Author: Ivor Horn, Kulleni Gebreyes
Publisher: World Economic Forum
Publication Year: 2022
Summary: In the following article, 2 healthcare industry professionals walk through 3 considerations all providers should make. The first big pitfall most providers fall into is assuming a generalizable patient base when the sampling only revolved around a certain group. For example, pharmaceutical trials nearly exclusively used white males as their test subjects, while generalizing their (positive) results to the entire population with adverse ripple effects. Second, data collection does not reflect some hidden variable that gets predicted upon adversely. For example, marginalized groups may be more likely to incur debt on loans, but they have repeatedly been preyed upon for payday loans, which are naturally more cumbersome to pay back without incurring some debt. Finally, using race-based models (or those based on some other sensitive group) leads to inequitable outcomes. For example, Black patients appeared less ill because their kidney filtration rates were higher on average, and they did not receive life-saving injections when they needed them.