DOI: https://doi.org/10.21203/rs.2.13425/v1
Purpose Residency programs have the intricate and complex role of training health care providers. But little is known about what residents and attendings consider norms of practice or the tensions among different values residents are expected to uphold. Thus, dialogical and reflective frameworks are being explored for resident learning.
Methods This study examined the use of facilitated conversations with groups of residents and attending physicians while reviewing video-recorded resident–patient interactions. The conversations were recorded, transcribed, and qualitatively analyzed.
Results A total of 24 residents and 10 attendings participated in conversations while separately and in parallel groups reviewing 15 resident sessions. Residents explicated the norms of practice and evaluated their performance, which often agreed with those of attending physicians in calling out important learning opportunities. When disagreement occurred, residents’ explications of their reasoning were often relevant and, via reflection and dialogue, helped clarify intentions that were not apparent in the videos. Residents and attendings both often judged actions on more than one domain of value. For instance, if a resident addressed problems, built relationships in a timely manner, and acted autonomously without jeopardizing the quality of care, she satisfactorily performed her duty.
Conclusions Practice norms and value struggles were addressed by participants during reviews, which provided a promising framework for learning and assessment. Also, the non-hierarchical structure opened space to acknowledge a diversity of positions and for tensions among values to be explicated.
The complex role of training doctors is handled through residency programs, which must support resident learning while still providing safe patient care, and residencies grapple with strategies that can simultaneously uphold the two ends.1–6 An increasing body of evidence supports the premise that resident engagement in reflection and self-assessment improves learning.7–11 There is also evidence that learning from peers may be equal to or more effective than learning from an attending.12 Furthermore, many viable learning models exist that center on learners, with non-hierarchical engagement and empowerment of participants.13–17
In contrast, learning has historically been viewed as adaptation, where the learner performs a behavior, then an evaluator gives feedback, and finally, the learner performs again.7,8,18 In this style of assessment, an evaluator observes an object and judges its behavior. Variations exist around who is performing the evaluation (e.g., attending, peer, self) and who is giving the feedback (attending or peer).7,8,19–21 Within the meta-theoretical framework of subject–object relations, self-assessment has been considered a “problem child,” and little is understood about its workings. Despite mounting evidence that self-evaluation enhances learning,7,8 it has often been criticized as inaccurate, unreliable, and “biased.”22–25
Engaging learners is not only effective but is arguably the right thing to do. Residencies should develop learning models that are not merely feedback-driven behavior adaptation but where, instead, learning focuses on growth and development scaffolded by competent others and led by learners themselves. Further, assessment can be viewed not as an external process focused on behavior but as a dialogical and reflective conversation where learners engage in understanding and refining their actions.
While there is a large body of literature on “best practices” from experts’ perspectives and on checklists used for judging behaviors from the perspective of observers, there is little known about the perspectives of participants in dialogical and reflective learning. This study helps fill the knowledge gap through 1) examining resident and attending views on norms of practice for primary care functions and 2) demonstrating the tensions caused by the range of values residents are expected to uphold while developing competency. These aims help in developing an inclusive and reflective learner–educator dialogue about resident competency and learning in clinical settings.
Because qualitative research investigates human phenomena that, by their nature, do not lend themselves to quantitative methods,26 we used a critical qualitative structure for this study. A qualitative study is also fitting because it allows for the inclusion of multiple perspectives (e.g., resident, faculty, and researchers) and the use of multiple data collection methods, such as video reviews and group discussions. The word “critical” is used to indicate shared concerns with social theory and an awareness of social structure, culture, power, and human agency in this inquiry.26 The study was reviewed and granted an exemption by the Indiana University IRB.
This study was conducted at a university-based family medicine residency program where residents and attending physicians provide full-scope family medical services through an outpatient clinic located in the heart of a metropolitan area. The Accreditation Council for Graduate Medical Education requires that residency training includes observation of resident interactions with patients. To that end, this residency implemented video reviews between August and November 2016, during which time 30 resident–patient interactions were recorded. These video recordings were first reviewed by groups of two to four faculty members and then by groups of three residents, one from each year of residency. The review sessions lasted approximately two hours, during which the group reviewed three videos, one from each resident. The separate interactions among faculty members and among the residents were all audiotaped.
