Effectiveness of DVD vs. group-initiated diabetes prevention on information uptake for high & low health literacy participants

Abstract

Objective:

This study evaluates the effectiveness of technology versus in-person, group-initiated diabetes prevention to enhance comprehension of learning objectives between patients with differing health literacy (HL).

Methods:

Evidence-based content through either a DVD (n = 217) or in-person, group class (n = 225) to initiate the intervention. A teach-back call was used to assess comprehension of, and reinforce, learning objectives. Chi-squared was used to determine differences between conditions (DVD vs Class) and HL levels (High n = 361 vs. Low n = 81) and regression analyses were used to examine relationships.

Results:

DVD participants performed significantly better across teach back questions (15.4 ± 2.5 v. 14.8 ± 2.6, p <0.01), demonstrated comprehension in fewer teach-back rounds (1.9 ± 0.7 v. 2.1 ± 0.7, p < 0.01), and answered more questions correctly on the first try (4.2 ± 1.6 v. 3.4 ± 1.8, p <0.01). Models for HL levels and modality by HL level were statistically significant (p < 0.01) favoring the DVD.

Conclusion:

Initiating a diabetes prevention program with the use of a DVD appears to be a superior option to in-person, class sessions. Teach-back and teach-to-goal strategies enables participants of both high and low health literacy levels to receive and confirm mastery of diabetes prevention objectives.

Practice Implications:

A teach-back call may improve information uptake increasing the likelihood of health behavior uptake.

1. Introduction

Currently in the United States, approximately 35% of the population have prediabetes, 38% are obese and approximately 40% have either impaired glucose tolerance or fasting glucose levels [1–3]. Additionally, most Americans have at least one other risk factor that has been observed to contribute to diabetes and/or cardiometabolic risk such as physical inactivity, family history of type 2 diabetes, hypertension, or high body mass index [4–7]. Annual medical costs incurred for an individual patient have been observed to range from $417 to $4117 for one to four risk factors, respectively, as well as experiencing diminished quality of life [2,7,8].

To address the growing number of adults with pre-diabetes, the Diabetes Prevention Program (DPP), a large multi-center trial, tested the ability of lifestyle intervention and modest weight loss to delay the onset of diabetes [9]. In short, the study found that lifestyle intervention resulted in better outcomes when compared to medication such as reduced incidence of T2DM (58% lower than placebo), improved uptake of physical activity (74% at 24 weeks) and weight reduction (50% achieved 7% or greater weight loss) [9].

Since then, researchers at several institutions have adapted the lifestyle intervention using technology-enhanced mediums, thus eliminating intensive direct provider-to-patient contact while allowing patients to receive intervention materials asynchronously while also automating goal-setting and feedback loops [10,11]. A meta-analysis published in March 2017 evaluated the use of different content delivery channels among diabetes prevention programs to address what methods may be most effective in-patient engagement lending to better weight loss [12]. Those programs that used the original DPP lifestyle intervention or adapted from that content, when combined with multiple modalities, displayed greater average weight loss (2.4 kg) than those that didn’t follow the evidence-based curriculum [12,13].

While these results are very promising, there is a gap in the literature related to the effectiveness of these interventions for participants with varying degrees of health literacy [11,14–23]—defined as the ability to acquire, synthesize and understand health information and services required to make decisions regarding an individual or community’s health [24,25]. Only two studies in the Bian et. al March 2017 meta-analysis evaluated participant health literacy, but neither used the concept as a moderating variable to determine if there were differential outcomes for patients with lower levels of health literacy [26,27]. It has been suggested that patient comprehension of the education material in interventions should be assessed and this may be especially important for interactive technology approaches where patients do not get regular interaction with a health professional who could respond to questions and clarify as material is being introduced [28–31]. Specifically, little is known about the degree to which technologically facilitated interventions leads to similar information uptake of key learning objectives, especially among participants of different health literacy levels [31–33].

While one could hypothesize that the ability to clarify educational content or the ability to do teach-back or teach-to-goal—mechanisms to ensure comprehension of the content [34–37]—could lead to reduced effectiveness of interactive technology interventions, there are also some reasons to hypothesize the opposite. Technological approaches may have several advantages for patients with lower health literacy such as repeatability. Most interactive technology-based interventions allow participants to review, play back, or re-do intervention activities. Similarly, most use auditory rather than text-based information delivery with images that reduce the need to read content. Finally, when comparing these approaches to in-person, group settings, many interactive educational components of a video or telephone call, may require more active participation of patients [35].

