Primary Care Diabetes 16 (2022) 381–386

Contents lists available at ScienceDirect

Primary Care Diabetes

journal homepage: www.journals.elsevier.com/primary-care-diabetes

Diabetes distress, emotional regulation, HbA1c in people with diabetes and

A controlled pilot study of an emotion-focused behavioral therapy
intervention in adults with type 2 diabetes
Emil F. Coccaro a, *, Tina Drossos b, David Kline c, Sophie Lazarus a, Joshua J. Joseph d,
Mary de Groot e
a
Department of Psychiatry and Behavioral Health, The Ohio State University Wexner Medical Center, Columbus, OH, USA
b
Department of Psychiatry and Behavioral Neuroscience, Pritzker School of Medicine, The University of Chicago, IL, USA
c
Department of Biomedical Informatics, Center for Biostatistics, The Ohio State University Wexner Regulation Medical Center, Columbus, OH, USA
d
Department of Medicine, Division of Endocrinology, Diabetes and Metabolism, The Ohio State University Wexner Medical Center, Columbus, OH, USA
e
Department of Medic ine, Indiana University School of Medicine, Indianapolis, IN, USA

A R T I C L E I N F O A B S T R A C T

Keywords: Aim: One potential barrier for people with diabetes to reach glycemic goals is diabetes distress. Accumulating
Diabetes Distress evidence suggests diabetes distress may be linked to individuals’ emotion regulation capacities. Thus, we con­
Emotional Regulation ducted two studies to elucidate a model for how emotion regulation impacts diabetes distress and A1c levels and
Hemoglobin A1c
determine preliminary effect size estimates for an intervention targeting poor emotion regulation on glycemic
Diabetes Self-Management
control.
Methods: Study I used structural equation modeling to assess the cross-sectional relationships between these
variables in a sample of 216 individuals with Type 1 and Type 2 diabetes. Study II built on findings from Study I
that highlighted the role of emotion regulation capacities in diabetes distress and A1c by conducting a pilot study
of an emotion-focused behavioral intervention compared to treatment as usual in a sample of individuals with
Type 2 diabetes.
Results: Study I examined two potential explanatory models with one of the models (Model II) showing a more
comprehensive view of the data revealing a total effect of poor emotional regulation of 42% of all effects on A1c
levels. Study II tested an emotion-focused behavioral intervention in patients with Type 2 diabetes compared to
treatment as usual and found medium sized reductions in A1c levels and smaller reductions in diabetes distress
that correlated with changes in emotion regulation.
Conclusions: These studies suggest that, in people with diabetes, elevated A1c levels and diabetes distress are
linked with poor emotion regulation. While the effect sizes from Study 2 are preliminary, an emotion-focused
behavioral intervention may reduce both A1c and diabetes distress levels, through improvements in emotion
regulation. Overall, these data suggest that targeting difficulties in emotion regulation may hold promise for
maximizing improvement in diabetes distress and A1c in individuals with diabetes.

1. Introduction distress associated with living with diabetes (e.g., chronic illness and
burden of treatment tasks). Diabetes distress is common, with nearly
Diabetes affects over 34.2 million Americans and is currently the 7th 42% reporting elevated diabetes distress scores, is inversely related with
leading cause of death in the US [1]. Despite advances in medication and both quality of life [4] and diabetes self-care behaviors [3] and mani­
device technology, less than 50% of people with diabetes achieve a fests an independent association with A1 C, over time [5,6].
glycemic target of A1 C < 7.0% (< 53 mmol/mol) [2,3]. One potential Options for treating diabetes distress have traditionally included
barrier to glycemic management is diabetes distress, the emotional diabetes education and psychological interventions based on cognitive

* Correspondence to: Department of Psychiatry and Behavioral Health, The Ohio State University Wexner Medical Center, 520 King Avenue, Columbus, OH 43210,
USA.
E-mail address: emil.coccaro@osumc.edu (E.F. Coccaro).

