Setting up a BSI Benchmarking Database for Comprehensive Cancer Centers

Abstract

ISSUE: There is currently no database which provides specific Blood Stream Infection (BSI) rates in cancer patients using current case definitions. The need for benchmarking in this specialty population prompted this collaboration. PROJECT: Eleven NCI designated cancer centers set out to determine the rate of BSI's in this population. Development of criteria for determining and stratifying BSI by their unique risks was essential. An initial data collection form included denominator data (service, patient days by unit and service) and the numerator (total number of infections). Facilities were assigned a designator to ensure anonymity in the final reports. RESULTS: Data were collected beginning in 2001, and submitted biannually. Reports were produced that compared BSI's from the participating facilities by service: Surgical, Bone Marrow Transplant (BMT), Hematology-Oncology and Pediatrics. Overall rates for facility comparison were provided. The initial data were useful but better demographics and risk factors would provide more meaningful data for evidenced based improvements in patient care. In 2004 work began on a web-based data entry for BSI's.Conference calls and 6 face-to-face meetings were held. The group gained consensus, a secured website was established and data entry via password was implemented. All patients who developed a BSI regardless of inpatient or outpatient status are included. Additionally primary oncology diagnosis, source of the culture, concurrent sites of infection, risk factors: chemotherapy, neutropenia and mucositis, type and location of vascular access devices and pathogen are in the database. All participants obtained a Limited Data Set Agreement in accordance with HIPAA guidelines. LESSONS LEARNED: Issues which required resolution to ensure that the data would be comparative included: a scientifically valid case-definition of a BSI in the neutropenic population, an accepted scoring system for mucositis and signs and symptoms of sepsis. For stratification purposes, defining populations such as BMT's, induction chemotherapy, were concerns. A major issue was the significance of a single positive blood culture with a generally "non-pathogenic" organism e.g. Coagulase negative staphylococcus, in the neutropenic population. Other issues included definitions of patient classification and how to collect denominator days.