The principles of patient-centered care have become widely accepted as central to the planning, delivery, and evaluation of health care. These include understanding the whole person, finding common ground, enhancing the clinician–patient relationship, and exploring health, disease, and the illness experience (Table 1).1 Furthermore, the medical education community, including accrediting agencies, are increasingly recognizing the importance of incorporating the principles of patient-centered care into medical education. To illustrate, one essential criterion Australian medical schools must meet for full accreditation specifies that “Learning and teaching methods in the clinical environment promote the concepts of patient-centred care.”2 Accordingly, one need is to develop effective strategies and methods to practically integrate these principles into medical curricula and training programs, including in the assessment of learners.
Medical educators are rapidly adopting the use of entrustable professional activities (EPAs) in the context of competency-based medical education.3 EPAs, defined as “a fundamental unit of professional practice that can be fully entrusted to a trainee once he or she has demonstrated the necessary competence to execute the specific activity unsupervised,”4 are increasingly forming the cornerstone of postgraduate training programs. EPAs focus on clinical tasks and provide a unique opportunity to integrate the principles of patient-centered care into medical assessment. In this study, we approached the growing field of EPA development using the principles of patient-centered care.
Given that only a limited selection of EPAs are intended to encompass what is core to a profession, it is imperative that a rigorous approach is taken in developing the content of each. No matter how carefully chosen, EPAs will be unusable and invalid without clear and relevant “EPA descriptors,” which previous authors have defined as specifications and limitations of the activity; expected knowledge, skills, and attitudes; associated competencies; sources of information to assess progress; and the level of entrustment.5 Published procedures for EPA development highlight the importance of engaging stakeholders, including senior clinicians, residents or fellows (i.e., “training doctors” in the Australian context), and course directors, in defining these descriptors.5,6 Currently, medical educators are using qualitative research methods such as interviews, focus groups, and Delphi procedures to ascertain stakeholder views when designing EPAs.4,5,7 Mapping EPAs to a curriculum containing a framework of competencies is also an established process in EPA development.4
To our knowledge, no studies have investigated the potential role of patients as stakeholders in EPA development. Consulting patients during the process of developing the content of EPAs aligns with the widely accepted principles of patient-centered care.1,8 According to the patient-centered care framework, patients and clinicians meet as complementary partners who bring different knowledge, needs, and concerns to the clinical encounter.9 An EPA developed after consulting with patients (as well as health care providers) could therefore represent both patient and clinician expectations of trainees; such an EPA could direct these physicians-in-training toward providing more patient-centered care.
Patients are a central stakeholder group in health care, so EPAs developed without their involvement may fail to capture what is truly important to patients themselves. EPA developers such as expert clinicians and trainees may implicitly acknowledge the importance of patient-centered care, but their views do not directly capture the patients’ perspectives and may be inaccurate.10–12 Research is needed to determine how to develop EPAs that align with the expectations of patients and clinicians alike, and that thus support best clinical practice.
Investigators studying EPA development have typically chosen a common, important clinical task as an exemplar to investigate and describe a particular method of EPA development.6,13,14 We chose “management of acute low back pain (LBP)” as our platform for researching the development of an EPA. We designed the EPA specifically for rheumatology fellows in training. LBP causes more global disability than any other condition,15 and managing it is a common, important task for doctors in multiple medical and surgical specialties.16
We reviewed the literature to identify previous research on the illness experiences and expectations of care for patients with LBP. Although research in this field is extensive, most studies have focused on chronic LBP in the outpatient setting.17 We concluded that to develop a relevant patient-centered EPA for hospital-based trainees caring for inpatients experiencing acute LBP, consulting such patients—those admitted to hospital with acute LBP—would be necessary.
In this qualitative study we aimed to examine what patient input adds to clinician input when designing EPAs. Our three research questions were as follows: (1) What are patient and clinician expectations of trainees when performing a common, important task such as managing acute LBP? (2) What are the similarities and differences in expectations between these groups? (3) What is the additional value of patient input when defining components of EPAs, including the descriptors (i.e., the specifications and limitations of the activity; expected knowledge, skills, and attitudes; associated competencies; sources of information to assess progress; and the level of entrustment)?
