Background: Systematic reviews (SRs) provide more precise estimates of test sensitivity and specificity than single studies. Their interpretation requires consideration of the impact of test results on patient management and consequences for patient outcomes. Objectives: To describe concepts for the inclusion of data about patient management as an extension to SRs of test accuracy. Methods: We apply standard epidemiological principles and present examples to define key concepts for reporting test impact on patient management in SRs of test accuracy. Results: Review authors should state assumptions about changes in management and consequences for patient outcomes due to detection of ‘extra’ true-positives (TP)/false-negatives (FN)/true-negatives (TN)/false-positives (FP) when comparing the sensitivity and specificity of two tests (Figure 1). a) If assumptions that all extra cases will receive the specified management change are straightforward, no further evidence about management is needed (Example 1). b). If uncertainty exists, additional data may be required to estimate what proportion of extra cases will receive a change in management (Example 2). These data may be found in accuracy studies and can be summarised to aid interpretation of SRs to clinical practice (Table 1). Unfortunately, they are often not clearly or adequately reported. The GRADE approach can then be used to judge evidence about the effects of these management changes on patient outcomes. Conclusions: Patient management cannot always be inferred from test accuracy results but are relevant for interpretation of these results, e.g. if the index test is more sensitive and less specific than the comparator, information about the proportion of extra TP who will receive a change in treatment may be important when weighing up treatment benefits against the harms of extra FP. Review authors can identify situations where empirical evidence about the changes in management will assist interpretation of accuracy results and present available data in a table with a summary estimate.