Home Healthcare CloudMedx AI Beats Human Doctors in a US Medical Exam

CloudMedx AI Beats Human Doctors in a US Medical Exam

Recently, CloudMedx Inc. revealed that the company’s AI took a tailored version of the US Medical Licensing Examination Step 1 some few weeks back whereby it beat human physicians in a simulated study.

Outperforming human doctors in the study marks a historic milestone, which shows that artificial intelligence can deliver insights that can help physicians as well as understand medical narratives and concepts.

Furthermore, it also showed that when human intelligence is combined with machine intelligence, the combined improved approach might produce improved outcomes for activities that would be computationally-difficult and time-consuming for human beings.

The simulated study shared a lot of similarities with USMLE, an examination that doctors take in a bid to obtain a license in the United States.

SEE MORE: Google’s DeepMind Beats Doctors at Detecting 50 Eye Diseases Just by Looking at Scans

The tailored examination consisted of questions such as case studies whereby a given instance of a patient is illustrated, and the party sitting for the exam, both the AI and humans, were required to apply medical reasoning and knowledge in answering multiple-choice questions.

The exam avoided simple facts that might be easily answered using a keyboard search, and instead, explained sophisticated scenarios.

For artificial intelligence (AI) to leverage this type of data analysis and produce insights is exceptional.

The intention of this study and CloudMedx is not providing any diagnosis, but rather offering tools for going through massive volumes of both unstructured and structured data drawn from health system records in a bid to produce predictive analytics that might be meaningful for clinicians or even their care teams.

CloudMedx AI takes the Modified Medical Exam

CloudMedx has created a clinical AI assistant with the potential of augmenting health systems’ revenue cycle management and clinical operations.

The system utilizes deep learning and natural language understanding (NLU).

It can also be embedded into Electronic Health Records for providing clinical insights, particularly within workflows in a bid to help hospital staff in improving documentation and operations, at per-patient level and population level.

Currently, CloudMedx is collaborating with some of the leading medical facilities in the US.

For benchmarking purposes, CloudMedx carried out a small-scale study by inviting several doctors to take the modified medical exam against the company’s medical resident AI.

SEE MORE: Hot AI Healthcare Startup Babylon Unveils a Chatbot that Beats GP at Medical Exams

SEE MORE: 10 Amazing Examples Of Natural Language Processing

The group was made up of physicians from one of the leading US-based medical universities.

The exam included 100 questions, whereby each was presented in the form of a case study featuring multiple choice answers.

For instance, a question may involve a situation whereby a patient explains his symptoms alongside past lab results, medications, and medical history.

The AI and the doctors involved were expected to use their medical experience, domain expertise, and medical knowledge to answer the questions.

The doctors were divided into three groups, and the task involved taking the examination to determine which group of doctors performed in the study.

Their scores were as follows:

  • Group 1. Human doctors: score 68-81(an average score of 75)
  • Group 2. CloudMedx AI: 85
  • Group3. Human physicians alongside CloudMedx AI: 91

CloudMedx AI performed better than all the human doctors involved in the study by attaining a score of 85%.

On the other hand, the score for humans was 68% and 81%, the lowest score and the highest score respectively.

However, the group with humans and AI outperformed the other two groups by attaining 91%.

In group 3, the AI suggested the most suitable answers for each question and the doctors either rejected or accepted the recommendation.

The group relied on combining their expertise and memory together with the suggestions from artificial intelligence (AI) in a bid to attain an overall higher score in comparison to the rest of the groups.

Human doctors took an average of 73 minutes to complete the exam, whereas the AI was done in less than 5 minutes.

READ MORE – Top 10 Ways Artificial Intelligence is Impacting Healthcare

The preliminary results of the study denote that artificial intelligence might augment human physicians, allowing their combined medical intelligence to bring efficiencies as well as boost the entire process.

When interviewing the last group consisting of the AI and human doctors, it was clear that AI helped in answering questions that left the human doctors on the fence.

CloudMedx is convinced that this will be the future of the healthcare sector- a machine-driven medicine that is assistive in nature as opposed to disruptive.

The lesson is that when both machines and humans work together, they can attain higher care delivery and efficiency.

Machine and human augmented intelligence working in unison

According to Cleveland Clinic Foundation’s Associate Chief Information Officer Dr. William Morris, a given CloudMedx partner is trying to boost clinical documentation and workflows.

“This is a very exciting technology that supports the vision of how Augmented Intelligence plays a pivotal role in the care of patients. While AI can master vast clinical knowledge, it is best when playing a supportive role to the clinician in their workflows and daily lives.”

This endeavor clearly indicates that machine and human augmented intelligence can offer a lot of reprieves, especially from the explosion of data as well as facilitate clinical documentation, operational efficiencies, and clinical workflows that cost care teams a lot of their time and effort.

For the past several years, there has been increased fear among people that AI will replace them in their workplaces.

While science fiction and movies reinforce this fear, the truth is that artificial intelligence ought to be considered as an augmented tool that can analyze massive data volumes and make sense out of it while allowing human beings to concentrate on creativity, critical decision making, and empathy.

The integrated approach involving machine and human augmentation can produce exceptional results, specifically for the healthcare space.

Alan Pitt, an MD working at the Barrow Neurological Institute, said: “AI should be viewed not as a threat by physicians, but as a supporting tool that helps them build efficiency in their workflows.

AI is often portrayed as technology marginalizing people. I see it differently. This is the first step showing another future for healthcare. It’s not man vs. machine but man and machine improving care.

The best performer was a doctor with AI. Moving forward providers will have a new tool to bring information to the bedside, I have hope AI will free providers to spend more time with patients by freeing them from their current role as scribes for the EMR.”

Given the appropriate technology and focus with respect to correct data sharing, security, and patient privacy, healthcare AI can be an excellent tool for both patients and doctors.

Consumers are currently seeking to gain access to their individual data, whereas doctors are asking for improved tools in their work to avoid being overburdened, and instead, boost care delivery.

Stefano Bini, an MD and Professor working at the University of California in San Francisco said:

“The premise that technology can be used to make humans smarter is not new. However, the fact that artificial intelligence can help doctors make better-informed decisions based on data than the smartest amongst them can make alone is pretty astounding.

This technology can make any quality disparity we see across geographic areas in our country essentially disappear. The cost savings alone would easily pay for the widespread deployment of the technology.”

CloudMedx is also partnering with other departments at the San Francisco-based University of California in an effort of automating workflows and improving clinical documentation in fields of orthopaedic surgery, liver cancer, and many others.

MORE – Data Science – 8 Powerful Applications

Subscribe to our newsletter

Signup today for free and be the first to get notified on the latest news and insights on artificial intelligence

KC Cheung
KC Cheung has over 18 years experience in the technology industry including media, payments, and software and has a keen interest in artificial intelligence, machine learning, deep learning, neural networks and its applications in business. Over the years he has worked with some of the leading technology companies, building and growing dynamic teams in a fast moving international environment.
- Advertisment -


AI Model Development isn’t the End; it’s the Beginning

AI model development isn’t the end; it’s the beginning. Like children, successful models need continuous nurturing and monitoring throughout their lifecycle. Parenting is exhilarating and, if...