With regards to information uptake relative to a specific modality, when patients receive information through an educational DVD, regardless of health condition, several studies have observed positive outcomes [38–40]. For example, in a group of sedentary older adults, a DVD-based intervention observed improvements in overall physical function [38]. Another program adapted their in-person weight loss intervention to be delivered via DVD and recorded an 83% completion rate of lessons and 6% average weight change at 12 months [41]. Furthermore, a DVD-mediated intervention was observed to have clinically-significant weight loss maintained 24 months after baseline, suggesting the potential of a DVD to initiate sustainable weight loss and behavior change [39]. When considering the use of a DVD delivery format compared to text-based information for patients with lower health literacy, coronary artery disease patients did not have significant worse clinical outcomes or health behaviors than their counterparts, suggesting a DVD can enhance knowledge retention, understanding of their condition, and how to best manage their health [40].

In contrast to technology-facilitated interventions, traditional patient education is primarily delivered through in-person and small group mechanisms. Participants or patients attending a small-group class have been observed to have positive results, as well. Researchers from Wake Forest University observed improved blood glucose, deceased insulin resistance, weight, and waist circumference in participants that had attended small-group class versus a standard care treatment group [41]. Those results parallel much of Seidel and colleagues achieved in their adapted group-based lifestyle diabetes prevention intervention in an urban, medically underserved neighborhood suggesting participants can engage in the core curriculum at a different pace and setting while being able to engage in behavior change lending favorable outcomes [42]. These advantages could enhance the ability of low health literacy participants to receive the information in more conducive manners due to the ability to interact with a trained medical professional, registered dietician or other class participants; however, largely uncertain is the degree to which participants can interact with the educational content to enhance their comprehension levels.

As we have documented, alternative hypotheses could be posed related to the benefits of DVD versus in-person, class-initiated diabetes prevention interventions for patients of varying health literacy levels. However, no research to date has compared the information uptake of key learning objectives when a diabetes prevention program is initiated with either a technology or in-person facilitated approach. The purpose of this study is to fill this gap by comparing the effectiveness of a DVD versus an in-person group-initiated diabetes prevention class to enhance patient comprehension of diabetes prevention program learning objectives based on healthy literacy status (i.e., high (HHL) & low health literacy (LHL)).

2. Methods

2.1. Research design

DiaBEAT-it! is an 18-month pragmatic hybrid-preference randomized controlled trial with primary aims to determine the reach, effectiveness, and cost of a technology-initiated diabetes prevention program when compared to an in-person initiated program and standard care diabetes prevention class [10]. The design allowed for participants to be initially assigned into one of two groups—choice of intervention or randomization into one of three conditions [10]. Participants in the randomized controlled trial (RCT; n = 334) were randomly assigned into one of three treatments—standard care (Class; n = 117), small-group in-person class-initiated intervention with interactive voice response follow-up (Class/IVR; n = 110) or DVD initiated intervention with interactive voice response follow-up (DVD/IVR; n = 107). Those assigned to the choice group (n = 264) could choose between the Class/IVR (n = 114) or the DVD/IVR conditions (n = 150) [10]. For the purposes of this study, participants were grouped according to the intervention received (Class/IVR= 224 or DVD/IVR = 257) independent of original group assignment (Choice vs. RCT) and were categorized as having adequate or high health literacy (HHL ≥ 4/6) versus possibly inadequate or low health literacy (LHL < 4/6) based on the validated Newest Vital Sign (NVS) health literacy assessment [43]. All participants were asked to complete an informed consent to participate at the baseline assessment. The study procedures were approved by the Carilion Clinic, Virginia Tech, and University of Nebraska Medical Center Institutional Review Boards and the protocol was registered at clinicaltrials.gov (NCT01262901).

2.2. Participant eligibility and recruitment

Participants were recruited through the Carilion Clinic Department of Family and Community Medicine in southwest Virginia. Patients over the age of 18 with a body mass index (BMI) greater than 25 were eligible to participate [10]. Patients with diabetes, that were pregnant or planning a pregnancy, those unable to read or communicate in English, or medically incapable were ineligible [10].

Initial baseline assessments (i.e. Height, weight, blood pressure, Dual X-Ray Absorptiometry, Health Literacy) were completed on the first study visit [10]. Health literacy levels were assessed via the validated Newest Vital Sign tool [10]. On the second study visit, the participants were assigned to or chose a program, and given educational materials to follow the design and objectives of the study [10].

