Summary: While there is clear potential to use ChatGPT in a clinical setting, researchers say the AI algorithm may not yet be a reliable way of replacing the family doctor, especially when it comes to making effective decisions about prescribing antibiotics for infections.
Source: University of Liverpool
Researchers from the University of Liverpool have tested whether the AI-powered chatbot ChatGPT could be used to make decisions about prescribing patients with antibiotics.
In a letter published in The Lancet Infectious Diseases, academics from the Institute of Systems, Molecular and Integrative Biology show that, while artificial intelligence can’t yet replace the family doctor, there is clear potential for technology to play a role in clinical practice.
The researchers presented ChatGPT with eight hypothetical infection scenarios which people would commonly consult their doctor about (such as a chest infection). They then assessed the advice delivered by the technology for its appropriateness, consistency and its impact on patient safety.
The assessment found that ChatGPT understood the scenarios and provided coherent answers, including disclaimers, and signposting patients to sources of advice. It also appeared to understand the need to only prescribe antibiotics when there was evidence of bacterial infection.
Interestingly, the AI tended to focus on the type of antibiotic prescribed in each scenario rather than other factors, reflecting the assumptions often initially made by doctors during consultation.
However, ChatGPT provided unsafe advice in complex scenarios and where important information was not explicitly provided.
Interestingly, the AI tended to focus on the type of antibiotic prescribed in each scenario rather than other factors, reflecting the assumptions often initially made by doctors during consultation.
Following the experiment, the researchers have now developed a checklist for standards that AI should meet in order to be considered for use in clinical practice in the future.
Co-author of the letter, Dr. Alex Howard said, “It was fascinating to see the potential of artificial intelligence in health care demonstrated through this experiment testing ChatGPT’s ability to give antibiotic treatment advice.”
“With the rise of antibiotic resistance posing a significant threat to global health, the ability of AI to provide accurate and safe treatment advice could revolutionize the way we approach patient care. We look forward to further exploration of this technology and its implications for the future of health care.”
Summary: Apremilast, an FDA-approved drug for the treatment of skin conditions including psoriasis and psoriatic arthritis, triggers increased activity in the nucleus accumbens, a brain area associated with regulating alcohol intake. Apremilast reduced drinking behaviors in mouse models with a genetic risk of alcohol use disorder.
Source: Oregon Health and Science University
Researchers from Oregon Health & Science University and institutions across the country have identified a pill used to treat a common skin disease as an “incredibly promising” treatment for alcohol use disorder.
The study was recently published in the Journal of Clinical Investigation.
On average, the people who received the medication, called apremilast, reduced their alcohol intake by more than half — from five drinks per day to two.
“I’ve never seen anything like that before,” said co-senior author Angela Ozburn, Ph.D., associate professor of behavioral neuroscience in the OHSU School of Medicine and a research biologist with the Portland VA Health Care System.
The lead author is Kolter Grigsby, Ph.D., a postdoctoral fellow in the Ozburn laboratory at OHSU.
Beginning in 2015, Ozburn and collaborators searched a genetic database looking for compounds likely to counteract the expression of genes known to be linked to heavy alcohol use. Apremilast, an FDA-approved anti-inflammatory medication used to treat psoriasis and psoriatic arthritis, appeared to be a promising candidate.
They then tested it in two unique animal models that have a genetic of risk for excessive drinking, as well as in other strains of mice at laboratories across the country. In each case, apremilast reduced drinking among a variety of models predisposed to mild to heavy alcohol use. They found that apremilast triggered an increase in activity in the nucleus accumbens, the region of the brain involved in controlling alcohol intake.
Researchers at the Scripps Research Institute in La Jolla, California, then tested apremilast in people.
The Scripps team conducted a double-blind, placebo-controlled clinical proof-of-concept study involving 51 people who were assessed over 11 days of treatment.
“Apremilast’s large effect size on reducing drinking, combined with its good tolerability in our participants, suggests it is an excellent candidate for further evaluation as a novel treatment for people with alcohol use disorder,” said co-senior author Barbara Mason, Ph.D., Pearson Family professor in the Department of Molecular Medicine at Scripps.
The clinical study involved people with alcohol use disorder who weren’t seeking any form of treatment, and Mason predicts that apremilast may be even more effective among people who are motivated to reduce their alcohol consumption.
“It’s imperative for more clinical trials to be done on people seeking treatment,” Ozburn said. “In this study, we saw that apremilast worked in mice. It worked in different labs, and it worked in people. This is incredibly promising for treatment of addiction in general.”
An estimated 95,000 people in the United States die every year from alcohol-related deaths, according to the National Institute on Alcohol Abuse and Alcoholism.
An estimated 95,000 people in the United States die every year from alcohol-related deaths, according to the National Institute on Alcohol Abuse and Alcoholism.
