The Effect of Bedside Rounds on Learning Outcomes in Medical Education: A Systematic Review.

PURPOSE: To determine if bedside rounds, compared with other forms of hospital ward rounds, improve learning outcomes in medical education.
METHOD: For this systematic review, the authors searched Ovid MEDLINE, Embase, and Scopus from inception through February 20, 2020. Experimental studies were included if they (1) compared bedside rounds to any other form of rounds in a hospital-based setting, and (2) reported a quantitative comparison of a learning outcome (e.g., learner reaction, knowledge, skills, behavior, health care delivery) among physicians-in-training (medical students, residents, fellows). Extraction elements were summarized using descriptive statistics and a narrative synthesis of design, implementation, and outcomes.
RESULTS: Twenty studies met inclusion criteria, including 7 randomized trials. All studies involved resident physicians, and 11 also involved medical students. The design and implementation of bedside rounds varied widely, with most studies (n = 13) involving cointerventions (e.g., staff education, real-time order entry).Of the 15 studies that reported learner satisfaction, 7 favored bedside rounds, 4 favored the control, and 4 were equivocal. Of the 4 studies reporting an outcome of learners' knowledge and skills, 2 favored bedside rounds and 2 were equivocal. Of the 8 studies that reported on learner behavior (e.g., bedside communication with patients), 5 favored bedside rounds, 1 favored the control, and 2 were equivocal. Finally, of the 14 studies that reported a health care delivery outcome (e.g., teamwork, rounding time), 8 favored bedside rounds and 6 were equivocal. Due to the high risk of bias and unexplained heterogeneity across studies, the overall strength of evidence was low.
CONCLUSIONS: In hospital-based settings, learners' satisfaction with bedside rounds is mixed. However, bedside rounds appear to have a positive effect on learner behavior and health care delivery. Given their potential value, additional research is needed to identify barriers to and facilitators of educationally successful bedside rounds.

For more than a century, bedside rounds—the process during which a physician–teacher and a learner discuss a hospitalized patient’s condition and plan of care in the patient’s presence—have been viewed as an ideal learning method in the hospital. [1-3] The practice of bedside rounds may provide opportunities for skill development, observation and feedback, and teamwork and communication while contributing to a culture of patient-centered care. [4-6] Additionally, the educational benefits of bedside rounds are rooted in principles of adult learning. For example, the theory of situated cognition emphasizes that learning is inseparable from doing and, by extension, that new knowledge is best acquired in a context where that knowledge can be readily applied. [7] As it applies to hospital ward rounds, situated cognition supports that learning to care for patients is best achieved in the presence of an actual patient.

Despite the rhetorical and theoretical benefits of bedside rounds, recent research has indicated that its use is decreasing. In the mid-20th century, even before the advent of bedside rounds as a specific educational strategy, teaching at the bedside was the norm. [8] However, a 2009 study showed that only 17% of teaching during ward rounds occurred in the patient’s room. [9] Time constraints, particularly in the context of resident duty-hour restrictions and an increasing reliance on technology, have driven educators and learners to seek other more “efficient” forms of rounding, such as conference room or hallway rounds. [5,10-13] More recently, efforts to reduce COVID-19 exposures, such as physical distancing, made bedside rounds infeasible.

The decline in bedside rounds has led to debate about their role as a teaching method in medical education. Experts and leaders have argued passionately for rounds to return to the bedside. [14,15] Others, including learners, have questioned the educational benefits of bedside rounds and whether bedside rounds should continue to be viewed as a standard of educational practice. [16,17] Given the present discourse around bedside rounds, a synthesis of evidence may help define their value in medical education and inform areas for future research. Therefore, we conducted a systematic review to answer the following question: in hospital-based settings, do bedside rounds improve learning outcomes compared with other forms of ward rounding in medical education?

Method

This study was conducted in accordance with the Preferred Reporting Items for Systematic Review and Meta-Analysis Protocols (PRISMA-P). [18] All steps in the study selection and data abstraction processes requiring subjective judgment were conducted independently and in duplicate by study investigators, with disagreements resolved by consensus.

