Methodology – artdiamondblog.com

Bayesian Updating, Not Clinical Trials, Is Key to Advancing Medical Knowledge

(p. D8) In the early pandemic era, for instance, airborne transmission of Covid-19 was not considered likely, but in early July the World Health Organization, with mounting scientific evidence, conceded that it is a factor, especially indoors. The W.H.O. updated its priors, and changed its advice.

This is the heart of Bayesian analysis, named after Thomas Bayes, an 18th-century Presbyterian minister who did math on the side. It captures uncertainty in terms of probability: Bayes’s theorem, or rule, is a device for rationally updating your prior beliefs and uncertainties based on observed evidence.

. . .

As Marc Lipsitch, an infectious disease epidemiologist at Harvard, noted on Twitter, Bayesian reasoning comes awfully close to his working definition of rationality. “As we learn more, our beliefs should change,” Dr. Lipsitch said in an interview.

. . .

But there is little point in trying to establish fixed numbers, said Natalie Dean, an assistant professor of biostatistics at the University of Florida.

“We should be less focused on finding the single ‘truth’ and more focused on establishing a reasonable range, recognizing that the true value may vary across populations,” Dr. Dean said. “Bayesian analyses allow us to include this variability in a clear way, and then propagate this uncertainty through the model.”

. . .

Joseph Blitzstein, a statistician at Harvard, delves into the utility of Bayesian analysis in his popular course “Statistics 110: Probability.” For a primer, in lecture one, he says: “Math is the logic of certainty, and statistics is the logic of uncertainty. Everyone has uncertainty. If you have 100 percent certainty about everything, there is something wrong with you.”

By the end of lecture four, he arrives at Bayes’s theorem — his favorite theorem because it is mathematically simple yet conceptually powerful.

“Literally, the proof is just one line of algebra,” Dr. Blitzstein said. The theorem essentially reduces to a fraction; it expresses the probability P of some event A happening given the occurrence of another event B.

“Naïvely, you would think, How much could you get from that?” Dr. Blitzstein said. “It turns out to have incredibly deep consequences and to be applicable to just about every field of inquiry” — from finance and genetics to political science and historical studies. The Bayesian approach is applied in analyzing racial disparities in policing (in the assessment of officer decisions to search drivers during a traffic stop) and search-and-rescue operations (the search area narrows as new data is added). Cognitive scientists ask, ‘Is the brain Bayesian?’ Philosophers of science posit that science as a whole is a Bayesian process — as is common sense.

. . .

Even with evidence, revising beliefs isn’t easy. The scientific community struggled to update its priors about the asymptomatic transmission of Covid-19, even when evidence emerged that it is a factor and that masks are a helpful preventive measure. This arguably contributed to the world’s sluggish response to the virus.

. . .

In 1650, Oliver Cromwell, Lord Protector of the Commonwealth of England, wrote in a letter to the Church of Scotland: “I beseech you, in the bowels of Christ, think it possible you may be mistaken.”

In the Bayesian world, Cromwell’s law means you should always “keep a bit back — with a little bit of probability, a little tiny bit — for the fact that you may be wrong,” Dr. Spiegelhalter said. “Then if new evidence comes along that totally contradicts your main prior belief, you can quickly ditch what you thought before and lurch over to that new way of thinking.”

“In other words, keep an open mind,” said Dr. Spiegelhalter. “That’s a very powerful idea. And it doesn’t necessarily have to be done technically or formally; it can just be in the back of your mind as an idea. Call it ‘modeling humility.’ You may be wrong.”

For the full story, see:

Siobhan Roberts. “Thinking Like an Epidemiologist.” The New York Times (Tuesday, August 4, 2020): D8.

(Note: ellipses added.)

(Note: the online version of the story has the same date as the print version, and has the title “How to Think Like an Epidemiologist.”)

“The Concept of Microaggressions” Is “Subjective by Nature”

(p. 25) Scott Lilienfeld, an expert in personality disorders who repeatedly disturbed the order in his own field, questioning the science behind many of psychology’s conceits, popular therapies and prized tools, died on Sept. 30 [2020] at his home in Atlanta.