2.3. Interventions

2.3.1. Small group diabetes prevention class (Class/IVR)

The in-person small group diabetes prevention class was offered twice a month lasting two hours and was led by a Carilion Clinic registered dietician [10]. As part of the curriculum, diabetes prevention objectives (i.e. appropriate physical activity, ideal food choices and portion sizes) were reviewed in addition to participants creating a personalized action plan to reduce weight by 10% over the course of twelve months [10]. The class was formulated to encourage discussion among participants on how to live a healthy lifestyle [10]. The in-person class session was followed by a teach-back/teach-to-goal (referred to as teach-back in the remainder of the article) call [36] that was intended to provide reinforcement for intervention learning objectives, review the personalized action plan, and prepare participants to receive follow-up IVR intervention calls [10].

2.3.2. DVD diabetes prevention intervention

A 60-minute DVD was designed to cover the same content and process of the in-person class session [10]. The DVD was a convenient form of media that can be reproduced at low cost while affording the participant the opportunity to watch the DVD multiple times, if necessary. Participants used the DVD to work through the development of an action plan to set their health behavior goals (i.e. physical activity, weight loss, fruits and vegetable consumption) and identify strategies and barriers to behavior change. About 4–5 days after watching the DVD, participants completed a teach-back call with a research assistant to review the action plan and reinforce the material presented in the DVD [10].

2.3.3. Teach back call

After attending the class or viewing the DVD, participants were asked to complete a teach-back call that included teach-to-goal opportunities and lasted 20 to 30 min. The call was designed to reinforce key learning objectives from the small group class or viewing of the DVD. A series of six questions were assessed using teach-back for each question to initiate the process of teach-to-goal to ensure information uptake. Question one asked participants to provide a description of factors that could help to prevent diabetes. Correct responses included reducing body weight, blood pressure, levels of LDL and triglycerides as well as increasing physical activity and healthful eating patterns. Each question had detailed responses that were used to determine if a participant answered correctly or not. Questions 2 through 6 focused on identifying the amount of weight loss necessary to reduce the risk of progressing into diabetes, the recommended amount and intensity of physical activity, appropriate resistance training activities, and the components of a MyPlate eating. Any question answered incorrectly was repeated for up to 3 rounds of assessment. After each round, participants reviewed components of their action plan.

Following previous studies in the literature, we selected three measures for assessing comprehension: teach back rounds completed, number of round one questions correct, and reverse score averages [34,44]. First, we calculated the number of teach-back rounds completed with fewer reflecting higher comprehension as a result of the DVD or Class. Second, we calculated the number of times each participant answered the questions correctly during the first round without need for further clarification. Scores ranged from zero to six with higher scores indicating better overall comprehension. Third, reverse scoring methods were applied by assigning a higher value for providing the correct answer in earlier rounds (i.e. Round 1 correct = 3, Round 2 correct = 2, Round 3 correct = 1, Incorrect in all 3 rounds = 0) and calculating a sum to gauge overall performance in the teach-back call (Tables 3–5). Scores ranged from zero to eighteen with higher scores indicating better comprehension and less overall rounds needed to complete all six questions. For instance, a score of 18 indicated a participant needed 6 overall rounds (responded every question correctly in the first round) to answer all six questions, a score of 17 indicated a participant needed 7 overall rounds, a score of 16 indicated a participant needed 8 overall rounds and so forth.

Table 3.

Live call reverse score.

