Educational attainment does not modify the effect of educational interventions on blood pressure control: a secondary analysis of data from a randomised trial

Abstract

Summary

Objectives

To assess whether the effects of community-based educational interventions to improve blood pressure, weight and health behaviours benefit participants with lower educational levels more than those with higher educational levels.

Design

Secondary data analysis.

Setting

Two 12-month community-based educational interventions, one led by trained peers and one delivered by health professionals.

Participants

A total of 403 hypertensive individuals, grouped by education (high school or less; 1–3 years college; 4 + years college).

Main outcome measures

Blood pressure, weight, physical activity and fruit and vegetable intake.

Results

We found that changes in blood pressure, weight and physical activity were similar across education levels; college graduates consumed more daily servings of fruits and vegetables at baseline (3.7 versus 3.6 for those with 12–15 years and 3.1 for those with < 12 years, p = 0.0112), and increased intake more after the intervention (+0.4 versus −0.1 and −0.1, p = 0.0142). The two methods of delivery – peer-led versus professional – had similar effects on all measures.

Conclusions

We conclude that educational interventions, whether delivered by peers or professionals, may improve chronic disease self-management among participants but do not confer greater benefits on participants with lower educational levels.

Keywords

Hypertension health education social conditions and disease health service research

Introduction

Hypertension is a common, highly treatable chronic illness, affecting 30.5% of adult men and 28.5% of adult women in the United States (US).¹ It is a major risk factor for heart disease and stroke; 69% of Americans who have a first heart attack and 77% of Americans treated for a first stroke have blood pressure above 140/90 mmHg.² Despite the effectiveness of drug therapy in controlling hypertension, fewer than half (46.5%) of US adults with hypertension achieve target blood pressure.¹

There is strong evidence that persons with lower socioeconomic status are more likely to have hypertension and less likely to achieve optimal blood pressure when treated.^3–10 Paulsen et al.⁵ investigated the association between socioeconomic status and blood pressure control in Denmark, a healthcare system with free access to care/treatment. They determined that patients who are under 65 years of age with an educational level of 10–12 years are more likely to have their blood pressure under control than patients with an educational level of <10 years. Poorer blood pressure control in low socioeconomic status populations might be due to the association between socioeconomic status and factors known to be related to blood pressure control, such as hypertension knowledge, medication adherence, diet, weight, self-efficacy and social support.^6–10

Numerous studies suggest that educational interventions can improve hypertension knowledge, health behaviours, and blood pressure control.^11–13 Researchers have shown a particular interest in interventions delivered by peer educators, which show promise as sustainable, cost-effective adjuncts to standard clinical care, especially for chronic conditions. For example, randomised trials of commercial, largely peer-directed weight loss programmes such as Weight Watchers have demonstrated efficacy.¹⁴ Similarly, the Chronic Disease Self-Management Program (CDSMP) developed by Lorig and her colleagues at Stanford University has improved functional status, self-efficacy, and/or healthcare use in a number of randomised trials.¹⁵ The usage of peer support to reduce hypertension has not been studied extensively, but preliminary evaluations of the National Heart, Lung and Blood Institute-sponsored program ‘Salud Para Su Corazon (Health for Your Heart)’ provide some reason for optimism.¹⁶

Given the evidence for the effectiveness of educational interventions in improving health behaviours, we hypothesised that such interventions might particularly benefit individuals with less education, since part of the disparity in hypertension prevalence and control is explained by differences in lifestyle-related risk factors.^6,8 Moreover, we hypothesised that blood pressure education delivered by trusted peers might be more effective than education delivered by health professionals, since peers are likely to convey health information in a less academic, more comprehensible manner.¹⁷

To test these hypotheses, we performed a secondary analysis of data from a randomised trial comparing two types of educational interventions promoting hypertension self-management among hypertensive Veterans: one delivered by trained peers and one delivered by health professionals. We first determined whether changes in blood pressure and lifestyle were related to participants' education. We then examined whether the relative effectiveness of peer-led versus professionally delivered education varied by level of education.

