Teaching patient-centered communication skills: a telephone follow-up curriculum for medical students

Background To encourage medical students’ use of patient-centered skills in core clerkships, we implemented and evaluated a Telephone Follow-up Curriculum focusing on three communication behaviors: tailoring education to patients’ level of understanding, promoting adherence by anticipating obstacles, and ensuring comprehension by having patients repeat the plans. Methods The intervention group consisted of two different cohorts of third-year medical students in longitudinal clerkships (n=41); traditional clerkship students comprised the comparison group (n=185). Intervention students telephoned one to four patients 1 week after seeing them in outpatient clinics or inpatient care to follow up on recommendations. We used surveys, focus groups, and clinical performance examinations to assess student perception, knowledge and skills, and behavior change. Results Students found that the curriculum had a positive impact on patient care, although some found the number of calls excessive. Students and faculty reported improvement in students’ understanding of patients’ health behaviors, knowledge of patient education, and attitudes toward telephone follow-up. Few students changed patient education behaviors or called additional patients. Intervention students scored higher in some communication skills on objective assessments. Conclusion A patient-centered communication curriculum can improve student knowledge and skills. While some intervention students perceived that they made too many calls, our data suggest that more calls, an increased sense of patient ownership, and role modeling by clerkship faculty may ensure incorporation and application of skills.

three specific, measurable patient-centered communication skills. These skills, currently taught in the University of California, San Francisco (UCSF) School of Medicine's pre-clerkship courses, help to tailor the educational message to the patient's level of understanding ('ask-teach-ask') (11), promote medical adherence by anticipating obstacles ('looking ahead') (12), and ensure patient comprehension of the plan by having the patient repeat the plan back in his or her own words ('closing the loop') (13). The purpose of the current study was to explore student and faculty perceptions of the TFC and to evaluate the impact of this reinforcing curriculum on MS3s' patient-centered communication skills. Our hypothesis was that reinforcing MS3s' patient-centered communication skills through a TFC would improve these skills in subsequent patient encounters and increase awareness of psychosocial barriers in safe care transitions.

Design and setting
We conducted this prospective, single-institution study of a curricular intervention using both quantitative and qualitative methodology at UCSF and obtained institutional review board approval.
The UCSF pre-clerkship curriculum introduces students to three patient-centered communication skills ('askteach-ask', 'looking ahead', and 'closing the loop'). The TFC was designed to build upon and reinforce these skills during the core clerkships. In the core clerkship curriculum, students are placed into different clerkship models based on their stated preferences. Models offered include 1) traditional clerkship rotations, 2) 6-month site-based programs where students complete a series of consecutive clerkships at one clinical site, and 3) a 12-month longitudinal integrated clerkship (LIC) (14) where students concurrently complete their core clerkships through discipline-specific preceptorships and continuity with a panel of patients they follow across settings. Students were not aware of the TFC or its implementation in the different curricular models when submitting their clerkship preferences.

Subjects
Study participants were MS3s on clerkship rotations in 2008Á2009 and 2009Á2010. In 2008Á2009, we implemented the TFC for MS3s in a 6-month site-based program (Intervention Group 1, or IG-1). Students in this program, called VALOR, completed their medicine, neurology, psychiatry, and surgery clerkships at a Veterans Affairs Hospital (15). In 2009Á2010, we expanded the TFC to include that year's MS3s in VALOR and the LIC (Intervention Group 2, or IG-2) (14). In total, 41 MS3s and seven faculty preceptors participated in the TFC during the two intervention years. MS3s completing traditional clerkship rotations (which did not include the TFC) comprised the comparison groups (n 090 in 2008Á2009 Comparison Group, CG-1; n 095 in 2009Á2010 Comparison Group, CG-2).

Intervention
The objectives of the TFC were to 1) reinforce students' use of three patient-centered communication skills ('askteach-ask', 'looking ahead', 'closing the loop') and 2) deepen students' understanding of patients' real-world psychosocial contexts and their obstacles to following action plans. Expected student learning outcomes included increased use of the skills in patient encounters and more attempts to help patients address obstacles to care plan adherence. We also hoped that students would assess patients' retention and understanding of their care plan from their last medical visits, reinforce clinical recommendations, and address obstacles.
Prior to implementation of the TFC, one author (JS) trained faculty preceptors, explaining their role in the curriculum and ensuring their comfort and proficiency with the three patient-centered communication skills; JS also introduced these skills to intervention group students (IG-1 & IG-2) and oriented them to the expected curricular activities (Appendices 1 and 2). All intervention students attended seminars (2 hours with scripted role plays) and made one to four calls to patients with individual and small group debriefs with preceptors.
IG-1 students selected four patients that they had seen in outpatient clinics or had discharged from inpatient care; they telephoned the patients at home approximately 1 week after their last encounter. Students followed up on clinical recommendations such as medication changes and referrals (Appendix 2), documented the telephone contact in the patient's medical record, and completed a written Post Telephone Call Exercise (PTCE; Appendix 1) after each telephone encounter. Based on IG-1 student feedback, we modified the curriculum to decrease the number of PTCEs required for IG-2 students, who were required to complete one PTCE sometime during the curriculum instead of after every telephone call. Students reviewed their completed PTCEs with faculty preceptors trained in the TFC.

