Risk Factors and Wellness Measures Associated with Prediabetes and Newly Diagnosed Type 2 Diabetes Mellitus in Hispanic Adults

Abstract

Background:

To characterize the associations of clinical risk factors, lifestyle factors, and wellness measures with prediabetes and new type 2 diabetes mellitus (T2DM) diagnosis in Hispanic adults and guide primary prevention.

Methods:

Sangre Por Salud Biobank enrolled 3733 Hispanic adults from Phoenix, AZ, United States, from 2013 to 2018. This analysis included participants with euglycemia, prediabetes, or new T2DM diagnosis (i.e., no prior T2DM diagnosis) at enrollment. Participants completed a baseline questionnaire on cardiometabolic risk factors and wellness measures and provided biometric measurements. The associations of factors and measures with odds (95% confidence interval) of prediabetes and new T2DM diagnosis were analyzed in logistic regression models.

Results:

Among 3299 participants with euglycemia (n = 1301), prediabetes (n = 1718), and new T2DM diagnosis (n = 280) at enrollment, 72% were women (n = 2376/3299). In adjusted models, most cardiometabolic risk factors were positively associated with prediabetes and new T2DM diagnosis, with stronger associations for new T2DM diagnosis. Obesity (body mass index ≥30 kg/m² vs. lower) was associated with higher odds of new T2DM diagnosis (3.14 [2.30–4.28]; P < 0.01) than prediabetes versus euglycemia (1.96 [1.66–2.32]; P < 0.01) and Interaction (P = 0.01). Similarly, waist circumference, family history of diabetes, and average systolic and diastolic blood pressure were associated with higher odds of new T2DM diagnosis versus euglycemia than prediabetes versus euglycemia. Using stepwise logistic regression modeling, a parsimonious model of age, family history of diabetes, waist circumference, diastolic blood pressure, passive tobacco exposure, and self-rated general health were associated with new T2DM diagnosis versus euglycemia.

Conclusions:

In Hispanic adults, modifiable cardiometabolic and lifestyle factors were associated with prediabetes and new T2DM diagnosis. Personalized interventions targeting these factors and measures could guide T2DM primary prevention efforts among Hispanic adults.

Introduction

In the United States in 2018, the prevalence of prediabetes and diabetes was 88 million and 34 million, respectively, with 90%–95% of diabetes cases attributed to type 2 diabetes mellitus (T2DM).¹ Of the 34 million adults with diabetes, 7 million were unaware of their diabetes status, and the proportion of newly diagnosed diabetes was higher among Hispanics (3.5%) compared to non-Hispanic whites (2.5%).¹ Describing the characteristics of Hispanic adults and risk factors for prediabetes and new T2DM diagnosis will guide the development of personalized interventions for T2DM primary prevention.

Previous studies have examined risk factors and risk scores to predict risk of T2DM.^2

–6 There are several biopsychosocial determinants of prediabetes and new T2DM diagnosis, which require characterization and quantification to guide primary prevention.⁶ Most scores were developed and/or validated in predominantly non-Hispanic cohorts. A recent systematic review identified a U.S. based study that evaluated performance of a diabetes risk score among Hispanic adults.⁵ This score is based on a noncontemporary Hispanic cohort (i.e., from the Second National Health and Nutrition Examination Survey, 1976–1980) and may not reflect the current burden of risk factors among Hispanic adults. It is unclear if the burden of risk factors has changed since the mid-late 1970s to the present time, which will have implications for screening and primary prevention.

Recently, the Northern Manhattan Study evaluated incident diabetes and showed that the association of risk factors [e.g., smoking, body mass index (BMI)] varied by race-ethnic groups.⁷ The ongoing Hispanic Community Health Study/Study of Latinos (HCHS/SOL) based in four U.S. cities examined risk factors associated with undiagnosed diabetes.⁸ In HCHS/SOL, 37% of adults had undiagnosed diabetes and new T2DM diagnosis. The odds of being undiagnosed were higher among women, overweight (vs. normal weight) adults, and those with dyslipidemia, among other factors. Adults with a family history of diabetes and those with hypertension had a lower odds of undiagnosed diabetes.⁸

To our knowledge, there are no reports on the association of diverse biopsychosocial risk factors with prediabetes. The HCHS/SOL study was based on recruitment in four U.S. cities (Chicago, Miami, New York, and San Diego), and it is not clear whether these results are applicable to Hispanic adults in other regions. In particular, the population of Arizona is 30% Hispanic; however, there is sparse information on risk factors associated with prediabetes and new T2DM diagnosis in this group.⁹ Paucity of this information may preclude the development of targeted primary prevention interventions.

To address these knowledge gaps and guide improvement in primary prevention, we examined the associations of demographic, cardiometabolic, and lifestyle factors and wellness measures with prediabetes and new T2DM diagnosis among Hispanic adults enrolled in the Sangre Por Salud (SPS) Biobank based in the metropolitan area of Phoenix, AZ.

Materials and Methods

Study cohort

The SPS Biobank was established through a partnership between the Mayo Clinic Center of Individualized Medicine and Mountain Park Health Center (MPHC), Phoenix, AZ, as described previously.¹⁰ Briefly, SPS Biobank enrolled 3733 self-identified Hispanic adults (ages 18–85 years), who completed a baseline lifestyle questionnaire and provided biometric measurements. In addition, participants provided a fasting blood sample processed and stored at −80°C, as well as access to their MPHC medical records. Participants were enrolled over a 5-year period from 2013 to 2018.

We divided the population into three groups, euglycemia, prediabetes, and new T2DM diagnosis, based on glycemic status at time of enrollment. Participants with new T2DM diagnosis did not have T2DM diagnosis before enrollment and received a “new T2DM diagnosis” at time of SPS enrollment. For this analysis, we excluded participants with prevalent T2DM (n = 434) at time of enrollment, to yield a cohort of 3299 participants.¹¹

This study followed the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guideline for cohort studies.

Biometric measurements

At the time of enrollment, blood pressure (average of three readings) and biometric measurements (weight, height, and waist circumference) were obtained. BMI (kg/m²) was calculated. Waist-height ratio was calculated as ratio of waist circumference to height. Waist circumference measurements were unverifiable for a subset of participants with euglycemia (n = 18 participants), prediabetes (n = 24 participants), and new T2DM diagnosis (n = 6 participants).

