This week, Peter Simons covers a study that found the validity of RCTs on schizophrenia and bipolar disorder treatments is very often severely compromised by publication bias and outcome reporting bias. He also covers a study that finds that no models, including clinical judgment or machine learning, are able to successfully predict suicide risk. Finally, he covers a paper that addresses the lack of guidance available for discontinuing psychiatric drugs.

• Misreporting Results and Publication Bias Common in Psychiatry Research
• Mental Health Professionals and AI Tools Fail to Predict Suicide
• Lack of Guidance Available for Discontinuing Psychiatric Drugs

Visit madinamerica.com for more news, articles and research updates.