[…] The wide variation in testing coverage, and hence data representativeness, is a major limitation in interpreting AMR rates globally and nationally and the results should be interpreted with caution. Complementary AMR surveillance approaches, such as prospective AMR prevalence surveys, should be considered to achieve nationally representative AMR data and trends to evaluate and inform the AMR response, especially in low-resource countries with weak routine surveillance. In addition to the variation in testing coverage, one gap identified in this report is the low proportion of CTAs (49%) performing external quality assurance in all clinical laboratories that serve the national AMR surveillance systems, which essentially means that not all test results are quality assured. Although current data cannot be used to calculate robust national and global resistance estimates for the combinations under surveillance, they do allow to identify combinations where resistance remains low and where high resistance is a cause for concern. This is particularly true where larger numbers of CTAs contributed data and where median global percentage resistance is consistent, regardless of coverage.