	Coefficient	SE	t
Constant**	18.2734	0.5806	31.4758
Choice vs. RCT	0.692	0.2266	0.3055
Modality and health literacy status	−0.1292	0.1255	−1.0297
Age	−0.0440	0.0095	−4.6259
Days between viewing or attending class	−0.0847	0.0302	−2.8047
R2 = 0.0948, F (4, 424) = 7.0218
Constant**	18.2383	0.5232	34.8591
Choice vs. RCT	0.1011	0.2231	0.4534
Class vs. DVD	−0.8402	0.2259	−3.7186
Age	−0.0405	0.0093	−4.3413
Days between viewing or attending class	−0.0807	0.0292	−2.7619
R2 = 0.1204, F (4, 424) = 10.4998
Constant**	15.6880	0.6654	23.5765
Choice vs. RCT	0.0160	0.2171	0.0738
Health literacy levels	1.9647	0.3520	5.5813
Age	−0.0314	0.0092	−3.4027
Days between viewing or attending class	−0.0631	0.0310	−2.0351
R2 = 0.1790, F (4, 424) = 15.0776
Constant**	14.1757	1.6777	8.4829
Choice vs. RCT	−0.01216	0.3315	−0.3666
Class/LHL vs. Class/HHL	1.2151	0.4167	2.9156
Age	−0.0442	0.0156	−2.8378
Days between viewing or attending class	−0.0572	0.0463	−1.2343
R2 = 0.1263, F (4, 208) = 6.2971
Constant**	14.6050	0.8257	17.6884
Choice vs. RCT	0.1399	0.2591	0.5398
DVD/LHL vs. DVD/HHL	2.6952	0.5872	4.5900
Age	−0.0152	0.0103	−1.4738
Days between viewing or attending class	−0.0632	0.0509	−1.2429
R2 = 0.2649, F (4, 211) = 7.0037
Constant	15.0278	1.8814	7.9876
Choice vs. RCT	−0.03483	0.6665	−0.5226
Class/LHL vs. DVD/LHL	0.1028	0.3296	0.3119
Age	−0.0289	0.0286	−1.0107
Days between viewing or attending class	0.0155	0.0492	0.3151
R2 = 0.0205, F (4, 72) = .8386
Constant**	18.7002	0.5076	36.8381
Choice vs. RCT	0.0632	0.2143	0.2949
Class/HHL vs. DVD/HHL	−0.5225	0.1099	−4.7526
Age	−0.0280	0.0091	−3.0801
Days between viewing or attending class	−0.1142	0.0409	−2.7952
R2 = 0.1681, F (4, 347) = 10.7243
Constant**	18.0633	.5236	34.4964
Choice vs. RCT	.0517	.2286	.2261
Age	−0.0443	.0095	−4.6599
Days between viewing or attending class	−0.0844	.0306	−2.7578
R2 = 0.0918, F (3, 425) = 8.9730

^**p < 0.001,

^{^}p < 0.05.

Table 5.

Number of round 1 questions correct.

	Coefficient	SE	t
Constant**	5.6543	0.3961	14.2768
Choice vs. RCT	0.0279	0.1621	0.1722
Modality and health literacy status	−0.1256	0.0844	−1.4875
Age	−0.0255	0.0065	−3.9153
Days between viewing or attending class	−0.0613	0.0204	−3.0111
R2 = 0.0864, F (4, 424) = 6.4029
Constant**	5.5872	0.3624	15.4170
Choice vs. RCT	0.0496	0.1594	0.3112
Class vs. DVD	−0.6581	0.1615	−4.0741
Age	−0.0228	0.0064	−3.5702
Days between viewing or attending class	−0.0581	0.0200	−2.9082
R2 = .1160, F (4, 424) = 10.5822
Constant**	3.8304	0.4463	8.5819
Choice vs. RCT	−0.0135	0.1555	−0.0866
Health literacy levels	1.3397	0.2160	6.2015
Age	−0.0170	0.0064	−2.6476
Days between viewing or attending class	−0.0465	0.0218	−2.1310
R = 0.1624, F (4, 424)= 15.7430
Constant**	2.0387	1.1501	1.7726
Choice vs. RCT	0.0157	0.2323	0.0675
Class/LHL vs. Class/HHL	1.0052	0.2756	3.6468
Age	−0.0234	0.0104	−2.2357
Days between viewing or attending class	−0.0349	0.0296	−1.1822
R = 0.1163, F (4, 208) = 7.1254
Constant**	3.3070	0.5418	6.1043
Choice vs. RCT	−0.0438	0.1942	−0.2257
DVD/LHL vs. DVD/HHL	1.6782	0.3493	4.8049
Age	−0.0054	0.0076	−0.7115
Days between viewing or attending class	−0.0514	0.0377	−1.3620
R = 0.2275, F (4, 211) = 6.7917
Constant	2.8861	0.9948	2.9011
Choice vs. RCT	−0.1131	0.4146	−0.2727
Class/LHL vs. DVD/LHL	−0.1291	0.2113	−0.6111
Age	−0.0048	0.0155	−0.3094
Days between viewing or attending class	0.0133	0.0348	0.3817
R = 0.0157, F (4, 72) = .1843
Constant**	5.9893	0.3697	16.2022
Choice vs. RCT	−0.0043	0.1618	−0.0267
Class/HHL vs. DVD/HHL	−0.3699	0.0819	−4.5138
Age	−0.0157	0.0066	−2.3790
Days between viewing or attending class	−0.0844	0.0214	−3.9494
R = 0.1501, F (4, 347)= 11.4096
Constant**	5.4502	.3669	14.8528
Choice vs. RCT	.0109	.1625	.0668
Age	−.0258	.0065	−3.9550
Days between viewing or attending class	−.610	.0208	−2.9296
R = 0.0805, F (3, 425) = 7.4941

^**p < 0.001,

^{^}p < 0.05.

