Running head: DSME EFFECTS ON HGA1C

Diabetes Self-Management Education Effects on Hemoglobin A1c
Briana Smith, BSN, RN
Judith Ochieng, PhD, DNP, MSN-ED, RN, FNP-BC
Arizona State University

1

DSME EFFECTS ON A1C AND DSM

2
Abstract

Diabetes, a common chronic condition, effects many individuals causing poor quality of
life, expensive medical bills, and devasting medical complications. While health care providers
try to manage diabetes during short office visits, many patients still struggle to control their
diabetes at home. Lack of diabetes self-management (DSM) is a potential barrier for people with
diabetes having to maintain healthy hemoglobin A1cs (HgA1c). In hopes of addressing this
concern, an evidenced-based intervention; diabetic education and phone calls, using the chronic
care model as its framework was implemented. The intervention targeted people with type II
diabetes at a transitional care setting. Measured variables included HgA1c and DSM. Statistically
significant improvements were seen in reported physical activity. Average improvements were
seen in HgA1c and DSM after three months of diabetes self-management education (DSME).
Attrition, cultural sensitivity, and increasing DSME hours should be further evaluated for future
projects.
Keywords: diabetes, diabetic patients, chronic care management, care management,
hemoglobin A1c

DSME EFFECTS ON A1C AND DSM

3

Diabetes Self-Management Education Effects on Hemoglobin A1c
Chronic care management (CCM) is a significant part of caring for patients with diabetes
as it contributes to better patient care and outcomes. Diabetes is a complex disease requiring
referrals, continuous education, and frequent medication adjustments. All of which are included
in CCM. With the health risks facing people with diabetes, it is important healthcare providers
seek alternative methods to care for people with diabetes.
There are approximately 422 million people living with diabetes worldwide, with a
predicted increase to 642 million by 2040 (World Health Organization [WHO], 2019; Zou et al.,
2018). About one in every five Americans aged 65 and older have been diagnosed with diabetes
(Hasche, Ward, & Schluterman, 2017). In Arizona, approximately one-third of people are
prediabetic and one in 10 are diabetic, representing 2.1 million and 720,000 people, respectively
(Diabetes Action Plan and Report, 2019). With about 34,000 being newly diagnosed yearly
(American Diabetes Association [ADA], 2014). Arizona spent an estimated $6.8 billion on
diabetes care in 2019 (Diabetes Action Plan and Report, 2019). In 2016, the prevalence of
diabetes in Yuma County, located in the southwest corner of Arizona, was 12.9% of the
population aged 20 years and older (Centers for Disease Control and Prevention, n.d.). The
county has seen diabetes rates double over the last decade (Yuma Regional Medical Center
[YRMC], 2016).
Transitional Care Services serves the Yuma Community providing patients with chronic
conditions, such as congestive heart failure, chronic obstructive pulmonary disease, and acute
myocardial infarctions, who need help transitioning home after a hospital discharge (YRMC,
2019). Their goal is to promote quality of life by enhancing knowledge and management of the
patient’s chronic conditions (YRMC, 2019). About 90-95% of patients are referred by the only

DSME EFFECTS ON A1C AND DSM

4

hospital in Yuma County, which had over 12,000 hospital and emergency room diabetes related
discharges (Contreras & Sandoval-Rosario, 2018). Although Transitional Care Services cares for
patients with complex chronic conditions, diabetes is not a disease they primarily focus on.
This information led to the clinically relevant PICO question, in adult patients diagnosed
with diabetes (P), how does CCM (I) compared to standard care (C) affect HgA1c (O)?
Literature review of current evidence included 10 critically appraised articles chosen
from CINHAL, PubMed, and Wiley (see Appendices A and B). Articles selected included five
randomized controlled trials, two cohort studies, one quasi-experimental, one observational study
with no control, and a case study. Level of evidence ranged from II-IV. All studies chosen had at
least one dependent variable (DV) measuring HgA1c. Independent variables showing significant
improvements in HgA1c were care coordination, telephone calls and education, especially
related to diabetes self-management (DSM).
It was determined the proposed evidence-based practice (EBP) project would use
diabetes self-management education (DSME) and telephone calls to implement CCM to type II
diabetic patients at Transitional Care Services. The measurable outcomes of the project were
DSM and HgA1c. The EBP project was informed by the Chronic Care Model (CCMo) because
evidence has shown it may improve diabetic outcomes, such as HgA1c (National Institute of
Diabetes and Digestive and Kidney Diseases [NIDDK], n.d.; see Appendix C). By applying the
elements of the CCMo, which are health systems, decision support, clinical information systems,
patient self-management support, and community resources, and delivery systems, the project
hoped to join informed, active patients and a prepared, proactive practice team to improve
diabetic outcomes (Improving Chronic Illness Care, 2019). Rosswurm and Larabee’s (1999)
model was chosen as the evidence-based model for this project to serve as guidance throughout

DSME EFFECTS ON A1C AND DSM

5

the process changes (see Appendix D). The model assists in changes that are healthcare specific
and strives for improved quality and outcomes.
Methods
Participants
Adults, 18 years or older, were identified using the electronical health record (EHR) at
Transitional Care Services with the target goal being 30 participants. Potential subjects of the
project met the following inclusion criteria: ≥ 18 years old, previously documented type II
diabetes diagnosis, previously documented HgA1c ≥ 6.5% in last month, English speaking, has
access to telephone calls for the duration of project, and able to sign consent. Exclusion criteria
includes: history of dementia, participating in other diabetic studies, and non-English speaking.
Once identified, a flyer was handed to potential subjects to avoid coercion. If the subject wished
to participate, consent was obtained. Ethical consideration for the project was processed and
approved by Arizona State University’s Institutional Review Board and Yuma Regional Medical
Center’s Innovation Council Advisory Board.
Study Design
All participants had a HgA1c collected from the EHR and completed a diabetes selfmanagement questionnaire (DSMQ) prior to intervention, which served as pretests. Diabetes
education was then initiated during the same visit. All participants were given the same
education by the same individual at individual times. Education included glucose management
(GM), dietary control (DC), physical activity (PA), and healthcare use (HU). All participants
were given a take home folder pertaining to the subject matter. Participants were given three
monthly phone calls to serve as a reminder of the lesson content provided at the educational visit.
After three months, participants had a new HgA1c collected from the EHR and complete a post-

DSME EFFECTS ON A1C AND DSM

6

DSMQ. Participants who did not have a new HgA1c recorded in EHR after three months or did
not complete a post-DSMQ were disqualified from the project. Measurable outcomes, HgA1c
and DSM, were statistically analyzed using a paired sample t-test.
Hemoglobin A1c
HgA1c is a blood test reflecting average blood sugars over three months (ADA, 2019).
The ADA (2019) recommends measuring HgA1c levels at least biannually if patients are
meeting treatment goals or quarterly if therapy has changed or glycemic goals are not met.
HgA1c was chosen as a measurable outcome because the ADA (2019) recognizes the blood test
as a standard of care due to its strong predictability value for diabetic complications. Although is
it recognized by the Centers for Disease Control and Prevention (2018) and the American
College of Physicians (2018) as an appropriate diabetic test, there are some limitations.
Conditions that affect red blood cell turnover might cause discrepancies in HgA1c (ADA, 2019).
Additionally, HgA1c has shown to have low sensitivity but high specificity. Measuring against a
single fasting glucose (≥126 mg per dL), the sensitivity and specificity of an HgA1c ≥6.5% for
detection of diabetes was 47% and 98%, respectively (Selvin, Steffes, Gregg, Brancati, &
Coresh, 2011). Three years later, repeated fasting glucose (≥126 mg/dL) showed sensitivity
increased to 67% and specificity remained high at 97% (Selvin, Steffes, Gregg, Brancati, &
Coresh, 2011).
Diabetes Self-Management Questionnaire
DSMQ has 16 questions pertaining to five subscales: GM, DC, PA, HU, and self-care
summary (SS) (Schmitt et al. 2013). SS is an overall measurement of perceived self-care. During
its evaluation, the DSMQ was found to be reliable with good factorial validity and a strong
correlation to HgA1c in patients with type I and II diabetes. It also had good concurrent validity

