Running head: DIABETES EDUCATION

1

Improving Glycemic Control in Patients with Type 2 Diabetes through Formal Education
Jessica Dixon
Arizona State University

Mentor: Judith Ochieng PhD, DNP, MSN-ED, RN, FNP-BC

DIABETES EDUCATION

2

Abstract
Background and Purpose: Over 30 million people in the United States (U.S.) have diabetes
mellitus, which comprises about 9% of the population, and about 90% of individuals with
diabetes have type 2 diabetes (Centers for Disease Control and Prevention [CDC], 2017). Adults
with type 2 diabetes at a local internal medicine clinic were consistently having high glycated
hemoglobin (HbA1C) levels, demonstrated by data collected from the electronic health record
(EHR), and there was no ordering process for referring patients to diabetes management
education and support (DSMES) services. The purpose of this project was to improve glycemic
control, demonstrated by lower HbA1C levels, and reach a diabetes education attendance rate of
62.5% at an internal medicine clinic in Chandler, Arizona.
Methods: An electronic health record (EHR) template was created and brief staff training was
completed to connect patients with diabetes in the community to a local formal diabetes
education program. HbA1C levels were measured before and three months after adults with type
2 diabetes mellitus (T2DM) received physicians’ orders for a DSMES program, and rates of
attendance to the program were calculated. Data was collected through the EHR and through
feedback from the DSMES program. Descriptive statistics were used in data analysis.
Outcomes: The participants’ results did not demonstrate significant differences in pre-referral
and post-referral HbA1C results after they were ordered DSMES services (p = .506). The
proportion of education attendance (30%) was lower than the project goal of 62.5%, but
increased from the clinic baseline.
Conclusions: EHR template implementation for referral to DSMES may increase rates of formal
diabetes education and improve glycemic control. Larger sample sizes, longer project periods,

DIABETES EDUCATION

3

alternative methods of communication, and increased follow-up of participants may be required
to produce significant results.
Keywords: diabetes mellitus, HbA1C, glycated hemoglobin, group-based education,
DSMES, type 2 diabetes

DIABETES EDUCATION

4

Improving Glycemic Control in Patients with Type 2 Diabetes through Formal Education
Uncontrolled type 2 diabetes mellitus (T2DM) can cause avoidable consequences such as
cerebrovascular accident (CVA), myocardial infarction (MI), renal failure, visual disturbance,
amputation of extremities, and death. Over 30 million people in the U.S. have diabetes, and
approximately 90% of them are affected by T2DM, which can be prevented, delayed, and treated
with healthy lifestyle modifications, such as healthy eating, weight loss, and exercise (CDC,
2017). Adults with diabetes at a local internal medicine clinic were having HbA1C levels that
were higher than the recommended level. After synthesizing the evidence of potential effects of
group-based education (GBE) on HbA1C levels and knowledge of disease (KOD), a project was
initiated to implement a staff-training program and an EHR template, to connect patients with
diabetes in the community to a local formal diabetic education program.
Background & Significance
Healthy People 2020 sets national goals to improve the outcomes for patients with
diabetes. One of the high-priority objectives is to decrease the number of patients with diabetes
who have HbA1C levels greater than 9% (Office of Disease Prevention and Health Promotion
[ODPHP], 2014). The organization has also set the goal for the number of patients receiving
formal diabetes education at 62.5%; this would be approximately 10% improvement from
current trends. This project aligns with national goals to improve glycemic control by decreasing
HbA1C levels in patients with T2DM at a local internal medicine clinic and strives to achieve a
62.5% rate of patients receiving formal diabetes education.
Quality improvement (QI) efforts by national organizations have spurred the creation of
measurement tools in the EHR. These tools allow providers to measure their patients’ progress in
major national health initiatives. One component of the system at a local internal medicine clinic

DIABETES EDUCATION

5

displays patients with a diagnosis code for diabetes, between 18 and 75 years of age, who have
an HbA1C level greater than 9, indicating poor disease control. Two hundred and forty patients
met the criteria in that age range at Lifeline Internal Medicine. According to this quality
indicator, approximately 60% of the 240 patients had poorly controlled diabetes. This indicates
over one half of their patient population in that age range had blood glucose levels that were
consistently at dangerous levels, putting them at risk for serious complications.
Soft data gathered from the staff at this internal medicine clinic echoed the statistic
generated from their EHR. They expressed concern for a large portion of their patient
population, which was not attaining adequate control of their diabetic disease processes. In
addition to high HbA1C levels, the staff has witnessed several other trends in their patients with
diabetes, which may be contributing to the lack of glycemic control. The patients presented to
follow-up appointments without logs of their daily blood glucoses that the provider had
requested, without knowledge pertaining to their disease processes and treatment plan, and did
not adhere to lifestyle modification recommendations. Furthermore, providers at the clinic were
not referring patients for any type of formal diabetes education. This data suggested that an
intervention to improve care of patients with T2DM would be beneficial for the clinic and led to
the initiation of this project.
The goals of this project were to lower HbA1C levels and increase rates of diabetes
education attendance to decrease the risk of diabetic complications, improve patients’ health
statuses, and reduce their future healthcare costs. This led to the PICO question, “In adults with
type 2 diabetes at Lifeline Internal Medicine, will formal group education, versus usual care,
improve glycemic control?”
Evidence Synthesis

DIABETES EDUCATION

6

An exhaustive search was conducted in the Cumulative Index of Nursing and Allied
Health Literature (CINAHL), PubMed, and JSTOR databases. Initial keywords in the search
strategy were diabetic, diabetes, compliance, compliant, adherence, adherent, and adult with
Boolean connectors. The keywords compliance, compliant, adherence, adult, and adherent were
subsequently removed after the most promising, potential intervention became more apparent.
The terms education, group, type 2, glycated hemoglobin, and knowledge were added as
keywords. Additional filters included studies in the past five years, full text availability, and
English language. After the search process was complete, pertinent studies were synthesized to
analyze current evidence related to the project goals.
Summaries of essential characteristics of the synthesized studies can be found in the
Evaluation Table (Appendix A). Many valid and reliable tools of measure were used for data
collection, and bias could be considered minimal in most and moderate in a few of the studies.
When referencing the Synthesis Table (Appendix B), the majority of the collected evidence can
be recognized as Level II or III evidence, including primarily randomized controlled trials
(RCTs). One systematic review was included. Only recent studies were included, with the
oldest study published in 2013. The number of participants ranged from 82 to 77,824.
A diverse group of authoring countries was included in appraisal, evaluation, and
synthesis of recent evidence. The synthesized components of the selected studies were focused
areas examining interventions to improve glycemic control, measured by HbA1C, and increase
KOD in patients with T2DM. Two of the studies also examined the effect of GBE on medication
adherence. In all of the studies, the attrition rates were less than 50%, and 8 of the studies had
attrition rates less than 30%. Most studies included an intervention consisting of GBE, and few
of the studies included complementary interventions, such as telephone calls, individual based

DIABETES EDUCATION

7

education (IBE) and home visits. To address the population of the studies, only one study
included patients with type 1 diabetes mellitus. All other participants were patients with T2DM.
As the resources spent on diabetic complications increase, researchers are working to
identify low-cost interventions to improve glycemic control and health maintenance in patients
with diabetes. Scripted phone calls to patients in a large randomized trial did not significantly
improve diabetic control compared to usual care (O’Connor et al., 2014). However, group
education has been an effective intervention, improving glycemic control more than home visits
(Santos et al., 2017). A research study in Brazil was conducted to evaluate patients with diabetes
mellitus (DM) “before and after” an educational program. The educational intervention
significantly increased the participants’ KOD, while also improving HbA1C levels and
adherence to their medication regimens (Figueira, Boas, Coelho, Freita, & Pace, 2016).
Patients’ attitudes toward their diabetes were investigated using questionnaires, revealing
that many people with DM may not understand the HbA1C level, and may need simpler
explanations of key diabetic concepts, with less medical jargon. Overall, participants were
trusting of their providers, but may be given insufficient or incorrect education by such
providers. Many patients revealed that they are using diabetic medication to avoid healthy
lifestyle modifications (Elliott, Harris, & Laird, 2016). A cross-sectional study in St Louis,
Missouri reported findings indicating most patients are unintentionally non-adherent to treatment
plans, suggesting educational interventions may be helpful and should be designed for patients
with limited health literacy (Fan, Lyons, Goodman, Blanchard, & Kaphingst, 2016).
Patients newly diagnosed with diabetes are not the only group who could benefit from
educational interventions. If patients are not meeting their glycemic goals, educational
reinforcement should be implemented, and patients with pre-diabetes may benefit from early