2.4. Data analysis

All participants that completed a teach back call were analyzed according to the intervention they selected or were randomized to, as well as their performance on NVS health literacy assessment (i.e. HHL or LHL). Descriptive statistics were computed for age, height, weight, BMI, income, sex, and insurance status. Comparisons using multiple linear regression techniques controlling for age, initial randomization for choice or RCT, days between viewing DVD or attending class and completing the teach-back call were conducted to determine the relationships between intervention condition, health literacy status, and comprehension. To control for hetero-skedasticity, White’s Robust Standard Errors adjustment procedures were calculated for number of round one questions correct, teach back rounds completed, and reverse score averages as a measure of overall performance to evaluate models by modality, health literacy level, as well as modality + health literacy level. The general regression model was $\text{y}={\beta }_0+{\beta }_1{\chi }_{\text{i}}+{\beta }_2{\chi }_{\text{i}}{}^2+{\beta }_3{\chi }_{\text{i}}{}^3+{\beta }_4{\chi }_{\text{i}}{}^4+{\epsilon }_{\text{i}}$ where i = 1, … n and class was coded as 1, DVD = 0, HHL= 1 and LHL = 0. The round to which all questions were completed answered correctly were analyzed by treatment groups using chi-square procedures (Table 6). All calculations were performed using IBM SPSS Statistics Version 23.0.

Table 6.

Round all questions answered correctly by treatment and HL group.

	Round 1		Round 2		Round 3		Didn’t get in any of the 3 rounds
	N	% Correct	N	% Correct	N	% Correct	N	% who missed
LHL1	5	6.2	37	45.7	25	30.9	14	17.3
HHL1	84	23.3	212	58.7	42	11.6	23	6.4
DVD2	54	22.9	141	59.7	27	11.4	14	5.9
Class2	35	17.0	108	52.4	40	19.4	23	11.2
DVD + LHL3,4	4	10.5	16	42.1	11	28.9	7	18.4
DVD+HHL3,4,6	47	25.1	120	64.2	14	7.5	6	3.2
Class + LHL3,5,	1	2.3	21	48.8	14	32.6	7	16.3
Class+HHL3,5,6	37	21.3	92	52.9	28	16.1	17	9.8

^{1, 3, 4}p < 0.001,

^{2, 5, 6}p < 0.05.

3. Results

Of 481 eligible participants, 442 (92%) completed a teach-back call with 225 (47%) and 217 (45%) receiving the DVD and class session, respectively. The average age of the entire sample was 52.3 years (±12.1) and 68% were female. DVD (50.8 ± 12.2 years) and class (53.9 ± 11.9 years) samples differed significantly on age. Over three quarters of the sample were Caucasian and 17% were African-American. Eighteen percent (n = 81) of the participants had low health literacy, conversely, 82% of the participants (n = 361) had adequate or high health literacy based on the Newest Vital Sign scores. Overall, 20% of those who chose or were assigned the DVD (n = 40) had LHL, comparable to the other LHL participants in the class treatment at 17% (n = 41). Participants with lower health literacy were significantly older (57.1 ± 11.9) than those with higher health literacy (51.2 ±11.9) and significantly more likely to be African-American (30%) when compared to other racial categories (14%). Finally, the duration between watching the DVD (4.3 ± 7.0 days) or attending the class (5.0 ± 6.0 days) and completing the teach-back call was not significantly different between groups. Table 1 contains descriptive information by health literacy level, modality and modality/health literacy level.

Table 1.

Participant Characteristics.