Evaluation measures
We focused on evaluating students' knowledge and skills, faculty and students' perceptions, and change in student behavior or application of knowledge and skills.

Survey
The IG-1 students completed a 15-item post-course survey assessing the usefulness and quality of the TFC and selfreported changes in their attitudes, knowledge, and skills regarding doctorÁpatient communication. Closed-ended questions used a five-point Likert scale (1 0strongly disagree, 50strongly agree). Faculty who worked with IG-1 students completed a 15-item post-course survey that combined Likert scale and open-ended questions to inquire about their impressions of student learning. Based on IG-1 student comments regarding the evaluation process, we streamlined evaluations to obtain focused verbal feedback from IG-2 students regarding the usefulness for learning and quality of the TFC.
Focus groups IG-1 students and their faculty participated in separate focus groups to share their experiences of participating in the TFC. Facilitators who were not involved in teaching the curriculum conducted the separate 1-hour focus groups with both students and faculty.

Standardized patient examination
To evaluate students' ability to apply skills that the TFC reinforced, we constructed a chronic illness case that provided several opportunities for students to demonstrate the three patient-centered communication skills. The case involved a 68-year-old man with a history of congestive heart failure and depression who presented with fatigue and an interest in reestablishing primary care. We included this case as one of three standardized patient (SP) cases in an existing clinical performance examination required for all MS3s. Following standard protocol for all SP examinations at our institution, the SPs received extensive training on all cases used in the exam, including case portrayal, checklist scoring, and exercises to establish inter-rater reliability. For all students, SPs completed a 34-item checklist on history-taking, physical examination, and communication items after each student encounter. SPs did not know which students participated in which clerkship programs and were not aware of the TFC. We compared the performance of students between the intervention and comparison groups on seven communication questions that specifically related to patient education skills (Table 3).

Quantitative data analysis Satisfaction surveys
We compiled intervention student and faculty evaluations of the curriculum and calculated mean scores for survey data.

Clinical performance data
We examined students' performance during the SP case and compared scores between intervention and comparison group students using a one-way analysis of variance (ANOVA). Multiple regression analyses were conducted to predict the type (intervention vs. comparison group) and intensity of the curricular experience (among those who made one, two, and four calls) on clinical performance (SP case).

Qualitative data analysis
We analyzed the focus group comments using open coding (16) to examine the qualitative data from the PTCEs. Four authors (GS, CC, AT, HCC) read seven PTCEs together to identify initial descriptive codes. Upon reaching agreement, three authors (GS, CC, HCC) reviewed the remaining PTCEs. Using the constant comparative method, we used multi-staged coding, beginning with open coding of raw data to develop key ideas. Axial coding then organized these categories into patterns, and finally, we used selective coding to develop theoretical formulations that linked key variables to themes. Discussion among investigators resolved analytic discrepancies.

