Weekend Long Reads: Remdesivir Is No Wonder Drug

by Kevin Schofield

In this column, I'll be giving you pointers to some of the most interesting articles and studies I've recently come across. I'll be aiming for things that are "less than a book, but more than a newspaper article" — readings that are a bit of a mental workout to take in but that expand our perspectives and make us better informed in our daily lives. I'll also try to pick items that share the joy of reading outside your area of expertise: articles not so technical and arcane that they are incomprehensible but that still give us a glimpse of how experts think about work in their own field.

This week: Is remdesivir a cure for COVID-19, should we trust the presidential polls this time, and is anyone other than the rich getting richer?

Is Remdesivir a Cure for COVID-19?

This week the FDA announced that it has approved the antiviral drug remdesivir as the first official COVID-19 treatment for patients who require hospitalization. But how well does it work? Should we think of it as a "cure"?

A recent study commissioned by the World Health Organization (WHO) looked at that question. It reports on the results of hospital trials of remdesivir and four other antiviral treatments in patients who were admitted for COVID-19. Unfortunately, the results were not great. The study found that the drug did not reduce the mortality rate among COVID patients, nor did it reduce the number of patients who needed to be ventilated. The one bright spot is that for those who survived, remdesivir may have slightly shortened their time to recovery. Digging in further, it looks like the drug helped "low-risk" patients a bit but may have harmed "high-risk" patients; specifically, those who were not on a ventilator when admitted did better when given remdesivir, and those who were already on a ventilator did worse.

Coming back to our questions: sadly, remdesivir doesn't work very well on COVID-19, and it's far from the cure that we're all hoping for. So keep wearing that mask, washing your hands, and social distancing while we wait for vaccines to arrive.

Repurposed antiviral drugs for COVID-19 — interim WHO SOLIDARITY trial results

How Americans' Income Has Changed in the Past Fifty Years

Last month the nonprofit RAND Corporation, which for decades has been famous for its comprehensive public policy research studies, published a deep dive into income trends in the United States between 1975 and 2018. We're all aware of the staggering extent to which the rich are getting richer, but this report does an amazing job of picking that apart and showing all of the sub-trends driving it. For starters, we can see in the chart below how much things changed after the "stagflation" period of the early 1970's: before that, income growth was fairly even across all Americans, but starting in the late seventies income growth shifted dramatically towards wealthier Americans.

The report highlights the racial, gender, educational, and geographic (rural vs. urban/suburban) disparities in income, and it presents a complex picture. For instance, among the 25% of people with the lowest earnings, the numbers aren't what you might expect: white men and Black men earn the same amount, and Black women make more than white women. In fact, in this quartile, white men's income (adjusted for inflation) has been stagnant since 1975.

(The "counterfactual" figures in the chart above are what the earnings would be today if they had tracked perfectly with GDP growth.)

Contrast that with the numbers for top 5% of earners, which are more in line with the figures we've learned to expect: race and gender lead to large disparities in income.

The report also shows that having a college degree makes a huge difference in earnings in every income bracket — as does living in an urban or suburban area.

After the first ten pages of introduction, this report is fifty pages of fascinating statistics. Some of it will make you angry, and some will make you scratch your head and wonder how that could be — or how we could have made so little progress on income equality in forty five years.

Trends in Income from 1975 to 2018

Can We Believe the Polls?

In late October 2016, political polls showed presidential candidate Hillary Rodham Clinton with a comfortable lead over Donald Trump. Then on election night we were all surprised by the outcome as Trump surged ahead and took the electoral college (even though he lost the popular vote). Now, less than two weeks from the next election, the polls are once again showing the Democratic candidate with a sizable lead. Should we trust the polls?

First of all, if by "trust the polls" you mean "don't bother voting because Joe's got it locked up," then no, definitely do not trust the polls, and go vote. Even if the polls are right, there are still many other important issues on the ballot that we need to decide. But the polls themselves are estimates based upon a relatively small sample of voters, and they could certainly be wrong again. That said, the political analysis site FiveThirtyEight (which four years ago this week was saying "hey folks, Trump still has a chance to win this thing") recently looked at how the professional pollsters have adjusted their tactics to prevent a repeat of 2016 this year. They noted that some of the problems four years ago stemmed from under-sampling certain demographics, and so the pollsters have made some adjustments to how they select individuals to survey to better balance the pool to be more reflective of the country. For instance, they now weight their samples by education, because educational attainment has become a significant predictor of how someone will vote. Many polls have also increased the number of cell phones versus landline telephones that they call, and some have even moved away from randomly dialing phone numbers toward using an address-based approach to ensure geographic representation. Some pollsters have also drawn samples from lists of registered voters in order to pre-select a certain mix of Democrats, Republicans, and independents.

But these adjustments don't guarantee that the polls will be accurate. Moving away from random selection towards weighting the results can introduce its own bias. Also, as we've already seen, the shift this year to mail-in ballots in many states has led to an increase in early voting — and potentially an equally large increase in voting overall. It's too early to know whether the polls are accurately capturing the effects of those shifts in voting patterns. Plus we don't know how the various efforts at suppressing voting will affect the election results; if suppression techniques are successful, it's entirely possible that the polls will more accurately represent the will of the people than the official ballot counts.

The bottom line remains that no one should feel complacent about how this election will turn out. So vote!

What Pollsters Have Changed Since 2016 — And What Still Worries Them About 2020

Kevin Schofield is a freelance writer and the founder of Seattle City Council Insight, a website providing independent news and analysis of the Seattle City Council. After graduating from Dartmouth College with a degree in computer science, he worked for Microsoft Corporation for 26 years. That culminated in a term as Chief Operations Officer for Microsoft Research, the division of the company focused on advancing the state of the art in computing. Upon leaving Microsoft in 2014 he decided to embrace his love of writing. Kevin volunteers at the Woodland Park Zoo in Seattle, where he also serves on the Board of Directors. He is also Vice Chair of the Board of Trustees of Harvey Mudd College in Claremont, California. Beyond writing, his personal passions are his twin daughters, photography, cooking, and playing the guitar. He lives in Seattle, Washington.

Featured image is attributed to Jernej Furman under a Creative Commons 2.0 license.