Using Peer Support to Aid in Prevention and Treatment in Prediabetes (UPSTART)

Kaiser Foundation Research Institute, National Institute of Diabetes and Digestive and Kidney Diseases (NIDDK)

An estimated 86 million adults in the United States have prediabetes, and low-income Latino and African American adults have disproportionately high rates compared to non-Hispanic adults. Structured lifestyle interventions can prevent or delay type 2 diabetes in these at-risk populations and now are widely offered at community organizations and health systems. Yet, uptake of and engagement in available formal programs is very low. Low-income adults in particular face multiple barriers to navigating, engaging in, and sustaining involvement in available programs and lifestyle behaviors found to decrease progression to diabetes. It is critically important to develop and evaluate innovative approaches to increase uptake, engagement, and maintenance of gains in diabetes prevention activities. Peer support has been shown in the investigators' and others' effectiveness trials to be a sustainable, effective approach for positive behavior change and improved outcomes in adults with diabetes and other chronic conditions. The study team's pilot work suggests such approaches are feasible and acceptable among low-income Latino and African American patients with prediabetes to prevent chronic disease and better navigate their health care systems to obtain healthy lifestyle counseling and support. However, such peer support models among Latino, African American, and other low-income adults with prediabetes have not yet been rigorously evaluated. Accordingly, the study will conduct a parallel, two-arm randomized controlled trial in primary care centers in two different health systems that serve multi-ethnic communities with a high concentration of Latinos and African Americans and diverse socio-economic backgrounds. The study will compare enhanced usual care (providing referrals to diabetes prevention programs and resources) with a model of a structured behavioral change intervention supplementing enhanced referral to programs and resources with peer support to help link adults with prediabetes to existing health system and community diabetes prevention programs, to support their engagement in formal programs, maintain achieved gains, and support participants to initiate and sustain healthy behaviors to prevent diabetes.

The intervention seeks to address the need to test in routine primary care evidence-based approaches to increase uptake, engagement, and maintenance of healthy behaviors necessary to decrease progression to diabetes among primary care patients with prediabetes, especially low-income and racial and ethnic minority adults with prediabetes. The study will conduct a parallel, two-armed, randomized controlled pragmatic clinical trial including adults with prediabetes at two primary care centers in two different health systems: Kaiser Permanente Northern California and the University of Michigan Health System. Each health system received IRB approval from its own IRB to accommodate site-specific nuances in its implementation. The trial will evaluate whether adding a 12-month predominantly telephone-based volunteer peer support program (UPSTART) to health care provider counseling and referral to diabetes prevention programs leads to greater improvements in A1c, the weight loss, and waist circumference than health care provider counseling and referral alone (Aim 1). The study will also compare differences in reported physical activity, diet, and enrollment and engagement in diabetes prevention programs as potential mediators as well as autonomous motivation, behavior-specific self-efficacy, patient activation, and perceived support; and moderators such as health literacy (Aim 2). To enhance adoption of the intervention by the two study health systems and dissemination to other health systems if effective, the study will evaluate costs and use an integrated RE-AIM and Consolidated Framework for Implementation Research (CFIR) framework to evaluate processes of intervention implementation in the two primary care settings (Aim 3). The study duration will be 5 years, to allow for peer supporter and patient recruitment, completion of the 12-month program, and assessment of outcomes at 6 months and at 12 months. The study will use mixed methods-i.e., the collection, analysis, and combining of both quantitative and qualitative data-to investigate elements important for implementation and dissemination. The study will gather data on how peer supporters, primary care clinic staff, and patients experience the intervention and how the experiences of participants together with the trial's results suggest the study should modify the intervention. Using this approach, the study aim to ensure that the intervention has the greatest possible likelihood of adoption in both UMHS and KPNC health systems should the study find it has positive effects on processes and outcomes of care. The UPSTART intervention is designed for peers to interact and provide support in a way that is autonomy supportive with the goal of enhancing autonomous motivation and self-efficacy for healthy behaviors. The study will hold periodic initial trainings for new peer supporters over the study period. In addition, peer supporters will receive: 1) routine, structured check-ins that will include monthly peer support group meetings (with option to call in) to allow exchange; 2) back-up support: offering peer supporters contact information for staff who they can call; and 3) continuing education and booster training at the monthly group meetings to enrich their skill sets and knowledge. The peer supporters were included in the total number of participant numbers for the trial and in the intervention arm in the study flow portion of this record because they were consented. Demographic characteristics of the peer supporters are also reported. However, they were not randomized; they were recruited to deliver the intervention to recruited participants with prediabetes whose outcomes we aimed to improve through the delivered intervention. The peer supporters' involvement was to deliver the intervention to the participants who were randomized to either receive the peer support intervention or to enhanced usual care. The study consented the peer supporters because they completed surveys at three time points; the intent was to look at whether delivering the intervention affected the peer supporters' own reported behaviors as exploratory only. There was and is no intention to compare coach data to primary, secondary, or any other outcome measures of the randomized participants in the trial. During the trial, another primary care center within the University of Michigan Health System was added. Because this center was within the University of Michigan Health System, and all contact information was identical, there was no reason to add another participating site.

