AI Outperforms Physicians in Managing Depression, Study Finds
Artificial intelligence (AI) has shown promising results in managing depression, outperforming primary care physicians (PCPs) in its ability to provide appropriate treatment recommendations without bias, according to a study conducted in Israel. The researchers emphasized that AI should complement human clinical judgement rather than replace it.
The study, published in Family Medicine and Community Health, aimed to determine the effectiveness of AI in addressing critical mental health challenges. Researchers Inbar Levkovich and Zohar Elyoseph compared the performance of AI system ChatGPT with recommendations made by 1,249 human PCPs in evaluating depressive episodes and suggesting treatments.
The findings revealed that ChatGPT provided more accurate suggestions for managing mild and severe depression, adhering to current clinical guidelines without displaying any biases based on gender or socioeconomic status (SES). In scenarios involving mild depression, PCPs often prescribed pharmacological treatments or a combination of medication and psychotherapy. In contrast, ChatGPT recommended psychotherapy in accordance with guidelines that state two weeks of psychotherapy is usually sufficient for mild depression, with the consideration of antidepressants only if there is no improvement or the symptoms are moderate to severe.
Regarding severe depression, ChatGPT consistently recommended the exclusive use of antidepressants. In contrast, PCPs often prescribed a mixture of antidepressants and anxiolytics/hypnotics. The researchers pointed out that ChatGPT’s therapeutic recommendations align with accepted guidelines for managing severe depression.
Furthermore, the study highlighted the absence of biases in ChatGPT’s recommendations based on gender or SES. Previous research has shown that PCPs can exhibit biases, being less likely to prescribe antidepressants to women compared to men. Levkovich and Elyoseph consider this finding to be of great potential, as ChatGPT’s ability to provide unbiased treatment recommendations can promote both quality and equity in mental healthcare.
However, the researchers stressed that ChatGPT should not replace human clinical judgement in diagnosing or treating depression. Rather, it should serve as a supportive tool that strengthens healthcare professionals in making well-informed clinical decisions.
The results of this study emphasize the potential of AI in addressing critical mental health challenges. By adhering to clinical guidelines and avoiding biases, AI systems like ChatGPT can provide accurate and equitable treatment recommendations for individuals struggling with depression. While AI shows promise in this field, human clinical judgement remains essential in providing comprehensive and personalized care to patients.
