Daily morning interprofessional rounds are a well-established keystone for high-quality care (1,2), but they are challenging to conduct, as teams must synthesize and discuss copious patient information and develop care plans under limited time and changing contexts. Critical care units experience increasingly high unit census and patient acuity, both of which are associated with acute hospital mortality (3).
Different contexts call for different responses. High census frequently necessitates shorter rounding patient encounters and high patient acuity requires longer encounters (4,5). Yet, current evidence-informed recommendations for patient rounds often omit consideration of these contextual factors (1,6,7). For instance, under high census, do teams unequivocally shorten discussions of the patient introduction/history, acute clinical status, or care plan to ensure timeliness in completing rounds? Such context-specific adaptations carry risk since shorter patient encounters associate with more information omissions (2). Discussion of the patient’s care plan is particularly important given that a shared understanding of treatment amongst providers is associated with improved outcomes (2). Thus, what information is trimmed from discussions during rounds requires thoughtful consideration.
There have been calls to better understand how context influences the length of discussions during rounds and when disproportionate time allocations may be preferred (6,8). However, how teams modify their discussions about essential topics during patient encounters when challenged with high census and patient acuity is unknown. We sought, firstly, to assess how contextual factors (i.e., census, patient acuity) modulate time spent discussing essential topics (i.e., introduction/history, acute clinical status, care plan) during individual patient encounters; and, secondly, to assess the impact of census and patient acuity on overall patient encounter durations. This work lays the foundation for future studies to determine the clinical impact of contextual factors and inform the need for, and design of, tailored rounding practices to ensure teams effectively adapt without impairing patient outcomes.
MATERIALS AND METHODS
Setting
Department of Critical Care Medicine at SickKids, Toronto, ON, Canada, in the medical-surgical PICU with a census comprising 18 licensed beds. Electronic records were used for all notes, orders, and laboratory data. The PICU medical staffing model consisted of one faculty attending (of six physicians who rotate weekly) with three to four fellows and two to three residents. A multidisciplinary team (i.e., attending physician, medical trainees, registered nurses, registered respiratory therapists, pharmacist, dietician) conducted morning rounds (with family present), which were run by the outgoing night fellow supervised by the attending physician, starting at 7:30 am, and were limited to 90 minutes. Although no standardized form of rounding took place, teams followed a traditional structure in which the night fellow presented three essential topics (with multidisciplinary input as required): patient introduction/history, acute clinical status, and care plan. Detailed functional clinical examinations were conducted before or after morning rounds by each clinician type. We observed patient rounds from February 2018 to April 2018, with an a priori target of observing a minimum of 160 patient encounters (2,7). The schedule ensured that each attending physician was observed at least twice. The Research Ethics Board (REB No. 1000059173) approved this study and waived the need for staff or patient consent.
Data Collection
Prior to formal data collection, two researchers completed 14 days of pilot observations, including 4 days with a PICU clinician to identify any clinically relevant information omissions or misinterpretations. Two researchers captured rounding durations until inter-rater reliability (kappa > 0.95) was achieved, after which one researcher collected data.
The event logging application, DELTA (Farzan Sasangohar, College Station, TX), was used to collect the duration of essential topics discussed and overall patient durations (excluding nonessential time, such as interruptions). Census (i.e., patients in the unit when rounds started under the responsibility of the attending physician) and patient acuity (represented by the Pediatric Logistic Organ Dysfunction score measured once per day) were collected from departmental metrics. Patient ICU length of stay (LOS) and rounding order were also collected and treated as potential confounding variables, as they have been shown to impact time allocations (5,9).
Data Analysis
Census and patient acuity were categorized as “low” or “high” and LOS as “short” or “long” compared with their respective means during the observation period. Rounding bed order was categorized as the “first four” or “last four” patients. After excluding outliers, we conducted a mixed analysis of variance with census (low vs high) and patient acuity (low vs high) as between-subject factors and essential topic (introduction/history vs acute clinical status vs care plan) as a within-subject factor to investigate how contextual factors modulate topic durations. Independent-sample t tests were conducted to compare mean overall encounter durations for individual patients as a function of contextual factor. Analyses were conducted using IBM SPSS Version 26 (IBM Corp., Armonk, NY), α less than 0.05 with Bonferroni correction for pairwise comparisons.