All residents and faculty at the residency program were included since it was conducted for educational purposes. For each resident, one to three videos were recorded, one of which was selected for review based on the quality of the recording. Residents were informed on the day they would be video-recorded. Only adult clinic patients with chronic or acute medical conditions who spoke English and were willing to be video-recorded were included. We used liberal inclusion criteria for the patients to assure findings could be generalized to other clinical settings and specialties. We excluded video recordings that were extremely short (< 7 minutes) and those with poor visual or sound quality. During the project’s life span, a total of 31 resident videos were reviewed. The researcher (MA), also a male attending physician, moderated both the attending group and resident group discussions. He was interested in this work to ensure adequate supervision of residents and to provide open space for learning and conversations. To ensure separation between the educational role of this tool and the research, the researcher conducted the data analysis independently and after moving to another institute.
For the purposes of this study and to ensure saturation, the researcher selected 15 residents from the population of 31 reviewed residents. This was a purposive sampling. For a resident to be included, recordings from both the attending physicians’ sessions and the residents’ sessions had to be available and of good quality in terms of content and duration. Further, the researcher aimed to include a diversity of clinical cases and a similar number of residents from each class. Each video review took between 30–45 minutes.
Interactions between residents or faculty during the respective review sessions were audio recorded, and each session followed the same pattern. First, videos were presented in two- to three-minute segments before pausing for discussion, which was moderated using opening questions or specific prompts. Examples of opening questions are “What do you see?” “What do you think?” and “What else could or should be done?” Examples of prompts include “How was the health condition managed?” and “What else could or should be done to manage this health condition?” All review session recordings were transcribed verbatim.
The analysis was completed using the systematic strategy of critical theory-based qualitative methods outlined by Carspecken,26 which includes four steps: 1) Writing low-level codes: These are word-tags listed to mark segments of the text and required little abstraction, so they stay close to the primary records. 2) Constructing meaning fields: This step includes articulating the possible meanings inferred from statements made by the participants. The field often consists of statements separated by “and” or “or.” 3) Developing a validity reconstruction: This reconstruction involves explicating meanings that can be understood and described in terms of validity claims. Validity claims are statements or utterances that are criticizable in the sense that they need reasons to be justified. They, in turn, can serve as reasons for other claims. This step clarifies assumptions and makes explicit presumed norms (i.e., normative claims) and positions about truth (i.e., objective claims) as well as feelings and intentions (i.e., subjective claims). 4) Organizing themes: This step involves an iterative, inductive and deductive process. The main concepts were developed and organized to answer the two main questions of this paper.
The researcher used Microsoft Word for coding and Microsoft Excel for the thematic organization. To enhance the study’s trustworthiness, the researcher used low inference in coding to maintain proximity between the content and the themes of analysis. The researcher also practiced reflexivity by writing memos on the positions of teacher and doctor separately and by discussing the interpretations with peers. The researcher has a master science in clinical research and did this work as part of his Ph.D. dissertation under the supervision of qualitative research advisors.
The 15 resident sessions that were reviewed included (5 female residents; 5 residents from each class year. A total of 24 residents (8 females; 8 from each class) and 10 attendings (6 females; 7 MDs, 2 psychologists, and 1 nurse practitioner) participated in the conversations and were included in the analysis.
From the perspectives of both attendings and residents, there are six critical norms of practice expected of residents:
1) Residents should prepare for the visit before entering the room. The participants valued knowing that a resident had reviewed charts and the paper slip that patients completed before the appointment. One attending commented, "I think it was good that he referenced the stuff the patient had already filled out. I think that lets the patient know that he’s taken time to review her stuff."
It was found it unacceptable when residents appeared to be unaware of why a patient was there or only guessed at the reason. One resident reflected, "I went to the computer and was trying to [look up] more information. I was thinking maybe I should have been more prepared before I stepped in." On the other hand, some residents argued that with a packed schedule, being prepared can be difficult. One resident explained, “sometimes it’s tough to look up 20 patients, and know everything.”
2) Negotiating the agenda for the visit is recommended. The participants valued residents starting a visit by setting the agenda, which can be done by illustrating and clarifying reasons for the appointment. An attending noted, "I think overall he’s done pretty well setting an agenda for the visit. He was finding out if the patient has any concerns and then setting a plan." Residents noticed that when there was no clear agenda, the visit would sometimes go out of control. A resident reflected, "[Five minutes into the visit] I was noticing that I didn’t have a set agenda. So, it needed to happen then since it wasn’t flowing naturally from the conversation."
3) Eliciting sufficient patient history is essential. The participants valued when residents gathered necessary information by asking the right questions. Taking a good history can mean screening for symptoms and risk factors, especially those associated with serious consequences or mental illnesses, which are often overlooked. It can also mean asking about social history and weight loss as well as medications. Residents and attendings both pointed out suboptimal practices in this regard. For example, one attending remarked, "I mean, if somebody has a [Patient Health Questionnaire] PHQ of 25 you have to ask, ‘Do you want to hurt yourself? Do you want to kill yourself or anybody else?’ ” Regarding the same incident where suicidality was overlooked, the resident reflected, "(I should have obtained) more information and gotten a full history. I had no idea she was suicidal. I completely glanced over the PHQ9. I didn’t catch that."
