This study, by Gwan et al., examines how healthcare professionals may classify pregnant individuals with diabetes as having “poor glucose control” as well as the implications of that label for perinatal outcomes. Researchers conducted a retrospective analysis of 1,433 pregnancies between 2018 and 2019 in the MHealth-Fairview system and used natural language processing (NLP) to identify instances where “poor glucose control” was documented in electronic health records. Researchers found that 10% of patients were labeled as having poor glucose control, often without objective evidence such as elevated A1c levels or documented abnormal glucose values. Physicians, particularly in obstetrics/gynecology and maternal-fetal medicine, were found to have used this designation most frequently. Patients of color, those with public insurance, and non-English/Spanish speakers were also more likely to receive this classification, even with comparable glycemic values to those who were not of color, had public insurance, and spoke fluent English. While those labeled as “poor glucose control” sometimes had earlier deliveries and more neonatal complications, differences largely disappeared when analyses were organized by actual glucose measures. Researchers emphasize that subjective language in charting may reflect bias and call for clearer, objective thresholds in clinical guidelines.