We conducted our investigation in accordance with the Standards for Reporting Qualitative Research (SRQR).18
Although the general philosophy of patient-centered care is widely accepted in health care, several different published definitions and models exist.19 For this study, we adopted Stewart and colleagues’ patient-centered framework.1 We considered Stewart and colleagues’ model to be the most appropriate for several reasons: (1) Although originating in the context of family medicine, the framework was designed to be relevant to all disciplines of medicine, as well as to nursing and allied health professions; (2) the framework remains up-to-date and is continuously revised, incorporating findings from recent studies in the field of patient-centered care1(pp4–7); and (3) Stewart and colleagues’ model comprises well-defined components as outlined in Table 1. This framework informed the participatory study design and our analysis of the data.
We conducted a qualitative interview and focus group study at a large 850-bed tertiary referral and teaching hospital in Sydney, Australia, from March through July 2015. We selected this setting because it features a busy emergency department servicing a large population and it offers on-site neurosurgical and rheumatology specialists likely to receive patients with LBP.
In the Australian health care system, medical graduates may commence internal medicine training after completing 24 months of generalist nonspecialty rotations. After an additional 24 months, physician trainees who are successful in their internal medicine examinations progress to an advanced training program in their subspecialty of choice and are referred to as “advanced trainees” or “fellows in training.” The generic term “training doctor” refers to a medical graduate who has not yet completed a postgraduate training program, but we have used the word “trainee” throughout this report for consistency.
Researcher characteristics, ethical approval, and study design
In accordance with SRQR guidelines, we have described our own roles and backgrounds here. C.E. is a rheumatologist with experience caring for patients with acute LBP. C.E. did not provide medical care for participants during the study period. W.H. is an experienced qualitative researcher and family physician. A.D. and H.P.M. are both rheumatologists. All four of us have substantive roles in medical education.
The South Western Sydney Local Health District Human Research Ethics Committee approved this research.
We conducted this study in two stages, illustrated in Figure 1: (1) Consulting patients and clinicians, and (2) Writing the EPA, based on patient and clinician input.
Stage 1: Patient and clinician views on trainee doctors managing acute LBP
Clinician focus group.
We invited clinicians from all health professions associated with the rheumatology service, including a musculoskeletal radiologist, to participate in a 60-minute multidisciplinary focus group. We invited them via e-mail and in person, and we offered no incentives. We chose the focus group method to enable the development and identification of ideas shared by participants through facilitated group discussion.20 Our use of a focus group also supported the exploration of team relationships within the multidisciplinary service. The discussion was audio recorded and transcribed verbatim.
During the focus group, participants discussed components of EPA design, as described by Aylward and colleagues6 (e.g., identifying barriers to assessment and possible settings for assessment). Next, we explored clinicians’ views on what they expected of a competent specialist trainee in rheumatology when caring for patients with acute LBP (see List 1 for focus group questions).
Questions for Clinicians and Patientsa
Key questions for clinicians participating in a focus group
- Consider the scenario of a patient attending the emergency department with acute low back pain. What would you expect a competent advanced trainee to do in this situation from beginning to end?
- The patient is admitted for further management. What would you expect of the trainee through the inpatient admission and on discharge?
- What would you consider required knowledge in this area? What reference or guideline would you recommend the trainee refers to in determining level of knowledge required?
- What makes a good trainee in terms of manner with patients? What would you be worried about?
- Where is this EPA learned, practiced, and could be assessed? What are some challenges to assessment?
- What methods or sources of information could be used to assess trainees?
Key questions for patients participating in one-on-one interviews
- Can you describe your experience when you had your episode of low back pain, from beginning to end?
- Which doctors were involved in your care? What did each doctor do or say?
- In your understanding, what types of doctors work in hospitals (e.g., junior, senior)? What do you expect from each “level” of doctor in managing your back pain?
- What qualities in a doctor do you consider important when treating patients with low back pain?
- Can you recall any particularly helpful things your doctor said or did? How would you have liked your management to be different?
- What advice would you give a training doctor on how to treat a patient with low back pain?
aThe authors conducted the focus group and interviews in 2015 to garner, respectively, health care provider and patient input into caring for patients admitted for acute low back pain to develop an entrustable professional activity for rheumatology trainees.
We chose the interview method for patient consultations to explore in-depth the range of patient experiences and perspectives. Adult patients (> 16 years old) who were admitted to the hospital between April and July 2015 with a primary diagnosis of acute LBP were invited to participate by the health care professionals treating them. Because we were designing this EPA specifically for hospital-based trainees managing acute LBP in the inpatient setting, we excluded outpatients from the study.