Methods

Study overview

We conducted a secondary analysis of data from POWER II, a cluster randomised controlled trial. In that study, hypertension control improved in both groups and the impact of the two interventions was similar overall.¹⁸ For the present analysis, we compared the change in blood pressure control, weight and health behaviours during the intervention period among participants with different levels of education. We then examined whether there was an interaction between the intervention's delivery mechanism (peer-led versus professionally led) and participants' educational level. The Zablocki Veterans Affairs Medical Center's Human Studies Subcommittee (IRB) approved the randomised controlled trial and our secondary data analysis (9392-08H); participants in the randomised controlled trial provided written informed consent.

Description of POWER II

POWER II participants were recruited from individual units of various Veterans' service organisations (e.g. American Legion, Veterans of Foreign Wars) in Southeast Wisconsin. These local units (typically called posts) meet monthly to conduct organisational business; 58 posts agreed to participate. All participating posts were provided with an automated blood pressure monitor, weight scale and pedometers. For the peer-led intervention, volunteer post members were trained to deliver 10-min educational sessions about blood pressure self-management at each monthly meeting for a total of 12 sessions. For the professionally delivered intervention, participants were invited to attend three 90-min self-management seminars that were held once per quarter and variously led by a physician, a pharmacist, a combat medic and an occupational therapist. Volunteer post members encouraged attendance at these seminars. Topics covered by the peer leaders in monthly sessions and those covered by the health professionals in the seminars were similar, including exercise, diet and the importance of blood pressure monitoring.

Randomisation occurred at the post level. Prior to alerting the posts of their random assignment, 404 hypertensive post members were recruited to participate in a study of the effect of the intervention on blood pressure control. Among these, 219 were at posts assigned to the peer-led intervention and 185 at posts assigned to the professionally led intervention.

Measures

Trained research assistants collected the following data from POWER II's 404 study participants: (1) resting blood pressure, (2) weight and (3) survey data, which included demographic questions (such as age, race, education, income), medical history and previously validated measures of health behaviour, including fruit and vegetable intake,¹⁹ sodium intake,²⁰ and physical activity levels,²¹ as well as self-efficacy,²² patient activation²³ and engagement in healthcare.²⁴ Details of the POWER II data collection methods have been described.^18,25 One participant did not provide his educational level, leaving 403 participants eligible for this analysis.

Analysis

We divided participants into three groups based on educational attainment: (1) those with no education beyond high school ( ≤ High school); (2) those with some college or technical school (1–3 years college); and (3) college graduates (4 + years college). We selected our cut points prior to looking at outcomes by group, seeking to have intuitive groups with similar numbers in each group. Statistical analysis of the POWER II data requires adjusting for the cluster sample structure of the data. Comparisons of continuous response variables were made with generalised linear models and adjusted for clustering using generalised estimating equations (GEE), resulting in an ANOVA-like hypothesis test. Comparison of frequency response variables was made using cluster-adjusted logistic or ordinal regression. Table 1 reports unadjusted means and frequencies with cluster adjusted p-values. Tables 2 and 3 report multivariable model based means (least square means) and p-values adjusting for education and baseline value of the reported difference (to control for regression to the mean). Where interactions were significant, pairwise multiple comparisons were made using a Bonferroni adjustment. All analyses were performed using Proc GENMOD and Proc SURVEYLOGISTIC in SAS version 9.3 (The SAS Institute, Cary, NC).

Table 1.

Baseline characteristics by level of education.