DIABETES EDUCATION

8

formal education. However, the best method for such education and the route to making
educational interventions more widely available are not yet clear (Beverly et al., 2013).
The synthesis of evidence reveals improved glycemic control and KOD in patients with
T2DM that received educational interventions, particularly GBE. With evaluation of the
significant effects that the educational interventions had on HbA1C levels and KOD, the data
suggests that patients with T2DM may benefit from GBE. The evidence supports the use of
GBE for patients with T2DM, but roughly half of the patients in Arizona are receiving proper
diabetes self-management education (ODPHP, 2014). A sustainable intervention to improve rates
of diabetic education is essential for the health of our patients and to reduce healthcare costs.
By increasing availability of diabetes education programs to patients in a local internal
medicine clinic, it was proposed that more of the clinic patients would receive formal diabetes
education, decreasing their HbA1C levels. The goal of this project was to increase the
availability of formal diabetes self-management programs to patients and significantly affect
their HbA1C, making it a low-cost, sustainable intervention to improve diabetic outcomes for
this clinic.
Theoretical Framework & Implementation Framework
Theoretical Framework
The Chronic Care Model (Figure 1) was selected as the theoretical framework to support
this project (Wagner, 1998). In the Chronic Care Model, resources and policies from the
community interact with health care organization to promote productive interactions between an
informed, activated patient and a prepared, proactive practice team. Inspiration from the selfmanagement support aspect of the Chronic Care Model spurred the desire to activate and inform
patients with T2DM at Lifeline Internal Medicine. This project also focused on the clinical

DIABETES EDUCATION

9

information systems aspect of the model, incorporating an additional innovative model, to prepare
the practice team at the clinic for project implementation and create an EHR template that would
streamline the project delivery.
Implementation Framework
Tidd and Bessant’s Process Model of Innovation (Figure 2) was chosen to apply the
synthesis of evidence to current practice (Dawson & Andriopoulos, 2017). This model is concise
and allows for customization. Four steps of the model were used to guide the project and
intervention, displaying the rationale for the employees at the office and serving as a tool for
potential barriers along the way. The first step of the model is search and assessment. In this
step, the area needing improvement is identified. At Lifeline Internal Medicine, the opportunity
was identified as diabetes self-management. The second step explores what the intervention will
be and why it is being done. In this case, we wanted to improve glycemic control through GBE.
The rationale for the intervention includes the soft data, including requests from patients and
staff and is based on an ethical foundation. The third step includes implementation. Specifically,
the staff received education regarding diabetes self-management programs, the new EHR
template for ordering patient education, and network education options to improve the care of
patients with diabetes. Participants’ HbA1C levels were measured pre- and post- GBE. The final
step of the framework captures the value in the innovation.
Methods
Human Subject Protection
Internal Review Board (IRB) approval was granted through Arizona State University
(ASU) prior to the initiation of this project. All clinic staff members had access to the data of the
project, through the password-protected EHR. The team leader extracted lab values and relevant

DIABETES EDUCATION

10

information for the project through the EHR. All paper information pertaining to the project was
stored in a lockable bag. The project leader maintained a master list to link the name of the
participant with a unique numerical study ID (participant 1, 2, 3, etc.,) and stored the master list
in a lockable bag. All project information was kept confidential. Only de-identified data was
retained after data collection. The master list connecting data with the participants’ identities was
destroyed at the end of data collection.
As consent for this project was in the pre-existing new patient paperwork for the
participants, the paperwork was stored electronically in the patients’ charts per clinic standard.
All patients received and signed this paperwork prior to being treated at the clinic. No data was
collected from the staff before or after staff training. There was no compensation or credit
offered to staff or patients to participate in the project.
Testing a patient’s HbA1C requires a blood draw. However, in this project, blood draws
were ordered per usual care, and the patient did not require any additional lab work to what
would have been ordered if the patient were not a participant in the project. The results that were
delivered to the provider were accessed through the EHR. No psychological or legal risks were
foreseeable for patients who participated in the events and data collection involved with this
project.
By participating in the education, the participants learned about their disease and how to
manage it. This could improve their glycemic control, improve their quality of life, and decrease
their risk of mortality. A large community-based population study included over 11,000
participants and found that in the participants with diabetes, an increase of 1% in HbA1C
concentration was associated with 40% increase in mortality from cardiovascular issues and 30%

DIABETES EDUCATION

11

increase in all-cause mortality (Sherwani, Khan, Ekhzaimy, Masood, & Sakharkar, 2016). A
reduction in the HbA1C concentration of 0.2% could decrease the patient’s mortality by 10%.
Population & Setting
This project took place at a local internal medicine clinic. The final study sample only
included patients with T2DM. Although the incidence of diabetes diagnoses in children is
increasing, only 0.18% of the U.S. population under 18 years old is diagnosed with diabetes
(CDC, 2017). This statistic sheds light on the population of interest, adults 18 years of age and
older, diagnosed with T2DM. Participants were required to speak English. Minors were
excluded from the study, as this clinic does not treat pediatric patients. No other specific
populations were targeted or excluded.
Since HbA1C levels can be affected by alterations in hemoglobin and other blood-related
alterations, HbA1C data from patients with a diagnosis code in their chart of anemia, end-stage
renal failure, or recent (last 3 months) blood transfusion were excluded from the analysis.
However, they were not excluded from referral to the education program, and their participation
was included when measuring the percentage of patients who completed the formal diabetes
education program. Participants needed a recent (6 months or less) HbA1C level to be included
in the project.
Project Description & Timeline
Practice and systematic changes were implemented to improve (increase) rates of formal
diabetes education and improve (decrease) HbA1C levels in patients with T2DM at Lifeline
Internal Medicine Clinic. A template was created in the existing EHR program to streamline the
process of ordering group-based diabetes management education for patients who might benefit
from formal diabetes education. Patients were referred to a local, established DSMES program,

DIABETES EDUCATION

12

which is recognized by the American Diabetes Association (ADA), to promote potential
insurance coverage and limit or abolish costs to patients.
Brief staff training was conducted for the staff members of Lifeline Internal Medicine to
inform them of an available community diabetes educational resource and how to order DSMES
for eligible patients using the new EHR fax template. The training was less than twenty minutes
and was scheduled when minimal patient appointments were scheduled. Each member of the
team was addressed individually to make sure he or she did not have any further questions and
all components of the project were clear.
The first outcome that was measured in this project was the HbA1C level of each
participant. Upon entering the bloodstream, glucose binds to hemoglobin in red blood cells. The
HbA1C level represents the percentage of red blood cells, containing hemoglobin that is coated
in glucose. HbA1C level over 6.5% indicates diabetes. The standard goal for the HbA1C level is
less than 7% (CDC, 2018). The World Health Organization (WHO) recognizes the HbA1C level
as diagnostic for diabetes (Florkowski, 2013). National organizations, such as those overseeing
government insurances and diabetic guidelines, have focused on the HbA1C level as a reliable
measure of diabetic control. The HbA1C level was measured using a blood sample. The patient
did not need to fast for the test, and it is routinely checked every 3 months in most patients with
T2DM at Lifeline Internal Medicine. HbA1C levels of the patients referred by the clinic were
measured prior to the referral for education, as this criterion is pertinent for the provider’s
consideration to refer the patient for education and for insurance coverage processes. Preintervention HbA1C levels were already on file, as they are used for the initial diabetes diagnosis
and are monitored at regular quarterly intervals.

DIABETES EDUCATION

13

Once glucose binds to hemoglobin in the blood stream, it is attached permanently until
the red blood cell dies. The average lifespan of a red blood cell is approximately 3-4 months, so
this was the ideal time for their second blood draw. Since exact dates for lab draws and program
activities were not the same for every patient, HbA1C levels collected in a timeframe of 2-4
months post-referral to the education program were allowed for inclusion in the data analysis.
HbA1C levels as early as 2 months post-referral date were included, since HbA1C is based on
weighted monthly averages. 50% of the HbA1C concentration is formed from glycaemia in the
most recent month (Florkowski, 2013). 25% is formed in the month prior to that, and the final
25% is formed in the month before that. There is not a phlebotomist at the clinic, so per usual
care, the patients were given an order from the provider for the tests, and they had samples
drawn at their regular laboratory. Variability in lab sites could have affected HbA1C results.
Variability in blood draw locations was considered an accepted risk for the project to be
sustainable, and so unnecessary inconvenience was not created for the patient.
Additionally, the project evaluated whether training the staff on the community resource
for diabetes education increased the number of patients receiving formal diabetes education. The
goal of this project will align with Healthy People 2020’s goal of 62.5% participation in formal
diabetes education (ODPHP, 2014). Patients were referred to the nearest diabetes selfmanagement education program at Mercy Gilbert Medical Center. Materials and education
distributed at the program were independent of this project. If patients attended one or more
hours of education, they were counted in the group that attended the education program. Followup consisted of usual care and two follow-up phone calls to remind patients to have their lab
work completed.
Data Collection & Analysis

DIABETES EDUCATION

14

After the staff training, data collection began to track the patients who were referred to
the community education program and the patients who attended the program. This was done by
searching the EHR for electronic fax referrals sent from the start date of recruitment through
October 31, 2019. All information was gathered through the EHR. Contact with the diabetes
education program to determine whether or not the patient completed the educational program
closed the loop. The ratio of participants who attended the program to number of participants
who were referred was essential to the data analysis and conclusion portions of the project. The
project attendance percentage was compared to the national goal percentage of 62.5%.
Budget
The proposed budget total for this project was $301.85 (Appendix E). Sources of funding
for the project included personal funds and clinic funds. Clinic funds accounted for the
productivity lost, and paper copies and equipment were purchased using personal funds. Other
non-calculated costs included the cost of the education to patients, if their insurance did not
cover the entire program. Efforts by the provider and community program were made to make
sure the patient receives coverage for the services, if they were eligible. Since the local DSMES
program provided pamphlets, the expense for printing pamphlets was avoided. Expenses for a
locking receptacle for project materials were decreased by purchasing a lock for a container the
project leader already had available for use.
Project Impact
On a local level, the patients, physicians, and staff at Lifeline Internal Medicine had
stakes in the success of this project. If the intervention was successful in improving glycemic
control, it could improve outcomes, decrease costs and complications, free up providers’ time,
and increase productivity in the workplace and the community. In the future, the results of this

