More than 25.8 million Americans have type 2 diabetes mellitus (T2DM). In 2012, T2DM cost the United States $245 billion in both direct and indirect medical costs [1]. Comparable to the general population, it has been estimated that 10% of the New York City population suffers from T2DM, and patients with diabetes-related disorders occupy half of the hospital beds in the city [2]. The patient-centered medical home (PCMH) model, which strives to provide comprehensive care and improve patient’s self-management skills through increasing engagement with health care providers (HCPs), is ideally suited for T2DM care [3-5].

Preparing patients for the PCMH model is challenging because of the inherent power differential between HCPs and patients. Efforts to prepare HCPs to practice in the PCMH model have included strategies for encouraging patient self-care activities and behavior change (eg, tailoring, brief negotiation, motivational interviewing) [6,7]. Interventions have also been developed to educate patients about disease management and increase involvement in their care [8-11]. However, both approaches fail to address the inherent asymmetry in the power dynamics of the physician-patient relationship [12-21]. Increasing patient activation is one method that has been suggested to overcome these barriers. Activated patients are knowledgeable about their health conditions, confident in their ability to manage these conditions, and maintain their health by seeking information and performing health promoting behaviors [22]. In general, more activated patients, as determined by a higher score on the Patient Activation Measure (PAM) [22,23], ask more questions during HCP visits [22-24] and perform more self-management behaviors including diet and exercise [22,23,25-28]. More specifically, patients with T2DM who have higher PAM scores report less difficulty in managing their diabetes than those with lower scores [28]. Furthermore, several interventions that increase patient activation have shown promise for improving outcomes [29-32], most notably in congestive heart failure [29] and T2DM [30].

Standardized patient (SP) training, a well-established, performance-based intervention, offers a compelling method of activating patients to become partners in their health care. Standardized patients are trained to reliably and validly assess HCP clinical competence [33-37] and, as a result, become more activated “real” patients who have higher expectations of HCPs and improved communication with HCPs [38-43]. Additionally, SPs have improved their own health behaviors in terms of weight loss [44] as well as HIV testing and sexually transmitted disease prophylaxis [45]. More recently, these same training methods have been used to train standardized health care providers (SHPs) to assess communication between HCPs of different disciplines (eg, doctors and nurses) as part of interprofessional education [46-48].

Our Patient Empowerment Program (PEP) seeks to adapt the successful SP methodology currently used in HCP education and translate it onto the patient side of the medical encounter [49]. To do so, the PEP incorporates ideas of shared decision making (SDM), involves role-playing, and helps patients develop the skills to give effective feedback to HCPs using validated checklists of observable behaviors. This intervention is both evidence based and theory supported and addresses the needs of multiple stakeholders, including health system quality leaders, HCPs, patient advocates, SP trainers, and patients. We focus on patients with T2DM because of its high prevalence in the population, the need for frequent contact with the health care system, and the numerous aspects of treatment (eg, lifestyle changes, medications, blood glucose monitoring, annual screenings, and so on) that patients must discuss with their HCPs.

The objectives of this study are to (1) assess the acceptability and feasibility of implementing the PEP across three urban clinical sites; (2) determine if a PEP will change patient preference for control in medical decision making, improve patient self-efficacy in patient-HCP communication, and increase patient activation; (3) determine if participation in the PEP is associated with an increase in diabetes self-management behaviors; and (4) explore the effect of the PEP on hemoglobin A_1c (HbA_1c) level.

The PEP is a randomized, wait-list control study (see Multimedia Appendix 1) that aims to enhance general communication skills so that patients can participate in SDM during office visits, become activated in their own health care, and better manage their medical conditions.

Patients were recruited in cohorts of 20 and then randomly assigned to attend the PEP intervention immediately (immediate intervention) or after a waiting period (deferred intervention). All participants completed a baseline assessment at the time of enrollment (T0) and will complete follow-up assessments 1 week (T2), 3 months (T3), and 6 months (T4) after completion of the PEP intervention (see Figure 1). Additionally, the deferred intervention group was assessed at the end of the waiting period (T1), immediately before being invited to attend the PEP intervention. The T1 assessment for the deferred intervention group was approximately 3-4 months after randomization and was timed to coincide with the T3 assessment of the immediate intervention group. A wait-list control design was chosen to create a control group with multiple time points of data collection (deferred intervention T0-T1) while still allowing for all participants who enrolled in the study to complete the PEP intervention and post-PEP follow-ups.