Participant characteristic		Highest level of education completed			p
Participant characteristic	Total, N (%)	≤ High school N = 168	1–3 years college N = 139	4 + years college N = 96	p
		n (%)	n (%)	n (%)
Age in years (µ, σ)	68.2, 10.2	69.1, 10.5	68.2, 9.6	66.6, 10.3	0.2534 ^a
Male gender	352 (87.3)	149 (88.7)	117 (84.2)	86 (89.6)	0.9853 ^b
Non-White race	15 (3.7)	5 (3.0)	8 (5.8)	2 (2.1)	0.9196 ^b
Employment status					0.0511 ^c
Retired	277 (68.7)	119 (70.8)	97 (69.8)	61 (63.5)
Employed	100 (24.8)	34 (20.2)	34 (24.5)	32 (33.3)
Other	26 (6.5)	15 (8.9)	8 (5.8)	3 (3.1)
Household income					<.0001 ^c
Less than $25,000	57 (14.1)	36 (21.4)	14 (10.1)	7 (7.3)
$25,000–34,999	64 (15.9)	25 (14.9)	32 (23.0)	7 (7.3)
$35,000–49,999	75 (18.6)	34 (20.2)	26 (18.7)	15 (15.6)
$50,000–74,999	70 (17.4)	28 (16.7)	20 (14.4)	22 (22.9)
$75,000 or more	64 (15.9)	16 (9.5)	18 (12.9)	30 (31.3)
Numeracy					0.0051 ^c
0	46 (11.5)	24 (14.4)	15 (10.9)	7 (7.4)
1	165 (41.3)	68 (40.7)	66 (47.8)	31 (32.6)
2	136 (34.0)	60 (35.9)	40 (29.0)	36 (37.9)
3	53 (13.3)	15 (9.0)	17 (12.3)	21 (22.1)
Health literacy (µ, σ)	61.9 ± 7.4	60.6 ± 8.1	62.0 ± 8.0	64.1 ± 3.8	<.0001 ^a
Self-efficacy (µ, σ)	31.6 ± 3.9	31.0 ± 3.6	31.6 ± 4.0	32.8 ± 4.0	0.0010 ^a
Preference for active role in health (µ, σ)	7.0 ± 2.9	6.0 ± 2.6	7.5 ± 2.9	8.3 ± 2.8	<.0001 ^a
Chronic conditions
Diabetes	111 (27.5)	52 (31.0)	40 (28.8)	19 (19.8)	0.0383 ^b
Renal disease	23 (5.7)	10 (6.0)	9 (6.5)	4 (4.2)	0.6204 ^b
Heart disease	127 (31.5)	59 (35.1)	42 (30.2)	26 (27.1)	0.2137 ^b
Stroke	33 (8.2)	12 (7.1)	15 (10.8)	6 (6.3)	0.9987 ^b
Smoking					0.0337 ^b
Current	46 (11.4)	24 (14.4)	16 (11.5)	6 (6.3)
Missing	33 (8.2)	22 (13.2)	8 (5.8)	3 (3.1)
Never	119 (29.6)	41 (24.6)	40 (28.8)	38 (39.6)
Past	204 (50.7)	80 (47.9)	75 (54.0)	49 (51.0)
Uncontrolled hypertension	183 (45.4)	83 (49.4)	61 (43.9)	39 (40.6)	0.0829 ^b
BMI ≧ 30 kg/m²	204 (50.6)	87 (51.8)	65 (46.8)	52 (54.2)	0.7808 ^a
Systolic BP mm Hg (µ, σ)	134.5 ± 15.6	134.9 ± 17.4	134.4 ± 14.2	133.7 ± 14.1	0.5182 ^a
Diastolic BP mm Hg (µ, σ)	72.4 ± 11.3	71.7 ± 12.4	72.7 ± 10.0	73.3 ± 11.2	0.2534 ^a
Weight, pounds (µ, σ)	210.1 ± 44.1	209.9 ± 46.6	205.8 ± 37.8	216.8 ± 47.4	0.3387 ^a
Servings of F&V (µ, σ)	3.4 ± 1.5	3.1 ± 1.4	3.6 ± 1.7	3.7 ± 1.4	0.0013 ^a
Physical activity in METS/week (µ, σ)	2938 ± 5216	3332 ± 6625	2782 ± 4323	2480 ± 3268	0.1835 ^a

BMI: body mass index; BP: blood pressure; F&V: fruits and vegetables; METS: metabolic equivalents.

^aCluster adjusted ANOVA test (general linear model using GEE) for linear education effect.

^bCluster adjusted logistic regression testing a linear education effect.

^cCluster adjusted logistic regression with cumulative logit link testing a linear education effect.

Table 2.

Change in outcomes at 12-month follow-up by educational level.

