Experts predict that the number of patients suffering from geriatric conditions and diseases will spike in the not-too-distant future, courtesy of an aging population. Unfortunately, the AAMC says that the U.S. could see a shortage of roughly 120,000 doctors by the year 2030. As the number of patients with complex medical needs grows, coupled with a shrinking number of doctors, artificial intelligence may be stepping in to fill in the gaps. This isn't just good news for older people, however—there are a lot of reasons why the shift toward automation in the medical industry might be a good thing. As with any new technology, medical algorithms come with pros and cons.
Healthcare is expensive, and, if anything, the cost is increasing. People of limited monetary means often also have limited access to care—those who rely on Medicaid, for example, are often very limited both by who is able to accept their insurance, and what specialists their state is willing to pay for. Using AI for diagnosis and routine monitoring provides a much less expensive alternative to a visit with a doctor, which can expand access to medical care for lower income patients.
While AI can improve access to medical care, there are questions about the quality of the patient experience. There is concern that the widespread implementation of AI-based care might increase the disparity between the rich and the poor, running the risk of shunting people in lower income brackets off to automated kiosks for the bulk of their care.
Jokes about doctors' handwriting abound, but there's a grain of truth to them—human errors happen. Whether they're from messy writing, confused transcription, poor communication, or missing a diagnostic clue here or there, the Institute of Medicine estimates that they end up costing between $17 billion and $29 billion per year. This doesn't even touch on the effect on patients, who end up suffering both physically and mentally from these errors. Artificial intelligence can make communication between patients, doctors, and pharmacies smoother and more efficient, improving patient care outcomes and reducing medical errors.
There is an oft-overlooked psychological aspect to seeking medical care. Patients visit the doctor because they are vulnerable—they are either acutely sick, injured, or managing a chronic health condition, all of which can be very intimidating to face alone. Many doctors are concerned about the lack of a “human touch” in an algorithm-led experience, fearing that it will deprive patients of the feeling that someone cares about them.
It takes an enormous amount of time and effort to analyze patient data. A busy hospital can have thousands of diagnostic images to interpret every week, which requires a tremendous amount of manpower. Artificial intelligence can help reduce this by automating the process, without a loss of accuracy. A study involving an algorithm used to interpret images of the retina found that it was able to diagnose eye diseases with over 94% accuracy—roughly the same as experts in the field. This means that patients can get accurate results and recommendations quickly.
Algorithms are only as good as their programming. It's often said that garbage data in equals garbage data out, and it's doubly true here. For example, an algorithm designed to spot early-stage melanoma is only as effective as the population it is tested on. If it is refined primarily using patients with light skin, it may not be effective when used to diagnose patients with dark skin. Diseases can also present differently in various populations, like cardiac symptoms in women versus men. This makes developing and implementing effective algorithms a challenge, particularly in mixed populations of patients.
Artificial intelligence presents a way for overworked doctors to improve their efficiency and accuracy. While there is still some fine-tuning needed when it comes to the patient experience, and legitimate fears that it may leave some communities vulnerable to substandard care, it can dramatically improve patient outcomes in other areas by increasing access to care, reducing errors, and providing quick, accurate analysis and recommendations.