Search strategy, eligibility criteria, and selection process

This systematic review was a planned extension of an existing review to assess the effect of bedside rounds on patient-centered outcomes. [19] With the aid of an experienced medical librarian, we created a search strategy using terms for setting or context (e.g., hospital, hospital medicine, inpatients), activity (e.g., rounds, bedside rounds, presentation), and participants (e.g., attending, resident, intern, medical student). We searched Ovid MEDLINE, Embase, and Scopus from inception through February 20, 2020. Details of our search strategy are available in Supplemental Digital Appendix 1 at http://links.lww.com/ACADMED/B223.

To ensure a comprehensive overview of bedside rounds medical education, we used broad eligibility criteria. Specifically, we selected the following definition of bedside rounds proposed by Gonzalo et al [20]:

A minimum of 2 physicians (including resident/fellow physicians-in-training) performing all 3 of the following in the presence of a hospitalized patient: (1) case presentation/history, (2) performance of at least one physical exam skill, and (3) discussion of the patient’s daily plan of care.

We included studies if they (1) used an experimental study design (randomized or quasi-experimental) comparing bedside rounds with any other form of ward rounds, (2) involved medical trainees (graduate or undergraduate), and (3) reported a quantitative learning outcome. Learning outcomes were categorized using a modified version of the framework described by Kirkpatrick and Craig, [21] as follows:

Level 1: learner reaction (e.g., satisfaction with rounds, satisfaction with teaching);

Level 2: learner knowledge and skills (e.g., standardized testing scores, performance during a simulated patient encounter);

Level 3a: learner behavior (performance “with real patients in a clinical context,” [22] e.g., assessment of resident communication at the bedside); and

Level 3b: health care systems/processes (e.g., duration of rounds, nurse–physician teamwork).

Given that previous systematic reviews have reported on the effect of bedside rounds on patient outcomes, we decided not to include studies that reported only patient-level (level 4) outcomes. [19,23] We reviewed full-text articles and conference abstracts, regardless of language or date of publication. If there was insufficient information to decide on inclusion, we contacted the study authors for clarification.

Data abstraction

We created a data abstraction instrument in Microsoft Excel 2010 (Microsoft Corporation, Redmond, Washington) based upon related research regarding interventions in hospital wards and medical education. [19,24-26] The tool was tested and iteratively refined. Extraction items included year, participants and setting, description of the bedside rounds intervention and its implementation (including cointerventions), evaluation methods, and learning outcomes.

We assessed methodological quality using tools adapted from the risk-of-bias assessment tool for nonrandomized controlled trials [27] and the Cochrane risk-of-bias assessment tool for randomized controlled trials. [28] These tools included assessments in standard categories such as confounding, selection bias, outcome measurement, and selective reporting.

Data synthesis

We initially planned to thematically group results within each Kirkpatrick outcome level for quantitative pooling and meta-analysis. However, early in the screening process, we identified a large degree of variability across studies regarding measures, even within similar outcome domains. As such, instead of quantitative pooling, we summarized the extraction elements using descriptive statistics and a narrative synthesis of design, implementation, and outcomes.

Results

Characteristics and quality of included studies

Our search strategy yielded 5,576 abstracts, 284 of which underwent full-text review (see Figure 1). Of these, 20 articles met full inclusion criteria. [20,29-47] Characteristics of the included studies are outlined in Supplemental Digital Appendix 2 at http://links.lww.com/ACADMED/B223. The majority of studies were conducted in the United States (n = 16, 80%) and used a nonrandomized experimental design (n = 13, 65%). All of the studies involved residents (n = 20, 100%), and many also involved medical students (n = 11, 55%). Additionally, many studies were interprofessional (n = 14, 70%) and were conducted on general medicine wards (n = 9, 45%).

