In response to the COVID-19 pandemic, we rapidly implemented a plasma coordination center, within two months, to support transfusion for two outpatient randomized controlled trials. The center design was based on an investigational drug services model and a Food and Drug Administration-compliant database to manage blood product inventory and trial safety.

A core investigational team adapted a cloud-based platform to randomize patient assignments and track inventory distribution of control plasma and high-titer COVID-19 convalescent plasma of different blood groups from 29 donor collection centers directly to blood banks serving 26 transfusion sites.

We performed 1,351 transfusions in 16 months. The transparency of the digital inventory at each site was critical to facilitate qualification, randomization, and overnight shipments of blood group-compatible plasma for transfusions into trial participants. While inventory challenges were heightened with COVID-19 convalescent plasma, the cloud-based system, and the flexible approach of the plasma coordination center staff across the blood bank network enabled decentralized procurement and distribution of investigational products to maintain inventory thresholds and overcome local supply chain restraints at the sites.

The rapid creation of a plasma coordination center for outpatient transfusions is infrequent in the academic setting. Distributing more than 3,100 plasma units to blood banks charged with managing investigational inventory across the U.S. in a decentralized manner posed operational and regulatory challenges while providing opportunities for the plasma coordination center to contribute to research of global importance. This program can serve as a template in subsequent public health emergencies.

To assess postoperative exposure to parenteral antibiotics and coded discharge diagnoses of infection as markers of nosocomial infection, postoperative morbidity, and potentially inappropriate antibiotic use after cesarean section.

Retrospective cohort study to compare automated markers with the criterion of record review.

Tertiary care hospital.

Women admitted to a large teaching hospital after April 15, 1987, and discharged before October 1, 1989, who underwent a nonrepeat, nonelective cesarean section and had received prophylaxis with a cephalosporin.

Antibiotic exposure and discharge diagnosis codes were obtained from a large electronic hospital data base. A sample of charts was reviewed to determine the presence of infection, other postoperative complications, and postoperative antibiotic exposure.

A total of 2,197 women who had undergone a nonrepeat nonelective cesarean section were included in the study cohort. These women were assigned to 6 subgroups based on postoperative antibiotic exposure status and discharge codes suggesting endometritis, other postoperative infection, or no infection. Review of 457 records indicated that the overall infection rate was 9%. Eight percent of all the patients had a coded diagnosis for infection, and 16% received some parenteral antibiotics after the first postoperative day. Exposure to at least 2 days of parenteral postoperative antibiotics was the best marker by which to discriminate between infected and uninfected patients, with a sensitivity of 81 %, a specificity of 95%, and a positive predictive value of 61% for detecting infection. The corresponding figures for coded diagnoses for infection had rates of 65%, 97%, and 74%, respectively. The combination of discharge codes and exposure to parenteral postoperative antibiotics resulted in a more accurate but less sensitive marker for nosocomial infections, with a positive predictive value of 94% and a sensitivity of 59%. The groups with discordant parenteral postoperative antibiotics exposure and discharge codes for infection were enriched for errors in coding, noninfectious morbidity, and unexplained antibiotic use. Less than 1% of the entire cohort had ≥ 2 days of parenteral postoperative antibiotics without any reason apparent in the medical record.

Parenteral postoperative antibiotic exposure determined from automated pharmacy records correlated with the results of the more labor-intensive manual review of medical records for the identification of nosocomial infection. In addition, information on antibiotic exposure combined with coded discharge diagnoses provided a rapid screen to identify subgroups of patients with higher rates of infectious and noninfectious morbidity, unexplained antibiotic use, and errors in discharge coding. Information derived from electronic data bases created for administrative purposes may be useful as a marker for infectious complications, inappropriate antibiotic prescribing, and other issues related to total quality hospital monitoring.