Antimicrobial use in the ICU: indications and accuracy--an observational trial

J Hosp Med. Nov-Dec 2012;7(9):672-8. doi: 10.1002/jhm.1964. Epub 2012 Aug 3.


Background: In intensive care unit (ICU) patients, signs of infection and inflammation are similar, making diagnosis of bacterial infections difficult. Antimicrobials may therefore be overused, contributing to development of antimicrobial-resistant bacteria.

Objectives: To measure the accuracy of clinician decisions to start antimicrobials; to correlate clinician certainty with the presence of infection; and to examine whether physiological variables correlate with clinician certainty.

Design: Prospective observational study.

Setting and patients: Patients staying >48 hours in a general ICU of a tertiary care hospital.

Measurements: The ICU clinician's certainty for the presence of infection was recorded when starting antimicrobials. An independent infectious diseases (ID) specialist determined if antimicrobials were required and if infection was present. Clinician antibiotic start decisions were tested for accuracy according to the ID determination for the presence of infection.

Results: Empirical antimicrobial therapy was justified by the presence of infection on 67/125 (54%) occasions. Clinician certainty for infection correlated well with the presence of defined infection (r(2) = 0.78), however, infection was defined on 6/19 (31%) occasions when ICU clinician certainty was low (≤2), and antimicrobials were prescribed even when clinician certainty was minimal. Antimicrobial course length was similar whether infection was defined or not (11.5 ± 9.2 vs 10.7 ± 9.1 days; P = 0.65). Physiological variables were not associated with clinician certainty of infection.

Conclusions: Antimicrobial therapy is probably overused in the ICU, possibly resulting from difficulties in diagnosis and the perceived greater risk of untreated infection when compared to the risks of potentially unnecessary antimicrobial therapy. Efforts to improve antimicrobial-related decision-making should be mandatory.

MeSH terms

  • Adult
  • Aged
  • Anti-Bacterial Agents / administration & dosage*
  • Anti-Bacterial Agents / therapeutic use
  • Anti-Infective Agents / therapeutic use
  • Bacterial Infections / drug therapy*
  • Bacterial Infections / microbiology
  • Drug Utilization / statistics & numerical data*
  • Female
  • Humans
  • Intensive Care Units / statistics & numerical data*
  • Male
  • Metronidazole / therapeutic use
  • Middle Aged
  • Practice Patterns, Physicians' / statistics & numerical data*
  • Prospective Studies
  • Risk Factors
  • Socioeconomic Factors
  • Tertiary Care Centers / statistics & numerical data


  • Anti-Bacterial Agents
  • Anti-Infective Agents
  • Metronidazole