Close Server: KOPWWW05 | Not logged in


Recognition of Undiagnosed Diabetes

This practice improvement initiative demonstrated potential for benefit.

Section Sponsored by:

Diabetes has reached epidemic proportions in the United States. The Centers for Disease Control and Prevention estimates that nearly 26 million Americans have diabetes and approximately 27% of them do not know it.1 Type 2 diabetes mellitus (T2DM) is a chronic metabolic disease characterized by hyperglycemia, impaired insulin secretion, insulin resistance, excessive hepatic glucose production, and abnormal fat metabolism. It can be undetected for many years and, during this time, significantly damage the vascular and nervous systems and organs.2 An estimated 79 million American adults have prediabetes, which raises the risk of heart disease, stroke and T2DM.3

Hyperglycemia in the hospital setting has been defined as any blood glucose level greater than 140 mg/dL (7.8 mmol/L).4,5 The American Diabetes Association recommends treating any levels above this that are significant and persistent during the hospital stay. HbA1c values greater than 6.5% suggest undiagnosed diabetes that preceded hospitalization.4 Because 7 million Americans have undiagnosed diabetes, early diagnosis and preventive measures are needed.1

Background and Significance

T2DM is the seventh leading cause of death in the United States.6 The incidence of T2DM has doubled over the last 30 years.7 Undiagnosed diabetes can cause progressive microvascular damage.8 At the time of T2DM diagnosis, approximately 20% of patients have diabetic retinopathy and 10% have nephropathy.8 T2DM accounts for 90% to 95% of all diagnosed cases of diabetes in adults, and it is on the increase in children and adolescents.9 One in 3 adults could have diabetes by 2050.1

Delayed diagnosis translates into more organ damage because effective therapy cannot be applied. Therefore, early screening of patients with elevated fasting plasma glucose (FPG) is essential. Many patients admitted to a hospital do not visit a primary care provider on a regular basis. Therefore, their diabetes is undiagnosed and may remain so for several more years.

HbA1c can be helpful as a screening and diagnostic tool in addition to FPG; it reflects the average level of glucose the red blood cell has been exposed to over its 120-day life span. Glucose in the blood binds irreversibly to a specific part of hemoglobin in red blood cells, forming HbA1c. Therefore, HbA1c is particularly valuable in distinguishing stress-induced hyperglycemia during a hospital stay from the sustained hyperglycemia characteristic of diabetes.10 By utilizing the HbA1c as a screening tool in combination with the FPG during a hospital stay, previously undiagnosed diabetes can be recognized, permitting earlier diabetes management and education.11

Study Design

The goal of this practice improvement project was to determine if an HbA1c blood test, when added to standard blood work on hospitalized patients with hyperglycemia (FPG ≥ 140 mg/dL), could help identify patients with previously undiagnosed T2DM.

The design was both retrospective and descriptive. A standardized procedure was used to identify patients with undiagnosed diabetes in a hospital setting in southwest Ohio. It sought to improve patient outcomes by identifying disease early through convenience sampling. If tests were positive, steps could be initiated by providers in collaboration with patients to promote healthy lifestyles.

This improvement project did not exclude any patient based on race or gender. The inclusion criteria were as follows: ages 20 to 65, FPG ≥ 140 mg/dL, nondiabetic by Epic Medical Health Record lab review (brand of electronic health record used at the facility), and not receiving corticosteroids or intravenous fluid therapy containing glucose. All patients who met the inclusion criteria between September 2012  and January 2013 were included.

I was the project leader, and I performed a two-step process. Step 1 evaluated the rate of elevated FPG in all hospitalized patients who had no prior history of T2DM, were not receiving corticosteroid therapy (during their hospital stays), and who were not receiving intravenous fluid therapy that contained glucose. Through retrospective study using Epic Medical Health Record lab review, the project leader used convenience sampling to identify random patients who met the inclusion criteria. I settled on 20 patients with an FPG ≥ 140 mg/dL, which suggests a high need for evaluation. Thus, by examining patient records until 20 patients with no prior diagnosis of T2DM were identified with FPG ≥ 140 mg/dL, the approximate rate of previously unidentified fasting hyperglycemia in this hospitalized population, i.e. the approximate frequency of this problem, could be estimated.

