A consultant experienced in developing predictive models for his clients to improve business operations, Everett Blake, was working with the Richardson Eye Care and Surgery Center (RECSC), which wanted to explore ways to improve its no-show rate, classifying a “no-show” as an appointment not kept without 24-hours’ notice. After running variables through an XGBoost model, Blake’s company had identified predictive variables, which included age and gender. However, Blake was uncomfortable with these two variables, fearing that, if they were used to identify potential no-shows (and consequently not get follow-up calls about their appointments in the future), this could be considered discriminatory and might lead to a different quality of care for different groups of patients. With this in mind, Blake had to decide what his recommendation to RECSC should be.