respondents re-
they treat their diabetes with diet, exercise, tered using stepwise criteria (P ⬍ 0.05). ported moderate levels of satisfaction
and medications. Other measures are de- Separate analyses were performed in the with the pain and the inflammation and
scribed below. type 1 and type 2 diabetic populations to bruising associated with insulin injec-
Burden of injections. Interference with see whether associations differed between tions (the scores for pain and inflamma-
eating and exercise was measured as the populations. tion/bruising were significantly higher
mean of two items asking, “How much than those for embarrassment, time
does the way you inject insulin interfere RESULTS needed, and ease of use, P ⬍ 0.001). A
with eating/exercising when you want?” quarter (24%) of respondents had a score
(response options: 1 ⫽ not at all, 2 ⫽ a Sample profile representing negative affect toward injec-
little, 3 ⫽ a moderate amount, 4 ⫽ a great The sample (n ⫽ 502) was 55% male, tions (they scored above the midpoint on
deal). The reliability of this scale was 73% white, 11% Hispanic, 11% African the scale), and 21% reported “often” wor-
moderate (␣ ⫽ 0.80). Interference with American, and 5% other race/ethnicity, rying about hypoglycemia.
activities of daily living was measured as a with a mean age of 55 years (Table 1).
count of the affirmative responses to the About half (51%) had attended college. Regression analysis
question, “Do your insulin injections have Only about one-third (38%) were pres- Over half (57%) of respondents reported
a negative effect on: social activities, rec- ently employed, and those who were not skipping insulin injections they knew
reational activities, sexual activity, work/ employed included 8% students and 8% they should take; 20% report skipping
career, family care-giving?” (possible disabled; the remainder were mostly re- them sometimes or often. Table 2 shows
range ⫽ 0 –5). Another measure of inter- tired or nonworking spouses. Median an- the results of the regression analysis of
ference was whether the respondent plans nual income of the sample was about intentional insulin omission frequency.
daily activities around insulin injections $35,000. Control variables (demographic and dia-
(1 ⫽ yes, 0 ⫽ no). Approximately one-third (32%) re- betes characteristics) accounted for 26%
Experience of injections. There were ported having been diagnosed by a health of the variance in intentional insulin
five measures—dissatisfaction with in- care professional as having depression. A omission. Older respondents, those who

