Increasing High-Value, Cost-Conscious Care Family Rounding Discussions Via an Educational Rounding Tool

The study used a single-center, nonrandomized, prospective pre/postinterventional design assessing HV3C FCR discussion performance, using a previously published and validated HV3C performance measurement tool,²²  before and after implementing HV3C Rounding and Quick Reference tools on a general pediatric hospitalist teaching service. We also created anonymous pre/postintervention surveys to assess providers’ perceived competence practicing and role-modeling HV3C, impact and usefulness of the tools, and suggestions for further tool refinement.

The project focused on the inpatient general pediatrics hospitalist teaching service, mainly during the 2022–2023 academic year (August 1, 2022–August 1, 2023), at a metropolitan, 136-bed, full-service children’s hospital with 7000 annual admissions.

With the exception of a few specialties, the service primarily manages a wide variety of conditions and complexity admitted to general- and intermediate-level care²³  units (∼2700 admissions/consults yearly). The service’s 3 teams each consist of a faculty hospitalist, senior resident, and 2 interns, plus rotating pediatric hospitalist fellows and medical students, which conduct in-room FCRs with the nurse and family as available (case management, social work, dietician, child–life interdisciplinary rounds occur with the medical team as a separate “table round” process before FCRs). A member of the medical team is logged onto the patient’s electronic medical record, entering discussed care-plan orders in real time. As part of the groups’ existing quality improvement endeavors, each team’s FCRs undergo a weekly convenience sampling of performance audits (sampling expectation: ≥2 unique patient rounding encounters per team per week), performed by trained, rotating hospitalist faculty not on service and departmental research coordinators, tracking initial versus subsequent patient encounters, whether rounds occurred inside the patient’s room, family presence, nursing presence, and participation in conversations, plus other factors.

All 20 hospitalists (15 pediatric, 5 medicine–pediatric) and that academic year’s trainees (33 pediatric, 32 medicine–pediatric residents, and 4 hospitalist fellows) participated in the project’s HV3C FCR performance assessment according to their inpatient general pediatric rotation schedules. At any given time, 1 to 2 transitional year, emergency medicine, or family physician residents also rotate through the service, who were included in the team-based FCR HV3C performance assessments but were excluded from the survey arm. Incoming 2023–2024 interns and fellows not present for the project’s majority also were excluded from postintervention surveys. Otherwise, all hospitalist faculty, remaining 2022–2023 fellows, and pediatric and medicine–pediatric residents were eligible to participate in the postintervention surveys.

Before the Rounding Tool’s introduction, respective trainee and faculty HV3C curriculums had been launched, centered around a recurring “high-value equation.” Topics included an introduction to HV3C (including health policy and payer components), clinical quality (including diagnostic testing), costs, and patient preference that occurred during a series of trainee noon lectures and hospitalist faculty development sessions, respectively. Sessions occurred during the project’s baseline performance period.

The Rounding Tool poster was constructed on the basis of a review of HV3C educational needs assessment, best-practice literature, and existing curricula at the time,^{9–17,24–28}  plus a cross-mapping of a 3-year data trend of the most common Current Procedural Terminology codes billed at our institution under residency teaching services and their respective, publicly posted cost data, as required by the Centers for Medicare and Medicaid Services that hospitals publicly provide pricing information.²⁹  The Rounding Tool poster was drafted as a bulleted list/conversation guide, and divided into quality, diagnostic testing, medication, care coordination, shared decision-making, and cost topics (Fig 1) that correlated to the above curriculum and cross-mapped to assessed HV3C performance measures²²  (see below). Draft versions underwent additional formatting and content review by the research team, plus the hospitalist group before implementation.