Biomarkers

Fasting glucose and hemoglobin A1c (HbA1C) were measured as described previously.¹⁰

Ascertainment of prediabetes and T2DM

Prediabetes was diagnosed by ≥1 confirmed laboratory criterion at time of enrollment: HbA1C 5.7%–6.4%; fasting plasma glucose 100–125 mg/dL; or 2-hr plasma glucose 140–199 mg/dL during 75 grams oral glucose tolerance test.¹¹

New T2DM was diagnosed by ≥1 confirmed laboratory criterion at time of enrollment: HbA1C ≥6.5%; fasting plasma glucose ≥126 mg/dL; or 2-hr plasma glucose ≥200 mg/dL during 75 grams oral glucose tolerance test, and no self-reported prior diagnosis of T2DM.¹¹

Euglycemia was defined by presence of all four criteria: HbA1C <5.7%, fasting plasma glucose <100 mg/dL, 2-hr plasma glucose <140 mg/dL (during 75 grams oral glucose tolerance test), and no self-reported prior diagnosis of T2DM.¹¹

Cardiometabolic risk factors, lifestyle factors, and wellness measures

Participants reported age, gender, demographics, cardiometabolic risk factors (hypertension, hyperlipidemia, blood pressure, obesity, family history of diabetes mellitus), lifestyle factors, and wellness measures (described in Supplementary Table S1). For a given risk factor or measure, ∼10% of participants did not report data; for frequency mild exercise, ∼40% of participants did not respond, likely because they responded to frequency of strenuous and moderate exercise. Obesity was determined by biometric measurements at the time of enrollment: BMI ≥30 kg/m² versus lower (nonobese); waist circumference [(≥90 cm for men or ≥80 cm for women) vs. lower (nonobese)] using cut-points for Latino populations¹⁰; and waist-height ratio ≥0.5 versus lower (nonobese). BMI was used as a measure of peripheral obesity and waist circumference and waist-height ratio as measures of central obesity.

Statistical analysis

Descriptive statistics were used to compare demographic, cardiometabolic risk factors, lifestyle factors, and wellness measures. Data are presented as median (range) for continuous variables and frequency (%) for categorical variables. Separate logistic regression models [odds ratio (95% confidence interval)] were used to compare factors in participants with prediabetes or new T2DM diagnosis and those with euglycemia (reference group). Model 1 adjusted for age and gender. Model 2 adjusted for model 1 covariates (age, gender) plus race (two categories: White or other), language spoken at home (three categories: Spanish, English, both Spanish and English), highest completed education (three categories: college or above, high school degree or below, other), and employment (full time, part time, not employed). The association of factors with glycemic status (prediabetes and new T2DM diagnosis) was compared by the probability value of the interaction term (factor × glycemic status) (P interaction). The stepwise logistic regression model (P < 0.05 to enter and stay) for prediabetes and new T2DM diagnosis incorporated model 1 risk factors that showed significant association (P < 0.05) with prediabetes and new T2DM diagnosis after adjustment for age and gender. Statistical analysis was completed using SAS^® 9.4 (SAS Institute, Inc.), and statistical significance was established at two-tailed P < 0.05.

Data availability

Data are available to researchers upon reasonable request.

Results

Among 3299 participants, we identified participants with euglycemia (n = 1301), prediabetes (n = 1718), and new T2DM diagnosis (n = 280) at time of enrollment. Compared to participants with euglycemia, participants with prediabetes and with new T2DM diagnosis were older (median [range] age: 35 [18–80] years for euglycemia vs. 44 [18–84] years for prediabetes vs. 47 [19–85] years for new T2DM diagnosis) and had fewer women (74.4% for euglycemia vs. 70.5% for prediabetes vs. 70.4% for new T2DM diagnosis). Compared to participants with euglycemia, participants with prediabetes and with new T2DM diagnosis had a higher frequency of cardiometabolic risk factors such as obesity and dyslipidemia, but generally similar lifestyle factors and wellness measures.

Cardiometabolic risk factors

Compared to participants with euglycemia, participants with prediabetes and with new T2DM diagnosis had a higher frequency of family history of diabetes, hypertension, hyperlipidemia, and obesity as defined by BMI, waist circumference, and waist-height ratio (P < 0.01, for all) (Table 1). The frequency of obesity (based on waist circumference) across all groups was >80%, including those with euglycemia.

Table 1.

Baseline Demographics, Clinical Risk Factors, Lifestyle Factors, and Wellness Measures, According to Glycemic Status at Enrollment

