Evaluating Physician Stress During Patient Handoffs: Analysis of Physician Heart Rate Variability,Patient Handoff Characteristics,and ED Metrics

Abstract

Patient handoffs represent a critical point in patient care, specifically in hospital Emergency Departments (EDs) where handoffs occur at every shift change. Previous studies have examined ED patient handoffs primarily by evaluating handoff communication methods and scheduling models that improve and reduce the prevalence of handoffs. In this study, we provided a quantitative analysis of physician stress by measuring the HRV of outgoing and oncoming physicians during patient handoffs. We recorded descriptive data of each observed patient handoff using the status of patient disposition as a measure of physician workload, and retrospectively gathered patient outcome and ED metric data for multivariate analysis. Results demonstrate that attending physicians experience increased HRV during handoff periods over non-handoff periods, with a significant increase observed in outgoing physicians (α = 0.05). This study creates a framework for future work that considers health care worker stress and evaluates ED handoffs via the stress and workload transferred between emergency physicians.

Keywords

Health care Stress Work physiology Emergency medicine

Introduction

Patient handoffs (also known as handovers or sign-outs) are well known to be critical points in patient care with regard to patient safety. (Arora and Johnson, 2006; Singh et al., 2007). Patients being handed off from one physician to another are a regular occurrence in hospital settings, both for hospital inpatients and patients in the emergency department (ED); however, handoff practices differ significantly between different departments. Generally, patient handoffs occur either within a department during shift change between physicians (inter-shift handoffs) or when inpatients are transferred from one department/unit to another (inpatient handoffs). Handoffs have been extensively studied in existing literature. Many of these studies that have examined patient handoffs in healthcare have primarily focused on communication efficacy during handoffs. This includes observational studies that evaluate the methods of communicating patient information during handoffs and interventional studies that propose mnemonic tools or alterations to standardize handoffs (Campbell et al., 2019; Heilman et al., 2016; Peterson et al., 2014; Smith et al., 2018; Watkins et al., 2014).

In the ED, inter-shift patient handoffs between physicians occur every shift change as outgoing physicians sign-out their current patients to newly arrived oncoming physicians. Other types of handoffs also occur within the ED when patients are initially transferred from emergency medical services (EMS) to the ED and when patients that are admitted to the hospital are transferred to the department that will ultimately provide their inpatient care. With respect to ED patient handoffs that occur between physicians, studies have similarly focused on communication evaluation and standardization during handoffs. For example, Heilman et al. (2016) explore the limitations of the I-PASS Handoff Program during inter-shift patient handoffs and recommend alterations to the standard I-PASS mnemonic (Illness severity, Patient summary, Action list, Synthesis by receiver, Summary by receiver) tailored to the ED specifically. Similarly, Cheung et al. (2010) evaluated the quality of patient handoff procedures and provide recommendations for future handoff standards that will improve patient safety. While the process through which patient handoffs occur has been extensively researched, the effects of handoffs on both patient outcomes and physician anxiety, stress, and burnout, is still largely unknown.

Studies on patient handoffs have generally used qualitative approaches, such as observations and interviews, to evaluate and alter patient handoff procedures. However, quantitative approaches have been utilized in the literature to provide alternative methods to alter handoff procedures. For instance, scheduling models have been developed to reduce the prevalence of inter-shift patient handoffs in the ED (Yoshida et al., 2019) and surveys have been widely used to monitor perceived physician stress and burnout in the ED (Durand et al., 2019; Howlett et al., 2015). Quantitative studies are advantageous in that they involve measurable processes and outcomes, and can be used for predictions, but do not represent much contextual detail about the environment, tasks, decisions, interactions and cognitive processes involved. Qualitative studies provide such context through descriptive narrative data but lack the advantages of quantifiable measures. Ideally qualitative and quantitative approaches should be combined in complementary ways such that the measurable and predictive affordances of quantitative approaches can be placed within the descriptive context of the system to support more informed decision making.