Results
In 2008Á2009, 18 students received TFC (IG-1), and 90 students comprised the comparison group (CG-1). Seventeen of the intervention students completed four PTCEs, and one student completed three. During 2009Á 2010, 33 students received TFC (IG-2), and 95 students constituted the comparison group (CG-2). Although IG-2 students were required to complete one PTCE, 1 student completed four, 7 completed two, and 16 completed one. Nine students' PTCEs were completed but misplaced. Students did not provide PTCEs for any additional calls they made. Over the 2 years of the study, a total of 105 PTCEs were submitted. Students reviewed their telephone follow-up exercises with their TFC faculty who included five internists and two psychiatrists with 1Á16 years of experience as preceptors.
Characteristics of telephone calls On PTCEs, students most frequently cited that their goals during the call were to check in about prescribed medications, follow-up appointments, and treatment plans discussed in their last encounter ( Table 1). Most of the reported telephone calls did not result in a change in medication or treatment. Students reported accurate patient comprehension and retention of information from their last contact in the majority of their calls. In half of the exercises, students reported good patient adherence to their management plan. In patients with non-adherence, students ascribed the problem to patients' poor understanding of the plan, situational or structural barriers, or the existence of co-morbid conditions. The majority of exercises revealed that students were surprised by unanticipated information elicited in their conversations with patients or caretakers during the calls (e.g., reasons for non-adherence, level of family involvement in care). Perceptions of the curricular experience Nearly all PTCEs (96%) contained student comments about the value of the follow-up exercise ( Table 2).
Positive comments reflected students' appreciation for follow-up information on their patients and opportunities to make additional interventions. Yes, patient did comprehend 58 64 No, patient did not comprehend 28 31 Patient comprehended some but not all of last contact 5 5 IV. Students' assessment of patient retention of last contact, e.g., information, plan (n 094) Yes, patient did retain 63 67 No, patient did not retain 21 22 Patient retained some but not all of last contact 10 11 V. Students' assessment of patient adherence to plan from last contact (n 0102) Yes, patient did adhere 51 50 No, patient did not adhere 39 38 Patient adhered to some but not all of plan from last contact 12 12 VI. Students' assessment of reasons for patient non-adherence to plan from last contact (n 048) Did not understand 15 31 Barriers to adherence 15 31 Co-morbidity 13 27 Other (e.g., did not remember; chose not to adhere) 5 11 VII. Content of call surprised students with unanticipated information (n 080) Students also noted how the calls led to a deepening of the relationship with the patient and that the patients expressed appreciation for the contact.
It really gave me a sense that I was in an ongoing partnership with the patient, and I think she appreciated the call as well.
Finally, students valued the opportunities to practice and ensure learning of the relevant communication skills.
Reflecting on the phone interview allowed me to think about the process of delivering information and in doing this, I realize that I am not as explicit with the communication techniques as I would like to be. Some students discussed how deciding to telephone more than the required number of patients depended on their sense of responsibility.
I was more likely to call if I knew they were coming back at some point-it's really my patient. I just didn't feel ownership of patients to the extent it seemed useful and even appropriate to call in a lot of cases.
Negative student comments focused on the requirements of the curriculum (preferred the calls to be elective, too many required calls) and logistical challenges of making telephone calls.
It's a reasonable learning activity . . . it is just difficult to schedule times to write everything up with so much other learning taking place throughout the week. LIC students, in particular, recommended that the PTCE occur early in the year when they are beginning to engage their patient panels.

Changes in attitudes, knowledge, and skills
On the post-course survey, the majority of students agreed or strongly agreed that they learned about patient health behaviors (77%), that their ability to provide patient education improved (71%), and that these communication skills will be relevant to future patient encounters (94%). However, only 41% of students reported having made changes to what they do during patient visits because of the TFC.
All five faculty preceptors for the IG-1 students agreed or strongly agreed on the post-course survey that the telephone follow-up calls were a valuable learning experience for students. Faculty preceptors believed that the exercise highlighted the gap between what clinicians think they have said and what patients understand, and that students learned even a single call could have significant impact on patient care. As one faculty member said, 'I thought it effectively illustrated how fragile a treatment plan can be and how powerful a telephone call [can] be in reinforcing a treatment plan'. Several faculty hypothesized why students found the curriculum challenging. Some felt that students may have perceived less learning from these exercises because they 'did not seem to fully appreciate the impact they made'. A related theme concerned role modeling, with another faculty stating, 'Students don't see attendings and housestaff making follow-up calls'.
Transfer of learning and ability to apply new skills and changes in behavior Data from PTCEs showed that a significant number of patients did not comprehend (31%) or retain (22%) information from their last clinical contact and were not adherent (38%) or did not understand (31%) their treatment plans. In response to these findings during their telephone follow-up calls, a number of students (35%) provided interventions during their calls to improve individual patients' care.
Of the seven communication outcome variables examined in the SP exercise, IG-1 students scored higher than the comparison group CG-1 on three techniques, and IG-2 students scored higher than CG-2 students on a different technique (Table 3). We found no differences between intervention and comparison groups when we separately compared the performance of the VALOR and LIC students in IG-2 with CG-2; combined IG-1 and IG-2 into one intervention group with the one compar-ison group (CG-1 ' CG-2) in a linear regression; or used regression analyses to look at performance of the IG-2 students in relation to the number of PTCEs completed.