care.diabetesjournals.org DIABETES CARE, VOLUME 33, NUMBER 2, FEBRUARY 2010 241


Correlates of insulin injection omission

Table 1—Sample profile

All respondents Type 1 diabetes Type 2 diabetes


N 502 114 388
Sex
M 55 41% 59%
F 45 59% 41%
Race/ethnicity
White 73 70% 74%
Hispanic 11 19% 9%
African American 11 8% 12%
Other 5 3 5
Current age (years) 54.9 ⫾ 13.9 46.9 ⫾ 15.6 57.5 ⫾ 12.2
Education
No college 49 24 44
Some college 28 38 26
College graduate 14 17 13
Graduate school 9 21 17
Employment
Full-time employment 28 33 27
Part-time employment 10 16 8
Not employed (student) 8 14 7
Not employed (disabled) 8 2 10
Other 46 35 48
Household income ($)
⬍15,000 23 23 23
15,000–24,999 11 5 12
25,000–34,999 10 10 10
35,000–49,999 13 11 14
50,000–74,999 16 12 17
ⱖ75,000 13 25 10
No answer 14 14 14
History of depression 32 30 33
History of cardiovascular disease 71 58 75
History of obesity 36 20 41
Duration of diabetes (years) 14.8 ⫾ 10.2 21.4 ⫾ 13.0 12.8 ⫾ 8.3
Diabetes care provider
Primary care physician 61 46 66
Endocrinologist 28 48 22
Nonphysician 11 6 12
Engage in physical activity 39 55 34
Follow healthy diet 55 65 52
Insulin injection device
Pen 30 33 29
Syringe 70 67 71
Change needle each use 56 50 58
Daily injection frequency 2.7 ⫾ 1.4 3.5 ⫾ 1.3 2.5 ⫾ 1.3
Plan daily activities around insulin injections 22 32 19
Interference with eating and exercise* 1.7 ⫾ 0.8 1.9 ⫾ 0.8 1.6 ⫾ 0.8
Interference with activity of daily living* 0.5 ⫾ 1.0 0.7 ⫾ 1.3 0.5 ⫾ 1.0
Dissatisfaction with time needed for injection† 2.0 ⫾ 0.9 1.9 ⫾ 0.9 2.1 ⫾ 0.9
Dissatisfaction with injection ease of use† 2.0 ⫾ 0.9 1.9 ⫾ 0.9 2.0 ⫾ 0.9
Dissatisfaction with injection pain† 2.3 ⫾ 0.9 2.2 ⫾ 0.9 2.3 ⫾ 0.9
Dissatisfaction with injection inflammation/bruising† 2.4 ⫾ 1.0 2.4 ⫾ 0.9 2.5 ⫾ 1.0
Dissatisfaction with injection embarrassment† 2.0 ⫾ 0.9 2.1 ⫾ 0.9 2.0 ⫾ 0.9
Negative affect toward injections‡ 1.8 ⫾ 0.9 1.6 ⫾ 0.7 1.9 ⫾ 0.9
Worry about hypoglycemia* 2.7 ⫾ 0.9 2.8 ⫾ 1.0 2.7 ⫾ 0.9
Skip insulin injections§ 1.8 ⫾ 0.8 1.7 ⫾ 0.7 1.8 ⫾ 0.8
Data are % or means ⫾ SD, unless otherwise stated. *1 ⫽ not at all, 2 ⫽ a little, 3 ⫽ a moderate amount, 4 ⫽ a great deal. †1 ⫽ very satisfied, 2 ⫽ satisfied, 3 ⫽
somewhat satisfied, 4 ⫽ not at all satisfied. ‡1 ⫽ strongly disagree, 2 ⫽ somewhat disagree, 3 ⫽ somewhat agree, 4 ⫽ strongly agree. §1 ⫽ never, 2 ⫽ rarely, 3 ⫽
sometimes, 4 ⫽ often.