Participants randomized to enhanced usual care will receive brief education and folder of information and resources.

Participants will be paired with a peer supporter. They will meet in person one time and then be in contact by phone or texting on a weekly basis for the first 6 months of a year. During the final six months of the year, the peer supporter and participant will be in touch at least monthly. The peer supporter will help link participants to existing health system and community diabetes prevention programs, support their engagement in formal programs, maintain achieved gains, and support participants to initiate and sustain healthy behaviors to prevent diabetes. The 95 peer supporters who consented to participate in the study are included in this arm, along with the participants who were randomized to this arm, in the participant flow. The peer supporters were not randomized to the arm; rather, they consented to deliver the intervention. Although 95 consented, only 68 met with at least one peer, the first step in intervention delivery.

An A1C test measures the percentage of red blood cells that have glucose-coated hemoglobin. This percentage indicates average blood sugar levels over the last 2-3 months. Higher percentages indicate higher levels of blood sugar levels, and an A1c between 5.7% and 6.4% is considered in the prediabetes range. The negative numbers for change over time indicate an improvement because they show a drop in percentages from the baseline percentages.

An A1C test measures the percentage of red blood cells that have glucose-coated hemoglobin. This percentage indicates average blood sugar levels over the last 2-3 months. Higher percentages indicate higher levels of blood sugar levels, and an A1c between 5.7% and 6.4% is considered in the prediabetes range. The negative numbers for change over time indicate an improvement because they show a drop in percentages from the baseline percentages.

Change in Frequency and Duration Participant Engages in Moderate to Vigorous Physical Activity as Shown by Minutes Per Week

This combined measure is based on participant-reported frequency and duration of moderate to vigorous physical activity, measured in minutes per week. Frequency and duration were initially registered separately. Two questions were used to calculated minutes of physical activity per week, asked in both the baseline and 12 months surveys: Q1. On average, how many days per week do you engage in moderate-to-vigorous physical activity? A1: 0-7, integer response only Q2: On average on those days, how many minutes do you engage in physical activity as this level? A2: open numeric response A1*A2 = minutes per week (Physical Activity value) Minutes of physical activity per week reported in 12 month survey - minutes of physical activity per week in baseline survey = change in minutes per week of physical activity

Recorded frequency of consumption in the last 30 days. Participants could report times per day, times per week, or times over past 30 days. These reports were converted to the number of times over past 30 days participants drank regular soda or pop that contained sugar.

PAM-13 is a 13-item scale that measures participant beliefs, perceived knowledge, and confidence for engaging in behaviors related to the participant's health condition. PAM-13 has a 0-100 scale, where a low score indicates low activation (disengaged and overwhelmed) and a high score indicates high activation (patient considers self their own advocate). The difference between the score at 12 months and the score at baseline was calculated.

Change in Participant's Perceived Confidence in Their Ability to Take Steps to Prevent Diabetes Using the Williams Perceived Competence Scale

Williams Perceived Competence Scale comprises 4 items measuring the participants' perception of their own ability to take steps to prevent diabetes. Responses for each individual question range from 1 (not at all true) to 7 (very true). The responses to all 4 items were averaged (range: 1-7), with higher scores on the scale indicating higher perceived competence in preventing diabetes.

Change in Autonomous Motivation to Prevent Diabetes Using the Treatment Self-Regulation Questionnaire (TSRQ - Autonomous)

The Treatment Self-Regulation Questionnaire (TSRQ) consisted of 11 items that measured the participants' perceptions of how true various reasons to take steps to help prevent diabetes may be to the participants. The TSRQ was split into two measures: autonomous and controlled. The autonomous measure consisted of 4 items that measured participants' autonomous (or internal) perceptions of how true various reasons to take steps to help prevent diabetes may be to the participants. Responses for each individual question range from 1 (not at all true) to 7 (very true). The responses to all 4-items are averaged (range: 1-7). Higher scores indicate participants agree with more internal reasons for preventing diabetes (ex: wanting to take responsibility for my own health).

Change in the Participant's Level of Social Support Related to Improving Their Own Health Behaviors Using the Change in Health-Specific Social Support Scale

The Health-Specific Social Support Scale consisted of 12 items that asked participants to indicate their level of agreement on a 6-point Likert scale ("Strongly disagree" [1] to "Strongly agree" [6]) with statements regarding accessibility of any contact they had with people in the past 3 months who were specifically also at risk for developing diabetes and could provide social support in attempts to prevent diabetes via healthy lifestyle changes. Scores for negative statements were reversed, and the total score was the sum of all scores. Total scores can range from 12 (minimum) to 72 (maximum). A higher score indicated a better outcome because it indicated a higher level of reported social support.

The Role a Participant's Peer Supporter Played in Assisting Them to Set and Reach Their Goals Using the Goal Setting and Intervention Tailoring Subscale of the Patient Assessment of Chronic Illness Care (PACIC) at 6 Months

The goal setting and intervention tailoring subscale of the Patient Assessment of Chronic Illness Care (PACIC) consisted of 5 questions that ask the participant to state how often their peer supporter engaged in supportive behaviors regarding setting and achieving goals in order to make healthy lifestyle changes. Each question had a range of 1 (None of the time) to 5 (Always). Averaged scores range from 1 (minimum) to 5 (maximum). A higher score indicated a better outcome because it indicated that the peer perceived their partner to be more supportive.