Characteristic	Euglycemia, n = 1301	Prediabetes, n = 1718	New T2DM diagnosis, n = 280	P
Age, years	35 (18–80)	44 (18–83)	47 (19–85)	<0.01
Women	968 (74.4)	1211 (70.5)	197 (70.4)	0.05
Race				0.01
White	1183 (90.9)	1610 (93.7)	260 (92.9)
Black or African American	3 (0.2)	0 (0)	0 (0)
Asian	0 (0)	2 (0.1)	0 (0)
Native Hawaiian or other Pacific Islander	1 (0.1)	0 (0)	0 (0)
American India or Alaska Native	9 (0.7)	1 (0.1)	0 (0)
Other	105 (8.1)	105 (6.1)	20 (7.1)
Highest education				<0.01
College or above	304 (24.7)	278 (17.2)	29 (11.0)
Below college	910 (74.0)	1320 (81.8)	230 (87.5)
Other	15 (1.2)	15 (0.9)	4 (1.5)
Employment				0.02
Full-time	418 (34.8)	564 (36.4)	76 (30.2)
Part-time	245 (20.4)	275 (17.8)	38 (15.1)
Not employed	539 (44.8)	710 (45.8)	138 (54.8)
Clinical risk factors
Family history of diabetes	497 (38.2)	828 (48.2)	187 (66.8)	<0.01
Average systolic BP, mmHg	113 (88–193)	120 (88–195)	128 (88–185)	<0.01
Average diastolic BP, mmHg	74 (52–105)	77 (55–121)	81 (46–122)	<0.01
Hypertension	132 (10.6)	276 (17.1)	71 (26.8)	<0.01
Hyperlipidemia	117 (9.4)	312 (19.3)	62 (23.3)	<0.01
Obesity (body mass index)	436 (33.8)	829 (48.7)	170 (60.7)	<0.01
Average waist circumference, cm	94 (62–148)	100 (62–152)	104 (82–152)	<0.01
Obesity (waist circumference)	1055 (82.2)	1577 (93.1)	268 (97.8)	<0.01
Obesity (waist-height ratio)	1061 (82.7)	1599 (94.4)	272 (99.3)	<0.01
Coronary artery disease	12 (1.0)	36 (2.2)	11 (4.1)	<0.01
TIA/stroke	9 (0.8)	5 (2.0)	12 (0.8)	0.15
Tobacco and alcohol
Lifetime smoking ≥100 cigarettes	208 (17.6)	343 (22.0)	57 (22.3)	0.01
Passive tobacco exposure at home and/or work	509 (40.9)	640 (39.2)	97 (37.2)	0.46
Alcohol consumption ≥2 times/month	90 (7.3)	113 (7.0)	12 (4.7)	0.31
Diet
Vegetable consumption ≥4 times/week	528 (46.7)	694 (44.9)	105 (42.7)	0.43
Fruit/fruit juice consumption ≥4 times/week	287 (25.5)	396 (25.9)	57 (23.3)	0.69
Potato consumption ≥4 times/week	177 (15.9)	252 (16.6)	28 (11.6)	0.14
Cake, sweet rolls, doughnuts, or Mexican sweet bread consumption ≥5 times/week	164 (14.7)	255 (16.8)	45 (18.7)	0.19
Flour tortillas (not corn), tacos, burritos, enchiladas, or pizza consumption ≥5 times/week	348 (30.7)	531 (34.4)	87 (35.4)	0.10
Fried chicken, refried beans, French fries, or fried potato consumption ≥5 times/week	292 (25.8)	423 (27.4)	61 (24.9)	0.53
Diet soft drink consumption ≥3 servings/day	38 (3.1)	47 (2.9)	8 (3.1)	0.96
Nondiet soft drink consumption ≥3 servings/day	105 (8.5)	124 (7.6)	29 (11.2)	0.15
Exercise
Strenuous exercise ≥5 times/week	156 (12.6)	157 (9.8)	21 (8.0)	0.02
Moderate exercise ≥5 times/week	202 (16.4)	241 (15.0)	31 (11.7)	0.15
Mild exercise ≥5 times/week	206 (26.7)	237 (25.6)	33 (23.6)	0.70
Wellness measures
General health				<0.01
Excellent or very good	389 (31.3)	374 (22.9)	45 (17.0)
Good	567 (45.6)	741 (45.4)	101 (38.3)
Fair or poor	288 (23.2)	516 (31.6)	118 (44.7)
Ability to take good care of health				0.17
A little or not confident	59 (5.2)	91 (5.9)	11 (4.5)
Somewhat confident	298 (26.3)	425 (27.5)	83 (33.6)
Completely or very confident	777 (68.5)	1031 (66.6)	153 (61.9)
Overall quality of life (rating ≤5)	127 (10.2)	177 (10.9)	36 (13.7)	0.26
Mental well-being (rating ≤5)	102 (8.2)	144 (8.9)	28 (10.6)	0.45
Physical well-being (rating ≤5)	184 (14.8)	254 (15.6)	47 (17.7)	0.47
Level of social activity (rating ≤5)	209 (16.8)	250 (15.4)	51 (19.2)	0.23
Spiritual well-being (rating ≤5)	140 (11.3)	142 (8.8)	29 (10.9)	0.07
Someone available when you need to talk				0.37
Most or all of the time	945 (76.0)	1196 (73.6)	198 (74.2)
Some of the time	188 (15.1)	289 (17.8)	42 (15.7)
None or a little of the time	111 (8.9)	140 (8.6)	27 (10.1)
Someone available to give good advice about a problem				0.12
Most or all of the time	962 (77.1)	1212 (74.5)	190 (71.4)
Some of the time	160 (12.8)	255 (15.7)	48 (18.0)
None or a little of the time	126 (10.1)	159 (9.8)	28 (10.5)
Someone available to provide emotional support				0.34
Most or all of the time	957 (76.9)	1234 (75.9)	203 (76.3)
Some of the time	153 (12.3)	236 (14.5)	40 (15.0)
None or a little of the time	134 (10.8)	156 (9.6)	23 (8.6)
Contact with someone you can trust and confide				0.79
Most or all of the time	949 (76.3)	1214 (74.8)	200 (75.2)
Some of the time	162 (13.0)	219 (13.5)	39 (14.7)
None or a little of the time	133 (10.7)	191 (11.8)	27 (10.2)

Data presented as median (range) for age, BP, waist circumference, and frequency (%) for others. Measures and survey questions are defined in Supplementary Table S1. Statistical significance based on Kruskal–Wallis test (continuous variables) and chi-squared test (categorical variables) at P < 0.05.

BP, blood pressure; T2DM, type 2 diabetes mellitus; TIA, Transient ischemic attack.

In logistic regression models adjusted for age and gender, participants with cardiometabolic risk factors had higher odds of prediabetes and new T2DM diagnosis compared to participants without these risk factors (Supplementary Table S2). These associations were generally preserved upon further adjusting model 1 covariates with race, language spoken at home, education, and employment (model 2) (Table 2). Most cardiometabolic risk factors showed stronger associations with odds of new T2DM diagnosis versus euglycemia than prediabetes versus euglycemia. For instance, obesity defined by BMI ≥30 kg/m² versus lower was associated with higher odds of new T2DM diagnosis versus euglycemia (3.14 [2.30–4.28]; P < 0.01, model 2) compared to prediabetes versus euglycemia (1.96 [1.66–2.32]; P < 0.01, model 2) (Interaction P = 0.01) (Table 2).

Table 2.

Association of Clinical Risk Factors with Odds of Prediabetes and New Type 2 Diabetes Mellitus Diagnosis, at Time of Enrollment

	Prediabetes vs. euglycemia, odds ratio (95% CI)	P	New T2DM diagnosis vs. euglycemia, odds ratio (95% CI)	P	Interaction P value
Hypertension	1.10 (0.85–1.41)	0.47	1.54 (1.04–2.27)	0.03	0.02
Coronary artery disease	1.25 (0.59–2.61)	0.56	2.27 (0.86–5.98)	0.10	0.18
Family history of diabetes	1.32 (1.12–1.56)	<0.01	2.46 (1.80–3.37)	<0.01	<0.01
Hyperlipidemia	1.56 (1.21–2.00)	<0.01	1.51 (1.02–2.25)	0.04	0.42
Transient ischemic attack/stroke	0.83 (0.31–2.24)	0.71	0.97 (0.25–3.74)	0.97	0.50
Average systolic BP (per 10 mmHg increment)	1.21 (1.13–1.29)	<0.01	1.53 (1.37–1.71)	<0.01	<0.01
Average diastolic BP (per 10 mmHg increment)	1.35 (1.23–1.50)	<0.01	1.99 (1.66–2.40)	<0.01	<0.01
Body mass index (per 5 kg/m² increment)	1.43 (1.33–1.53)	<0.01	1.14 (1.68–2.17)	<0.01	<0.01
Average waist circumference (per 5 cm increment)	1.18 (1.15–1.22)	<0.01	1.34 (1.26–1.42)	<0.01	<0.01
Obesity (body mass index)	1.96 (1.66–2.32)	<0.01	3.14 (2.30–4.28)	<0.01	0.01
Obesity (waist circumference)	2.26 (1.73–2.95)	<0.01	7.84 (3.05–20.14)	<0.01	<0.01
Obesity (waist-height ratio)	2.69 (2.03–3.58)	<0.01	16.58 (3.96–69.47)	<0.01	<0.01

Odds ratio from model 2 adjusted for model 1 variables (age, gender) plus race, language spoken at home, highest education, and employment. Measures and survey questions are defined in Supplementary Table S1. Model 1 results are shown in Supplementary Table S2.