242 DIABETES CARE, VOLUME 33, NUMBER 2, FEBRUARY 2010 care.diabetesjournals.org


Peyrot and Associates

Table 2—Multiple regression analysis of frequency of insulin injection omission diabetes and is common in ⬃20% of these
individuals. Intentional insulin omission
B SE ␤ Significance among adults varies with a number of de-
mographic and disease characteristics. It
Female sex* ⫺0.045 0.067 ⫺0.028 0.499 also is associated with indicators of per-
White race/ethnicity† ⫺0.006 0.073 ⫺0.003 0.936 ceived burden and the experience of injec-
Current age (years) ⫺0.015 0.003 ⫺0.256 ⬍0.001 tions as painful and embarrassing.
Education‡
Some college ⫺0.001 0.076 ⫺0.001 0.990 Demographic and disease factors
College graduate ⫺0.023 0.100 ⫺0.010 0.818 We found that respondents with higher
Graduate school 0.305 0.121 0.109 0.012 household income, but not individuals
Employment§ with more education, were less likely to
Not employed (student) 0.273 0.125 0.094 0.030 skip insulin injections they knew they
Not employed (disabled) ⫺0.317 0.120 ⫺0.107 0.009 should take. This may reflect easier access to
Part-time employment ⫺0.136 0.104 ⫺0.051 0.192 medications and supplies among individu-
Full-time employment ⫺0.108 0.078 ⫺0.061 0.166 als with higher income, but it is also likely
Household income ($) ⫺0.068 0.017 ⫺0.176 ⬍0.001 that higher socioeconomic status is associ-
History of depression 0.048 0.072 0.028 0.504 ated with more access to diabetes educa-
History of cardiovascular disease ⫺0.028 0.072 ⫺0.016 0.699 tion, higher health literacy, greater control
History of obesity ⫺0.062 0.070 ⫺0.038 0.371 over one’s daily routines, and better prob-
Type 2 diabetes㛳 0.428 0.092 0.226 ⬍0.001 lem-solving skills (19). Our study appears
Duration of diabetes 0.001 0.004 0.019 0.686 to be among the first to identify an associa-
Diabetes care provider¶ tion between socioeconomic status and in-
Endocrinologist ⫺0.085 0.077 ⫺0.048 0.270 sulin omission. Future research should seek
Nonphysician ⫺0.169 0.102 ⫺0.066 0.096 to identify potential mediators of this rela-
Engage in physical activity ⫺0.093 0.071 ⫺0.057 0.186 tionship, i.e., what links lower socioeco-
Follow healthy diet ⫺0.157 0.069 ⫺0.098 0.023 nomic status to insulin omission.
Use insulin pen# ⫺0.001 0.074 ⫺0.001 0.989 Contrary to earlier reports (3,4), we
Needle change frequency ⫺0.049 0.070 ⫺0.031 0.479 found no racial/ethnic differences in inten-
Daily injection frequency 0.118 0.025 0.205 ⬍0.001 tional insulin omission. This may be be-
Plan daily activities around insulin injections 0.247 0.079 0.129 0.002 cause we did not have enough nonwhite
Interference with activity of daily living 0.112 0.033 0.146 0.001 respondents to examine the different racial/
Dissatisfaction with injection pain 0.130 0.042 0.149 0.002 ethnic groups separately. Alternatively, this
Dissatisfaction with injection embarrassment 0.146 0.041 0.169 ⬍0.001 may be due to our controlling for income
*Reference category ⫽ male. †Reference category ⫽ nonwhite. ‡Reference category ⫽ no college. §Reference and education in the analysis, thereby elim-
category ⫽ other not employed. 㛳Reference category ⫽ type 1 diabetes. ¶Reference category ⫽ primary care inating the confounding of race/ethnicity
physician. #Reference category ⫽ use syringe. with socioeconomic status.
Much prior research has suggested that
intentional insulin omission is common
were disabled, those with higher house- related pain or embarrassment intention- among female adolescents with type 1 dia-
hold income, and those who followed a ally skipped insulin injections more often. betes, serving as a weight control strategy
healthy diet were significantly less likely Separate analyses using the variables and sometimes linked to eating disorders
to skip injections, whereas subjects who listed in Table 2 were performed among (10). We found that students (who were
were students and had the highest educa- subjects with type 1 diabetes and subjects younger than nonstudents) were more
tion or type 2 diabetes and subjects who with type 2 diabetes (results not shown). likely to skip injections they knew they
took more injections were significantly Because there were more participants should take, but this behavior was not more
more likely to skip injections. with type 2 diabetes, the overall model common among women than it was among
Four measures of injection burden most closely resembled that for type 2 di- men. We found no overall association be-
and experience had significant indepen- abetes. Only two variables significant in tween age and intentional insulin injection
dent associations with higher levels of in- the overall model were not significant in omission among patients with type 1 diabe-
tentional insulin omission and accounted the type 2 model— being a student and tes, suggesting that patients with type 1 di-
for an additional 10% of the variance. following a healthy diet. Only three vari- abetes “age-out” of this behavior by early
These included two aspects of interfer- ables were significant in the type 1 mod- adulthood, when they complete their edu-
ence (planning one’s activities around in- el—following a healthy diet, number of cation. Ascertaining the validity of this in-
sulin injections and injections interfering daily insulin injections, and interference terpretation would require following youth
with activities of daily living) and two as- with activities of daily living (being a stu- with type 1 diabetes as they age into adult-
pects of the injection process itself (pain dent had a P value of 0.056). hood to determine change in rate of insulin
and embarrassment). Respondents who omission.
planned daily activities around insulin in- CONCLUSIONS — This study sug- Our finding that, among individuals
jections, those who said that taking injec- gests that intentional omission of insulin with type 2 diabetes, older respondents
tions interfered with activities of daily injections that should be taken occurs in the were less likely to skip insulin injections is
living, and those who reported injection- majority of adults using insulin to treat their consistent with earlier studies (3,4). This