Participants thought it was not useful to ask open-ended questions with patients who are not linear. On the other end of the spectrum, they also found it inappropriate to fail to ask for more explanation or assess symptoms in a rigid, formulaic way. While attendings and residents sometimes agreed about what to ask and how, that was not always the case. For example, an attending once remarked, "I think [the resident] is patient, I probably would not ask any open-ended questions. I’d ask yes/no type questions, especially when she starts off on those tangents." But, during the resident reviewing the session, the resident had a different position and explained, "I think I got all the information that I wanted, but it took a little while. I could have potentially just kept asking/redirecting questions."
4) Patients should have proper examinations. The participants valued a resident telling patients what to expect before a thorough exam, with attention to aspects that would be relevant to the problem at hand. They also appreciated when findings were explained in real-time. One peer resident remarked, "She explained what she was doing [in] real-time. She initially started with medical lingo, but then broke it down a bit to more layman’s terms."
Both residents and attendings paid attention to inappropriate examination techniques, and attendings were particularly concerned about insufficient exams or lack of diligence. An attending pointed out, "Probably not getting nose to nose with the patient when you do your eye exam. When he pulled away, he checked her left eye with his right eye!"
5) Making the diagnosis correctly is of primary importance.The participants expressed satisfaction when residents considered differential diagnoses. Residents were particularly impressed when a peer had recognized the root of the problem even with a poorly presented history. All the participants criticized incorrect diagnoses or relying solely on history without an adequate physical exam. Regarding one incident, the attending said, "If the patient comes in to us and the fracture’s not healing, I’m not thinking diabetes. (Another attending) I’m thinking, ‘Where’s your splint?!’ ” During the resident session, the resident shared that he had given less than optimal management plan. They said, "I did the x-ray, and it was a month out, and he still, it wasn’t healing, so then I thought maybe he had diabetes. They gave him a finger splint, so I told him to use it at work to prevent any further injury. After work, I told him to take it off and try exercising a little bit!"
Residents and attendings did not always maintain an agreement. For example, while attendings criticized a resident not being thorough, a resident mentioned reliance on gestalt feelings. For example, incident, the attending once was concerned, "He sort of nailed in on carpal tunnel very fast, which I think is probably at the very top, differential. It’s almost as if he had the diagnosis in his mind, and he was trying to confirm it with the questions that he asked." But, during the resident review session, the resident explained, "I was going with more of a gestalt feeling about it as opposed to following history, physical, etc. It doesn’t sound like there was any traumatic history. It doesn’t sound like there was any compromise in terms of function, no ER visits, no x-rays, nothing that was red flags to me at that point. I felt comfortable doing that."
6) Carrying out proper management of health conditions. The participants expected residents to prescribe the right medicine, develop a targeted plan, and do whatever else was needed, including providing information and counseling patients appropriately. Participants noted that residents should make their thoughts explicit to patients. They also should ensure that communication is at levels appropriate for the patient’s educational attainment. For example, an attending remarked, "I think it was good that she used specific instructions and wrote them down for the family." Regarding the same interaction, a peer resident commented, "The counseling was good. It was specific and giving clear instructions at the appropriate educational level."
The participants criticized when residents did not thoroughly address problems, gave hurried or inaccurate recommendations, or developed plans without measures to ensure follow-through. They also considered a resident’s avoidance of complicated health problems unacceptable. One resident reflected, “He has a panic attack, but I don’t really want to address that, because he’s not going to get benzos from me. That’s what was going through my mind. ‘Let’s just try and glaze over this’ because I don’t want to give him benzos. I recognized it as a panic attack, but I also recognized it in my head that that would warrant benzos, which I don’t want to give him. I wanted to gloss over that and say, ‘Are you still seeing your counselor? Good. She’ll take care of that.’ Which isn’t good."
Attendings further criticized residents for failing to use specific communication strategies. However, the reviewed residents were occasionally able to rationally justify not using a particular approach within the context of the interactions with the patients. For example, an attending exclaimed, "It would be helpful to have the patient use teach-back, ‘Can you tell me what the top three things are that I’ve asked you to do to take care of this wound?’ ” But, the resident provided contradictory, yet valuable insight when he said, "[The patient] was finishing my sentences, almost. That’s how I knew, ‘Okay, he knows what I’m saying.’ ”
The analysis identified five themes related to tensions between differing values.
