Optimising Blood Culture Collection Techniques to Minimise Contamination Rates in a Busy Public Sector Paediatric Department in South Africa: A Quality Improvement Project

Abstract

Background Blood culture (BC) is the gold standard in diagnosing sepsis in hospitalised children. A limitation of BC is contamination with skin commensals, which gives rise to false positive results and may impact on health care expenses related to laboratory costs and increased hospital length of stay. We undertook a Quality Improvement Project (QIP) to evaluate BC contamination rates, laboratory costs and length of hospital stay in children admitted in the general paediatric wards at Chris Hani Baragwanath Hospital in Soweto, South Africa. Methods The QIP was implemented in three successive time periods each lasting two months (01 De- cember 2022 to 31 May 2023). The study population included all children aged 0-14 years from whom a BC was collected as part of routine clinical care on the day of admission. Pe- riod 1 was a pre-intervention period. During Period 2 (intervention), a Standard Operating Procedure (SOP) was implemented and admitting clinicians were taught how to adhere to the SOP. During unobserved BCs in Period 2, clinicians were expected to obtain BCs ac- cording to the SOP. All BCs obtained during Period 3 (post-intervention) were unobserved. Contamination rates, length of hospital stay and laboratory costs were compared between study periods. Results During Period 1 the contamination rate was 9.0% (66 of 732 BCs), compared to 5.5% (54 of 974 BCs) during Period 2; P=0.007. In Period 2, observed BCs (n=250) had a lower contamination rate compared to unobserved BCs (n=720) (3.9% versus 6.1%; P=0.253). In Period 3 BC contamination rebounded to 9.7% (71 of 732 BCs). In Period 1, the average cost of BC processing was R483.20 (IQR R483.20-R649.16), compared to R361.16 (IQR R361.16- R361.16) in Period 2 (p < 0.001). Cost increased to R480.56 (IQR R480.56-R480.56) in v Period 3. Children with contaminated BCs had longer hospital stay (median 5.0 (IQR 3.0 to 8.0) days) compared to those with negative BCs (4.0 (IQR 3.0 to 6.0) days; P < 0.001). Conclusions Implementation of a SOP significantly decreased the BC contamination rate. However, with- out reinforcement of correct technique, lower contamination rates were not sustained. Pa- tients with negative BCs had reduced hospital length of stay and laboratory costs compared to those with contaminated cultures, which demonstrates the economic burden of contami- nated BCs.