Gatekeeping under Congestion : An Empirical Study of Referral Errors in the Emergency Department

Using data from over 300,000 visits to an emergency department (ED), we study the accuracy of gatekeeping decisions (admission to the hospital or discharge home) and the effectiveness of a second gatekeeping stage (a clinical decision unit (CDU) in our context) at reducing gatekeeping errors. While only 9.9% of patients in our sample are routed through the CDU, we find that had the unit not been in place during the observation period, then the rates of unnecessary hospitalization and wrongful discharge of patients from the ED would have increased by 14.3% and 29.6%, respectively. We find that a key contributor to the success of the two-stage gatekeeping system is admission control, which ensures that only those higher-risk patients who stand to benefit most from being placed into the CDU are referred there. In fact, we estimate that had patients instead been allocated at random to the CDU, then the unit's effectiveness would have reduced by more than half. We also investigate a critical trade-off in operating a two-stage system: resources must be split between the two stages, increasing congestion in the first stage. We demonstrate in our context, though, that the combination of an ED with a CDU performs better than a commensurately enlarged ED without a CDU that pools the capacity of both stages. Together, our results provide insights as to the circumstances under which adding additional stages to a gatekeeping system may be advantageous