Discussion
Our evaluation of the TFC yielded encouraging but mixed results which may be best understood through the framework of the first three levels of Kirkpatrick's model of evaluation, modified for medical education (17Á19):

Level 1-Perceptions of the curricular experience:
Students enjoyed the curriculum overall and saw the value of TFC for their personal learning and for their patients' care.

Level 2-Changes in attitudes, knowledge, and skills:
Calls allowed students to assess their patients' comprehension, retention, and adherence after inperson encounters Á important clinical feedback data typically not available to students. Both students and faculty reported improvement in students' understanding of patients' health behaviors, knowledge of and skills in communication related to patient education, and attitudes toward telephone follow-up in patient care. 3. Level 3-Ability to apply new skills and changes in behavior: Few students reported changing their patient education behaviors. In addition, few made Combining data from both years, no significant difference was found in the communication behaviors of the intervention and comparison groups. Interestingly, IG-1 students, who each completed 4 PTCEs, performed better than IG-2 students, who completed fewer PTCEs. It is possible that we might have seen a difference had we required IG-2 students to complete more PTCEs. Students may not always know or want what is best for their learning (20), and our modifications to TFC to improve satisfaction may have resulted in negative learning outcomes.
Another explanation for the different findings over 2 years might be a change in the comparison group students from one academic year to another. The CG-2 students improved in two of the parameters that had shown significant differences between CG-1 and IG-1 students. There were no other curricular changes to explain the comparison group's skill improvements. Although it is possible that elements of the TFC could have generalized to non-participants, the etiology for this baseline improvement is unclear and deserves further study.
It is interesting that, despite self-reported gains in knowledge and skills and perceived improvements to their patients' care, few students made telephone follow-up calls to additional patients to further their communication skills. A number of students acknowledged that they did not telephone additional patients because they did not feel ownership over those patients' care. This sense may have limited their opportunities to practice communication skills and cement further learning. Yet, the TFC provides a tangible way for MS3s to assume authentic roles in reducing communication gaps that adversely affect patient care (9). It is possible that students may develop improved communication skills by further promoting continuous patient relationships in their core clerkships, thereby increasing their sense of accountability and responsibility for patient care (21Á25).
Another factor in students not making additional calls may relate to a lack of modeling. Focus group faculty mentioned that students do not observe residents or faculty calling patients in between visits. In addition, telephone calls to patients are not a routine expectation of clerkships and students are not formally evaluated on them. These factors may result in a hidden curriculum (26) that suggests to students telephone follow-up calls are not a typical or essential part of real practice. To ensure that students do not view certain activities as curricular busy-work and that skills learned in the classroom are reinforced during the clerkship experience, modeling from clinical supervisors is critical.
Although students demonstrated no measurable change in their behavior with subsequent patients or in a formal assessment, students seemed to have positive effects on patient care while participating in the TFC. In their telephone follow-up calls, students made a wide range of interventions in response to significant rates of poor patient retention, comprehension, and adherence to treatment plans. The TFC resulted in immediate benefits to patients and improved quality of care, and this arguably supports further implementation of this curriculum. Perhaps evidence of tangible effects on improved patient care can be leveraged to change the hidden curriculum, improve student engagement in the curriculum, and increase student practice of telephone follow-up.

Limitations
We introduced TFC into only one medical school, although the clerkships occurred at multiple private and public hospital and clinic settings. While the demographics of the intervention and comparison group students do not differ (14,15), it is unclear if one group valued patientcentered communication skills more than the comparison group. Presumably, students electing these longitudinal programs may hold increased interest in continuity of care, thereby introducing selection bias. There were non-equal numbers of data points in each of the 2 years of the curriculum. The IG-1 students made four phone calls and completed four PTCEs each, whereas IG-2 students completed a minimum of one phone call (based on the required one PTCE). This change in the degree of experience did not permit a direct comparison between the samples of the 2 years, although it does suggest that more experience with phone calls may improve skills to a greater degree. Future research would benefit from directly ascertaining patients' experiences of their follow-up telephone encounters with students.
One important challenge in medical education is to change learners' behavior for the benefit of their patients. This study demonstrates that a patient education curriculum can improve student knowledge and skill acquisition, and that more intense or different interventions may be needed to affect student behavior change. While students perceived the number of telephone calls required to be excessive, objective results suggest that more calls may be necessary to ensure incorporation and application of skills. Though we focused on improving the outcomes of our students' future patients as the ultimate curricular goal, this study suggests that this curricular intervention may also have improved the quality and safety outcomes of their current patients.