The Quick Reference was designed as an easily-accessible, distilled companion to the Rounding Tool in the form of a small, 2-sided laminated card that participants could keep in back of their name tags, which contained a quick response (QR) code to the Rounding Tool poster, and key HV3C FCR terms/reminders presented in mnemonic format (CHEETAH, after a similar, animal-themed mnemonic of another existing, educational project), plus condensed cost data from the Rounding Tool (Fig 2). The Quick Reference also underwent hospitalist group review and approval.

A week before launch, participants were oriented to and received the Rounding Tool and Quick Reference at respective hospitalist administrative and residency program director meetings. In addition to being available via the Quick Reference’s QR code, the Rounding Tool was also posted in resident and hospitalists work rooms, plus available on a network drive accessible to all participants, containing other commonly referenced educational and clinical protocols.

HV3C FCR performance measurement occurred over an ongoing 6-month each, with baseline and intervention periods using a 10-item, similarly named High-Value Care Rounding Tool developed and published by McDaniel et al, which had proven interrater reliability,²²  and has been leveraged in subsequent studies to measure HV3C performance.^16,23  We integrated the 10 High-Value Care Rounding Tool items into our FCR audit questions and process as part of the service’s broader FCR quality auditing (Table 1). FCR auditors were trained on this tool’s use according to resources provided by the authors.²²  Because interrater reliability had been validated as part of the tool’s design, auditors’ interrater performance was not compared for this study. All responses were anonymous.

After a previous study exploring the impacts factors such as in-room versus hallway rounding, initial versus subsequent patient encounter, and family and nursing presence and participation had upon providers’ FCR HV3C performance,¹⁶  we also examined these factors.

Anonymous provider baseline and postinterventions surveys, by faculty and trainee tiers, were given electronically via a QR code, which participants voluntarily took at project orientation and wrap-up meetings. Weekly reminders occurred for 3 weeks at respective hospitalist and resident educational meetings, plus e-mail for both phases. Questions captured basic respondent characteristics, and explored faculty/trainees’ perceived importance of HV3C, comfort practicing HV3C, the frequency of faculty role-modeling HV3C, and trainees’ competence practicing HV3C that were adopted from other published HV3C education and needs assessment survey research at the time.^9,10,15,30  Postsurveys also assessed respondents’ usage, ease of use, and most helpful sections of the Rounding Tool and Quick Reference on a 5-point Likert scale, plus 1 free-text option asking for additional feedback or suggestions.

Given that our audit process did not capture HV3C conversations that possibly occurred outside assessed FCR encounters, and that not all 10 items may apply to every encounter, a 100% performance in all HV3C items was not realistic. These factors also limited setting a specified absolute target performance; therefore, for each measure, we sought to achieve a relative significant mean performance increase compared with baseline after our intervention. On the basis of variable baseline performance data, the most conservative sampling calculations requiring a minimum of 147 intervention samples were applied across all items.

For underlying FCR factors, χ² (α = .05) was used to assess baseline and intervention frequency differences. A 2-sided t test was used to compare associations between HV3C average performance and FCR variables. Analysis of variance was originally intended to compare mean survey Likert responses, but because of lower responses, Fisher’s exact tests were used instead to assess response frequency differences with 2-sided p-values reported. SAS 9.4 was used for statistical calculations. This study was approved by the University of Illinois College of Medicine, Peoria, institutional review board.

During each 6-month period, 197 baseline and 157 intervention phase encounters were audited. All FCRs occurred inside the patient’s room. The frequency of initial versus subsequent patient encounters (19.9% vs 18.6%, P = .8), family presence (97.5% vs 95.0%, P = .2), nursing presence, (49.8% vs 47.8%, P = .7), and nursing active participation (22.3% vs 21.0%, P = .8) was statistically comparable between baseline and intervention phases. There was no significant association between initial versus subsequent patient rounding encounters. Family presence was positively associated with higher performance for item #1 (see Table 1 for full item description) pertaining to providing anticipatory guidance (57.1% vs 23.1, P = .02), #9 customizing care plans (73.5% vs 46.2%, P = .03), #10 addressing family concerns (73.8% vs 30.8%, P < .001), and total items addressed (4.7 vs 3.2, P = .02). Nursing presence alone was significantly associated with higher HV3C performance on item #7 regarding avoiding low-value services (28.3% vs 18.8%, P = .03), #8 regarding the need for hospitalization (93.1% vs 81.2%, P < .001), and total items addressed (5.0 vs 4.3, P = .007). Nursing participation beyond presence alone had no significant associations with higher HV3C performance. Adjusted baseline and intervention HV3C performance comparisons were calculated with the above significantly associated underlying family and nursing presence factors (Table 1).

