Brief Report

The impact of prompt engineering in large language model performance: a psychiatric example