Opt-Out Protocol Decreases Unnecessary Antibiotic Use for Suspected Sepsis

Patients (N=767) with suspected sepsis were enrolled in a patient-level randomized trial, which was conducted at 10 sites in the United States between 2018 and 2020. Researchers aimed to determine whether the use of an opt-out protocol decreases unnecessary antibiotic treatment for suspected sepsis. Patients were randomly assigned in a 1:1 fashion to receive either the opt-out protocol (n=383) or usual care (controls; n=384).

The intervention involved a pharmacist-led discussion with the clinician to encourage antibiotic discontinuation using opt-out language. Clinicians were asked to provide rationales and de-escalation plans for patients in whom antibiotic treatment was continued as opposed to per-protocol antibiotic cessation. The primary outcome was the number of antibacterial days of therapy (DOT) at 30 days following enrollment.

Among patients included in the intervention and control cohorts, the median age was 63 (IQR, 49-73) and 66 (IQR, 53-76) years, 49% and 45% were women, 49% and 51% were White, 32% and 29% were hospitalized within the past 90 days, and the median Elixhauser comorbidity score was 11 (IQR, 5-19) and 11 (IQR, 4-20), respectively.

The most common diagnoses indicating antibiotic treatment among the patients included urinary tract, skin and soft tissue, intra-abdominal, and respiratory infections.

The overall length of hospitalization was shorter among patients in the intervention cohort compared with those in the control cohort (median, 5 vs 6 days, respectively).

Following study enrollment, patients in the intervention vs control cohorts had similar DOT (mean, 8.2 vs 8.3 days) and similar rates of de-escalation plans instituted at day 5 following blood culture collection (57.5% vs 52.9%).

Compared with patients in the control cohort, the number of patients who continued receiving antibiotics (odds ratio [OR], 0.68; 95% CI, 0.47-0.98) or rank 3 extended-spectrum antibiotics (OR, 0.73; 95% CI, 0.55-0.98) following enrollment was decreased among those in the intervention cohort.

A total of 133 and 157 major safety events occurred among patients in the intervention and control cohorts, respectively. The most common safety signals among patients in the intervention vs control cohorts were hospital  (15.9% vs 14.8%), recurrence of suspected sepsis (7.8% vs 7.8%), intensive care unit admission (6.8% vs 8.6%), and reinitiation of inpatient antibiotics after more than 48 hours of no antibiotics (4.2% vs 4.2%).

At 30 days, the opt-out protocol was associated with the probability of a better 6-level desirability of outcome ranking (DOOR) and a better treatment response, with adjustments for duration of antibiotic risk (RADAR) profiles (OR, 0.52; 95% CI, 0.48-0.56).

Antibiotics were discontinued among 59 patients in the intervention cohort. Of these patients, most (97%) were treated in a medicine service. None of the patients in the intervention cohort who discontinued antibiotics were treated in infectious diseases or specialty surgery services, compared with 2% and 10% of those who continued antibiotic treatment, respectively. The most common rationale for continuing antibiotics was for the treatment of a localized infection, which was reported in 76% of the patients.

This study was limited by the inclusion of patients from both community and academic hospitals with varied resources and antibiotic stewardship practice.

According to the researchers, “Future stewardship interventions must more directly address diagnostic uncertainty and perceptions of safety when addressing antibiotic decision-making for patients with suspected sepsis.”

Disclosure: Multiple authors declared affiliations with industry. Please see the original reference for a full list of disclosures.