Selective labels occur when label observations are subject to a decision-making process; e.g., diagnoses that depend on the administration of laboratory tests. We study a clinically-inspired selective label problem called disparate censorship, where labeling biases vary across subgroups, and unlabeled individuals are imputed as “negative” (i.e., no diagnostic test = no illness). Inspired by causal models of selective labels, we propose Disparate Censorship Expectation Maximization (DCEM). We theoretically analyze how DCEM mitigates disparate censorship. We validate DCEM on synthetic data, showing that it improves bias mitigation (area between ROC curves) without sacrificing discriminative performance (AUC) compared to baselines. We achieve similar results in a sepsis classification task using clinical data.