Alliance for a Healthy Border: Factors Related to Weight Reduction and Glycemic Success

Abstract

We examined the factors related to success in achieving weight reduction and glycemic control in Alliance for a Healthy Border (AHB), a chronic disease prevention program implemented from 2006 to 2009 through 12 federally qualified community health centers serving primarily Hispanics in communities located along the US–Mexico border region. We analyzed data from Phase I of AHB using logistic regression to examine the determinants of success in achieving weight reduction and glycemic control among the participants in AHB programs. Factors affecting weight reduction success were sex, age, employment status, income, insurance, diabetes, baseline body mass index (BMI), smoking status, family history of diabetes, session type, program duration, and physical activity changes. Factors affecting achievement of glycemic success included sex, age, employment status, diabetes, baseline BMI, family history of diabetes, program duration, and physical activity changes. We found that the AHB interventions were more successful in reducing participants' HbA1c level than BMI. In addition to sociodemographic factors, participants with better baseline health conditions (ie, participants without diabetes or family history of diabetes, normal BMI, former smokers) were more likely to achieve success after the interventions. Of the 4 key features defining each of the 12 interventions, session type and program duration were associated with success. Within a relatively short time period, physical activity improvements had a stronger effect on weight reduction and glycemic success than improvements in dietary habits. The effectiveness of diabetes and cardiovascular disease prevention programs can be improved substantially by considering these factors during program design and structure. (Population Health Management 2012;15: 90–100)

Introduction

D iabetes is a serious epidemic that affects more that 220 million people worldwide.¹ In the United States, it is estimated that over 23.6 million people, or 7.8% of the population, have diabetes.² Hispanics, the largest and fastest growing minority group in the United States, are disproportionately affected by diabetes and its complications compared to other ethnic and racial groups³; diabetes prevalence and mortality rates for Hispanics are 2 to 5 times higher than those of other ethnic and racial groups.³ According to the National Diabetes Prevention and Control Program, 1 in 10 (2 million) Hispanic adults in the United States have diabetes, while it is estimated that 1 in 3 Hispanic adults live with diabetes but are unaware of their condition.⁴ Hispanics with type 2 diabetes are 2 times more likely to develop diabetes-related complications than their non-Hispanic white counterparts.⁵

The burden of diabetes is especially alarming in US–Mexico border communities, a region with a predominantly Hispanic population (80% of the population in border counties).⁶ The Hispanic population is known to have high rates of obesity,^7

–10 low income levels,^9
–11 low education levels,^9
–11 and low levels of health care coverage,¹² all of which are risk factors related to type 2 diabetes. The effects are already evident in the US–Mexico border region; the US border population has a 16.1% diabetes prevalence rate and a 13.6% prediabetes prevalence rate.⁶

Given that the border population is at high risk for diabetes, it is important to implement well-designed prevention programs to reduce such risk. Although research shows that diabetes is caused partly by nonmodifiable genetic factors,^13,14 it has been shown that modifiable lifestyle factors also play an important role in the prevention and control of diabetes.^15

–20 However, participation in prevention programs does not translate into successful outcomes for everyone. Understanding the elements that contribute to certain individuals being more or less successful than others can help with the design of prevention programs that address the needs of those who are least likely to improve their modifiable risk factors. The purpose of this study was to examine the determinants of success in achieving improved weight and glycemic outcomes among participants in chronic disease prevention programs at community health centers located in the US–Mexico border region.

Alliance for a Healthy Border (AHB) is a chronic disease prevention program (2006–2008) that provided resources for nutrition and physical activity education programs at 12 federally qualified community health centers located along the US–Mexico border in Texas, New Mexico, Arizona, and California. The initiative was sponsored by Pfizer Inc. AHB's goals were to reduce modifiable risk factors associated with diabetes and cardiovascular disease, to develop or modify existing prevention programs that target the Hispanic population, and to identify and promote best practices in the prevention of these diseases. To achieve these goals, participants were recruited into the programs through promotions at health fairs, flyers at clinics, provider referrals, and word of mouth. Participants then partook in interactive group or individual class sessions (taught by certified community health workers [promotoras]) on chronic disease prevention, nutrition, and physical activity. Regarding curricula, some centers adapted well-established health programs such as Pasos Adelante (Steps Forward), some centers modified an existing curriculum (eg, Salud para Su Corazón [For the Health of Your Heart]), and some centers used curriculum developed in-house such as Medir para Vivir (Measure to Live). Program duration ranged from 5 to 24 weeks. Although each educational program facilitated in these centers had a unique curriculum, program length, and delivery method, all were based on culturally-sensitive programs that focused on improving nutrition, increasing physical activity, and preventing and managing diabetes and cardiovascular disease appropriately. More detailed information describing 2 representative programs is provided in Table 1.

Table 1.

Representative Programs

Representative Program A	• Expanded its diabetes self-management program to include prevention of diabetes and cardiovascular disease.
Duration: 12 weeks Session Type: Group	• Intervention activities included weight loss and management sessions supported by physical activity, healthy eating, stress management, and smoking cessation.
	• Promotora (community health workers) led educational classes under the supervision of a health educator at clinic facilities.
	• Twelve educational sessions were delivered. Each class was limited to 22 participants.
	• Participants were contacted by phone prior to each educational session or support session.
	• To qualify for a completion certificate, a framed picture, and to participate in a celebration party, patients were encouraged to attend at least 8 of the 12 sessions.
	• Upon completion of the intervention, 12 weekly support sessions were conducted to provide patients with an opportunity to support each other through positive reinforcement and constructive feedback.
Representative Program B	• Program based on the Su Vida, Su Corazón (Your Life, Your Heart) curriculum, with added modules on chronic disease management and group activities excluded.
Duration: 8 weeks Session Type: Individual	• One-on-one sessions were conducted by promotora during biweekly home visits. Each participant received a total of 4 home visits.
	• Each home visit was confirmed by phone.
	• During each visit the next module of the intervention was discussed. The patient would discuss his/her progress as well as challenges encountered. Each visit included measuring blood pressure, blood sugar, weight, hip and waist.
	• The health educator would meet with the promotora to discuss any problems/questions they were not able to handle, so any issues could be solved by the next home visit.
	• Follow-up phone calls were made to patients who are due for their 6-month follow-up.