DIABETES EDUCATION

15

project may be applicable when discussing potential interventions for patients who present with
pre-diabetes, as well. In 2015, 33.9% of adults, and nearly half of the population above 65 years
of age, had pre-diabetes (CDC, 2018).
In the state of Arizona, in 2010, only 51.4% of patients with diabetes received education
regarding the disease process (ODPHP, 2014). By educating the staff on community resources
available to their patients and increasing availability of diabetes education programs to patients
in a local internal medicine clinic, it was proposed that more of the clinic patients would receive
formal diabetes education, decreasing their HbA1C levels. If the training program increased the
availability of formal diabetes self-management programs to patients or significantly affected
their HbA1C, it would be a low-cost, sustainable intervention to improve diabetic outcomes for
this clinic. This could lead to initiation of cost-effective strategies to decrease complications and
healthcare costs from diabetes.
T2DM complications exhaust resources and contribute to significant healthcare costs
globally. The cost of care for DM in the U.S. is $327 billion per year (ADA, 2018). The majority
of the money is used for inpatient hospital care and prescriptions to treat complications of the
disease. The majority of T2DM care in the U.S. is paid for by government insurances, such as
Medicare, Medicaid, and the military (ADA, 2018).
In 2014, more money was spent on DM discharges than any other emergency department
visit or hospital stay in Arizona (Arizona Department of Health Services [AZDHS], 2018). Over
$8 million dollars was charged for DM for those hospital visits, more than six times the amount
charged for CVA. Results of this project will contribute to the knowledge that could guide
interventions to decrease complications of diabetes, increase rates of diabetes education, and
decrease costs of care in patients with T2DM.

DIABETES EDUCATION

16
Results

Descriptive statistics were used to describe the sample and outcome variables. The
physicians at an internal medicine clinic in Chandler, Arizona ordered DSMES services through
the created EHR template for 10 patients during the recruitment period. The sample consisted of
10 participants (N=10). Twenty percent of participants who were referred for the DSMES
program completed the 10-hour educational program. Ten percent of participants completed at
least 1 hour of the educational program. Thirty percent of participants who were referred
declined the program, 30% did not respond to two or more attempts at contact, and 10% of
participants no-showed to the scheduled program.
The average age of the subjects was 61.10 (SD = 8.96). Ages ranged from 45 to 74 years.
Each gender had an observed frequency of 5 (50%). Frequencies and percentages for insurance,
race, and gender are presented in Table 1.
Table 1
Frequency Table for Gender, Race, and Insurance
Variable
Gender
1
2
Missing
Race
1
2
3
5
Missing
Insurance
1
2
3
Missing

n

%

5
5
0

50
50
0

3
1
3
3
0

30
10
30
30
0

3
1
6
0

30
10
60
0

DIABETES EDUCATION

17

The observations for pre-referral HbA1C in the group that did not attend the DSMES
program had an average of 9.56% (SD = 2.24). The observations for post-referral HbA1C in the
group that did not attend the DSMES program had an average of 7.83% (SD = 1.00). The
summary statistics can be found in Table 2.
Table 2
Summary Statistics Table for Pre- and Post-Referral HbA1C Variables in Participants who Did
Not Attend Education
Variable
M
SD n
Min
Max
SEM
Post_HbA1C
7.83
1.00 3
0.58
6.70
8.60
Pre_HbA1C
9.56
2.24 7
0.85
7.00
12.90
Note. '-' denotes the sample size is too small to calculate statistic.

Skewness
-0.58
0.38

Kurtosis
-1.50
-1.15

The observations for pre-referral HbA1C in the group that attended the DSMES program
had an average of 8.63 (SD = 2.42). There were insufficient observations to calculate summary
statistics for post-referral HbA1C. The summary statistics can be found in Table 3.
Table 3
Summary Statistics Table for Pre- and Post-Referral HbA1C Variables in Participants who
Attended Education
Variable
M
SD n
SEM
Min
Max
Post_HbA1C
7.70
- 1
7.70
7.70
Pre_HbA1C
8.63
2.42 3
1.40
6.90
11.40
Note. '-' denotes the sample size is too small to calculate statistic.

Skewness
0.64

Kurtosis
-1.50

A two-tailed paired samples t-test was conducted to examine whether the mean difference
of pre-referral HbA1C and post-referral HbA1C results from the group that did not attend the
education program were significantly different from zero. The result of the two-tailed paired
samples t-test was not significant based on an alpha value of 0.05, t(6) = 2.22, p = .068,
indicating the null hypothesis cannot be rejected. The results are presented in Table 4.

DIABETES EDUCATION

18

Table 4
Two-Tailed Paired Samples t-Test for the Difference between Pre-Referral HbA1C and PostReferral HbA1C in the Group that Did Not Attend Education
Pre_A1C
Post_A1C_imputed
M
SD
M
SD
t
p
9.56
2.24
7.88
0.90
2.22
.068
Note. N = 7. Degrees of Freedom for the t-statistic = 6. d represents Cohen's d.

d
0.84

A two-tailed paired samples t-test was conducted to examine whether the mean difference
of pre-referral HbA1C and post-referral HbA1C results was significantly different from zero in
the group that attended the education program. The result of the two-tailed paired samples t-test
was not significant based on an alpha value of 0.05, t(2) = 0.90, p = .465, indicating the null
hypothesis cannot be rejected. The results are presented in Table 5.
Table 5
Two-Tailed Paired Samples t-Test for the Difference Between Pre-Referral HbA1C and PostReferral HbA1C in Group that Attended Education Program
Pre_A1C
Post_A1C_imputed
M
SD
M
SD
t
p
8.63
2.42
7.49
0.34
0.90
.465
Note. N = 3. Degrees of Freedom for the t-statistic = 2. d represents Cohen's d.

d
0.52

Effect of Group Education on HbA1C
A mixed model analysis of variance (ANOVA) with one within-subjects factor and one
between-subjects factor was conducted to determine whether significant differences exist among
pre-referral HbA1C and post-referral HbA1C between the group that attended the educational
program and the group that did not attend the educational program, using imputed values for
post-referral HbA1C missing values. The results were examined based on an alpha of 0.05. The
main effect for education was not significant F(1, 8) = 0.48, p = .506, indicating the two
education groups were similar. The main effect for the within-subjects factor was not
significant F(1, 8) = 3.96, p = .082, indicating the values of pre-referral HbA1C and post-referral

DIABETES EDUCATION

19

HbA1C were all similar. The interaction effect between the within-subjects factor and education
was not significant F(1, 8) = 0.14, p = .719, indicating that for all combinations of the withinsubjects factor and the education groups, the strength of the relationship between the outcome
and the interaction of education does not change significantly.
Comparison of Program Attendance to Goal
A one-proportion z-test was conducted to examine whether education attendance could
have been produced by a probability distribution with a proportion of 0.625. The result of the one
proportion z-test was significant based on an alpha value of 0.05, z = -2.12, p = .034, CI = [-0.63,
-0.02], indicating the null hypothesis can be rejected. This suggests that education attendance is
unlikely to have been produced by a distribution with a proportion of 0.625. The proportion of
education attendance is most likely lower than 0.625. The confidence interval (α = 0.05) for
the proportions of education attendance is -0.63 to -0.02. Table 6 presents the results of the one
sample proportion z-test.
Table 6
One Proportion z-Test for Education Attendance and a Test Proportion of 0.625
Samples
Responses
n
Proportion
Education Attendance

3

10

0.3

SD

SE

0.46

0.15

Note. z = -2.12, p = .034, CI for α = 0.05: [-0.63, -0.02].
Discussion
Although the results were not considered significant after data analysis, when considering
limitations, this project reinforces the findings of other studies that GBE may improve
knowledge of disease and HbA1C levels in patients with T2DM. As other studies pointed out,
the best method for such education and the route to making educational interventions more
widely available are not yet clear (Beverly et al., 2013).