https://doi.org/10.1016/j.pcd.2022.03.002
Received 31 May 2021; Received in revised form 1 February 2022; Accepted 2 March 2022
Available online 11 March 2022
1751-9918/© 2022 Primary Care Diabetes Europe. Published by Elsevier Ltd. All rights reserved.
E.F. Coccaro et al. Primary Care Diabetes 16 (2022) 381–386

behavioral therapy, typically targeting depressive symptoms [3,7] even assessed, by finger-stick, in the Kovler Diabetes Program using a point of
though diabetes distress is only modestly associated with depression care instrument (Siemens) as in our previous study [14]. A1c levels were
scores [8]. Unfortunately, these approaches have resulted in only not normally distributed and were log transformed for analysis.
modest effects on A1c levels [7]. One missing element to addressing Assessment of Diabetes Distress. Three items from the Diabetes
diabetes distress may be a lack of acknowledgement of individual dif­ Distress Scale [18] and three from the Problem Areas in Diabetes [19]
ferences in the ability to regulate one’s negative emotions. and were used to screen for Diabetes Distress. This screen demonstrated
Emotional regulation consists of the experience, processing, under­ very good internal consistency (α = 0.89) and correlated significantly
standing, and coping with emotion [9]. Problems in emotion regulation with a quality of life measure (r = − 0.41, p < 0.001) as does the full
are manifest by feeling too much (or too little) emotion in response to Diabetes Distress Scale [3].
daily life events, and/or in the reactivity/lability of emotion referred to Assessment of Diabetes Self-Care. Both studies included a question­
as Emotion Regulation-Experience. Difficulty in identifying, evaluating, naire related to diabetes self-care which was assessed with the Self-Care
and controlling the expression of emotion in an appropriate manner is Inventory-Revised (SCI-R) [20]. The SCI-R is a 15-item questionnaire,
referred to as skill in emotion management. Emotion scored on a 0–4 Likert scale (ranging from “never” to “always”),
Regulation-Experience and Emotion Regulation-Skill are inversely assessing diabetes self-care in the past one to two months.
related, and the presence of poor Emotion Regulation-Skill increases as Assessment of Emotional Regulation. Four questionnaires related to
Emotion Regulation-Experience (i.e., negative emotionality) since the negative emotionality and skill at regulating negative emotion were
sub-optimal degree of Emotion Regulation-Skill cannot “reign in” the used in this study. Negative emotionality was assessed with the six-item
experience of negative emotion. “Negative Emotional Intensity”; e.g., “my friends would probably say
A relationship between emotion regulation and glycemic manage­ I’m a tense or ‘high-strung’ person”) scale of the Affect Intensity Mea­
ment is supported by studies on the impact of emotional states, and sure (AIM) [21] and the eight-item “Anxiety-Depression Lability” (e.g.,
chronic stress on circulating glucose levels [10,11]. Specific to Emotion “there are times when all I can think about is how worthless I am and
Regulation-Skill, one study in = adults without diabetes reported that then very soon afterwards all I can think about are the things that I am
enhancing positive emotional states reduce, while enhancing negative worried about”) scale form the Affect Lability Scales (ALS) [22]. Skill at
emotional states increase, circulating glucose, specifically in those regulating negative emotion was assessed by using the “Clarity of
“Poor” Emotion Regulation [12]. Consistent with these findings, recent Emotion” (11 items; e.g., “I usually know my feels about a matter”) and
studies in individuals with Type 1 diabetes (T1D) [13] and Type 2 dia­ “Repair of Emotion” (6 items; e.g., “When I become upset I remind
betes (T2D) [14] report significant correlations with measures of nega­ myself of all the pleasures in life”) from the Trait-Meta Mood (TMM; [9])
tive emotional experience and skill at modulating negative emotion. In a questionnaire.
larger sample, we have also shown that poor emotion regulation is
strongly associated with diabetes distress [15]. If so, an explicit focus on 2.2. Statistical analysis
emotion regulation skills may improve outcomes for diabetes distress
interventions. To date only the T1REDEEM [16] study in individuals Study 1 involved a series of analyses including a descriptive analysis,