The AHB programs were implemented and assessed using a pre-post-post study design. Survey instruments were based on questions from the Centers for Disease Control and Prevention's Behavioral Risk Factor Surveillance System and the Community Tracking Study Household Survey. Survey instruments were administered at the beginning of the intervention, at program end, and at 6 months post program end. Pre- and post-program clinical health outcomes and anthropometric measures were also collected.

Methods

The success of program participants was assessed using 2 outcome measures: weight and glycemic outcomes. We defined weight outcomes as a success if participants with a normal baseline body mass index (BMI; BMI ≥18.5 and <25) achieved any weight reduction; or if participants who were overweight or obese (BMI ≥25) at baseline reduced their weight by at least 5%. A 5% reduction in weight is generally considered to be clinically significant in the literature.^17,21,22 The glycemic outcome was defined as a success if glycosylated hemoglobin (HbA1c) levels were maintained at below 6 for participants with baseline HbA1c levels lower than 6. For participants with a baseline HbA1c level ≥6 but less than <7.1, success was a 5% HbA1c reduction; success was a 10% reduction for those with baseline levels ≥7.1 but <7.6; and success was a 15% reduction for those with baseline levels ≥7.6.²³

We included 23 predictor variables in our study that were categorized into sociodemographic factors, baseline health conditions, center and program characteristics, and dietary and physical activity factors. The sociodemographic factors we considered were sex, age, marital status, birth country, employment status (categorized into 4 groups: employed for wages or self-employed; unemployed, students, or those unable to work; homemakers; retired), education level, income, and health insurance coverage status (categorized as “yes” if respondents had any kind of health insurance coverage [eg, private health insurance, prepaid plans such as health maintenance organizations, or government plans such as Medicare]).

Baseline health condition factors considered were self-reported health status, diabetes diagnosis (yes group consisting of type 1 or type 2 diabetes, and diabetes but do not know the type); BMI categories (normal:BMI ≥18.5 and <25; overweight: BMI ≥25 and <30; obese: BMI ≥30); smoking status; family history of diabetes (parents, siblings, and/or children); activity limitations resulting from physical, mental, or emotional problems; hypertension diagnosis by a health professional; and high cholesterol diagnosis by a health professional.

Center and program characteristic factors included were an overall program quality measurement (high versus low), session type (group versus individual education sessions), program duration in weeks, and curriculum (adapted versus developed). Overall program quality measurements were made by an evaluation team based on the following factors: percentage of participants who completed the program and surveys at program end and 6 months later, level of organization, and quality of interaction with the teams at each community health center.

Dietary and physical activity factors included those that indicated participants' improvement of dietary habits (labeled as “improved” if respondent's dietary score increased at least 6 points on the 12-item healthy habits scale, otherwise as “not improved”), fruit and vegetable intake (labeled as “improved” if respondent maintained or met the 5-a-day recommendation for fruit and vegetable consumption, otherwise as “not improved”), and physical activities (labeled as “improved” if respondent maintained or met Healthy People 2010 physical activity recommendations, otherwise as “not improved”). The healthy habits scale is a 12-item scale that includes statements that reflect the level of engagement in healthy dietary habits related to fat/cholesterol intake, salt/sodium intake, and other general habits such as the utilization of nutrition labels and smaller food portions. Items were based on a subset of the "My Family Habits" survey questions of the Salud para su Corazón ("For the Health of Your Heart") initiative. Healthy People 2010 physical activity recommendations: being moderately active for at least 30 minutes 5–7 days a week, or vigorously active for at least 20 minutes 3–7 days a week.

During Phase I of AHB (January 2006–December 2007), each of 12 centers had approximate actual expenses of $160,000. Upon completion of this phase, 2304 participants completed baseline surveys with the program and had baseline measurements; 1549 finished the end-of-program interview and collection of measurements; and 1161 finished the post-6-month interview and collection of measurements. Table 2 lists the rates of enrollment and completion at the program end and at post 6 months. These rates were estimated by demographic and center characteristics. Comparing the retention rates over time, it seems reasonable to assume that data were missing completely at random. We included participants in our analysis if their data were complete for the corresponding outcome and 23 predictors. Therefore, for the end-of-program time point, data from 599 participants were used to evaluate weight outcomes and data from 545 participants were used to assess glycemic outcomes. For the post-6-month time point, data from 540 participants were used to evaluate weight outcomes and data from 488 participants were used to assess glycemic outcomes. Participants with complete data had almost identical distributions on all outcome and predictor variables when compared to all survey participants.

Table 2.