CI, confidence interval.

Lifestyle factors

Participants with euglycemia, prediabetes, and new T2DM diagnosis reported similar lifestyle factors, including tobacco and alcohol use, food frequency, and exercise (Table 1). Across all groups, only ∼45% of participants reported vegetable consumption ≥4 times/week, whereas 25%–35% reported flour-based or fried food consumption ≥5 times/week. Across all groups, only ∼10% of participants reported strenuous exercise ≥5 times/week, and ∼25% reported mild exercise ≥5 times/week.

In logistic regression models adjusted for age and gender (model 1), participants with vegetable consumption ≥4 times/week versus lower had similar lower odds of prediabetes (0.83 [0.71–0.98]; P = 0.03, model 1) and new T2DM diagnosis (0.72 [0.53–0.98]; P = 0.04, model 1) (Interaction P = 0.63, model 1) (Supplementary Table S3). Upon further adjustment using model 2, this association was abrogated for prediabetes versus euglycemia (0.86 [0.72–1.02]; P = 0.08, model 2) but preserved for new T2DM diagnosis versus euglycemia (0.72 [0.52–1.00]; P = 0.05, model 2) (Interaction P = 0.47, model 2) (Table 3). Participants with diet soft drink consumption ≥3 servings/day versus lower showed no association with odds of prediabetes versus euglycemia, but a two-fold higher odds of new T2DM diagnosis versus euglycemia, an association preserved for new T2DM diagnosis versus euglycemia (1.99 [1.20–3.30]; P = 0.01) in model 2 (Table 3 and Supplementary Table S3). In models 1 and 2, alcohol consumption ≥2 times/month versus lower showed no association with prediabetes, but 50% lower odds of new T2DM diagnosis (0.49 [0.25–0.97]; P = 0.04) (Interaction P = 0.17, model 1). Participants with passive tobacco exposure at home/work had lower odds of prediabetes versus euglycemia but not new T2DM diagnosis versus euglycemia (model 2). Self-reported strenuous, moderate, or mild exercise ≥5 times/week versus lower was not associated with odds of prediabetes or new T2DM diagnosis (Table 3).

Table 3.

Association of Lifestyle Factors with Odds of Prediabetes and New Type 2 Diabetes Mellitus Diagnosis, at Time of Enrollment

	Prediabetes vs. euglycemia, odds ratio (95% CI)	P	New T2DM diagnosis vs. euglycemia, odds ratio (95% CI)	P	Interaction P value
Tobacco and alcohol
Lifetime smoking ≥100 cigarettes vs. lower	0.95 (0.77–1.19)	0.68	0.92 (0.62–1.35)	0.66	0.69
Passive tobacco exposure at home and/or work	0.84 (0.71–0.99)	0.04	0.74 (0.54–1.01)	0.06	0.75
Alcohol consumption ≥2 times/month vs. lower	0.88 (0.64–1.21)	0.43	0.49 (0.25–0.98)	0.04	0.17
Dietary measures
Vegetable consumption ≥4 times/week vs. lower	0.86 (0.72–1.02)	0.08	0.72 (0.52–1.00)	0.05	0.47
Fruit/fruit juice consumption ≥4 times/week vs. lower	1.05 (0.87–1.28)	0.61	0.74 (0.50–1.08)	0.11	0.14
Potato consumption ≥4 times/week vs. lower	1.01 (0.80–1.28)	0.91	0.52 (0.31–0.87)	0.01	0.04
Cake, sweet rolls, doughnuts, or Mexican sweet bread consumption ≥5 times/week vs. lower	1.12 (0.89–1.41)	0.35	1.23 (0.81–1.86)	0.32	0.63
Flour tortillas (not corn), tacos, burritos, enchiladas, or pizza consumption ≥5 times/week vs. lower	1.11 (0.93–1.33)	0.25	1.13 (0.81–1.58)	0.46	0.82
Fried chicken, refried beans, french fries, or fried potato consumption ≥5 times/week vs. lower	0.95 (0.78–1.14)	0.57	0.77 (0.54–1.11)	0.16	0.53
Diet soft drink consumption ≥3 servings/day vs. lower	1.41 (0.87–2.29)	0.16	1.60 (0.67–3.83)	0.29	0.14
Nondiet soft drink consumption ≥3 servings/day vs. lower	1.17 (0.87–1.58)	0.31	1.99 (1.20–3.30)	<0.01	0.91
Exercise
Strenuous exercise ≥5 times/week vs. lower	0.80 (0.62–1.04)	0.10	0.68 (0.40–1.14)	0.14	0.49
Moderate exercise ≥5 times/week vs. lower	1.06 (0.84–1.33)	0.63	0.80 (0.50–1.26)	0.33	0.21
Mild exercise ≥5 times/week vs. lower	0.96 (0.75–1.22)	0.72	0.67 (0.41–1.11)	0.12	0.39

Odds ratio from model 2 adjusted for model 1 variables (age, gender) plus race, language spoken at home, highest education, and employment. Risk measures and survey questions are defined in Supplementary Table S1. Model 1 results are shown in Supplementary Table S3.

Wellness measures

Participants with euglycemia, prediabetes, and new T2DM diagnosis reported similar measures of mental, physical, and spiritual well-being, level of social activity, and social supports (Table 1).

In logistic regression models, participants with self-reported “fair or poor” versus “good” general health had a higher odds of new T2DM diagnosis versus euglycemia (1.60 [1.13–2.26]; P < 0.01, model 2) compared to prediabetes versus euglycemia (1.12 [0.92–1.37]; P = 0.26, model 2) (Interaction P < 0.01, model 2) (Table 4 and Supplementary Table S4). Conversely, participants with “excellent or very good” versus “good” health had 30% lower odds of prediabetes versus euglycemia and new T2DM diagnosis versus euglycemia in models 1 and 2. Low rating of physical well-being (rating ≤5 vs. 6–10) showed no association with prediabetes versus euglycemia, but 1.5-fold higher odds of new T2DM diagnosis versus euglycemia. Other measures, including spiritual well-being and availability of emotional support, showed no association with prediabetes or new T2DM diagnosis (Table 4 and Supplementary Table S4).