We defined acute LBP, in accordance with previous research, as an episode of activity-limiting LBP (with or without pain referred into either lower limb), lasting for a least one day, and present for less than three months.21 We continued to recruit patient participants until we reached thematic saturation of patient-generated themes.
After obtaining informed written consent, we conducted semistructured telephone interviews with patient participants one to two weeks after their discharge from the hospital. We developed interview prompts (see List 1) from the LBP literature15–17,21 and iterative discussion with an expert panel comprising family physicians and rheumatologists. Interview topics covered an account of the hospital admission and the patient’s interactions with and expectations of health care professionals, particularly trainees. Interviews were audio recorded and transcribed verbatim.
We applied the Framework Method to summarize, compare, and systematically analyze interview and focus group data from clinicians and patients22 using qualitative data analysis software (NVivo QSR, Version 10, Melbourne, Australia). The method provides a structured and transparent approach to categorizing and comparing qualitative data between different participant groups. The defining feature of the Framework Method is an explicit “matrix output” where rows represent cases, and columns represent coding categories. Each cell in the matrix contains summarized data, making responses between different participants directly comparable whilst retaining a sense of how their responses fit within the dataset as a whole. We followed the stages of data analysis described by Gale and colleagues22 and outlined them below.
After reading transcripts to become familiar with them, one of us (C.E.) coded the data using both predefined coding categories informed by the theoretical framework of patient-centered care, and open coding to categorize new themes as they emerged. Next, we organized the codes and data into a matrix as described above, with study participants along one axis, and coding categories along the other. Coding categories classified data as follows: (1) patient and clinician expectations of care provided by a training doctor, and (2) patients’ hospital admission and illness experiences. Each matrix cell contained a summary of input from each participant group within a coding category, hyperlinked to the actual data. One of us (W.H.) supervised the processes of data collection and analysis throughout the study, whilst two of us (A.D. and H.P.M.) independently reviewed the framework matrix once data collection was complete. All of us then reviewed the cells to identify final themes, as well as similarities and differences between patient and clinician input.
Stage 2: Developing the EPA
Beginning with a blank EPA template, we methodically reviewed Stage 1 results to determine how patient and clinician input could inform each EPA descriptor. As mentioned, descriptors include (1) specifications and limitations of the activity, (2) expected knowledge, skills, and attitudes, (3) the most relevant domains of competence, (4) sources of information to assess progress, and (5) the level of entrustment.4,5 During this process, we continually referred to the framework matrix and themes from Stage 1. For example, if patients expected doctors to be gentle and caring, we included “being gentle, showing care” as an attitude in the “knowledge, skills and attitudes” descriptor. Following established procedures in EPA development, we also used the interview and focus group findings to map the EPA to relevant competencies in the Royal Australasian College of Physicians (RACP) rheumatology advanced training curriculum, the training standard for trainees in our study setting.23
We included any content in areas where clinicians and patients had similar views in the EPA. If the views of the two groups were discordant, we applied Stewart and colleagues framework of patient-centered care1 to incorporate patients’ ideas. For example, in keeping with Stewart and colleagues’ “finding common ground” principle, the task description would include the explicit expectation that trainees discuss the relevant issue with patients.
We returned the Stage 1 findings and draft EPA to all study participants for feedback, either in written format, or through telephone interviews and face-to-face meetings. We incorporated their feedback into subsequent versions of the EPA before finalizing it.
To address the question of what patient input adds to an EPA, we tracked, throughout the process of developing the EPA, whether themes informing the EPA descriptor originated from patients or clinicians, or both. This tracking enabled us to identify EPA components specifically informed by patients.
Stage 1: Patient and clinician views on management of acute LBP
The focus group comprised 4 rheumatologists, 2 rheumatology clinical nurse specialists, 2 rheumatology trainees, a physiotherapist, and a musculoskeletal radiologist. The 10 focus group participants ranged in age from 24 to 60. Of these 10 participants, 4 were men.
We reached thematic saturation of patient-generated themes after interviewing 14 patients: 9 self-identified women and 5 self-identified men. Three patients declined to participate in the study (1 cited feeling unwell, and 2 did not specify a reason). The mean age of patients was 61 (range 36–87).
We describe the views of clinicians and patients below in three themes: (1) diagnosis and treatment, (2) teamwork and practicing under supervision, and (3) compassionate care and communication. For each theme, we firstly outline the areas of overlap between clinician and patient expectations of trainees, and then describe areas where patients provided distinct and unique views.