RESULTS
Sixteen morning rounds were observed comprising 248 patient encounters. However, 165 patient encounters remained after excluding those with missing data (e.g., patient LOS). Mean census was 16.1 patients, mean patient acuity was 9.0, and mean LOS was 3.6 days. Patient LOS and rounding bed order did not impact patient encounter durations (p > 0.05).
Essential Topic Durations × Contextual Factors
Essential topic durations varied significantly as a function of census and patient acuity (F[2,322] = 3.4; p < 0.04). Regardless of census or patient acuity, duration of the patient introduction/history did not significantly change (Table 1). When census was high versus low, acute clinical status discussions decreased for both low (–49.5% change; p = 0.001) and high (–42.6% change; p = 0.001) acuity patients. Durations of care plan discussions reduced when census was high versus low, for low acuity patients (p < 0.001), with this result showing the greatest percent change (–54.7%). No duration changes were observed between high versus low census, for care plan discussions for high acuity patients.
TABLE 1. -
Mean Durations of Essential Topics As a Function of Census and
Patient Acuity
| Essential Topics |
Low Acuity |
% Change |
p
|
High Acuity |
% Change |
p
|
| Low Census |
High Census |
Low Census |
High Census |
| Patient information/history |
00:55 (00:45) |
00:40 (00:26) |
–27.3% |
NS |
00:49 (00:43) |
00:36 (00:28) |
–26.5% |
NS |
| Acute clinical status |
01:39 (01:07) |
00:50 (00:34) |
–49.5% |
0.001 |
02:02 (01:26) |
01:10 (00:50) |
–42.6% |
0.001 |
| Care plan |
02:52 (01:48) |
01:19 (00:45) |
–54.7% |
< 0.001 |
02:58 (02:09) |
02:30 (01:50) |
–15.7% |
NS |
NS = not significant.
Means reflect duration in minutes and seconds. sds are in parentheses. % change reflects the difference in patient encounter durations between low- vs high-unit census and patient acuity.
Overall Encounter Durations × Contextual Factors
Mean overall encounter durations were shorter when census was high (mean = 03:41, sd = 01:59) versus low (mean = 05:36, sd = 02:23, t[160] = 5.59, p < 0.001) but longer when patient acuity was high (mean = 05:00, sd = 02:54) versus low (mean = 04:09, sd = 01:55, t[162] = –2.15, p < 0.04).
DISCUSSION
Our results demonstrate that contextual factors (i.e., census, patient acuity) disproportionately modulate the durations of discussions of essential topics during PICU morning rounds in an unit with a constrained duration. No changes were observed concerning the duration of introduction/history discussions across high or low census or patient acuity. However, when census was high, teams shortened their discussions of: 1) acute clinical status, regardless of patient acuity and 2) care plans, for low but not high acuity patients. Additionally, our findings affirm that high census and low patient acuity associate with shorter patient encounter durations (4,5).
This study fills a gap by bringing into focus how contextual factors alter discussions of essential topics during rounds that may not be recognized or acknowledged by clinicians, nor taken into consideration by healthcare systems. Our findings are particularly relevant given the recommendations aimed at modifying the rounding structure and standardization of patient encounters (e.g., via checklists) (6,10), with no explicit consideration for contextual factors, such as census and acuity. While standardized structures contribute to effective rounding processes (11,12), different contexts necessitate different responses. More research is needed to understand how best to maintain the benefits of standardization (e.g., not skip fundamentals) while safely adapting to context. The propensity is to adapt by extending the duration of rounds during high census, which can have downside effects (e.g., delayed tasks/orders). Alternatively, teams could recognize the situational context they are in and adapt by tailoring rounding time to ensure comprehensive discussions about critical patient information and the treatment plan (10). Solutions may include sharing or deferring information in other ways (e.g., share patient data electronically, defer discussions not requiring multidisciplinary input to after rounds).