. . .

He . . . received blowback when he touched a nerve. In 2017, he published a critique of the scientific basis for microaggressions, described as subtle and often unwitting snubs of marginalized groups. (For instance, a white teacher might say to a student of color, “My, this essay is so articulate!”) Dr. Lilienfeld argued that the concept of microaggressions was subjective by nature, difficult to define precisely, and did not take into account the motives of the presumed offender, or the perceptions of the purported victim. What one recipient of the feedback might consider injustice, another might regard as a compliment.

The nasty mail rolled in, from many corners of academia, Dr. Lilienfeld told colleagues.

“There was no one like him in this field,” said Steven Jay Lynn, a psychology professor at Binghamton and a longtime collaborator. “He just had this abiding faith that science could better us, better humankind; he saw his championing as an opportunity to make a difference in the world. He enjoyed stepping into controversial areas, it’s true, but the motives were positive.”

For the full obituary, see:

Benedict Carey. “Scott Lilienfeld, 59, Psychologist Who Questioned Science of Psychology, Dies.” The New York Times, First Section (Sunday, October 18, 2020): 25.

(Note: ellipses, and bracketed year, added.)

(Note: the online version of the obituary has the date Oct. 16, 2018, and has the title “Scott Lilienfeld, Psychologist Who Questioned Psychology, Dies at 59.”)

Expense of Clinical Trials Reduce the Incentive to Re-Purpose Old, Cheap, Off-Patent Vaccines

(p. A5) “Retrospective studies are great and they provide some hints, but there are caveats,” said Dr. Shyam Kottilil, a professor of medicine with the Institute of Human Virology at the University of Maryland School of Medicine. “It’s very difficult to establish causality.”

Interest in the cross-protective effects of vaccines has led to efforts to repurpose old vaccines that may have potential to provide at least transient protection against the coronavirus until a specific vaccine against SARS-CoV-2 is developed and proven safe and effective, he said.

“But nobody knows whether this approach will work unless we test them,” Dr. Kottilil said. “To endorse this, you need to do really good randomized clinical trials.” There is little incentive for private companies to invest in expensive trials because the old vaccines are cheap and off-patent, he added.

For the full story, see:

Roni Caryn Rabin. “Are Past Vaccinations a Shield? It’s Doubtful.” The New York Times (Thursday, July 30, 2020): A5.

(Note: the online version of the story has the date July 29, 2020, and has the title “Old Vaccines May Stop the Coronavirus, Study Hints. Scientists Are Skeptical.”)

Arthur Ashton’s Serendipitous Invention of Optical Tweezers

(p. B11) Arthur Ashkin, a physicist who was awarded a 2018 Nobel Prize for figuring out how to harness the power of light to trap microscopic objects for closer study, calling his invention optical tweezers, died on Sept. 21 [2020] at his home in Rumson, N.J.

. . .

Dr. Ashkin’s discovery was serendipitous.

In 1966, he was head of the laser research department at Bell Labs, the storied New Jersey laboratory founded by the Bell Telephone Company in 1925, when he went to a scientific conference in Phoenix. There, in a lecture, he heard two researchers discuss something odd that they had found while studying lasers, which had been invented six years earlier: They had noticed that dust particles within the laser beams careened back and forth. They theorized that light pressure might be the cause.

Dr. Ashkin did some calculations and concluded that this was not the cause — it was most likely thermal radiation. But his work reignited a childhood interest in the subject of light pressure.

Light pushes against everything, including people, because it comprises tiny particles called photons. Most of the time the pressure is utterly insignificant; people, for one, feel nothing. But Dr. Ashkin thought that if objects were small enough, a laser might be used to push them around.

. . .

Then, in 1986, he and several colleagues, notably Steven Chu, achieved the first practical application of optical tweezers when they sent a laser through a lens to manipulate microscopic objects. Their results were published in another paper in Physical Review Letters. Dr. Chu began using the tweezers to cool and trap atoms, a breakthrough for which he was awarded a one-third share of the Nobel Prize in Physics in 1997.