Diagnosis and treatment: Areas of overlap.
Clinicians and patients alike firstly expected trainees to arrive at an accurate diagnosis and to appreciate the impact of LBP on the patient’s life. Clinicians clearly articulated the required level of knowledge expected of trainees to safely treat patients with acute LBP. Such knowledge includes awareness of current LBP management guidelines, epidemiology of LBP, lower limb myotomes and dermatomes, and clinical pharmacology of common analgesic medications. Further skills specified by clinicians included lower limb neurological examination and interpretation of spinal imaging. Both patient and clinician participants strongly emphasized the importance of excluding serious underlying pathology as illustrated by the following comments:
I would like them [trainees] to eliminate all possibilities … is it from your back, is it possibly cancer? (Patient 5)
I want them to take a history that included … red flag features, weight loss, fever, constitutional symptoms … (Rheumatologist 2)
Both groups also emphasized the importance of functional recovery in the management of LBP. Patients and clinicians expected trainees to facilitate treatment that relieves pain and helps patients to return to the premorbid activities of daily living.
I’m a mother, and I want to get back to doing everything. (Patient 4)
Whatever the cause, we aim for the same thing … to reduce pain and improve mobility. To get them back to the function which was there. (Rheumatologist 3)
Ensuring safe discharge from hospital and continuity of care was also important to clinicians and patients alike. Both groups highlighted providing patient education, specifying clear medication plans, and recommending outpatient follow-up as key components of care.
Diagnosis and treatment: Unique input from patients.
We noted discordance between patients and doctors regarding the role of spinal imaging in diagnosing acute LBP. Patients perceived history and examination as providing insufficient evidence to explain their symptoms, and they expected doctors to investigate with imaging studies. On the other hand, clinicians, citing current guidelines,24 felt trainees should organize spinal imaging only if signs indicated they were necessary. One patient’s comments illustrate the patient perspective:
[T]hat’s what I was really upset about. You’re in the hospital, you’re there. I know it’s busy and there’s people worse off, but if you’re there and you’ve got this excruciating pain, it would have been really nice if you’d been taken away and X-rayed, and before you left home you could say, here you are, the X-rays shows this, don’t do this. I have absolutely no idea why they didn’t send me for an X-ray. (Patient 4)
Another difference between patients and clinicians was that patients expected their follow-up appointments to be organized and booked, rather than simply receiving a general instruction to see their usual doctor or to “return if things get worse”—as illustrated by two patients’ comments:
That’s what we are disappointed with … we had no papers, no follow-up, nothing from the hospital. They said goodbye and that was it. (Patient 3)
They just gave me a letter to take to my local doctor, and mentioned I should see a physiotherapist. But no actual plan … my hospital experience wasn’t satisfying, to be honest. (Patient 11)
Teamwork and practicing under supervision: Areas of overlap.
Clinicians expected trainees to consult supervisors regarding significant clinical decisions, particularly if they could affect patient safety. They expected trainees to provide accurate information to supervisors, and they would check to ensure this was happening. One physician explained:
You need time of observation, direct observation, and that builds your trust in terms of the clinical decision.… Does your assessment concord with what you’ve been told by the trainee? Is there reliability as far as you can tell? (Rheumatologist 3)
In concordance with clinicians’ expectations, patients acknowledged trainees’ limited experience and their need to work under supervision. They expressed trust in the health care system to adequately teach and supervise trainees. Patients expected trainees to regularly communicate with their supervisors when making decisions. They noted:
[I expect] … that they’re competent, and they know what they’re doing and have been properly trained. (Patient 10)
They’re learning, I can’t expect them to make the decisions, they listen to the specialist … they’re people, everyone can make errors. (Patient 5)
I would’ve hoped that the junior doctor may have gone to their senior doctor and said, “Look, there is some extreme pain there, I think we should run some scans.” (Patient 12)
Teamwork and practicing under supervision: Unique input from patients.
Members of both the patient and clinician groups highlighted the importance of trainees functioning as multidisciplinary team members. In addition to physicians and trainees working in high-functioning teams, however, patients expected to be informed and updated of the team’s decisions and plans—in other words, to be a part of the team themselves, actively participating in their own care. For example, one patient commented,
If the doctor had come in and said, “I’ve had a discussion with the physiotherapist,” if there was teamwork, that would have been great. (Patient 4)
Compassionate care and communication: Areas of overlap.
