 Machine learning, ML, offers the potential to analyze vast amounts of data without making prior assumptions about variables or relationships. However, its use in healthcare data has been met with mixed success due to the, black box, nature of some algorithms and the fact that data issues are often present in large administrative data sets. Good data and analytical design are essential to ensure accurate results from ML-based studies. This paper discusses common misconceptions about ML, the corresponding truths, and suggestions for incorporating these methods into healthcare research. This article was authored by Geron Arbett, Cole Brokamp, Jareen Meinzender, and others.