DSME EFFECTS ON A1C AND DSM

7

when compared to Summary of Diabetes Self-Care Activities Measure. Overall, internal
reliability was good with a Cronbach’s α coefficient of 0.84. Its subscales were mostly
acceptable (GM: 0.77; DC: 0.77; PA: 0.76; HU: 0.60).
Statistical Analysis
Statistical analysis began after data collection was finalized using Intellectus Statistics. A
two-tailed paired samples t-test was conducted to examine whether the mean difference of DVs
were significantly different from zero based on an alpha value of 0.05. Based on Shapiro-Wilk
test and Levene’s test, all DVs’ normality assumptions and homogeneity of variances were met.
Results
Demographics
In total, 29 participants were recruited. By final data collection, there were nine subjects
who completed the intervention (see Appendix E). The most frequently observed age range was
65 years and older (n = 5, 56%). Most subjects were male (n = 8, 89%). Most subjects identified
as Hispanic/Latino/Spanish (n = 8, 89%). Most subjects had been diagnosed with diabetes
greater than 10 years ago (n = 5, 56%). The majority of subjects’ highest level of education was
high school (n = 7, 78%).
Hemoglobin A1c
There were mean improvements in pre- and post-HgA1c for final subjects, 8.57% (SD =
1.92) and 8.29% (SD = 1.77), respectively. The result of the two-tailed paired samples t-test was
not significant, t(8) = 0.57, p = .587.
Diabetes Self-Management

DSME EFFECTS ON A1C AND DSM

8

Each individual subscale of the DSMQ was statistically analyzed. Each subscale was first
given a 10-point scale score. The scale score value was used to calculate the two-tailed paired
samples t-test of each subscale.
Glucose Management. There were mean improvements in pre-GM and post-GM, 6.30
(SD = 3.01) and 7.11 (SD = 2.37), respectively. The result of the two-tailed paired samples t-test
was not significant, t(8) = -0.70, p = .507.
Dietary Control. There were mean improvements in pre-DC and post-DC, 5.07 (SD =
2.34) and 7.12 (SD = 1.56), respectively. The result of the two-tailed paired samples t-test was
not significant, t(8) = -0.71, p = .500.
Physical Activity. There were mean improvements in pre-PA and post-PA, 6.30 (SD =
3.98) and 8.40 (SD = 2.02). The result of the two-tailed paired samples t-test was significant, t(8)
= -2.56, p = .034.
Healthcare Use. There were mean decreases in pre-HU and post-HU was 9.39 (SD =
0.94) and mean of post-HU was 10 (SD = 0). The result of the two-tailed paired samples t-test
was not significant, t(8) = -1.89, p = .095.
Self-Care Summary. The mean of pre-SS was 9.39 (SD = 3.53) and mean of post-SS
was 10 (SD = 3.11). The result of the two-tailed paired samples t-test was not significant, t(8) =
0.61, p = .559.
Project Impact
By using the CCMo as the project’s conceptual framework, the project was able to
combine aspects of the community, such as self-management support, and health systems,
specifically the EHR, to produce proactive providers. The project encouraged providers to focus

DSME EFFECTS ON A1C AND DSM

9

on diabetes. A chronic disease which was not a primary focus for providers at the clinic prior to
the project.
The framework supported informed, activated patients. Most results were not statistically
significant. Yet, on average, subjects had lower Hg A1c levels and reported better GM, DC, PA,
and HU. Furthermore, most subjects had been living with diabetes for 10 or more years and
reported never receiving DSME prior to the project. Additional notable reports included:
receiving their first diabetic eye exam after 10 or more years of diabetes diagnosis, increasing
their daily physical activity, and keeping food and blood glucose logs.
Project Sustainability
Since phone calls were already apart of the clinic’s workflow and care management of
patients, the project was perceived to have high sustainability moving forward. Nurses at the
clinic conduct weekly phone calls with patients, which is more frequent than the project required.
In addition, the initial DSME visit was approximately 20 minutes. Fortunately, the clinic’s
patient volume and schedule flexibility allowed for this block of time. Seldomly providers were
delayed seeing their patients. The sustainability of this project would require additional supplies
for DSME folder packets, employee hourly pay, and time for education. This additional cost
could be sustained by available reimbursement of DSME from entities, such as the Centers for
Medicare and Medicaid Services and the CDC.
Discussion
The project did improve HgA1c levels and DSM with the use of DSME as CCM, but
statistically significant improvements in HgA1c levels were not yielded. Statistically significant
improvement was seen in reported PA. There was a worsening of reported SS.

DSME EFFECTS ON A1C AND DSM

10

Findings were congruent to previous literature suggesting significant reduction in HgA1c
levels are found in those offered greater than 10 hours of DSME services (Beck et al., 2017).
Over the course of the intervention, the project provided about two hours of DSME per subject.
Limitations
The project sample size was small due to large attrition. Subjects were disqualified
because they did not have a post-HgA1c value in the EHR to collect, they did not answer phone
calls, or they did not perform a post-DSMQ. Social determinants could have played a factor in
high rate of attrition. Evidence suggests Latino populations, especially men, struggle with shame
of illness and lack of interest in health (Testerman & Chase, 2018).
The project had a short interventional period. Furthermore, the three month period was
over several holidays. Some subjects expressed they had overly indulged in culturally traditional
foods over the holiday season.
Recommendations
Recommendations to retain subjects include incentivizing the completion of the project.
Contacting subjects once a month may have lost the interest of subject’s participation without
incentivization. Having scheduled phone calls could help retain subjects by avoiding missed
phone calls.
Increasing DSME hours to greater than 10 could help yield significant results. Increasing
the hours of DSME could give opportunity to measure greater intervals of time, such as six-,
nine-, and 12-months. This may give insight to sustainability of the project.
Most subjects were Hispanic with the highest level of education being high school.
Subjects could have benefited from culturally centered DSME. In addition, many patients at the

DSME EFFECTS ON A1C AND DSM

11

clinic only spoke Spanish, which prevented them from being eligible participants. Further studies
could target Spanish speakers.
Conclusion
CCM is a vital part of any chronic disease. In those with diabetes, CCM is an ongoing
process that supports individuals with diabetes through the lifelong process of DSM. Tools that
help individuals meet their HgA1c goals should be promoted to reduce diabetic complications.
DSME, a component of CCM, has been shown to reduce Hg A1c levels. Additionally, DSME
has been shown to have a positive impact on diabetes-related costs and complications. While the
benefits of DSME have been demonstrated in the literature, low utilization of DSME remains.
Efforts to improve DSME should be explored for improving CCM and lowering Hg A1c.
This project showed DSME can be used to help improve HgA1c and DSM. Although
statistical significant were not yielded in HgA1c and most subscales of DSM, average
improvements were seen in mostly all DV. Attrition rates, cultural sensitivity, DSME hours
provided, and length of project intervention should be further evaluated to produce significant
results.

DSME EFFECTS ON A1C AND DSM

12
Reference

American Diabetes Association. (2014). The burden of diabetes in Arizona [PDF file]. Retrieved
from http://main.diabetes.org/dorg/PDFs/Advocacy/burden-of-diabetes/arizona.pdf
American Diabetes Association. (2019). Standards of medical care in diabetes – 2019 abridged
for primary care providers. Clinical Diabetes, 37(1), 11-34. https://doi.org/10.2337/cd180105
American College of Physicians. (2018). ACP recommends moderate blood sugar control targets
for most patients with type 2 diabetes. Retrieved from https://www.acponline.org/acpnewsroom/acp-recommends-moderate-blood-sugar-control-targets-for-most-patientswith-type-2-diabetes
Beck, J., Greenwood, D. A., Blanton, L., Bollinger, S. T., Butcher, M. K., Condon, J. E…Wang,
J. (2017). 2017 national standards for diabetes self-management education and support.
The Diabetes EDUCATOR, 43(5). doi: 10.1177/0145721717722968
Centers for Disease Control and Prevention. (n.d). Diagnosed diabetes, total, adults aged 20+
years, age-adjusted percentage, Arizona, 2015. Retrieved from
https://gis.cdc.gov/grasp/diabetes/DiabetesAtlas.html#
Centers for Disease Control and Prevention. (2018). All about your A1c. Retrieved from
https://www.cdc.gov/diabetes/managing/managing-blood-sugar/a1c.html
Contreras, O. A. & Sandoval-Rosario, M. (2018). Diabetes in Arizona: The 2018 burden report
[PDF file]. Retrieved from https://www.azdhs.gov/documents/prevention/tobaccochronic-disease/diabetes/reports-data/diabetes-burden-report-2018.pdf
Diabetes Action Plan and Report 2019, 2258, 53rd AZ Legis. (2019).