The Role a Participant's Peer Supporter Played in Assisting Them to Set and Reach Their Goals Using the Goal Setting and Intervention Tailoring Subscale of the Patient Assessment of Chronic Illness Care (PACIC) at 12 Months

The goal setting and intervention tailoring subscale of the Patient Assessment of Chronic Illness Care (PACIC) consisted of 5 questions that ask the participant to state how often their peer supporter engaged in supportive behaviors regarding setting and achieving goals in order to make healthy lifestyle changes. Each question had a range of 1 (None of the time) to 5 (Always). Averaged scores range from 1 (minimum) to 5 (maximum). A higher score indicated a better outcome because it indicated that the peer perceived their partner to be more supportive.

Participant's Perceived Autonomy Support From Their Peer Supporter Using the Health Care Climate Questionnaire (HCCQ, Long Form) at 6 Months

The Health Care Climate Questionnaire (HCCQ, long form) consisted of 15 items that asked the participant to rank their agreement with statements that indicated degree of autonomy support (Strongly Disagree [1] to Strongly Agree [7]). Averaged scores ranged from 1 (minimum) to 7 (maximum). A higher the score indicated a better outcome because it indicated that the peer perceived their partner to be more autonomy supportive.

Participant's Perceived Autonomy Support From Their Peer Supporter Using the Health Care Climate Questionnaire (HCCQ, Long Form) at 12 Months

The Health Care Climate Questionnaire (HCCQ, long form) consisted of 15 items that asked the participants to rank their agreement with statements that indicated degree of autonomy support (Strongly Disagree [1] to Strongly Agree [7]). Total scores ranged from 1 (minimum) to 7 (maximum). A higher score indicated a better outcome because it indicated that the peer perceived their partner to be more autonomy supportive.

Recorded frequency of consumption in the last 30 days. Participants could report times per day, times per week, or times over past 30 days. These reports were converted to the number of times over past 30 days participants drank sugary drinks besides pop or soda.

Recorded frequency of consumption in the last 30 days - participants could report times per day, times per week, or times over past 30 days, which was converted to times over past 30 days

One measure that recorded frequency of consumption in the last 30 days - participants could report times per day, times per week, or times over past 30 days, which was converted to times over past 30 days

One measure that recorded frequency of consumption in the last 30 days - participants could report times per day, times per week, or times over past 30 days, which was converted to times over past 30 days

One measure that recorded frequency of consumption in the last 30 days - participants could report times per day, times per week, or times over past 30 days, which was converted to times over past 30 days

One measure that recorded frequency of consumption in the last 30 days - participants could report times per day, times per week, or times over past 30 days, which was converted to times over past 30 days

One measure that recorded frequency of consumption in the last 30 days - participants could report times per day, times per week, or times over past 30 days, which was converted to times over past 30 days

Change in Controlled Motivation to Prevent Diabetes Using the Treatment Self-Regulation Questionnaire (TSRQ - Controlled)

The Treatment Self-Regulation Questionnaire (TSRQ) consists of 11 items that measure participants' perceptions of how true various reasons to take steps to help prevent diabetes may be to the participants. The TSRQ is split into two measures: autonomous and controlled. The controlled measure consists of 7 items that measure participants' controlled (or external) perceptions of how true various reasons to take steps to help prevent diabetes may be to the participants. Responses for each individual question range from 1 (not at all true) to 7 (very true). The responses to all 7-items are averaged (range: 1-7). Higher scores indicate participants agree with more external behaviors or reasons for preventing diabetes (ex: feeling pressure to do so from others).

Inclusion Criteria: Peer supporter: prior A1c 5.7 - 6.4 in last 3 years, with most recent being either lass than 5.7 or 0.4 points less than prior; AND/OR prior BMI >=25 or >=23 if Asian American in last three years with most recent 2% lower body weight;AND/OR prior completion of a wellness or diabetes prevention program; AND/OR completion of initial 6 months of UPSTART intervention as participant. Patient participant: no prior dx of DM or current use of anti-hyperglycemic medication; BMI >=25 m2/kg or >=23 if Asian; and A1c 5.7 - 6.4. Exclusion Criteria: serious psychiatric disorder in past 24 months pregnant or planning pregnancy substance abuse in past 12 months diagnosis of dementia, bipolar, schizophrenia, personality disorder other serious concerns rendering possible development of diabetes unimportant to potential participant.

Heisler M, Kullgren J, Richardson C, Stoll S, Alvarado Nieves C, Wiley D, Sedgwick T, Adams A, Hedderson M, Kim E, Rao M, Schmittdiel JA. Study protocol: Using peer support to aid in prevention and treatment in prediabetes (UPSTART). Contemp Clin Trials. 2020 Aug;95:106048. doi: 10.1016/j.cct.2020.106048. Epub 2020 Jun 1.