Rates (%) of Enrollment and Completion by Demographic Factors and Center Characteristics

	Start (N = 2304)	Program end (N = 1549)	6 months post (N = 1161)
Sociodemographic
Sex
Male	20.1	18.9	18.5
Female	78.3	79.5	80.1
Missing	1.6	1.6	1.4
Age
45 years or younger	32.2	28.0	24.8
45–64 years	49.8	52.4	56.1
65 years or older	17.7	19.3	18.9
Missing	0.3	0.3	0.3
Marital status
Married	62.9	62.3	64.3
Single	33.3	34.0	32.0
Missing	3.7	3.7	3.7
Birth country
US	20.2	19.0	18.2
Mexico	76.6	78.1	79.4
Missing	3.2	2.8	2.4
Employment
Employed & self-employed	30.6	29.7	29.0
Unemployed & unable to work	21.8	20.3	18.8
Homemaker	35.2	36.9	38.5
Retired	11.4	12.3	12.7
Missing	1.0	0.8	1.0
Education
Middle school or less	46.0	48.5	47.0
High school grad or some	37.2	35.5	36.4
College (3 or 4 years)	13.8	13.2	13.7
Missing	3.0	2.8	2.8
Income
Less than $10,000	42.0	43.2	41.9
$10,000 to less than $20,000	31.2	31.7	32.0
$20,000 to less than $30,000	13.7	14.2	15.2
$30,000 and more	6.8	6.1	6.5
Missing	6.4	4.8	4.4
Insurance
Yes	40.4	41.8	44.3
No	53.9	53.8	52.7
Missing	5.6	4.3	3.0
Center Characteristics
Overall program quality^*
Low quality	18.1	15.6	11.6
High quality	81.9	84.4	88.4
Group
Individual	28.9	28.0	28.3
Group program	71.1	72.0	71.7
Program duration
5 weeks	7.4	7.3	9.0
8 weeks	18.4	16.3	13.0
9 weeks	17.9	18.9	19.0
10 weeks	26.2	28.0	28.5
12 weeks or longer	29.9	29.5	30.4
Curriculum
Adapted	69.4	66.0	59.9
Developed	30.6	34.0	40.1

Total N is the number of participants who had both survey and anthropometric measurements at that time and previous time point(s).

Overall program quality measurements were made by the evaluation team based on assessments of the following factors: percentage of participants who completed the program and surveys at program end and 6 months later, level of organization, and quality of interaction with the teams at each community health center.

We utilized descriptive analysis to explore the success rates for weight reduction and glycemic outcomes. Chi-square tests were used to study the association of each predictor and outcome pair. Logistic regression models were constructed to identify predictors of success and study their effects on outcomes. We used Nagelkerke R ² as an approximate measure of effect size. The significance levels 0.05, 0.01, and 0.001, were used.

Results

Overall, the AHB program produced an 18.0% (95% confidence interval [CI]: 14.9%–21.1%) weight reduction success rate at program end, which increased to 27.6% (95%CI: 23.3%–30.7%) at 6 months post program. There was a 48.8% (95%CI: 43.8%–52.2%) glycemic success rate at program end, which decreased to 47.3% (95%CI: 42.6%–51.4%) at 6 months post program. Table 3 summarizes distributions of the 23 factors and their respective association with weight and glycemic outcomes. For sociodemographic predictors, participants more likely to be successful were younger, employed, or self-employed, high school graduates or those with some high school education, and/or with incomes of $30,000 or more. For baseline health factors, successful participants reported good, very good, or excellent health; no diabetes diagnosis or family history of diabetes; normal BMI; normal HbA1c level; no limiting physical, mental, or emotional problems; no hypertension; and/or no high cholesterol. For center and program characteristics predictors, success rates were higher among participants enrolled in good quality programs, and/or programs that lasted 9 to 10 weeks. Finally, participants who improved their fruit and vegetable intake and physical activity levels achieved higher rates of success.

Table 3.

Rates of Weight Reduction and Glycemic Success by 23 Factors at Program End and 6 Months Later

