We developed a computerized decision support system for prescription of antibiotics to inpatients. The purpose of the study is to assess the performance of the system in different wards, in three different hospitals, in three countries.

Antibiotic treatment for suspected moderate to severe bacterial infections is usually initiated empirically, prior to identification of the causative pathogen. Appropriate treatment, that is matching in-vitro susceptibilities of subsequently isolated pathogens, reduces the overall fatality rate of severe infections with adjusted odds ratios varying between 1.6 and 6.9. In the same studies, 20-50% of patients were given inappropriate empirical antibiotic treatment.

We developed a computerized decision support system (TREAT) based on a causal probabilistic network to improve antibiotic treatment of inpatients. The aims of the system were to improve the rate of appropriate antibiotic treatment, thereby reducing mortality, and to route antibiotic use towards ecologically economical antibiotics as determined by local resistance profiles. The system can be calibrated to different locations.

The TREAT system was tested in a multi-center observational cohort study. The study proved the system safe and effective. TREAT prescribed appropriate antibiotic treatment to 70% of patients, 58% of whom were treated appropriately by physicians. TREAT used a narrow antibiotic formulary and at lower costs, mainly lowering costs assigned by the model to future resistance. The system performed well in three different countries (Israel, Italy and Germany).

We then proceeded to assess the effect of TREAT on the management of inpatients in these sites in a cluster randomized controlled trial. We used wards as the unit of randomization to avoid contamination through education of users by the system, and to benefit from the interaction of TREAT with the ward as a whole.

Comparison: the TREAT system was installed in intervention wards and its use was offered to physicians at the time of empirical antibiotic treatment. Physicians were asked to inspect TREAT's result interface. The final choice of antibiotic treatment was theirs. Control wards had no access to the system. We assessed outcomes in intervention vs. control wards with regard to patient outcomes, appropriateness of antibiotic treatment and antibiotic costs.

Inclusion Criteria:

• Patients from whom blood cultures were drawn.
• Patients prescribed antibiotics (not for prophylaxis).
• Patients fulfilling sepsis diagnostic criteria.
• Patients with a focus of infection.
• Patients with shock compatible with septic shock.
• Patients with febrile neutropenia

Exclusion Criteria:

• HIV positive patients with a current (suspected or identified) opportunistic disease and/or AIDS defining illness currently or within the past six months
• Organ or bone marrow transplant recipients
• Children <18 years; suspected travel infections or tuberculosis
• Pregnant women
• Re-entries