In a critical assessment of antibiotic resistance in Germany, researchers have unveiled significant limitations in utilizing hospital consumption data to measure the efficacy of antibiotics used in treatment. This study, conducted by a team from the University Hospital Carl Gustav Carus Dresden, highlights that contradictory data sets prevent a clear understanding of how antibiotic consumption affects resistance development, particularly in the case of E. coli.
Over a five-year period from 2015 to 2019, the study analyzed ecological and clinical data through various computational models, including linear regression and artificial neural networks. This robust investigation revealed that, although overall antibiotic use in Germany has been stable, the composition of antibiotic consumption has changed significantly. Notably, fluoroquinolone usage decreased by approximately 6% and cephalosporin usage declined by about 4%, while penicillin consumption increased by around 10%.
The study utilized data from the ADKA-if-DGI-project, which aggregates antibiotic consumption data from over 300 acute care hospitals throughout Germany. Additionally, antibiotic resistance data for E. coli was sourced from the Antibiotika-Resistenz-Surveillance (ARS) database, maintained by the Robert Koch Institute. While this wealth of data has the potential to unveil patterns in antibiotic resistance, the researchers noted major discrepancies in findings, which they attribute to the aggregation of resistance data at the federal state level.
"None of our models is stable against small deviations," wrote the authors of the article, reflecting concerns about the reliability of the models developed from this aggregated data. The findings indicate that there is a need for more granular data, including monthly or quarterly antibiotic consumption and resistance reports at both the hospital and departmental levels, which would provide a clearer picture of the dynamics in play.
The nonlinear trends observed in ciprofloxacin resistance levels illustrate these challenges clearly; resistance peaked at around 20% in both the first quarter of 2015 and the second quarter of 2017, before dropping to about 17% by the end of 2019. In contrast, cefotaxime resistance levels fluctuated around 12%, demonstrating the inconsistencies and lack of clarity regarding the impact of antibiotic consumption on resistance development.
In examining the results from the linear and artificial neural network models, the study concluded that overall, the models did not perform well when applied to data aggregated at the regional level. This discrepancy further supports the researchers' call for the establishment of a systematic, detailed, and time-sensitive approach to data collection regarding antibiotic consumption and resistance. "This highlights the importance of systematically collecting and providing matched antibiotic consumption and corresponding resistance data at the hospital and department level across the country," emphasized the authors.
Given that antibiotic resistance is a pressing global health issue—predicted to cause 10 million deaths annually by 2050 without effective countermeasures—this study underscores the urgency of refining data collection methods. By enhancing the granularity and reliability of antibiotic monitoring systems, researchers believe that better models can be developed, leading to more effective antimicrobial stewardship programs, which are crucial in combating antibiotic resistance.
The implications of this analysis extend beyond the borders of Germany, offering insights that could aid other nations in refining their strategies against antimicrobial resistance. The researchers stress that without coordinated international efforts to enhance data quality and collection methodologies, the global healthcare community may struggle to manage the growing threat posed by resistant infections.