	RCT						Choice						Overall
	n = 198						n = 244						n = 442
	Class/IVR			DVD/IVR			Class/IVR			DVD/IVR			LHL	HHL	Class/IVR	DVD/IVR	Choice	RCT
	Overall	LHL	HHL	Overall	LHL	HHL	Overall	LHL	HHL	Overall	LHL	HHL
	n = 104 μ (SD)	n = 21 μ (SD)	n = 83 μ (SD)	n = 94 μ (SD)	n = 15 μ (SD)	n = 79 μ (SD)	n = 113 μ (SD)	n = 22 μ (SD)	n = 91 μ (SD)	n = 131 μ (SD)	n = 23 μ (SD)	n = 108 μ (SD)	n = 81 μ (SD)	n = 361 μ (SD)	n = 217 μ (SD)	n = 225 μ (SD)	n = 244 μ (SD)	n = 198 μ (SD)
Agea, b, g, j, q	52.3 (12.2)	58.5 (10.4)	50.7 (11.9)	51.7 (12.2)	51.3 (15.7)	52.1 (11.7)	55.8 (11.6)	60.7 (9.3)	54.1 (11.7)	49.9 (11.9)	56.0 (11.8)	48.6 (11.7)	57.1 (11.9)	51.2 (11.9)	50.8 (12.2)	53.9 (11.9)	52.5 (12.1)	52.1 (12.1)
Weightc, d, r	231.3 (45.5)	218.5 (35.1)	234.7 (49.8)	239.8 (56.5)	243.2 (54.1)	239.5 (55.8)	220.3 (41.1)	212.2 (50.4)	222.4 (38.6)	226.3 (37.5)	220.5 (36.1)	227.7 (37.6)	221.9 (44.3)	230.6 (45.5)	232.2 (46.0)	225.7 (44.5)	223.7 (39.2)	235.5 (51.4)
BMIe, s	37.8 (7.9)	35.6 (4.7)	38.2 (8.4)	38.4 (7.7)	37.4 (7.6)	38.7 (7.7)	35.8 (6.1)	34.7 (5.9)	36.1 (6.1)	36.4 (5.0)	36.6 (5.0)	36.4 (5.0)	36.0 (5.7)	37.2 (6.9)	37.3 (6.3)	36.7 (7.0)	36.1 (5.5)	38.1 (7.8)
	n (%)	n (%)	n (%)	n (%)	n (%)	n (%)	n (%)	n (%)	n (%)	n (%)	n (%)	n (%)	n (%)	n (%)	n (%)	n (%)	n (%)	n (%)
% Female	67 (67.7)	12 (12.1)	57 (57.6)	65 (65.7)	9(9.1)	54 (54.5)	74 (69.2)	16 (15)	62 (57.9)	94 (68.6)	17 (12.4)	73 (53.2)	54 (66.7)	246 (68.1)	153.0 (68.0)	147.0 (67.7)	168 (68.9)	132 (66.7)
% Minori-iesh, k, n	21 (21.2)	4 (4.0)	17 (17.2)	20 (20.2)	7 (7.1)	13 (13.1)	17 (15.9)	7 (6.5)	12 (11.2)	24 (17.5)	7(5.1)	15 (10.9)	25 (30.8)	57 (15.8)	42 (18.7)	40 (18.5)	41 (16.8)	41 (20.7)
% Uninsuredf, l, o,	1 (1.0)	1 (1.0)	0 (0)	2 (2.0)	1 (1.0)	1 (1.0)	3 (2.8)	2(1.9)	2 (1.9)	7 (5.1)	3 (2.2)	3 (2.2)	7 (8.6)	6 (1.7)	8 (3.6)	5 (2.3)	10 (4.1)	3 (1.5)
% low-incomei, m, p	19 (19.2)	4 (4.0)	16 (16.2)	22 (23.2)	7 (7.1)	15 (15.2)	26 (24.2)	10 (9.3)	17 (15.9)	31 (22.6)	10 (7.3)	20 (14.6)	31 (38.3)	68 (18.8)	52 (23.1)	47 (21.7)	57 (23.4)	42 (21.3)

^aChoice only, p < 0.001.

^b-fAll 4 treatment arms, (b) p < 0.01; (c and d) p < 0.05; (e) p < 0.001; (f) p < 0.001.

^{g, h}LHL v. HHL within Choice, (g) p < 0.05; (h) p < 0.05.

^i-lLHL v. HHL within RCT, (i) p < 0.001; (j) p < 0.01; (k) p < 0.05; (l) p < 0.001.

^m-pLHL v. HHL all treatments, (m) p < 0.001; (n) p < 0.001; (o) p < 0.01; (p) p < 0.001.

^qClass v. DVD, p < 0.01;

^{r, s}Choice v. RCT, (r) p < 0.01; (s) p < 0.01.

