How AI Could Cure Healthcare if it isn’t Blocked at Birth

The Miracle Drug

Two in five Americans and over a quarter of British are obese. Glucagon-like peptide-1 (GLP-1) agonists, such as Ozempic and Zepbound, have been effective in reducing weight. Studies also claim these drugs lower blood pressure, reduce the risk of heart and kidney disease, and improve fatty liver disease. They don’t.

Casey Means is the nominee for US Surgeon General, a role known as the nation’s doctor. She dropped out of her ninth and final year of residency aged 31 because she says the healthcare system lacks the incentives, time and training to solve problems. Means believes that a faulty metabolic system triggered by poor diet and lifestyle is the primary cause of most chronic disease. She wants to redesign the US healthcare system around prevention rather than a lifetime of medication.

Reforming healthcare may seem a strange topic for a newsletter with roots in helping small businesses. Yet an evolving theme of The Profit Elevator is how AI will reshape industries creating opportunities for fleet-of-foot companies. While redesigning industries is the work of established companies, or well-funded disruptors, the impact is felt by all the businesses in a supply chain.

As health is the number one use case for AI-powered research, healthcare is a starting point for thinking about how this change may unfold.

Root Cause Analysis  

Means describes treating a woman of her age with eleven different chronic conditions. The brief was to conduct a chargeable physical examination, schedule additional tests, prescribe a medication, and ensure there was no immediate threat to life. All within 15 minutes. The patient was on more than 11 medications with almost no coordination between her numerous physicians.

The mitochondria in cells are our body’s battery. When a group of cells don’t function properly we call this disease. In the brain this might be depression, Alzheimer’s or a stroke. Elsewhere it is heart disease, kidney failure, or reproductive issues. A faulty battery is most often caused by irregular blood sugar, which is a result of diet and lifestyle.

Means determined that reducing weight would reduce all instances of disease in her patient. This is a theory requiring research. Artificial intelligence is capable of spotting patterns in complex data systems that it would take humans too long to notice. The right research programme could decide how many chronic conditions are traceable to poor food and limited exercise. GLP-1 agonists, which address weight but not disease, may not be the best treatment.

This work would not take anybody’s job. The analysis isn’t being done. That does not mean, however, that doing it would uproot the healthcare industry.

The Diffusion Dilemma

The provision of healthcare is deeply embedded throughout the political economy. Pharmaceutical companies are the dominant advertisers on television, sponsors of medical research, and major political lobbyists. A preventative health system that stopped people needing drugs would collapse the profits of big pharma, university hospitals, media and lobbyists. It would uproot the food industry from subsidised cereal farming to distribution of cheap produce in supermarkets. A drug company or food retailer is the largest employer in 39 of 50 US states.

Kapoor and Narayan, whom we met in What AI Can and Cannot Do, argue that AI is just like any other technology. Its impact is determined by the speed with which it diffuses through an economy. Healthcare has the vast datasets best suited to AI-powered research. Yet the speed with which this will change the economy depends on human factors such as money, regulations and political will.

Technology companies have profits to protect as much as drug companies do. They will supply AI hardware and services to health researchers provided they continue to be funded. This will not absorb all the AI capacity being built however, meaning a continued push to find analysis-heavy workloads to perform in other industries.

To date, finance has been a popular area for exploration, but it is highly regulated. Hedge funds and trading firms adopt the latest technology in search of a hair’s breadth advantage over the competition. The mass market for retail products, in contrast, is governed by regulations on transparency and working in the best interests of clients. AI might help retail investment firms run with greater efficiency, but using a black box to make investment decisions could be years away. Again, it is not the capability of AI that constrains its adoption. It is the speed with which humans allow it be adopted.

Casey Means is 37 years old. Her decision to not qualify as an Ear, Nose and Throat surgeon means she lacks board certifications. Her medical licence has lapsed, meaning she has limited patient-care experience. The food and pharma lobby may therefore be able to block her appointment. Diffusion of AI-powered, holistic healthcare research throughout the economy would fall at the first hurdle. In 10 years’ time your AI doctor may be handing you the same disease-specific treatment as today.

Questions to Ask and Answer

1. Which parts of my supply chain are free from regulatory bottlenecks?

2. What change would be most beneficial in those areas?

3. Would my business gain or lose from this change?

Here are 3 ways I can help

1. Book a consultation to talk about AI.

2. Explore how to manage an AI project.

3. Explore use cases using NVIDIA accelerated computing.