Figure 1

Preferred Reporting Items for a Systematic Review and Meta-Analysis Protocols (PRISMA-P) [18] flow diagram for a systematic review of the literature on the effect of bedside rounds on learning outcomes in medical education, inception to February 2020. aTotal exceeds the number of excluded articles because some were excluded for more than 1 reason. bDefined as a quantitative outcome that could be categorized into Kirkpatrick level 1 through level 3b.

The majority of studies (n = 17, 85%) demonstrated a high risk of bias in at least 1 domain, and 7 studies (35%) were at high risk for bias in multiple domains. [31,32,39,41-43,45] Supplemental Digital Appendixes 3 and 4 (available at http://links.lww.com/ACADMED/B223) summarize the methodological quality of the included studies. The 2 most frequent sources of bias were deviation from the intended intervention and selective reporting, which were present in 10 (50%) of the studies. Deviation from the intended intervention was often related to low adherence rates to bedside rounds interventions and unbalanced cointerventions between intervention and control groups. Selective reporting was typically related to reporting 1 of several potential outcome measures and analyses. For example, many studies used surveys as the primary measurement instrument and reported item-level analyses without a prespecified analytic plan.

Description and implementation of bedside rounds

The structure and process of bedside rounds varied widely among the 18 (90%) studies that described their implementation. For example, McNeil et al described a bedside rounds intervention that involved bedside shift handover between incoming and outgoing emergency resident physicians and had a mean duration of 17.5 minutes per round. [40] Alternatively, Gonzalo et al described a bedside rounds intervention that involved a case presentation from the resident, physical examination, and discussion of the patient’s plan of care with an average duration of 16 minutes per patient. [20]

In addition to the variability of the process of bedside rounds, there was also variability in how bedside rounds were implemented as part of a broader set of interventions. The majority of studies (n = 13, 65%) implemented bedside rounds with 1 or more cointerventions, including staff education/training (n = 8, 40%), [20,31,34,36,39,41,43,46] communication script/safety checklists (n = 5, 25%), [31,35,43,45,46] prerounds huddle/determination of rounding order (n = 4, 20%), [30,34,35,39] and real-time order entry during rounds (n = 4, 20%). [36,39,42,46] Cointerventions were most often not balanced between intervention and control rounds—that is, the bedside rounds group often had additional interventions but the control group did not—leading to the potential for bias when trying to assess the specific contribution of bedside rounds to the outcome.

Six studies [20,30,34,36,39,40] (30%) reported an adherence rate for bedside rounds; the median adherence rate among these studies was 67.9%, with 2 studies reporting an adherence rate > 90% [30,40] and 1 study reporting an adherence rate < 50%. [20]

Effect of bedside rounds on learning outcomes

Level 1: Reactionary outcomes.

Fifteen studies (75%) reported a reactionary learning outcome (Table 1). [20,29-33,35,37-40,43,44,46,47] Outcomes and measures in this domain varied widely but often related to satisfaction with or preference for bedside rounds, or to the perceived educational value of bedside rounds. The results of 7 of these studies favored the intervention, 4 favored the control, and 4 were equivocal. The results were mixed in the 3 studies with a moderate risk bias as well as 12 studies with a high risk. For instance, in the study by Landry et al, [37] learners reported less comfort asking and being asked questions during bedside rounds but no difference in overall satisfaction, and in the study by Seo et al, [47] nearly all learners preferred control rounds. Alternatively, in the study by McNeil et al, [40] learners reported significantly higher satisfaction with the quality of education during bedside rounds.

Table 1

Summary of Articles Examining Reactionary Learning Outcomes (n = 15) in a Systematic Review of the Effect of BR on Learning Outcomes, February 2020

Level 2: Learner knowledge and skills.

Four studies (20%) reported an outcome within the domain of learners’ knowledge and skills; 2 favored bedside rounds [42,43] and 2 demonstrated equivocal results (Table 2). [30,33] Three of the studies used a subjective measure (i.e., self-reported outcome measures addressing trainees’ perception of learning), [30,33,42] and the other used an objective measure (i.e., medical students’ shelf exam scores). [43] All studies reporting an outcome in the knowledge and skills domain were deemed to be at high risk of bias.

