Improving Relationships Using Motivational Interviewing (IRMIT)

A Multilevel Approach to Improving Patient-Clinician Relationships and Outcomes in African American Patients With Diabetes Using Artificial Intelligence-Enhanced Motivational Interviewing Training

This research aims to optimize patient-clinician relationships through motivational interviewing training in a Federally Qualified Health Center (FQHC) to improve health outcomes for African American patients with Type 2 diabetes.

Motivational Interviewing (MI) is an evidence-based approach that has been demonstrated to increase both the patient engagement and activation necessary for effective management of chronic conditions, but is underutilized due to the inadequate training of and implementation by health professionals. This research aims to optimize patient-clinician relationships through MI training in a Federally Qualified Health Center (FQHC) to improve health outcomes for African American patients with Type 2 diabetes. MI training and cuing in this study will make use of a human-artificial intelligence (AI) teaming solution, Real-time Assessment of Dialogue in Motivational Interviewing (ReadMI), a tool that uses natural language processing to provide automatic and immediate metrics relevant for patient-clinician interaction.

All participating medical residents in the study will receive the same motivational interviewing training. Patients of these medical residents will be recruited to provide data on their perceptions of their relationship with their physician, as well as health indicators, but there is no direct intervention for patients themselves. The target number of patients is 200.

Participating medical residents will all receive two successive interventions, motivational interviewing (MI) booster training and the use of a dashboard that provides MI metrics. The booster training will make use of an artificial intelligence tool developed by our study team, Real-time Assessment of Dialogue in Motivational Interviewing (ReadMI), that produces metrics on important conversational skills (e.g., talking time, use of open-ended questions). The dashboard will produce these metrics during clinical encounters as a way to provide cuing for the MI approach.

The training in motivational interviewing that medical residents will receive will make use of an artificial intelligence measurement tool, Real-time Assessment of Dialogue in Motivational Interviewing (ReadMI), that produces a spectrum of relevant MI metrics.

The Patient-Doctor Relationship Questionnaire-9 (PDRQ-9) will be used to assess the patient's subjective perception of the quality of the patient-clinician relationship.The PDRQ-9 consists of 9 statements, for which there are response options ranging from 1 (not at all appropriate) to 5 (totally appropriate). Scores range from 9 to 45 with higher scores being more desirable.

The Patient Psychological Safety Scale, developed by the investigators, contains 13 items for which responses can range from 1 (Strongly disagree) to 5 (Strongly agree). Four items are reverse-scored. The minimum value is 13 and the maximum value is 65. Higher scores mean greater perceptions of psychological safety, a desired state.

The Patient Activation Measure (PAM-13) will be used to assess the patient's knowledge, skills, and confidence which is integral to managing their own health. The PAM-13 consists of 13 statement for which each has four response options, ranging from 1 (Disagree strongly) to 4 (Agree strongly), as well as a not applicable (n/a) option. Scores range from 13 to 52, with higher scores being more desirable.

This measures the amount of blood sugar (glucose) attached to hemoglobin, and functions as an indicators of how well diabetes is being controlled.

Changes in the percentage of time the resident speaks in role plays in motivational interviewing practice.

Clinician motivational interviewing skills will be measured using Real-time Evaluation of Dialogue in Motivational Interviewing (ReadMI), software that produces communication metrics, including the percentage of time the clinician speaks versus the patient speaking time. Percentage can range fro 0% to 100%, with the most desirable range being 33%.

Changes in the ratio of open-ended to closed-ended questions used by the resident in role plays in motivational interviewing practice.

Clinician motivational interviewing skills will be measured using Real-time Evaluation of Dialogue in Motivational Interviewing (ReadMI), software that produces communication metrics, including the ratio of open-ended to closed-ended questions. Ratios can range from 0 to infinity, with the most desirable ration being in the 5:1 or higher range.

Changes in the ratio of reflections to questions used by the resident in role plays in motivational interviewing practice.

Clinician motivational interviewing skills will be measured using Real-time Evaluation of Dialogue in Motivational Interviewing (ReadMI), software that produces communication metrics, including the ratio of reflective statements to questions. Ratios can range from 0 to infinity, with the most desirable ratio being greater than 1.

Changes in the "spirit of motivational interviewing" rating given to the resident by the facilitator in role plays in motivational interviewing practice.

The "spirit of motivational interviewing" rating refers to the facilitator's perception of the degree to which the clinician communicates with the patient in a manner that is aimed as creating a need for change on the part of the patient, highlights the patient's autonomy and choice, and minimizes default use of education and advice. Ratings are on a 1 to 10 scale, with 10 being the most desirable rating.

Inclusion Criteria: Eligible participants are residents in Family Medicine and Internal Medicine residency programs at Wright State University who will be recruited to participate in the study. Exclusion Criteria: None Note: Patients of these participating residents will be invited to participate in the study by providing survey and health data at specified intervals, but these patients will not be the direct recipients of any intervention in this study. Eligible patients will be self-identifying African American individuals ages 18 and over with a diagnosis of Type 2 diabetes, and are patients of either Family Medicine or Internal Medicine residents participating in this study. The target number of patients is 200.

Deidentified group data will be made available to other researchers but no individual participant data will be shared.

Vasoya MM, Shivakumar A, Pappu S, Murphy CP, Pei Y, Bricker DA, Wilson JF, Castle A, Hershberger PJ. ReadMI: An Innovative App to Support Training in Motivational Interviewing. J Grad Med Educ. 2019 Jun;11(3):344-346. doi: 10.4300/JGME-D-18-00839.1. No abstract available.

Hershberger PJ, Pei Y, Bricker DA, Crawford TN, Shivakumar A, Vasoya M, Medaramitta R, Rechtin M, Bositty A, Wilson JF. Advancing Motivational Interviewing Training with Artificial Intelligence: ReadMI. Adv Med Educ Pract. 2021 Jun 4;12:613-618. doi: 10.2147/AMEP.S312373. eCollection 2021.