1) Addressing problems while building relationships. The participants judged visits heavily on the success of addressing problems while simultaneously building a relationship as a goal in and of itself. Attendings considered that residents did well if they spent time addressing what mattered to a patient’s health, while residents gave differing explanations for omissions in this regard. One attending remarked, "In a patient you’ve never seen before, you’re trying to build this relationship from the start. It would be helpful if you were showing that you were engaged and passionate about what’s going on with the patient." Regarding the same session, the reviewed resident commented, "I felt that it was important to address that problem being that it’s my first OB visit with her. I felt that if I focused on the pregnancy and were not addressing her complaint, I probably wouldn’t be doing justice to her. Also in terms of building that rapport, it would probably reflect poorly on that."
2) Performing duties within a reasonable time frame. Participants considered residents to be doing a good job if they dealt with multiple problems or built rapport in a short period of time. They recognized the potential conflict between time constraints and building rapport. They appreciated the residents who were time-efficient without compromising rapport. Participants criticized time-consuming habits such as using pen and paper for notes or attempting to address all problems in one visit. Regarding one visit, an attending commented, “It’s just a lot longer than it needs to be. It was an hour by the time he went to talk to [a preceptor]." The reviewed resident had a similar position when they saw their video, "I completely agree with [peer]. The stuff I spent 10 minutes talking about, I could have probably done three minutes. I can improve on that."
3) Completing tasks on the computer without compromising conversation. The participants extensively discussed the use of computers in exam rooms. They looked positively on residents who asked for a patient’s permission before using the computer. Further, they thought that computers are best kept to one side so they would not form a barrier and residents could stay engaged while typing. Residents who are adept in this area maintain eye contact and sit facing patients, preferably at eye level. While they complete visit-related tasks on the computer, they maintain a genuine conversation, evidenced by reassuring sounds, paraphrasing, using reflection, and using teach-back, among other techniques. Participants agreed that computers could be particularly helpful for sharing results such as X-rays with patients or for obtaining information, such as medication dosages. However, they criticized when residents stared at the computer, acted in a hurry, or took detailed notes rather than brief reminders, especially when patients shared intimate information. For example, one attending commented, "It feels like the visit is not focused on her as much as. He is on his computer, and I feel that ‘I am trying to get my notes done,’ that’s the impression I got. ‘Just want to get over it and get done,’ kind of feeling." In a separate session, the reviewed resident remarked, “I’m wondering, ‘What am I looking at?’ All the time, when I’m looking at on the screen. Yeah, I look at it too much."
4) Supporting patient autonomy while asserting boundaries. The participants accepted that residents can lay the groundwork for expectations and set boundaries and that residents can practice medicine only within the limits of their values and beliefs. They noted that residents could respectfully say no to prescribing certain medications, such as controlled substances. Here is one attending observing, "He may be trying to lay all this groundwork so he can say, ‘This is why I’m not going to give you what you want.’ The patient’s using [drugs]. ‘I’m prepping you for this. [This] is why I’m not going to do what I already said I wasn’t going to do.’ ” Similarly, a peer remarked on the same encounter that his peer, "[was] setting boundaries."
Participants expected residents to acknowledge a patient’s interest in treatment. Residents should support patients in making appropriate changes when the patient is ready. Participants felt that patients should be given options so that doctors can identify their true preferences. Here is one peer resident commenting, "She figured out the patient preference and gave her all the options." Also, participants criticized residents for not confronting patients who avoided engaging in a serious conversation about their health.
