Study Shows Ambient AI Tools Could Enhance Clinical Consultations

A new study led by AI and healthcare researchers in the UK has found that the use of ambient AI devices could improve the quality of medical appointments and produce high-standard clinical documentation.

The study published in the Future Healthcare Journal suggests that ambient AI, which focuses on creating systems capable of solving problems in real-time and at scale, could shorten medical procedures, reducing workloads and potentially enhancing patients’ clinical experiences.

EHRs and Physician Burnout: A Challenge in Modern Medicine

The implementation of electronic health records (EHRs), which are computerized compilations of patient’s medical records, has led to several advantages in modern medicine, including enhanced clinical documentation.

However, these increasing documentation efforts have been associated with mental health complications in healthcare professionals, contributing to rising cognitive load and physician fatigue.

According to data from the American Medical Association, 48.2% of physicians in the U.S. reported experiencing exhaustion in 2023.

Additionally, research investigating physician burnout in consultant doctors in Ireland revealed that nearly half of clinicians participating in the survey (42%) suffered from high levels of work-related stress. The study also found that emotional exhaustion may lead doctors to experience more serious conditions such as anxiety and depression.

In this context, there are significant concerns that the deteriorating mental health of physicians could compromise the current standard of patient care, ultimately diminishing the overall effectiveness of healthcare delivery.

Study Methodology and Results

The study was conducted by researchers from Tortus AI, a platform designed to improve clinical operations through integration with Electronic Health Records, in collaboration with Great Ormond Street Hospital for Children (GOSH) in London, UK.

The objective of the study was to evaluate the effectiveness of an end-to-end ambient AI tool in enhancing the quality of notes generated during clinical consultations, comparing it against standard EHR procedures. Researchers also aimed to determine the device’s capacity to improve the quality of those consultations while reducing their duration.

To do this, the study utilized the AI tool to ambiently listen and summarize audio from 47 clinical consultations into clinic notes by using speech-to-text and Large Language Models.

The AI tool recorded the speech, redacted Personal Health Information (PHI), transferred the recording to a local cloud provider for transcription and then reinstated the PHI to generate clinic notes.

Professional medical actors simulated patient interactions alongside real clinicians from GOSH, allowing for a direct comparison of consultations with and without the AI tool in a simulated clinic environment.

Results demonstrated that the AI tool significantly enhanced documentation quality and operational productivity. Additionally, doctors acknowledged the device’s transcription efficiency, user-friendliness, accessibility and simplicity in note-taking as well as its potential to ease the cognitive burden on healthcare professionals.

“AI-produced documentation achieved higher SAIL scores, with consultations 26.3% shorter on average, without impacting patient interaction time. Clinicians reported an enhanced experience and reduced task load”.

Researchers led by Dr. Shankar Sridharan, Consultant Fetal and Paediatric Cardiologist, Cardiac Network Lead for GOSH and visiting Consultant Paediatric Cardiologist for Whipps Cross Hospital and The Royal London Hospital, Whitechapel.

AI and Medicine: A Promising Yet Uncertain Future

The study’s results emphasize that AI has the potential to improve clinical consultations and documentation, assisting clinicians in advancing healthcare practices not only in this field but also in areas including cancer treatment, protein stability research and vaccine production, among others.

However, the application of AI in modern medicine is rising certain concerns, including patient data privacy, device clinical validation and inherent biases. These challenges are currently hindering AI from integrating reliably into contemporary healthcare settings.

Therefore, regulating the medical use of these technologies and addressing issues involving inaccuracy and biases will be essential for AI to effectively settle in contemporary medicine. This would allow clinicians to safely incorporate AI as a tool to complement and improve their work while safeguarding patient needs, potentially enhancing modern healthcare practices.