When considering participants who completed the intervention via the DVD versus class we found that there were significant differences in the reverse score performance (DVD-15.4 ± 2.5; Class-14.8 ± 2.6; F(3,425) = 13.72, p < 0.001), number of teach-back rounds (DVD-1.9 ± 0.7; Class-2.1 ±0.7; F(3, 425) = 5.98, p < 0.001) and number of round 1 questions correct(DVD-4.2 ± 1.6; Class-3.4 ± 1.8; F(3425) = 20.95, p < 0.001) (See Table 2). Based on health literacy level we found consistently that participants with HHL performed better across the outcomes (See Table 6). Finally, when considering intervention modality by health literacy status we found that the DVD delivery resulted in superior comprehension for HHL participants across all outcomes. However, DVD versus class differences for participants with LHL were not significant and approximately, 18% and 16% of DVD and class LHL participants did not achieve the teach-to-goal purpose after the final round of teach-back was completed (Table 6). In the analysis of teach-back rounds, number of round 1 questions and reverse score performance, every predictor variable mentioned above was significant except for class/ LHL vs. DVD/LHL (Tables 3–5).

Table 2.

Mean Comprehension Outcome Scores.

	Overall	DVD	Class	DVD		Class
				High health literacy	Low health literacy	High health literacy	Low health literacy
Reverse score average	15.4 (2.5)	15.9 (2.3)	14.8 (2.6)	16.3 (1.7)	13.5 (3.3)	15.2 (2.5)	13.3 (2.6)
Number of round 1 questions correct	3.8 (1.7)	4.2 (1.6)	3.4 (1.8)	4.5 (1.4)	2.8 (1.9)	3.7 (1.7)	2.4 (1.4)
Number of teach-back rounds	2.0 (0.7)	1.9 (0.7)	2.1 (0.7)	1.8 (0.6)	2.3 (0.8)	2.0 (0.7)	2.4 (0.6)

^*Note: Significance between modalities or health literacy level addressed by Table 3–6.

4. Discussion and conclusion

4.1. Discussion

The need for interventions that include strategies to address participants with varied levels of health literacy is well documented [28,30]. Consistent with other research, our study found that even when information is presented using clear communication strategies during an initial intervention session, it may not be enough to ensure information uptake, in our case, related to diabetes prevention objectives. In fact, less than 21% of all participants were able to demonstrate complete comprehension of the materials during the first round of questioning, indicating the importance of additional rounds of material reinforcement even for individuals with higher health literacy [45–47]. Also, like previous research, outside of the context of a diabetes prevention intervention, when information uptake is evaluated, researchers have observed improved comprehension over multiple rounds of teach-to-goal educational assessment [45–47].

In the review by Bian et al. 2017, the use of multiple health education modalities relative to single health education modality interventions to deliver diabetes prevention lessons was observed to lead to greater participant weight loss [12]. However, none of the multiple modality interventions reviewed measured information uptake through teach-back or teach-to-goal or used an initial teach-back call to evaluate uptake of key learning objectives [11,19,23,26,48]. The lack of strategies focusing on enhanced information uptake may help explain the levels of attrition (38%−57%) and variability in weight outcomes across studies [12,50–52].

When teach-back and teach-to-goal methods are utilized, positive outcomes have been observed [46,47,49]. For example, in executing an informed consent procedure with teach-to-goal strategies, proportions of marginal and inadequate health literacy participants were nearly equivalent after two rounds of assessment [47]. In an asthma administration education program by providers tailored towards low health literacy patients, 59, 21, and 10 percent of patients needed one, two, or three additional rounds of teach-to-goal education, respectively [47]. The latter study has suggested that through increasing information uptake, patient engagement may be more likely through enhanced self-efficacy of the behavior lending to a greater likelihood of behavioral uptake and health outcome achievement [30,45]. While reporting on the relationship between teach-back strategies and health outcomes is beyond the scope of this paper, our results support the importance of multiple opportunities for presenting health information to individuals, regardless of health literacy levels. Indeed, the initial 21 percent of participants that had achieved the learning objectives as demonstrated by the first teach back opportunity, grew to over 90 percent demonstrating this achievement by the completion of the third round of teach back.

Perhaps our most interesting and actionable finding was that the DVD initiated diabetes prevention intervention was superior to supporting patient uptake of information when compared to the in-person initiated version. It is not clear why this might be, but as we proposed earlier, it is possible that the DVD gave participants multiple opportunities to review the material over time. Similar to the Paasche-Orlow et al. [45] and Sudore et. al [47]. studies, the difference between modalities was reduced over time as a result of the teach-to-goal strategies used. A fruitful area for additional research would be to determine the potential mechanism that underlies the superiority of the DVD or other interactive technology-based interventions when compared to in-person sessions.