As the healthcare industry is facing increasing staff shortages (Allan and Aldebron, 2008) in part due to increased physician stress, workload, and burnout (Humphries et al., 2014), it is key to not only improve patient handoffs but understand the effects of handoff characteristics, such as the total number of patients handed off, the relative acuity levels of those patients, and the crowding levels of the ED during the handoff period, on patient outcomes and staff wellbeing. Previous studies have shown the effects of physician stress and workload on patient safety. For instance, Weigl et al. (2017) identified a direct relationship between interruptions and ED physician and nurse perceived stress. Similarly, studies have examined ED physician stress using biological markers such as HRV and cortisol levels and have identified how the ED environment increases the stress of resident and attending physicians (Arnetz et al., 2017; Girishan Prabhu et al., 2020; Janicki et al., 2021). However, in the context of ED patient handoffs, we have not found previous work that quantitatively defines the relationship between qualitative handoff characteristics and patient outcomes and physician stress. This study aimed to assess physician stress and workload during inter-shift patient handoffs in the ED, using physiological measures, and identify relationships between these measures, qualitative handoff characteristics, and ED metric and patient outcome data.

Methods

This study was conducted at an 814-bed tertiary referral hospital and academic center in the southeastern United States. The ED at this hospital has over 100,000 annual patient visits (as of 2020) and is comprised of 5 “pods” which each contains its own patient beds and clinical teams and is specialized for certain types of care. Because the hospital is an academic center, clinical teams consist of attending physicians, resident physicians, nurses, and technicians. The ED pod where this study was conducted contains 18 patient bays and specializes in treatment for chest-pain and trauma overflow.

Data relating to the characteristics of the handoffs and physician stress were collected observationally. Handoff characteristics included the relative patient load transferred between the physicians during handoffs, and the nature of patients in terms of the state of their disposition (defined in Table 1). Physician stress via heart rate variability (HRV) was recorded for both outgoing and oncoming physicians in 2-hour periods using the Empatica E4 wristwatch, a device which records heart rate data that can be converted to HRV. HRV has been extensively reviewed in existing literature as a surrogate measure for cognitive stress (Kim et al., 2018; McDuff et al., 2014; Shaffer and Ginsberg, 2017). ED metric data and patient outcome data were gathered retrospectively following the completion of all live observations. ED metric data refers to overcrowding scores, number of patient visits, and descriptive statistics for patient outcomes during a given time period. Patient outcome data refers to attributes such as patient length of stay (LOS), patient mortality, and patient acuity. These data were collected by author SF during November and December of 2022.

Table 1.

Index system for the state of patient dispositions during ED patient handoffs.

Index Level	State of Patient Disposition	Explanation
5	Patient is boarding. Outgoing physician did not record primary history.	The patient has been admitted to an inpatient unit and is waiting to be physically relocated to said inpatient unit.
4	Patient is admitted to inpatient care or discharged	The patient disposition will be assigned by the outgoing physician.
3	Patient disposition expected	The patient is expected to be admitted or discharged in the near future but will not have their disposition assigned by the outgoing physician.
2	Patient disposition dependent on in-progress tests or interventions	The patient has received testing, imaging, and/or treatment that could result in either discharge or admittance based on the result.
1	Patient disposition unknown	The patient has recently arrived at the ED and has either not been seen by the outgoing physician or the outgoing physician does not yet have a disposition for the patient.

Observations consisted of 4-hour periods centered on an inter-shift patient handoff in the designated pod. Three shift changes occurred each day of the week at 7am, 3pm, and 11pm with the attending physician ending their shift at the given time and being replaced by a new attending physician at that same time. Oncoming attending physicians had their heart rate recorded for 2 hours at the beginning of their shift, at the start of the patient handoff. Similarly, outgoing attending physicians had their heart rate recorded beginning with 2 scheduled hours remaining in their shift and ending immediately after the handoff. Each handoff was comprised of the transfer of patient care and patient care related responsibilities from an outgoing physician who was ending their shift to an oncoming physician who was beginning their shift. For outgoing physicians, HRV recording began 2 hours prior to the end of their shift and ended after the handoff of patient care to the oncoming physician was complete. For oncoming physicians, HRV recording began at the start of their shift, prior to the handoff of patient care from the outgoing physician to the oncoming physician, and recording ended 2 hours into their shift.