DIABETES EDUCATION

20

Limitations
The sample size was a limitation of the project, which was evident in the data analysis
portion of the project. Since there were several missing values for post-referral HbA1C
observations, missing values in both groups were imputed, since there were not at least three
values for the group that attended the education program.
This project did not meet the goal education attendance rate of 62.5%. The proportion of
attendees was found to be significantly different from the goal of 62.5%. Communication with
patients was a barrier in the link to the education program for the clinic and the DSMES
program. Forty-three percent of participants who did not attend the education program were left
voicemails multiple times, but never responded, eventually resulting in a “send-back” to the
provider. Forty-three percent answered, but declined the education program when contacted
about attending. The final 7% of the group that did not attend the education program did not
show up to the initial scheduled education session. In the future, an alternative method of
contacting patients may be beneficial in increasing the attendance rate.
There was a significant decrease in one participant’s HbA1C level in the group that did
not attend the education program (33% decrease from pre-HbA1C level). It was noted that this
participant initiated subcutaneous injections in their diabetes treatment plan at the time of referral
to the diabetes education program, which could have had an effect on post-referral HbA1C levels
in the group that did not attend the DSMES program.
Project duration could have limited the results of the study. A few participants were due
for follow-up at the time of data collection, so they could have completed their lab work shortly
after data collection took place. Project implementation took place during winter months, with
data collection taking place just after the beginning of the new year. Since many people indulge

DIABETES EDUCATION

21

in carbohydrate-rich foods during this time and may defer follow-ups due to holiday obligations,
a longer project period may be beneficial.
Strengths
This project used a technological modality to order diabetes education for participants.
The National Diabetes Education Program, which was established in a partnership between the
CDC and U.S. Department of Health and Human Services (HHS) National Institutes of Health
(NIH) encourages self-management of diabetes that is sustainable and promotes technological
advances to link patients, providers, and communities to strategies and support that encourage
sustainable self-management (HHS, NIH, National Institute of Diabetes and Digestive and
Kidney Diseases [NIDDK], 2014).
Although no significant change was found in the pre- and post-referral means of HbA1C
in patients that attended the group education program verses those that did not attend the
education program, the participant who attended the program and completed lab work posteducation did demonstrate a decrease in the HbA1C level of 3.7%, which was a 33% decrease
from the pre-referral HbA1C level.
Although the project did not meet the national attendance goal, it did increase the number
of patients at the clinic who attended a formal diabetes education program. Since there was no
referral process for diabetes education in place for the project site, this project created a link
between the provider, the patients, and the DSMES program, which resulted in 30% of
participants attending at least 1 hour of the DSMES program they were referred to. Twenty
percent of participants attended all 10 hours of the DSMES program.
Sustainability

DIABETES EDUCATION

22

The EHR template was implemented for ordering DSMES to simplify the process for
providers and staff, and to promote sustainability. Providers will be able to use the template
after program completion to create the link between patients and the DSMES program in the
community. The template could also be used for additional DSMES sites in the future.
Implications for Practice
Medicare and Medicaid, along with many other insurance plans, will cover up to 10 hours
of DSMES the first year (Warshaw, 2018). However, even if a person does not participate in all
10 hours, they may only receive coverage for up to two hours of education in the subsequent
years. Healthcare providers must take advantage of these 10 hours of available education for
their patients in their first year after diagnosis to help them learn how to manage their diabetes by
linking patients to these services and encouraging them to follow through with them.
Insurance companies may require certain information to cover the services, so providers
must make sure they are meeting all criteria when ordering educational services for patients with
diabetes. Templates in the EHR can be used to remind providers of steps they should take with
patients who have T2DM to address all significant areas of concern. Templates can also be used
to make sure the pertinent information is included in orders and referrals to promote
reimbursement for provider services, including other services or support they may order for the
patient. To meet the demands of evolving reimbursement strategies and the increasing
prevalence of patients with T2DM, providers will need to be more proactive, while utilizing the
EHR’s specialized features, such as templates to reduce documentation load, to encourage
ongoing support for patients with T2DM. Providers will need to make more collaborative efforts
to link patients to community resources to manage and support various aspects of those patients’
care.

DIABETES EDUCATION

23

The results of this project suggest that alternative modalities for communicating with
patients may be beneficial, as communication via telephone was a barrier to contacting patients
for the clinic and the DSMES program. Subsequent studies could use other means of
communication, such as text messages, e-mails, or cellular phone applications to coordinate care
with patients. Finding a reliable means of communication will be essential. Since 70% of
participants did not complete the recommended post-referral HbA1C, they may not be receiving
optimal care for their T2DM, and successful alternatives for reminding these patients to follow
up would benefit patients and providers.
Conclusion
Although comparisons of pre-referral and post-referral means of participants who
attended DSMES and participants who did not attend DSMES were not significant, using
imputed data for missing values, there was a decrease in post-referral HbA1C in the participant
who attended DSMES and completed recommended post-referral HbA1C lab work. Patients
may not be following up with lab work or providers for optimal treatment of their T2DM.
Implementing an EHR template and training staff to use the template can streamline the ordering
process for DSMES and increase the attendance of formal diabetes education in patients, but
alternative methods to telephone calls need to be considered for patients with T2DM.

DIABETES EDUCATION

24
References

American Diabetes Association. (2018). The cost of diabetes. Retrieved from
http://www.diabetes.org/advocacy/news-events/cost-of-diabetes.html
Arizona Department of Health Services. (2018). Diabetes in Arizona: The 2018 burden report.
Retrieved from https://www.azdhs.gov/documents/prevention/tobacco-chronicdisease/diabetes/reports-data/diabetes-burden-report-2018.pdf
Beverly, E. A., Fitzgerald, S. M., Brooks, K. M., Hultgren, B. A., Ganda, O. P., Munshi, M., &
Weinger, K. (2013). Impact of reinforcement of diabetes self-care on poorly controlled
diabetes. The Diabetes Educator, 39(4), 504–514. doi:10.1177/0145721713486837
Centers for Disease Control and Prevention. (2017). About diabetes. Retrieved from
https://www.cdc.gov/diabetes/basics/diabetes.html
Centers for Disease Control and Prevention. (2018). All about your A1C. Retrieved from
https://www.cdc.gov/diabetes/library/features/a1c.html
Dawson, P., & Andriopoulos, C. (2017). Managing change, creativity & innovation. London:
Sage.
Elliott, A. J., Harris, F., & Laird, S. G. (2016). Patients’ beliefs on the impediments to good
diabetes control: A mixed methods study of patients in general practice. British Journal
of General Practice, 66(653), e913–e919. doi:10.3399/bjgp16x687589
Fan, J. H., Lyons, S. A., Goodman, M. S., Blanchard, M. S., & Kaphingst, K. A. (2016).
Relationship between health literacy and unintentional and intentional medication
nonadherence in medically underserved patients with type 2 diabetes. The Diabetes
Educator, 42(2), 199–208. doi:10.1177/0145721715624969

DIABETES EDUCATION

25

Figueira, A. L. G., Boas, L. C. G. V., Coelho, A. C. M., Freitas, M. C. F., & Pace, A. E. (2017).
Educational interventions for knowledge on the disease, treatment adherence and control
of diabetes mellitus. Revista Latino-Americana de Enfermagem, 25.
https://doi.org/10.1590/1518-8345.1648.2863
Florkowski, C. (2013). HbA1c as a diagnostic test for diabetes mellitus: Reviewing the
Evidence. The Clinical Biochemist, 34(2), 75–83.
Improving Chronic Illness Care. (n.d.) The chronic care model. Retrieved from
http://improvingchroniccare.org/index.php?p=Chronic+Care+Model&s=124
O’Connor, P. J., Schmittdiel, J. A., Pathak, R. D., Harris, R. I., Newton, K. M., Ohnsorg, K. A.,
… Steiner, J. F. (2014). Randomized trial of telephone outreach to improve medication
adherence and metabolic control in adults with diabetes. Diabetes Care, 37(12), 3317–
3324. doi:10.2337/dc14-0596
Office of Disease Prevention and Health Promotion. (2014). Diabetes. Healthy People 2020.
Retrieved from https://www.healthypeople.gov/2020/data-search/Search-the-Data#topicarea=3514;
Santos, J. C., Cortez, D. N., Macedo, M. M. L., Reis, E. A., Reis, I. A., & Torres, H. C. (2017).
Comparison of education group strategies and home visits in type 2 diabetes mellitus:
Clinical trial. Revista Latino-Americana de Enfermagem, 25. doi:10.1590/15188345.2315.2979
Sherwani, S. I., Khan, H. A., Ekhzaimy, A., Masood, A., & Sakharkar, M. K. (2016).
Significance of HbA1c test in diagnosis and prognosis of diabetic patients. Biomarker
insights, 11, 95–104. doi:10.4137/BMI.S38440

DIABETES EDUCATION

26

U.S. Department of Health and Human Services, National Institutes of Health, National Institute
of Diabetes and Digestive and Kidney Diseases. (2014). NDEP diabetes strategic plan for
2014-2019. Retrieved from https://www.niddk.nih.gov/healthinformation/communication-programs/ndep/about-national-diabetes-educationprogram/ndep-strategic-plan-2014-2019
Wagner, E., H. (1998). Chronic disease management: What will it take to improve care for
chronic illness? Effective Clinical Practice (1), 2-4.
Warshaw, H. (2018). Is it time to remodel diabetes-self management education and support?
Evidence Based Diabetes Management. Retrieved from
https://www.ajmc.com/journals/evidence-based-diabetes-management/2018/september2018/is-it-time-to-remodel-diabetes-selfmanagement-education-and-support

