A Trainee-Led Quality Improvement Project Using Change Management Theory to Improve Bedside Rounding

Our institution is a university-affiliated, tertiary care children’s hospital with 160 pediatric resident physicians (“residents”). The hospital has ∼650 inpatient beds and >29 000 admissions annually. Residents rotate on 4 general pediatrics teams for 2- to 4-week blocks and are overseen by 1 to 2 pediatric hospital medicine attending physicians (“attendings”). Residents are present in-house 24 hours per day, 7 days per week. General pediatrics teams admit newborn to young adult patients with a variety of general pediatric conditions. Resident subspecialty teams (eg, cardiology, rheumatology, etc) and attending-only general pediatrics teams were excluded from this project.

The initial team selected for the intervention has patients primarily located on a single unit with a maximum of 16 patients. The team is staffed by 2 to 3 second-year residents and 1 attending. The 3 additional teams have a different structure, in which first-year residents are supervised by second- and third-year residents. Some teams have patients housed in 1 single unit, whereas others are responsible for patients on multiple units within the hospital.

In the fall of 2021, the rounding structure primarily involved the team conducting rounds in the hallway outside patients’ rooms with variable participation from patients, families, and nursing staff. First-year residents were expected to review data and examine the patients before rounds. Physical examinations were often performed by the attending before or after rounds without the full team present. In an initial stakeholder assessment, attending physicians reported the frequent need to change the plan after independently examining patients because of the discovery of new information.

The QI project team was led by pediatric hospital medicine fellows and included residents and care team assistants (CTAs) with the mentorship of a pediatric hospital medicine attending. CTAs are nonclinical members of the inpatient care team who provide administrative support and collect data for multiple unit-based and hospital-wide quality initiatives. The team also received support from a QI analyst who assisted with project management and QI methodologies. The team applied the Institute for Healthcare Improvement Model for Improvement and selected Dr John Kotter’s “Eight Steps to Transforming Your Organization” as a change management theory to support the adoption of QI interventions (Table 1).^14,16 

Key stakeholders were identified, including residents, attending physicians, and representatives from the institution’s Family Advisory Council (FAC) and Youth Advisory Council (YAC). The FAC consists of 15 to 20 individuals identified as experienced family partners from many different backgrounds. The YAC consists of ∼30 adolescents who are current or previous patients at the hospital. The goal of the 2 councils is to provide patient and caregiver perspectives to improve patient- and family-centered care.

Stakeholder input was collected through a combination of individual meetings, focus groups, and electronic surveys. After meeting with leadership from both the FAC and YAC, a targeted focus group guide was generated. Notably, the FAC focus group included representation from family members who use a language other than English for health care communication. Questions were modified for an attending focus group after a discussion with QI leadership from the Section of Hospital Medicine. Discussion with team members and residency program leadership informed the development of a REDCap (Research Electronic Data Capture, Nashville, TN) survey, which was administered to all residents.

Stakeholder feedback informed the creation of a process map and driver diagram (Fig 1). Three primary drivers of FCR were identified: knowledge, culture, and logistics. The first primary driver, knowledge, highlighted the lack of a common definition of FCR, variable stakeholder experience in conducting rounds at the bedside, and a lack of awareness of the practice of FCR from patients and families. Rounding away from the bedside became an institutionalized practice for both attending physicians and residents during the coronavirus disease 2019 pandemic, informing the second primary driver of culture. Logistics, the third primary driver, included barriers such as lack of bedside space, patients located on multiple floors, and concern for limited personal protective equipment availability. Test of change ideas to address key drivers were identified by team members and stakeholders. An impact/effort matrix was used to select interventions, prioritizing those anticipated to have a high impact.

Twelve PDSA cycles were performed from December 1, 2021 to February 28, 2023 (Table 2) to address the key drivers of knowledge, culture, and logistics. Knowledge-related interventions focused on education (eg, faculty development lecture, preservice electronic communication). Culture was addressed through interventions that shared, normalized, and institutionalized the team’s operational definition of FCR (eg, highlighting success stories of champions, spreading to additional units). Logistics-related interventions were designed to reduce barriers to implementation (eg, electronically distributed tip sheet for residents and attending physicians).

CTAs directly observed rounds and recorded the following data in an Excel (Microsoft Corporation, Redmond, WA) spreadsheet for each patient: bedside round completion, caregiver attendance, nursing attendance, and duration of rounds. Data were collected on weekdays from October 2021 to February 2023, except national holidays. Patients admitted to services other than general pediatrics were excluded. The trainee team met with a QI analyst biweekly to review data. Data from October 1, 2021 to December 1, 2021 were used to establish a 2-month baseline proportion of bedside rounding, which was compared with that from the intervention period (December 2, 2021–February 28, 2023).