Descriptive observation data during patient handoffs were recorded using a prototype handoff evaluation tool. This tool was designed in collaboration with emergency medicine physicians and researchers to act as a surrogate measure for workload as represented by the patient load transferred from one physician to another during handoffs. The main goal of the tool was to measure and quantify the relative workload transferred between physicians during handoffs using a method that did not require explicit clinical knowledge surrounding patient diagnoses, illnesses, treatment methods, etc. Instead, this tool relies on the observer’s ability to identify the state of a given patient’s disposition as a surrogate measure for workload. During handoffs, the observer would listen to each patient “handed off” by physicians and index the state of each patient’s disposition as discussed by the physicians on a scale of 1 to 5, paralleling the 5-point Emergency Severity Index (ESI). This disposition index system is further explained in Table 1.

The patient handoff evaluation tool also recorded: the total duration, the duration and total number of interruptions, and start and end times of each patient handoff observed. ED metric data is currently being retroactively gathered from the months of November and December 2022 to include other covariates and patient outcome measures in the data analysis. These data are provided through the participant hospital’s data support core. The covariate variables gathered include the national emergency department over-crowding score (NEDOCS) and the overall patient acuity in the selected ED pod during observations while the patient outcome measures include length of stay (LOS) and mortality. These data are currently being made available to the research team and will be included in further analyses.

Participants

Participants were selected exclusively amongst emergency medicine attending physicians that regularly worked in the designated pod in the hospital ED. Attending physicians were chosen based on their extensive experience in emergency medicine and familiarity with the handoff process. Attending physicians were recruited for this observational study using selective sampling via email correspondence. Participants were each awarded $50 online gift cards for their participation in this study. This study was approved by the institutional review boards of the collaborating institutions of each of the authors.

Thirty-one participants were recruited for this study during the months of October and November 2022. Each participant is an attending physician in emergency medicine at the hospital’s parent healthcare system and rotates working in several EDs across the Prisma Health system. The mean age for participant physicians across all observations was 42.7 years old (σ = 9.5) during observations while the average years of practice for each physician (defined as the time since completion of residency) was 15 years of practice (σ = 6.2). Of the 31 participants who participated in the observational study, 22 are males and 9 are females.

Data

Physician HRV data were collected across 36, 2-hour interval observations. HRV was calculated based on heart rate data captured through the Empatica E4 wristwatch. The inter-beat interval (IBI) is the time period between consecutive beats of the heart (Empatica, 2023) and HRV is defined as the natural variation between IBIs. HRV was calculated as the root mean square of successive differences (RMSSD) of IBI measures recorded with the Empatica E4 across varying time intervals (Farnsworth et al., 2022).

Results

Observations

This observational study was conducted in 1 of the 5 ED pods at the designated hospital. This pod was chosen for observation due to 2 factors: (i) shifts consisted of a single attending physician and a single resident physician, (ii) there was no overlap between outgoing and oncoming physician shifts to ensure one-to-one handoffs. The study was comprised of 18 observed handoffs recorded during the months of November and December of 2022. Three participants each participated in 2 observations at different points during the study while one participated in 3 separate observations. In total, 36 HRV periods were recorded consisting of 18 outgoing observations and 18 oncoming observations with these observations being comprised entirely by the 31 participants previously described.

Initial Analyses

Analyses were conducted to first test for differences between physician HRV during the handoff period (defined as the hour before or after handoff, for outgoing and oncoming physicians respectively) and physician HRV recorded during the non-handoff period (the 2^nd-to-last hour of an outgoing physician’s shift or the 2^nd hour of an oncoming physician’s shift). Of the 36 observations, 33 (15 outgoing and 18 oncoming) were included in the analysis as 3 were deemed unusable due to observation errors and technical malfunctions. Using a within-subjects design, a one-sample t-test was conducted to test for a significant change in HRV between the handoff period and non-handoff period for both outgoing and oncoming physicians. Similar t-tests were also performed for the different shift-change times observed (7am, 3pm, and 11pm). The results of each of these tests are displayed in Table 2.