with Type 1 diabetes included a psychological intervention (OnTrack) a Bayesian graphic model, and a comparison of two hypothesized
involving an explicit focus on emotion management. OnTrack contained models. For the descriptive analysis, a negative Emotion Regulation-
four hour-long videos over the course of the trial. While it yielded a large Experience variable was created by taking the mean of Z-scores for
reduction in diabetes distress (d = 1.06) it was only associated with a AIM “Negative Emotional Intensity” and ALS “Anxiety/Depression
small relationship between change in diabetes distress and A1 C (r = Lability”. An Emotion Regulation-Skill variable was created by taking
0.14, p = 0.01). Later analysis [17] suggested that “emotion manage­ the mean of Z-scores of TMM “Clarity of Emotion” and TMM “Repair of
ment” may be critical to reducing diabetes distress and, possibly, A1 C Emotion”). In turn, these were similarly combined into a composite
levels and that such interventions likely require a more intense emotion variable reflecting Emotional Regulation with high scores reflecting
regulation intervention [7] and that should be targeted to individuals poor emotion regulation. Statistical analysis of these data involved chi-
with difficulties in emotion regulation [17]. square, t-test, ANOVA, all at a two-tailed alpha level of 0.05. For the
In this paper, we report on two studies. The aim of Study 1 was to Bayesian analysis we tested the fit of hypothesized relationships be­
further explicate the relationships between variables of A1c, diabetes tween Emotion Regulation, Diabetes Distress, Self-Care, A1c levels in the
distress, emotion regulation, and self-care variables through the analysis context of a Bayesian network model. The model jointly estimated each
of cross-sectional data from individuals with Type 1 and Type 2 diabetes. edge in the graph as a linear regression model adjusted for age, sex, and
Study 2 was a natural extension of our findings from Study 1 and its aim ethnicity. Non-informative flat prior distributions were used for all
was to conduct a pilot study to determine preliminary estimates of effect regression coefficients and gamma prior distributions with shape and
sizes of an emotion regulation focused behavioral therapy intervention scale parameter of 0.1 were used for all precision parameters. All
on diabetes distress and A1c levels in individuals with Type 2 diabetes. continuous variables were standardized to have mean 0 and variance 1
to allow regression coefficients to be interpreted as estimates of partial
2. Methods: study 1 correlation. Analogous to frequentist significance testing at the 0.05
level, associations were noted as significant if the posterior probability
Study Participants. Participants were recruited from individuals with that the coefficient was greater than 0 was less than 0.025 (negative
Type 1 and Type 2 diabetes receiving care at the Kovler Diabetes Center association) or greater than 0.975 (positive association). Models were fit
program at the University of Chicago Medical Center between 2012 and using a Markov chain Monte Carlo algorithm implemented in R using
2016 and have been previously described [15]. After giving informed NIMBLE [23]. The algorithm was run for 50,000 iterations, discarding
consent agreeing that their data would be used for research purposes the first 20,000 as burn-in. Convergence was assessed visually using
without identifying them, study participants were evaluated by clinical trace plots. Posterior distributions were summarized using the posterior
psychology externs and interns with a structured clinical assessment and mean, 95% credible interval (CI), and the posterior probability the effect
a series of other assessments relevant to diabetes (see below). The study is greater than 0. Finally, we tested two hypothesized models. The first
was approved by the University of Chicago Institutional Review Board. model tested was a linear model from poor emotion regulation to dia­
betes distress to self-care to A1c levels (Fig. 1a). The second model tested
2.1. Assessments included poor emotion regulation to diabetes distress with self-care as a
mediating variable between poor emotion regulation and diabetes
Assessment of Glycosylated Hemoglobin (A1c). A1c levels were distress to A1c level (Fig. 1b).

