“Medical algorithms are less transparent, far more complex, more likely to reflect preexisting human bias, and more apt to evolve (and fail) over time than medical devices in the past…. ‘Do no harm’ must apply— even to computer algorithms”.
– SciAm January 2022
Medical algorithms are not credentialed, but physicians are.
Physicians are credentialed by State Medical Boards of Registration, hospitals, and several federal and state regulatory agencies. The computer programming of medical AI (Artificial Intelligence, or “machine learning”) is often touted as able to make diagnoses and prognosis like a physician. Medical algorithms (the AI programming) are NOT approved (“credentialed”) by the FDA despite its authority to approve “medical devices” and to require clinical trials of new treatments. One such algorithm is the CDS (Clinical Decision Support) system used by 100 health systems to help detect early signs of sepsis and spur early referral of patients to complex care sites. That one CDS algorithm proved to actually discourage referrals of millions (yes, millions) of Black patients to medically appropriate care because of unrecognized bias in the traditional data base upon which it was built.
– SciAm January 2022
Speaking of sepsis in Blacks.
An economist study has proposed that the arrival of severe malaria in our South in the 1680’s caused the farmers to shift from using indentured servants from Europe susceptible to malaria to importing African slaves who were already immune to malaria because of their inherited sickle cell blood which reduces the effects of malaria . Records show that slaves from areas of Africa where malaria was more endemic were worth more on the auction bloc. A tag line of the article could be, “Did infection cause the civil war?”
– American Economic Journal, Esposito, E.
Elephants are teaching roboticist how to program AI.
An elephant’s trunk tip is a versatile, precision “picking up” tool which roboticist have been studying in order to write computer programs for their robot’s hands. It turns out that the elephant controls its complex boneless trunk’s movements with only 6 simple maneuvers. All the roboticist have to do is program their robot to mimic the elephants use of the 40,000 muscles in its trunk!
– Sci Am December 2021.
A new use of mRNA (a medical AI – machine learning – story)
This is NOT about vaccines!
A study of 1800 pregnant women measuring the amount and types of mRNA circulating in their blood resulted in more accurate predictions of the fetal gestational age than ultrasound. Also certain combinations of mRNA analyzed by AI actually predicted the onset of preeclampsia in the mother at 14 weeks of gestation more accurately than traditional predictors; a real benefit to early treatment to avoid potential complications for both mother and fetus. Some circulating fetal mRNAs in the mother’s blood were also associated with specific organs’ development in the fetus and could be used to monitor previously undetected fetal abnormalities.
– Nature January 2022
A breathalyzer for diseases (more medical AI)
The human breath contains over 800 chemical compounds. Some can be detected by our (or a dog’s) sense of smell; like acetone in an out-of-control diabetic’s breath, nitric oxide elevation in inflammatory lung disease, or increased aldehydes related to lung cancer. In March 2020 researchers from the Israeli Institute of Technology were able to detect COVID-19 in exhaled breath in 100% of COVID-test positive people with 95% accuracy in a study in Wuhan, China. In 2021 NASA spent almost $4 million repurposing its “E-nose” that scans the air for potentially dangerous chemicals in the International Space Station to detect COVID.
– Sci Am December 2021
More home-based medicine
A home-based, patient-administered, 10-15 minute cognitive test called SAGE (Self-Administered Gerocognitive Examination) using 11 items to detect early dementia is as reliable and as accurate as longer tests in clinicians’ offices. [If use of SAGE is added to the advances of telemedicine, at-home COVID testing, and the lowered colonoscopy and PSA age-limit recommendations, us old men may never have to go to a doctor’s office again.]
– Alzheimer’s Res. Ther. December 2021
Speaking of Telemedicine
The founder of a health care technology platform (Firefly Health) trying to innovate one aspect of telemedicine, internet prescriptions, asks the question: “Why do you have to visit your doctor to get a prescription?” He proposes the scrapping of a law regulating Internet prescriptions passed in the 2008 wake of the death of an 18 year old from an overdose of Vicodin obtained from an internet pharmacy. According to Jonathan Bush, founder of Firefly Health and the founder of Athenahealth (a “revenue boosting office computer system” for physicians’ offices), the language in the law requiring a medical evaluation “with the patient in the physical presence of the practitioner” is outdated and a “real barrier to the growth of digital health care. . . Many of the pill mills in Florida that spawned this legislation were largely in-person clinics, not telemedicine. . . . Online providers could safely use computer programs now used to prevent money laundering (Know Your Customer)”. [Using anti-money laundering techniques to safeguard telemedicine prescriptions suggests that “the businessfication of medicine” as predicted by NEJM Editor Dr, Arnold Relman way back in the 1980’s has really come full circle.]
– BosGlobe December 12, 2021