The primary limitations of our study include the short duration and the lack of health or behavioral outcomes associated with learning objective comprehension. It is unlikely that simply providing a DVD or in-person session would lead to sustained changes in behavior, weight, and diabetes risk. However, as part of a larger trial and intervention, our findings may be generalizable to other contexts and health promotion outcomes—we demonstrated that the DVD approach could improve initial information uptake and that the use of a teach-back and teach-to-goal strategy can be used to reinforce key learning objectives. An additional possible explanation for our findings could be that in-person class sessions had variable implementation fidelity which could influence the results.

4.2. Conclusion

The use of a DVD may produce superior uptake of learning objectives when compared to an in-person class and participants with LHL typically perform worse on assessment of information uptake regardless of implementation modality. Nevertheless, we identified that a teach-back call may enhance information uptake of diabetes prevention learning objectives in diabetes prevention programs, especially among participants with lower health literacy. Finally, many of the participants with higher health literacy were able to improve comprehension through the reinforcing structure of the teach-back and teach-to-goal call. Teach-back strategies may be important components to be considered for future diabetes prevention programs, independent of delivery method, to ensure participants, independent of health literacy level, fully comprehend the materials and learning objectives being covered.

4.3. Practice implications

A teach-back call has many practical implications—easy-to-complete, pragmatic, efficient, and it may enhance a provider’s ability to help a patient comprehend important information related to health behaviors needed to prevent the onset of T2DM. Furthermore, a teach-back call may enhance engagement of participants in diabetes prevention interventions, especially LHL members, due to greater information understanding, thus improving the likelihood of health behavior uptake. As such, clinical interventions may observe a greater proportion of the patient population achieving the primary or secondary outcomes of weight loss or improvements in preventive behaviors such as better nutrition or more physical activity.

Table 4.

Teach-back rounds.

	Coefficient	SE	t
Constant^	1.5278	0.1477	10.3442
Choice vs. RCT	0.0253	0.0650	0.3890
Modality and health literacy status	0.0267	0.0325	0.8234
Age	0.0071	0.0025	2.8767
Days between viewing or attending class	0.0090	0.0060	1.4843
R2 = 0.0245, F (4, 424) = 2.8009
Constant^	1.5352	0.1370	11.2064
Choice vs. RCT	0.0187	0.0645	0.2905
Class vs. DVD	0.1731	0.0652	2.6542
Age	0.0063	0.0025	2.5812
Days between viewing or attending class	0.0081	0.0058	1.3995
R2 = 0.0390, F (4, 424) = 4.5016
Constant**	2.0595	0.1699	12.1187
Choice vs. RCT	0.0363	0.0633	0.5731
Health literacy levels	−0.4038	0.0875	−4.6140
Age	0.0045	0.0024	1.8550
Days between viewing or attending class	0.0045	0.0063	0.7206
R2 = 0.0720, F (4, 424) = 8.0491
Constant^	2.7948	0.4456	6.2719
Choice vs. RCT	0.0154	0.0937	0.1641
Class/LHL vs. Class/HHL	−0.3445	0.1175	−2.9326
Age	0.0045	0.0035	1.2805
Days between viewing or attending class	0.0040	0.0082	0.4845
R2 = 0.0587, F (4, 208) = 3.9194
Constant^	2.1042	0.2310	9.1076
Choice vs. RCT	0.0435	0.0873	0.4991
DVD/LHL vs. DVD/HHL	−0.4635	0.1377	−3.3671
Age	0.0029	0.0035	0.8450
Days between viewing or attending class	0.0040	0.0101	0.3934
R2 = 0.0809, F (4, 211) = 3.3813
Constant	2.1138	0.3585	5.8955
Choice vs. RCT	0.1109	0.1611	0.6884
Class/LHL vs. DVD/LHL	0.0463	0.0830	0.5577
Age	0.0042	0.0059	0.7094
Days between viewing or attending class	−0.0119	0.0086	−1.3826
R2 = 0.0490, F (4, 72) = .5885
Constant^	1.4807	0.1558	9.5009
Choice vs. RCT	0.0253	0.0687	0.3691
Class/HHL vs. DVD/HHL	0.0931	0.0348	2.6789
Age	0.0035	0.0026	1.3388
Days between viewing or attending class	0.0151	0.0087	1.7323
R2 = 0.0430, F (4, 347) = 3.3393
Constant^	1.5713	.1369	11.4745
Choice vs. RCT	.0289	.0649	.4462
Age	.0071	.0024	2.9198
Days between viewing or attending class	.0089	.0061	1.4534
R2 = 0.0227, F (3, 425) = 3.4785

^**p < 0.001,

^{^}p < 0.05.