382
E.F. Coccaro et al. Primary Care Diabetes 16 (2022) 381–386

Fig. 1. Model I (1a) and Model II (1b) with correlation and standardized regression coefficients. Asterisks indicate a statistically significant result in the standard
error of the mean (SEM).

3. Results: study 1
Table 1
Characteristics of Adult Study Participants with Type 1 and 2 Diabetes in Study
Participant Characteristics. Two-hundred-sixteen adult study par­
1.
ticipants took part in this study. The sample was split between those
with Type 2 (n = 136) and Type 1 (n = 80) diabetes and their charac­ VARIABLES ALL T2D T1D P (T2D
(N = 216) (N = 136) (N = 80) vs. T1D)
teristics are listed in Table 1. While participants with Type 2 diabetes
differed from those with Type 1 diabetes, in age, ethnicity, current in­ DEMOGRAPHIC
VARIABLES
come, years with diabetes, insulin dependence, and self-care, no sig­
Age (Years) 49.3 + 17.5 58.1 + 12.8 34.3 + 13.9 < 0.001
nificant differences were observed between the groups in sex, A1c levels, Sex (% Female) 59.8 60.3 56.3 = 0.560
or in diabetes distress, emotion regulation-experience, or emotion Race (% Non-White) 51.9 69.9 21.3 < 0.001
regulation-skill scores. Income (% < $20 K / 38 / 27 / 35 42 / 32 / 26 31 / 19 / 50 = 0.012
Descriptive Analyses. Zero-order correlations (Table 2) suggested $20–60 K / > $60 K)
DIABETES RELATED
highly significant, medium-sized, relationships between poor emotional VARIABLES
regulation and diabetes distress and A1c, and between self-care and A1c A1c Level 7.9 + 1.8 7.8 + 1.6 8.0 + 1.9 = 0.583
levels. Given the inter-relationships among these variables we pro­ Years With Diabetes 34.3 + 19.2 45.3 + 13.6 15.7 + 11.5 < 0.001
ceeded to fit the data to two models. Insulin Dependent (%) 76.9 63.2 100.0 < 0.001
Diabetes Distress 6.7 + 6.0 6.8 + 6.6 6.4 + 4.9 = 0.551
Network Model Analyses. A simple linear model (Model I), demon­
Screen
strated a significant PATH coefficient from poor emotion regulation to Diabetes Self-Care 49.7 + 9.0 48.0 + 8.7 52.7 + 8.8 < 0.001
diabetes distress (0.39, 95% CI: 0.26–0.51), a smaller, but non- (SCI-R)
significant, PATH coefficient for diabetes distress to self-care (− 0.15, EMOTION
95% CI: − 0.01 to − 0.28), and a statistically significant, PATH coeffi­ REGULATION
VARIABLES
cient from self-care to A1c (− 0.22, 95% CI: − 0.09 to − 0.35); Fig. 1a. A Global Negative 32.7 + 8.9 32.7 + 8.5 32.5 + 9.6 = 0.420
second model (Model II) allowing for mediation effects among the var­ Emotionality
iables also fit these data. This model demonstrated a significant Direct (Emotion
PATH coefficient from poor emotion regulation to A1 C of 0.14 (95% CI: Regulation-
Experience)
0.02–0.28) and a significant Indirect Path coefficient from poor emotion
Global Emotional Skills 46.3 + 8.5 45.9 + 8.3 46.9 + 8.9 = 0.585
