Wednesday, June 3, 2009

Beryl Benderly's article on Sheri Sangji: Wrong on the facts, wrong on the policy

Beryl Lieff Benderly is apparently a science writer of some repute; she writes the "Taken for Granted" column of ScienceCareers. While most of her work seems pretty harmless, her recent article on the Science Careers website about the UCLA / Sheri Sangji case was published (in slightly different form) on Slate, where it has gotten a few diggs and a raft of comments in "The Fray", Slate's discussion forum.

Wrong on the facts: The article is wrong on a variety of facts, especially the introductory paragraph. And I quote:
A few days after Christmas of 2008, a young technician in a biochemistry laboratory at the University of California-Los Angeles began to transfer a tablespoon of t-butyl lithium from one container to another. T-butyl lithium is pyrophoric, meaning it ignites on contact with air, but Sheri Sangji wasn't wearing a protective lab coat—instead, she had on a flammable synthetic sweatshirt. Somehow the stuff spilled onto her clothing, and she was engulfed in flames. Sangji died from her burns 18 days later, and UCLA officials bemoaned the "tragic accident" that killed her.
Sigh. Where to begin?

Sheri Sangji was practicing synthetic organic chemistry in Professor Patrick Harran's lab; Professor Harran is a prominent chemistry professor, not a biochemist. tert-butyl lithium would never be used for biochemistry procedures, a fact that seems to have escaped Ms. Benderly.

Her writing strongly suggests that Ms. Sangji was literally using a tablespoon to perform the transfer. While 20 mL is indeed close to a tablespoon of material, the fact is that Sheri Sangji was using a 60 mL syringe with nitrogen pressure. Her writing mischaracterizes and trivializes the skill required and the danger involved in the transfer procedure she was doing. She also seems to be suggesting that Ms. Sangji was doing this during some sort of inventory procedure, whereas she was actually attempting to perform a chemical reaction.

Later in the article, Ms. Benderly characterizes Ms. Sangji's death as the result of a "chemical spill", which again brings to mind the incorrect notion that her death was a result of exposure to a toxic compound, not fire.

Even the headline ("Explosions in the Lab") and subhed ("What can be learned from the death of a young biochemist at UCLA?") are wrong.* There's no evidence there was an explosion (although there was clearly a fire) and again, Sheri Sangji was not a biochemist.

*I'll note that I'm holding my fire about the stupid picture of the dude with the tie holding the test tube of food coloring and water.

Why is academia less safe than industry? It's the students, stupid: Here are the things that Ms. Benderly blames for academia's poor safety record: Cal/OSHA's lack of oversight over students, federal funding agencies, powerless EH&S offices and PIs. She misses the forest for the trees. While all of these contribute to the problem, there is a single problem that encompasses all of them: every few years, the best-trained lab workers leave in a process called "graduation" or "getting a real job."

Why is academia inherently less safe than industry? Three words: experience, experience, experience. In the typical academic lab, the average age of the workers is in their mid-20's. In the typical industrial lab, the average age is somewhere north of 30 or so. Let's look at this differently: how many years had Sheri Sangji been working in an organic chemistry lab? I'll make a rudimentary and generous guess -- somewhere around 2-3 years. (I'm guessing this is way high, even with her two published papers.) What is the average experience level in a typical grad school-level lab? Probably 5-6 years (counting undergrad research, summers in the lab, time in grad school.) What's the typical level of experience in an industrial environment? In the lab I used to work in (big pharma), it was at least 10 years (undergrad, grad and work experience) and probably more.

This points to a fundamental issue with academic lab safety that Ms. Benderly misses completely: because the students and workers are so inexperienced, the oversight from the university should be higher (or at least, as high) than industry, not lower. And yet, the situations are completely reversed. This is wrong and needs to change -- but she doesn't talk about it in her article. Instead...

Uh, yeah, that'll work: What does Ms. Benderly prescribe? That "NIH and NSF should insist - on pain of dismissal, like the chemical companies - that researchers follow all accepted safety procedures", that PIs and funding agencies cannot "ever again tolerate the risk of harm to lab workers", and that the provost on down must make safety "a requirement for career advancement and hiring... and tenure(!)" and last (but not least!), that funding agencies must make "a good safety record and evidence of safety awareness real conditions for getting and keeping grants."

Another big sigh -- some of this is really naive and some of this is weapons-grade nonsense. First of all, NIH and NSF cannot reach down into university labs and kick out that Bangladeshi postdoc who insists on running columns without lab glasses and side shields on. They don't do firing decisions at that level, nor do I think it practical to do so. Zero tolerance? Yeah, that'll work -- just ask Derek Lowe.

It's the last two suggestions that are really bad: that tenure and grants be awarded on the basis of a "good safety record". A "good safety record" is hard to detect -- the absence of a bad one is far, far easier to find or fake. If such a system were to be implemented, what would happen? That's right -- the PIs and graduate students now have strong financial incentives to keep sweeping near misses, close calls and outright accidents under the rug. Hell, profs would buy a bigger broom, if it meant keeping their jobs, grants and promotions!

Triple sigh: There's real work to be done to revolutionize lab safety in academic chemistry. There's a long-form article to be written on the Sheri Sangji case that will shed light on what happened that day (Kim Christensen has written most of it.) There's an investigation that should happen into the incident, by a government agency with real subpoena power. There's good commentary to be made. This article by Beryl Lieff Benderly is none of those things.


  1. I have to disagree with one conclusion. Many, many graduate students work in the DOE labs, and the number of accidents is far lower than in academic labs. Safety is taken as the primary concern, science is second. However, it means we get less work done.

  2. Interesting, Mitch. Why do you think that is? Is it just the emphasis on safety from higher ups (professors, etc.)? Do you folks have higher permanent staff ratios than a typical academic setup?

  3. A lab tech in Somorjai's lab placed isopropanol in a waste container of nitric acid. The bottle exploded and the lab director suspended Somorjai's pay for 3 months (the lab director, Alivisatos, is Somorjai's son in law AWKWARD...). Actions like that tend to decrease the accidents that abound in government labs.

  4. Yeah, that would do it. I'm impressed -- wouldn't you like to have been in on Thanksgiving dinner?