Measuring the Impact of an Interactive Communication Skills Curriculum on Internal Medicine Residents

This trial employed a cluster-randomized design. The residency program randomly divides the residents into 5 groups when they arrive as interns. These groups receive lectures together for all three years of residency and have similar schedules. The investigators used this predetermined structure to randomize the residents' communication course for this study. Two of the five groups were selected using a random number generator to receive the control course and three of the five groups received the intervention course. Although participation in the research study was voluntary, attending the assigned communication course was a mandatory educational activity for the residents.

Control (standard) Course:

The control course consisted of three hour-long didactic lectures on delivering bad news, discussing prognosis, and talking to patients about pain. The majority of time in these lectures was spent on didactic material. Little, if any, time was spent practicing skills or on interactive skill building. The three parts of the course were delivered over a 4-month period.

Intervention Course:

The intervention course also consisted of three hour-long sessions on communication delivered over a 4-month period. It differed from the control course in subject matter and pedagogy. Subject matter focused on specific skills used in communication: reflective listening, responding to emotion, and providing information within a broad range of communication scenarios. In terms of pedagogy, the intervention course was interactive with a focus on skills practice, communication drills, and improvisation to engage learners. The sessions consisted of fifteen minutes of lecture and 45 minutes of skills practice.

Faculty, senior residents, and fellows at the Center for Excellence in Healthcare Communication (CEHC) facilitated the control and intervention courses.33 Instructors at the CEHC taught facilitators best practices in adult learning theory, small and large group facilitation, and effective communication skills. Two trained facilitators led each course. The same facilitators taught the control and intervention courses.

PARTICIPANTS

Resident Participants:

Residents were recruited from Cleveland Clinic's Internal Medicine Residency program from November 2017 through April 2018. A total of 120 residents were eligible to participate in the study as they were post-graduate (PGY) 2 and 3 residents scheduled to receive mandatory educational curriculum on communication. Residents were invited to participate in the study first by an e-mail and again in person immediately prior to the start of the course. Residents received an information sheet about the study which explained that both the residents and their patients would be surveyed. Residents who completed the baseline surveys were considered enrolled, and their patients were sampled as detailed below. Residents were given a $10 gift card in compensation for their time and effort. While participation in the study was optional, all residents received either the intervention or control course as part of their mandatory educational curriculum. Participating residents provided basic demographic data and information about previous communication training. They also completed the Jefferson Scale of Empathy (JSE) and Maslach Burnout Inventory (MBI), validated tools that measure empathy and burnout respectively.

Patient Participants:

Patients who had an appointment scheduled with an enrolled resident in their primary care clinic were identified through the electronic medical record (EMR). Patients with appointments scheduled at all eight clinics in our health care system were eligible to participate in the study to ensure demographic representation. Eligible patients were age 18 or older, had completed an appointment with a resident in the trial within 2 months of the resident finishing the course, and were proficient in English.

Patients were recruited by mail within two weeks of their appointment and were compensated with a $10 gift card. Survey packets contained a survey invitation cover letter, the study information sheet, the survey, and a stamped return envelope. The cover letter identified the resident the patient saw and the location of their medical appointment, to clarify which medical provider to assess in the survey. The survey consisted of the Communication Assessment Tool, the Patient Activation Measure (PAM13), and demographic questions. Three attempts were made to contact non-responders.

MEASURES

Primary Outcome - Communication Assessment Tool:

The primary outcome was a validated measure of communication, the Communication Assessment Tool (CAT). The CAT is a fifteen item survey measuring physician communication as perceived by patients. Each item is rated from 1 "Poor" to 5 "Excellent". The first fourteen items measure specific aspects of communication, and the last item measures the quality of communication overall. Because surveys such as this are subject to a ceiling effect, scores were also analyzed as the percent of questions that received the maximum score - also known as a "top box" score.

Secondary Outcomes:

In addition, the investigators evaluated whether better communication shaped patient health outcomes. Before and after the patient's clinic appointment, the investigators examined the patient's EMR and extracted information about depression, blood pressure, and admittance history. Depression was evaluated using PHQ-9 scores recorded in the EMR, a common measure of depression with well documented internal consistency and construct validity. The PHQ-9 consists of nine questions, each scored from 0 to 3 for a total score of 0 to 27, with higher scores indicating more severe depression. Scores between 5 and 10 suggest mild depression, 10-14 suggest moderate depression, and 15-27 suggest severe depression. Missed appointments ("no-shows") and cancelled appointments were recorded as the frequency per month. Hospitalizations were recorded as the frequency of hospitalization for 1, 3 and 6 months after the patient's appointment with the internal medicine resident.

Finally, patients also completed the 13-item Patient Activation Measure (PAM13), a standardized measure to assess patients' knowledge, skill, and confidence to manage their own health. Scores range from 0 (low activation) to 100 (high activation). The PAM13 was included in the patient survey.

STATISTICAL ANALYSES Summary statistics were performed for baseline characteristics both at the resident and patient levels. Frequencies with proportions were used for categorical data and means with standard deviations for continuous data. To examine the associations between intervention and control groups, the investigators used chi-square tests for categorical data, and t-tests for continuous data. At the patient level, the investigators then assessed the primary outcome, the CAT score, using multivariable logistic regression, and adjusted for gender, age, race, income, PAM-13 scores, physician's emotional exhaustion as reported on the MBI, JSE score, and year of residency. Results of models were summarized as odds ratios together with 95% Wald Confidence Intervals. All tests were 2-sided and a p<0.05 was considered statistically significant. The investigators used SAS 9.4 for all statistical analyses.