5) Acting autonomously without jeopardizing patient care. The participants all understood that residents could encounter situations where they do not know what to do. In these cases, they expected residents to formulate plans and, if they are unsure how to proceed, ask their attending. Here is one example as an intern reflected, "[The patient is] on two SSRIs! I think I was a little confused at that point. I was like, ‘why! [But] this is my first visit, who am I to question them. But, this doesn’t seem right’. I went to go follow up with one of the attendings and get their opinion. In the back of my mind, I’m like ‘I don’t think you should be on two different SSRIs.’ ”
Residents were criticized for not seeking instruction with complicated conditions and, instead, making decisions on their own. On the other hand, attendings criticized residents who gave up their autonomy by deferring to supervising doctors for decisions. The participants expected residents to rely less on preceptors as they advanced in their training. They also pointed to situations when engaging the attending could have empowered residents to advocate for theier patient. Without seeking support, the resident was left to mange patients alone, and at times, with substandard plans. In one example, the attending noted, "She goes and shoots herself, and they find out she was just at the doctor’s office - ‘cause I’m sure [the preceptor] didn’t get this as, ‘Hey! She’s got major depression with suicidal ideation.’ ” In another example, the attending noted, "It seems like the patient is saying, ‘This is painful, I’m not able to be on [light duty], there is no such thing as light duty unless he [was] totally taken off of the floor, which requires a physician’s note, which he’s never had. Everything he does cause more trauma to his finger. It seems strange that the patient was saying, ‘I need your help. I need your help!’ And [the resident] is like, ‘Oh well, okay, well…’ ”
To our knowledge, this is the first study to synthesize the viewpoints of both attendings and residents regarding the norms of practice and the tensions between expected values in primary care residency training. Residents’ actions are better understood and judged from the participants’ perspectives and not the observers’. The responses demonstrate that a resident’s work is judged adequate when it simultaneously upholds many values, and only a participant or a person taking the position of a participant can elaborate these norms and judge an action’s adequacy. These insights are often missed in studies that rely on observations alone because making explicit these norms and values is only possible through participants’ reflections and critique.
Prior research has focused on presenting the work of doctors as sets of technical skills that can be judged, often out of context, by an observer. Our study indicates that residents must be judged in the context of unique interactions with specific patients. Further, we show that, rather than the application of one specific skill, what matters to residents and attendings is action judged in its own context. Roter et al. suggested that evidence of improved outcomes can be seen in the increased use of certain types of questions (open-ended vs. closed-ended, relationship building statements, etc.) and letting patients talk more.7 Our study demonstrates that what determines adequacy is a complex process that involves the evaluator taking the position of the acting doctor and contemplating what she could or should do in that specific patient interaction. Thus, our first practical recommendation is to implement more dialogical and reflective approaches in residency training instead of relying solely on observation and feedback. Previous research has shown that less than optima practices take place while residents are training;27–29 our study complements prior knowledge in this area and adds to understanding what kinds of omissions residents make. In our study, residents were criticized because they did not adequately perform some of the basic tasks of the visit, such as not eliciting sufficient information from a suicidal patient, not diagnosing the reason for a patient’s fracture not healing, or at times not giving the right advice. They also acknowledged that they, accidentally or consciously, did not address the patient’s concerns. Furthermore, at times they aimed to address too many problems and ended up with a partially adequate job for a few complaints. The most common theme behind these criticized practices, however, was that residents often did not know what they did not know. Residents are building competencies, but they are not always competent. Our study showed that when faced with uncertainty, residents quite often acted and made decisions without asking for support from a supervisor. Therefore, our second practical recommendation is to advance the conversation on supervision strategies that directly involve a more competent other in the care of patients, even when residents advance in training.
Our study has many strengths. Unlike studies that relied on standardized patients or structured settings,11,30–32 we used actual patient interactions in the natural context. When including the perspectives gained from the video recordings to allow for observation of exactly what took place, this research method provided more depth in the data collection methods, especially with the triangulation of perspectives from attending physicians. However, our work is not without limitations. First, the study was done in one institution, and learning can only be transferrable with caution to other institutes that may have different cultures and norms. Second, while we believe the 15 included patients represent a fairly robust sample, as a naturally occurring experiment, the selection of cases may have been determined by the contingency of what patients came into the office on the days of recording.
This project is not yet complete. Future work will aim to explore how learning takes place when residents engage in a model of reflection and dialogue over time. We expect to see resident performance developing to better adhere to the explicated norms. We also anticipate seeing more robust critical reflections and dialogues on these norms and values.
Participants in the conversations, both residents and attendings, addressed practice norms, and highlighted their struggles around the tensions between different values. As this model of video reviews engages residents in genuine reflections and conversations with peers, it provided a promising framework for learning and self-assessment.
Funding: none
Author’ contributions: MA is the sole author of this work.
Acknowledgments: The author wishes to thank Phil Carspecken, Barbara Dennis, William Philips, Chad Lochmiller, and David Estelle. The paper has been previously used as a section of a PhD thesis in Inquiry Methodology at Indiana University. Defense date: 4/20/2018.
Competing Interests: The authors declare no competing interests.
Ethics approval and consent to participate: The study was reviewed and granted an exemption by the Indiana University IRB in August 2016. Reference number: 1607804494. As an educational research, verbal consent was obtained. The work was presented in a residency meeting to inform participants. The IRB approved this way of obtaining consent because the work is primarily educational, and the research part is regarding conversations already carried out for education.
Consent to publish: not applicable.
Availability of data and Materials: The datasets used and/or analysed during the current study available from the corresponding author on reasonable request.