DSME EFFECTS ON A1C AND DSM

13

Improving Chronic Illness Care. (2019). The chronic care model. Retrieved from
http://www.improvingchroniccare.org/index.php?p=The_Chronic_Care_Model&s=2
National Institute of Diabetes and Digestive and Kidney Diseases. (n.d.) Chronic care model.
Retrieved from https://www.niddk.nih.gov/health-information/communicationprograms/ndep/health-professionals/practice-transformation-physicians-health-careteams/team-based-care/chronic-care-model
Rosswurm, M. A, & Larrabee, J. H. (1999). A model for change to evidence-based practice.
Journal of Nursing Scholarship, 31(4), 317-322. https://doi.org/10.1111/j.15475069.1999.tb00510.x
Schmitt, A., Gahr, A., Hermanns, N., Kulzer, B., Huber, J., & Haak, T. (2013). The Diabetes
Self-Management Questionnaire (DSMQ): Development and evaluation of an instrument
to assess diabetes self-care activities associated with glycemic control. Health Quality
Life Outcomes, 11(138). https://doi.org/10.1186/1477-7525-11-138
Selvin, E., Steffes, M. W., Gregg, E., Brancati, F. L., & Coresh, J. (2011). Performance of A1C
for the classification and prediction of diabetes. Diabetes Care, 34(1), 84-89. doi:
10.2337/dc10-1235
Testerman, J., & Chasse, D. (2018). Influences on diabetes self-management education
participation in low-income, Spanish-speaking, Latino population. Diabetes Spectrum,
31(1), 47-57. https://doi.org/10.2337/ds16-0046
World Health Organization. (2019). Diabetes [Fact sheet]. Retrieved from
https://www.who.int/news-room/fact-sheets/detail/diabetes

DSME EFFECTS ON A1C AND DSM

14

Yuma Regional Medical Center. (2016). 2016 Yuma County Community health needs assessment
[PDF file]. Retrieved from https://www.yumaregional.org/EmergeWebsite/media/YumaDocuments/Community-Health-Needs-Assessment_1.pdf
Yuma Regional Medical Center. (2019). Transitional Care Services. Retrieved from
https://www.yumaregional.org/Medical-Services/Transitional-Care-Services
Zou, Q., Qu, K., Lou, Y., Yin, D., Ju, Y., & Tang, H. (2018). Predicting diabetes mellitus with
machine learning techniques. Frontiers in Genetics, 9(515). doi: 10.3389/
fgene.2018.00515

DSME EFFECTS ON A1C AND DSM

15
Appendix A

Table 1
Evaluation Table
Citation

Holtrop et al.
(2017). Diabetic
and obese patient
clinical outcomes
improve during a
care management
implementation
in PC.
Funding:
NIDDK
Bias: No
conflicts
recognized
Country: USA

Theory/Con
ceptual
Framework

Design/Method/P
urpose

Sample/Setting

CCMo

Design: Pairmatched cluster
randomized trial

N – 1,392
IG – 696
CG – 696

Purpose: To
understand how
individual
practices would
implement care
management, its
successes and
effects on those at
risk of DM due to
obesity.

Demographics:
M Age – 54.8
M/F – 50.4%/49.6%
Setting: PC practices that
are physician-owned
medical group in
southeast Michigan
Inclusion: active pt at
study practices during
study period, 18 years or
older, diagnosis of type 2
DM or obesity

Major Variables

Measurem Data Analysis
ents/Instru
mentation

IV1 – care
management
which includes
staffing
improvements
and new care
management
software and
modifications to
EMR

As pts
presented
for care,
clinical data
and
laboratory
test were
collected

DV1 – A1C
DV2 – weight

Diabetics:
DV1 –
Baseline
IG – M=8.4, SD = 0.4
CG – M=7.4, SD=0.4
12 months
IG –
M=7.5, SD=0.1
CG –
7.4, SD=0.5
Unadjusted
CI - -0.8 (-1.4,-0.3)
Adjusted
CI - -0.5 (-1.0, -0.04)

DV3 – BP
DV 4 – LDL
DV5 – BMI

Exclusion: had less than
12 month life expectancy,
non-English speaking,
lived in nursing or group
home, had substance

Paired t test,
McNemar’s
chi-square
test, StuartMaxwell
symmetry
test, linear
mixed effects
model, linear
regression

Findings/Results

DV6 - AU

DV2 –
Baseline
IG – M=234.1, SD = 8.3
CG – M=213.7, SD=6.9
12 months
IG –
M=230.4, SD=6.0
CG –
M=209.8, SD=9.0

Level or Evidence/Decision for
Use/Application to Practice
LOE – Level I
Strengths – RCT design
Weakness – only 10 practices
participated, which 5 received
intervention, variability in
baseline risks factors vs
comparison pt, especially BMI
and A1C for diabetics
Conclusions – Findings add to
the growing EB for the
effectiveness of CM as an
effective clinical practice with
regard to improving DM and
obesity related outcomes
Feasibility/Applicability –
findings consistent with
literature, recommended for
diabetic pts because of
significant improvements in

Key: A1C – hemoglobin A1C; ACE – angiotensin-converting enzyme inhibitor; ADA – American Diabetic Association; ARB – angiotensin-receptor blocker;
AU – microalbumin; BMI – body mass index; BP – blood pressure; CCC – chronic care coordinator; CCMo – chronic care model; CHC – community health
center; CHW – community health worker; CI – confidence interval; CM – care management; CMS – Centers for Medicare and Medicaid Services; Com. –
community; Cr – serum creatinine; DM – diabetes; DOHMH – Department of Health and Mental Hygiene; DSME – diabetes self-management education;
DSMS – diabetes self-management support; DV – dependent variable; Dx – diagnosis; EB – evidence based; EHR – electronic health record; EMR – electronic
medical record; Endo – endocrinology; HTN – hypertension; IV – independent variable; KDIS – Key Drivers Implementation Scales; LDL – low-density
lipoprotein; LOE – level of evidence; M – median; MCC – multiple chronic conditions; MEMS – medications event monitoring system; M/F – male/female; N
– sample; n – subgroup of sample size; N/A – not applicable; NIDDK – National Institute of Diabetes and Digestive and Kidney Diseases; Opht –
ophthalmology; PCMH – patient centered medical home; PC – primary care; PCP – primary care provider; PHQ-9 – Patient Health Questionnaire; POCT –
point of care testing; Pt – patient; QI – quality improvement; RCT – randomized controlled trial; SD – standard deviation; TP - telephone

DSME EFFECTS ON A1C AND DSM
Citation

Theory/Con
ceptual
Framework

Design/Method/P
urpose

Sample/Setting

abuse, psychiatric illness,
or cognitive impairment,
had DM or impaired
glucose tolerance due to
chronic glucocorticoid
use, polycystic ovary
syndrome, pituitary
lesion, or pancreatectomy.

16
Major Variables

Measurem Data Analysis
ents/Instru
mentation

Findings/Results

Unadjusted
CI – 0.2 (-9.1,9.5)
Adjusted
CI - -2.2 (-5.3,0.7)

Level or Evidence/Decision for
Use/Application to Practice
A1C, but will require training
and therefore, funding.