Currently, there are three medications approved for alcohol use disorder in the United States: Antabuse, which produces an acute sensitivity akin to a hangover when alcohol is consumed; acamprosate, a medication thought to stabilize chemical signaling in the brain that is associated with relapse; and naltrexone, a medication that blocks the euphoric effects of both alcohol and opioids.
Through better equipment, improved knowledge of head injury and better medical care, we can mitigate the effects of concussion at all levels of football
Several years ago, I was at a pre-season football practice at a high school where I was working with the team on a concussion research project. The players were lined up in two rows facing each other and with little more instruction from the coach than, “on the whistle, hit the man across from you,” grave concern rushed through my mind.
Professional sports get the lion’s share of attention, but concussions are a significant medical issue at all levels of play. While roughly 1,700 athletes play professional football each year, over three million children and adolescents in the United States play the same game. This squarely places concussions as a significant public health concern in homes from coast to coast.
For decades, athletes, parents, coaches and medical professionals considered concussion a temporary injury with no long-term consequences. Athletes commonly played through it as a sign of toughness. Our thinking on concussion shifted around 2005, when the brain tissue pathology report of Mike Webster, a retired National Football League player, was published. The report was the first to identify chronic traumatic encephalopathy (CTE) in a football player and suggested the disease was linked to concussions the player had suffered on the field. Since then, public attention has focused on this link between brain injury and blows to the head, and much of that conversation has centered around football.
Participation in any sport carries injury risk, and concussion will always be part of that calculus. This was never more evident than early in the 2022 season, when Miami Dolphins quarterback Tua Tagovailoa was concussed while playing against the Cincinnati Bengals, his contorted hands on full display in front of a national audience. Concussion prevention and care is better now than it has ever been, but short of ending all sports, we need to do more to prevent them. This means improvements in equipment, better training for coaches and players, and better medical care.
In the early 1900s football was played without helmets, but severe injuries, like skull fractures, lead to the sporadic use of leather helmets in the 1920s. The plastic shell helmet was invented in 1939 and became mandatory a few years later. The first face mask entered the game in the 1950s, and foam padding wasn’t added until the 1960s. Over the years, engineers updated the basic design with more robust face masks and newer internal padding such as air bladders and gel padding. Until recently, these changes were all made with an eye toward reducing the most severe brain injuries such as brain bleeds and skull fractures, and they have been largely successful. But as our understanding of concussion evolved to understand its significance as an injury, so too did the scrutiny of player equipment.
The fundamental problem is that regardless of how much impact force is mitigated by protective equipment, we cannot stabilize the brain inside the skull following an impact – likely making a concussion-proof helmet an impossible engineering challenge. That hasn’t stopped helmet improvements, however. With a new eye toward concussion prevention, researchers at Virginia Tech introduced a helmet rating system specific to concussion risk in 2011. The five-star system (one being the worst, five the best) rates the ability of the helmet to reduce concussion risk. This gives consumers tangible evidence to make an informed purchase. In its first year, only a single helmet was awarded a five-star rating, but now, the most recent rankings list more than 25 five-star helmets. Modern helmets use advanced shell materials that flex upon impact, have moveable panels to absorb forces, and multi-layered padding that responds to different impact velocities. Companies will continue to improve helmets as new materials become available, guided by the newest science.
But engineering limitations should not stop us from closely examining other factors that can reduce concussion risk. For example, in one study of head impacts and concussion across five college football seasons, nearly 50 percent of concussions occurred during the four-week pre-season; the rest occurred over the next 12 weeks of in-season play. Such data led the NCAA to reduce the number of allowable full contact pre-season practice sessions. Others have shown that reducing the number of high school football practices in which contact is allowed reduces head impacts by as much as 46 percent. Moving the kickoff line forward to the 40-yard line reduced the number of times the ball is run back by the receiving player. This is one of the riskiest plays of the game in terms of concussion.
As players get older and start playing contact football, the coaching staff must teach appropriate tackling technique—not leading with the head and wrapping their opponent up with their arms. Learning how to do this properly has the added benefit of more playing time (i.e., you can’t play if you’re injured). In addition, having medical providers with training in concussion management at practices and competitions can help identify and rapidly remove injured players, a known factor in reducing injury severity and the time out of sport.
While injury prevention is the best approach, concussions will always be a part of sport participation. This is broader than just football. This injury occurs in all sports and affects athletes regardless of sex. Women, who make up about 45 percent of college athletes, tend to report more concussions in sex-comparable sports, but represent only 20 percent of the medical literature. Some researchers have speculated women are more likely in general to report a medical concern to a health provider, but they also have smaller neck musculature relative to their head mass limiting their ability to stabilize their head when hit. Some data indicate an increased risk of concussion at different points in the menstrual cycle. This all suggests that concussion is not unique to football and unless we intend to ban all sports, it is urgent that we continue working to better understand concussions and how to prevent them.