Table 4.

Association of Wellness Measures with Odds of Prediabetes and New Type 2 Diabetes Mellitus Diagnosis, at Time of Enrollment

Characteristic	Prediabetes vs. euglycemia, odds ratio (95% CI)	P	New T2DM diagnosis vs. euglycemia, odds ratio (95% CI)	P	Interaction P value
General health
Excellent or very good vs. good	0.79 (0.65–0.96)	0.02	0.64 (0.42–0.97)	0.04	0.04
Fair or poor vs. good	1.12 (0.92–1.37)	0.26	1.60 (1.13–2.26)	<0.01	<0.01
Ability to take good care of health
Little or not confident vs. somewhat confident	1.16 (0.78–1.73)	0.46	0.79 (0.36–1.77)	0.57	0.36
Completely or very confident vs. somewhat confident	0.77 (0.63–0.93)	<0.01	0.65 (0.46–0.92)	0.02	0.65
Overall quality of life (rating ≤5 vs. higher)	0.97 (0.75–1.27)	0.84	1.08 (0.69–1.69)	0.75	0.43
Mental well-being (rating ≤5 vs. higher)	0.93 (0.69–1.24)	0.62	1.02 (0.62–1.68)	0.94	0.50
Physical well-being (rating ≤5 vs. higher)	1.25 (0.99–1.57)	0.06	1.52 (1.02–2.27)	0.04	0.50
Level of social activity (rating ≤5 vs. higher)	0.85 (0.68–1.06)	0.16	0.95 (0.63–1.42)	0.80	0.42
Spiritual well-being (rating ≤5 vs. higher)	0.81 (0.62–1.07)	0.14	1.21 (0.75–1.95)	0.44	0.18
Someone available when you need to talk
Most or all the time vs. some of the time	0.85 (0.68–1.06)	0.14	1.00 (0.66–1.53)	0.98	0.87
None or little of the time vs. some of the time	0.85 (0.60–1.19)	0.33	1.05 (0.56–1.97)	0.88	0.67
Someone available to give good advice about a problem
Most or all the time vs. some of the time	0.80 (0.63–1.02)	0.07	0.67 (0.44–1.00)	0.05	0.54
None or little of the time vs. some of the time	0.76 (0.54–1.06)	0.11	0.58 (0.32–1.07)	0.08	0.70
Someone available to provide emotional support
Most or all the time vs. some of the time	0.86 (0.68–1.10)	0.23	0.90 (0.58–1.39)	0.63	0.58
None or little of the time vs. some of the time	0.79 (0.57–1.11)	0.18	0.59 (0.31–1.13)	0.11	0.43
Contact with someone you can trust and confide
Most or all the time vs. some of the time	0.91 (0.72–1.17)	0.47	0.96 (0.61–1.51)	0.86	0.65
None or little of the time vs. some of the time	1.08 (0.78–1.50)	0.64	0.96 (0.51–1.81)	0.91	0.58

Risk factor model for prediabetes and new T2DM diagnosis

Using stepwise logistic regression, we identified a parsimonious model associated with prediabetes versus euglycemia and new T2DM diagnosis versus euglycemia. We identified five factors associated with prediabetes and six factors associated with new T2DM diagnosis versus euglycemia. Factors associated with both prediabetes and new T2DM diagnosis were age, waist circumference, diastolic blood pressure, and passive tobacco exposure. Hyperlipidemia was associated with prediabetes, whereas family history of diabetes and general health were associated with new T2DM diagnosis (Table 5).

Table 5.

Stepwise Logistic Regression Model for Prediabetes and New Type 2 Diabetes Mellitus Diagnosis, at Time of Enrollment

	Prediabetes vs. euglycemia, odds ratio (95% CI)	P	New T2DM diagnosis vs. euglycemia, odds ratio (95% CI)	P
Age at enrollment, per 10 year increment	1.69 (1.56–1.82)	<0.01	2.08 (1.79–2.42)	<0.01
Family history of diabetes (yes vs. no)	—	—	1.99 (1.41–2.81)	<0.01
Hyperlipidemia (yes vs. no)	1.35 (1.04–1.75)	<0.01	—	—
Waist circumference, per 5 cm increment	1.16 (1.12–1.20)	<0.01	1.25 (1.18–1.34)	<0.01
Average diastolic BP, per 10 mmHg increment	1.15 (1.04–1.28)	<0.01	1.64 (1.33–2.02)	<0.01
General health (fair/poor vs. good)	—	—	1.74 (1.19–2.53)	<0.01
Passive tobacco exposure at home/work (yes vs. no)	0.75 (0.63–0.89)	<0.01	0.57 (0.40–0.82)	<0.01

Risk factors selected for stepwise logistic regression model were selected from risk factors with significant association (P < 0.05) in model 1. For stepwise regression, risk factors were allowed to enter and stay if P < 0.05.

Some risk factors showed no association with prediabetes versus euglycemia (family history of diabetes; general health) and with new T2DM diagnosis versus euglycemia (hyperlipidemia) and their odds ratios are not reported.

Risk measures and survey are questions defined in Supplementary Table S1.

Discussion

In this study, self-identified Hispanic adults with cardiometabolic risk factors, including a family history of diabetes, peripheral obesity, and central obesity, had higher odds of new T2DM diagnosis versus euglycemia compared to prediabetes versus euglycemia. Most lifestyle factors showed no association with odds of prediabetes and new T2DM diagnosis, except for vegetable consumption and passive tobacco exposure, which showed similar lower odds with both. High confidence in ability for self-care was associated with similar lower odds of prediabetes and new T2DM diagnosis, whereas other wellness measures showed no association. Taken together, this study identified modifiable cardiometabolic risk factors and lifestyle factors amenable to tailored interventions to reduce the risk of prediabetes and new T2DM diagnosis.

