Local antibiotic resistance surveillance in nursing homes using lot quality assurance sampling: a feasibility study

Abstract

Objectives: To assess the feasibility of lot quality assurance sampling (LQAS) for defining high prevalence of antibiotic-resistant (ABR) uropathogens in Dutch nursing homes.

Methods: In this cross-sectional study, we constructed 'lots' of residents with urinary tract infections (UTIs) across three Dutch nursing home organizations (21 locations, 2095 beds) between February and July 2023. Resistance thresholds for common antibiotics used in empirical antibiotic therapy (EAT) were set at 20% (upper) and 5% (lower). Per organization ≥44 strains of E. coli/Klebsiella spp. were consecutively sampled; ≥ 5 resistant strains indicated 'high' ABR prevalence. Aggregate data were used in an adapted, meta-regression model to identify structural ABR determinants. Three multidisciplinary focus groups were conducted to identify implementation requirements for LQAS-based ABR surveillance.

Results: From 298 urine specimens 132 E. coli/Klebsiella strains were identified. Across all three organizations, amoxicillin-clavulanic acid (first-choice EAT for pyelonephritis) was classified as 'high', while seven other antibiotics, including three EAT agents, had varying classifications. Higher ABR was associated with higher proportions of rehabilitation/short-stay care beds, higher number of shared bathrooms, higher hospital admission rates and higher antibiotic usage. Focus groups revealed that defining a ABR upper threshold for LQAS was difficult, that multiple factors influenced the choice of EAT (e.g. side effects), and that the role of ABR prevalence in EAT-related decision making was unclear.

Conclusions: Using LQAS to classify E. coli/Klebsiella ABR in nursing homes was feasible within 4 months. To effectively inform local EAT, consensus is needed on the threshold of maximum acceptable level of ABR.