Outcome measure	≦ High school	1–3 years college	4 + years college	p
SBP mm Hg Mean Δ (95% CI)	−4.1 (−6.2, −2.0)	−4.0 (−6.3, −1.7)	−5.3 (−8.0, −2.5)	0.7518
DBP mm Hg Mean Δ (95% CI)	−1.7 (−3.0, −0.3)	−1.6 (−3.1, −0.1)	−2.6 (−4.4, −0.9)	0.6219
Weight, pounds Mean Δ (95% CI)	0.0 (−1.7, 1.6)	1.5 (−0.2, 3.3)	−1.7 (−3.7, 0.4)	0.0700
Physical activity (METS per week, % increase, 95% CI)	6 (−20, + 39)	52 (14, 103)	62 (15, 128)	0.0802
Sodium intake score Mean Δ (95% CI)	−0.2 (−0.5, 0.1)	0.0 (−0.3, 0.3)	0.0 (−0.4, 0.3)	0.6910
Servings of F&V Mean Δ (95% CI)	−0.1 (−0.3, 0.1)	−0.1 (−0.3, 0.2)	0.4 (0.1, 0.7)	0.0142
Self-efficacy Mean Δ (95% CI)	0.1 (−0.5, 0.7)	0.6 (0.0, 1.2)	0.5 (−0.2, 1.3)	0.3846
Engagement with care Mean Δ (95% CI)	0.1 (−0.2, 0.5)	0.2 (−0.2, 0.6)	0.4 (−0.1, 0.8)	0.7975
Patient activation Mean Δ (95% CI)	0.6 (−1.2, 2.4)	1.2 (−0.7, 3.1)	1.9 (−0.4, 4.2)	0.6628

SBP: systolic blood pressure; Δ: change; DBP: diastolic blood pressure; CI: confidence interval; METS/wk.: metabolic equivalents/week; F&V: fruits and vegetables.

Table 3.

Change in blood pressure, weight and diet by education and intervention.^a

Endpoints	Intervention group	Highest level of education completed
Endpoints	Intervention group	≤ High school	1–3 years college	4 + years college
SBP mm Hg^b Mean Δ (95% CI)	Peer-led	−5.64 (−8.61,−2.66)	−3.16 (−6.38,0.05)	−4.90 (−8.35,−1.46)
SBP mm Hg^b Mean Δ (95% CI)	Professional	−2.44 (−5.53,0.64)	−4.93 (−8.23,−1.63)	−5.82 (−10.39,−1.24)
Weight, pounds^c Mean Δ (95% CI)	Peer-led	−1.04 (−3.25,1.17)	1.77 (−0.64,4.18)^c	−3.97 (−6.54,−1.41)^c
Weight, pounds^c Mean Δ (95% CI)	Professional	1.04 (−1.28,3.35)	1.24 (−1.22,3.70)	2.55 (−0.89,5.99)^c
Servings of F&V^b Mean Δ (95% CI)	Peer-led	0.02 (−0.29, 0.34)	−0.02 (−0.35, 0.32)	0.55 (0.20, 0.91)
Servings of F&V^b Mean Δ (95% CI)	Professional	−0.22 (−0.55, 0.10)	−0.11 (−0.45, 0.23)	0.10 (−0.36, 0.55)

Δ: change in measure; CI: confidence interval; SBP: systolic blood pressure; F&V: fruits and vegetables.

^aThere is no significant interaction between education and intervention for any endpoint.

^bThere is no significant Education-by-Intervention interaction effects for change in SBP and F&V servings (p = 0.27, 0.60).

^cThere is a significant Education-by-Intervention interaction effect for change in weight (p = 0.016). Within the 4 + years college group, peer-led lost more weight than professional (p = 0.0025). Within the peer-led group, those with 4 + years college lost more weight than those with 1–3 years of college (p = 0.0015). These may be considered significant at the 0.05 level with a Bonferroni adjustment for nine comparisons.

Results

Reflecting the demographics of the participating organisations, the 403 study participants were primarily older (mean age 68.2 years), white (96%) and male (87%). Table 1 presents the baseline characteristics of participants by educational level. While baseline systolic blood pressure and weight were similar across education groups, educational level was positively associated with our measures of literacy,^26,27 patient activation, self-efficacy and preference for participation in healthcare, as well as daily servings of fruits and vegetables. Conversely, educational level was inversely associated with the likelihood of smoking. Other measures were similar across groups.

