Impact of simple screening criteria on utilization of low-yield bacterial stool cultures in a children's hospital

Objective. To determine diagnostic yield of stool cultures for Salmonella, Shigella, Campylobacter jejuni, Yersinia enterocolitica, and Escherichia ca ll O157:H7 (SSCYE) among hospitalized children and to develop guidelines fo r appropriate use of these tests. Setting. Tertiary care pediatric hospital. Design. Computerized records from the Microbiology Laboratory from January 1992 to December 1996 were reviewed retrospectively to collect data on the number of stool cultures performed in inpatients and outpatients, the lengt h of hospital stay at the time cultures were sent, and diagnostic yield of cultures in hospitalized patients. A detailed review of medical records of all patients with a stool pathogen isolated after 3 days of hospitalization was also undertaken. The results from this retrospective analysis were used to develop guideline s to reduce unwarranted stool cultures and to educate medical care provider s in the appropriate use of these tests.; The impact of these guidelines on reduction in the volume of stool cultures performed on hospitalized patien ts was measured prospectively from January 1998 to June 1998. Results. A total of 27 110 stool cultures for SSCYE were performed in the 5 -year study period. Of the 14125 cultures from inpatients, 174 (1.2%) were positive. Among the cultures from inpatients, 9378 (66%) were from patients hospitalized for >3 days. Only 13 (.14%) were positive. Of these 13 cultur es, 4 represented nosocomial infections, whereas the remaining 9 cultures e ither were sent to document clearance from a patient known previously to be infected with an enteric pathogen (7), or were attributed to delayed testi ng in individuals admitted with a diarrheal illness (2). Introduction of guidelines to reject all SSCYE cultures from patients hospi talized for >3 days who did not meet specified criteria was associated with an overall reduction of 689 (43%) in the volume of tests performed in the 6-month period evaluated. This included 497 fewer cultures ordered and 192 cultures that were ordered but rejected because screening criteria were not met. Only II (5.4%) of 203 cultures sent >3 days after admission were proc essed because they met clinical criteria for testing. None were positive. E stimated cost savings were $50 163/year. Conclusions. Stool cultures for SSCYE among hospitalized patients have very low diagnostic yield and are extremely overutilized. Simple guidelines, su ch as rejecting (with few exceptions) cultures from patients hospitalized f or >3 days, can reduce substantially such unnecessary testing.