In this issue of JAMA Psychiatry, Olfson et al1 report a very comprehensive study of the use of benzodiazepines in the United States. This study confirmed what has been found in several other countries, that benzodiazepines are used predominantly in elderly persons, mostly women, and for long periods of time.2,3 The older the patient, the longer the drug is used. This would not be a major issue if benzodiazepines were truly useful (preferably life-saving) and reasonably risk-free. They have saved many lives by being used to attempt suicide instead of barbiturates that are much more toxic4; however, for their main indications of insomnia and anxiety, benzodiazepines fare little better than placebos after a few weeks of treatment. After an initial improvement, the effect wears off and tends to disappear. At that point, what happens when patients try to stop taking benzodiazepines is that they experience withdrawal insomnia and anxiety. The usual conclusion is “you see, they work. When I stop them, I get worse.” Initially, patients get better before returning to the pretreatment state and then get worse than before treatment began when they attempt to stop taking benzodiazepines. After a few weeks of treatment, patients are actually worse off than before they started (or at least not better) and cannot stop taking the drug.