At 12 months, we had data from 378 of the 403 participants; 13 died, 4 moved, 5 withdrew, and 3 were lost to follow up. Mean systolic blood pressure decreased by 4.4 mm Hg overall, while overall weight, daily servings of fruits and vegetables and physical activity were unchanged. We present changes in outcomes measures from baseline to 12 months by educational level in Table 2. The changes in systolic blood pressure, weight and metabolic equivalents/week of physical activity were similar in all education groups. However, fruits and vegetables intake improved more among participants with the highest level of education (+0.4 servings/daily) than in the low (−0.1 servings/daily) and medium education groups (−0.1 servings/daily), (p = 0.0142).

There was no pattern to the interactions between the effect of level of education and that of intervention approach on changes in the outcome measures. We present changes in systolic blood pressure, weight and daily servings of fruits and vegetables in Table 3. Although the impact of peer-led compared with professionally delivered education on weight was greatest in the most highly educated group, this pattern was not seen for other behaviour changes. Moreover, persons with the lowest levels of education had a more favourable change in weight than those with intermediate levels of education. This lack of trend is surprising, and we have no explanation other than a possible type I error.

Discussion

Principal findings

The results of our analysis do not support the hypothesis that educational interventions may be particularly useful in populations with lower levels of education; they also give no indication that persons with less education do not benefit. Across a number of outcomes, less and better educated individuals had similar changes in systolic blood pressure or health behaviours. This finding was true for both the peer-delivered and professionally delivered educational interventions. Similarly, we found no evidence that persons with less education derived more benefit from peer-delivered education than they did from professionally led educational sessions.

We note that this study had a number of strengths. We examined a range of outcomes that were either objective (blood pressure, weight) or based on validated instruments. Our educational intervention was also efficacious, generating a significant improvement in hypertension knowledge²⁸ during the intervention, as well as improved blood pressure control.

Several limitations of our study warrant consideration. First, our study population was relatively homogeneous – all participants were Veterans from southeastern Wisconsin; most were older white men. However, this homogeneity made it less likely that educational status would be confounded by other characteristics associated with response.

Second, we had relatively few participants with very limited education; only 26 participants did not have at least a high school education. Thus, we cannot determine how individuals with only primary education respond to an educational intervention. Furthermore, despite our effort to recruit peer educators who shared similar background and experience with the participants, our peer educators tended to be better educated than the average participants (data not shown). Nevertheless, they were members of the same relatively small (10–40 members) group of Veterans that met on a monthly basis, suggesting they were truly peers. However, due to their different educational backgrounds, they might not have delivered the material in a way that would facilitate learning among participants with less formal schooling.

Third, although our intervention duration occurred over an entire year, we were addressing long-standing behaviour patterns, and the number of educational contacts was limited. It may be that a longer or more intense intervention would do more to alleviate disparities related to educational background.

Finally, we note that our study was relatively small, so that a small difference in the impact of our intervention on blood pressure control could have been missed. However, the confidence intervals around our estimates of change in blood pressure in each educational group suggest any effect would be small.

Relation to other studies

Our study is the first to examine this question among persons with hypertension, but it is consistent with other research, which has also suggested that the impact of educational interventions is similar across educational backgrounds, and may actually be greater among persons with more education. Bosma et al.²⁹ found that among patients with type II diabetes mellitus or chronic obstructive lung disease and mild to moderate depressive symptoms, a self-management intervention only improved depressive symptoms among persons with more education.

Educational interventions may be unable to mitigate socioeconomic disparities in health because they fail to address important environmental factors which are associated with lower socioeconomic status. For example, individuals with less education are more likely to live in neighbourhoods with fewer stores selling fresh produce, which could result in lower fruits and vegetables intake.³⁰ Moreover, as seen in our study, they are likely to have lower levels of health literacy, numeracy, and self-efficacy and be less actively involved in their own healthcare. These factors may limit the impact of educational interventions directed at chronic disease self-management.

Our finding that individuals with less education did not benefit more from peer-led education contrasts with the results of Rhee et al.¹⁷ who found that adolescents from lower socioeconomic backgrounds benefited significantly more from peer-led asthma education than the adult-led asthma education. It is difficult to compare our study with this study, which used a different form of intervention for a different disease in a different population. Moreover, the advantage of peer education in the asthma study was confined to asthma quality of life; objective measures of pulmonary function did not improve in either group. It may be that quality of life is more apt to respond to peer education. In the present study, we focused on health behaviours and blood pressure control.