DIABETES EDUCATION

27
Appendix A
Evaluation Table

Citation

Conceptual
Framework

Design/Metho
d

Sample/Setting

Odgers-Jewell,
K. et al. (2017).
Effectiveness of
group-based
selfmanagement
education for
individuals with
type 2 diabetes:
A systematic
review with
meta-analysis
and metaregression

“Promoting
and
supporting
positive
selfmanagement
behaviors”,
self-efficacy
inferred

Design:
Systematic
review of
RCTs, cluster
randomized
trials, and
controlled
clinical trials

N=8,533
n=4416 (IG)
n=4117 (CG)

Funding: None
Bias: Most
studies
classified as
having mod risk
of bias;
performance

Purpose: To
determine the
efficacy of
GBE vs. IBE or
UC for
improving
outcomes in
patients with
T2DM

Setting: 32
studies (68%)
in primary care
settings, 15
studies (32%)
in secondary or
tertiary settings
Inclusion: ≥ 18
y/o, face-toface GBE,
T2DM, ≥ 4
participants, >
1 hour
intervention
Exclusion:
pregnant

Major
Variables &
Definitions
IV: GBE
DV1: HbA1C
DV2: DM
knowledge
Other DVs
included FBG,
body weight,
WC, BP, lipid
levels, selfefficacy

Measuremen
t

Analysis

Findings

Decision for
Use

Various
questionnaires
(validated)specifics not
included,
biometric
measurements

RevMan, Excel,
meta-analysis
using
DerSimonian and
Laird, I-squared,
meta-regression
using stat
statistical
software

GBE is
more
effective
than UC
and IBE.

LOE: LOE I

95% CI,
Significant
difference
in reducing
HbA1C
(p=0.0002),
significant
difference
in
increasing
diabetic
knowledge
(p=0.01)

Strengths:
AMSTAR
quality
assessment:
high quality
review (10/11),
statistically
significant
results, only
included
patients with
T2DM
Weaknesses:
Most studies
had mod risk
of bias. 13/47
studies had
possible
conflict of
interest. High

KEY: BMI-body mass index; BP-blood pressure; CCT-clinically controlled trial; CCM-Chronic Care Model; CG-control group; CIconfidence interval; CM-Conversation Maps; DKN-A-Diabetes Knowledge Scale; DKQ-Diabetes Knowledge Questionnaire; DVdependent variable; DM-diabetes mellitus; EMR-electronic medical record; FBG-fasting blood glucose; GBE-group-based
education; HbA1C-glycated hemoglobin; HDL-high-density lipoprotein; HTN-hypertension; HV-home visits; IBE-individual-based
education; IG-intervention group; IV-independent variable; LDL-low-density lipoprotein; LOE-level of evidence; MA-mean age;
MAT-Measure of Adherence to Treatments; MP-medication persistence; MPR-medication possession ratio; PMA-primary
medication adherence; RCT-randomized clinical trial; SCT-Social Cognitive Theory; SPSS-Statistical Package for Social
Sciences;T2-type 2; UC-usual care; WC-waist circumference

DIABETES EDUCATION

28

bias, detection
bias

women, IBE,
type 1 DM

heterogeneity
of some metaanalysis

Country: 38%
of studies from
U.S., 13% from
UK, 11% Italy

Citation

Conceptual
Framework

Design/Metho
d

Sample/Setting

Merakou, K. et
al. (2015).
Group patient
education:
Effectiveness of
a brief
intervention in
people with
T2DM in
primary health
care in Greece:
A CCT

Inferred
CCM with
an emphasis
on selfmanagement

Design: CCT

N=193
n=55 (CG)
n=138 (IG)

Funding:
unknown

Purpose: To
determine the
efficacy of
GBE using
CMs sessions
for people with
T2DM vs. IBE
by primary
healthcare
provider

Setting: Health
Centre of the
Primary
Healthcare
Clinic in
Markopoulo,
Greece
Demographics
:

Major
Variables &
Definitions
IV: GBE using
CM
DV1: HbA1C
DV2: BMI
DV3: lipid
levels
Conversation
Maps:
Learning
About
Diabetes
consists of
educational
tools that

Measuremen
t

Analysis

Findings

Biochemical
markers:
HbA1C, LDL,
HDL,
Triglycerides,
anthropometr
y

SPSS, chi-square
and Fisher’s
exact, Student’s ttest, paired t-tests,
analysis of
variance, analysis
of covariance

Short
education
intervention
s using
CMs may
be more
effective
than IBE in
controlling
HbA1C.
95% CI,
Significant
difference
in reducing

Feasibility:
Recommended
for use in
practice,
supporting
GBE use to
improve
HbA1C and
DM
knowledge.
Decision for
Use
LOE: LOE II
Strengths:
measurement
methods,
T2DM only
Limitations:
small town,
familiarity
with research
team, may not
be
representative
of general

KEY: BMI-body mass index; BP-blood pressure; CCT-clinically controlled trial; CCM-Chronic Care Model; CG-control group; CIconfidence interval; CM-Conversation Maps; DKN-A-Diabetes Knowledge Scale; DKQ-Diabetes Knowledge Questionnaire; DVdependent variable; DM-diabetes mellitus; EMR-electronic medical record; FBG-fasting blood glucose; GBE-group-based
education; HbA1C-glycated hemoglobin; HDL-high-density lipoprotein; HTN-hypertension; HV-home visits; IBE-individual-based
education; IG-intervention group; IV-independent variable; LDL-low-density lipoprotein; LOE-level of evidence; MA-mean age;
MAT-Measure of Adherence to Treatments; MP-medication persistence; MPR-medication possession ratio; PMA-primary
medication adherence; RCT-randomized clinical trial; SCT-Social Cognitive Theory; SPSS-Statistical Package for Social
Sciences;T2-type 2; UC-usual care; WC-waist circumference

DIABETES EDUCATION

29
MA of
CG=63.8,
58.2% men,
41.8% women;
MA of
IG=67.2,
53.6% men,
46.4% women

Bias: selection,
allocation of
participants not
concealed
Country:
Greece

encourage
interactivity in
small GBE.

HbA1C
(p=0.003)

Feasibility:
Could be
implemented
for many
people at lowcost with
limited
personnel,
more
applicable in
primary care
setting with
established
patientprovider
relationships

Inclusion:
T2DM, regular
patients in the
setting, ≥ 18
y/o,

Citation

Conceptual
Framework

Design/Metho
d

Reaney, M. et
al. (2013).
Impact of CM
education tools
versus regular

Inferred
CCM with
an emphasis
on selfmanagement

Design: RCT
Purpose: To
determine
whether CM-

Exclusion:
HTN, other
serious heart
disease, stroke,
kidney disease,
mental
disorder,
insulin-use,
complications
Sample/Setting

N=681
n=330 (IG)
n=351 (CG)

population,
exclusion
criteria,
excluded
insulin-users

Major
Variable &
Definitions
IV: CM-based
education
DV1: DM
knowledge
DV2: HbA1C

Measuremen
t

Analysis

Findings

Decision for
Use

ADKnowl
questionnaire
(validated),
biometric
testing

ANCOVA,
Wilcoxon rank
sum, mixedmodel repeated
measures,

DM
knowledge
and HbA1C
improved in

LOE: LOE II
Strengths:
High retention
rate, vague

KEY: BMI-body mass index; BP-blood pressure; CCT-clinically controlled trial; CCM-Chronic Care Model; CG-control group; CIconfidence interval; CM-Conversation Maps; DKN-A-Diabetes Knowledge Scale; DKQ-Diabetes Knowledge Questionnaire; DVdependent variable; DM-diabetes mellitus; EMR-electronic medical record; FBG-fasting blood glucose; GBE-group-based
education; HbA1C-glycated hemoglobin; HDL-high-density lipoprotein; HTN-hypertension; HV-home visits; IBE-individual-based
education; IG-intervention group; IV-independent variable; LDL-low-density lipoprotein; LOE-level of evidence; MA-mean age;
MAT-Measure of Adherence to Treatments; MP-medication persistence; MPR-medication possession ratio; PMA-primary
medication adherence; RCT-randomized clinical trial; SCT-Social Cognitive Theory; SPSS-Statistical Package for Social
Sciences;T2-type 2; UC-usual care; WC-waist circumference

DIABETES EDUCATION
care on
diabetes-related
knowledge of
people with T2
diabetes: A
randomized,
controlled study

30
based group
education
would increase
DM related
knowledge at 6
months vs.
those who
received UC

Country:
Germany, Spain
Funding: Eli
Lilly &
Company
Bias: openlabel study with
potential
recruitment bias
and potential
questionnaire
result bias

Citation

Setting: 19
sites in
Germany, 14
sites in Spain
Demographics
:
MA of
CG=61.9,
52.4% male,
47.6% female;
MA of
IG=62.0,
54.2% male,
45.8% female

Pearson’s x2,
Fisher’s exact

Other DVs
included
satisfaction
with care,
patient
empowerment,
self-care, lipid
levels, goal
attainment,
emotional
distress, and
vital signs

In Spain,
scores for
DM
knowledge
were higher
for CM than
UC
(p<0.001).
In
Germany,
the opposite
was true
(p<0.001).