regulation to A1 C via diabetes distress of 0.13 (95% CI: 0.06–0.20) for a (Emotion
Total Effect of poor emotion regulation on A1 C of 0.27 (95% CI: Regulation-Skill)
0.10–0.44). The model also demonstrated a significant Direct Path co­
efficient from diabetes distress to A1 C of 0.32 (95% CI: 0.18–0.46) and a
significant Direct PATH coefficient from self-care to A1 C of 0.19 (95% respectively). More so than Model I, Model II suggests that the total
CI: 0.06–0.32). Compared with Model I, poor emotion regulation and effect of poor emotion regulation on A1 C (0.14 direct and 0.12 indirect
diabetes distress have less relevance on A1 C via self-care (0.10 and 0.11, via diabetes distress) is: (a) substantial (0.26) and almost as large as the

383
E.F. Coccaro et al. Primary Care Diabetes 16 (2022) 381–386

Table 2 regarding events, thoughts, feelings, and behaviors and how these relate
Zero-Order Correlations for Poor Emotion Regulation, Diabetes Distress Screen, to diabetes management. More specifically, this intervention teaches
Diabetes Self-Care, and HbA1c in Study 1. participants to identify emotions and understand their purpose; to use
Poor Emotion Diabetes Distress Diabetes specific cognitive restructuring strategies to change negative thoughts
Regulation Screen Self-Care into more adaptive, realistic thinking; to reduce emotional reactivity
Diabetes Distress 0.41 (p < 0.001) through mindfulness and acceptance of emotional experiences along
Screen with relaxation and distress tolerance strategies; and to appropriately
Diabetes Self-Care -0.13(p = 0.05) - 0.12(p = 0.07) express and communicate their thoughts and feelings, strategies that are
A1c 0.03(p = 0.698) 0.31(p < 0.001) - 0.21
practiced and role-played with participants. Each session was followed
(p = 0.002)
by homework assignments that were reviewed at the beginning of the
next session (manual is attached in Supplemental Materials).
direct effect of diabetes distress on A1 C (0.32), (b) together, poor Statistics. The focus of the Study 2 pilot clinical trial was to deter­
emotion regulation and diabetes distress account for 71% (0.46/0.65) of mine preliminary effect sizes (Cohen’s d) for the comparison of the
direct effects on A1 C, (c) the direct effect of self-care on A1 C accounts for emotion regulation focused behavioral therapy intervention, vs. treat­
29% (0.19/0.65) of all direct effects on A1 C and, (d) the indirect effect ment as usual, on both A1c levels and Regimen-Related Diabetes Distress
from poor emotion regulation and diabetes distress to A1c is very small Scale scores, and to explore relationships between changes in A1c levels
(i.e., 0.019 and 0.021, respectively). and Regimen-Related Diabetes Distress Scale, Emotion Regulation-
Experience, Emotion Regulation-Skill, and Self-Care scores by Pearson
4. Methods: study 2 correlation.