In the interim, several medical and organizations have created concussion guidelines that did not exist in the early 2000s, and every four years a group of international experts meets to review the relevant medical literature and make recommendations on the best approach to concussion care. The latest guidelines are due later this year. Like all medical advances, change will be incremental and often hard to perceive, but concussion prevention and care is better now than it has ever been. Regardless, many athletes will do just about anything to stay on the field and “play through the pain.” It is imperative to create a culture that supports athletes and encourages them to come forward when injuries happen, empowers them to be their own health advocates, and allows them to protect their overall brain health while playing competitively and safely.
Change is happening in how we view sports-related concussions. In my 20 years of doing concussion research, I can attest that we now take concussion seriously as an injury, better understand impact biomechanics and other risk factors, are making strides to prevent it, and are improving the management and recovery process. Sports are an integral part of American culture, and they give millions of children much-needed exercise. As those children become adults and make sport their hobby or even vocation, ensuring they play safely at all levels is essential.
A professor explains why he is allowing students to incorporate ChatGPT into their writing process instead of banning the new technology
To read the news, the sanctity of everything from college application essays to graduate school tests to medical licensing exams is imperiled by easy access to advanced artificial intelligence like ChatGPT, the AI chatbot that can produce remarkably clear, long-form answers to complex questions. Educators in particular worry about students turning to ChatGPT to help them complete assignments. One proposed solution is to roll back the clock to the 20th century, making students write exam essays using pen and paper, without the use of any Internet-connected electronic devices. The University of California, Los Angeles, where I teach, is considering making it an honor code violation to use ChatGPT for taking an exam or writing a paper.
That’s the wrong approach. This semester, I am telling the students in my class at the UCLA School of Law that they are free to use ChatGPT in their writing assignments. The time when a person had to be a good writer to produce good writing ended in late 2022, and we need to adapt. Rather than banning students from using labor-saving and time-saving AI writing tools, we should teach students to use them ethically and productively.
To remain competitive throughout their careers, students need to learn how to prompt an AI writing tool to elicit worthwhile output and know how to evaluate its quality, accuracy and originality. They need to learn to compose well-organized, coherent essays involving a mix of AI-generated text and traditional writing. As professionals working into the 2060s and beyond, they will need to learn how to engage productively with AI systems, using them to both complement and enhance human creativity with the extraordinary power promised by mid-21st-century AI.
In addition to the sound pedagogical reasons for treating ChatGPT as an opportunity and not a threat, there are practical ones as well. It simply isn’t feasible to effectively ban access to this technology. Honor code or not, many students will be unable to resist the temptation to seek AI assistance with their writing. And how would an educational institution enforce a ChatGPT ban? While there are toolsaimed at detecting text produced by AI, future versions of AI will get better at emulating human writing—including to the point of emulating the style of the particular person who is using it. In the resulting arms race, the AI writing tools will always be one step ahead of the tools to detect AI text.
Enforcement of a ChatGPT ban would also inevitably produce the injustice of false positives and false negatives. Some students who use ChatGPT despite a ban would, through luck or thanks to careful-enough editing of AI-generated text, avoid having their writing flagged as AI-assisted. Worse, some students would be falsely accused of using ChatGPT, triggering enormous stress and potentially leading to punishment for a wrong they did not commit.
And what of the argument that learning to write well provides benefits that go well beyond writing? Writing a good essay from scratch requires careful, often painstaking, thought about organization, flow and communication. Learning to write without AI does indeed promote focused, disciplined thinking. But learning to successfully combine unassisted and AI-assisted writing to create truly good essays also requires these qualities.
Writing is a craft worthy of enormous respect, one which few of us ever master. But most students don’t aspire to become professional writers. Instead, they are preparing for careers where they will write to further goals beyond the production of writing. As we do today, they will write to communicate, explain, convince, memorialize, request and persuade. AI writing tools, when properly used, will help them do those things better.
When I was a middle and high school student in the late 1970s and early 1980s, I was told that professional success required good “penmanship” and the ability to perform long division by hand. By the time I entered the professional workforce in the late 1980s, technology advances had rendered those skills obsolete. Education culture can be very slow to change, as evidenced by the fact that many schools today still force children to learn long division—a task they will never have to perform anywhere outside of school. With AI writing, educators should stay ahead of the technology curve, as opposed to lagging decades behind it.
The upshot: I am helping my students to prepare for a future in which AI is simply another technology tool as opposed to a novelty. I am also telling them that they are solely and fully responsible for the writing they turn in bearing their name. If it’s factually inaccurate, that’s on them. If it’s badly organized, that’s on them. If it’s stylistically or logically inconsistent, that’s on them. If it’s partially plagiarized, that means that they have committed plagiarism.
In short, I’m encouraging my students to become responsible, aware users of the AI technologies that will play a profoundly important role over the course of their careers. The AI writing, so to speak, is on the wall.
ARYAN'S BLOG 