Dr. Ashkin, it was clear, was irked that the Nobel committee had not recognized his foundational work in awarding the prize. But he had already begun to use the tweezers for a different purpose: trapping live organisms and biological material.

Other scientists thought this application would not work, as he explained in an interview with the Nobel Institute after he was awarded the prize in 2018.

“They used light to heal wounds, and it was considered to be deadly,” he said. “When I described catching living things with light, people said, ‘Don’t exaggerate, Ashkin.’”

. . .

Dr. Ashkin was awarded one-half the 2018 physics prize, . . . . In so doing he became, at 96, the oldest recipient of a Nobel Prize at the time.

. . .

Dr. Ashkin’s retirement from Bell Labs did not stop him from continuing his research. When he received word of his Nobel Prize, he was working on a project in his basement to improve solar energy collection. Asked if he was going to celebrate, he said: “I am writing a paper right now. I am not about celebrating old stuff.”

For the full obituary, see:

Dylan Loeb McClain. “Arthur Ashkin, 98, Dies; Nobel-Winning Physicist.” The New York Times (Tuesday, September 29, 2020): B11.

(Note: ellipses, and bracketed year, added.)

(Note: the online version of the obituary was updated Oct. 5, 2020, and has the title “Arthur Ashkin, 98, Dies; Nobel Laureate Invented a ‘Tractor Beam’.”)

The essay about Aoyagi mentioned above is:

Severinghaus, John W. “Takuo Aoyagi: Discovery of Pulse Oximetry.” Anesthesia & Analgesia 105, no. 6 (Dec. 2007): S1-S6.

“Greatness in Science Often Comes From the Well-Prepared Mind Turning a Chance Observation Into a Major Discovery”

(p. 27) Takuo Aoyagi, a Japanese engineer whose pioneering work in the 1970s led to the modern pulse oximeter, a lifesaving device that clips on a finger and shows the level of oxygen in the blood and that has become a critical tool in the fight against the novel coronavirus, died on April 18 [2020] in Tokyo.

. . .

Mr. Aoyagi’s contribution to medical science was built on decades of innovation and invention. In an essay about Mr. Aoyagi, John W. Severinghaus, a professor emeritus of anesthesia at the University of California, San Francisco, wrote in 2007 that Mr. Aoyagi’s “dream” had been to detect oxygen saturation levels without having to draw blood.

. . .

But he soon ran into a problem. Blood does not flow smoothly like an open tap, but pulses through the body irregularly, thus preventing an accurate recording of dye levels. The problem, however, turned out to be an opportunity. By devising a mathematical formula to correct for this “pulsatile noise,” he created a device that measured oxygen levels with greater accuracy than before.

“Greatness in science, often, as here, comes from the well-prepared mind turning a chance observation into a major discovery,” Dr. Severinghaus wrote.

For the full obituary, see:

John Schwartz and Hikari Hida. “Takuo Aoyagi, 84; Invented Medical Device.” The New York Times, First Section (Sunday, May 3, 2020): 27.

(Note: ellipses, and bracketed year, added.)

(Note: the online version of the obituary was updated June 20, 2020, and has the title “Takuo Aoyagi, an Inventor of the Pulse Oximeter, Dies at 84.”)

The essay about Aoyagi mentioned above is:

Severinghaus, John W. “Takuo Aoyagi: Discovery of Pulse Oximetry.” Anesthesia & Analgesia 105, no. 6 (Dec. 2007): S1-S6.

“You Can’t Wait for Somebody to Make a Giant Study”

(p. A6) In April [2020], researchers published an article in the Journal of the American Medical Association suggesting many Covid-19 patients with respiratory distress might require a different treatment approach than typically used for ARDS.

. . .

Maurizio Cereda, an anesthesiologist and head of the surgical ICU at the Hospital of the University of Pennsylvania, said doctors normally use standardized tables to match the level of oxygen in the blood with the amount of PEEP needed. Penn tends to use a table with lower PEEP values, he said, but even those lower levels seem to damage the lungs of some of his Covid-19 patients. As a result, he disregards the table entirely at times, he said, even though some in his institution disagree with his approach.