Table 2

Summary of Articles Examining Learner Knowledge and Skills (n = 4) in a Systematic Review of the Effect of BR on Learning Outcomes, February 2020

Level 3a: Learner behavior.

Eight studies (40%) reported the effect of bedside rounds on learner behavior, with the majority (n = 5, 63%) favoring the bedside rounds intervention while 1 favored control, and 2 were equivocal (Table 3). [29-31,33,40,41,44,46] Outcomes in this domain most often included self-reported ratings of communication skills by learners. Only McNeil et al [40] used an objective measure of learners’ behavior by having trained observers categorize the amount of time residents spent discussing differential diagnosis and other specific aspects of clinical care during rounds.

Table 3

Summary of Articles Examining Learners’ Behavior (n = 8) in a Systematic Review of the Effect of BR on Learning Outcomes, February 2020

Level 3b: Health care delivery.

Fourteen studies (70%) measured the effect of bedside rounds on health care delivery, [20,29-31,34-36,39,40,42-46] with the majority (n = 8) favoring the intervention and the rest being equivocal (n = 6) (Table 4). The most common outcomes in this domain were the duration of rounds (n = 8) and interprofessional teamwork (n = 8), particularly between nurses and physicians. Of the studies reporting on the duration of rounds, 4 were equivocal [20,39,40,42] and 4 favored bedside rounds. [29,31,34,46] Of the studies measuring teamwork, 7 favored bedside rounds [30,31,34,42-45] and 1 was equivocal. [35]

Table 4

Summary of Articles Examining Health Care Delivery (n = 14) in a Systematic Review of the Effect of BR on Learning Outcomes, February 2020

Discussion

We found that bedside rounds may improve learners’ knowledge and skills and have a positive effect on learner behavior and health care delivery in clinical learning environments. When compared with alternative forms of ward rounds, bedside rounds appear to improve both the communication skills of medical students and residents and the teamwork culture on hospital units. However, learners’ perception of and satisfaction with bedside rounds is mixed. These findings should be interpreted with caution, given the differences in the design and implementation of bedside rounds, variable adherence to the intervention, and the high risk of bias within the studies.

The results of this review build upon a body of literature highlighting the benefits of bedside teaching for the development of clinical knowledge, skills, and behaviors in medical education. Specifically, previous work by Peters and ten Cate [1] and Aldeen and Gisondi [13] have reviewed the literature regarding bedside teaching across a variety of clinical learning environments, both inpatient and outpatient. These reviews reached similar conclusions that rounding at the bedside offers distinct advantages for teaching certain clinical skills, such as history taking and the physical examination. Additionally, both articles articulated the potential for bedside encounters to promote learner professionalism and patient-centered communication. The results of our review provide more definitive evidence that bedside rounds, when compared with conference room or hallway rounds, improve the communication abilities of medical students and residents.

Despite these encouraging results, perceptions of bedside rounds are mixed among physician trainees. Our results reinforce an important and frequently observed dichotomy in medical education: while students and residents acknowledge the value in the presence of the patient, they often prefer to learn away from the bedside. [6,48-50] Both the learners and their supervising physicians have noted an underlying discomfort during bedside rounds. [6,11] The reasons cited for preferring to round away from the bedside vary but include worrying about making a mistake and looking incompetent or about causing patient discomfort; lacking freedom to ask questions and discuss particularities of a clinical case; and having concerns regarding efficiency. [48-50] It is encouraging that the results of this review, and others, [19,23] provide evidence that bedside rounds do not negatively affect patients or prolong rounds and may, in fact, improve these outcomes. However, understanding learners’ internal reluctance will be important if the educational benefits of bedside rounds are to be fully realized.

Interpretation and implications of the findings

Our objective at the outset of this review was to answer a pragmatic question asked by many clinician–educators: is the effort of bedside rounds worth it? However, the variable and in some cases weak evidence for bedside rounds suggests that it may be more useful to shift the focus from “do bedside rounds work?” to deeper questions of “when, how, and why do bedside rounds work?” [51] To answer these questions, we suggest that clinician–educators and education researchers apply conceptual frameworks to make sense of the variability in outcomes following bedside rounds.