Peripheral and central obesity were confirmed as major cardiometabolic risk factors associated with higher odds of prediabetes and new T2DM diagnosis. The high frequency of obesity among participants with prediabetes and new T2DM diagnosis agrees with reports of the high frequency of obesity among Hispanic adults and among those with prevalent T2DM.^12
–14 These results are consistent with the positive association between BMI (per standard deviation increment) and incident diabetes among Hispanic adults in the Northern Manhattan Study.⁷ The NIH/SOL study showed a higher odds of undiagnosed diabetes among overweight participants (BMI 25.0–29.9 kg/m² vs. <25.0 kg/m²); however, participants with class I obesity (30.0–34.9 kg/m² vs. <25.0 kg/m²) and class II obesity (≥35.0 kg/m² vs. <25.0 kg/m²) showed no association with odds of undiagnosed diabetes, in contrast to the positive association in our study.⁸ The reasons for the differences between the two studies are unclear and may point to different roles for obesity in the pathogenesis of undiagnosed diabetes in different Hispanic cohorts. Furthermore, the reference groups used in both studies differed. The NIH/SOL study used BMI <25.0 kg/m² as the reference group, whereas this study used ≥30 kg/m² versus lower to define obesity. However, these differences should magnify the estimates in the NIH/SOL study and reduce the estimates in this study. Using stepwise regression models in this study, BMI was not selected in the final parsimonious models. Instead, waist circumference was selected for prediabetes and new T2DM diagnosis. This suggests that different measures of obesity (BMI and waist circumference) may have different associations with prediabetes and new T2DM diagnosis and require further investigation. The strong association of obesity with prediabetes and new T2DM diagnosis is concerning given their association with increased risk for incident cardiovascular disease (common complication of T2DM) and mortality.^15

–18

Despite the higher burden of cardiometabolic risk factors among participants with prediabetes and new T2DM diagnosis, unhealthful lifestyle factors showed no association with odds of prediabetes and new T2DM diagnosis. This study relied on self-reported lifestyle factors and did not characterize daily caloric consumption or food sources, which may have differed between the groups. Furthermore, our survey asked about recent lifestyle factors (e.g., within 30 days before enrollment) and legacy effects of previous lifestyle habits (i.e., more than 30 days before enrollment) not captured in lifestyle questions that may have been captured in clinical risk factors such as obesity and hypertension, which showed positive associations with odds of prediabetes and new T2DM diagnosis. The majority of participants, including those with euglycemia, did not meet dietary or exercise recommendations associated with lowering the risk for prediabetes and new T2DM diagnosis.^19,20 The low adherence to healthful lifestyle/behavior may be determined by, in part, low health awareness, lack of culturally competent lifestyle recommendations, and patient preferences for lifestyle/behavior.^21

–27 Diet soft drinks may appear to be “healthful” alternatives to nondiet drinks; however, this study builds on previous reports of the higher cardiometabolic risk associated with use of non-nutritive sweeteners.²⁸ We assessed the association of individual lifestyle factors with odds of prediabetes and new T2DM diagnosis; however, a composite lifestyle score may show stronger association with prediabetes and new T2DM diagnosis. Developing a composite score is beyond the scope of the current study but should be the topic of a future study. The lower odds of prediabetes and new T2DM diagnosis among participants with versus without passive tobacco exposure was intriguing and in contrast to previous reports that showed higher risk.^29

–32 These results must be interpreted cautiously and require further investigation.

Interventions to reduce the risk of prediabetes and new T2DM diagnosis depend on the interplay between genetic and environmental factors. In this study, family history of diabetes was associated with a higher odds of prediabetes and new T2DM diagnosis, which contrasted with the lower odds observed in the NIH/SOL study.⁸ Family history of diabetes may point to shared genetic and environmental (e.g., similar dietary habits, physical activity) factors. In this regard, the Viva La Familia Study of Hispanic participants reported the genetic determinants of obesity and body composition traits among children.^33,34 While genetic factors may influence risk of type 2 diabetes,^35
–37 lifestyle interventions are also effective in reducing the risk of diabetes.^38,39 While there is an urgent need to investigate and implement interventions to reduce the risk of T2DM among Hispanic adults, these interventions must be culturally sensitive and competent to increase uptake and adherence.^24
–26

Previous studies have incorporated cardiometabolic risk factors and lifestyle factors into models to predict current or future risk of T2DM.^{2,4,5,40
–42} Most scores were developed in non-Hispanic cohorts and did not evaluate the contribution of wellness measures. In this regard, this study incorporated cardiometabolic risk factors, lifestyle factors, and wellness measures and identified key determinants associated with prediabetes and new T2DM diagnosis. Our observation that strenuous, moderate, or mild physical activity (≥5 times per week vs. lower) was not associated with odds of prediabetes and new T2DM diagnosis was surprising, but was consistent with a systematic review, in which only one of three studies specific to Hispanic adults showed an association between physical activity and T2DM risk reduction.⁴³ Self-reported physical activity may not reflect true activity, and devices that monitor and track activity may better reflect physical activity.

The American Diabetes Association recommends the use of nonlaboratory risk scores to identify current risk of diabetes.³ The performance of this and other scores in Hispanic cohorts requires further study. Furthermore, the role of metabolic biomarkers in predicting risk of incident diabetes needs further investigation as they may be better composite measures of underlying metabolic pathways (based on cardiometabolic risk factors, lifestyle factors, and wellness measures) associated with increased risk of T2DM. Ultimately, health, wellness, and disease onset are better managed by a supportive health care system that responds to and addresses the health needs of its individuals. The National Institute of Minority Health and Health Disparities (NIMHD) has developed a research framework of domains and levels of individual-level and community-level influence.^44,45 Efforts to improve health and reduce the risk of prediabetes and new T2DM diagnosis among Hispanic adults must be contextualized in this framework.

In this study, the majority of participants were women. Previous studies have investigated barriers among Hispanic men and women for participation in research studies and clinical trials. These studies showed that Hispanic men, more than women, expressed concern about potential physical harm and perceived personal benefit from participating in research studies.^46,47 Further studies are required to identify barriers for participation in research, in particular, among Hispanic men. In addition, the overall low enrollment of Hispanic men and women, compared to White adults, has raised concern that race/ethnic factors may not be incorporated in individualized treatment decisions. A recent systematic review synthesized published strategies to improve recruitment of minority populations, which will be necessary for the development of individualized screening and treatment strategies.⁴⁸

This study has limitations. Most cardiometabolic risk factors, lifestyle factors, and wellness measures were self-reported and subject to recall bias, a common limitation of studies using self-reported questionnaires. Lifestyle questions (e.g., diet, exercise) were not based on documented calorie counts or physical activity. We did not have complete data for all factors and measures. Subjects were asymptomatic at the time of enrollment, and new cases of diabetes were attributed to T2DM, in contrast to type 1 diabetes, which typically manifests in adolescents (age <18 years) and with ketoacidosis or severe insulin deficiency. This study is based on adults in one U.S. state and requires investigation in other cohorts. Despite this, the study has several strengths. We enrolled a large number of participants and used English and Spanish language questionnaires to obtain detailed demographic and risk factor information. We obtained information on wellness measures, which was incompletely characterized in previous studies. We developed parsimonious models of risk factors associated with odds of prediabetes versus euglycemia and new T2DM diagnosis versus euglycemia.