DV3 –
Baseline
IG – M=127.0, SD = 2.0
CG – M=127.5, SD=0.7
12 months
IG –
M=127.0, SD=2.6
CG – M=125.8, SD=3.7
Unadjusted
CI – 1.8 (-2.1,-5.7)
Adjusted
CI – 2.1 (-2.1, -6.2)
DV6 –
Baseline
IG – M=26.6, SD = 4.0
CG – M=24.3, SD=8.7
12 months
IG –
M=21.1, SD=4.5
CG –
27.9, SD=11.7
Unadjusted
CI - -9.1 (-26.3,8.1)
Adjusted
CI - -1.3 (-14.0, -11.4)

Key: A1C – hemoglobin A1C; ACE – angiotensin-converting enzyme inhibitor; ADA – American Diabetic Association; ARB – angiotensin-receptor blocker;
AU – microalbumin; BMI – body mass index; BP – blood pressure; CCC – chronic care coordinator; CCMo – chronic care model; CHC – community health
center; CHW – community health worker; CI – confidence interval; CM – care management; CMS – Centers for Medicare and Medicaid Services; Com. –
community; Cr – serum creatinine; DM – diabetes; DOHMH – Department of Health and Mental Hygiene; DSME – diabetes self-management education;
DSMS – diabetes self-management support; DV – dependent variable; Dx – diagnosis; EB – evidence based; EHR – electronic health record; EMR – electronic
medical record; Endo – endocrinology; HTN – hypertension; IV – independent variable; KDIS – Key Drivers Implementation Scales; LDL – low-density
lipoprotein; LOE – level of evidence; M – median; MCC – multiple chronic conditions; MEMS – medications event monitoring system; M/F – male/female; N
– sample; n – subgroup of sample size; N/A – not applicable; NIDDK – National Institute of Diabetes and Digestive and Kidney Diseases; Opht –
ophthalmology; PCMH – patient centered medical home; PC – primary care; PCP – primary care provider; PHQ-9 – Patient Health Questionnaire; POCT –
point of care testing; Pt – patient; QI – quality improvement; RCT – randomized controlled trial; SD – standard deviation; TP - telephone

DSME EFFECTS ON A1C AND DSM
Citation

Solorio et al.
(2014). Impact of
chronic care
coordinator
intervention on
diabetes of care
in a community
health center
Funding:
University of
Washington
Royal Research
Fund
Bias:
observational
study based on
retrospective
study and may
include bias due
to confounding
factors
Country: USA

Theory/Con
ceptual
Framework

Design/Method/P
urpose

CCM

Design:
Retrospective
cohort study
design
Purpose: to
evaluate the
impact of CCC
intervention on
quality of DM
care within the
CHC,
predominantly
low-income
Hispanic and nonHispanic white pt

Sample/Setting

N – 1,483
IG – 664
CG – 819
Demographics:
M Age – 50-59
M/F – 48.8%/ 51.2%

17
Major Variables

Measurem Data Analysis
ents/Instru
mentation

IV1 – at least 1
CCC visit, that
includes case
management, care
coordination, and
self-management

Data
collection
through
EMR

Setting: Sea Mar CHC
that provides PC services
to predominantly lowincome Hispanics and
non-Hispanic white pt in
the Washington area

DV1 – process of
care, including
A1C tested at
least twice taken
3 months apart,
LDL, AU, retinal
eye exam, and
foot exam

Inclusion: established dx
of DM type 2 in EMR in
the past 12 months,
current Sea Mar pt with
clinic visit between
2/1/2009 and 9/30/2009,
ages 18-69 years old, have
at least 2 visits at the same
clinic in last year, speak
English or Spanish

DV2 –
intermediate
outcomes of DM
care, including
A1C < 7.0 %,
LDL < 100
mg/dL, BP <
130/80 mmHg

Exclusion: older than 69
years old, DM type 1,
pregnant, history of organ
transplant, Cr 2.5 mg/dL,
dementia, and terminal
illness

DV3 – health
care utilization,
including number
of PC visits, at
least once referral
to opht, and at

Propensity
score analysis
to reduce
effect of
selection bias,
linear mixed
effects model
during 12
month preand
postenrolleme
nt , R
statistical
software, chisqaure test of
homogeneity,
two-sample ttest

Findings/Results

A1C –
Baseline
CG – M=8.0, SD= ±1.6
IG – M=8.4, SD= ±1.6
p<0.001
DV1 –
A1C measurements: CI 2.63(1.88, 3.68), p < 0.001;
AU screen: CI- 2.94 (2.07,
4.17), p < 0.001;
Retinal exam: CI - 2.27
(1.59, 3.25), p < 0.001;
Foot exam: CI - 5.22 (3.42,
7.98(, p < 0.001
DV2 –
A1C < 7%:
CI - 0.70 (0.39, 1.27), p =
0.242;
A1C last value: CI - 0.06
(0.02, 0.13, p = 0.151;
BP: CI - 0.99 (0.69, 1.42), p
= 0.968;

Level or Evidence/Decision for
Use/Application to Practice
LOE – Level IV
Strengths – large sample
Weakness – observational study
prone to bias, no data on BMI,
income, marital status,
employment, education, alcohol
use and time with DM, missing
weight and height on some
participants, data of duration of
CCC visits is missing
Conclusions – CCC is
suggested and may benefit pt
with DM type 2 by improving
receipt of DM services
Feasibility/Applicability – Due
to significant findings in
increases in DM services with
CCC, diabetic pt may benefit
from CCC. Therefore, making
use for CCC.

DV3 –
PCP visit: CI -1.39 (1.28,
1.51), p < 0.001;
Endo referral: CI - 0.88
(0.30 - 2.60), p = 0.818;
Opht referral: CI - 1.59
(0.86, 2.94), p = 0.142

Key: A1C – hemoglobin A1C; ACE – angiotensin-converting enzyme inhibitor; ADA – American Diabetic Association; ARB – angiotensin-receptor blocker;
AU – microalbumin; BMI – body mass index; BP – blood pressure; CCC – chronic care coordinator; CCMo – chronic care model; CHC – community health
center; CHW – community health worker; CI – confidence interval; CM – care management; CMS – Centers for Medicare and Medicaid Services; Com. –
community; Cr – serum creatinine; DM – diabetes; DOHMH – Department of Health and Mental Hygiene; DSME – diabetes self-management education;
DSMS – diabetes self-management support; DV – dependent variable; Dx – diagnosis; EB – evidence based; EHR – electronic health record; EMR – electronic
medical record; Endo – endocrinology; HTN – hypertension; IV – independent variable; KDIS – Key Drivers Implementation Scales; LDL – low-density
lipoprotein; LOE – level of evidence; M – median; MCC – multiple chronic conditions; MEMS – medications event monitoring system; M/F – male/female; N
– sample; n – subgroup of sample size; N/A – not applicable; NIDDK – National Institute of Diabetes and Digestive and Kidney Diseases; Opht –
ophthalmology; PCMH – patient centered medical home; PC – primary care; PCP – primary care provider; PHQ-9 – Patient Health Questionnaire; POCT –
point of care testing; Pt – patient; QI – quality improvement; RCT – randomized controlled trial; SD – standard deviation; TP - telephone

DSME EFFECTS ON A1C AND DSM
Citation

Swietek et al.
(2018). Do
medical homes
improve quality
of care for
persons with
multiple chronic
conditions?
Funding:
Agency for
Healthcare
Research and
Quality
Bias: regression
model used to
reduce bias
Country: USA

Theory/Con
ceptual
Framework

PCMH

Design/Method/P
urpose

Design: quasiexperimental
Purpose:
examine the
association
between PCMH
enrollment and
receipt of diseasespecific quality
measures for
nonelderly
Medicaid
beneficiaries

Sample/Setting

N – 208,122
IG – 145,145
CG – 62,977
Demographics:
M Age – 43.91
M/F – 32.4%/ 67.6%
Setting: Com. Care of
North Carolina, regional
PC

18
Major Variables

least 1 referral to
endo
IV1 – PCMH
enrollment
DV1 – A1C
testing
DV2 – attention
for nephropathy
DV3 – liver
function tests

Inclusion: ages 18-64
years old; at least 2
chronic conditions that
included: DM, asthma,
hyperlipidemia,
hypertension, major
depression and
schizophrenia; pt with at
least partial Medicaid
eligibility; have at least 2
outpatient or emergency
department visits or at
least 1 inpatient visit for
given condition

DV4 – eye
examinations

Exclusion: Dual
Medicare and Medicaid
enrollees

DV8 – any
psychotherapy

DV5 – Lipid
profile
DV6 – ACE or
ARB
DV7 – SABA
overuse, which is
4+ canister
equivalents in 3
months

Measurem Data Analysis
ents/Instru
mentation

Dataset that
links
Medicaid
claims with
other
administrati
ve data
sources

t-test, chisquare, linear
probability
model, fixedeffects model

Findings/Results

DV1 –
CG – M=61.5
IG – M=82.1
p<0.001
DV2 –
CG – M=30.3
IG – M=43.5
p<0.001
DV3 –
CG – M=20.7
IG – M=25.4
p<0.001
DV4 –
CG – M=30.0
IG – M=44.2
p<0.001

Level or Evidence/Decision for
Use/Application to Practice

LOE – Level III
Strengths – large sample
Weakness – Not generalized
population, PCHM was only
defined as any enrollment in a
year which may not capture the
effects of extended duration of
PCMH
Conclusions – PCMH model
may improve quality of care for
pt with MCC
Feasibility/Applicability –
Significant findings show
PCMH could have benefits to pt
with MCC, which shows
feasibility.