This research is an important first step to understanding the impact that time discussing essential topics has on the quality of patient care. For instance, given that high acuity patients are associated with acute hospital mortality (3), is it maladaptive to reduce discussion of acute clinical status updates for these patients when census is high? Does a shared understanding of the care plan weaken amongst providers if discussion durations decrease for low acuity patients, and if so, does this negatively impact patient outcomes as the literature suggests (2)? Future research should investigate how these adaptations and other contextual factors (e.g., interruptions, family engagement) impact the quality of bedside rounds and patient outcomes. These findings will help inform the need for, and development of, interventions that are tailored to efficient rounding practices in varying contexts.
Study limitations include: lack of assessment of rounding adaptations on patient outcomes; the Hawthorne effect; single site with a self-imposed rounding target (90 min) and shorter patient encounters compared with those reported in literature (2,5–7); and variability across attendings leading rounds may have impacted discussion durations.
CONCLUSIONS
Contextual factors (e.g., census, patient acuity) disproportionately impact time spent discussing essential patient topics during rounds. Yet, characterization of how contextual factors impact rounding practices is not reflected in current rounding guidelines. Consequently, organizations may miss important considerations leading to a decrease in the quality of rounding discussions about critical patient information and their care plan. Our findings provide needed detail regarding how teams adapt their discussions of essential topics and establishes a foundation for consideration of varying contextual factors in the design of rounding guidelines.
ACKNOWLEDGMENTS
We thank the patients, their families, and the participating healthcare professionals for participating in our work. Finally, we thank T. Adam, RN, L. Kolodzey, J. Tomasi, K. Fiaes, and the project committee members: M. Chen, RN, J. Hubbert, RN, R. Kirsch, MD, W. Seto, Pharm, C. Sperling, RRT, J. Stillman, RN, and P. Sutton, RN.
REFERENCES
1. Lane D, Ferri M, Lemaire J, et al.: A systematic review of evidence-informed practices for patient care rounds in the ICU. Crit Care Med 2013; 41:2015–2029
2. Artis KA, Bordley J, Mohan V, et al.: Data omission by physician trainees on ICU rounds. Crit Care Med 2019; 47:403–409
3. Wilcox ME, Patel A, Rowan KM: Higher ICU capacity strain is associated with increased acute mortality in closed ICU. Crit Care Med 2020; 48:709–716
4. Beck K, Albert E, Johnsen A, et al.: Pediatric intensive care unit rounding improvements. BMJ Qual Improv Rep 2016; 5:u208902.w3600
5. Sung N, Weiss C, Rho J, et al.: Rounding order as a predictor of individual patient rounding times. Am J Respir Crit Care Med 2012; 185:A6562
6. Abraham J, Kannampallil TG, Patel VL, et al.: Impact of structured rounding tools on time allocation during multidisciplinary rounds: An observational study. JMIR Hum Factors 2016; 3:e29
7. Vats A, Goin KH, Villarreal MC, et al.: The impact of a lean rounding process in a pediatric intensive care unit. Crit Care Med 2012; 40:608–617
8. Christianson MK, Ilan R: End-of-round time compression in physician handoff sessions-reply. JAMA Intern Med 2013; 173:1033–1034
9. Jones L, Kannampallil T, Franklin A: End-of-round time compression in physician handoff sessions-letter. JAMA Intern Med 2013; 173:1033–1034
10. Stollings JL, Devlin JW, Lin JC, et al.: Best practices for conducting interprofessional team rounds to facilitate performance of the ICU liberation (ABCDEF) bundle. Crit Care Med 2020; 48:562–570
11. Knoll M, Lendner I: Nurses’ perspective on interprofessional
communication on an intensive care unit. Pflege 2008; 21:339–351
12. Collins SA, Bakken S, Vawdrey DK, et al.: Model development for EHR interdisciplinary information exchange of ICU common goals. Int J Med Inform 2011; 80:e141–e149
