Prompting vs Checklist to Reduce Empiric Antibiotics

In Review

Weiss CH, et al. "A clinical trial comparing physician prompting with an unprompted automated electronic checklist to reduce empirical antibiotic utilization". Crit Care Med.. 2013. 41(11):2563-9.
PubMed • Full text • PDF

Clinical Question

Among critically ill patients, does face-to-face prompting compared to an unprompted electronic health record checklist reduce empirical antibiotic utilization?

Bottom Line

Comparing unequal interventions, having a person stop the healthcare team during rounds and prompt them to consider antibiotic indication and stewardship was superior to an electronic checklist that the team did not have to complete in improving antibiotic use.

Major Points

Previous work has demonstrated that face to face prompting is effective at reducing the use of empiric antibiotics, lowering mortality and length of stay in the ICU. Use of a checklist may be effective in reminding clinicians to reassess therapy. In this trial by Weiss and colleagues published in 2013, daily ICU rounds were augmented with either an EHR pop-up prompt or an in-person prompt on the need to continue empiric antibiotics. Published in 2013, after enrolling 296 patients, in-person prompting was more effective at decreasing patient days receiving antibiotics and narrowing spectrum. The delivery of a required EHR checklist is less effective than face-to-face reminders possibly because of pop-up burden or lack of perceived usefulness of the process of reviewing appropriate antibiotics. From this trial we see that daily reminders and prompting are beneficial and that daily review of antibiotics are a helpful way to aid in antimicrobial stewardship.

Guidelines

As of March 2016, no guidelines include findings associated with this trial.

Design

• Single center, Random allocation (by coin flip), parallel group
• N=296
  • Electronic Health Record (n=125)
  • Live Prompting (n=171)
• Setting: Medical intensive care unit in a tertiary care teaching facility, closed unit, each team consisting of: 1 attending, 1 fellow, 1 pharmacist, several residents/interns
• Enrollment: 27 June to 7 October 2011
• Primary Outcome:
  • Empiric antibiotic duration
  • Proportion of antibiotic days receiving empiric antibiotics
• Secondary Outcomes:
  • Indication for antibiotics
  • Antibiotics used
  • Culture results
  • ICU and hospital length of stay (LOS)
  • APACHE IV predicted mortality and LOS

Population

Inclusion Criteria

• ≥1 day of empirical antibiotics

Exclusion Criteria

• Patients transferred to or from a different ICU service
• Any medical ICU (MICU) re-admissions without an intervening hospital discharge (first MICU admissions were included)

Baseline Characteristics

Presented as EHR (n=125) vs. Prompted (n=171)

• Age (yr), mean (sd): 62.6(17.6) vs. 60.0(17.8)
• Gender[male], no.(%): 66(52.8) vs. 88(51.5)
• Race, no. (%)
  • White: 52(42.3) vs. 80(47.9)
  • African American: 36(29.3) vs. 44(26.4)
  • Hispanic: 11(8.9) vs. 17(10.2)
  • Asian: 2(1.6) vs. 5(3.0)
  • Other or declined: 22(17.9) vs. 21(12.6)
• Location prior to medical ICU, no. (%)
  • Emergency Department: 67(54.0) vs. 92(53.8)
  • General medical ward: 53(42.7) vs. 64(37.4)
  • Outside hospital transfer: 4(3.2) vs. 15(8.8)
• Diagnosis, no. (%)
  • Sepsis: 36(29.0) vs. 66(38.6)
  • Pneumonia: 29(23.4) vs. 31(18.1)
  • Obstructive airways disease: 4(3.2) vs. 9(5.3)
  • Other respiratory: 17(13.7) vs. 12(7.0)
  • GI hemorrhage: 5(4.0) vs. 4(2.3)
  • Metabolic: 8(6.5) vs. 1(0.6)
  • Neurologic: 2(1.6) vs. 8(4.7)
  • Drug intoxication/withdrawal: 1(0.8) vs. 5(2.9)
  • Other: 22(17.7) vs. 35(20.5)
• Diagnosis, no. (%)
  • Sepsis or pneumonia: 65(52.4) vs. 97(56.7)
  • All other diagnoses 59 (47.6) 74 (43.3)
• Sepsis sub-diagnosis, no. (%)
  • Pulmonary: 5(13.9) vs. 18(27.3)
  • GI: 1(2.8) vs. 4(6.1)
  • Urinary: 10(27.8) vs. 9(13.6)
  • Soft tissue: 4(11.1) vs. 7(10.6)
  • Other or unknown: 16(44.4) vs. 28(42.4)
• Mechanical ventilation within 24 hr of ICU admission, no. (%): 49(39.5) vs. 56 (32.8)
• Hospital discharge disposition among survivors, no. (%)
  • Home: 81(85.3) vs. 104(74.3)
  • Skilled nursing or rehabilitation facility: 10(10.5) vs. 30(21.4)
  • Long-term acute care hospital: 2(2.1) vs. 5(3.6)
  • Short-term acute care hospital: 2(2.1) vs. 1(0.7)
  • APACHE IV predicted length of stay, d: 5.6(1.5) vs. 5.6(1.7)
  • APACHE IV predicted mortality, no.(%): 42.8(34.2) vs. 46.7(27.5)
• Patient-days with intervention: 166 vs. 55(72%)

