Methods

Data Source

ANRESIS collects and analyses anonymous antibiotic resistance data submitted weekly or monthly by a representative network of Swiss clinical microbiology laboratories. The data are generated during routine medical care (privacy statement) and are reported as delivered by the laboratories. ANRESIS performs no data validation. All laboratories providing data for this report are approved by Swissmedic and participate in at least one external quality control program. Most laboratories use semi-automated systems, based on EUCAST guidelines. Quantitative resistance data, such as minimal inhibitory concentrations (MIC) or disc diffusion diameters (DD), are delivered by a subset of laboratories only. Therefore, all analyses are based on the interpreted quantitative resistance test results (susceptible, susceptible with increased exposure, resistant; S-I-R). 

The dataset covers approximately 90% of annual hospitalisation days in Switzerland (see interactive graph for details) and at least 50% of outpatient practitioners. Available epidemiological variables allow stratification by healthcare setting (hospital vs. outpatient), age group and anatomical location. During data upload, automated algorithms, are applied to identify duplicates and perform other quality checks.

Interpretation considerations: As analyses are based on interpreted qualitative resistance data ('S-I-R'), changes in breakpoints over time may affect observed resistance trends. Moreover, according to EUCAST, different breakpoints may be used for different types of infections. The number and type of participating laboratories change over time, which may affect the results. Annual isolate counts submitted to ANRESIS are available in a stacked bar chart.

Data Analysis

Details on individual analyses are provided under "additional information" in the relevant figures. Results from reference laboratories and screening results are labelled by the laboratories and excluded from most analyses, as are duplicate results (same patient, same microorganism, same calendar year). Because patient identifiers are laboratory-specific, duplicate isolates originating from different laboratories could not be identified and removed. Most analyses were restricted to isolates recovered from blood cultures or cerebrospinal fluid to facilitate comparison with international surveillance data.

Resistance proportions were calculated as the number of resistant isolates divided by the total number of isolates tested. An isolate was classified as resistant to an antibiotic group if resistance was detected to at least one antibiotic within that group. The Wilson score method was used for the calculation of the 95% confidence interval of the resistance proportions.

For the calculation of incidence rates, only invasive isolates from hospitals submitting at least one sample per bed and year were included. Isolate counts were extrapolated to the Swiss population based on the proportion of national hospital bed-days covered by participating hospitals. Extrapolation was stratified according by linguistic region and by university vs. non-university hospital status. Incidences rates are reported as cases per 100'000 population. An exact Poisson confidence interval based on the gamma parameterisation of the exact (Garwood) Poisson interval was used, which generalises the chi-squared formulation to fractional event counts.

The Resistance Monitoring programme in livestock and meat is yearly defined for specific pathogens from pigs, cattle, poultry and turkey. Stratified random samples are taken since 2016. Sampling is spread evenly throughout each year, based on a sampling plan established for meat inspections. Samples are collected at the two largest poultry slaughterhouses as well as at the four largest pig and five largest cattle slaughterhouses. Every slaughterhouse taking part in the program collected a number of samples proportional to the number of animals of the species slaughtered per year. Samples are sent to the national reference laboratory for antimicrobial resistance ZOBA, Vetsuisse Faculty, University of Bern, for further analyses. In accordance with the European legislation, meat samples are taken from fresh, chilled, packed and untreated meat sold at the retail level. Samples are collected in all Swiss cantons throughout the year. The applied sampling scheme considered each canton’s population density and the market shares of the retailers. Moreover, the proportion of imported and domestically produced meat within each meat category was included in the sampling plan.

The Resistance Monitoring programme in diseased animals was initiated in 2019 and consolidated in 2022 to include pathogen/animal and indication combinations which are of relevance in veterinary medicine. The strains are isolated from diagnostic submissions of nearly all Swiss veterinary diagnostic laboratories (university, cantonal, private) in order to achieve good representativeness of the results.

Determination of the minimal inhibitory concentrations (MICs) are performed at the Swiss national reference laboratory (ZOBA).