Data collection and analysis

How are the data collected and processed?

EARS-Net is based on routine clinical antimicrobial susceptibility data from local and clinical laboratories reported to ECDC by appointed representatives from the Member States. The data originate form national AMR surveillance initiatives and/or laboratory networks. Only data from invasive isolates (blood and cerebrospinal fluid) are included in EARS-Net.

What surveillance data are collected?

EARS-Net performs surveillance of antimicrobial susceptibility of seven bacterial pathogens commonly causing infections in humans:

  • Escherichia coli 
  • Klebsiella pneumoniae 
  • Pseudomonas aeruginosa
  • Acinetobacter species
  • Streptococcus pneumoniae
  • Staphylococcus aureus 
  • Enterococcus faecalis
  • Enterococcus faecium

The EARS-Net reporting protocol defines the panels of antimicrobial agent combinations under surveillance for each species. In addition, the EUCAST guidelines for the detection of resistance mechanisms and specific types of resistance of clinical and/or epidemiological importance explain the mechanisms of resistance and describe the recommended methods of detection for key species–antimicrobial group combinations. For further details on EARS-Net data collection and analysis, please refer to the latest EARS-Net reporting protocol.

Publication of results

Interpretation of results

Surveillance data on antimicrobial resistance should be interpreted with caution, especially regarding inter-country comparisons, but also national trends. A number of factors can influence data quality and introduce bias to the data, resulting in over- as well as underestimation of resistance percentages. Some of the most important potential sources of bias in EARS-Net are explained below.

Population coverage

Population coverage varies among reporting countries. Some countries report data from large national surveillance systems with a high national coverage, while other countries report data from a smaller subset of local laboratories and hospitals.

For countries reporting data from only a small number of hospitals and laboratories located in one specific geographical area, the sample may not be representative for the whole country. Likewise, national trends may not be representative of regional situations as pooled data could mask variations at the local level.

For some countries, the population under surveillance is not constant and may change over the years due to variations in the number of participating laboratories. To control for this potential bias in trend analyses, an additional sensitivity analysis including a subset of data originating only from laboratories reporting for all the previous four years is provided for all national trend analyses.

Sampling

EARS-Net data are exclusively based on invasive isolates from blood or cerebrospinal fluid. The clinical relevance of indicator organisms isolated from these sites is undisputable. This restriction prevents some of the inconsistencies that arise from differences in clinical case definitions, different sampling frames or heterogeneous healthcare utilisation that would otherwise confound the data analysis if isolates from all anatomical sites were accepted. However, invasive isolates may not be representative of isolates of the same bacterial species from other type of infections, i.e. urinary tract infections, pneumonia, wound infections, etc.

Case ascertainment of patients with bloodstream infections (BSIs) is strongly linked to diagnostic practices and the frequency with which blood cultures are taken. Therefore, variations in blood culture frequency (non-differential sampling) result in an increasing uncertainty when comparing resistance percentages between hospitals and countries. Extrapolations of EARS-Net data as a measure of BSI incidence could therefore underestimate the true value in countries with low blood culture frequency.

Differential sampling can occur if blood cultures are only performed after empirical treatment shows no adequate therapeutic response. Predictably, this will lead to an overestimation of the resistance percentage by not including susceptible BSI isolates from the denominator.

Laboratory routines and capacity

The use of guidelines for clinical breakpoints varies among countries in Europe, and in some instances even between laboratories in the same country. At present, many European laboratories are changing from using Clinical and Laboratory Standards Institute (CLSI) to EUCAST clinical guidelines. As a result, the interpretation of AST results may vary, at least for resistance mechanisms resulting in minimum inhibitory concentrations (MICs) close to the breakpoints. In addition, clinical breakpoints may change over time, as breakpoints may be revised. As quantitative data (i.e. disk diffusion zone diameters or MIC values) are not provided by all participating laboratories, only the reported S, I, and R results are considered for the analyses.

The ability to identify the microorganism and its associated antimicrobial susceptibility pattern may differ among laboratories. All laboratories providing data for EARS-Net are offered participation in an annual external quality assessment (EQA) to assess the reliability of the laboratory test results.

For more information on the EARS-Net EQA and laboratory performance, see the EARS-Net Annual Report.