JMIR Publications

ABSTRACT

Background:

Antimicrobial resistance (AMR) is a global health emergency, disproportionately impacting sub-Saharan Africa where fragmented laboratory information systems (LIS) and inconsistent antimicrobial susceptibility testing (AST) practices limit both clinical decision-making and surveillance capacity. At the University Teaching Hospital of Butare (CHUB) in Rwanda, a baseline Laboratory Assessment of Antibiotic Resistance Testing Capacity (LAARC) identified systemic gaps, including 0% AST panel standardization, 17% cumulative antibiogram generation capacity, and no enforcement of testing quality rules. A standards-based digital infrastructure integrating CLSI- and EUCAST-compliant AST panels into OpenClinic GA with bidirectional WHONet export was implemented in 2024. The implementing study expressly deferred quantitative key performance indicator (KPI) evaluation to allow sufficient post-implementation observation time.

Objective:

To quantify the impact of the implemented infrastructure on five KPI domains: (1) AST data capture volume and specimen throughput; (2) data standardization and ease of analysis; (3) laboratory turnaround time (TAT); (4) testing quality, including detection and resolution of specimen-antibiotic incompatibilities, method violations, and intrinsic resistance violations; and (5) antimicrobial resistance (AMR) profiles, including appropriate use of screener discs.

Methods:

A retrospective pre-post study compared Q1 2024 (pre-implementation; n=505 culture-positive specimens, legacy LIS export) with Q1 2026 (post-implementation; n=3,669 specimens, WHONet-format export) at CHUB. Outcomes included specimen throughput, nomenclature and data structure standardization, TAT (specimen receipt to results release, days), testing quality flags (Nitrofurantoin/Nalidixic acid on non-urine specimens; vancomycin by disc diffusion on staphylococci; ampicillin on intrinsically resistant Enterobacteriaceae), screener disc utilization (Pefloxacin, Cefoxitin), and resistance rates for key pathogens against CLSI 2024 breakpoints.

Results:

Specimen throughput increased 7.3-fold (+626%). Unique organism names decreased from 76 to 39 (-49%), with 11 non-standard spellings of Escherichia coli alone pre-implementation reduced to one WHO-standard entry. All 7,465 post-implementation records carried CLSI 2024 annotations versus 0% pre-implementation. Critical data-analysis fields, age categories, department, TAT timestamps, WHONET antibiotic codes, absent pre-implementation were fully present post-implementation. Pre-implementation errors included: 6 instances of Nitrofurantoin or Nalidixic acid on non-urine specimens, 4 instances of ampicillin on intrinsically resistant Enterobacteriaceae (including 1 spurious susceptible result), and 44 vancomycin results by disc diffusion on staphylococci and streptococci (16 on Staphylococcus aureus), an unreliable method per CLSI. All resolved to zero post-implementation. Screener disc use was transformed: Pefloxacin (fluoroquinolone surrogate for Enterobacteriaceae) was absent pre-implementation and introduced in 95 tests post-implementation; Cefoxitin (MRSA surrogate) was used without a systematic inferential framework pre-implementation (n=81, no inferential display) and formalised under the screener protocol post-implementation (n=33, MRSA rate 24.2%). Median TAT from specimen receipt to results release was 2.80 days (IQR 2.03-3.77). Among key pathogens, Escherichia coli ciprofloxacin resistance was 54.2% and ceftriaxone resistance 42.2%, while carbapenem susceptibility was preserved (imipenem resistance 1.5%).

Conclusions:

Implementation of a CLSI/EUCAST-compliant digital AST and surveillance infrastructure at CHUB produced measurable, clinically meaningful improvements across all five KPI domains, yielding CHUB's first institution-wide cumulative antibiogram in GLASS-compatible WHONet format. These results fulfill the deferred quantitative evaluation and provide a replicable evidence base for tertiary institutions across sub-Saharan Africa.