Unbiased Analysis of Today's Healthcare Issues
Do patients prefer AI or physician responses to their questions. Based on a study in JAMA Internal Medicine by Ayers et al (2023), the answer is AI: Of the 195 questions and responses, evaluators preferred chatbot responses to physician responses in 78.6% (95% CI, 75.0%-81.8%) of the 585 evaluations… Chatbot responses were rated of significantly…
Victoria Krakovna of LessWrong provides some examples where artificial intelligence (AI) algorithms didn’t work out exactly as planned. In fact, she has put together a master list of these AI issues. Below I have listed some of these examples that are more (or less) related to health: Cancer. AI trained to classify skin lesions as…
From an interesting article in Scientific American: When we look back on the past 175 years, the manipulation of matter and energy stands out as a central domain of both scientific and technical advances. Techno-scientific innovations have sometimes delivered on their promises and sometimes not. Of the biggest advances, three really did change our lives—probably…
In the past I have written about the limited evidence that wellness program work. So we should see that employers are abandoning these programs or at least scaling back? According to one recent study, however employers are actually increasing investments in wellness programs. The study surveys employers from over 500 companies with more than 100…
Interesting article on how human intelligence and artificial intelligence can be seen as complements rather than substitutes. Human+AI combinations have been shown to perform better than human or AI alone. Why? AIs are best at choosing answers. Humans are best at choosing questions. The advent of AI is not a zero sum game. The symbiotic…
Interesting video from Wired on how artificial intelligence (AI) software may help improve patient care in the future.
The site 99% Invisible has an interesting post on the invention of the stethoscope had how it transformed medicine. René Laennec actually felt that patient’s accounts of their own disease were still important, but the quest for objective information about disease was underway, and the stethoscope was just the beginning. Now we have X-rays, CT…
Should health insurers cover orphan drugs? Although the clear answer appears to be yes, the issue is tricky. An orphan drug is one which treats a limited number of people. In Europe, this designation generally applies to patients with a disease with an overall prevalence between 5 and 10,000 individuals. In order to incentivize innovators…
An article by Ajay Agrawal and Avi Goldfarb (“The Simple Economics of Machine Intelligence“) provide an interesting perspective on how machine learning will affect employment with a nice example from the health care sector. All human activities can be described by five high-level components: data, prediction, judgment, action, and outcomes. For example, a visit to the…
Many people have feared that robots, computers, and other forms of technology will take people’s jobs. Conventional wisdom holds that healthcare workers have been largely immune from these technological changes to date. However, a new paper from Lu, Rui and Seidmann examines whether technological advances–specifically computerized provider order entry (CPOE)–affects employment at nursing homes. Using…