MediScript

Problem: Healthcare professionals struggle to accurately document patient consultations due to traditional note-taking methods being time-consuming, error-prone, and requiring significant effort. This can result in incomplete or inconsistent records, negatively affecting the quality of care provided. Target Audience: Health institutions and healthcare professionals. Market Size: According to recent market research, the global medical transcription market is projected to reach $3.7 billion by 2029. This demonstrates the vast opportunities that lie ahead for our AI transcription tool. Solution: MediScript is a web-based platform that records conversations between physicians and patients, and provides a summary of symptoms, diagnosis, and treatment. MediScript consists of two main steps. The first step is recording and converting patient-doctor conversations to text using Speech Recognition AI. For that, we used OpenAI’s Whisper API. Second, feeding the transcription along with the patient's information to a large language model to summarize the symptoms and suggest a diagnosis and treatment. We used Google’s Bard, a conversational generative AI. We feed Bard with the patient information including chronic diseases and allergies along with the conversation. Bard extracts symptoms and suggests a diagnosis and its treatment. Benefits: Reduces time and effort of patient visits. Improves transcription accuracy. Provides standardized transcriptions. Better collaboration between healthcare professionals. Improves healthcare. Future Work: Integrating ICD-10-AM (6th Edition), a standardized classification system of all diagnoses and symptoms, to provide an accurate diagnosis. Using DrugBank, a comprehensive and up-to-date medication database, to ensure accurate prescribing by considering the latest pharmaceutical information and guidelines. Support multiple languages and dialects. Sending a report of the visit to the patient's phone and email.