Introduction of our Rounding and Quick Reference tools was associated with an overall significant right shift average performance increase (Fig 3) of 5.8 total items being addressed per encounter (median: 6, mode: 6) compared with our baseline of 3.8 (median: 4, mode: 3). Eight of the 10 items showed a significant performance improvement (Table 1). In terms of underlying FCR factors’ impact upon performance, our interventions contributed to a higher, but only significant, performance increase when family was present for items #1, #9, #10, and the total. Our intervention had a significant, independent positive effect, even when nursing was absent, for item #8 addressing the need for hospitalization, and total items addressed.

For respective provider baseline and postintervention surveys, 80% and 60% of faculty (n = 20 each phase), and 23% (n = 69) and 40% (n = 48; excluding 2022–2023 graduated residents/fellows) of eligible trainees completed surveys, with Table 2 outlining results. There were no differences by group demographics between baseline versus postintervention, faculty versus trainees, nor the distribution of HV3C perception question responses. Comparable to other studies,¹⁰  25% of faculty reported no previous HV3C education before our intervention, with most reporting various departmental/hospital (75%), residency (13%), and/or fellowship (6%) lectures and/or training. All respondents consistently perceived HV3C as important. Respondent postintervention perceptions all improved, with significant improvements in faculty and trainees’ perception of faculty role-modeling HV3C discussions, trainees’ comfort practicing HV3C, and faculty’s perception of trainees’ preparedness making HV3C decisions. Among respondents, 67% of faculty and 42% of trainees (40% and 17% of eligible faculty/trainee participants overall) reported using our educational tools with equal frequencies, with no one reporting “always” using either the Quick Reference or the Rounding Tool poster. Trainees more commonly referred to the tools before, whereas faculty more commonly used the tools during FCRs. Although the Rounding Tool poster had higher ratings, both were generally perceived as easy to use and helpful among faculty and trainees alike, with trainees and faculty finding specific sections variably more useful, and most faculty noting both helped change their medical management. No free-text feedback responses were received regarding suggestions to improve either.

This study supports that implementing a directive clinical resource guide, in the forms of evidence and HV3C curricula-based interventions, can significantly improve pediatricians’ discussions, as well as perceived competence and comfort navigating HV3C conversations with families. Additionally, given our institution’s established process that patient care plan orders are updated in real time during FCRs, it is reasonable that such discussions directly translated to care, which will be an area of future study to confirm. Both the Rounding Tool poster and Quick Reference interventions contained several factors found in a systematic review of interventions known to be effective in supporting HV3C practice and education alike, including knowledge transmission of cost data, evidence-based practice approaches, addressing family values, and encouraging a supportive educational environment with reflective practices and faculty role-modeling.²⁵  The only item we measured that did not show improvement was praising the medical team for avoiding unnecessary treatment,²²  which our intervention did not specifically address.

