The Startups Applying AI To Transform Healthcare

In the U.S. economy, health care is perhaps the most significant sector. At the same time, health care is the U.S. economy's most fractured sector.

It is the largest: the United States spends nearly $ 4 trillion per year on healthcare. Recommended Carbon Emissions Problem For You Deep Learning Here Is How The U.S. Should Regulate Artificial Intelligence Deepfakes Go To Wreak Havoc On Society.

A constellation of data points — recent physical symptoms, blood pressure, genetic makeup, chemical composition of the bloodstream, and so on — can be collected that tells the definitive storey of a person’s health, taken together and compared against population-level patterns.


Using computer vision to identify health conditions in medical images has become perhaps the most widely referenced health-care use case for AI. It is easy to understand why: examining a medical scan to determine whether there is a tumour, a skin lesion, a retinal disease or some other indication is a clear example of classification of objects, exactly what deep learning excels.

As AI legend Geoff Hinton famously declared in 2016, "Radiologists should stop training now. It’s just completely obvious that within 5 years, deep learning is going to do better than radiologists.”

Patient Intake and Engagement
Another area where AI will improve care delivery is patient intake and involvement, a critical part of the journey through healthcare.

Recent advances in the processing of natural language have made AI-based conversational interfaces possible which can automate patient screening and navigation care. Patients, for example, can share symptoms and questions through a text message and receive automated clinical guidance in response. Similarly, AIs that communicate with patients on an ongoing basis can be developed to ensure they remain engaged and compliant with their treatment regimen.

Remote Health
Eko has built a platform of proprietary sensors and machine learning algorithms which can monitor the cardiopulmonary vital signs of patients remotely to detect heart and lung problems early. Eko 's AI detects heart problems significantly more accurately than do human physicists using a stethoscope. For example , general practitioners detect atrial fibrillation with accuracy of 70-80 percent, while Eko 's algorithms do so with accuracy of 99 per cent.

“We can increase the doctor’s judgement about cardiac and pulmonary diagnoses with data analysed in seconds from tens of thousands of past patient exams,” said Connor Landgraf, CEO of Eko.

In-Hospital Care
Gauss Surgical uses computer vision for monitoring blood loss during childbirth as one example. Visual estimation of blood loss by human clinicians is notoriously imprecise, and haemorrhage is the leading preventable cause of maternal mortality. Gauss’ AI solution at one hospital system led to a 4x increase in the recognition of hemmorrhage and a 34 percent reduction in delayed bleeding interventions.


AI can play a key role here. As anyone who has dealt with the healthcare system knows, it is plagued by waste and inefficiency. The application of AI to the administrative side of healthcare may seem unglamorous compared to clinical or life sciences use cases.

But there is a huge opportunity here for value creation. Revenue Cycle Management (RCM) is one administrative function which is particularly challenging and important for health care providers. The application of machine learning to automate many of these rotary tasks is a promising set of companies. Read From The Source