		Weight Reduction Success		Glycemic Success
	Percent	Program end	6 months post	Program end	6 months post
Overall Success Rates		18.0	27.6	48.8	47.3
Sociodemographic
Sex
Male	19.7	23.7	33.0	42.5	46.2
Female	80.3	16.6	26.4	50.3	47.6
Age			P = 0.047	P = 0.000	P = 0.000
45 years or younger	23.7	23.2	36.0	68.8	72.2
45–64 years	58.1	15.8	25.8	48.1	42.7
65 years or older	18.2	18.3	22.7	26.0	33.0
Marital status
Married	68.4	18.0	27.3	48.5	47.1
Single	31.6	18.0	28.2	49.4	47.8
Birth country
US	19.2	16.5	22.1	44.8	38.5
Mexico	80.8	18.4	28.9	49.8	49.4
Employment			P = 0.007	P = 0.000
Employed & self-employed	30.4	19.2	36.4	64.6	52.7
Unemployed & unable to work	18.2	19.3	21.6	39.6	42.3
Homemaker	39.6	16.9	21.8	46.3	50.0
Retired	11.9	16.9	31.9	31.3	34.8
Education				P = 0.000
Middle school or less	49.2	16.3	25.8	40.8	42.3
High school grad or some	35.1	18.6	27.5	59.5	53.7
College (3 or 4 years)	15.7	22.3	33.3	51.8	50.0
Income				P = 0.004
Less than $10,000	42.9	17.5	27.2	49.0	43.3
$10,000 to less than $20,000	33.1	17.2	23.3	42.9	47.2
$20,000 to less than $30,000	16.9	16.8	30.2	48.9	48.8
$30,000 and more	7.2	27.9	42.9	75.7	67.6
Insurance
Yes	43.7	16.8	29.0	47.3	45.9
No	56.3	19.0	26.4	50.0	48.6
Baseline Health Factors
Self-reported health condition			P = 0.013	P = 0.000	P = 0.031
Good, very good, and excellent health	37.7	21.2	33.9	59.6	52.0
Fair	50.8	16.1	24.4	45.1	47.0
Poor	11.5	15.9	6.7	31.3	32.1
Diabetes		P = 0.040	P = 0.015	P = 0.000	P = 0.000
No	58.8	20.7	31.7	66.0	63.6
Yes	41.2	14.2	22.2	26.8	28.6
Body mass index		P = 0.000	P = 0.000	P = 0.037
Normal	8.2	73.5	68.0	63.0	63.0
Overweight	31.9	11.5	17.5	52.2	44.4
Obese	59.9	13.9	26.5	44.9	46.4
Baseline HbA1c		P = 0.006	P = 0.000	P = 0.000	P = 0.000
HbA1c <6	40.7	23.5	38.8	81.1	77.8
HbA1c ≥6	59.3	14.5	20.5	26.6	27.2
Smoking
Nonsmoker	70.8	17.0	27.3	51.2	47.3
Former smoker	19.0	24.6	30.4	40.0	44.2
Current smoker	10.2	13.1	24.1	49.1	54.0
Family history of diabetes			P = 0.003	P = 0.003	P = 0.002
With family history	62.9	17.0	25.9	44.3	42.0
Without family history	31.2	21.9	32.5	59.6	59.7
Unknown	5.8	8.6	20.6	41.2	43.8
Limited by problems			P = 0.045	P = 0.001	P = 0.029
Yes	24.5	17.7	21.0	36.3	39.1
No	75.5	18.1	29.9	52.9	50.3
Hypertension				P = 0.000	P = 0.003
No	41.4	19.9	28.3	61.0	55.6
Yes	58.9	16.7	27.1	40.4	41.7
High cholesterol					P = 0.046
No	34.1	18.1	26.1	54.4	53.6
Yes	65.9	18.0	28.4	46.0	44.1
Center Characteristics
Overall program quality			P = 0.045		P = 0.011
High quality	11.4	14.7	16.7	41.5	31.6
Low quality	88.6	18.5	29.0	49.8	49.4
Group
Individual	27.4	22.0	22.3	44.8	44.6
Group program	72.6	16.6	29.6	50.4	48.4
Program duration				P = 0.001
5 weeks	4.2	4.0	17.9	28.0	64.3
8 weeks	11.4	11.8	20.4	32.8	43.6
9 weeks	18.7	14.3	28.6	60.7	46.0
10 weeks	25.4	20.4	33.1	54.1	49.2
12 weeks or longer	40.4	21.5	26.7	47.5	45.2
Curriculum			P = 0.044	P = 0.040
Adapted	54.4	18.7	31.0	52.9	50.4
Developed	45.6	17.2	76.8	44.1	43.6
Dietary and Physical Activity Factors
Improved dietary habits
No	53.6	15.3	24.0	49.3	45.7
Yes	46.4	21.2	31.0	48.2	49.0
Improved fruit and vegetable intake			P = 0.049
No	62.1	18.3	24.4	49.3	46.8
Yes	37.9	17.6	32.0	48.1	48.0
Improved physical activities		P = 0.047	P = 0.025
No	62.1	15.6	24.4	46.3	48.4
Yes	37.9	22.0	33.5	52.9	45.2

P values from chi-square tests of association are provided for those factors significantly (at 0.05 significance level) associated with the corresponding dependent variables.

In the next section, we further identify significant predictors of success and study their effects by using all 23 independent factors in a logistic regression model. Because baseline HbA1c levels are closely related to diagnosis of diabetes, we excluded baseline HbA1c from the logistic regression, while retaining all of the other 23 factors.

Effects of predictors on weight reduction success

The results from logistic regressions for weight reduction and glycemic success are presented in Table 4. At program end, the effect size for the logistic model with 23 hypothesized factors was 0.489, indicating that the model explained 48.9% of the variation of weight reduction success. The hypothesized factors were found to jointly predict weight reduction success (likelihood ratio [LR] χ2(36) = 198.80, P < 0.001), and the model was able to correctly predict 89.6% of the weight reduction successes and failures. When holding other factors constant, we found female participants were less likely to succeed in weight reduction than males (odds ratio [OR] = 0.33). Participants 45–64 years of age were less likely to achieve weight reduction success (OR = 0.38) compared to those who were 45 years of age or younger. When holding other factors constant, retired participants were less likely to achieve weight reduction success (OR = 0.17) than employed or self-employed participants.

Table 4.

Results from Logistic Regression for Both Weight Reduction and HbA1c Outcomes at 2 Time Points: at Program end and 6 Months Later