“You can’t wait for somebody to make a giant study,” Dr. Cereda said. “You are alone with your clinical observation. A lot of people don’t feel comfortable with that because they want to have big guidelines. People seem to be afraid they’re going to do something wrong.”

. . .

At Maimonides Medical Center in Brooklyn, critical-care and emergency-medicine doctor Cameron Kyle-Sidell said he was initially seeing much higher mortality rates from Covid19 patients on ventilators than he would have expected from classic ARDS, possibly because physicians were sticking to PEEP levels used to treat traditional ARDS.

“There are people who are treating this the way they would have treated any other ARDS,” he said. “Then there’re people on the flip side—and I am on that flip side—that think you should treat it as a different disease than we treated in the past.”

For the full story, see:

Sarah Toy and Mark Maremont. “Doctors Split on Best Way To Treat Coronavirus Cases.” The Wall Street Journal (Thursday, July 2, 2020): A6.

(Note: ellipses, and bracketed year, added.)

(Note: the online version of the story has the date July 1, 2020, and has the title “Months Into Coronavirus Pandemic, ICU Doctors Are Split on Best Treatment.” The online version quoted above includes a couple of added sentences quoting Dr. Cereda, beyond the single sentence quoted in the print version.)

Natural Experiments Are Equal to Randomized Double-Blind Clinical Trials in Showing Causality

(p. B6) . . . randomized controlled trials are the gold standard in medicine. Using randomization (by, say, flipping a coin to assign patients to a new treatment or not) is the best way to determine whether treatments work.

Unfortunately, randomized trials take time — which is a problem when doctors need answers now. So doctors and public health officials have been turning to available real-world data on patient outcomes and trying to make sense of them.

. . .

“Large-scale randomized evaluations have been less common in economics, prioritizing the need for economists to identify often creative but sometimes narrow natural experiments to estimate the causal effects of treatments,” said Amitabh Chandra, an economist at the Harvard Business School and the Kennedy School of Government.

Ashish Jha, recently appointed the dean of the Brown University School of Public Health, said that while “natural experiments have causal interpretations, typical associational studies in medicine do not, which may make some medical researchers less comfortable interpreting the results.”

. . . Most doctors can relate to recent comments by the Food and Drug Administration director Stephen Hahn in last week’s congressional pandemic hearing. “In a rapidly moving situation like we have now with Covid-19,” he said, decisions are made “based on the data that’s available to us at the time.”

For the full commentary, see:

Anupam B. Jena and Christopher M. Worsham. “THE UPSHOT; What Coronavirus Researchers Can Learn From Economists.” The New York Times (Thursday, July 2, 2020): B6.

(Note: ellipses added.)

(Note: the online version of the commentary has the date June 30, 2020, and has the same title as the print version.)

Open Offices Reduce Productivity and Spread Diseases

(p. B4) When historians of the early 21st century look back on the pre-Covid era, one of the absurdities they might highlight is the vogue for gigantic, open-plan offices. The apotheosis of this trend of breaking down barriers between co-workers must surely be Facebook Inc.’s 433,555-square-foot Frank Gehry-designed open-plan office at its headquarters in Menlo Park, Calif. Opened in 2015, it’s now a ghost town, a monument to offices vacated by the pandemic.

Cramming cavernous spaces with as many desks as they could hold might have increased serendipitous interactions, but it almost certainly reduced productivity and helped spread communicable diseases, including coronavirus.

. . .

Cue the “dynamic workplace,” a pivot away from the open plan, built on the idea that with fewer employees coming to work on any given day, offices can offer them more flexibility of layout and management.

While open offices and dynamic workplaces share similar components—privacy booths and huddle rooms to escape the hubbub, cafe-like networking spaces, etc.—they’re philosophically distinct. One is intended to be a place where people come (at least) five days a week, and get most of their work done on site. The other is planned for people rotating in and out of the office, on flexible schedules they have more control over than ever.

. . .