DV5 –
CG – M=51.0
IG – M=70.72
p<0.001
DV6 –
CG – M=53.3
IG – M=78.6
p<0.001

Key: A1C – hemoglobin A1C; ACE – angiotensin-converting enzyme inhibitor; ADA – American Diabetic Association; ARB – angiotensin-receptor blocker;
AU – microalbumin; BMI – body mass index; BP – blood pressure; CCC – chronic care coordinator; CCMo – chronic care model; CHC – community health
center; CHW – community health worker; CI – confidence interval; CM – care management; CMS – Centers for Medicare and Medicaid Services; Com. –
community; Cr – serum creatinine; DM – diabetes; DOHMH – Department of Health and Mental Hygiene; DSME – diabetes self-management education;
DSMS – diabetes self-management support; DV – dependent variable; Dx – diagnosis; EB – evidence based; EHR – electronic health record; EMR – electronic
medical record; Endo – endocrinology; HTN – hypertension; IV – independent variable; KDIS – Key Drivers Implementation Scales; LDL – low-density
lipoprotein; LOE – level of evidence; M – median; MCC – multiple chronic conditions; MEMS – medications event monitoring system; M/F – male/female; N
– sample; n – subgroup of sample size; N/A – not applicable; NIDDK – National Institute of Diabetes and Digestive and Kidney Diseases; Opht –
ophthalmology; PCMH – patient centered medical home; PC – primary care; PCP – primary care provider; PHQ-9 – Patient Health Questionnaire; POCT –
point of care testing; Pt – patient; QI – quality improvement; RCT – randomized controlled trial; SD – standard deviation; TP - telephone

DSME EFFECTS ON A1C AND DSM
Citation

Theory/Con
ceptual
Framework

Design/Method/P
urpose

Sample/Setting

19
Major Variables

Measurem Data Analysis
ents/Instru
mentation

DV9 – assertive
community
therapy

Chamany et al.
(2015). TP
intervention to
improve DM
control: a
randomized trial
in New York City
A1c registry.
Funding: Albert
Einstein College
of Medicine
Bias: None
identified
Country: USA

CCM

Design: RCT
Purpose: 1) to
evaluate the
incremental effect
of patientcentered TP
intervention on
the M A1C levels
beyond that
achieved with
print materials
mailed to pts and
providers by the
DOHMH registry
intervention; 2)
determine what
patient
demographic and
psychosocial
factors mediate
the effect of the
interventions; and
3) provide
estimates of
implementation
costs of the TP

N – 941
IG – 443
CG – 498
Demographics:
M Age – 56.3, SD 11.7
M/F – 36.3%/ 63.7%

IV1 –
Telephonic:
between 4-8
phone calls each
year for health
behavior
counseling to
improve A1C

Setting: South Bronx
Inclusion: pts with DM
who speak English and/or
Spanish and reside in the
South Bronx; > 18 years,
with DM, who become
part of the NYC registry
by virtue of having a
reported A1C >7% to the
DOHMH
Exclusion: < 18 years;
A1C < = 7 %; refuses
informed consent and
HIPAA consent; cognitive
dysfunction as assessed by
TP; does not read or speak

CG – standard
registry: letters
from the
DOHMH to
promote
improved A1C
and give lists of
Bronx resources
for healthier food
and activities
DV1 – A1C
DV2 – DM selfcare activities

Findings/Results

Level or Evidence/Decision for
Use/Application to Practice

DV7 –
CG – M=7.8
IG – M=10.4
p<0.001
DOHMH
Registry;
self-report;
Morisky
Medication
Adherence
four-item
scale;
Summary
of Diabetes
Self-Care
Activities
; PHQ-9;
Well-Being
scale of the
WHO

Two-tailed ztest; Mann–
Whitney U te
st; Sobel test;
Stata, version
12.1 MP

DV1 –
Baseline
IG – M=9.3, SD = 2.1,
n=443
CG – M=9.1, SD=2.0,
n=498
12 months
IG – M=8.4, SD=1.9,
n=334,
CG –
8.6, SD=2.0, n=360
Statistically significant, p
<0.05

LOE – Level II
Strengths – randomized
Weakness – missing primary
outcome data for 26.3% of
participants; not generalized and
focuses on low-incomes, mostly
Latinos with DM with TP access
Conclusions – TP intervention
delivered by health educators
can be an effective tool to
improve DM control in diverse
populations, specifically for
those with worse metabolic
control identified using a
registry.
Feasibility/Applicability – The
intervention is low cost and lowintensive, making it feasible and
applicable.

Key: A1C – hemoglobin A1C; ACE – angiotensin-converting enzyme inhibitor; ADA – American Diabetic Association; ARB – angiotensin-receptor blocker;
AU – microalbumin; BMI – body mass index; BP – blood pressure; CCC – chronic care coordinator; CCMo – chronic care model; CHC – community health
center; CHW – community health worker; CI – confidence interval; CM – care management; CMS – Centers for Medicare and Medicaid Services; Com. –
community; Cr – serum creatinine; DM – diabetes; DOHMH – Department of Health and Mental Hygiene; DSME – diabetes self-management education;
DSMS – diabetes self-management support; DV – dependent variable; Dx – diagnosis; EB – evidence based; EHR – electronic health record; EMR – electronic
medical record; Endo – endocrinology; HTN – hypertension; IV – independent variable; KDIS – Key Drivers Implementation Scales; LDL – low-density
lipoprotein; LOE – level of evidence; M – median; MCC – multiple chronic conditions; MEMS – medications event monitoring system; M/F – male/female; N
– sample; n – subgroup of sample size; N/A – not applicable; NIDDK – National Institute of Diabetes and Digestive and Kidney Diseases; Opht –
ophthalmology; PCMH – patient centered medical home; PC – primary care; PCP – primary care provider; PHQ-9 – Patient Health Questionnaire; POCT –
point of care testing; Pt – patient; QI – quality improvement; RCT – randomized controlled trial; SD – standard deviation; TP - telephone

DSME EFFECTS ON A1C AND DSM
Citation

Edelman et al.
(2015). Nurse-led
behavioral
management of
DM and HTN in
the com.
practices: a
randomized trial.
Funding:
NIDDK
Bias: None
identified
Country: USA

Theory/Con
ceptual
Framework

CCM

Design/Method/P
urpose
intervention for
comparison with
the print
intervention.
Design: RCT
Purpose: To
assess the
effectiveness of
nurse behavioral
management of
DM and HTN in
com. practices
among pts with
both diseases.

Sample/Setting

20
Major Variables

Measurem Data Analysis
ents/Instru
mentation

IV1 – 12 calls
over 2 years:
from a nurse
experienced in
DM and HTN
management;
calls were
tailored to pts’s
DM- and HTNbehavioral
barriers

Clinical
data from
visits and
POCT

Findings/Results

Level or Evidence/Decision for
Use/Application to Practice

English or Spanish; no
DM
N – 377
IG – 193
CG – 184
Demographics:
M Age – 59.6, SD – 10.7
M/F – 45.1%/ 54.9%
Setting: Practice-based
research network of com.
PC practices
Inclusion: adult pts with
both DM 2 and HTN and
receiving care at 1 of 9
com. fee-for-service
practices; A1C ≥ 7.5%
but could have wellcontrolled HTN and had
to be taking medications
for both

CG – 12 calls not
tailored or
interactive: calls
involved health
issues unrelated
to DM or HTN

Exclusion: DM type 1;
inability to receive a
telephone intervention in
English, participations in
another diabetes or HTN

DV2 – BP: taken
at each visit, 2
measures 5
minutes apart and
were averaged

DV1 – A1C:
measured by
fingerstick

Linear mixed
model;
covariance
model;
Wilcox ranksum test;
generalized
estimating
equation
model

DV1 –
Baseline
IG – M=9.2, SD = 1.5,
n=193
CG – M=9.0, SD=1.4,
n=184
24 months
IG – M=8.6,
CG – 8.5
CI (-0.3%, 0.5%), p=0.50 –
not significant

LOE – Level II
Strengths – blinded,
randomized
Weakness – intervention was
ineffective
Conclusions – telephonic nurse
case management did not lead
to improvement in A1c or SBP.
Feasibility/Applicability –
Small gains in clinical
outcomes may add up to an
important public health impact
over a large population, the
study of a modest intervention
by traditional trial methods may
not be feasible.