	Weight Reduction Success		Glycemic Success
	Program end	6 months post	Program end	6 months post
Sociodemographic
Sex (ref: Male)
Female	0.33^**	0.68	1.34	0.27^**
	CI: 0.15–0.72			CI: 0.12–0.58
	Wald:7.68			Wald: 11.56
Age (ref: 45 years or younger)
45–64 years	0.38^*	0.56^*	0.60	0.14^***
	CI: 0.18–0.80	CI: 0.32–0.98		CI: 0.07–0.29
	Wald: 6.49	Wald: 4.19		Wald: 28.58
65 years or older	2.54	0.32^*	0.17^***	0.06^***
		CI: 0.12–0.86	CI: 0.07–0.45	CI: 0.02–0.18
		Wald: 5.06	Wald: 13.21	Wald: 24.95
Marital status (ref: Married)
Single	1.38	1.07	1.52	1.16
Birth country (ref: US)
Mexico	1.80	1.54	0.91	1.97
Employment (ref: Employed & self-employed)
Unemployed & unable to Work	1.56	0.60	0.48^*	1.87
			CI: 0.29–0.93
			Wald: 4.79
Homemaker	1.27	0.49^*	0.43^**	1.34
		CI: 0.27–0.87	CI: 0.24–0.77
		Wald: 6.00	Wald: 7.93
Retired	0.17^*	1.27	0.53	1.01
	CI: 0.04–0.74
	Wald: 5.52
Education (ref: Middle school or less)
High school grad of some	0.62	0.68	1.37	1.36
College (3 or 4 years)	0.76	0.69	0.66	0.81
Income (ref: less than $10,000)
$10,000 to less than $20,000	1.10	0.68	0.60	1.37
$20,000 to less than $30,000	1.12	0.98	0.62	0.72
$30,000 and more	9.21^**	1.87	1.87	0.73
	CI: 2.50–33.99
	Wald: 11.11
Insurance (ref: Yes)
No	2.30^*	0.97	0.67	1.21
	CI: 1.02–5.16
	Wald: 4.05
Baseline Health Factors
Self-reported health condition (ref: good, very good and excellent health)
Fair	0.89	0.80	0.98	1.70
Poor	1.26	0.48	0.98	1.32
Diabetes (ref: No)
Yes	0.41^*	0.73	0.15^***	0.09^***
	CI: 0.20–0.84		CI: 0.09–0.26	CI: 0.05–0.17
	Wald: 5.85		Wald: 51.71	Wald: 57.58
Body mass index (ref: Normal)
Overweight	0.01^***	0.07^***	0.66	0.20^**
	CI: 0.00–0.02	CI: 0.03–0.15		CI: 0.07–0.57
	Wald: 69.14	Wald:41.35		Wald: 9.15
Obese	0.01^***	0.12^***	0.41^*	0.34^*
	CI: 0.00–0.20	CI: 0.05–0.25	CI: 0.18–0.94	CI: 0.13–0.92
	Wald: 74.27		Wald: 4.41	Wald: 4.52
Smoking (ref: nonsmoker)
Former smoker	2.95^**	1.04	0.59	1.13
	CI: 1.39–6.28
	Wald: 7.87
Current smoker	0.83	0.80	0.70	2.18
Family history of diabetes (ref: with family history)
Without Family History	0.69	1.03	1.17	1.99^*
				CI: 1.14–3.47
				Wald: 5.89
Unknown	0.10^*	0.76	0.58	0.41
	CI: 0.01–0.88
	Wald: 4.34
Limited by problems (ref: Yes)
No	0.65	0.97	1.27	1.33
Hypertension (ref: No)
Yes	1.08	1.36	0.70	0.83
High cholesterol (ref: No)
Yes	0.85	1.33	0.90	0.96
Center Characteristics
Overall program quality (ref: Low quality)
High quality	0.66	1.40	1.19	1.86
Session type (ref: Individual session)
Group session	0.24^**	1.34	1.63	1.62
	CI: 0.10–0.59
	Wald: 9.68
Program duration (ref: 8 weeks)
5 weeks	0.53	0.58	0.40	5.80
9 weeks	2.54	0.80	4.07^*	0.37
			CI: 1.36–12.20
			Wald: 6.285
10 weeks	2.31	0.74	1.42	0.40
12 weeks or longer	10.69^**	1.58	2.23	0.44
	CI: 2.02–56.64
	Wald: 7.75
Curriculum (ref: Adapted)
Developed	0.46	0.54	0.65	0.69
Dietary and Physical Activity Factors
Improved dietary habits (ref: No)
Yes	1.85	1.31	0.76	1.03
Improved fruit-vegetable intake (ref: No)
Yes	1.18	1.30	0.85	0.65
Improved physical activities (ref: No)
Yes	4.19^***	1.68^*	1.91^**	0.78
	CI: 2.16–8.12	CI: 1.05–2.71	CI: 1.18–3.08
	Wald: 17.92	Wald: 4.58	Wald: 7.02

CI, confidence interval; ^***significant at 0.001; ^**significant at 0.01; ^*significant at 0.05.

Participants who reported income levels of $30,000 or more were more likely to achieve weight reduction success than those with less than $10,000 in income (OR = 9.21). Participants without health insurance coverage were more likely to achieve weight reduction success than participants with health insurance coverage (OR = 2.30). Participants diagnosed with diabetes were less likely to achieve weight reduction success than those not diagnosed with this chronic health condition (OR = 0.41). Both overweight and obese participants were less likely to succeed in weight reduction than normal weight participants (OR = 0.01). Former smokers were more likely to have weight reduction success than nonsmokers (OR = 2.95). Participants with no known family history of diabetes were less likely to achieve weight reduction success than those who reported family history of diabetes (OR = 0.10). Controlling other factors in the model, respondents who participated in group-based programs were less likely to succeed in weight reduction than those who participated in individual-based programs (OR = 0.24). Participants in programs that lasted 12 weeks were more likely to succeed than those in programs that lasted 8 weeks (OR = 10.69). Participants with improved physical activity levels were more likely to succeed than those without improved physical activities (OR = 4.19).

Six months after the program end, the effect size for the logistic model was 0.261 for weight reduction success. The hypothesized factors were found to jointly predict weight reduction success (LR χ2(36) = 107.09, P < 0.001) and yielded a correct prediction rate of 77.4%. Controlling for other factors in the model, participants 45–64 years of age were less likely to achieve weight reduction success relative to those who were 45 years of age or younger (OR = 0.56); participants who were 65 years of age and older were less likely to achieve weight reduction success relative to those who were 45 years of age or younger (OR = 0.32). Homemakers were less likely to achieve weight reduction success than those who were employed or self-employed (OR = 0.49). Overweight (OR = 0.07) and obese participants (OR = 0.12) were less likely to succeed in weight reduction compared to normal weight participants. Participants with improved physical activity levels were more likely to succeed than those without improved physical activity levels (OR = 1.68), when controlling other factors in the model.

Effects of predictors on glycemic success

At the end of the program, the effect size for the logistic model was 0.437 for glycemic success. The hypothesized factors were found to jointly predict glycemic success (LR χ2(36) = 211.70, P < 0.001), and to correctly predict 73.2% of the glycemic successes and failures. Participants who were 65 years of age and older were less likely to achieve glycemic success than those 45 years of age or younger (OR = 0.17) when other factors were held constant. Compared to employed or self-employed participants, homemakers and those who were unemployed or unable to work were less likely to achieve glycemic success (OR = 0.48 and 0.43, respectively). Participants diagnosed with diabetes were less likely to achieve glycemic success than those who were not diagnosed with the disease (OR = 0.15). Obese participants were less likely to succeed in glycemic reduction (OR = 0.41) than normal weight participants. Compared to participants in programs lasting 8 weeks, participants in programs lasting 9 weeks were more likely to achieve glycemic reduction (OR = 4.07). Participants with improved physical activity levels were more likely to succeed in glycemic reduction than those without improved physical activity levels (OR = 1.91) when controlling other factors in the model.

