AI Chatbot Outperforms GPs in Depression Treatment, Eliminating Gender and Socio-Economic Biases
A recent study published in the British Medical Journal reveals that an AI chatbot called ChatGPT may be more effective than general practitioners (GPs) in treating depression, particularly due to its ability to eliminate biases related to gender and socio-economic factors. Researchers found that ChatGPT does not differentiate between genders or economic statuses when determining suitable treatment plans for patients with mild and severe depression. Additionally, the chatbot demonstrates a greater adherence to clinical guidance compared to GPs.
The study compared treatment plans recommended by 1,249 GPs in France with those provided by ChatGPT for patients manifesting depressive symptoms like sadness, sleep problems, and loss of appetite. The research found that there was no significant difference in the treatment approaches suggested by the AI system for men and women, as well as white collar and blue collar workers.
Conversely, GPs showed a tendency to prescribe fewer antidepressants to women compared to men. Moreover, they commonly prescribed antidepressants without therapy to blue collar workers, while white collar workers received a combination of both. In contrast, the latest version of the AI bot, ChatGPT-4, closely aligned with clinical guidance by recommending therapy to 97.5% of hypothetical patients with mild depression, surpassing the 4.3% referral rate by GPs.
Notably, GPs often favored a drugs-focused approach, choosing medications as the principal therapeutic strategy for approximately 50% of cases. They also frequently prescribed medication combined with therapy for a third of patients. Conversely, when it came to severe depression cases, ChatGPT consistently recommended therapy and antidepressant medication for every patient, while GPs followed this approach for less than half of their cases (44.4%) and opted for medication alone in 40% of instances.
Researchers from Oranim Academic College in Israel and Imperial College London highlighted that GPs predominantly relied on antidepressants, even for cases of mild to moderate depression, indicating a pharmacological approach. The study also revealed that GPs were more inclined to prescribe anti-anxiety and sedative-hypnotic drugs, such as benzodiazepines, despite the addictive nature and recommendation against their use as a first-line treatment. In fact, these drugs were regularly prescribed alongside antidepressants or on their own in one out of eight cases. ChatGPT, on the other hand, did not exclusively prescribe these addictive drugs but occasionally recommended them alongside antidepressants for 30% of patients prescribed medication.
The study acknowledged the inherent potential of AI chatbots in enhancing clinical decision-making by adhering to accepted treatment guidelines and eliminating gender and socio-economic biases present among doctors. However, the researchers also emphasized the need for further research to refine AI recommendations for severe cases and consider potential risks and ethical concerns.
According to a survey conducted by the British Standards Industry, approximately half of the population believes that AI has the capacity to assist in tackling the record-high NHS waiting lists, which currently stand at 7.75 million. Respondents expressed optimism about AI’s potential to address staffing shortages in the healthcare sector and improve diagnostic accuracy.
In conclusion, the study’s findings suggest that ChatGPT, the AI chatbot, possesses qualities that make it a potential superior alternative to GPs for treating depression. By eliminating biases related to gender and socio-economic factors, and by diligently adhering to clinical guidance, ChatGPT demonstrates promise in enhancing patient care. Nonetheless, further research and consideration of potential risks and ethical implications are necessary to capitalize fully on the capabilities of AI chatbots like ChatGPT.