Background: Due to the rising incidence of ESBL-producing bacterial infections, the use of carbapenems has increased over recent decades. Carbapenems are part of the group of last-resort antimicrobials and are used widely as empirical therapy, which is contributing to the growing rate of antimicrobial resistance (AMR). De-escalation has been proven to be a successful tool in antimicrobial stewardship programmes (ASPs) in minimising the occurrence of AMR and decreasing the use of antimicrobials. The purpose of the study was to find the reasons why prescribers do not de-escalate from empiric carbapenem therapy. Methods: This retrospective quantitative study was conducted in a private hospital in South Africa. The infection markers and cultures of these patients were considered. Results: De-escalation was practiced in 17% of the patients. Empiric carbapenem therapy was started in 11.2% of patients and the most prescribed carbapenem was ertapenem (62.4%). Cultures were available in 71.1% of the study population. De-escalation was not performed in 83% of patients, mostly since their infection markers decreased with carbapenem therapy (45.9%) or because of culture unavailability (28.9%). Conclusion: The study came to the conclusion that prescribers do not want to de-escalate once their patients are improving on current treatment or if there are no cultures available.

Due to the rising incidence of ESBL infections, the use of carbapenems has increased over the past decades. Carbapenems are part of the last resort antimicrobials and are used widely as empirical therapy, which is contributing to the growing rate of antimicrobial resistance (AMR). De-escalation has been proven to be a successful tool of antimicrobial stewardship programmes (ASPs) in minimising the occurrence of AMR and decreasing the use of antimicrobials. The purpose of the study was to find the reasons why prescribers do not de-escalate from empiric carbapenem therapy. This retrospective quantitative study was conducted in a private hospital in South Africa. The infection markers and cultures of these patients were considered. De-escalation was practiced in 17% of the patients. Empiric carbapenem therapy was started in 11.2% of patients and the most prescribed carbapenem was ertapenem (62.4%). Cultures were available in 71.1% of the study population. De-escalation was not performed in 83% of patients, mostly since their infection markers decreased with carbapenem therapy (45.9%) or because of culture unavailability (28.9%). The study came to the conclusion that prescribers do not want to de-escalate once their patients are improving on current treantment or if there are no cultures available.