Six months after the programs ended, the effect size for the logistic model was 0.476, for glycemic success. The hypothesized factors were found to jointly predict glycemic success (LR χ2(36) = 202.52, P < 0.001) with a correct prediction rate of 75.2%. When holding other factors constant, female participants were less likely to succeed in glycemic reduction than males (OR = 0.27). Compared to participants who were 45 years of age or younger, those 45–64 years of age (OR = 0.14) and those who were 65 years of age and older (OR = 0.06) were less likely to achieve glycemic success. Participants diagnosed with diabetes were less likely to achieve glycemic success than those without a diabetes diagnosis (OR = 0.09). Overweight (OR = 0.20) and obese (OR = 0.34) participants were less likely to succeed in glycemic reduction than normal weight participants. Holding other factors constant, participants who reported no family history of diabetes were more likely to achieve glycemic success than those with a family history of diabetes (OR = 1.99).

Discussion

This study investigated the factors of success in reducing weight and glycemic levels at 2 time points: end of program and post 6 months. Our findings about sociodemographic factor effects on the outcomes of interventions for diabetes and obesity are consistent with the results in the existing literature.^24

–27 For example, older participants were less likely to achieve weight reduction success,²⁵ and men were more likely to achieve weight reduction goals.²⁵ AHB participants who were employed or self-employed, and those who earned $30,000 or more, were more likely to succeed than their counterpart groups. These results are consistent with extant research^25,26,27 that people with low socioeconomic status are a disadvantaged population who need more attention in terms of health education and prevention.

Baseline health factors may affect the effectiveness of diabetes prevention interventions.^{23,28,29

–34} We found that participants with better baseline health conditions, (ie, participants without diabetes, normal BMI, former smokers, without a family history of diabetes) were more likely to achieve success after the interventions. Some of our findings are similar to those reported in the literature (eg, participants with lower baseline BMI were more likely to achieve the weight loss goal,²⁵ overweight individuals were less likely to decrease their HbA1c levels²³), but some are different from findings in other studies (eg, more obese individuals lost more weight³⁵). These divergent results may arise from differences in the populations studied as well as from differences in the sets of predictors employed. Our findings are consistent with studies in the literature,^36,37,38 indicating that the at-risk Hispanic population in border communities would benefit from preventive programs and self-management diabetes education.

One important goal of the AHB program was to identify feasible and efficient preventive practices for the border Hispanic population. We used 4 program and center characteristics to capture the variation in intervention features across participating community health centers. Two of the 4 features, overall program quality and curriculum, affected participant retention rates remarkably. Programs classified as high-quality interventions, as well as programs that rely on internally developed curriculum to meet the needs of the corresponding population, had relatively high retention rates. Furthermore, 2 of those 4 program characteristics—session type and program duration—were associated with success at the 2 time points. Individual programs worked better than group programs, based on weight reduction success at the end of a program. This finding is consistent with the Diabetes Prevention Program (DPP) intensive lifestyle intervention²⁵ and the Finnish diabetes prevention study,²⁰ indicating that working with participants individually is an effective way to deliver diabetes education. Compared to programs that lasted 8 weeks, program duration effects were significant for programs that lasted 9 weeks (for glycemic success) and 12 weeks (for weight reduction). Programs that lasted 9 or 10 weeks yielded either the highest or the second highest success rates (Table 3), except for the glycemic success rate at post 6 months, for which the 5-week program yielded the highest success rate. Shorter programs (5-week duration) were less likely to show glycemic success at program end, but more likely to result in significant HbA1c reductions at 6 months post program end. This apparent inconsistency may be because a period of 5 weeks is usually insufficient to reveal any significant changes in HBA1c levels, which usually reflect blood glucose levels over the preceding 2 to 3 months (thus, the low success rates at the end of short-duration programs).

The last group of predictors in our study dealt with dietary habits and physical activity. Studies in the literature show that dietary habits and physical activity are the core modifiable factors that are essentially the pathway to some other modifiable factors such as body weight and HbA1c level.^{18,20,39
–41} The AHB program measured dietary habits and physical activity at baseline, end of program, and at 6 months post program. Thus, we included the dietary habits and physical activity improvements between program start and the evaluation points in our analysis. Compared to their stationary measurements at the evaluation time point, the dynamic improvements were better mediator factors contributing to weight and HbA1c levels reductions. Our results showed that, within a relatively short period, physical activity improvement had a significant effect on weight reduction and glycemic control, whereas improvements in dietary habits and fruit and vegetable intake did not.

Compared to other prevention programs in the literature, the AHB interventions performed well in terms of glycemic control. For example, AHB produced a 2.0 ± 11.7% decrease in initial glycemic levels for overweight/obese participants at program end, compared to a 0.1 ± 0.7% decrease in initial HbA1c at the end of year 1 in the Finnish diabetes prevention study.²⁰ In addition, AHB interventions helped 27% of participants with abnormal baseline HbA1c achieve success in terms of HbA1c level reduction at both program end and post 6-month follow-up. On the other hand, for overweight/obese participants, AHB interventions resulted in a 13.2% weight reduction success rate and a 1.9 ± 3.6% reduction of initial body weight at program end, compared to the 49% weight loss success rate and 6.9 ± 4.5% reduction of initial body weight at the end of core curriculum in the DPP.²⁸ We suspect that the success in reducing BMI might require longer time (as in the DPP and the Finnish diabetes prevention study) than the intervention period in our study.

