Datasets used to train clinical models for stroke and diabetes have been described as flawed. It is unclear if this issue is related to recent developments in academic publishing, including defamation verdicts and restrictions on authors with retractions. The quality of datasets used in clinical modeling is a concern, but the specific implications and outcomes are not clearly defined.