A well-dramatized telling of the Salk polio vaccine's origins. It's a gripping read, demonstrating the widespread terror of polio season among parents and kids, even telling the sad stories of a few polio victims who played roles in the vaccine's testing and development.
But sometimes it felt like the author focused too much on the researchers' human drama---Salk's killed-virus vaccine vs. Sabin's live-virus! Which will win? Will one fail dramatically like the Brodie vaccine? What catty snipes will they make next?---and didn't explain why each scientist felt his belief was justified.
This story had such potential to illustrate why researchers & statisticians design studies the way they do---why the trials have to be just so in order for the sample results to reflect something meaningful about the whole population. That concern is touched on in this book, but shallowly. Still, the book is worth reading as it is.
There's also some good history about FDR and his support of the fight against polio. I hadn't known he developed polio as an adult, after already being a (not-so-successful?) politician/candidate.
Moving, too, were the mentions of Salk's insurance inquiries (how much would it cost to insure yourself and lab colleagues against possible self-infection with polio? a test-tube could break or virus samples could splatter on you at any time, as they did when the Life photographers visited... oof...)
...and of his early vaccination of his own family (you have to be seriously confident of your results to give your own sons this vaccine!).
Favorite parts and statistical questions:
* p.5: "Facts were precious enough commodities when you needed them; why violate them once you'd got them in hand?"
* p.67: I didn't know March of Dimes was started to raise funds for polio, nor that it was a pun on a popular newsreel March of Time.
* p.57-58: Pacifists during WW2 were not dismissed from the draft---they were put to work as vaccine guinea pigs!
* p.72-73: Salk also worked earlier on a flu vaccine, and he ran into the same problems we have today of quick-changing flu strains from year to year. I never thought about the intense "epidemiological reconnaissance work" or "viral scouting" needed to stay ahead of each year's flu. What a great job title: Viral Scout.
* p.117: "It was the zero-sum nature of the virus game that the only way to prevent more blameless children from streaming into the upper floors [of the polio hospital] was to sicken and kill the blameless monkeys in the basement. Somewhere, perhaps, was the person who could tease out all the moral threads of that arrangement, but polio scientists, as a rule, could not turn their minds to the task. Once the virus was beaten, all the creatures---the ones upstairs and the ones below---could be left in peace. Until someone offered a better deal, this was the one they would have to take."
* p.228: Sabin, a "rival" scientists, thinks a live-virus vaccine would work better than Salk's killed-virus vaccine (but doesn't have one ready yet himself). When Congress deliberates on Salk's field trial, Sabin says: "I for one, would strongly oppose large-scale work on hundreds of thousands of children based on the work of any one investigator..." which is a reasonable argument. I wish the author had spent a bit more time explaining Sabin's POV and rationale, instead of presenting it as just a personal conflict between two guys. Also, how did Salk's preliminary trials differ from the massively-failed Brodie-Park-Kolmer vaccine trials a few decades before? Clearly they did---but how?
* p.234: Nice quick summary of double-blind trial vs observational controls. But what justified Salk's faith in his vaccine, without the gold standard of a double-blind trial? Just the small-scale trial results around p.187? When is lab-science proof good enough for medicine, vs. statistical proof of a massive trial?
* p.238: If Salk thought it's unethical to give placebos (and hence preferred observational controls over a double-blind trial)... then why was it ethical to give the drug to 2nd-graders but not at all to 1st- or 3rd-graders? Why is refusal to treat any better than placebo, which would at least give you sounder statistical evidence?
* p.242: Tommy Francis (Salk's former supervisor/mentor) led the design and analysis of the major field trial, and he convinced Salk to go for a double-blind rather than observational-control study. But how? The argument presented here is from authority: "If Francis, of all people, said a double-blind trial was the best possible one he could conduct, Salk would accept that." Surely there was much more to Francis-convincing-Salk than that!
(Also---logistically they had to end up doing a dual-mode trial: double-blind in some states, observational-controls in others. What are the statistical methods for combining such results?)
* p.245: Another curious moral question, once Salk was convinced of his vaccine's value (but before field tests were complete), about whether to give some to family & friends. Obviously there's not yet enough vaccine to share with all the strangers sending letters requesting it---does that make it immoral to share a small amount with "a handful of children in their charmed circle and not to those who happened to fall outside it"?
* p.262-3: Failure during manufacturing led to arguments over new testing procedures. 54 monkeys vs 350 monkeys per lot, and 11 safe lots required in a row before releasing any---what statistical arguments justified these numbers? I wish the author gave a bit more space for the NIH statisticians' response to Rivers' comment: "As far as I'm concerned, you can take your pencil and paper and shove them up your ass."
* p.277: Large-scale implementation of a trial is a mess. This section should be required reading for people who want to do massive data collection. Technical issues with syringes and manufacturing; personal issues with getting the kids to the shot facilities; even confusion over the simple fact that blood samples must be drawn from the same kids who got the shots!
* p.280: Salk, impatient to change a manufacturing procedure: "Data from these experiments have not yet cooled off. But it is said that to await certainty is to await eternity."
* p.311: Salk again, this time waiting out a possible manufacturing flaw: "The people of Idaho have had a tragic experience. Our deep concern has not been a secret. But only when the evidence is available will the state be in a position to draw a sound conclusion."
(I like that---"a sound conclusion"---that's what experimental design is all about: you can't always know if your conclusion is correct, but at least it can be soundly drawn.)