Background The rate of unscheduled return visits is often used as a quality-of-care indicator in EDs, although its validity is not yet fully established. Our aim was to identify the characteristics of return visits that may be attributed to problems in quality of care.

Methods Retrospective paired review of medical charts in a random sample of return visits during the 72 hours following discharge from the ED in three hospitals of Andalusia, Spain in 2013. Charts were reviewed by senior medical physicians to determine which return visits reflected quality-of-care problems. Time frame for return visit, index and return visit acuity, disposition and diagnosis were compared with determine which variables were associated with a quality problem. Sensitivity and specificity for each variable to indicate a quality problem were determined.

Results We studied the causes of 895 return visits, finding that 65 (7.3%) were due to inadequate quality of care in the index visit. Potentially avoidable return visits were more common in more severely ill patients, in those with greater severity in the return than in the index visit and in patients hospitalised after the return. The combination of this three variables presented sensitivity 66% and specificity 68% in identification of quality-related returns.

Conclusions The overall level of return visits cannot be considered a valid indicator of quality of care. However, certain specific variables, including the level of severity of the patient's condition or the discharge destination following the return visits, could be considered valid in this respect.