For example, one way of interpreting the mixed reactions of trainees toward bedside rounds is through the lens of self-determination theory (SDT), a dominant motivational theory in the field of psychology, which posits that motivation (defined here as “the general desire or willingness ... to do something”) exists upon a spectrum from amotivation to external motivation to internal motivation. [52] According to SDT, a person’s motivation for a specific activity is dependent on how that activity meets 3 psychological needs: (1) autonomy, (2) competence, and (3) relatedness, and the greater the degree to which an activity meets these needs, the higher the level of intrinsic motivation a person will have for that activity. [52]

Clinical teachers can use the SDT framework to understand and foster learners’ motivation to participate in bedside rounds. [53] Specifically, attending physicians should consider using methods that support learners’ autonomy, competence, and relatedness during bedside rounds. For instance, to promote a sense of autonomy, support the attending–trainee relationship, and demonstrate respect and partnership in the decision-making process, they could engage learners in a dialogue regarding rounding strategy at the start of a clinical rotation and solicit learners’ preferences for rounds. Role modeling the processes of bedside rounds may also drive intrinsic motivation by promoting a sense of competence, particularly for medical students. Identifying additional barriers to and testing facilitators of motivation for bedside rounds are important directions for future research.

The results of this review also reveal that there is wide variability in the design, implementation, and outcomes of bedside rounds. To illustrate, consider the following comparison: the bedside rounds described by McNeil et al [40] involved only physicians and were conducted during shift turnover in an emergency department, whereas the bedside rounds described by Huang et al [34] involved an entire interprofessional health care team and included a checklist of essential safety items and a focus on preparation for dismissal. When considering these disparate structures and processes, it should be no surprise that the outcome measures and results achieved varied. A useful framework for making sense of this variability is the structure-process-outcome (S-P-O) model of health care quality first described by Avedis Donabedian more than 50 years ago. [54] As the model applies to bedside rounds, the structure of the bedside encounter (e.g., personnel involved, available technology, ergonomics) and the processes occurring during the encounter (e.g., case presentation, safety checklist, physical examination) will directly affect the outcomes achieved.

Viewing bedside rounds through the S-P-O framework suggests that teachers and learners should be careful to clarify the desired learning outcomes of bedside rounds. Physicians and trainees (along with patients and interprofessional health care team members) should ask themselves what they are trying to accomplish in doing bedside rounds. Is the primary goal of this bedside rounds to introduce the patient to new health care team members? Is it to meet with a family caregiver to clarify the nuances of a patient’s history of present illness? Is it to ensure that junior trainees learn the proper technique for an important part of the physical examination? By beginning with the end in mind, clinicians, learners, and patients can create a shared mental model of bedside rounds and avoid unnecessary tensions resulting from competing objectives. Once the members of the team have identified that desired outcome, they can work backward to ensure the structure and processes of bedside rounds support it.

Limitations

This review has limitations. First, we focused specifically on bedside rounds involving physician learners in the hospital setting. As such, our results cannot be applied to other clinical settings (e.g., the outpatient office) or different populations of learners (e.g., nursing or pharmacy students). Second, because bedside rounds are a heterogeneous construct, we made subjective judgments during article selection and data abstraction that may be a source of bias. However, we attempted to mitigate this risk of bias by developing a rigorous article selection and data abstraction process that included duplicate review. Third, we only included articles reporting quantitative learning outcomes and did not evaluate qualitative research studies.

Conclusions

Bedside rounds appear to positively affect learner knowledge and skills, but learners seem to have mixed preferences for bedside rounds compared with other forms of ward rounding. Given the variety of ways that bedside rounds can be implemented, future research should work to identify how specific design features of bedside rounds can be used to facilitate desired learning outcomes.

Acknowledgments

The authors wish to thank Patricia Erwin, MLS, for creating the search strategy and Danielle Gerberi, MLS, AHIP, for updating the search.