In summary, we studied a large cohort of Hispanic adults in metropolitan Phoenix, AZ, to evaluate the associations of cardiometabolic risk factors, lifestyle factors, and wellness measures with prediabetes and new T2DM diagnosis. This study shows the high burden of cardiometabolic risk factors that are amenable to targeted interventions. Although lifestyle risk factors across groups (euglycemia, prediabetes, and new T2DM diagnosis) were generally similar, there was an overall low adherence to healthful lifestyles known to reduce the risk of prediabetes and new T2DM diagnosis. The results provide a foundation to develop, measure, and implement targeted interventions to improve cardiometabolic risk factors and reduce the risk of prediabetes and new T2DM diagnosis. There is an urgent need for such targeted, culturally competent interventions to improve T2DM primary prevention among Hispanic adults, who currently experience a disproportionate burden of diabetes and its cardiovascular sequelae, compared to White adults. Such efforts will help to reduce the health disparities associated with the primary prevention of T2DM.

Conclusions

In this study of self-identified Hispanic adults in Phoenix, AZ, we described the associations of cardiometabolic risk factors, lifestyle factors, and wellness measures with prediabetes and new T2DM diagnosis. The risk factor model identified modifiable risk factors amenable to targeted intervention. Future work in SPS will examine the associations of cardiometabolic risk factors, lifestyle factors, and wellness measures with incident T2DM and its cardiovascular sequelae. These studies will guide efforts to develop and test culturally competent interventions for Hispanic adults.

Footnotes

Acknowledgments

The authors are grateful to Mountain Park Health Center (MPHC) personnel for assistance in developing the Sangre Por Salud Biobank, in particular, Dr. Davinder Singh, Mrs. Valentina Hernandez, RDN, and Mrs. Crystal Gonzalez. The authors are grateful to Mrs. Giovanna Moreno-Garzon, SPS Senior Program Coordinator, for support and guidance in navigating resources for data and sample collection within the SPS Biobank. The authors are grateful to the Mayo Clinic Center for Individualized Medicine for financial support of the SPS Biobank. The authors acknowledge the dedication and interest of participants who enrolled in the Biobank.

Ethics Approval

This study was approved by the Mayo Clinic Institutional Review Board (19-004408) and conducted in accordance with ethical standards of the institution and the Helsinki Declaration as revised in 2013.

Author Disclosure Statement

No conflicting financial interests exist.

Funding Information

Sagar B. Dugani was supported by the Robert and Elizabeth Strickland Career Development Award; Eleanna De Filippis was supported by KL2 TR002379-02-01, CTSA UL1 TR002377 NCATS/National Institutes of Health James A. Ruppe Career Development Award in Endocrinology, Mayo Clinic Center for Individualized Medicine, and MCA-Research Program Funds FP0090043; Michelle M. Mielke was supported by National Institute on Aging grant R01 AG49704.

Supplementary Material

Supplementary Table S1

Supplementary Table S2

Supplementary Table S3

Supplementary Table S4

References

CDC. National Diabetes Statistics Report 2020. Published 2020. Accessed at https://www.cdc.gov/diabetes/pdfs/data/statistics/national-diabetes-statistics-report.pdf on July 1, 2020.

Noble

, Mathur

, Dent

, et al. Risk models and scores for type 2 diabetes: Systematic review. BMJ, 2011; 343:d7163.

Bang

, Edwards

, Bomback

, et al. Development and validation of a patient self-assessment score for diabetes risk. Ann Intern Med, 2009; 151:775–783.

Kengne

, Beulens

JWJ

, Peelen

, et al. Non-invasive risk scores for prediction of type 2 diabetes (EPIC-InterAct): A validation of existing models. Lancet Diabetes Endocrinol, 2014; 2:19–29.

Juarez

, Gonzalez

, Agne

, et al. Diabetes risk scores for Hispanics living in the United States: A systematic review. Diabetes Res Clin Pract, 2018; 142:120–129.

Spanakis

, Golden

. Race/ethnic difference in diabetes and diabetic complications. Curr Diab Rep, 2013; 13:814–823.

Kulick

, Moon

, Cheung

, et al. Racial-ethnic disparities in the association between risk factors and diabetes: The Northern Manhattan Study. Prev Med, 2016; 83:31–36.

Casagrande

, Menke

, Aviles-Santa

, et al. Factors associated with undiagnosed diabetes among adults with diabetes: Results from the Hispanic Community Health Study/Study of Latinos (HCHS/SOL). Diabetes Res Clin Pract, 2018; 146:258–266.

U.S. Census. United States Census Bureau. Published 2019. Accessed at https://www.census.gov/programs-surveys/geography/guidance/geo-areas/urban-rural/2010-urban-rural.html on July 1, 2020.

10.

Shaibi

, Singh

, De Filippis

, et al. The Sangre Por Salud Biobank: Facilitating genetic research in an underrepresented Latino community. Public Health Genomics, 2016; 19:229–238.

11.

American Diabetes Association. Classification and Diagnosis of Diabetes: Standards of Medical Care in Diabetes-2018. Diabetes Care, 2018; 41(Suppl 1):S13–S27.

12.

Velasco-Mondragon

, Jimenez

, Palladino-Davis

, et al. Hispanic health in the USA: A scoping review of the literature. Public Health Rev, 2016; 37:31.

13.

Aguayo-Mazzucato

, Diaque

, Hernandez

, et al. Understanding the growing epidemic of type 2 diabetes in the Hispanic population living in the United States. Diabetes Metab Res Rev, 2019; 35:e3097.

14.

Schneiderman

, Llabre

, Cowie

, et al. Prevalence of diabetes among Hispanics/Latinos from diverse backgrounds: The Hispanic Community Health Study/Study of Latinos (HCHS/SOL). Diabetes Care, 2014; 37:2233–2239.

15.

Gnatiuc

, Alegre-Díaz

, Wade

, et al. General and abdominal adiposity and mortality in Mexico City: A prospective study of 150 000 adults. Ann Intern Med, 2019; 171:397–405.

16.

Daviglus

, Talavera

, Avilés-Santa

, et al. Prevalence of major cardiovascular risk factors and cardiovascular diseases among Hispanic/Latino individuals of diverse backgrounds in the United States. JAMA, 2012; 308:1775–1784.

