Comparison of indications for hysterectomy in our clinic with recommendations of the artificial intelligence program

Objective: This study aims to compare and analyze the medical data and clinical indications of patients who have been decided for hysterectomy in our clinic with the recommendations of the artificial intelligence program ChatGPT. The recommendations generated by ChatGPT, based on scientific articles and studies, are used to assess its potential and limitations in clinical decision-making processes. The research investigates how effective and reliable AI-based systems can be in planning major surgical interventions such as hysterectomy. Furthermore, the study examines how AI technology can be integrated into current medical practices and its extent of contribution to clinical decision-making processes. In this context, the study aims to provide significant insights into the future use of AI in the medical field. Materials and Methods: The study was conducted on a total of 87 patients aged between 40-65 years, who applied to the Akdeniz University Department of Obstetrics and Gynecology and were decided for hysterectomy between June 1, 2023, and November 1, 2023. The detailed anamneses and information of the patients included in the study were systematically collected and evaluated by the research staff of our clinic. The collected data were entered into the AI program ChatGPT, aiming to determine the most effective treatment option suitable for each patient's scenario. During this process, the program was requested to interpret the medical condition of each patient, considering the current literature, and to explain the reasons for its recommendations. This methodology was designed to assess the contribution of AI to clinical decision-making processes and to examine its potential effectiveness. Results: The data of 87 patients who presented to the Akdeniz University Department of Obstetrics and Gynecology and were decided for hysterectomy between June 1, 2023, and November 1, 2023, were analyzed in this study. The average age of the patients was found to be 48.71, with the most common complaints being irregular menstruation (31.00%) and heavy menstrual bleeding (HMB) (27.60%). The most frequent additional disease histories were obesity (13.80%) and hypertension (HT) (14.90%). A low proportion of patients had a history of breast cancer and Tamoxifen usage (5.70% and 4.60% respectively). When comparing the treatment options recommended by the AI program ChatGPT with the decisions of clinicians, it was observed that the program recommended hysterectomy in 70.10% of the cases. Other recommended treatments included myomectomy (10.30%), hysteroscopy (8.00%), and medical treatment (4.60%). These recommendations correlated with the clinical and demographic characteristics of the patients. Notably, the AI's recommendations for hysterectomy were highly consistent with situations involving abnormal uterine bleeding and the presence of myomas. Conclusions: The alignment between ChatGPT's recommendations and clinical decisions demonstrates the potential of AI in medical decision-making processes. However, the presence of some differences between the recommendations of AI and actual clinical practices highlights that AI should not yet be used as an independent decision-making tool and should continue to be employed as a supportive technology in medical applications. Furthermore, the study brings to light the potential and limitations of AI in medical decision-making processes. Despite the recommendations of AI programs being based on medical data and current literature, the importance of clinical experience and evaluating the individual condition of the patient is emphasized. In conclusion, this study demonstrates that AI-based systems can be an effective support tool in clinical decision-making processes. However, the use of such systems should be to assist and complement the clinical decisions of physicians. These findings provide a foundation for better understanding the future role and integration of AI in the medical field. Key words: Artificial Intelligence, Abnormal Uterine Bleeding, Hysterectomy, gynecology