Inclusion: 1875 y/o adults,
T2DM, poor
disease
management, in
need of
education

Conceptual
Framework

Design/Metho
d

Sample/Setting

Major
Variable &
Definitions

both
groups.

Measuremen
t

Analysis

Note:
78.3% of
those in
Germany
and 13.5%
of those in
Spain had
non-CM
structured
education
during the
study.
Findings

inclusion
criteria defined
by provider
Limitations:
Open-label
study, studied
participants
only 6 months
after
intervention, 2
countries
studied,
baseline low
HbA1C levels
Feasibility:
CM could be
more effective
in areas where
structured DM
education is
not already a
major part of
UC.

Decision for
Use

KEY: BMI-body mass index; BP-blood pressure; CCT-clinically controlled trial; CCM-Chronic Care Model; CG-control group; CIconfidence interval; CM-Conversation Maps; DKN-A-Diabetes Knowledge Scale; DKQ-Diabetes Knowledge Questionnaire; DVdependent variable; DM-diabetes mellitus; EMR-electronic medical record; FBG-fasting blood glucose; GBE-group-based
education; HbA1C-glycated hemoglobin; HDL-high-density lipoprotein; HTN-hypertension; HV-home visits; IBE-individual-based
education; IG-intervention group; IV-independent variable; LDL-low-density lipoprotein; LOE-level of evidence; MA-mean age;
MAT-Measure of Adherence to Treatments; MP-medication persistence; MPR-medication possession ratio; PMA-primary
medication adherence; RCT-randomized clinical trial; SCT-Social Cognitive Theory; SPSS-Statistical Package for Social
Sciences;T2-type 2; UC-usual care; WC-waist circumference

DIABETES EDUCATION
Santos, J. C. et
al. (2017).
Comparison of
education group
strategies and
home visits in
T2DM: Clinical
trial
Country:
Brazil
Funding:
Federal
University of
Minas Gerais
Bias: cluster
stratification
used to
minimize bias

Inferred
Selfmanagement
focused
CCM

31
Design: RCT
Purpose: To
compare
adherence and
empowerment
of self-care and
glycemic
control in HV
and GBE
strategies

N= 238
n= 34 (IG, HV)
n= 93 (IG,
GBE)
n=111 (CG)
Setting: 10
primary care
sites in Brazil
(clusters)
Demographics
:
MA=57.8 y/o,
77.4% female
Inclusion:
T2DM, 30-80
y/o
Exclusion:
Chronic T2DM
complications,
Discontinuing
criteria: <6
GBE meeting
participation,
<4 home
meetings

IV1: GBE
IV2: HV
DV1: HbA1C
DV2: ESM
DV3: DES-SF
Other
variables
included
lipids, FBG,
BMI

Self-care
questionnaire
(ESM)
Diabetes
Empowermen
t Scale-Short
Form (DESSF), HbA1C

SPSS, ShapiroWilk, Anova,
paired Student’s
t-tests, Wilcoxon,
Mann-Whitney

Similar
results
between
HV and
GBE for
adherence
to self-care
and
empowerme
nt.
Significant
improveme
nts in
HbA1C in
GBE
(p=0.0000)
vs. home
visit
(p=0.9900).

LOE: LOE 1I
Strengths:
Allows
replication of
educational
strategies in
primary care,
demonstrates
importance of
structured
education
strategies
Limitations:
Intellectual
capacity of
participants not
considered,
inhomogeneity
of disease time
among groups,
specific
location of
study (small
area in Brazil)
Feasibility:
Cost-effective
recommendatio
ns for primary

KEY: BMI-body mass index; BP-blood pressure; CCT-clinically controlled trial; CCM-Chronic Care Model; CG-control group; CIconfidence interval; CM-Conversation Maps; DKN-A-Diabetes Knowledge Scale; DKQ-Diabetes Knowledge Questionnaire; DVdependent variable; DM-diabetes mellitus; EMR-electronic medical record; FBG-fasting blood glucose; GBE-group-based
education; HbA1C-glycated hemoglobin; HDL-high-density lipoprotein; HTN-hypertension; HV-home visits; IBE-individual-based
education; IG-intervention group; IV-independent variable; LDL-low-density lipoprotein; LOE-level of evidence; MA-mean age;
MAT-Measure of Adherence to Treatments; MP-medication persistence; MPR-medication possession ratio; PMA-primary
medication adherence; RCT-randomized clinical trial; SCT-Social Cognitive Theory; SPSS-Statistical Package for Social
Sciences;T2-type 2; UC-usual care; WC-waist circumference

DIABETES EDUCATION

32
care health
sites

Citation

Conceptual
Framework

Design/Metho
d

Sample/Setting

Hwee, J. et al.
(2014).
National
diabetes
education
through group
classes leads to
better care and
outcomes than
individual
counseling in
adults: A
populationbased cohort
study.

Inferred
CCM

Design:
Populationbased cohort
study

N=77,824
n=12,234
(IG,GBE)
n=55,761
(IG,IBE)
n=9,829 (IG,
both IBE and
GBE)

Country:
Canada
Bias: Possible
confounding
bias

Purpose: To
determine if
quality of care
or DM
complications
differed
between GBE
and IBE.

Setting:
Ontario
healthcare
administrative
databases used
to extract info
detailing
utilization of
selfmanagement
education
programs by
residents in
Ontario

Major
Variable &
Definitions
IV1: GBE
IV2: IBE
DVs:
hospitalization
s & ED visits
for
hypo/hypergly
cemia, foot
ulcers,
cellulitis;
claims for 2+
HbHb tests;
claim for 1+
lipid test,
claim for 1+
eye exam for
retinal
screening, Rx
for antihypertension
meds, oralglucose
lowering

Measuremen
t

Analysis

Findings

Decision for
Use

Administrativ
e claims
extracted from
databases,
prescriptions

Chi-squared,
ANOVA, Logitbased generalized
estimating
regression
models, SAS
version 9.3

Fewer ED
visits/hospit
alizations
for DM
complicatio
ns
(p<0.001),
higher rates
of adequate
lab testing
and statin
use
(p<0.001)
among
those in the
GBE group

LOE: LOE IV
Strengths:
large sample
size, majority
of population
had T2DM
Limitations:
many people
not included
with unknown
educational
formal
(15,561),
varying
amounts of
education,
contained both
type 1 and
T2DM,
intensity of
education

KEY: BMI-body mass index; BP-blood pressure; CCT-clinically controlled trial; CCM-Chronic Care Model; CG-control group; CIconfidence interval; CM-Conversation Maps; DKN-A-Diabetes Knowledge Scale; DKQ-Diabetes Knowledge Questionnaire; DVdependent variable; DM-diabetes mellitus; EMR-electronic medical record; FBG-fasting blood glucose; GBE-group-based
education; HbA1C-glycated hemoglobin; HDL-high-density lipoprotein; HTN-hypertension; HV-home visits; IBE-individual-based
education; IG-intervention group; IV-independent variable; LDL-low-density lipoprotein; LOE-level of evidence; MA-mean age;
MAT-Measure of Adherence to Treatments; MP-medication persistence; MPR-medication possession ratio; PMA-primary
medication adherence; RCT-randomized clinical trial; SCT-Social Cognitive Theory; SPSS-Statistical Package for Social
Sciences;T2-type 2; UC-usual care; WC-waist circumference

DIABETES EDUCATION

33

Funding: grant
from
Physicians’
Services Inc.
Foundation of
Ontario and by
Ontario
Ministry of
Health and
Long-Term
Care
(MOHLTC)

Demographics
:
MA GBE=58.8
MA IBE=59.2
MA both=58.0

agents, and
insulin

could have
been
underestimated
as this study
only accounted
for 1 year

Exclusions:
died in the
same 1 year of
study (1,682),
<18 y/o,
unknown
educational
format (15,561)

Citation

Conceptual
Framework

Design/Metho
d

Sample/Setting

O’Connor, et al.
(2014).
Randomized
trial of
telephone
outreach to
improve
medication
adherence and
metabolic
control in adults
with diabetes.

none stated
or easily
inferred

Design:
pragmatic RCT

N=2378
n=1,220 (IG)
n=1,158 (CG)

Purpose: to
determine the
efficacy of
phone calls to
DM patients
who were
above
recommended
levels for
lipids, BP, or

Setting: several
large multispecialty
medical groups
Demographics
:
IG MA 61.67,
51.1% female;

Major
Variable &
Definitions
IV: structured
telephone
interview
contact
DV1: PMA
DV2: MP
DV3: MPR
DV4: HbA1C,
BP, or LDL
PMA: Fill of
medication
refill within 60

Measuremen
t

Analysis

Findings

Prescription
fills, HbA1C,
BP, LDL, data
extracted for
EMR

Primary and
secondary
analysis
performs; intentto-treat analysis,
per-protocol
analysis, post-hoc
analysis. Specific
measures not
provided.

The
intervention
failed to
produce any
significant
changes in
PMA, MP,
MPR,
HbA1C,
BP, or
LDL.