All patients had been admitted through the emergency department, where the first blood draw was performed. Since some patients may not have been fasting depending on the time of admission, the project leader used the FPG of the next day to ensure an 8-hour fasting period. All information about blood test results was obtained from blood specimens that had already been drawn routinely for the purposes of patient care.

Step 2 (prospective data) identified 20 patients with fasting hyperglycemia (FPG ≥ 140 mg/dl). The providers for each of the 20 patients were notified of the fasting hyperglycemia and, per written/phone order, agreed to add an HbA1c blood test to the existing laboratory orders. The laboratory of the hospital was informed of the project and utilized the blood of the assigned patients for testing HbA1c. The method used for measuring HbA1c is considered valid because it met National Glycohemoglobin Standardization Program requirements. Once results were received from the laboratory, the project leader notified the provider of the results. The provider notified the patients of the results of the added HbA1c and the possible need for treatment, education or other interventions.

I used the Health Belief Model as a guideline to address the importance of screening with the HbA1c to help identify possible T2DM. The Health Belief Model was chosen because it is a method to assess patients' perceptions of modifying factors, as well as their belief about the seriousness of diabetes (i.e., complications) and susceptibility (i.e., elevated blood glucose) and help guide them through a health education program to initiate behavior change.12 It was used to address the importance and benefits of screening (here, HbA1c) to prevent or detect diabetes.12

Correlation analysis of the HbA1c results to each patient's FPG was performed to determine if an added HbA1c was useful in diagnosing diabetes in previously undiagnosed patients. In addition, the estimated average glucose (eAG) was calculated using the formula: 28.7 x A1c% - 46.7. The eAG is easier for the patient to understand since the unit of measurement (mg/dL) is the same as what they are used to seeing with other "blood sugar" numbers.13

Statistical analysis was performed using Microsoft Excel. Step 1 results were presented in a frequency distribution table with relative frequency and a cumulative frequency polygon. Step 2 results were presented in a table and as a correlation scatter plot.

This project was approved by the institutional review board at the Mercy Jewish Hospital in Cincinnati, Ohio, and the Medical University of South Carolina, where the project leader was a DNP student. All collected data was stored on a password-protected secure network storage site at the academic site. There were no physical, psychological, social, legal or other risks to the participants involved in this project. The collected data for this project and results were only presented to the project leader and provider with a HIPAA agreement in place.


Step 1, retrospective data from the electronic lab review, showed that a review of 733 patient records was required to identify 20 patients who presented with hyperglycemia defined as an FPG ≥ 140 mg/dL. Most patients presented with an FPG in the range of 140 mg/dL to149 mg/dL (35%) followed by the range of 180 mg/dL to 189 mg/dL (20%). Step 2, prospective data, found high frequencies of prediabetes (55%) and moderate frequencies of diabetes (30%) compared to low frequencies of non-diabetes (15%) in the 20 patients. Calculation of eAG was performed using the formula: 28.7 x HbA1c (%) - 46.7. The FPG was then correlated to the HbA1c and revealed a strong linear relationship between FPG and HbA1c, which was statistically significant (r = 0.977, n = 20, p ? 0.001). Through this intervention, it was concluded that an added HbA1c is useful in diagnosing diabetes in previously undiagnosed patients.


This practice improvement initiative showed that an added HbA1c to existing ordered blood work in patients with a FPG ≥ 140 mg/dL can help detect prediabetes and diabetes in undiagnosed patients. Although the frequency of patients was relatively low in Step 1, the case findings in Step 2 presented with high frequencies of prediabetes and moderate frequencies of diabetes. Considering the prevalence of T2DM in the general U.S. population, the finding that 30% of included patients had diabetes is high.14 Although the sample size was small and the age group limited to 20 to 65, results may have been missed if an HbA1c had not been added. The diabetes identified with lower range HbA1c results especially benefitted from this project because they may have otherwise been overlooked.