Because results of previous studies [24] (including Study 1) suggest 5. Results: study 2
that levels of A1c and diabetes distress have important relationships with
poor emotion regulation, we conducted a pilot randomized clinical trial Participant Characteristics. Of the thirty study participants screening
of an emotion regulation focused behavioral therapy intervention in positive for Study 2, twenty began the trial and completed A1c level and
adults with Type 2 diabetes of at least one year duration. behavioral assessments at baseline and end of trial (or ten weeks later for
Study Participants. Between October 2017 and September 2018, treatment as usual participants). The remaining ten, equally randomized
forty adults of both sexes between ages 21 and 65 years, with Type 2 to the two intervention groups, did not return for baseline assessments
diabetes, and persistently elevated A1c levels (> 7.0% as documented in or intervention, and were lost to follow-up; see Consort diagram in
the medical record), were identified and approached to participate in a supplemental file. At baseline, the twenty study participants had high
pilot, randomized, clinical trial to assess the effect of an emotion regu­ A1c levels (9.5 + 2.1%), high Regimen-Related Diabetes Distress Scale
lation focused behavioral therapy intervention, compared with treat­ scores (3.6 + 1.4), high Emotion Regulation-Experience scores (37.3 +
ment as usual, on A1c levels and diabetes distress. All potential 8.3), and low Emotion Regulation-Skill scores (42.1 + 7.9). No differ­
participants received their diabetes care at the Kovler Diabetes Center at ences between intervention groups were observed in demographic or in
the University of Chicago. The trial was approved by the University of other variables of interest (Table 3).
Chicago Institutional Review Board and was registered with Clin­ Response to Emotion Regulation Focused Behavioral Therapy and
icalTrial.gov (NCT03553680). Treatment as Usual. At endpoint, Regimen-Related Diabetes Distress
Assessments. Assessments for Study 2 were the same as in Study 1 Scale scores (mean + sd) fell by 4.0 + 8.6 points (vs. 1.4 + 4.5 points)
except that the full Diabetes Distress Scale was used in this second study. resulting in a smaller than medium-sized (d = 0.38) net improvement
Screening, Entry, and Randomimzation for Pilot Clinical Trial. After favoring the emotion regulation focused behavioral therapy interven­
initial identification, potential study participants were consented for tion (Fig. 2, left). Similarly, A1c levels fell by 1.3 + 1.4% in the inter­
study and further screened with the Diabetes Distress Scale [18] to vention (vs. 0.6 + 1.4% with treatment as usual) participants resulting
identify participants with elevated Regimen-Related Diabetes Distress in a medium-sized (d = 0.53) net improvement in A1c levels favoring the
scores, and with the four emotion regulation assessments, described emotion regulation focused behavioral therapy intervention (Fig. 2,
above, to identify participants with elevated Emotion right). Changes in A1c levels were significantly correlated with changes
Regulation-Experience, and reduced Emotion Regulation-Skill, scores. in Regimen-Related Diabetes Distress Scale [r = 0.48 (CI: 0.05–0.76),
Regimen-Related Diabetes Distress scores > 2 were designated as p < 0.032], in Emotion Regulation-Skill [r = − 0.58 (CI: − 0.18 to
elevated as in previous studies [6]; Emotion Regulation-Experience
scores > 25 were designated as elevated, and Emotion Regulation-Skill Table 3
scores < 48 were designated as reduced (n.b., respectively, these Baseline Characteristics of Adult Study Participants with Type 2 Diabetes in
cut-off scores were above, and below, the mean scores of 247 medically Study 2.
and psychologically healthy controls in our research program). Based on EFBT TAU
these data, thirty participants met study entry criteria (i.e., elevated A1c (N = 10) (N = 10)
levels, elevated Regimen-Related Diabetes Distress, elevated Emotion Demographic Variables
Regulation-Experience, and reduced Emotion Regulation-Skill, scores) Age 54.5 + 10.8 54.2 + 13.5
and were randomized 1:1 to receive an emotion regulation focused Sex ( F / M) 8/2 7/3
therapy intervention or to continue with treatment as usual (control). Ethnicity (AA / White) 9/1 10 / 0
Income ( < $20 K / > $20 K) 6/4 4/6
Emotion Regulation Focused Behavioral Therapy Intervention. Our Variables Relevant to T2D
emotion regulation focused therapy intervention was developed and Glycosylated Hb (A1c) 9.8 + 1.5 9.3 + 2.6
manualized by adapting materials targeting Emotion Regulation- Years with Diabetes 14.3 + 11.1 17.5 + 12.3
Experience and Emotion Regulation-Skill [25–27]. The intevention Insulin-Dependent (Yes / No) 9/1 6/4
Regimen-Related Diabetes Distress (R-DD) 3.6 + 1.4 3.6 + 1.5
was designed to cover a wider range of skills than typical cognitive
Diabetes Self-Care (SCI-R) 45.0 + 7.5 44.2 + 7.9
behavioral therapy by targeting specific deficits in emotional awareness Behavioral Variables
and emotion management that likely underlie diabetes distress. Each of Emotion Regulation-Exp (Negative Emotional 38.2 + 9.0 36.3 + 7.9
the ten (45 min) sessions was delivered by a masters-level clinical psy­ Experience)
chologist (TP) trained in its use and supervised by doctoral-level clinical Emotion Regulation-Skill (Emotional Regulatory 40.4 + 10.2 43.8 + 4.6
Skill)
psychologists (TD, AB). The intervention began with psychoeducation

