Do experts make biased decisions? In this study, we examine the decision-making of emergency department physicians who must decide whether to admit or discharge patients. We find evidence suggesting that these physicians are subject to biases in their decision-making. Specifically, we discovered that a physician's decision for a given patient can be correlated with her decision for the previous patient, even if the two patients are unrelated. After accounting for all relevant factors, we found that (1) when two consecutive patients are dissimilar, the physician's decisions for them are nearly independent; (2) when two consecutive patients are similar in all respects, the physician tends to make the same decision for both; and (3) when two consecutive patients differ in severity level but are similar in other ways, the physician tends to make opposite decisions for them. Our findings align with the memory and attention mechanism, which suggests that more recent and similar experiences are easier for decision-makers to recall and can affect their decision-making. The similarity or differences between the decision maker's recalled experiences and the current decision problem can be exaggerated and thus lead to biased decisions, even among experts. Being aware of this mechanism can help experts in the business area to avoid decision biases and make better decisions.