Table 2.

One-Sample t-tests of HRV difference scores for physicians during the handoff period vs. during the non-handoff period.

Description	N	Mean (s)	Standard Deviation	Degrees of Freedom	t-statistic	p-value (two-tailed)	Effect Size (Cohen’s)
Outgoing Physicians	15	0.00794	0.12098	14	2.542	0.023**	0.656
7am	5	0.01188	0.12741	4	2.085	0.105	.932
3pm	5	0.00560	0.00740	4	1.963	0.166	0.757
11pm	5	0.00634	0.01630	4	0.87	0.434	0.389
Oncoming Physicians	18	0.02062	0.06259	17	1.398	0.180	0.329
7am	6	0.02670	0.07588	5	0.862	0.428	0.352
3pm	6	0.04446	0.07375	5	1.477	0.200	0.603
11pm	6	−0.00930	0.01792	5	−1.271	0.260	−0.519

denotes significance (α = 0.05).

This analysis revealed that only outgoing physicians when combined across all shift-change times had a significant increase (α = 0.05) in HRV (lower stress) from the non-handoff period to the handoff period. For oncoming physicians, although the difference wasn’t significant, there was a notable trend towards higher HRV during the handoff period when compared to the handoff period, the second hour of their shift. In general, increases were also demonstrated from the non-handoff period to the handoff period for each group, except for oncoming physicians that began their shift at 11pm.These physicians had a trend towards lower HRV (an increase in cognitive stress) from the handoff period (the first hour of their shift) to the non-handoff period (the second hour of their shift).

Further tests were conducted to compare the mean HRV records between outgoing and oncoming physicians’ during both the non-handoff period and the handoff period. These results are shown in Table 3.

Table 3.

Independent Samples t-test of HRV differences between oncoming and outgoing physicians.

Description	Mean Outgoing (s)	Standard Deviation Outgoing	Mean Oncoming (s)	Standard Deviation Oncoming	Degrees of Freedom	t-statistic	p-value (two-tailed)	Effect Size (Cohen’s d)
Handoff Period	0.10994	0.02997	0.10431	0.04945	31	0.385	0.703	.135
non-Handoff Period	0.10429	0.03703	0.09392	0.02820	31	31	0.368	0.319

The result of the independent samples t-test (shown in Table 3) reveals that there were no significant differences detected in this data between oncoming and outgoing physicians both in the handoff period and the non-handoff period. However, both periods demonstrate an increase in HRV for outgoing physicians during the same period as the oncoming physicians.

Discussion

Our initial analysis has shown that outgoing physicians have significantly higher HRV in the handoff period (the last hour of their shift) compared to the non-handoff period (the penultimate hour of their shift). Although not significant, there was a statistical trend toward higher HRV, i.e. lower cognitive stress, during the handoff period (first hour) for oncoming physicians. This is a notable finding as it indicates lower stress for the oncoming and outgoing physicians during the hour in which their shifts overlap during the patient handoff. While this may partly be attributable to the outgoing physician “winding down” at the end of their shift and the oncoming physician “ramping up” at the start of their shift, it does call for additional investigation into how varying periods of overlap during patient handoffs affect physician stress in the ED. In conjunction with previous work that has demonstrated similar decreases in HRV amongst ED resident physicians when compared between working and non-working hours (Janicki et al., 2021), these findings pose avenues for future work that closely analyzes changes in physician stress when beginning and ending ED shifts.

Future analyses will aim to contextualize pre-existing claims concerning ED physician stress and patient handoffs in the literature with quantitative findings. Time series analysis will be used to better understand how the ED physician stress changes during their shifts, as current findings support the claim that physicians experience increased stress towards the center of their shift. Analysis between each of the three shift times will provide greater understanding of the effects of circadian rhythms on physician stress, which has previously been documented in the literature with respect to overnight shifts (Adams et al., 1998; Çalişkan et al., 2016; Machi et al., 2012). Further analyses will be used to holistically understand the relationships between different ED metrics, patient handoff characteristics, physician characteristics, and stress (as HRV) through the inclusion of covariates and confounding variables in keeping with existing quantitative ED studies (Wrenn et al., 2010). Additionally, interviews with ED physicians involved in the observed handoffs will be conducted to comprehensively understand their perspectives on stress, workload, and coping mechanisms to gain additional context relating to the measures collected.