Clinicians as well as patients highlighted the importance of trainees demonstrating care and consideration towards patients; that is, both groups valued showing empathy, comforting, listening to patients’ concerns, and communicating clearly.
Compassionate care and communication: Unique input from patients.
Although clinicians highlighted the importance of compassionate care in principle, patients drew on their personal experiences to provide vivid descriptions of desired qualities and behaviors in trainees, as illustrated below.
It’s something they should have learnt by then … if you’re going to be a doctor, you’ve got to talk to people and you’ve got to have empathy.… Not just say, ‘Oh, it’s just back pain.’ (Patient 4)
Of note, patients wanted trainees to be as gentle as possible when conducting a physical examination given their severe pain. When patients felt they were not treated with compassion and respect by health practitioners, they reflected angrily on their experience. One patient remembered:
He said, “You’ll be seen to in due course … you’ve only got sciatica.” How the hell does he know what I’ve got, when he didn’t even ask? Every time I talk about this, it just makes me very angry … it makes my blood boil. (Patient 1)
Furthermore, patients wanted trainees to respect and take their personal circumstances and social context into account when addressing their medical problems. One patient commented:
I just felt like I was nuisance I suppose, I just don’t think they took much notice of me. As a person I mean. As an elderly woman. (Patient 3)
Patients with LBP who had negative experiences with trainees almost invariably cited the problem of poor patient communication. They wanted doctors to discuss in detail their diagnosis and proposed management plans. They also wanted specific education and information about their LBP prognosis and treatment. Two comments illustrate the importance of clear, detailed communication:
You know, that’s where my disappointment was with them, because of the lack of communication. (Patient 5)
When I went onto the new medication, they explained what it was and how it would relieve me … it helped me feel confident that they knew what they were doing. (Patient 13)
Stage 2: Developing the EPA
We used these findings—based on patient and clinician input—to write an EPA titled “Management of acute LBP.” Table 2 provides explanations of the approach we took to develop each descriptor, highlights components informed by patient interviews, and explains the origin (patient or clinician input) of each EPA descriptor.
To our knowledge, this is the first study to explore the role of patient input in EPA design. We have demonstrated the unique value of patient input and have described an effective and rigorous method to incorporate patient expectations of trainees into an EPA. A qualitative research approach enabled us to draw on the unique knowledge, needs, and concerns of patients and clinicians alike.
We found that patients and clinicians had several overlapping expectations of trainees, which directly informed our EPA descriptors. Examples include excluding serious underlying pathology causing LBP and focusing on functional recovery when planning the treatment.
We also noted, however, that multiple components of the managing acute LBP EPA were primarily informed by findings from the patient interviews. Drawing from their firsthand experience of LBP, patients reported distinct expectations of trainees which were not reported by clinicians.
One example is the “Required attitudes” section of the EPA. Whilst clinicians referenced generic qualities such as compassion toward patients, patients provided detailed descriptions of desirable and observable behaviors (e.g., share the plan with the patient, be gentle during the physical exam). Descriptions of undesirable behaviors, such as being dismissive of patient concerns, were equally useful in determining the required trainee attitudes. These measurable qualities inspired by patients highlight the utility of patient input in EPA development. In sum, patient input allowed us to describe in practical terms how trainees should be expected to apply the principles of patient-centered care in the context of acute LPB.
Notably, we observed discordance between clinician and patient views regarding the role of spinal imaging, which is consistent with at least one other study investigating patient expectations in LBP care.25 We sought to reconcile this lack of agreement by using a patient-centered approach. That is, rather than simply recommending that trainees follow accepted imaging guidelines, the EPA specifies that trainees should discuss the role of spinal imaging with patients to help overcome the difference in care expectations. In Stewart and colleagues’ patient-centered care framework, this process is known as “finding common ground,”1 and it is crucial to achieving consensus with patients in developing a care plan. For the EPA, we chose the word “partnering” to highlight the importance of trainees working with patients to reach shared treatment goals. The EPA thus guides trainees to consider the patient’s perspective, and directs supervisors to assess this component of the task.