Research on hot-desking in office spaces, for example—where employees give up a dedicated space in favor of first-come-first-serve seating—finds that it decreases socialization and trust. This happens because employees figure they might never again see the person they sit next to on a given day, says Dr. Sander. In other studies, employees complain they can’t find their colleagues, that it’s a hassle to find a new spot to work every day, and that such arrangements ignore humans’ innate territoriality and desire to make a space their own.

For the full commentary, see:

Christopher Mims. “Goodbye, Open Office. Hello, ‘Dynamic Workplace.” The Wall Street Journal (Saturday, September 12, 2020): B4.

(Note: ellipses added.)

(Note: the online version of the commentary has the same date and title as the print version.)

How “Blind” Is a Double-Blind Trial When Volunteers Know the Side-Effects of the Vaccine?

(p. A8) George Washington University had vaccinated 129 people since its share of the trials started. I would be No. 130. Altogether, Moderna planned to enroll 30,000 people in its trial. Half would be given the actual vaccine and half would get the placebo. The protocol called for two shots spaced a month apart.

Finally, it was time for my injection, which is when things got a little weird.

“We have to leave you now, because this is a double-blind study and we are blinded,” Dr. Malkin said. “You’ve been randomized.”

Before I could ask her to translate what she had just said, she was gone, and two nurses arrived with my vaccine. The first nurse left, and the second nurse, Linda Witkin, asked whether I was right-handed or left-handed, then proceeded to inject my right arm.

“Which one are you giving me, the vaccine or the placebo?” I asked. She gave me a look, clearly not pleased with my questioning.

. . .

With the Moderna trial, the side effects reported so far have been typical: fever, chills, muscle and joint soreness.

. . .

The night after my shot, I took my temperature: 97.5. I felt under my arms for glandular swelling and felt only mild joint pain.

. . .

“You all gave me the placebo, didn’t you?” I demanded of Dr. Diemert on Wednesday, during my one-week checkup. “I cannot believe I went through all of this and got the placebo.”

He told me that the actual vaccine shot was more “viscous” than the placebo, which was why neither he nor Dr. Malkin could be in the room when I got it, because they would have been able to easily determine. And so he really couldn’t answer because the double-blind program is meant to protect doctors like him from patients like me. He said I wasn’t to badger Ms. Witkin, if I ever even saw her again. He also said that most people reacted more to the second shot than the first one.

I texted the peanut gallery, “I feel no different.”

For the full story, see:

Helene Cooper. “From Reporting on Ebola to Being a Volunteer in a Covid-19 Vaccine Trial.” The New York Times (Saturday, September 12, 2020): A8.

(Note: ellipses added.)

(Note: the online version of the story has the date Sep. 11, 2020, and has the title “Covering Ebola Didn’t Prepare Me for This: I Volunteered for the Covid-19 Vaccine Trial.”)

Quick Less Precise, but Repeated, Covid-19 Tests Can Be Better Than Slow Precise Tests

(p. A1) Public health experts are increasingly calling for a shift in thinking about Covid-19 testing: It is better to get fast, frequent results that are reasonably accurate than more precise results after dayslong delays.

. . .

Covid-19 tests that don’t require a lab tend to be less sensitive than “gold standard” laboratory-based tests, meaning they are likely to miss more cases. But many public health experts now say that repeat testing can make up for the loss of sensitivity, and such testing could quickly identify the most infectious people and help bring transmission to heel as workplaces and schools resume in-person operations and as influenza season looms.

. . .

(p. A6) “When we looked ahead, we realized we needed a paradigm shift from the still-needed diagnostic tests to the screening tests,” said Jonathan Quick, managing director for pandemic response, preparedness and prevention at the Rockefeller Foundation, which released a report in July [2020] calling for a massive scale-up in quick, cheap tests for Covid-19 screening. “As a practical matter, that meant making much more of a new kind of test,” Dr. Quick said.

Most Covid-19 diagnostic testing in the U.S. is processed in laboratories and uses a technique called rt-PCR that searches for the virus’s genetic material and amplifies it. The tests are incredibly sensitive but expensive to run, and the process often requires shipping samples from a test site to a lab.

. . .