Based on the findings presented in this study, future work should examine changes in ED physician stress that occur in response to various stimuli. This includes changes in HRV that occur throughout the course of an entire shift, the relationship between stress and consecutive days worked, and stress perceived by physicians when exposed to different handoff methods. Further work is planned that utilizes a mixed methods approach of incorporating interviews with this stress data to increase the context and comprehensiveness of this work through the combination of quantitative and qualitative methods.

Planned Analyses

Further analyses are planned to incorporate both the handoff evaluation tool and the ED metric data gathered concurrently with observations. Time series analysis will be utilized regarding the HRV data to identify trends that occur for both outgoing and oncoming physicians throughout the duration of their shifts. Analysis of variance (ANOVA) will be used to compare HRV measures between the time of shift-change (7am v 3pm v 11pm). Linear regression modeling will be used to incorporate sign-out evaluation measures, ED metrics, and physician age and experience data into a predictive model that predicts the HRV of a physician during sign-out. Lastly, analysis of covariance (ANCOVA) will be used to control for the covariates and confounding variables recorded during observations when comparing between oncoming and outgoing physicians and shift-times with respect to physician HRV. Each of these analyses are expected to be conducted during the Spring of 2023.

Limitations

Due to the exploratory and observational nature of this study, there are inherent limitations in the presented findings and claims. The HRV data presents limitations due to the Empatica E4 wristwatch technology. Specifically, the gathered IBI data had been pre-cleaned by the Empatica, removing any irregularities in the data, namely ectopic and abnormal beats. Likewise, some participant physicians adjusted the Empatica wristwatch, sometimes resulting in break of contact between the sensor and their wrist. In part because of these factors, there is significant missing data that varies between each participant. Similarly, given the data collected from this exploratory work, future studies may find more poignant results due to larger participant pools. Lastly, while many covariates and confounding variables will be included in the final analyses, the unpredictable nature of the ED leads to significant variability in observational studies. Nevertheless, the findings of this work as a pilot study are intended to provide a basis for future observations, experiments, and interventions to improve the efficacy of patient handoffs and reduce physician stress in the ED.

Footnotes

Acknowledgements

The study is supported by the Prisma health Transformative Seed Grant awarded to authors SH, RP, KT and SF.

References

Adams

S. L.

Roxe

D. M.

Weiss

Zhang

Rosenthal

J. E.

(1998). Ambulatory blood pressure and Holter monitoring of emergency physicians before, during, and after a night shift. Academic Emergency Medicine, 5(9), 871-877.

Arnetz

B. B.

Lewalski

Arnetz

Breejen

Przyklenk

. (2017). Examining self-reported and biological stress and near misses among Emergency Medicine residents: a single-centre cross-sectional assessment in the USA. BMJ open, 7(8), e016479.

Çalişkan Tür

Toker

İ.

Şaşmaz

C. T.

Hacar

Türe

. (2016). Occupational stress experienced by residents and faculty physicians on night shifts. Scandinavian journal of trauma, resuscitation and emergency medicine, 24, 1-8.

Campbell

Dontje

. (2019). Implementing bedside handoff in the emergency department: a practice improvement project. Journal of Emergency Nursing, 45(2), 149-154.

Cheung

D. S.

Kelly

J. J.

Beach

Berkeley

R. P.

Bitterman

R. A.

Broida

R. I.

… White

M. L.

(2010). Improving handoffs in the emergency department. Annals of emergency medicine, 55(2), 171-180.

Durand

A. C.

Bompard

Sportiello

Michelet

Gentile

(2019). Stress and burnout among professionals working in the emergency department in a French university hospital: Prevalence and associated factors. Work, 63(1), 57-67.