Feasibility:
Less resource
intensive and
cost effective
interventions
may be able to
improve
patient care
with GBE.
Decision for
Use
LOE: LOE II
Limitations:
Between 6678% of
participants
had already
filled their
prescription
prior to the
intervention.
Low power
after post-hoc

KEY: BMI-body mass index; BP-blood pressure; CCT-clinically controlled trial; CCM-Chronic Care Model; CG-control group; CIconfidence interval; CM-Conversation Maps; DKN-A-Diabetes Knowledge Scale; DKQ-Diabetes Knowledge Questionnaire; DVdependent variable; DM-diabetes mellitus; EMR-electronic medical record; FBG-fasting blood glucose; GBE-group-based
education; HbA1C-glycated hemoglobin; HDL-high-density lipoprotein; HTN-hypertension; HV-home visits; IBE-individual-based
education; IG-intervention group; IV-independent variable; LDL-low-density lipoprotein; LOE-level of evidence; MA-mean age;
MAT-Measure of Adherence to Treatments; MP-medication persistence; MPR-medication possession ratio; PMA-primary
medication adherence; RCT-randomized clinical trial; SCT-Social Cognitive Theory; SPSS-Statistical Package for Social
Sciences;T2-type 2; UC-usual care; WC-waist circumference

DIABETES EDUCATION
Country: USA

34
glucose and
had been
prescribed a
new med to
treat the
elevation

Bias: none
identified
Funding:
Agency for
Healthcare
Research and
Quality

CG MA 62.04,
52.3% female
Inclusion:
18-75 y/o, 15
months care at
center prior to
enrollment,
new medication
that meets
criteria in last
180 days

Citation

Conceptual
Framework

Design/Metho
d

Sample/Setting

Figueira, A. et
al. (2017).
Educational
interventions
for knowledge
on the disease,
treatment
adherence, and
control of
diabetes
mellitus.).

SCT

Design:
randomized
intervention
study with
single
comparison
group

N=82
n= 82

Purpose: to
analyze the
effect of
educational
interventions

Setting:
outpatient
clinic
Demographics
: 48 women
(58.5%), Mage
60.43

days of
prescription
date

analysis to
detect
significant
changes in
LDL and BP.

MP: 2 or more
fills in 180
days of
medication

Major
Variable &
Definitions
IV: GBE
using CM
DV1: DM
knowledge
DV2:
medication
treatment
adherence
DV3:
glycemic
control

Measuremen
t

Analysis

Findings

DKN-A,
MAT,
electronic
system
HbA1C

R version 3.02;
Excel; paired
Wilcoxon test,
KomolgrowSmirnov and
Leven tests

All
dependent
variables
improved
significantl
y with GBE
using CM
including
DM
knowledge
of disease
(p<0.001),
medication

Feasibility:
Intervention
not likely to
produce
significant
changes, more
resourceintensive than
alternative
options
Decision for
Use
LOE: LOE IV
Strengths:
There is a
general lack of
studies on CM
as an education
tool and SCT
as a conceptual
framework for
diabetes
education.

KEY: BMI-body mass index; BP-blood pressure; CCT-clinically controlled trial; CCM-Chronic Care Model; CG-control group; CIconfidence interval; CM-Conversation Maps; DKN-A-Diabetes Knowledge Scale; DKQ-Diabetes Knowledge Questionnaire; DVdependent variable; DM-diabetes mellitus; EMR-electronic medical record; FBG-fasting blood glucose; GBE-group-based
education; HbA1C-glycated hemoglobin; HDL-high-density lipoprotein; HTN-hypertension; HV-home visits; IBE-individual-based
education; IG-intervention group; IV-independent variable; LDL-low-density lipoprotein; LOE-level of evidence; MA-mean age;
MAT-Measure of Adherence to Treatments; MP-medication persistence; MPR-medication possession ratio; PMA-primary
medication adherence; RCT-randomized clinical trial; SCT-Social Cognitive Theory; SPSS-Statistical Package for Social
Sciences;T2-type 2; UC-usual care; WC-waist circumference

DIABETES EDUCATION

35
on knowledge
of disease,
medication
treatment
adherence, and
glycemic
control in
patients with
DM

Country:
Brazil
Bias: possible
setting selection
Funding:
University of
Sao Paulo

Citation

Conceptual
Framework

Design/Metho
d

Torres, H. C. et
al. (2018).
Evaluation of
the effects of a
diabetes
educational
program: A

Creating
awareness
through
self-care

Design: cluster
RCT
Purpose:
determine
efficacy of an
educational
program of

Inclusion:≥40
y/o, diagnosed
with T2DM
Exclusion:<40
y/o, lesion on
lower limbs,
previous
amputation of
lower limbs,
under
hemodialysis,
wheelchair,
CBA, psych
diseases,
participating in
another GBE,
cultural factors
affecting ability
to understand
instruments
Sample/Setting

N=341
n=170 (IG)
n=171 (CG)
Setting:
primary
healthcare units
(8)

treatment
adherence
(p=0.0318),
and
glycemic
control
(p=0.0321).

Significant
results
Limitations:
small sample
size, many
excluding
criteria for
participants
Feasibility:
Allows active
participation of
group
members using
CM.

Major
Variable &
Definitions
IV:
educational
program
DV1: HbA1C
DV2: lipid
levels

Measuremen
t

Analysis

Findings

Decision for
Use

HbA1C, lipid
panels

SPSS, R version
3.0.1,ShapiroWilk, Box-Cox
transformation,
Bonferroni
correction, chisquare with Yates
correction,

There were
statistically
significant
changes in
the HbA1C
over time in
both
groups. The

LOE: LOE II
Strengths:
Statistically
significant
results

KEY: BMI-body mass index; BP-blood pressure; CCT-clinically controlled trial; CCM-Chronic Care Model; CG-control group; CIconfidence interval; CM-Conversation Maps; DKN-A-Diabetes Knowledge Scale; DKQ-Diabetes Knowledge Questionnaire; DVdependent variable; DM-diabetes mellitus; EMR-electronic medical record; FBG-fasting blood glucose; GBE-group-based
education; HbA1C-glycated hemoglobin; HDL-high-density lipoprotein; HTN-hypertension; HV-home visits; IBE-individual-based
education; IG-intervention group; IV-independent variable; LDL-low-density lipoprotein; LOE-level of evidence; MA-mean age;
MAT-Measure of Adherence to Treatments; MP-medication persistence; MPR-medication possession ratio; PMA-primary
medication adherence; RCT-randomized clinical trial; SCT-Social Cognitive Theory; SPSS-Statistical Package for Social
Sciences;T2-type 2; UC-usual care; WC-waist circumference

DIABETES EDUCATION
randomized
clinical trial

36
patients with
DM

Funding:
educational
grants
Bias: selection
Country:
Brazil

Demographics
:
Average age
60.6
Mostly female
participants
(69% of CG
and 74.7% of
IG)
Inclusion: age
30-70,
diagnosed with
DM

Note:
educational
program
consisted of
GBE, home
visits, and
phone calls

McNemar, tStudent test, tStudent-Welch
test
5% significance
level

means were
significantl
y lower in
the
intervention
group
(p<0.05).

Measuremen
t

Analysis

Findings

DKN-A,
SDSCA,
PAID

SPSS,
KolmogorovSmirnov,
Shapiro-Wilk
tests, Friedman’s,
Multiple
Comparison,
Mann Whitney

Significant
improveme
nts in scores
related to
knowledge
of disease
in both
groups (IBE

Exclusion:
illiteracy,
complications
of disease

Citation

Conceptual
Framework

Design/Metho
d

Sample/Setting

Imazu, M.
(2015).
Effectiveness of
individual and
group
interventions
for people with
T2 diabetes.

Inferred
CCM

Design:
comparative,
longitudinal
and prospective
study (nonrandomized)

N=150
n=75 (IBE)
n=75 (GBE)
Setting: private
healthcare
service

Major
Variable &
Definitions
IV1: IBE
IV2: GBE
DV1:
knowledge of
disease
questionnaire
scores

Limitations:
wide range of
exclusion
criteria, 27.3%
attrition
Feasibility:
Flexible
strategies for
awareness and
self-care for
DM may be
helpful for
patients. May
be difficult to
replicate
specific
program,
involving
multiple
methods.
Decision for
Use
LOE: LOE III
Strengths:
Limitations:
nonrandomized,
high rates of

KEY: BMI-body mass index; BP-blood pressure; CCT-clinically controlled trial; CCM-Chronic Care Model; CG-control group; CIconfidence interval; CM-Conversation Maps; DKN-A-Diabetes Knowledge Scale; DKQ-Diabetes Knowledge Questionnaire; DVdependent variable; DM-diabetes mellitus; EMR-electronic medical record; FBG-fasting blood glucose; GBE-group-based
education; HbA1C-glycated hemoglobin; HDL-high-density lipoprotein; HTN-hypertension; HV-home visits; IBE-individual-based
education; IG-intervention group; IV-independent variable; LDL-low-density lipoprotein; LOE-level of evidence; MA-mean age;
MAT-Measure of Adherence to Treatments; MP-medication persistence; MPR-medication possession ratio; PMA-primary
medication adherence; RCT-randomized clinical trial; SCT-Social Cognitive Theory; SPSS-Statistical Package for Social
Sciences;T2-type 2; UC-usual care; WC-waist circumference

