Substance misuse is not consistently addressed within the scope of routine medical practice. More than 90% of primary care physicians fail to diagnose substance misuse when presented with early symptoms of alcohol or drug misuse in an adult patient. Screening, Brief Intervention, and Referral to Treatment (SBIRT) is an integrated, evidence-based approach aimed at delivering early intervention in medical settings for drug and alcohol misuse. An integral component of SBIRT is motivational interviewing (MI), a brief, evidence-based, interventional practice that has been demonstrated to be highly effective in triggering change in high-risk lifestyle behaviors. MI is a patient-focused conversation between a practitioner and a patient that reinforces a patient's motivation to make positive changes in any targeted health behavior. Due to ineffective training, MI is underutilized. The MI approach is very difficult to teach to practitioners who are accustomed to taking a directive approach with patients, asking closed-ended questions (which require yes/no answers), and doing most of the talking. To implement MI and SBIRT widely, there is a critical need to improve the MI training process in graduate medical education.

Our research team has been developing and testing a training tool that uses natural language processing to provide Real-time Assessment of Dialogue in Motivational Interviewing (ReadMI). ReadMI is a low cost, ultraportable solution to enable instantaneous MI training assessment and analysis. It makes use of the latest advances in deep learning–based speech recognition and mobile and cloud computing technologies. In real time, ReadMI produces a spectrum of metrics for MI skills evaluation, including the number of open- and closed-ended questions asked, provider versus patient conversation time, use of emotion words, number of reflective statements, and use of a change ruler, which are all integral parts of MI. One of the innovative properties of ReadMI is that it will provide real-time feedback in the form of a buzzer and a light to trainees whenever they ask 3 consecutive closed-ended questions. ReadMI can facilitate resident skill development in MI as this software-based training solution analyzes their responses and provides immediate feedback. Our central hypothesis is that ReadMI will produce significantly better MI performance than traditional MI training.

The current version of ReadMI automatically produces the complete transcripts of the MI dialogue with over 92% accuracy and reports on provider versus patient conversation time (over 95% accuracy) and the number of open- and closed-ended questions (over 92% accuracy). The preliminary results demonstrate the significant benefit of making ReadMI results available in real time to the trainer and trainees. In role-play training sessions with a simulated patient (figure), ReadMI produces specific metrics that a trainer can share with the resident for immediate feedback. This can be particularly effective for situations in which the resident is doing most of the talking, primarily asking closed-ended questions, or ignoring emotion words used by the patient. Given the time constraints on targeted skill development faced by most residency programs, ReadMI decreases the need to rely on subjective feedback and more time-consuming video review to illustrate important teaching points.

figure

Screenshot of ReadMI Interface During Training With a Medical Resident

figure

Screenshot of ReadMI Interface During Training With a Medical Resident

Close modal

We plan to test ReadMI in a randomized controlled trial, comparing residents who receive MI training with ReadMI technology with residents who receive the usual and customary MI training. ReadMI has the potential to transform MI training, because the prospective applications of ReadMI extend far beyond substance misuse detection and intervention to include chronic disease management. ReadMI will likely improve health care quality by better equipping physicians as decision support agents in their efforts to facilitate patients' health-related behavior change.