384
E.F. Coccaro et al. Primary Care Diabetes 16 (2022) 381–386

follow-up before actual entry in the trial. Accordingly, the results of

Study 2 must be seen as preliminary and subject to Type 1 Error. This
said, the results of Study 2 are consistent with Model II in Study 1.
These findings are consistent with previously reported relationships
between A1 C and measures reflective of skill at emotion regulation in
adults with Type 2 [14] and Type 1 diabetes [13] as well as results from
another small study reporting improvements in Emotion
Regulation-Skill scores [25] with reductions in A1c levels [28] in adults
with Type 2 diabetes using a group emotional regulation skill behavioral
therapy based intervention. In addition to what improved emotion
regulation skill might do in reducing glucose levels [12], enhancement
of emotion regulation skill may also lead to a more healthy psycholog­
ical state enabling adults with Type 2 diabetes to better use positive
coping strategies to improve glycemic management. Further study in a
larger sample, using a non-emotion focused attentional control, to
evaluate the efficacy and generalizability of these findings are
warranted.

7. Conclusion
Fig. 2. Effect of Emotion Regulation Focused Behavioral Therapy (EFBT) vs.
Treatment as Usual (TAU): Change from Baseline to End-Trial for Regimen- Data from both studies, above, suggest that diabetes distress in adults
Related Diabetes Distress Scale scores (left) and A1c levels (right) and as a is linked with heightened negative emotionality (Emotion Regulation-
function of effect size (d). Experience) and reduced skill at emotional regulation (Emotion
Regulation-Skill) in adults, both of which are related to elevated A1c
− 0.81), p = 0.007] and in Emotion Regulation-Experience [r = 0.45 levels and that these relationships are stronger than that with diabetes
(CI: 0.01–0.74), p = 0.048] scores but non-significant for changes in self-care. Given other studies examining the construct of emotionality
Self-Care [r = 0.09 (CI: − 0.37 to 0.51), p = 0.712], scores. [17,29], and preliminary results from small treatment studies targeting
emotionality/emotional regulation [25,28], these data suggest that
6. Discussion diabetes distress and A1c may be improved, especially, in those with
diabetes and difficulties with emotionality.
The data from Study 1 suggest that a linear model from poor
emotional regulation to elevated diabetes distress, to reduced self-care, Funding
to elevated A1c levels is a viable model for the relationship between
these variables and A1c. As such, these results are replicative of the same This work was supported in part by a pilot grant from the University
model tested with T1-REDEEM study data which reported coefficients of of Chicago Center for Diabetes Translation Research (via NIDDK P30
0.36 from “poor emotion management” to diabetes distress, 0.19 from DK092949) to EFC. In addition, this work was supported by the Kovler
diabetes distress to “skipped insulin boluses”, and of 0.23 from “skipped Diabetes Center at the University of Chicago.
boluses” to A1c [17]. However, the more comprehensive, interactive,
model (Model II) provides greater explanatory power regarding the ef­ Disclosure
fects of these variables on A1c. Specifically, the effect of poor emotion
regulation (indirect only) effect on A1c was 0.013 in Model I, while its The supporting bodies for this project had no role in the design,
total (direct and indirect) effect on A1c was twenty-fold higher, at 0.26, interpretation, analyses or interpretation of the data presented. The
in Model II. Similarly, the effect of elevated diabetes distress (indirect authors declare that there is no conflict of interest associated with this
only) on A1c was 0.033 in Model I while its total effect was 0.34 in Model manuscript.
II. Reduced self-care’s (direct) effect on A1c was significant in both
models at 0.22 in Model I and 0.19 in Model II. Model II, however, CRediT authorship contribution statement
demonstrates that both poor emotional regulation and elevated diabetes
distress are also associated with increased A1c at a level, respectively, EFC and TD designed the studies. Data analyses were performed by
42%, and 79% greater than that for reduced self-care. If so, it is likely EFC and DK, EFC wrote the first draft of the manuscript and all authors
that interventions targeting poor emotion regulation and/or elevated contributed to discussion, reviewed, and edited the final manuscript.
diabetes distress should have an effect on reducing A1c levels, which set
the stage for the pilot clinical trial in Study 2. Acknowledgements
The data from Study 2 suggest our emotion-focused behavioral
therapy intervention may have clinically meaningful effects on both The authors wish to acknowledge Joselyn Gomez, B.A. for her efforts
Regimen-Related Diabetes Distress scores and A1c levels in adults with in coordinating the data collection of both studies and acknowledge
Type 2 Diabetes, likely by improving emotional regulation as assessed in Tiffany Potts, M.S. and Andrea Busby, Ph.D. for their work in delivering
these studies. Changes in A1c, overall, were significantly correlated with (TP) and in the supervision (AB) of the emotion regulation focused
changes in Regimen-Related Diabetes Distress, Emotion Regulation- behavioral therapy intervention. Finally, we wish to acknowledge sup­
Experience, and Emotion Regulation-Skill scores but not with changes port from the Kovler Regulation Diabetes Center at the University of
in Self-Care scores. While changes in the latter had no effect on A1c, it is Chicago in the conduct of these studies.
possible that our emotion-focused behavioral therapy intervention im­
proves self-care through behaviors not reflected by the measure we used. Conflict of interest
A limitation of Study 1 is that it is a cross-sectional study and in­
ferences from the model-fitting analyses and need to be confirmed in a The supporting bodies for this project had no role in the design,
longitudinal study. Limitations for Study 2 include its small sample size, interpretation, analyses or interpretation of the data presented. The
and the fact that one-third of the randomized participants were lost to authors declare that there is no conflict of interest associated with this

385
E.F. Coccaro et al. Primary Care Diabetes 16 (2022) 381–386

manuscript. [13] D. Ruiz-Aranda, Z.L. García-Linares, E. Castellano-Guerrero, A.M. Martínez-Brocca,

