Antimicrobial resistance: mapping the burden, trends and evaluating interventions across the UK healthcare economy

Abstract

Background: Antimicrobial resistance (AMR) is a growing threat to global health security. Escherichia coli is a frequent cause of bacteraemia and one of the key microorganisms that contributes to the burden of AMR. Methods: This PhD first explores the overall scale of the burden of AMR amongst Gram- negative bacteraemia in the United Kingdom (UK), across five key organisms (E. coli, Serratia spp., Proteeae spp., Klebsiella spp., and Enterobacter spp.). The chapter uses data from a surveillance programme for AMR at the British Society for antimicrobial Chemotherapy (BSAC) to undertake a trend analysis for the aforementioned organisms. The PhD then focuses on AMR in E. coli causing bacteraemia, using data from a mandatory and voluntary national laboratory surveillance dataset at Public Health England (PHE). The mandatory surveillance is for reporting the incidence of all reported E. coli bacteraemia cases in the country and the voluntary surveillance is for reporting AMR in E. coli causing bacteraemia at Public Health England (PHE). The PhD explores three methods for analysis of AMR trends and generates forecasts for 11 key antibiotics which are monitored for AMR (beta-lactam inhibitor combinations [co-amoxiclav, piperacillin/tazobactam], fluoroquinolones [ciprofloxacin], third generation cephalosporins [cefpodoxime, ceftazidime, cefotaxime, ceftriaxone], aminoglycosides [gentamicin] and carbapenems [ertapenem, imipenem, meropenem]) . The PhD then goes on to investigate prevalence of antimicrobial and multidrug resistance (MDR) and determine risk factors for both, using the aforementioned mandatory and voluntary surveillance datasets from PHE. The study includes a multilevel analysis to investigate variation in AMR that exists between regions. Finally, ecological interrupted time series modelling is used to conduct an evaluation of two national system-level interventions under the Quality Premium (QP) initiative implemented in England to curtail antibiotic prescribing and AMR associated with E. coli bacteraemia. Results: Trend analysis for AMR in E. coli-causing bacteraemia in England showed increases in resistance to key antibiotics in community- and hospital-onset bacteraemia cases between 2013 to 2018. Risk factors for AMR in E. coli-causing bacteraemia to 11 key antibiotics were assessed in both care settings. Risk factors for AMR in community-onset cases were elderly patients (>65 years), male patients and patients from the Northeast of England after adjustment for potential confounders. Analysis of hospital-onset bacteraemia showed an association between resistance of E. coli isolates to at least one antimicrobial agent and samples taken in 2018, isolates taken from patients in regions in the most deprived areas and from male patients after adjustment for potential risk factors. In both community- and hospital- onset cases, the odds of MDR were highest amongst isolates taken from patients aged over 65 and those residing in the East of England. The odds of MDR E. coli bacteraemia cases were higher in isolates taken from bacteraemia cases that occurred in 2018, from male patients and from patients residing in urban settings. This PhD work has also demonstrated that whilst the national incentive-orientated interventions have succeeded in reducing antibiotic prescribing, resistance in E. coli causing bacteraemia, though attenuated following the implementation of these interventions, remains on an upward trajectory. Finally, an analysis of a second national intervention to prevent incidence of E. coli bacteraemia has been successful for hospital-onset cases but showed no impact for community-onset cases. The implementation of this national intervention did however succeed in reducing MDR in E. coli causing bacteraemia. Conclusion: This is the first study in England to conduct trend analyses, analyses for measures of association and national intervention evaluation using linked data spanning a range of epidemiological datasets, including PHE and BSAC data, for AMR associated with E. coli bacteraemia. This PhD analyses various age groups and English regions using multiple statistical methods. This PhD also provides a lens into the strengths and limitations of routine surveillance data for E. coli AMR in England. Finally, this PhD makes recommendations for interventions and additional future research to investigate risk factors further, to curtail the impact of AMR in all strata of the healthcare economy.