Implications and future work

Despite these caveats, we believe our findings are important. In a large population of hypertensive patients, we found nothing to suggest that educational interventions alone will reduce disparities in chronic disease self-management. Indeed, similar to Bosma et al.,²⁹ the trends we observed suggest that persons with greater baseline education are more likely to benefit from such programs, exacerbating socioeconomic health disparities.

Future studies should examine whether more intensive self-management programs, such as the 15 contact-hour peer-led Chronic Disease Self-Management Program developed by Lorig et al.¹⁵ would diminish disparities in self-management skills. In the interim, efforts should continue to address environmental factors that make it harder for disadvantaged groups to self-manage chronic conditions.

Disclaimer

The views expressed in this article are those of the authors and do not necessarily reflect the position or policy of the Department of Veterans Affairs or the United States government.

Footnotes

Declarations

Funding

This work was supported by a grant from the Department of Veterans Affairs, Veterans Health Administration, Office of Research and Development, Health Services Research and Development (Project IAB 06-086-2, Working with Veterans Service Organizations to Improve Blood Pressure).

Guarantor

JW.

Ethical approval

Trial Registration: ClinicalTrials.gov NCT00571038/

Contributorship

KZ generated the hypothesis, conducted the literature review, wrote the first draft, and provided feedback on subsequent revisions; DE performed the statistical analyses and created the tables; KE recruited, consented, and collected data from participants in the randomised controlled trial, and ensured adherence to study protocols. She edited and proofread all drafts of the paper for content and clarity; JW designed the randomised controlled trial, oversaw its conduct, and analysed and interpreted the data. He assisted KZ with the design, conduct, and write-up of the secondary data analysis. He reviewed and edited all drafts of the paper.

Acknowledgements

We would like to acknowledge the support we received from the staff in our local Research Office.

Provenance

Not commissioned; peer-reviewed by Anthony Laverty.

References

Guo

Zhang

Walton

. Trends in prevalence, awareness, management, and control of hypertension among United States adults, 1999 to 2010. J Am Coll Cardiol 2012; 60: 599–606.

Lloyd-Jones

Adams

Brown

. Heart disease and stroke statistics – 2010 update: a report from the American Heart Association. Circulation 2010; 121: e46–e215.

Kiely

Gross

Kim

Lipsitz

. The association of educational attainment and SBP among older community-living adults: the Maintenance of Balance, Independent Living, Intellect, and Zest in the Elderly (MOBILIZE) Boston Study. J Hypertens 2012; 30: 1518–1525.

Kaplan

Huguet

Feeny

McFarland

. Self-reported hypertension prevalence and income among older adults in Canada and the United States. Soc Sci Med 2010; 70: 844–849.

Paulsen

Anderson

Munck

Larsen

Hansen

Jacobsen

. Socio-economic status influences blood pressure control despite equal access to care. Fam Pract 2012; 29: 503–510.

Minor

Wofford

Wyatt

. Does socioeconomic status affect blood pressure goal achievement? Curr Hypertens Rep 2008; 10: 390–397.

Luepker

Rosamond

Murphy

Sprafka

Folsom

McGovern

. Socioeconomic status and coronary heart disease risk factor trends: the Minnesota Heart Survey. Circulation 1993; 88(5 Pt 1): 2172–2179.

Darmon

Drewnowski

. Does social class predict diet quality? Am J Clin Nutr 2008; 87: 1107–1117.

Matthews

Räikkönen

Gallo

Kuller

. Association between socioeconomic status and metabolic syndrome in women: testing the reserve capacity model. Health Psychol 2008; 27: 576–583.

10.

Miyaki

Song

Taneichi

Tsutsumi

Hashimoto

Kawakami

. Socioeconomic status is significantly associated with dietary salt intakes and blood pressure in Japanese workers (J-HOPE Study). Int J Environ Res Public Health 2013; 10: 980–993.

11.

Dawes

Kaczorowski

Swanson

Hickey

Karwalajtys

. The effect of a patient education booklet and BP ‘tracker’ on knowledge about hypertension: a randomized controlled trial. Fam Pract 2010; 27: 472–478.

