This study explores the potential of Large Language Models (LLMs) to enhance the accuracy of Automatic Speech Recognition (ASR) systems in medical transcription. Using the PriMock57 dataset, which includes 57 mock consultations, the research evaluates the effectiveness of LLMs in improving Word Error Rate (WER), Medical Concept WER (MC-WER), and speaker diarization accuracy. The study compares zero-shot and Chain-of-Thought (CoT) prompting techniques to enhance ASR outputs. Results show that LLMs, particularly through CoT prompting, significantly improve diarization accuracy and MC-WER, achieving state-of-the-art performance. LLMs also enhance semantic textual similarity, preserving the contextual integrity of clinical dialogues. The study highlights the dual role of LLMs in augmenting ASR outputs and independently excelling in transcription tasks. The findings suggest that LLMs can transform medical ASR systems, leading to more accurate and reliable patient records. The research also demonstrates that LLMs can be effectively used in low-resource settings, combining affordable audio equipment with LLM post-correction via an API. The study's results indicate that LLMs can significantly improve the accuracy of medical transcriptions, particularly in capturing medical concepts and enhancing semantic coherence. The findings underscore the importance of LLMs in improving the accuracy and reliability of medical transcription systems.This study explores the potential of Large Language Models (LLMs) to enhance the accuracy of Automatic Speech Recognition (ASR) systems in medical transcription. Using the PriMock57 dataset, which includes 57 mock consultations, the research evaluates the effectiveness of LLMs in improving Word Error Rate (WER), Medical Concept WER (MC-WER), and speaker diarization accuracy. The study compares zero-shot and Chain-of-Thought (CoT) prompting techniques to enhance ASR outputs. Results show that LLMs, particularly through CoT prompting, significantly improve diarization accuracy and MC-WER, achieving state-of-the-art performance. LLMs also enhance semantic textual similarity, preserving the contextual integrity of clinical dialogues. The study highlights the dual role of LLMs in augmenting ASR outputs and independently excelling in transcription tasks. The findings suggest that LLMs can transform medical ASR systems, leading to more accurate and reliable patient records. The research also demonstrates that LLMs can be effectively used in low-resource settings, combining affordable audio equipment with LLM post-correction via an API. The study's results indicate that LLMs can significantly improve the accuracy of medical transcriptions, particularly in capturing medical concepts and enhancing semantic coherence. The findings underscore the importance of LLMs in improving the accuracy and reliability of medical transcription systems.