Aside from attending the PEP intervention, all participants continued to receive their usual care, including all scheduled primary care and specialty clinic appointments, HbA_1c monitoring, and all medications as directed by their physicians.

Bellevue Hospital Center, Gouverneur Health, and Woodhull Medical Center are all within the Health and Hospitals Corporation, a large public hospital system in New York City. The adult ambulatory care clinics at these sites serve patients from the local community as well as immigrants from around the world, with English, Spanish, Mandarin, Cantonese, and Bengali among the most common languages spoken. Approximately one-third of patients have Medicaid or are without health insurance and the most common diagnoses in the clinic include obesity, hypertension, heart disease, diabetes, and asthma.

We recruited patients with T2DM who presented to the adult ambulatory care clinic at each site. Patients eligible for inclusion in the study (1) were at least 18 years old, (2) had a diagnosis of T2DM for at least 1 year, and (3) their most recent HbA_1c level was between 6.5% and 11%. Patients were excluded if they (1) were unable to speak English or (2) had a major psychiatric illness that impaired their ability to care for themselves (eg, schizophrenia, uncontrolled bipolar disorder, or uncontrolled depressive disorder).

Research assistants (RAs) screened all of the clinic appointments each day to identify patients who were eligible for the study. Research assistants then briefly explained the study to all eligible patients who came to their appointment and obtained informed consent from those who decided to enroll (see Figure 1). After obtaining informed consent, RAs read participants a set of questionnaires to complete the initial (T0) assessment (see Table 1).

After 20 participants were enrolled in the study, a Web-based random number generator was used to generate a list of 20 integers between one and two. The list was refreshed until it contained 10 ones and 10 twos. Based on this list, participants were assigned in order of their study identification number to the immediate intervention (1) or deferred intervention (2) group. All participants in the immediate intervention group were contacted by phone to schedule the PEP intervention sessions and all participants in the deferred intervention group were contacted by phone and informed that they would be recontacted in 3-4 months to schedule their PEP intervention sessions.

Participants were compensated for participating in the study. They received US $10 and a MetroCard with US $10 at each of the two PEP intervention sessions. They also received US $60 and a MetroCard with US $10 for completing the study if they attended a post-PEP focus group.

In order to increase retention of deferred intervention participants, they received a letter from our program and an educational handout about T2DM in the mail approximately every 3 weeks. Handouts were sent in the same order and all handouts came from the Diabetes Care and Education website [50]. In total there were four mailings: (1) Nutrition to Help Manage your Diabetes and Weight, (2) Eating Healthy on a Lean Budget, (3) Know Your Blood Sugar Numbers, and (4) Managing and Preventing Hypoglycemia. All four of these handouts were short, only 1-2 pages in length, and contained general information about T2DM that participants should have already obtained from their HCPs as part of their routine treatment. Therefore, we do not expect that receiving these handouts will cause a significant difference in the two randomized groups or otherwise affect the results of the study.

Outcome measures for the study primarily consist of self-report questionnaires assessing health attitudes and behaviors (see Table 1). Because of the low literacy of the patient population, RAs read all questionnaires to the participants and then recorded their responses. All questionnaires were asked in the same order and a standardized protocol was created so that all RAs asked and clarified questions in the same manner. Outcome measures were assessed at the time of enrollment (T0), after the waiting period (T1: deferred intervention group only), and postintervention after 1 week (T2), 3 months (T3), and 6 months (T4).

Our PEP is a two-session course led by a clinical health psychologist (LA) and assisted by research staff (JP, SK, and CS). Each session is 2 hours in length and the two sessions are held approximately 1 week apart. The PEP intervention is designed to be a group experience, with 2-6 participants attending each session. Ideally, the same participants who attended session 1 will return for session 2, but because of the limited availability of both research staff as well as participants, it is expected that this may not always be the case.

