Time of Day Affects the Probability of Antibiotic Initiation for Hospitalized Patients with Sepsis

By Betty Tran, MD, MSc

Associate Professor of Medicine, Division of Pulmonary and Critical Care Medicine, Northwestern University Feinberg School of Medicine, Chicago

SYNOPSIS: In this retrospective cohort study of patients with hospital-onset sepsis, the probability of antibiotic initiation was lowest at shift changes and gradually declined overnight compared to during the day shift.

SOURCE: Ginestra JC, Kohn R, Hubbard RA, et al. Association of time of day with delays in antimicrobial initiation among ward patients with hospital-onset sepsis. Ann Am Thorac Soc 2023;20:1299-1308.

Prior studies have reported an association between antibiotic delays and mortality in community-acquired sepsis.^1,2 These delays may be related to the time of day, an effect that has been associated with suboptimal care delivery in the outpatient setting.^3,4 The authors of this study hypothesized that the probability of antibiotic initiation for patients with hospital-acquired sepsis would decrease over the time course of provider shifts similar to the decline seen in the outpatient setting.

The primary cohort consisted of 1,672 ward patients with hospital-onset sepsis from five acute care hospitals in the University of Pennsylvania system. Sepsis onset was defined as the time that the earliest of the following three criteria, based on Centers for Disease Control and Prevention Adult Sepsis Event (ASE) criteria, was met: 1) blood culture collection, 2) initiation of broad-spectrum antimicrobials within 48 hours before/after blood culture collection and continued for ≥ 96 hours, and 3) new organ dysfunction measured by a simplified version of the Sequential Organ Failure Assessment (SOFA) ≥ 1. Day shift was 7 a.m. to 7 p.m.; night shift was 7 p.m. to 7 a.m. The primary analysis tested the association between hour of day (primary exposure) and receipt of a new broad-spectrum intravenous antimicrobial during that hour as a binary variable (primary dependent variable). Secondary analyses tested for associations between hour of sepsis onset and antimicrobial initiation as a continuous variable, binary variables as starting within one and three hours of sepsis onset, and in-hospital mortality. Repeat analyses and post-hoc analyses incorporated SOFA score at sepsis onset to assess whether the associations observed persisted in patients with high severity of illness.

The median time from admission to sepsis onset was 7.3 days (interquartile range [IQR], 3.8 to 12.7 days), with equal distribution throughout the 24-hour day. The unadjusted median time to antimicrobial initiation after sepsis discovery was 4.1 hours (IQR, 0.4-23.3 hours), and, overall, 320 (19%) patients died while in the hospital. The probability of antimicrobial initiation varied almost fivefold throughout the day, with nadirs at 7 a.m. and 7 p.m., and declined progressively over the course of the night shift (13.4% at 9 p.m. to 3.2% at 6 a.m.). In secondary analyses, the standardized predicted median time to antimicrobial initiation was 3.2 hours (IQR, 2.5-3.8 hours) for sepsis onset during day shift hours compared to 12.9 hours (IQR, 10.9-14.9 hours) for patients with sepsis onset during night shift hours, a fourfold difference. The standardized predicted probability of antimicrobial initiation within one and three hours after sepsis onset overall declined after 9 p.m. until midnight, rebounded around 1 a.m., but then declined during the later hours of the night shift with a nadir at 6 a.m. There was no correlation between increased mortality and times of day with longer times to antimicrobial initiation. These findings were consistent among patients with the highest SOFA scores at sepsis onset.

COMMENTARY

This study adds to the growing body of literature reporting that time of day is a significant factor in care delivery and medical decision-making. The three main periods of concern noted in the study include: 1) at shift change, 2) night shift overall vs. day shift, and 3) an hourly cumulative effect seen over the course of the night shift.

There are likely pragmatic reasons for these observations. Shift changes for clinicians require time away from direct patient care and may delay timely communication between nurses and clinicians, sepsis recognition, and antimicrobial initiation when there are competing responsibilities and tasks. Night shifts overall have lower staff-to-patient ratios, which can result in less time for chart review, bedside evaluation, and patient interactions. Finally, cumulative physical and cognitive, particularly decision-making, fatigue over the course of a night shift may impede sepsis recognition and delay antimicrobial initiation as a result. Additionally, in certain circumstances, major decisions regarding ordering new antimicrobials may be delayed until morning rounds by the day team.

These detailed findings provide the groundwork for future validation studies in other settings and also highlight a need to improve support and care processes overnight, as these hours are quite vulnerable to several factors that can hinder optimal care delivery. These factors may be individual to specific hospitals, wards, and/or other settings, but improvements in staffing levels, formal and informal rounds and communication overnight, and strategies to alleviate physical and decision-making fatigue are likely to be appreciated universally.

REFERENCES

1. Seymour CW, Gesten F, Prescott HC, et al. Time to treatment and mortality during mandated emergency care for sepsis. N Engl J Med 2017;376:2235-2244.
2. Liu VX, Fielding-Singh V, Greene JD, et al. The timing of early antibiotics and hospital mortality in sepsis. Am J Respir Crit Care Med 2017;196:856-863.
3. Hsiang EY, Mehta SJ, Small DS, et al. Association of primary care clinic appointment time with clinician ordering and patient completion of breast and colorectal cancer screening. JAMA Netw Open 2019;2:e193403.
4. Linder JA, Doctor JN, Friedberg MW, et al. Time of day and the decision to prescribe antibiotics. JAMA Intern Med 2014;174:2029-2031.