A simple nurse-based prompt increases screening and prevention counseling for diabetes☆

1. Introduction 

Type 2 diabetes is an increasing problem in the United States. Approximately 17 million Americans have diabetes and an estimated 12 million US citizens have pre-diabetes (impaired fasting glucose or impaired glucose tolerance) [1], [2]. Diabetes increases the risk of many medical disorders including cardiovascular disease, renal failure, amputations, blindness and early death [3], [4], [5]. Numerous successful interventions have been demonstrated to delay or prevent both the onset of diabetes and the complications associated with diabetes using both lifestyle modification and medications [6], [7], [8], [9], [10], [11], [12], [13]. Both the ADA and the American Academy of Clinical Endocrinologists advocate early screening and intervention [14], [15]. Despite these recommendations, recent reports indicate that nearly 5 million Americans have undiagnosed diabetes and nearly half of these persons have levels of hemoglobin A1c in the range associated with diabetes complications [1], [2], [16].

There have been many reports concerning the costs and strategies for screening for diabetes. However the conclusions suggest that routine population screening is too costly, not successful, or not substantial [17], [18], [19], [20], [21]. Efforts to improve diabetes screening and early intervention have had mixed results. In a recent survey in Montana 48% of persons with two diabetes risk factors and 37% of persons with three risk factors had not been screened for diabetes during the past 3 years [22]. Unfortunately, opportunistic rather than population screening based on increased risk for diabetes is the more common method for identifying persons with diabetes [23], [24]. A large population-based screening study using a self-administered risk chart distributed by mail, followed by office-based glucose testing, identified only 19% of persons with undiagnosed diabetes [25].

Although there are a number of reports of community-based screening, there are few published reports of efforts to improve office-based screening and intervention in order to increase preventive counseling in persons at high-risk for diabetes [20], [24], [26]. Fasting glucose (FG) or oral glucose tolerance test (OGTT) are the recommended screening tests for diabetes, however, these are often not used in primary care practice. In a recent office-based HMO study spanning 3 years, only 90 (1%) of 8286 patients 45 years old and older were screened with a fasting glucose and only 11 (0.1%) received an OGTT. Additionally, only 38% of the patients with abnormal results received appropriate follow-up testing and treatment [23].

Prompts have been shown to improve the delivery of preventive services in a number of settings including primary care offices [27], [28], [29]. Based on these findings, we hypothesized that having a nurse score a diabetes risk survey and prompt the physicians for patients at increased risk for diabetes would result in increased screening and preventive counseling for diabetes. The purpose of this study was to assess the impact of a simple nurse-based prompt on screening and intervention rates for persons at increased risk for diabetes in an outpatient setting.

3. Results 

Of the 1438 subjects who were recruited from 10 outpatient primary care practices, 1176 met eligibility criteria. In all practices the nurses took on average less than 20s to determine the BMI and score the prompt.

The basic anthropometric characteristics of eligible subjects are shown in Table 1. A total number of 597 intervention and 579 control subjects were eligible for this investigation. Although there were no significant differences in ADA risk scores, some differences were found between groups. Overall, the intervention group was older than the control group (p<0.001). Compared to the intervention group, the control group was heavier as defined by BMI (p<0.001). There were gender and racial/ethnic differences between intervention and control groups. The proportions of women and Blacks were higher in the control group compared to the intervention group (p<0.001).

Table 1. Characteristics of studied variables

		Intervention group	Control group	p-value
	Sample size (n)	597	579
	Age (y)	52.9±16.7	46.5±14.7	<0.001
	BMI (kg/m2)	29.5±6.3	31.1±6.7	<0.001
	ADA risk score	3.85±1.98	3.87±1.96	0.889
	Degree of adiposity (%)
	Normal	22.9	17.3	0.001
	Overweight	35.1	29.2
	Obese	42.0	53.5
	Gender (%)
	Men	35.7	26.6	<0.001
	Women	64.3	73.4
	Race/ethnicity (%)
	White	76.2	49.4	<0.001
	Black	19.9	47.7
	Others	3.9	2.9

ADA, American diabetes association.

Comparative analyses of ADA diabetes risk factors in the intervention and control groups are shown in Table 2. There were significantly more subjects with an age of ≥45 years, a history of vascular disease, and an immediate family member with diabetes in the intervention compared with the control group. The control group had more subjects with BMI ≥25 (kg/m2), and more at risk subjects (defined as non-Whites) compared with the intervention group. There were more females who had babies weighing more than 9lbs in the intervention group compared with the control group.

Table 2. Diabetes risk factors (in %) of intervention and control groups

	Risk factor	Intervention group	Control group	p-value
	BMI greater than 25 (kg/m2)	77.1	82.7	0.024
	Age ≥45 (years)	67.7	48.0	<0.001
	At-risk race	23.8	50.6	<0.001
	HDL<35 (mg/dl) or TG>200 (mg/dl)	19.1	18.3	0.333
	Polycystic ovarian syndromea	1.0	2.1	0.226
	History of hypertension	80.5	77.3	0.333
	History of IFG or IGT	0.8	0.6	0.663
	History of vascular disease	13.2	5.9	<0.001
	Immediate family with diabetes	37.0	44.6	0.009
	Have had baby weighing greater than 9lbsa	17.8	12.5	0.036
	Have had gestational diabetesa	3.1	4.2	0.408
	History of frequent urination	19.8	16.2	0.115
	History of blurred vision	12.9	11.6	0.488
	History of pain while walking	31.8	32.3	0.862

IFG, impaired fasting glucose; IGT, impaired glucose tolerance.

a Values are for women (n=808).