Key: A1C – hemoglobin A1C; ACE – angiotensin-converting enzyme inhibitor; ADA – American Diabetic Association; ARB – angiotensin-receptor blocker;
AU – microalbumin; BMI – body mass index; BP – blood pressure; CCC – chronic care coordinator; CCMo – chronic care model; CHC – community health
center; CHW – community health worker; CI – confidence interval; CM – care management; CMS – Centers for Medicare and Medicaid Services; Com. –
community; Cr – serum creatinine; DM – diabetes; DOHMH – Department of Health and Mental Hygiene; DSME – diabetes self-management education;
DSMS – diabetes self-management support; DV – dependent variable; Dx – diagnosis; EB – evidence based; EHR – electronic health record; EMR – electronic
medical record; Endo – endocrinology; HTN – hypertension; IV – independent variable; KDIS – Key Drivers Implementation Scales; LDL – low-density
lipoprotein; LOE – level of evidence; M – median; MCC – multiple chronic conditions; MEMS – medications event monitoring system; M/F – male/female; N
– sample; n – subgroup of sample size; N/A – not applicable; NIDDK – National Institute of Diabetes and Digestive and Kidney Diseases; Opht –
ophthalmology; PCMH – patient centered medical home; PC – primary care; PCP – primary care provider; PHQ-9 – Patient Health Questionnaire; POCT –
point of care testing; Pt – patient; QI – quality improvement; RCT – randomized controlled trial; SD – standard deviation; TP - telephone

DSME EFFECTS ON A1C AND DSM
Citation

Theory/Con
ceptual
Framework

Design/Method/P
urpose

Sample/Setting

21
Major Variables

Measurem Data Analysis
ents/Instru
mentation

IV1 – DM
knowledge/infor
mation: 12 DM
education
modules over 12
week period
based on
guidelines from
ADA

EMR and
clinical
visits

Findings/Results

Level or Evidence/Decision for
Use/Application to Practice

study, or living in an
assisted living facility.
Egede. (2017)
Telephonedelivered
behavioral skills
intervention for
African American
adults with type 2
DM: an RCT
Funding:
National Institute
of
Health/NIDDK
Bias: None
identified
Country: USA

Informationmotivation
behavioral
skills model

Design: RCT
Purpose: To
assess the
efficacy of a
combined
telephonedelivered
education and
behavioral skills
intervention in
reducing
hemoglobin A1C
levels in African
Americans with
type 2 DM

N – 255
IG – knowledge: 63,
skills: 65, combined: 63
CG – 64
Demographics:
M Age – 50-64
M/F – 55.3%/44.7 %
Setting: Medical
University of South
Carolina (general internal
medicine, endo, family
medicine, and com. PC
clinics) and the Ralph H.
Johnson Veterans
Administration Medical
Center, both located in
Charleston, South
Carolina.
Inclusion: ≥18 years old;
dx of type 2 DM and A1C
≥9% at screening visit;
self-identified as Black or
African American; taking
at least 1 oral medication
for DM, HTN, or
hyperlipidemia and must
be willing to use the

IV2 –
motivation/behavi
oral: pt
activation, pt
empowerment,
and behavioral
skills training
delivered via 30
minute phone call
ever week for 12
weeks
IV3 – combined:
receives weekly
telephonedelivered DM
knowledge/infor
mation, pt
activation, pt
empowerment,

Chi-square;
ANOVA;
ANCOVA;
longitudinal
model

DV1 –
Baseline
IG –
Knowledge:
M=9.3, SD = 1.5, n=63
Skills: M=9.2, SD = 2.1,
n=65
Combination: M=9.2, SD =
1.9, n=63
CG – M=9.3, SD=2.1, n=64
12 months (Differences in
levels of A1C)
IG –
Knowledge:
CI – 0.49(-0.13, 1.11),
p=0.123 – not significant;
Skills: CI – 0.23(-0.38,
0.83), p=0.456 – not
significant;
Combination: CI – 0.48(0.10, 1.07), p=0.105 – not
significant
CG – reference group

LOE – Level II
Strengths – targets vulnerable
population; no RCT in this
populations; telephone calls are
efficacious
Weakness – eligibility between
screening time and baseline visit
varied causing drop in eligible
pts; staff turnover was high
during study, especially among
health educators
Conclusions – combined
education and skills training did
not achieve greater reductions in
A1C at 12 months compared to
CG, educations alone, or skills
training alone.
Feasibility/Applicability –
Because telephone calls are low
cost and nursing staff that are
not mastered prepared are doing
education makes this study
feasible. Modifications must be
made to show significant
changes in A1C.

Key: A1C – hemoglobin A1C; ACE – angiotensin-converting enzyme inhibitor; ADA – American Diabetic Association; ARB – angiotensin-receptor blocker;
AU – microalbumin; BMI – body mass index; BP – blood pressure; CCC – chronic care coordinator; CCMo – chronic care model; CHC – community health
center; CHW – community health worker; CI – confidence interval; CM – care management; CMS – Centers for Medicare and Medicaid Services; Com. –
community; Cr – serum creatinine; DM – diabetes; DOHMH – Department of Health and Mental Hygiene; DSME – diabetes self-management education;
DSMS – diabetes self-management support; DV – dependent variable; Dx – diagnosis; EB – evidence based; EHR – electronic health record; EMR – electronic
medical record; Endo – endocrinology; HTN – hypertension; IV – independent variable; KDIS – Key Drivers Implementation Scales; LDL – low-density
lipoprotein; LOE – level of evidence; M – median; MCC – multiple chronic conditions; MEMS – medications event monitoring system; M/F – male/female; N
– sample; n – subgroup of sample size; N/A – not applicable; NIDDK – National Institute of Diabetes and Digestive and Kidney Diseases; Opht –
ophthalmology; PCMH – patient centered medical home; PC – primary care; PCP – primary care provider; PHQ-9 – Patient Health Questionnaire; POCT –
point of care testing; Pt – patient; QI – quality improvement; RCT – randomized controlled trial; SD – standard deviation; TP - telephone

DSME EFFECTS ON A1C AND DSM
Citation

Theory/Con
ceptual
Framework

Design/Method/P
urpose

Sample/Setting

MEMS cap and bottle for
1 year; speak English;
access to a telephone for
the 12 week period
Exclusion: mental
confusion; participations
in other DM clinical trials,
alcohol/drug
abuse/dependence; active
psychosis or acute mental
disorder; life expectancy
< 6 months.

Halladay et al.
(2014) More
extensive
implementation
of the CCM is
associated with
better lipid
control in DM.
Funding:
Agency of
Healthcare
Research and

CCM

Design:
observational
study

N – 42 practices
IG – N/A
CG – N/A

Purpose: This
study examines
whether higher
KDIS scores are
associated with
improved diabetes
outcomes.

Setting: 42 PC practices
in North Carolina
Inclusion: participated
with a practice coach for
at least 13 months starting
in February 2008 or later;
submitted clinical data
reports in months

22
Major Variables

Measurem Data Analysis
ents/Instru
mentation

Findings/Results

Level or Evidence/Decision for
Use/Application to Practice

and behavioral
skills
CG – standard
care with general
health education
DV1 – A1C at 12
months
DV2 – costeffectiveness and
change in
physical activity,
diet, medication
adherence, and
self-monitoring
of blood glucose
in 12 months
IV1 – 4 key
drivers: registries,
planned care
template,
protocols, and
self-management
support
CG – standard
practice: without
drivers

Clinical
data and
KDIS data

Logistic
regression;
odds ratio;
extrabinomial
variation in
linear model

DV1 –
Baseline
IG – 23 (37%), n=42
12 months
IG – 4 – not significant

LOE – Level IV
Strengths – innovative
approach for QI
Weakness – Short length of data
(2-3 years), was not significant
Conclusions – Practices that
implement key changes may
achieve improved patient
outcomes in LDL control among
their pts with diabetes.