care.diabetesjournals.org DIABETES CARE, VOLUME 33, NUMBER 2, FEBRUARY 2010 243


Correlates of insulin injection omission

suggests that there are parallel aging-out reasons, including their receiving more experience of injecting insulin as painful
processes among individuals with type 1 assistance with care, or making a greater and embarrassing (but not dissatisfaction
and type 2 diabetes, but in type 2 diabe- effort to compensate for poor health. with time needed, ease of use, or skin in-
tes, this process takes place later in the life Two aspects of patients’ treatment flammation/bruising). There are numer-
course (almost all people with type 2 di- regimens were associated with increased ous device-related strategies for reducing
abetes are diagnosed as adults). Ascertain- insulin omission—respondents who took pain and embarrassment, including insu-
ing the validity of this interpretation more injections each day and those who lin pens, finer gauge needles, injection
would require following adults with type did not follow a healthy diet were more ports, needleless injectors, and other in-
2 diabetes as they age to determine likely to skip injections. That dietary non- jection assistance devices. However, we
change in rate of insulin omission. adherence is associated with insulin non- have found that patients do not feel that
Having type 2 diabetes was itself as- adherence is not surprising. More their health care providers are giving
sociated with higher levels of intentional frequent injection omission among indi- them adequate assistance in managing
omission of insulin injections. The beta viduals taking more injections could re- these problems, even when they raise the
for this variable (0.226) was approxi- flect the frequently reported finding that issue with their providers (18).
mately twice the size of the unadjusted eta more complex regimens are associated We note that the measure of negative
(0.095), reflecting the fact that control- with lower levels of adherence (6). It affect toward injections was significantly
ling for confounding factors (such as age might also be that the impact of skipping associated with insulin omission until dis-
and number of daily injections) revealed a a shot is reduced among individuals who satisfaction with injection embarrassment
stronger underlying association. The in- take more shots. and pain were entered into the model (re-
dependent association of type 2 diabetes sults not shown). This suggests that ad-
with increased insulin omission may re- Insulin and injection-related factors dressing pain and embarrassment may
flect the fact that patients with type 2 di- Our study suggests that insulin omission reduce not only insulin omission, but also
abetes have a residual insulin response, is affected by the perceived burden of in- the emotional burden of injections, thereby
reducing the immediate consequences of sulin therapy (i.e., having to plan one’s enhancing psychological well-being. It is in-
omitting an injection. Thus, these indi- life around insulin injections and feeling teresting that worry about hypoglycemia
viduals may feel less vulnerable to the ef- that the insulin regimen interferes with did not predict intentional omission of in-
fects of skipping insulin injections they activities of daily living such as social ac- sulin injections, even though worry about
know they should take. Interestingly, tivities, work-related activities, and family hypoglycemia was high in the study popu-
whereas duration of diabetes was associ- care-giving responsibilities). We offer one lation. This suggests that patients may ad-
ated with the frequency of insulin injec- caveat regarding our findings; we do not dress this worry by eating more or lowering
tion omission, regression analysis believe that the behavior of planning one’s insulin doses rather than by skipping injec-
revealed that duration of diabetes did not day around insulin injections actually in- tions altogether.
make an independent contribution to this creases the level of insulin injection omis-
behavior. That is, although insulin omis- sion, but we do believe that feeling that one Study strengths and limitations
sion may be less common among individ- has to plan around one’s injections is asso- Strengths of the study include the large
uals with longer duration of diabetes, this ciated with higher frequency of skipping in- sample of diabetic patients drawn from a
is likely a function of other factors such as sulin injections one should take. That is, general population and the fact that the
age and type of diabetes rather than of when there is a conflict between scheduling sample was weighted to be nationally rep-
duration per se. of treatment and life activities, one can ei- resentative. However, patients volunteer
The associations of insulin injection ther plan one’s activities in a way that re- for the panel from which respondents
omission with other health conditions duces this conflict or deal with the conflict were drawn and may not be representa-
were examined. Surprisingly, history of by ignoring treatment needs. Reducing the tive of all patients (e.g., they may be more
depression was not associated with insu- perceived burden of insulin injections may adherent with their treatment regimens).
lin omission; this contradicts findings require more effort from health care provid- Limitations of the study include the
from studies of general adherence (9,20) ers. As we have suggested elsewhere, pro- fact that there was no objective measure of
and of insulin omission among adoles- viders must find out what the specific issues insulin use (e.g., pharmacy records).
cents (10). However, because current de- are for each patient and work with that pa- Moreover, while our measure of insulin
pression was not assessed, any concurrent tient to develop solutions that will work for omission was very specific (i.e., skipping
association was likely to be lost. Others him or her (24). injections that respondents knew they
have shown that depression symptom We note that the measure of interfer- should take), it is possible that some re-
scores fluctuate substantially over rela- ence with eating and exercise was signifi- spondents included injections they did
tively short periods of time; individuals cantly associated with insulin injection not skip intentionally, but rather simply
with elevated depression symptoms at a omission until interference with activities forgot to take. Respondents might also
given point in time are likely to not report of daily living was entered into the model have included scheduled injections that
elevated symptoms 6 months later (results not shown). Thus, while interfer- were appropriately skipped because a
(21,22). In addition, depression symp- ence with eating and exercise might be meal was not eaten or blood glucose levels
toms across the whole range of severity part of the burden of insulin therapy, in- were very low. This could explain (at least
symptoms have been shown to predict terference with other aspects of daily liv- in part) the association between following
regimen adherence more powerfully than ing had a more substantial association a healthy diet and fewer skipped injections.
diagnosed depression (23). Being dis- with insulin omission. More disciplined eating behavior reduces
abled was associated with less insulin Our study suggests that insulin omis- meal skipping, which is an often-cited rea-
omission; this may be due to a variety of sion may be affected by the immediate son for skipping insulin injections. This is