There are some important limitations to our study. First, our findings were based on complete data after cases with missing values were eliminated. Another limitation is that there was no direct control comparison group for the AHB intervention. In addition, the follow-up period of 6 months may be relatively short in terms of being able to detect weight reduction/gain. One future work direction is to evaluate the AHB program success in terms of the dietary and physical activities that are commonly considered to be modifiable factors for preventive practices. In this study we used them as predictors for the 2 intermediate modifiable factors: weight and HbA1c level.

In conclusion, AHB interventions successfully improved weight reduction and glycemic outcomes among participants. Further, we conclude that AHB interventions were more successful in reducing participants' HbA1c level than BMI within a relatively short period. Our results suggest that future preventive practices and programs can be developed and tailored for various target groups according to the aforementioned factors. Improved, culturally appropriate preventive practices and self-management diabetes education along with timely access to health care can reduce or delay the onset of diabetes and its complications. These programs are particularly needed in predominantly Hispanic communities characterized by poor access to health care and to health prevention programs, such as those communities located along the US-Mexico border.

Footnotes

Acknowledgments

The authors thank Donna Jackson, BSc, and Violeta Diaz Avilez, PhD, for their work on program coordination and assistance in data management and research.

Author Disclosure Statement

Drs. Wang, Ghaddar, Brown, and Pagán, and Ms. Balboa disclosed no conflicts of interest. This study was funded by a competitive grant from South Texas Border Health Disparity Center, the University of Texas-Pan American, 2009–2010.

References

World Health Organization. Diabetes. http://www.who.int/mediacentre/factsheets/fs312/en/. 2010 July 23.

Centers for Disease Control and Prevention. National Diabetes Fact Sheet: General Information and National Estimates on Diabetes in the United States, 2007. Atlanta, GA: U.S. Department of Health and Human Services, Centers for Disease Control and Prevention, 2008.

Centers for Disease Control and Prevention. Prevalence of diabetes among Hispanics—Selected areas, 1998–2002. MMWR Morb Mortal Wkly Rep, 2004; 53:941–944.

Office of Minority Health and Health Disparities. Hispanic Health Program—National Diabetes Prevention and Control Program. http://www.cdc.gov/omhd/populations/HL/HHP/Diabetes.htm. 2010 July 23.

Lavery

, Armstrong

, Wunderlich

, Tredwell

, Boulton

. Diabetic foot syndrome: Evaluating the prevalence and incidence of foot pathology in Mexican Americans and non-Hispanic whites from a diabetes disease management cohort. Diabetes Care, 2003; 26:1435–1438.

Centers for Disease Control and Prevention. U.S.–Mexico Border Prevention and Control Project. http://www.cdc.gov/diabetes/projects/border.htm. 2010 July 22.

Centers for Disease Control and Prevention. Self-reported prevalence of diabetes among Hispanics—United States, 1994–1997. MMWR Morb Mortal Wkly Rep, 1999; 48:8–12.

Giltinan

, Grunwald

, Hill

, Jatkauskas

, King

, Sharp

. Association of anthropometric measures with risk of diabetes and cardiovascular disease in Hispanic and Caucasian adolescents. Prev Med, 2003; 37:611–616.

Melnik

, Hosler

, Sekhobo

, Duffy

, Tierney

, Engelgau

, Geiss

. Diabetes prevalence among Puerto Rican adults in New York City, NY, 2000. Am J Public Health, 2004; 94:434–437.

10.

Burke

, Williams

, Narayan

, Leibson

, Haffner

, Stern

. A population perspective on diabetes prevention: Whom should we target for preventing weight gain? Diabetes Care, 2003; 26:1999–2004.

11.

Centers for Disease Control and Prevention. Diabetes among Hispanics—Los Angeles County, California, 2002–2003. MMWR Morb Mortal Wkly Rep, 2003; 52:1152–1155.

12.

Gary

, Narayan

, Gregg

, Beckles

, Saaddine

. Racial/ethnic differences in the healthcare experience (coverage, utilization, and satisfaction) of US adults with diabetes. Ethn Dis, 2003; 13:47–54.

13.

O'Rahilly

, Barroso

, Wareham

. Genetic factors in type 2 diabetes: The end of the beginning? Science, 2005; 307:370–373.

14.

McCarthy

, Froguel

. Genetic approaches to the molecular understanding of type 2 diabetes. Am J Physiol Endocrinol Metab, 2002; 283:E217–E225.

15.

Kaprio

, Tuomilehto

, Koskenvuo

, Romanov

, Reuanen

, Eriksson

, Stengard

, Kesaniemi

. Concordance for type 1 (insulin-dependent) and type 2 (non-insulin-dependent) diabetes mellitus in a population-based cohort of twins in Finland. Diabetologia, 1992; 35:1060–1067.

16.

Pan

, Li

, Hu

, Wang

, Yang

, An

, Hu

, Lin

, Xiao

, Cao

, Liu

, Jiang

, Wang

, Zheng

, Shang

, Bennett

, Howard

. Effects of diet and exercise in preventing NIDDM in people with impaired glucose tolerance. The Da Qing IGT and diabetes study. Diabetes Care, 1997; 20:537–544.

17.

Tuomilehto

, Lindström

, Eriksson

, Valle

, hamalainen

, Ilanne-Pankka

, Keinanen-Kiukaanniemi

, Laakso

, Louheranta

, Rastas

, Salminen

, Uusitupa

. Finnish Diabetes Prevention Study Group. Prevention of type 2 diabetes mellitus by changes in lifestyle among subjects with impaired glucose tolerance. N Engl J Med, 2001; 344:1343–1350.

18.