Interventions

• EHR - Attending/fellow shown once how to access the Electronic Health Record (EHR) and was never prompted again. Question in EHR written as “Can antibiotic coverage be stopped or narrowed today?” with radio buttons for yes/no
• Live prompting - Team was given a paper checklist of 6 items (Vent Wean, Abx, CVC, Foley, DVT, SUP). Medical resident not directly involved in care listened to rounds Monday-Friday and if antibiotics not addressed, asked Attending or Fellow: 1) “Why is the patient on (antibiotic)?” and 2) “The (test [e.g., blood culture]) was (negative/positive) for (X [e.g., bacteria]). Do you plan to continue (antibiotic)?”

Outcomes

Comparisons are EHR (n=125) vs. live prompting (n=171) [p-Value] (SD/IRQ as specified)

Primary Outcomes

Patient days while on Empiric Antibiotics

70% vs. 63% [p=0.002]

Empiric Antibiotic Duration Days (IRQ)

3 (2-4) vs 3 (1-5) [NS]

Proportion of Antibiotic Therapy that was Empiric (SD)

0.83 (0.27) vs. 0.78 (0.27) [NS]

Total Days Antibiotic Duration (IRQ)

4 (2-8) vs. 4 (2-7) [NS]

Secondary Outcomes

Indication for Antibiotics

NOT REPORTED

Abx used

NOT REPORTED

Culture results

NOT REPORTED

ICU LOS, Days )

2.8(1.7–6.5) vs. 2.6(1.5–6.9)[NS]

Hospital LOS (Days)

9.6(5.9–15.8) vs. 11.8(5.9–22.8)[NS]

Hospital Mortality, no.(%)

30 (24.0) vs. 30 (17.5) [NS]

Standardized mortality ratio based on APACE score (90% CI)

0.70(0.47–1.0) vs. 0.64(0.43–0.92)[N/A]

Ventilator-free and alive at day 28, days

20.3(0–25.9) vs. 21.9(13.5–25.7)[NS]

Narrowing of antibiotics spectrum (Patient days, %)

16, 9.6% vs. 24, 43.6% [P<0.001]

Subgroup Analysis

Empiric ABx duration and Hospital Mortality, 1-3 days = reference

4-6 days = OR 1.3, 95% CI 0.55–3.1, P=NS

≥ 7 days = OR 3.2, 95% CI 1.5–6.8, P=0.002

Empirical antibiotic outcomes according to likelihood of infection

• Definite/probable infection
  • Patient-days on which empirical antibiotics were used, no. (%): NS
  • Proportion of antibiotic-days on which empirical antibiotics were used, mean (SD): NS
  • Empirical antibiotic duration (d), median [IQR]: NS
• Possible/empirical infection
  • Patient-days on which empirical antibiotics were used, no. (%): 452(73) vs. 637(66) P=0.002
  • Proportion of antibiotic-days on which empirical antibiotics were used, mean (SD): NS
  • Empirical antibiotic duration (d), median [IQR]: NS

Criticisms

• Small sample, unclear if balanced (no states done on demographics, patients not randomized)
• Team was randomized but cross contamination of teams by changing Attendings (5) and Fellows (1)
• Trial 2 years prior used similar 6 interventions and the “culture” in the ICU could have changed
• EHR checklist available in both groups
• No penalty or reminder to complete EHR
• Question was imbedded in the middle of the EHR (right in front of the “APACHE IV form” question)

Funding

• Supported in part by grant T32HL076139-07 from National Heart Lung and Blood Institute
• Several authors declared financial support which may represent a conflict of interest