Key: A1C – hemoglobin A1C; ACE – angiotensin-converting enzyme inhibitor; ADA – American Diabetic Association; ARB – angiotensin-receptor blocker;
AU – microalbumin; BMI – body mass index; BP – blood pressure; CCC – chronic care coordinator; CCMo – chronic care model; CHC – community health
center; CHW – community health worker; CI – confidence interval; CM – care management; CMS – Centers for Medicare and Medicaid Services; Com. –
community; Cr – serum creatinine; DM – diabetes; DOHMH – Department of Health and Mental Hygiene; DSME – diabetes self-management education;
DSMS – diabetes self-management support; DV – dependent variable; Dx – diagnosis; EB – evidence based; EHR – electronic health record; EMR – electronic
medical record; Endo – endocrinology; HTN – hypertension; IV – independent variable; KDIS – Key Drivers Implementation Scales; LDL – low-density
lipoprotein; LOE – level of evidence; M – median; MCC – multiple chronic conditions; MEMS – medications event monitoring system; M/F – male/female; N
– sample; n – subgroup of sample size; N/A – not applicable; NIDDK – National Institute of Diabetes and Digestive and Kidney Diseases; Opht –
ophthalmology; PCMH – patient centered medical home; PC – primary care; PCP – primary care provider; PHQ-9 – Patient Health Questionnaire; POCT –
point of care testing; Pt – patient; QI – quality improvement; RCT – randomized controlled trial; SD – standard deviation; TP - telephone

DSME EFFECTS ON A1C AND DSM
Citation

Theory/Con
ceptual
Framework

Design/Method/P
urpose

Quality/National
Institutes of
Health/National
Institute of
Environmental
Health Sciences

Country: USA
Carrasquillo et al.
(2017). Effect of
a com. health
worker
intervention
among Latinos
with poorly
controlled type 2
DM.
Funding:
National Heart,
Blood, and Lung
Institute, National
Center for
Advancing
Translational
Sciences and the
National
Institutes on

Sample/Setting

Major Variables

10,11,12, and submitted
another clinical date
report at some point
during their second year
of participation with their
coach.

DV1 – number of
practices with pt
at with A1C < 9%

Exclusion: Not noted

Bias: Lack of
study design may
lead to bias.

23
Measurem Data Analysis
ents/Instru
mentation

Findings/Results

Level or Evidence/Decision for
Use/Application to Practice

Feasibility/Applicability –
Needs stronger study design to
be feasible and applicable.

DV2 – number of
practices with pt
with BP <130/80
DV3 – number of
practices with pt
with LDL <100

CCM

Design: RCT
Purpose: To
compare a CHW
intervention with
enhanced
usual care

N – 300
IG – 150
CG – 150
Demographics:
M Age – 55.2, SD – 7.0
M/F – 45%/ 55%

IV1 – A 1-year
CHW
intervention
consisted of home
visits, telephone
calls, and grouplevel activities.

Setting: 2 public hospital
outpatient clinics in
Miami-Dade County,
Florida

CG – enhanced
usual care

Inclusion: A1C >8.0%

DV2 – LDL

Exclusion: dx with type 2
DM < 6 months
previously, self-reported
type 1 DM, dx with type 2

DV1 – SBP

DV3 – A1C

EMR,
telephone
calls

2-tailed t test,
generalized
estimating
equation
model, chisquared test

DV3 –
Baseline
IG – M=9.3, SD = 2.1,
n=150
CG – M=9.3 SD=1.9, n=150
12 months (Adjusted)
IG – CI - -0.51% (-0.94, 0.09) - significant

LOE – Level II
Strengths – single-blinded
RCT, correlates with previous
evidence
Weakness – does not provide
evidence on which part of the
intervention helped lower A1C
Conclusions – Both groups
showed a statistically significant
reduction of HbA1c at 6 and 12
months following baseline.
Feasibility/Applicability –
Although CHW are not
expensive compared to the
average diabetic treatment,

Key: A1C – hemoglobin A1C; ACE – angiotensin-converting enzyme inhibitor; ADA – American Diabetic Association; ARB – angiotensin-receptor blocker;
AU – microalbumin; BMI – body mass index; BP – blood pressure; CCC – chronic care coordinator; CCMo – chronic care model; CHC – community health
center; CHW – community health worker; CI – confidence interval; CM – care management; CMS – Centers for Medicare and Medicaid Services; Com. –
community; Cr – serum creatinine; DM – diabetes; DOHMH – Department of Health and Mental Hygiene; DSME – diabetes self-management education;
DSMS – diabetes self-management support; DV – dependent variable; Dx – diagnosis; EB – evidence based; EHR – electronic health record; EMR – electronic
medical record; Endo – endocrinology; HTN – hypertension; IV – independent variable; KDIS – Key Drivers Implementation Scales; LDL – low-density
lipoprotein; LOE – level of evidence; M – median; MCC – multiple chronic conditions; MEMS – medications event monitoring system; M/F – male/female; N
– sample; n – subgroup of sample size; N/A – not applicable; NIDDK – National Institute of Diabetes and Digestive and Kidney Diseases; Opht –
ophthalmology; PCMH – patient centered medical home; PC – primary care; PCP – primary care provider; PHQ-9 – Patient Health Questionnaire; POCT –
point of care testing; Pt – patient; QI – quality improvement; RCT – randomized controlled trial; SD – standard deviation; TP - telephone

DSME EFFECTS ON A1C AND DSM
Citation

Theory/Con
ceptual
Framework

Design/Method/P
urpose

Minority Health
and Health
Disparities

Funding:
National
Coordinator for
Health
Information
Technology,
North Carolina
Regional
Extension Center
Cooperative
Agreement, The
North Carolina
Health and
Wellness Trust
Fund

Major Variables

Measurem Data Analysis
ents/Instru
mentation

Findings/Results

DM younger than 25
years old, were enrolled in
intervention studies,
planned to move from the
county within the next
year

Bias: None noted
Country: USA
Cykert et al.
(2016).
Meaningful use in
chronic care
improved DM
outcomes using
PC extension
center model

Sample/Setting

24

Primary care
extension
center
model/CCM

Design: cohort
study
Purpose: to
evaluate the
effectiveness QI
of EHR on
diabetes

N – 50 practices
IG – 50
CG – N/A
Demographics:
N/A
Setting:
Inclusion: practices that
signed up for Regional
Extension Center for
Health Information
Technology services and
agreed to implement a
certified EHR system,
perform QI through onsite
practice facilitation using
DM chronic care
measures, and work
toward achievement of
CMS-defined meaningful
use of their certified EHR.

Level or Evidence/Decision for
Use/Application to Practice
insurance plans may not cover
their services.

IV1 – QI:
provided to
practices with a
coach and
practice team
engagement at the
site, or web-based
communication
DV1 – percentage
of diabetic pts
who achieved
A1C< 7%
DV2 – percentage
who remained
with HGB A1C >
9% for each
practice site

EMR,
onsite
practice
facilitation

Bivariate
analysis,
linear
regression
model, KDIS
scores

DV1 –
Baseline
IG – M=41.6, SD = 16.7,
n=50
6 months (EHR + practice
facilitation)
IG – M = 51.3, SD = 16.0,
n=45
6 moths (HER +practice
facilitation + Meaningful
Use
IG – M = 60.0, SD = 11.6,
n=29
DV2 –
Baseline
IG – M=21.6, SD = 11.8,
n=50
6 months (EHR + practice
facilitation)
IG – M = 20.1, SD = 13.3,
n=45
6 moths (EHR +practice
facilitation + Meaningful
Use

LOE – Level IV
Strengths – QI proven to be
successful in DM management
Weakness – No control
Conclusions – Practice
facilitation that provided EHR
and QI coaching support showed
important improvements in
diabetes outcomes in practices
that achieved meaningful use of
their EHR systems.
Feasibility/Applicability – if
grant money can be rewarded
this is feasible. Study is
applicable since HER are highly
used in practices.

