From page $393,$ "Analytics in Practice"

The East Carolina University $($ECU$)$ Student Health Service $($SHS$)$ provides health care services and wellness education to enrolled students. Patient volume consists almost entirely of scheduled appointments for non$-$urgent health care needs. In a recent academic year, $35,050$ appointments were scheduled. Patients failed to arrive for over $10\%$ of these appointments. The no$-$show problem is not unique. Various studies report that no$-$show rates for health service providers often range as high as $30\%$ to $50\%.$

To address this problem, a quality$-$improvement $($QI$)$ team was formed to analyze an overbooking option. Their efforts results in developing a novel overbooking model that included the effects of employee burnout resulting from the need to see more patients than the normal capacity allowed. The model provided strong evidence that a $10\%$ to $15\%$ overbooking level produces the highest value. The overbooking model was also instrumental in alleviating staff concerns about disruption and pressures that result from large numbers of over scheduled patients. At a $5\%$ overbooking rate, the staff was reassured by model results that predicted $95\%$ of the operating dates with no patients being over scheduled; in the worst case, eight patients would be over scheduled a few days each month. In addition, at a $10\%$ overbooking rate, the model predicted that during $85\%$ of the operating dates per month, no patients would be over scheduled; a maximum of $16$ over scheduled patients would rarely occur.

Based on the model, the SHS implemented an overbooking policy and overbooked by $7\mathrm{.}3\%$ with plans to increase to $10\%$ in future semesters. The SHS director estimated that the savings from overbooking during the first semester of implementation would be approximately $$95,000.$