DIABETES EDUCATION

37
Purpose:
Compare
effectiveness of
GBE vs. IBE
on knowledge
of disease,
quality of life,
and self-care
practices

Country:
Brazil
Funding: not
identified
Bias: Selection
bias due to
convenience
sampling

Demographics
:
MA 60 years,
56% female,
80% white
Inclusion:
enrolled in the
“Chronic
Patient
Mentoring
Program”, >18
y/o, diagnosed
with T2DM,
with or without
comorbidities

Citation

Conceptual
Framework

Design/Metho
d

Sample/Setting

McEwan, M.
M. et al. (2017).
Effects of a
family-based
diabetes
intervention on
behavioral and
biological

Selfefficacy

Design: RCT

N=157
n=83 (IG)
n=74 (CG)

Purpose:
Analyze effects
of a familybased selfmanagement

Setting: urban
Hispanic
neighborhoods

DV2: quality
of life
questionnaire
scores
DV3: self-care
practices
questionnaire
scores

SDSCA:
Summary of
Diabetes SelfCare Activities
questionnaire
(validated)
PAID:
Problem Areas
in Diabetes
questionnaire
(validated)
Major
Variable &
Definitions
IV: familybased diabetes
intervention
DV1:
Family-based
intervention:
12-week

p=0.003,
GBE
p=0.008).
Significant
improveme
nts in
quality of
life in IBE
(p=0.007).
Significant
improveme
nt in selfcare
practices in
GBE
(p<0.01).

attrition,
especially in
GBE, presence
of
comorbidities
as uncontrolled
variables
Feasibility:
The use of
combination of
IBE and GBE
may be
beneficial for
different areas
in the DM
management
and education
components.

Measuremen
t

Analysis

Findings

Decision for
Use

DKQ, Newest
Vital Sign,
weight
balanced
scale,
Diabetes SelfCare
Activities

SPSS; Chi-square
and t-tests,
ANOVA

Significant
effect of
familybased
intervention
s on DM
selfmanagemen

LOE: LOE II
Strengths:
Intervention
increased selfmanagement,
similar

KEY: BMI-body mass index; BP-blood pressure; CCT-clinically controlled trial; CCM-Chronic Care Model; CG-control group; CIconfidence interval; CM-Conversation Maps; DKN-A-Diabetes Knowledge Scale; DKQ-Diabetes Knowledge Questionnaire; DVdependent variable; DM-diabetes mellitus; EMR-electronic medical record; FBG-fasting blood glucose; GBE-group-based
education; HbA1C-glycated hemoglobin; HDL-high-density lipoprotein; HTN-hypertension; HV-home visits; IBE-individual-based
education; IG-intervention group; IV-independent variable; LDL-low-density lipoprotein; LOE-level of evidence; MA-mean age;
MAT-Measure of Adherence to Treatments; MP-medication persistence; MPR-medication possession ratio; PMA-primary
medication adherence; RCT-randomized clinical trial; SCT-Social Cognitive Theory; SPSS-Statistical Package for Social
Sciences;T2-type 2; UC-usual care; WC-waist circumference

DIABETES EDUCATION
outcomes for
Mexican
American
adults
Bias: Potential
selection/recruit
ment
Funding: grant
from National
Institute for
Minority Health
and Health
Disparities
Country: USA

38
support
intervention

in Arizona
border region
Inclusion:
Mexican
American,
T2DM at lest 1
year, 35-74 y/o,
had 1 family
member
participating
>18 y/o who
lived with or
saw weekly
Exclusion: DM
education in the
past 1 year
Demographics
: MA=53.53
y/o, 85%
female, 93.6%
overweight
Family
members’
MAs=47.27,
72.6% female

program
including 12
hours of
support group
sessions, 3
weekly 2-hour
home visits,
and 3 weekly
20- minute
phone calls

Questionnaire,
Self-Efficacy
for Diabetes
Scale, Fat,
Fruit, and
Vegetable
questionnaire
International
Physical
Activity
Questionnaire

t, exercise,
selfefficacy,
distress. No
significant
changes in
HbA1C.

geographical
region
Limitations:
More timeintensive and
potentially
costly
interventions
involved. Lowincome sample
Feasibility:
Culturally
relevant
family-based
education may
be beneficial.
However,
grocery
certificates
were dispersed
for incentive
for participants
at data
collections.
“Booster
sessions” may
be necessary to
maintain
glycemic
control.

KEY: BMI-body mass index; BP-blood pressure; CCT-clinically controlled trial; CCM-Chronic Care Model; CG-control group; CIconfidence interval; CM-Conversation Maps; DKN-A-Diabetes Knowledge Scale; DKQ-Diabetes Knowledge Questionnaire; DVdependent variable; DM-diabetes mellitus; EMR-electronic medical record; FBG-fasting blood glucose; GBE-group-based
education; HbA1C-glycated hemoglobin; HDL-high-density lipoprotein; HTN-hypertension; HV-home visits; IBE-individual-based
education; IG-intervention group; IV-independent variable; LDL-low-density lipoprotein; LOE-level of evidence; MA-mean age;
MAT-Measure of Adherence to Treatments; MP-medication persistence; MPR-medication possession ratio; PMA-primary
medication adherence; RCT-randomized clinical trial; SCT-Social Cognitive Theory; SPSS-Statistical Package for Social
Sciences;T2-type 2; UC-usual care; WC-waist circumference

DIABETES EDUCATION

39
Appendix B

OdgersJewell et
al.
2017
I
review
yes

Merakou,
K. et al.

Reaney,
M. et al.

2015
II
CCT
yes

2013
II
RCT
yes

Year
LOE
Design
T2DM
only
Attrition
# of
Participants
Country

N/A
8,533

0%
193

8.2%
681

Assorted

Greece

Intervention

GBE

GBE

Germany/
Spain
GBE

GBE effect on
medication
adherence
GBE effect on
hospital visits
for diabetes
complications
GBE effect on
KOD

Synthesis Table
Santos, Hwee, J. O’Connor, Figueira,
J. C. et
et al.
et al.
A. et al.
al.
2017
2014
2014
2017
II
IV
II
IV
RCT
PBC
RCT
SCRI
yes
T1DM
yes
yes
included
0%
21%
12%
28%
238
77,824
2,378
82

Torres,
H. C. et
al.
2018
II
CRCT
yes

Imazu,
M.
2015
III
CLPS
yes

McEwan,
M. M. et
al.
2017
II
RCT
yes

27%
341

32%
150

45%
157

Brazil

Canada

U. S.

Brazil

Brazil

Brazil

IBE,
ME,
GBE

TC

GBE

E

IBE,
GBE

GBE,
HV, TC

U.S.
(Arizona)
GBE, HV,
TC

N/A

N/A

N/A

N/A

N/A

N/A

N/A

N/A

NSE




N/A

N/A

N/A

N/A



N/A

N/A

N/A

N/A

N/A



N/A



N/A

N/A

N/A



N/A



N/A

Key: -decreased; -increased; CLPS- comparative, longitudinal, & prospective study; CRCT- cluster randomized controlled trial;
E-education; GBE-group based education; HbA1C-glycated hemoglobin; HV-home visits; IBE-individual based education; KODknowledge of disease; ME-mixed GBE & IBE; N/A-not applicable in this study; NSE-no significant effect; PBC-population based
cohort study; SCRI-single comparison randomized intervention; T2DM-type 2 diabetes mellitus; TC-telephone calls; U.S.-United
States

DIABETES EDUCATION
GBE effect on
HbA1C



40






N/A

NSE





N/A

NSE

KEY: BMI-body mass index; BP-blood pressure; CCT-clinically controlled trial; CCM-Chronic Care Model; CG-control group; CIconfidence interval; CM-Conversation Maps; DKN-A-Diabetes Knowledge Scale; DKQ-Diabetes Knowledge Questionnaire; DVdependent variable; DM-diabetes mellitus; EMR-electronic medical record; FBG-fasting blood glucose; GBE-group-based
education; HbA1C-glycated hemoglobin; HDL-high-density lipoprotein; HTN-hypertension; HV-home visits; IBE-individual-based
education; IG-intervention group; IV-independent variable; LDL-low-density lipoprotein; LOE-level of evidence; MA-mean age;
MAT-Measure of Adherence to Treatments; MP-medication persistence; MPR-medication possession ratio; PMA-primary
medication adherence; RCT-randomized clinical trial; SCT-Social Cognitive Theory; SPSS-Statistical Package for Social
Sciences;T2-type 2; UC-usual care; WC-waist circumference

DIABETES EDUCATION

41
Appendix C
Figure 1

Reprinted with permission from ACP-ASIM Journals and Books (Wagner, 1998)

DIABETES EDUCATION

42
Appendix D
Figure 2

(Dawson & Andriopoulos, 2017)

DIABETES EDUCATION

43
Appendix E
Budget
Activities

Design and print pamphlets for
education program (200)
Design & print educational
materials for staff meeting
Related physician loss of
productivity for staff meeting cost
Related staff loss of productivity
for staff meeting cost
Food items for staff meeting
Locking box and files for patient
paper files related to project
TOTAL

Cost
$40
$2
$160
$14.85
$25
$60

$301.85