M.A. Gutiérrez-Colosía, Emotional abilities and HbA1c levels in patients with type
1, Diabetes 93 (2018) 118–123.
Appendix A. Supporting information [14] E.F. Coccaro, T. Drossos, L. Phillipson, HbA1c levels as a function of emotional
regulation and emotional intelligence in patients with type 2 diabetes, Prim Care
Supplementary data associated with this article can be found in the Diabetes 10 (5) (2016) 334–341.
[15] E.F. Coccaro, et al., Emotional regulation and diabetes distress in adults with type 1
online version at doi:10.1016/j.pcd.2022.03.002. and Type 2 diabetes, Diabetes Care 44 (1) (2021) 20–25.
[16] L. Fisher, et al., T1-REDEEM: a randomized controlled trial to reduce diabetes
References distress among adults with type 1 diabetes, Diabetes Care 41 (2018) 1862–1869.
[17] L. Fisher, et al., Emotion regulation contributes to the development of diabetes
distress among adults with type 1 diabetes, Patient Educ. Couns. 101 (1) (2018)
[1] Centers for Disease Control, National Diabetes Statistics Report 2020: Estimates of 124–131.
Diabetes and Its Burden in the United States. 2020: Atlanta, GA. [18] W.H. Polonsky, L. Fisher, Assessing psychosocial distress in diabetes: development
[2] S.S. Casagrande, et al., The prevalence of meeting A1C, blood pressure, and LDL of the diabetes distress scale, Diabetes Care 2 (3) (2005) 626–631.
goals among people with diabetes, Diabetes Care 36 (8) (2013) 2271–2279. [19] B.E. McGuire, T.G. Morrison, Short-form measures of diabetes-related emotional
[3] M.M. Carper, et al., The differential associations of depression and diabetes distress distress: the problem areas in diabetes scale (PAID)-5 and PAID-1, Diabetologia 53
with quality of life domains in type 2 diabetes, J. Behav. Med. 37 (3) (2014) (1) (2010) 66–69.
501–510. [20] K. Weinger, et al., Measuring diabetes self-care: a psychometric analysis of the self-
[4] B. Levya, et al., The relative impact of diabetes distress vs depression on glycaemic care inventory-revised with adults, Diabetes Care 28 (2005) 1346–1352.
control in Hispanic patients following a diabetes self-management education [21] R.J. Larsen, E. Diener, Affect intensity as an individual difference characteristic: a
intervention, Ethn. Dis. 21 (2011) 322–327. review, J. Res. Personal. 21 (1987) 1–39.
[5] L. Fisher, et al., Clinical depression versus distress among patients with type 2 [22] P.D. Harvey, B.R. Greenberg, M.R. Serper, The affective lability scales:
diabetes, Diabetes Care 30 (3) (2007) 542–548. development, reliability, and validity, J. Clin. Psychol. 45 (1989) 786–793.
[6] D. Hessler, L. Fisher, R.E. Glasgow, Reductions in regimen distress are associated [23] P. de Valpine, et al., Programming with models: writing statistical algorithms for
with improved management and glycemic control over time, Diabetes Care 37 (3) general model structures with NIMBLE, J. Comput. Gr. Stat. 26 (2017) 403–417.
(2014) 617–624. [24] L. Fisher, et al., Toward effective interventions to reduce diabetes distress among
[7] J. Sturt, et al., Effective interventions for reducing diabetes distress: systematic adults with type 1 diabetes: Enhancing Emotion regulation and cognitive skills,
review and meta-analysis, Int. Diabetes Nursing 12 (2015) 40–55. Patient Educ. Couns. 102 (2019) 8.
[8] J.L. Wardian, et al., Is it distress, depression, or both? exploring differences in the [25] B.M. Yalcin, et al., The effects of an emotional intelligence program on the quality
diabetes distress scale and the patient health questionnaire in a diabetes specialty of life and well-being of patients with type 2 diabetes mellitus, Diabetes Educ. 34
clinic, Clin. Diabetes 37 (2) (2019) 124–130. (2008) 1013–1024.
[9] P. Salovey, et al., Emotional attention, clarity and repair. Exploring emotional [26] K. Tchanturia, Cognitive Remediation and Emotion Skills. Training ((CREST)),
intelligence using the trait meta mood scale, in: J.W. Pennebaker (Ed.), Emotion, South London and Maudsley NHS Trust & Institute of Psychiatry King’s College,
Disclosure and Health, American Psychological Assosciation, Washington, 1995, London, UK, 2015.
pp. 125–154. [27] Applying Emotional, in: J. Ciarrochi, J.D. Mayer (Eds.), Intelligence: A
[10] N. Hermanns, C. Scheff, B. Kulzer, Association of glucose levels and glucose Practitioner’s Guide, Psychology Press, New York, NY, 2007.
variability with mood in type 1 diabetic patients, Diabetologia 50 (2007) 930–933. [28] T.F. Karahan, B.M. Yalcin, The effects of an emotional intelligence skills training
[11] C.E. Lloyd, et al., Association between stress and glycemic control in adults with program on anxiety, burnout and glycemic control in type 2 Diabetes mellitus
type 1 (insulin-dependent) diabetes, Diabetes Care 22 (8) (1999) 1278–1283. patients, Turk. Klinikleri J. Med. Sci. 29 (1) (2009) 16–24.
[12] K. Niven, et al., Achieving the same for less: improving mood depletes blood [29] L. Fisher, et al., Toward effective interventions to reduce diabetes distress among
glucose for people with poor (but not good) emotion control, Cogn. Emot. 27 (1) adults with type 1 diabetes: enhancing emotion regulation and cognitive skills,
(2013) 133–140. Patient Educ. Couns. 102 (8) (2019) 1499–1505.

386