Eriksson

, Lindgarde

. Prevention of type 2 (non-insulin dependent) diabetes mellitus by diet and physical exercise. The 6-year Malmö feasibility study. Diabetologia, 1991; 34:891–898.

19.

Knowler

, Barrett-Connor

, Fowler

, Hamman

, Lachin

, Walker

, Nathan

. Diabetes Prevention Program Research Group. Reduction in the incidence of type 2 diabetes with lifestyle intervention or metformin. N Engl J Med, 2002; 346:393–403.

20.

Lindström

, Louheranta

, Mannelin

, rastas

, Salminen

, Eriksson

, Toumilehto

. The Finnish Diabetes Prevention Study (DPS): Lifestyle intervention and 3-year results on diet and physical activity. Diabetes Care, 2003; 26:3230–3236.

21.

Valsamakis

, Chetty

, Anwar

, Banerjee

, Barnett

, Kumar

. Association of simple anthropometric measures of obesity with visceral fat, and the metabolic syndrome in male Caucasian and Indo-Asian subjects. Diabet Med., 2004; 21:1339–1345.

22.

Teixeira

, Going

, Houtkooper

, Cussler

, Metcalfe

, Blew

, Sardinha

, Lohman

. Pretreatment predictors of attrition and successful weight management in women. Int J Obes Relat Metab Disord, 2004; 28:1124–1133.

23.

Sarkadi

, Rosenqvist

. Field test of a group education program for type 2 diabetes: Measures and predictors of success on individual and group levels. Patient Educ Couns, 2001; 44:129–139.

24.

Van Wye

, Kerker

, Matte

, Chamany

, Eisenhower

, Frieden

, Thorpe

. Obesity and diabetes in New York City, 2002 and 2004. Prev Chronic Dis, 2008; 5:A48.

25.

Wing

, Hamman

, Bray

, Delahanty

, Edelstein

, Hill

, Horton

, Hoskin

, Kriska

, Lachin

, Mayer-Davis

, Pi-Sunyer

, Regensteiner

, Venditti

, Whylie-Rosett

. Diabetes Prevention Program Research Group. Achieving weight and activity goals among diabetes prevention program lifestyle participants. Obes Res, 2004; 12:1426–1434.

26.

Evans

, Newton

, Ruta

, MacDonald

, Morris

. Socio-economic status, obesity, and prevalence of type 1 and type 2 diabetes mellitus. Diabet Med, 2000; 17:478–480.

27.

Maty

, Everson-Rose

, Haan

, Raghunathan

, Kaplan

. Education, income, occupation, and the 34-year incidence (1965–99) of type 2 diabetes in the Alameda County Study. Int J Epidemiol, 2005; 34:1274–1281.

28.

Bowie

, Cho

, Juon

, Rodriguez

. Factors associated with overweight and obesity among Mexican Americans and Central Americans: Results from the 2001 California Health Interview Survey. Prev Chronic Dis, 2007; 4:A10.

29.

Haire-Joshu

, Glasgow

, Tibbs

. Smoking and diabetes. Diabetes Care, 1999; 22:1887–1898.

30.

Lewis

, MacLeod

, McKnight

, Morris

, Peden

, Prescott

, Walker

. Royal College of Physicians of Edinburgh Diabetes Register Group. Predicting vascular risk in type 1 diabetes: Stratification in a hospital based population in Scotland. Diabet Med, 2005; 22:164–171.

31.

Ruggiero

, Rossi

, Prochaska

, Glasgow

, de Groot

, Dryfoos

, Reed

, Orleans

, Prokhorov

, Kelly

. Smoking and diabetes: Readiness for change and providers advice. Addict Behav, 1999; 24:573–578.

32.

Wallerstien

. Population genetics in minority children with type 2 diabetes mellitus. J Pediatr Endoctinol Metab, 2002; 15:485–486.

33.

Will

, Galuska

, Ford

, Mokdad

, Calle

. Cigarette smoking and diabetes mellitus: Evidence of a positive association from a large prospective cohort study. Int J Epidemiol, 2001; 30:540–546.

34.

Zhao

, Ford

, Li

, Mokdad

. Compliance with physical activity recommendations in US adults with diabetes. Diabet Med, 2008; 25:221–227.

35.

Wing

, Phelan

. Strategies to improve outcomes and predictors of success. Eckel

. Obesity: Mechanisms and Clinical Management. Philadephia, PA: Lippincott, Williams, and Wilkins, 2003; 415–435.

36.

Andres

, Olson

, Wiebe

, Bean

, DiGregorio

, Guillermina

, Ortiz

. Diabetes prevalence and treatment adherence in residents living in a colonia located on the West Texas, USA/Mexico border. Nurs Health Sci, 2008; 10:195–202.

37.

Oladele

, Barnett

. Racial/ethnic and social class differences in preventive care practices among persons with diabetes. BMC Public Health, 2006; 6:259.

38.

Brown

, Garcia

, Kouzekanani

, Hanis

. Culturally competent diabetes self-management education for Mexican Americans: The Starr County Border Health Initiative. Diabetes Care, 2002; 25:259–268.

39.

Miller

3rd , Erlinger

, Young

, Jehn

, Charleston

, Rhodes

, Wasan

, Appel

. Results of the Diet, Exercise, and Weight Loss Intervention Trial (DEW-IT) Hypertension, 2002; 40:612–618.

40.

Daubenmier

, Weidner

, Sumner

, Mendell

, Merritt-Worden

, Studley

, Ornish

. The contribution of changes in diet, exercise, and stress management to changes in coronary risk in women and men in the multisite cardiac lifestyle intervention program. Ann Behav Med, 2007; 33:57–68.

41.

Cramer

, Sibley

, Bartlett

, Kahn

, Loffredo

. An adaptation of the diabetes prevention program for use with high-risk, minority patients with type 2 diabetes. Diabetes Educ, 2007; 33:503–508.