Key: A1C – hemoglobin A1C; ACE – angiotensin-converting enzyme inhibitor; ADA – American Diabetic Association; ARB – angiotensin-receptor blocker;
AU – microalbumin; BMI – body mass index; BP – blood pressure; CCC – chronic care coordinator; CCMo – chronic care model; CHC – community health
center; CHW – community health worker; CI – confidence interval; CM – care management; CMS – Centers for Medicare and Medicaid Services; Com. –
community; Cr – serum creatinine; DM – diabetes; DOHMH – Department of Health and Mental Hygiene; DSME – diabetes self-management education;
DSMS – diabetes self-management support; DV – dependent variable; Dx – diagnosis; EB – evidence based; EHR – electronic health record; EMR – electronic
medical record; Endo – endocrinology; HTN – hypertension; IV – independent variable; KDIS – Key Drivers Implementation Scales; LDL – low-density
lipoprotein; LOE – level of evidence; M – median; MCC – multiple chronic conditions; MEMS – medications event monitoring system; M/F – male/female; N
– sample; n – subgroup of sample size; N/A – not applicable; NIDDK – National Institute of Diabetes and Digestive and Kidney Diseases; Opht –
ophthalmology; PCMH – patient centered medical home; PC – primary care; PCP – primary care provider; PHQ-9 – Patient Health Questionnaire; POCT –
point of care testing; Pt – patient; QI – quality improvement; RCT – randomized controlled trial; SD – standard deviation; TP - telephone

DSME EFFECTS ON A1C AND DSM
Citation

Theory/Con
ceptual
Framework

Design/Method/P
urpose

Bias: None noted

Funding:
Bristol-Myers
Squibb
Foundation

Major Variables

Measurem Data Analysis
ents/Instru
mentation

CCMo

Design: empirical
case study,
retrospective

N – 173
IG – 173
CG – N/A

Purpose: to
measure the
implementation
and effects of a
multisite
coordinated care
approach that
delivered DSME
and DSMS for
disadvantaged pts

Demographics:
N/A
Setting: 4 PMCHs in
Jacksonville, Florida,
Athens County, Ohio,
Oklahoma City,
Oklahoma, and Nashville,
Tennessee

Bias: None
identified

Inclusion: PCMH had to
be a part of Together on
DM

Country: USA

Exclusion: Not noted

Findings/Results

Level or Evidence/Decision for
Use/Application to Practice

IG – M = 15.4, SD = 6.2,
n=29

Exclusion: practices that
had participated in QI
programs

Country: USA
Sepers et al.
(2015).
Measuring the
implementation
and effects of a
coordinated care
model featuring
DSME within 4
PCMH.

Sample/Setting

25

IV1 – DSME and
coordinated care:
accredited DSME
program with pttailored curricula,
DSMS that
targets unique
needs of
underserved
populations,
enhanced access
and linkage to
care services, and
practice changes
aimed at
improving quality
of DM clinical
care
CG – N/A
DV1 – A1C

SPSS
Statistics
for
Windows,

Pairedsample t test,
Pearson
productmoment
correlation
coefficient

DV1 –
Baseline
IG – M=9.1, SD = 2.4
6 months
IG – M=8.5, SD = 2.1
p = 0.01, significant

LOE – Level IV
Strengths – pt and staff
satisfaction implementing
intervention
Weakness – no control group
Conclusions – DSME and
DSMS within coordinated care
settings have the potential to
improve PCMH practice and
associated clinical health
outcomes for populations
experiencing health disparities.
Feasibility/Applicability – pts
and staff shared high satisfaction
with DSME within the PCMH
setting, making this intervention
applicable. Testing of the
intervention at multiple sites can
be costly.

DV2 – BMI
DV3 – BP
DV4 - LDL

Key: A1C – hemoglobin A1C; ACE – angiotensin-converting enzyme inhibitor; ADA – American Diabetic Association; ARB – angiotensin-receptor blocker;
AU – microalbumin; BMI – body mass index; BP – blood pressure; CCC – chronic care coordinator; CCMo – chronic care model; CHC – community health
center; CHW – community health worker; CI – confidence interval; CM – care management; CMS – Centers for Medicare and Medicaid Services; Com. –
community; Cr – serum creatinine; DM – diabetes; DOHMH – Department of Health and Mental Hygiene; DSME – diabetes self-management education;
DSMS – diabetes self-management support; DV – dependent variable; Dx – diagnosis; EB – evidence based; EHR – electronic health record; EMR – electronic
medical record; Endo – endocrinology; HTN – hypertension; IV – independent variable; KDIS – Key Drivers Implementation Scales; LDL – low-density
lipoprotein; LOE – level of evidence; M – median; MCC – multiple chronic conditions; MEMS – medications event monitoring system; M/F – male/female; N
– sample; n – subgroup of sample size; N/A – not applicable; NIDDK – National Institute of Diabetes and Digestive and Kidney Diseases; Opht –
ophthalmology; PCMH – patient centered medical home; PC – primary care; PCP – primary care provider; PHQ-9 – Patient Health Questionnaire; POCT –
point of care testing; Pt – patient; QI – quality improvement; RCT – randomized controlled trial; SD – standard deviation; TP - telephone

CHRONIC CARE MANAGEMENT AND A1C

26
Appendix B

Table 2
Synthesis Table
Author

Holtrop

Solorio

Swietek

Chamany

Edelman

Egede

Year

2017

2014

2018

2015

2015

2016

LOE

II

IV

III

II

II

II

Design

RCT

CS

QE

RCT

RCT

Sample Size

1,392

1,483

208,122

941

377

Age (Mean)

54.8

50-59

43.9

56.3

% Male

50.4

48.8

32.4

36.3

Improve A1C

X

X+

X+

X+

CCC

X

X

X

Staff ∆

X

EMR ∆

X

Demographics
59.6
45.1
Findings
X

Halladay

Cykert

Sepers

2014

Carrasquil
lo
2017

2016

2015

IV

II

IV

IV

RCT

OS

RCT

CS

CC

255

42

300

50

173

50-64

N/A

55.2

N/A

N/A

55.3

N/A

45

N/A

N/A

X

X+

X+

Interventions

DSME
TP Call
Education
Registries

X

X

X
X

X

X
X

X

X
X

X

X
X

Home Visits

X

Group
Activities
CHW

X
X

Key: A1C – hemoglobin A1C; CC – controlled case study; CCC – chronic care coordinator; CHW – community health worker; CS – Cohort study; DSME –
diabetes self-management education; EMR – electronic medical record; LOE – level of evidence; N/A – not applicable; OS – observational study; QE – quasiexperimental; RCT – randomized controlled trial; TP – telephone; + - significantly improved; ∆ - modifications

CHRONIC CARE MANAGEMENT AND A1C
Appendix C
Figure 1
Chronic Care Model

27

DSME EFFECTS ON A1C AND DSM

28
Appendix D

Figure 2
Rosswurm and Larabee’s Model

DSME EFFECTS ON A1C AND DSM

29
Appendix E

Table 3
Demographics
Variable
RACE/ETHINICITY
HISPANIC/LATINO/SPANISH
WHITE
Missing
HISTORY OF DIABETES DIAGNOSIS
>10 YEARS
1-5 YEARS
0-1 YEAR
5-10 YEARS
Missing
GENDER
MALE
FEMALE
Missing
AGE
45-54
>65
Missing
EMPLOYMENT
UNEMPLOYED
RETIRED
Missing
EDUCATION
HIGH SCHOOL
NO FORMAL
Missing
Note. Due to rounding errors, percentages may not equal 100%.

n

%

8
1
0

88.89
11.11
0

5
1
2
1
0

55.56
11.11
22.22
11.11
0

8
1
0

88.89
11.11
0

4
5
0

44.44
55.56
0

2
7
0

22.22
77.78
0

7
2
0

77.78
22.22
0

DSME EFFECTS ON A1C AND DSM

30
Appendix F
Budget

Phase
Preparation

Activities
Print copies of project
overview for staff (qty 30)
Print copies of consent,
evaluation, and educational
material for participants (qty
30)
Educational session at clinic
for staff for 30 min:
site
snacks
time of presenter (project
director)

Delivery

Site

Evaluation

Educational session (project
director)
Monthly phone calls by
project director (30 min/call
x 3 months)
Front staff scheduling patient
for visit(10 min/call x 30
patients)
Review and analysis of
results (10 hours plus
software)

Cost

Subtotal

$0.60 x 30

$18

$3 x 30

$90

Total

$30

$0
$15
$15

$0

$0

$15 x 20
hours
$15 x 30
hours

$300

$12 x 5 hours

$60

$20 x 10
hours + $60

$260

$450

$1,208

Budget Justification: Potential revenue and benefits of project exceeds costs. Decreasing A1c
levels could decrease number of diabetes related visits to hospital and emergency room visits.
Alongside, meeting quality measures set forth by Yuma Regional Medical Center.
Possible funding: Transitional Care will fund part of the costs, such as site and front staff. Project
director will volunteer time and provide funding for all other cost.