A broader sampling including patients older than 65 might produce an even higher prevalence of diabetes. However, because this population qualifies for Medicare and therefore a preventive yearly exam, diabetes is more likely be screened for and diagnosed outside the hospital setting. The project leader used a convenience sample and did not monitor patients daily, which may have influenced the Step 1 data. The data may provide a basis for altering standards of care within the hospital and identifying undiagnosed prediabetes and diabetes to ensure early recognition and early intervention for this chronic disease.

The Health Belief Model was appropriate for initial screening because it focuses on the short-term constructs of cost/benefit and perceived threat to initiate action. Although not part of the study design, the patients in Step 2 reported that they found education with the Health Beliefs Model beneficial and "eye opening."

In summary, this project initiative of 40 patients admitted to a hospital with hyperglycemia provided pilot data about the prevalence of undiagnosed diabetes. Most patients with impaired plasma glucose have a predisposition to acquire undiagnosed prediabetes and diabetes. Therefore, it is important to screen patients with a FPG ≥ 140 mg/dL for prediabetes and diabetes and to initiate timely education and management with an appropriate educational model. 


1. Centers for Disease Control and Prevention. Number of Americans with diabetes rises to nearly 26 million.

2. Cox ME, Edelman D. Tests for screening and diagnosis of type 2 diabetes. Clin Diab. 2009;27(4):132-138.

3. Centers for Disease Control and Prevention. National Diabetes Prevention Program. Prediabetes. Am I at risk?

4. American Diabetes Association. Standards of medical care in diabetes, 2012. Diabetes Care. 2012;35(Suppl 1):S11-S63.

5. Moghissi ES, et al. American Association of Endocrinologists and American Diabetes Association consensus statement on inpatient glycemic control. Diabetes Care. 2009;32(6):1119-1131.

6. National Center for Chronic Disease Prevention and Health Promotion. National diabetes fact sheet, 2011.

7. Fox CS, et al. Trends in the incidence of type 2 diabetes mellitus from the 1970s to the 1990s. The Framingham Heart Study. Circulation. 2006;113(25):2914-2918.

8. McCulloch D.K. Screening for diabetes mellitus. Available with subscription at

9. National Diabetes Education Program (NDEP). The facts about diabetes: a leading cause of death in the US.

10. Dugger L, et al. Can hemoglobin A1c testing in hospital settings help the early identification of diabetes? Clin Nurse Spec. 2011;25(4):171-175.

11. American Diabetes Association. American College of Endocrinology and American Diabetes Association consensus statement on inpatient diabetes and glycemic control. Endocrine Practice. 2006;12(4):458-468.

12. Glanz K, et al. The Health Belief Model. In: Health Behavior and Health Education: Theory, research and practice. 4th ed. San Francisco, CA: Jossey Bass; 2008: 46-50.

13. American Diabetes Association. Estimated average glucose.

14. National Diabetes Information Clearinghouse. National diabetes statistics, 2011.  

Diana Cizmadija is a family nurse practitioner who practices at Evendale Family Medicine in Cincinnati. She has completed a disclosure statement and reports no relationships related to this article.


Articles Archives


Email: *

Email, first name, comment and security code are required fields; all other fields are optional. With the exception of email, any information you provide will be displayed with your comment.

First * Last
Title Field Facility
City State

Comments: *
To prevent comment spam, please type the code you see below into the code field before submitting your comment. If you cannot read the numbers in the below image, reload the page to generate a new one.

Enter the security code below: *

Fields marked with an * are required.

View the Latest from ADVANCE


Back to Top

© 2017 Merion Matters

660 American Avenue Suite 300, King of Prussia PA 19406