244 DIABETES CARE, VOLUME 33, NUMBER 2, FEBRUARY 2010 care.diabetesjournals.org


Peyrot and Associates

consistent with the fact that the relationship of Adults with Diabetes using Diabetes Med- 13. Mollema ED, Snoek FJ, Pouwer F, Heine
between our diet measure and insulin omis- ication by Type of Medication, United States, RJ, van der Ploeg HM. Diabetes Fear of
sion is present only among individuals with 1997–2003. Data computed by the person- Injecting and Self-Testing Questionnaire:
type 1 diabetes where closer matching of nel of the Division of Diabetes Translation, a psychometric evaluation. Diabetes Care
National Center for Chronic Disease Pre- 2000;23:765–769
food and insulin is required. vention and Health Promotion, CDC. Avail-
Finally, our study probably underes- 14. Mollema ED, Snoek FJ, Ader HJ, Heine
able online at http://www.cdc.gov/diabetes/ RJ, van der Ploeg HM. Insulin-treated di-
timated the level of insulin nonadherence statistics/meduse/fig2.htm. 2008 abetes patients fear of self-injecting or fear
in this population because it did not cap- 2. Nathan DM, Buse JB, Davidson MB, Ferran- of self-testing: psychological comorbidity
ture instances in which patients took an nini E, Holman RR, Sherwin R, Zinman B. and general well-being. J Psychosom Res
injection but gave less than the amount of Medical management of hyperglycemia in 2001;51:665– 672
insulin they knew was needed for optimal type 2 diabetes: a consensus algorithm for 15. Snoek FJ, Skovlund SE, Pouwer F. Devel-
glucose control. the initiation and adjustment of therapy: a opment and validation of the insulin
consensus statement from the American Di- treatment appraisal scale (ITAS) in pa-
abetes Association and the European Asso-
Clinical implications tients with type 2 diabetes. Health Qual
ciation for the Study of Diabetes. Diabetes
Our findings make clear that while most Life Outcomes 2007;5:69 –75
Care 2009;32:193–203
patients did not report regular omission 16. Martinez L, Consoli SM, Monnier L, Si-
3. Cramer JA, Pugh MJ. The influence of in-
of insulin injections, a substantial number sulin use on glycemic control: how well mon D, Wong O, Yomtov B, Gueron B,
did. Thus, our findings suggest that it is do adults follow prescriptions for insulin? Benmedjahed K, Guillemin I, Arnould B.
important to identify patients who omit Diabetes Care 2005;28:78 – 83 Studying the Hurdles of Insulin Prescrip-
insulin and to be aware of the potential 4. Morris AD, Boyle DIR, McMahon D, tion (SHIP): development, scoring, and
MacDonald TM, Newton RW, for the initial validation of a new self-adminis-
risk factors identified here. Lack of per-
DRTS/MEMO Collaboration. Adherence to tered questionnaire. Health Qual Life
sonal resources (especially income) is one Outcomes 2007;5:53– 63
potential warning sign. Although much insulin treatment, glycemic control, and
ketoacidosis in insulin-dependent diabetes 17. Peyrot M, Skovlund SE, Landgraf R. Epi-
attention has been focused on insulin demiology and correlates of weight worry
mellitus. Lancet 1997;350:1505–1510
omission among adolescents with type 1 5. Lee WC, Balu S, Cobden D, Joshi AV, Pa- in the multi-national Diabetes Attitudes,
diabetes, our findings suggest that among shos CL. Medication adherence and the Wishes and Needs (DAWN) Study. Curr
adults, individuals with type 2 diabetes associated health-economic impact among Med Res Opin 2009;25:1985–1993
are at higher risk. Patients who are not patients with type 2 diabetes mellitus con- 18. Rubin RR, Peyrot M, Kruger DF, Travis
adhering to other elements of the treat- verting to insulin pen therapy: an analysis of LB. Barriers to insulin injection therapy:
ment regimen, especially diet, also may third-party managed care claims data. Clin patient and health care provider perspec-
be at risk for insulin omission. For pa- Ther 2006;28:1712–1725 tives. Diabetes Educ 2009;35:1014 –1022
tients who report injection-related 6. Rubin RR. Adherence to pharmacologic 19. Link BG, Phelan JC, Miech R, Westin EL.
problems (interference with daily activ- therapy in patients with type 2 diabetes The resources that matter: fundamental
mellitus. Am J Med 2005;118(Suppl. 3): social causes of health disparities and the
ities, injection pain, and embarrass-
27S–34S challenge of intelligence. J Health Soc Be-
ment), providers should consider 7. Wagner TH, Heisler M, Piette JD. Pre- hav 2008;49:72–91
recommending strategies and tools for scription drug co-payments and cost- 20. Lin EH, Katon W, Von Korff M, Rutter C,
addressing these problems to prevent related medication underuse. Health Simon GE, Oliver M, Ciechanowski P,
insulin omission. This may contribute Econ Policy Law 2008;3:51– 67 Ludman EJ, Bush T, Young B. Relation-
to improved treatment adherence and 8. Paes AH, Bakker A, Soe-Agnie CJ. Impact ship of depression and diabetes self-care,
consequent clinical outcomes. of dosage frequency on patient compli- medication adherence, and preventive
ance. Diabetes Care 1997;20:1512–1517 care. Diabetes Care 2004;27:2154 –2160
9. Gonzalez J, Peyrot M, McCarl A, Collins 21. The Diabetes Prevention Program Re-
Acknowledgments — This work was sup- E, Serpa L, Mimiaga M, Safren S. Depres-
search Group. Depression symptoms and
ported by an unrestricted grant from Patton sion and diabetes treatment nonadher-
antidepressant medicine use in Diabetes
Medical Devices, which also made the data ence: a meta-analysis. Diabetes Care 2008;
Prevention Program participants. Diabe-
available to the authors. L.B.T. is a shareholder 31:2398 –2403
10. Olmsted MP, Colton PA, Daneman D, tes Care 2005;28:830 – 837
in Patton Medical Devices, is a co-chair of the
scientific advisory committee for Patton Med- Rydall AC, Rodin GM. Prediction of the 22. Peyrot M, Rubin RR. Persistence of de-
ical Devices, and works as a consultant for Pat- onset of disturbed eating behavior in pressive symptoms in diabetes. Diabetes
ton Medical Devices. D.F.K. is an advisor for adolescent girls with type 1 diabetes. Care 1999;22:448 – 452
and owns stock in Patton Medical Devices. No Diabetes Care 2008;31:1978 –1982 23. Gonzalez JS, Safren SA, Cagliero E, Wex-
other potential conflicts of interest relevant to 11. Anderson RT, Skovlund SE, Marrero D, Le- ler DJ, Delahanty L, Wittenberg E, Blais
this article were reported. vine DW, Meadows K, Brod M, Balkrishnan MA, Meigs JB, Grant RW. Depression,
R. Development and validation of the Insu- self-care, and medication adherence in
lin Treatment Satisfaction Questionnaire. type 2 diabetes: relationships across the
References Clin Ther 2004;26:565–578 full range of symptom severity. Diabetes
1. Centers for Disease Control and Preven- 12. Peyrot M, Rubin RR. Validity and reliability Care 2007;30:2222–2227
tion (CDC) National Center for Health of an instrument for assessing health-related 24. Peyrot M, Rubin RR. Behavioral and psy-
Statistics, Division of Health Interview quality of life and treatment preferences: the chosocial interventions in diabetes: a con-
Statistics, data from the National Health Insulin Delivery System Rating Question- ceptual review. Diabetes Care 2007;30:
Interview Survey. Age-Adjusted Percentage naire. Diabetes Care 2005;28:53–58 2433–2440

care.diabetesjournals.org DIABETES CARE, VOLUME 33, NUMBER 2, FEBRUARY 2010 245