“I think there’s a sense of desperation that we need to do something else,” Ashish Jha, dean of Brown University’s School of Public Health, said at a media briefing in August [2020].

. . .

Antigen tests are better at identifying cases when people have more virus in their system—meaning they will likely find people when they are most infectious, said Michael Mina, an epidemiologist at the Harvard T.H. Chan School of Public Health and an advocate of low-cost, widely available at-home testing that can be done on a paper strip.

. . .

The FDA also has said that rapid tests should have comparable accuracy to PCR diagnostic tests—a requirement that some public health specialists and companies say is overly stringent for surveillance testing.

An FDA official noted sensitivity rates lower than PCR might be acceptable, depending on how the test results are used. The agency has allowed for antigen tests with a sensitivity rate of 80% or better, the official said. “You can even have lower than 80% sensitivity” if it is a recurring or serial test.

For the full story, see:

Brianna Abbott, and Thomas M. Burton. “Speed Over Precision Favored in Covid Tests.” The Wall Street Journal (Wednesday, September 9, 2020): A1 & A6.

(Note: ellipses, and bracketed year, added.)

(Note: the online version of the story has the date Sep. 8, 2020, and has the title “Public Health Officials Pursue Covid-19 Tests That Trade Precision for Speed.” Where there are differences between the print and online versions, the passages above follow the online version.)

The report by The Rockefeller Foundation mentioned above is:

The Rockefeller Foundation. “National Covid-19 Testing & Tracing Action Plan.” Thurs., July 16, 2020.

Randomized Controlled Trials Can Obscure “Nuances and Complexities”

(p. A17) You’ve probably heard of the “gold standard”—randomized controlled trials—for evaluating new pharmaceutical therapies, including for Covid-19. Many treatments that showed promise in other studies have turned up muddy results in randomized controlled trials. But that doesn’t mean they’re necessarily ineffective. Doctors and regulators need to consider the totality of medical evidence when treating patients.

. . .

“Randomized trials for some purposes is the gold standard, but only for some purposes,” Harvard’s Donald Berwick, a former health adviser to President Barack Obama, said in an interview with GNS Health Care CEO Colin Hill in 2013. “Context does matter. We’re learning in a very messy world, and the context that neatens up that world may make it hard to know how to manage in the real world.”

. . .

As Thomas Frieden, who directed the Centers for Disease Control and Prevention under Mr. Obama, wrote in a 2017 New England Journal of Medicine article: “Elevating RCTs at the expense of other potentially highly valuable sources of data is counterproductive.” Such limitations affect their use for “urgent health issues, such as infectious disease outbreaks.” He added: “No study design is flawless, and conflicting findings can emerge from all types of studies.”

. . .

Some experts have dismissed the antimalarial hydroxychloroquine, or HCQ, even though more than a dozen observational studies have found it beneficial. A retrospective observational study of Covid-infected nursing-home residents in France, for instance, found those treated with HCQ and azithromycin were 40% less likely to die.

But a few randomized controlled trials found no benefit. A Spanish randomized trial of HCQ for prophylaxis found it didn’t reduce risk of illness among a large group of people exposed in nursing homes, households and health-care settings. Yet two-thirds of the subjects “reported routine use of masks at the time of exposure,” so they were probably less likely to be infected. Nursing-home residents, who may be less likely to wear masks, were 50% less likely to become sick if they took HCQ. But this finding was statistically insignificant, because the trial included only 293 residents.

. . .

Another problem with Covid-19 randomized trials: Patients at different stages of an illness are often assigned the same dosage. Trials don’t reveal differences in how patients respond to a drug at different dosages or illness severity.

Observational studies can do so. Consider a large study by the Mayo Clinic, which found no overall benefit among patients who received a higher-antibody convalescent plasma versus a lower one. Yet the researchers reported a 37% reduction in mortality among patients under 80 who weren’t on a ventilator and received a high-antibody plasma within three days of hospitalizations.

A randomized trial might have obscured these nuances and complexities, denying doctors important information about treatment options. Randomized controlled trials can yield important insights, but it is a medical mistake and a disservice to patients to dismiss other types of evidence.