Empatica (2023). E4 Data – IBI Expected Signal. Empatica Support. https://support.empatica.com/hc/en-us/articles/360030058011-E4-data-IBI-expected-signal#:~:text=The%20inter%2Dbeat%20interval%20(abbreviated,estimate%20the%20instantaneous%20heart%20rate.

Estryn-Behar

Doppia

M. A.

Guetarni

Fry

Machet

Pelloux

… Prudhomme

(2011). Emergency physicians accumulate more stress factors than other physicians–results from the French SESMAT study. Emergency Medicine Journal, 28(5), 397-410.

Farnsworth

Fischer

Soehn Ambrose. Krosschell

(2022). Heart Rate Variability – How to Analyze ECG Data. Imotions. https://imotions.com/blog/learning/best-practice/heart-rate-variability/.

10.

Girishan Prabhu

Taaffe

Pirrallo

Shvorin

. (2020). Stress and burnout among attending and resident physicians in the ED: a comparative study. IISE Transactions on Healthcare Systems Engineering, 11(1), 1-10.

11.

Heilman

J. A.

Flanigan

Nelson

Johnson

Yarris

L. M.

(2016). Adapting the I-PASS handoff program for emergency department inter-shift handoffs. Western Journal of Emergency Medicine, 17(6), 756.

12.

Howlett

Doody

Murray

LeBlanc-Duchin

Fraser

Atkinson

P. R.

(2015). Burnout in emergency department healthcare professionals is associated with coping style: a cross-sectional survey. Emergency Medicine Journal, 32(9), 722-727.

13.

Kim

H. G.

Cheon

E. J.

Bai

D. S.

Lee

Y. H.

Koo

B. H.

(2018). Stress and heart rate variability: a meta-analysis and review of the literature. Psychiatry investigation, 15(3), 235.

14.

Janicki

A. J.

Frisch

S. O.

Patterson

P. D.

Brown

Frisch

(2021). Emergency medicine residents experience acute stress while working in the emergency department. Western Journal of Emergency Medicine, 22(1), 94.

15.

Machi

M. S.

Staum

Callaway

C. W.

Moore

Jeong

Suyama

… Hostler

(2012). The relationship between shift work, sleep, and cognition in career emergency physicians. Academic Emergency Medicine, 19(1), 85-91.

16.

McDuff

Gontarek

Picard

(2014, August). Remote measurement of cognitive stress via heart rate variability. In 2014 36th annual international conference of the IEEE engineering in medicine and biology society (pp. 2957-2960). IEEE.

17.

Peterson

S. M.

Gurses

A. P.

Regan

(2014). Resident to resident handoffs in the emergency department: an observational study. The Journal of emergency medicine, 47(5), 573-579.

18.

Shaffer

Ginsberg

J. P.

(2017). An overview of heart rate variability metrics and norms. Frontiers in public health, 258.

19.

Smith

C. J.

Buzalko

R. J.

Anderson

Michalski

Warchol

Ducey

Branecki

C. E.

(2018). Evaluation of a novel handoff communication strategy for patients admitted from the emergency department. Western Journal of Emergency Medicine, 19(2), 372.

20.

Watkins

L. M.

Patrician

P. A.

(2014). Handoff communication from the emergency department to primary care. Advanced emergency nursing journal, 36(1), 44-51.

21.

Weigl

Beck

Wehler

Schneider

(2017). Workflow interruptions and stress atwork: a mixed-methods study among physicians and nurses of a multidisciplinary emergency department. BMJ open, 7(12), e019074.

22.

Wrenn

Lorenzen

Jones

Zhou

Aronsky

(2010). Factors affecting stress in emergency medicine residents while working in the ED. The American journal of emergency medicine, 28(8), 897-902.

23.

Yoshida

Rutman

L. E.

Chen

Shaffer

M. L.

Migita

R. T.

Enriquez

B. K.

… Mazor

S. S.

(2019). Waterfalls and handoffs: a novel physician staffing model to decrease handoffs in a pediatric emergency department. Annals of emergency medicine, 73(3), 248-254.