In terms of the underlying factors’ impacts upon HV3C performance, in contrast to earlier studies finding no association,¹⁶  we found significant associations between higher HV3C performance with family, as well as nursing presence for select HV3C discussion topics. Although our intervention showed better performance rates for items pertaining to addressing anticipatory guidance, incorporating family values to customize care plans, and addressing family concerns even when family was absent, given the required family engagement inherent to these HV3C topics to be truly effective, the fact that our intervention only showed a significant performance increase when family was present is not necessarily unexpected or disappointing in the greater context of HV3C FCR goals. Additionally, given the comparatively small number of absent families (although a testament of success for FCR goals overall), we may not have been able to detect any significant differences associated with our intervention independently. Interestingly, nursing presence alone had a positive impact on HV3C performance regarding avoiding low-value services and discussion about need for ongoing hospitalization, which other studies similarly have noted the positive influence nursing can have on such waste and cost discusions.¹⁸  Although these items’ HV3C performance was higher when nursing was present (again not an unexpected expectation of FCRs overall), our intervention had a significant, independent, positive impact for these measures even when nursing was absent.

Consistent with these performance trends, survey respondents generally agreed that both our Rounding Tool Poster and Quick Reference positively changed medical management, with faculty noting increased cost consideration when making care plans, trainees noting significant improvements in faculty role-modeling HV3C discussions, and trainees feeling significantly more comfortable practicing HV3C. These findings also likely relate to faculty assessing trainees as significantly more competent in HV3C decision-making.

In terms of our tools’ utility ratings, although generally positive, interpretation is limited because of lower overall responses received, in combination with a lower number of respondents who reported using the tools. Given the nature of a teaching service and the respective iterative and cumulative discussions that take place between faculty and trainees during each cumulative FCR encounter, even if not all the care team used the tools, as long as 1 did, the tools may have had some positive impact even for nonusers. Given that faculty reported using both tools more frequently than trainees, and specifically more commonly used the more detailed Rounding Tool poster during rounds, possibly as a teaching tool, in combination with trainees noting increased faculty HV3C role-modeling, all lend support to the Tools’ group and organizational-based,³¹  educational and practice benefits. To explore the above, in addition to auditing HV3C performance, future studies should objectively audit the tools’ actual use and by whom during FCRs, plus any carry-over effects previous HV3C FCR conversations may have into subsequent rounding encounters. Although there was no consistent trend in which sections were most useful by faculty and trainees, both found the cost information generally helpful, which is consistent with other studies noting providers’ and institutions’ lacking cost data knowledge and transparency toward supporting HV3C practices.^13,14,32,33 

This study has several methodological strengths including its likely generalizability to other tertiary care, inpatient, general pediatric teaching services (after updating institution-specific cost data), its research and curricula-based Rounding Tool Poster design, use of a validated HV3C performance measurement tool,²²  and adoption of previously published HV3C provider survey perception and competence questions.^9,10,15,30 

Study results are limited to a singular institution and clinical service. Future multicenter studies across other services (ie, ICUs) are needed to support our intervention’s generalizability. Additionally, a pre/post versus randomized controlled trial was chosen to minimize disruption to existing clinical services and participant schedules. However, the pre/post design may be influenced by external, unmeasured factors that were different between the 2 time periods. The convenience sampling and FCR-focused design of our project did not capture HV3C discussions that may have occurred at other times. Because of the anonymous nature of our data collection, our study could not comment on whether our intervention and HV3C FCR discussions may have correlated to patient complexity, length of stay, satisfaction, ordered medications, diagnostic testing, and other measures. Surveys may have been subject to participant bias, given their voluntary nature, but faculty response rates were generally acceptable and representative of the group (aside from baseline gender). Trainee response rates were fewer despite reminders, especially for the baseline, which limited statistical analysis, and may have been because of combined “resident fatigue” given that several resident survey-based quality improvement projects were occurring at the same time; plus, baseline surveys occurred during the busy respiratory illness season.

Adoption of an evidence-based, best-practice HV3C clinical resource guide helps improve provider performance and consistency discussing HV3C topics with families during FCRs, significantly improving HV3C teaching, role-modeling, and competence.

We thank Drs Trina Croland, Bhavana Kandikattu, and Rawan Musaitif for their administrative support of this project, plus our pediatric hospitalists, fellows, and resident colleagues for participating in this project.