After completion of the PEP intervention, participants will be followed up for 6 months and will complete repeat assessments at 1 week (T2), 3 months (T3), and 6 months (T4) after the PEP intervention (see Table 1). The T2 assessment will take place either at the beginning of the focus group, to avoid any bias introduced by discussing the PEP intervention again, or by telephone for any participants who do not come to the focus group. The T3 and T4 follow-ups will occur by telephone and any concurrent HbA_1c available at these times will be collected from the EMR.

Prior studies have pilot-tested interventions to increase patient activation and found that PAM scores increased by a range of 4-8 points [30,31]. This difference was statistically significant, is thought to be clinically significant, and translates to a moderate effect size, Cohen’s d=0.48. Given the repeated-measures design of the study, setting alpha = .05 and beta = .2, and using the effect size calculated from the literature, the required sample size will be n=36 patients. It is our goal to have 40 participants complete the PEP intervention.

In anticipation of significant participant dropout over the course of the study, we aimed to enroll 40 patients from Bellevue Hospital Center, 20 from Gouverneur Health, and 20 from Woodhull Medical Center. From these patients, it was our goal that 20 from Bellevue Hospital Center and 10 each from Gouverneur Health and Woodhull Medical Center would complete the PEP intervention. Because of these concerns about dropout, all participants who were randomized to the immediate intervention group but unable to attend the PEP intervention at that time were contacted for the T1 assessment with the deferred intervention group and invited back to attend the PEP intervention as part of that group (see Figure 1).

First, descriptive statistics including mean, standard deviation, range, skewness, and kurtosis for all continuous variables as well as frequencies for all categorical variables will be calculated at each time point (T0-T4). Because of concerns about the relatively small sample size and low literacy/health literacy, the normality of all continuous variables will be assessed both graphically with histograms as well as statistically using the Shapiro-Wilk and Kolmogorov-Smirnov tests, and transformations will be made if necessary. Then we will assess for any baseline (T0) differences between two separate groups: (1) the randomly determined immediate intervention and deferred intervention groups and (2) all participants who attended the PEP intervention and those who did not. Both of these analyses will be conducted using chi-square test for categorical variables and either independent samples t test or Wilcoxon rank sum test for continuous variables.

Next, analyses will be conducted within each intervention group to evaluate for any effect of the PEP intervention on the primary, secondary, and exploratory outcomes using chi-square test for categorical variables (eg, PAM level, PCS) and either repeated-measures t test or Wilcoxon signed rank test for continuous variables (eg, PAM score, AURA, Krantz, SF-12v2 Health Survey, DSMQ, and HbA_1c). Initial assessment (T0) will be compared with each of the post-PEP assessments (T2-T4) for each variable. For the deferred intervention group, initial assessment (T0) will be compared with the assessment at the end of the waiting period (T1), which will then be compared with each post-PEP assessment (T2-T4).

After analyzing each group individually, the groups will be combined in order to have a larger sample size for pre- and post-PEP comparisons. Pre-PEP intervention data for these analyses will consist of data from enrollment (T0) for participants in the immediate intervention group and data collected at the end of the waiting period (T1) for participants in the deferred intervention group. Repeated-measures analyses will be conducted on the primary, secondary, and exploratory outcomes as described above and pre-PEP data (either T0 or T1) will be compared with each post-PEP assessment (T2-T3).

Finally, multivariate analyses will be conducted using a series of generalized mixed-effects models to determine what effect patient characteristics, group membership, and time have on each continuous outcome measure. In each model, the dependent variable will be the outcome measure (eg, PAM, AURA, Krantz, SF-12v2 Health Survey, DSMQ, and HbA_1c). Independent variables will include participant characteristics such as age, sex, race/ethnicity, education level, annual income, number of years with T2DM, and health literacy as measured by the NVS. Other independent variables will include group (immediate vs deferred intervention) and time (T0-T4).