About a third of all deaths in U.S. hospitals are caused by sepsis, killing some 270,000 Americans every year. Speed is everything in treating the condition: The sooner hospitals can discover cases, the better the outcomes will be. On the strength of their own testing, commercial electronic health record vendor Epic introduced their sepsis model to speed detection, which has now been implemented in hundreds of hospitals across the United States. But because the model is proprietary, the amount of publicly available information about it has been limited, and to date there have been no independent validations.
Until now. A large study involving 27,697 people hospitalized in the University of Michigan medical system from 2018–2019 found the model predicts sepsis poorly, failing to detect it two-thirds of the time, making many false positive predictions, and falling far short of the company’s self-reported numbers. In response to a comment request, Epic said through a spokesperson that the model has triggered warnings for thousands of people that might otherwise have been missed, and they criticized the paper’s approach. “They did not take into account the analysis and required tuning that needs to occur prior to real world deployment to get optimal results,” the company said. JAMA Internal Medicine
The withering gulf between the discovery of new technologies and their eventual commercial availability is commonly called the “Valley of Death”—an unfortunate concept when it comes to medical discoveries, where that distance is too often actually measured in human lives. According to a perspective this week by the chief medical officer for the health company Accolade and echoed in an oft-quoted statistic, it takes 17 years on average for new drugs and treatment strategies to make it into the clinic. The coronavirus pandemic has shattered that pace, and one of its long-term legacies we can celebrate could be to reduce that time through the advent of technologies like telehealth and decision support systems—notwithstanding the troubles with the tool for predicting sepsis. MedPage Today
A new trimodal technique for imaging the human brain is described in a paper this week from researchers at the University of Illinois, Urbana-Champaign. It combines fMRI, EEG, and a near-infrared imaging technique called EROS, which was developed at the same university to track neuronal activity near the surface of the brain. Together the three complementary approaches promise to better capture how the brain responds to stimuli. Human Brain Mapping
A clinical study of 80 overweight, post-menopausal women on a 12-week, very-low-calorie liquid diet has shown that even though severe calorie restriction leads to weight loss and improves metabolic health, it also decreases the abundance of bacteria in the gut, impairs nutrient absorption, and removes growth restrictions on the pathogenic bacteria C. difficile. During the diet, half the women were limited to less than 800 calories per day (about what you would consume in a single meal or one large milkshake), and half were not. The researchers, from Charité Universitätsmedizin Berlin and the University of California, San Francisco, also showed that when they transplanted stool samples from the five people in the study who lost the most weight into the guts of mice, those mice also lost a significant amount of weight and showed a similar gut restructuring. Nature
Sticks and stones may break your bones but followers come tweeting after. According to psychologists at the University of Cambridge, ideological invective is the best way to go viral. Examining more than 2.7 million recent tweets and Facebook posts published by media outlets and members of Congress, they found, regardless of party affiliation, opponent-slamming political posts were twice as likely to be shared. In fact, when U.S. members of Congress used negative language across the two channels, shares increased by up to 45 percent per word. Each positive word, on the other hand, decreased shares by 2–5 percent. PNAS
A systematic review of 315 randomized, controlled clinical trials of 14 different dietary supplements and alternative weight-loss therapies found only limited high-quality evidence that any of them work—in part due to methodological inconsistencies in trial designs. In a related perspective, the authors call on federal and state regulatory authorities to protect consumers by better policing the accuracy of marketing claims and preventing false ones. That’s a high hill to climb. Lack of evidence does not currently matter as far as the U.S. Food and Drug Administration is concerned. Dietary supplements are not regulated as drugs at all but instead allowed to be marketed so long as their ingredients are generally regarded as safe—a low bar for the FDA and an even lower one for all the rabid, dietary-supplement snake oil salesmen of the world. Obesity
Subscribe to our free weekly newsletter.