Following the intervention, the percent of intervention versus control subjects with diet, exercise and weight control plans were 28.8 versus 3.8, 23.6 versus 3.8 and 6.4 versus 3.1, respectively (p<0.001). The percent of intervention versus control subjects receiving FG screening was 11.6% versus 1.4% (p<0.001). The impact of the nurse prompt on FG screening, diet plans, exercise plans, and weight loss plans are shown in Table 3. Before adjusting for other independent variables (Model I), the nurse prompt was associated with 10.2, 7.4, 6.8, and 1.5 increased odds of FG screening, diet, exercise, and weight loss plans, respectively. In the multiple logistic regression analyses (Model II), adjusting for age, BMI, gender, race/ethnicity and ADA risk score, the nurse prompt was also associated with 9.3, 6.1, 7.4, and 1.9 increased odds of FG screening, diet, exercise, and weight loss plans, respectively. Adjusting for other covariates, female gender and increased age were also associated with increased odds of FG screening, diet, and exercise plans (p<0.01).

Table 3. Relationship between simple nurse prompt and odds of receiving fasting glucose screening, diet plan, exercise plan, and weight loss plan

		Diet Plan		Exercise plan		Weight loss plan		FG Screening
		OR, 95% CI	p	OR, 95% CI	p	OR, 95% CI	p	OR, 95% CI	p
	Model I: univariate
	Prompt	7.4, 4.5–12.4	<0.001	6.8, 5.0–15.6	<0.001	1.5, 1.0–2.8	<0.001	10.2, 4.6–22.3	<0.001
	Model II: multivariate
	Prompt	6.1, 3.5–10.7	<0.001	7.4, 4.0–13.9	<0.001	1.9, 1.1–3.7	0.035	9.3, 3.6–24.0	<0.001
	Age	1.3, 1.1–1.8	<0.001	1.4, 1.3–1.7	<0.001	1.2, 1.0–1.4	0.092	1.5, 1.3–1.6	0.001
	BMI	1.4, 1.0–1.7	0.050	1.1, 0.9–1.1	0.257	1.2, 1.1–1.3	0.014	1.0, 0.9–1.4	0.993
	Sex	2.3, 1.5–3.5	<0.001	2.3, 1.5–3.5	<0.001	1.2, 0.6–2.4	0.574	2.0, 1.2–3.3	0.008
	Race	0.9, 0.5–1.5	0.648	0.7, 0.4–1.1	0.127	0.4, 0.2–0.8	0.016	1.4, 0.7–2.7	0.299
	Score	1.1, 0.9–1.2	0.739	1.0, 0.9–1.1	0.663	0.9, 0.8–1.2	0.686	0.9, 0.8–1.1	0.342

OR, odds ratio from the logistic regression models; p, p-value; CI, confidence intervals; FG, fasting glucose; sex, female; race, African American; score, ADA risk score.

Data from the chart review also revealed that following the intervention, 70.9% of the subjects in the intervention group had a notation of high-risk for diabetes on their chart whereas only 29.1% of the subjects in the control group had such a notation (p<0.001).

4. Discussion 

This study employed a simple office-based intervention which significantly improved office-based management of patients at risk for diabetes at three levels: identification, as evidenced by documentation in the medical record; diagnosis by fasting glucose testing; and provision of advice regarding diet, exercise and weight reduction. The intervention employed three strategies for improving diabetes screening and treatment which have been advocated by the ADA: improving health professional education regarding the standards of diabetes care, using reminders and prompts, and training nurses to use algorithms [14]. Implementation of the intervention required approximately 30min of nurse training and orientation of the physician providers, which was usually accomplished over lunch. Scoring of the prompt and calculations of BMI required minimal additional nursing time (only about 20s per patient). This simple, low cost, office-based intervention shows promise for increasing the identification and treatment of patients with diabetes and prediabetes in primary care settings.

As the prevalence of obesity and diabetes continue to increase in the United States, it is imperative that primary care professionals improve their ability to provide early screening and delivery of services for diabetes prevention in persons at high-risk for diabetes. Early detection and intervention are also needed to decrease the high number of individuals who are already experiencing the complications of diabetes at the time of diagnosis [31], [32], [33], [34], [35]. Implementation of early screening and lifestyle change counseling in primary care settings is challenging because of the time limitations of primary care visits and the many competing demands for screening and prevention counseling [36]. The low cost, time efficient intervention employed in this project was effectively implemented, despite these limitations. Once physician and nurse training have been completed, the cost of ongoing screening is equal to the cost of copy paper and the brief amount of nurse time necessary to score the survey. This makes this intervention easily sustainable for most outpatient practices.

This prompt was also implemented in one practice with an electronic medical record (EMR) without difficulty. The paper prompt was scanned into the chart on the same day and became a permanent part of the electronic record. An automated electronic prompt could prove to be an even more effective tool. Further studies are needed to determine the effectiveness of a prompt embedded in an EMR on diabetes screening and intervention rates. The ADA recommends screening at 3-year intervals in all persons aged ≥45 (particularly those with BMI ≥25) as well as in younger persons who have elevated BMI and another risk factor [14]. Future studies should include chart prompts for periodic re-screening of patients who screen negative. Again, an automated electronic prompt within an EMR seems a simple method for accomplishing this goal.

5. Conclusion 

In our trial, a simple nurse-based prompt was an effective instrument that significantly increased documentation of diabetes risk, fasting glucose screening, and the delivery of preventive services, including advice to improve diet, increase exercise and lose weight. Further studies are needed to develop more effective office-based tools to increase screening and early intervention for persons at risk for diabetes.