12.

Iso

Shimamoto

Yokota

Sankai

Jacobs

Jr Komachi

. Community-based education classes for hypertension control: a 1.5-year randomized controlled trial. Hypertension 1996; 27: 968–974.

13.

Morisky

Levine

Green

Shapiro

Russell

Smith

. Five-year blood pressure control and mortality following health education for hypertensive patients. Am J Public Health 1983; 73: 153–162.

14.

Heshka

Anderson

Atkinson

Greenway

Hill

Phinney

. Weight loss with self-help compared with a structured commercial program: a randomized trial. JAMA 2003; 289: 1792–1798.

15.

Lorig

Sobel

Stewart

Brown

Bandura

Ritter

. Evidence suggesting that a chronic disease self-management program can improve health status while reducing hospitalization: a randomized trial. Med Care 1999; 37: 5–14.

16.

Balcázar

Alvarado

Hollen

Gonzalez-Cruz

Pedregón

. Evaluation of Salud Para Su Corazón (Health for your Heart) – National Council of La Raza Promotora Outreach Program. Prev Chronic Dis 2005; 2: A09.

17.

Rhee

Belyea

Hunt

Brasch

. Effects of a peer-led asthma self-management program for adolescents. Arch Pediatr Adolesc Med 2011; 165: 513–519.

18.

Whittle

Schapira

Fletcher

Hayes

Morzinski

Laud

. A randomized trial of peer-delivered self-management support for hypertension. Am J Hypertens 2014; 27: 1416–1423.

19.

Thompson

Kipnis

Subar

Krebs-Smith

Kahle

. Evaluation of 2 brief instruments and a food-frequency questionnaire to estimate daily number of servings of fruit and vegetables. Am J Clin Nutr 2000; 71: 1503–1510.

20.

Hopkins

Williams

Kuida

Stults

Hunt

Barlow

. Predictive value of a short dietary questionnaire for changes in serum lipids in high-risk Utah families. Am J Clin Nutr 1989; 50: 292–300.

21.

Craig

Marshall

Sjöström

Bauman

Booth

Ainsworth

. International physical activity questionnaire: 12-country reliability and validity. Med Sci Sports Exerc 2003; 35: 1381–1395.

22.

Schwarzer

Jerusalem

Generalized self-efficacy scale. In: Weinman

Wright

Johnston

(eds). Measures in health psychology: a user's portfolio. Causal and control beliefs, Windsor, UK: NFER-NELSON, 1995, pp. 35–37.

23.

Hibbard

Stockard

Mahoney

Tusler

. Development of the Patient Activation Measure (PAM): conceptualizing and measuring activation in patients and consumers. Health Serv Res 2004; 39(4 Pt 1): 1005–1026.

24.

Krantz

Baum

Wideman

. Assessment of preferences for self-treatment and information in health care. J Pers Soc Psychol 1980; 39: 977–990.

25.

Patterson

McGinley

Ertl

Morzinski

Fyfe

Whittle

. Location and organizational features: what type of veteran communities participate in health programs? Prog Commun Health Partnersh 2012; 6: 141–152.

26.

Davis

Long

Jackson

Mayeaux

George

Murphy

. Rapid estimate of adult literacy in medicine: a shortened screening instrument. Fam Med 1993; 25: 391–395.

27.

Lipkus

Samsa

Rimer

. General performance on a numeracy scale among highly educated samples. Med Decis Making 2001; 21: 37–44.

28.

Schapira

Fletcher

Eastwood

Hayes

Patterson

Ertl

. The development and validation of the Hypertension Evaluation of Lifestyle and Management (HELM) knowledge scale. J Clin Hypertens 2012; 14: 461–466.

29.

Bosma

Lamers

Jonkers

van Eijk

. Disparities by education level in outcomes of a self-management intervention: the DELTA trial in The Netherlands. Psychiatr Serv 2011; 62: 793–795.

30.

Zenk

Lachance

Schultz

Mentz

Kannan

Ridella

. Neighborhood retail food environment and fruit and vegetable intake in a multiethnic urban population. Am J Health Promot 2009; 23: 255–264.