From March 10, 2022 to February 28, 2023, attending physicians and residents on the initial team were invited to provide feedback regarding their perception of bedside rounds anonymously through a voluntary REDCap survey. In addition to addressing rounding location preference, the surveys assessed perceptions of the impact of bedside rounds on daily workflow, the need to change the patient’s plan of care later in the day, and communication with patients and caregivers. Attending physicians were asked about the impact on teaching. Residents were asked if bedside rounding decreases family questions in the afternoon. If an attending or resident rotated on a team multiple times, they could have completed >1 survey.

The primary outcome measure was the average weekly percentage of bedside rounds. Rounds were considered to be conducted at the patient’s bedside if the resident and attending entered the patient’s room and discussed the assessment and plan at the bedside, regardless of caregiver presence. For families that use a language other than English, bedside rounds were counted if interpreter services were used during rounds. For each team, the average weekly percentage of bedside rounds was calculated as the total number of rounds conducted at the bedside divided by the total number of patients per week.

The average weekly percentage of caregiver attendance on rounds was a secondary outcome measure. Caregiver attendance was defined as a caregiver being present during the patient’s assessment and plan discussion, whether virtually, at the bedside, or at another location. Patients without a caregiver present were excluded from this measure. Balancing measures included the average weekly percentage of nurse attendance on rounds, the average daily rounding time, and the perceptions of bedside rounds by residents and attending physicians. The average weekly percentage of nurse attendance on rounds was defined as a nurse joining at any time during rounds to discuss the patient, regardless of location. Average daily rounding time was defined as the total time in minutes from the start of rounds for the first patient to the completion of rounds for the last patient.

The outcome and process measures were analyzed by using statistical process control p-charts. The balancing measures were analyzed via a p-chart (nursing attendance) and an x-bar chart (rounding time). Special cause variation was defined according to standard Associates in Process Improvement rules.¹⁷  The balancing measures of attending and resident perceptions were monitored monthly via descriptive statistics of Likert-scale responses and free-text responses to the survey were reviewed.

This QI initiative was reviewed and determined to not meet the criteria for human subject research by the hospital’s institutional review board.

For the initial team, a total of 13 522 patient days were included in the analysis, including 1723 patient days during the baseline period and 11 799 during the intervention period. The average weekly percentage of rounding at the patient’s bedside increased from 18% to 86% after PDSA cycle 2 and reached the highest percentage at 89% at the start of the next academic year in July 2022, revealing improvement consistent with special cause variation (Fig 2). The average weekly percentage of caregiver attendance increased from 66% to 99%, and subsequently to 100% after PDSA cycle 2, revealing improvement consistent with special cause variation (Supplemental Fig 4). The average weekly percentage of nursing attendance during rounds decreased from 79% to 73% after PDSA 2 (Supplemental Fig 4). The average daily rounding time increased from 131 minutes to 166 minutes after PDSA cycle 2, revealing an increase consistent with special cause variation, which was sustained throughout the project (Supplemental Fig 4).

We received 32 survey responses from attending physicians for the initial team with a response rate of 70% (Supplemental Fig 5). Seventy-seven percent indicated a preference for rounding at the bedside compared with the previous model (ie, hallway rounds) with 23% indicating no preference (neutral). Per the qualitative attending responses, perceived benefits included improved workflow, quality of the daily plan, communication with the health care team and patients and families, and increased opportunities for teaching and feedback. Thirty-one survey responses were received from residents (response rate 37%). Feedback on rounding preference was variable, with 48% preferring bedside rounds, 13% neutral, and 39% preferring the previous model. Resident perceptions regarding decreased afternoon interruptions and improved overall workflow were similarly variable; however, 81% of resident responses indicated bedside rounds reduced later changes to the plan and 68% felt bedside rounds improved the quality of the plan. A common theme among the open-ended resident survey responses was that their experience of FCR varied depending on the attending. Residents identified more positive experiences when attending physicians demonstrated effective modeling or teaching on FCR, and conversely identified negative experiences when attending physicians were less facile at redirecting conversations with families or managing time on rounds.

Spreading the intervention to the remaining 3 teams resulted in an increase in weekly bedside rounding on each respective team, with an overall increase from 7% to 54% (Fig 3). Notably, a decrease in the average weekly percentage of bedside rounding that resulted in special cause variation was observed during a period of high hospital census on the initial team and Team 2 in the winter of 2022 (Figs 1 and 3).