17.

Daviglus

, Pirzada

, Durazo-Arvizu

, et al. Prevalence of low cardiovascular risk profile among diverse Hispanic/Latino Adults in the United States by age, sex, and level of acculturation: The Hispanic Community Health Study/Study of Latinos. J Am Heart Assoc, 2016; 5:e003929.

18.

Daviglus

, Pirzada

, Talavera

. Cardiovascular disease risk factors in the Hispanic/Latino population: Lessons from the Hispanic Community Health Study/study of Latinos (HCHS/SOL). Prog Cardiovasc Dis, 2014; 57:230–236.

19.

AHA. Life's Simple 7. Accessed at https://www.heart.org/en/professional/workplace-health/lifes-simple-7 on July 1, 2020.

20.

Evert

, Dennison

, Gardner

, et al. Nutrition therapy for adults with diabetes or prediabetes: A consensus report. Diabetes Care, 2019; 42:731–754.

21.

Rosas

, Lv

, Lewis

, et al. A Latino patient-centered, evidence-based approach to diabetes prevention. J Am Board Fam Med, 2018; 31:364–374.

22.

O'Brien

, Moran

, Tang

, et al. Patient perceptions about prediabetes and preferences for diabetes prevention. Diabetes Educ, 2016; 42:667–677.

23.

Rosal

, Borg

, Bodenlos

, et al. Awareness of diabetes risk factors and prevention strategies among a sample of low-income Latinos with no known diagnosis of diabetes. Diabetes Educ, 2011; 37:47–55.

24.

Camacho Á, González

, Castañeda

, et al. Improvement in depressive symptoms among Hispanic/Latinos receiving a culturally tailored IMPACT and problem-solving intervention in a Community Health Center. Community Ment Health J, 2015; 51:385–392.

25.

Sanchez

, Killian

, Eghaneyan

, et al. Culturally adapted depression education and engagement in treatment among Hispanics in primary care: Outcomes from a pilot feasibility study. BMC Fam Pract, 2019; 20:140.

26.

, Amirehsani

, Wallace

, et al. A family-based, culturally tailored diabetes intervention for Hispanics and their family members. Diabetes Educ, 2016; 42:299–314.

27.

, Amirehsani

, Wallace

, et al. Perceptions of barriers in managing diabetes: Perspectives of Hispanic immigrant patients and family members. Diabetes Educ, 2013; 39:494–503.

28.

Azad

, Abou-Setta

, Chauhan

, et al. Nonnutritive sweeteners and cardiometabolic health: A systematic review and meta-analysis of randomized controlled trials and prospective cohort studies. CMAJ, 2017; 189:E929–E939.

29.

Wang

, Ji

, Liu

, et al. Passive smoking and risk of type 2 diabetes: A meta-analysis of prospective cohort studies. PLoS One, 2013; 8:e69915.

30.

Wei

, Meng

, Yu

. A meta-analysis of passive smoking and risk of developing type 2 diabetes mellitus. Diabetes Res Clin Pract, 2015; 107:9–14.

31.

Oba

, Goto

, Mizoue

, et al. Passive smoking and type 2 diabetes among never-smoking women: The Japan Public Health Center-based Prospective Study. J Diabetes Investig, 2020; 11:1352–1358.

32.

Jiang

, Chang

, Ziogas

, et al. Secondhand smoke, obesity, and risk of type II diabetes among California teachers. Ann Epidemiol, 2019; 32:35–42.

33.

Butte

, Cai

, Cole

, et al. Viva La Familia study: Genetic and environmental contributions to childhood obesity and its comorbidities in the Hispanic population. Am J Clin Nutr, 2006; 84:646–654.

34.

Butte

, Comuzzie

, Cole

, et al. Quantitative genetic analysis of the metabolic syndrome in Hispanic children. Pediatr Res, 2005; 58:1243–1248.

35.

Abdullah Said

, Verweij

, Van Der Harst

. Associations of combined genetic and lifestyle risks with incident cardiovascular disease and diabetes in the UK biobank study. JAMA Cardiol, 2018; 3:693–702.

36.

Meigs

, Shrader

, Sullivan

, et al. Genotype score in addition to common risk factors for prediction of type 2 diabetes. N Engl J Med, 2008; 359:2208–2219.

37.

Udler

, McCarthy

, Florez

, et al. Genetic risk scores for diabetes diagnosis and precision medicine. Endocr Rev, 2019; 40:1500–1520.

38.

Corona

, Flores

, Arab

. Trends in evidence-based lifestyle interventions directed at obese and overweight adult Latinos in the US: A systematic review of the literature. J Community Health, 2016; 41:667–673.

39.

Knowler

, Barrett-Connor

, Fowler

, et al. Reduction in the incidence of type 2 diabetes with lifestyle intervention or Metformin. N Engl J Med, 2002; 346:393–403.

40.

Yokota

, Miyakoshi

, Sato

, et al. Predictive models for conversion of prediabetes to diabetes. J Diabetes Complications, 2017; 31:1266–1271.

41.

Leal

, Morrow

, Khurshid

, et al. Decision models of prediabetes populations: A systematic review. Diabetes Obes Metab, 2019; 21:1558–1569.

42.

Herman

, Smith

, Thompson

, et al. A new and simple questionnaire to identify people at increased risk for undiagnosed diabetes. Diabetes Care, 1995; 18:382–387.

43.

Boyer

, Churilla

, Ehrlich

, et al. Protective role of physical activity on type 2 diabetes: Analysis of effect modification by race–ethnicity. J Diabetes, 2018; 10:166–178.

44.

NIMHD. NIMHD Research Framework. Accessed at https://nimhd.nih.gov/about/overview/research-framework/nimhd-framework.html on June 19, 2020.

45.

National Institutes of Health. National Institute on Minority Health and Health Disparities (NIMHD). Published 2019. Accessed at https://www.nih.gov/about-nih/what-we-do/nih-almanac/national-institute-minority-health-health-disparities-nimhd on June 19, 2020.

46.

Arevalo

, Heredia

, Krasny

, et al. Mexican-American perspectives on participation in clinical trials: A qualitative study. Contemp Clin Trials Commun, 2016; 4:52–57.

47.

Ceballos

, Knerr

, Scott

, et al. Latino beliefs about biomedical research participation: A qualitative study on the U.S.-Mexico border. J Empir Res Hum Res Ethics, 2014; 9:10–21.

48.

Ibrahim

, Sidani

. Strategies to recruit minority persons: A systematic review. J Immigr Minor Health, 2014; 16:882–888.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.04 MB

0.03 MB