Our finding that patients have expectations not reported by clinicians supports the literature suggesting that physicians cannot accurately identify patients’ concerns without directly enquiring about them.10–12 Our approach also addresses an accepted criterion for good assessment—that is, to include patients’ contributions in the shared understanding of competence and performance.26
In line with the principles of patient-centered care, we held the patients’ perspectives in the same regard as those of clinicians, equally valuing their different expertise. Charles L. Bardes9 has eloquently captured the value of this approach: “Patient and physician must therefore meet as equals, bringing different knowledge, needs, concerns, and gravitational pull but neither claiming a position of centrality.”
Interestingly, during our patient interviews, we gained insight into how patients make entrustment decisions. Patients placed trust in the health care system as a whole, as well as in individual trainees, on the assumption that the latter were adequately supervised. Some patients did not feel trainees were consulting their supervisors enough, and in these cases, they expressed their dissatisfaction and lack of confidence in the trainee. This finding raises important questions for future work. For example, how may an understanding of patients’ entrustment decisions be incorporated into training programs where program directors seek to produce patient-centered doctors? Can patients act as assessors for EPAs? Indeed, another set of investigators has explored and applied the potential role of patients in assessing doctor–patient interactions outside the EPA context.27
EPAs were designed as a novel assessment tool in competency-based medical education, which authors have defined as “an approach to preparing physicians for practice that is fundamentally oriented to graduate outcome abilities and organized around competencies derived from an analysis of societal and patient needs”28 [emphasis ours]. We propose that consulting patients should be an integral part of determining what should be in an EPA by incorporating empirical measures of what patients view as their actual needs.
Further supporting the incorporation of patient perspectives in EPA design are recommendations by key groups such as the UK General Medical Council, which has asserted that “it is widely agreed in principle that patients should contribute to deciding what outcomes medical students need to meet in order to graduate.”29
We acknowledge several limitations in our study. The approach we took was time-consuming as it entailed multiple steps—collecting and analyzing data, iteratively reviewing and confirming the EPA—however, given that relatively few EPAs can play a central role in the final stage of a specialty training program, we feel that a rigorous process is justified. Also, we submit that primary data collection may not be required where there is published research on the patient perspective for particular conditions or tasks.
Interviewing patients post discharge may engender recall bias; nonetheless, we chose this approach (1) because most participants were initially in severe pain requiring opioid analgesics, interfering with their ability to interview; (2) to ensure participant confidentiality and privacy; and (3) to encourage participants to speak freely outside the episode of care and have the opportunity to reflect on their hospital experience.
We acknowledge that individual clinicians’ perspectives may not be identified in a focus group setting; however, focus groups represent a time-efficient and rigorous way to capture initial views as well as an effective means of exploring team relationships and communication, which turned out to be important to patients. Also, each clinician had the opportunity to provide individual feedback on the Stage 1 findings and draft EPA.
Finally, this was a single-site study, with a specific patient population. Although the specificity facilitated an in-depth understanding of the patient perspective, these factors limit the generalizability of our findings. Despite this, we found patient and clinician expectations of trainees aligned closely with relevant domains of competence in the RACP training curriculum.23
Our methods were intentionally rigorous so that we could demonstrate the unique value of patient input in EPA design; however, we acknowledge different ways to potentially develop “patient-centered EPAs.” At a curriculum development level, patient-centered frameworks such as the “Compassionate, collaborative care model” can be mapped to EPAs, as recently proposed by Lown and colleagues.30 For individual EPAs, reviewing the literature on the relevant patient illness experience and patient expectations can be very valuable. If primary data are required, consumer organizations may be a helpful resource for reviewing or helping to develop EPAs. Furthermore, already-developed EPAs can be continually revised on the basis of feedback from both supervisors and patients. Using Stewart and colleagues’ framework1 to review existing EPAs and incorporate published data may provide insights that are relevant to patient-centered care. In this study, we found that directly interviewing patients enabled the translation of a theoretical model of patient-centered care into measurable and assessable—practical and concrete—EPA descriptors. Ultimately, the optimal method for including patient input will be determined by the EPA topic, existing literature, and available resources.
In conclusion, patients provide valuable and unique input that often complements that of clinicians. Such input can be incorporated into practical, observable descriptors of an EPA. Our methodology has led to the development of a more patient-centered EPA, aligning best clinical practice with patient expectations. Other medical educators, seeking to apply the principles of patient-centered care to EPA-based assessment in other professional activities and settings, may want to consider adopting our approach. Further research is needed to test the effect of using a patient-centered approach on the design of other EPAs and, in turn, on learning.