By combining a QI framework and change management principles, we increased the percentage of bedside rounds from 18% to 89% on an initial team and sustained change over 12 months, exceeding the first SMART aim goal of 50%. When spread to the remaining 3 teams, we achieved the second SMART aim and demonstrated an increase from 7% to 54% for all general pediatrics resident teams.

Although previous QI work has highlighted the benefits of change management, to our knowledge, this is the first example of combining Kotter’s 8-step theory with the Model for Improvement in the initial project design. With this approach, we not only achieved the staged SMART aims but increased resident team membership from 9 to 16 members (10% of the residency program) with representation from all resident classes. This QI initiative reveals the successful incorporation of change management theory to promote culture change and impact a primary outcome.

On the initial team, the practice of bedside rounding increased rapidly. We hypothesize that the early success was due to the project design being grounded in 2 established frameworks. Whereas the Model for Improvement enabled the team to identify goals, measures, and interventions, the use of change management theory allowed the team to proactively avoid common pitfalls regarding change (eg, lack of buy-in, lack of sponsorship). We believe this approach influenced the project’s sustainability because of our focus on culture change, which will enable continued improvement after all involved trainees have graduated.

For the secondary outcome measure, we predicted an increase in caregiver attendance at rounds, which was observed regardless of rounding location. Because only patients with caregivers present were included in this measure, the move to the bedside naturally increased caregiver attendance. Nursing attendance decreased, which was likely multifactorial. Because the initial team cared for patients on >1 unit, coordinating with nurses outside of the primary unit was challenging. In addition, rounding in the hallway may have made rounds more accessible and provided a visual cue to nurses that the team was rounding. Initial key drivers focused on physician achievement of bedside rounds and increasing nursing presence on rounds likely requires interventions targeting nursing-focused drivers.¹⁸  In addition, delayed engagement with nursing stakeholders and high baseline nursing attendance preintervention (79%) may have impacted this measure. Notably, the baseline nursing participation was consistent with previously cited averages of nursing attendance on FCR.³ 

The duration of rounds increased by an average of 2.2 minutes per patient, consistent with previous FCR studies.^4,19,20  However, the increased time burden to residents may have been offset by the deferment of preround physical examinations for eligible patients in PDSA cycle 2. By deferring preround examinations, timely plans were made on the basis of team bedside examinations. Subsequently, residents reported a decreased need to change the plan later in the day. Attending physicians also described an improved workflow.

Although the project was trainee-driven, resident rounding preferences were variable throughout the intervention. Residents noted that a positive rounding experience depended not only on conducting rounds at the bedside but also on the attending physician’s proficiency and comfort with FCR. However, the low resident response rate (37%) may not be representative of the perceptions of the entire residency class. Additionally, we did not account for duplicate responses when attending physicians and residents rotated on the team multiple times throughout the project time period, therefore, possibly inflating the response rate.

The generalizability of the findings may be limited because of characteristics specific to the initial intervention team (eg, resident experience level, small rounding team). The spread of the interventions to the remaining teams resulted in a variable increase in bedside rounding, likely reflecting team-specific characteristics and the need for ongoing intervention. Perceptions of bedside rounding from nurses, patients, and families were not obtained. Including these perspectives may contribute to a more holistic understanding of the intervention. Although measures including bedside rounds, caregiver attendance, and nursing attendance are components of FCR, each measure was examined independently; therefore, the primary outcome measure of bedside rounds serves as a proxy for FCR. We did not include the measurement of additional elements of FCR (eg, removal of medical jargon, age-appropriate involvement of patients) or components that may enable assessment of rounding quality. Lastly, this project may be challenging to replicate without a designated team member to complete data collection.

The next steps include the spread of the intervention to resident subspecialty teams to standardize bedside rounds on all resident teams and the optimization of other aspects of FCR quality, including nursing participation. Additionally, patients and families who use a language other than English for health care communication experience further variability in FCR.^21–24  Future work may explore a deeper understanding of the patient and family experience, especially for those who communicate using a language other than English.

Through our thoughtful project design, we achieved a sustained increase in the percentage of bedside rounds on an initial general pediatrics resident team and successfully spread to 3 additional general pediatrics resident teams, indicating a culture change within the Section of Hospital Medicine. This initiative reveals that the incorporation of a change management theory within a QI framework may facilitate intervention adoption and outcome sustainability.

We wish to acknowledge Melanie Katrinak and Megan Roman for their project management and QI support, Kristin McNaughton for her support of manuscript development, and Dr Rebecca Tenney-Soeiro for her mentorship.