bad science

In my last post, I noted that there were, among the projects, a number of studies of how family history affected risk of various diseases. These were built around some methods that Liz and I developed for querying genealogical databases, primarily the Utah Population Database, fondly referred to as the UPDB. The UPDB provided huge numbers of connections between people over multiple generations, but was hit-or-miss with the disease data that was needed to connect people with the same condition. The cancer data was excellent, thanks to longstanding connections to the statewide cancer registry. Heart disease, stroke, diabetes, Alzheimers, and many (most) others, though, were only available with the help of clinical enterprises, most of which had financial and competitive interests that frequently impeded research.

Over time, we developed some approaches that allowed for reasonable conclusions to be drawn about smaller clinical datasets, and a steady trickle of physicians brought data to us for analysis. Among these were a number of fine physician-scientists with whom it was a pleasure to work.

Not all of them, however. Shortly before we left Utah for Louisville, I worked with an ophthalmologist who had a fairly typical clinical series of patients with age-related macular degeneration, a leading cause of visual impairment and blindness among older folks.

As was the norm, I identified the relatives of his patients, including which of them were also affected, and computed the relative risks (compared to controls), and the usual supporting statistics, wrote the bulk of the Methods and Results sections of the resulting paper, and contributed commentary about the wider implications. For some reason he felt my contribution was not worth noting, however, and he sent off his draft, got it accepted and published, without my name on the author list.

If you're not listed as an author, of course, you don't participate in any correspondence with the journal or its editors. Nor is it the practice of any university oversight committee to query all the faculty not listed as authors to make sure no one was left off the list. So I didn't see this paper until it came out. I wrote a strongly-worded email to this ophthalmologist (he's the senior author--the one listed last), cc'd to higher-ups at the university, and extracted a promise to revise the online publication to correctly reflect my contribution. The principle mattered more to me than this particular publication, and I promptly forgot about the entire incident.

Until I saw this headline. It seems that what goes round comes round, at least if you go far enough around. I especially liked this line from The Scientist: "Zhang’s poor record of disclosing conflicts of interest accompanies a sketchy past of putting patients and research participants at risk during his work." 

That and apparently being some kind of spy. In comparison, my beef about authorship is pretty small. But I think the view into his character afforded by that minor incident was clear enough, and I was not at all surprised that someone who can so easily step on a colleague has no compunction about abusing his patients.

[By the way, it's clear in the first link above that he lied about contacting the journal. Saves me the embarrassment of having co-authored a paper with him.]

Projects

The reality of my projects has been called into question--recently, but also on many prior occasions--and, as a more-or-less recently retired professor, I am here to defend them.

In academics, as in life, projects can take many forms. For me and many others in biomedical research, the classic pattern has been: 1) develop an idea; 2) do some preliminary research at little or no cost; 3) if the idea seems to have some support, try to get funding to do a larger-scale study; 4) publish the results; 5) iterate steps 3 and 4 for as long as possible, perhaps an entire career.

For the record, you can see how this looks in my case by clicking here. The number of publications is less important than the number of citations, which indicate that someone has read your paper and thinks that the research that they have done is in some way indebted to the previous work described in your paper. They may disagree vehemently with your conclusions, but they can't ignore your work. Scientific consensus arises from scientific argument; nothing gets you ahead in your career faster than disagreeing with the conventional wisdom and actually being right. Actually, "faster" isn't the right adjective. "Farther" is better. Lister, Mendel, Gallileo, McClintock, and many others did not achieve success in the short term. (The trick is that the conventional wisdom in science is very often right.)

I've started, finished, and contributed to all kinds of scientific projects over the years, ranging from assessments of the degree to which fallout from nuclear weapons testing caused cancer and other diseases in Utahns living downwind, to (many) studies of the utility of family history as a predictor of disease risk, to studies of molecular markers of aging . . . and numerous other projects inspired by collaborations with physicians, fellows, and students over the years.

The majority of projects don't get past step 1 or 2, but that doesn't mean they don't have merit. In fact, the early stages of putting together ideas and data, doing preliminary statistical analysis and seeing early results is usually more fun and rewarding than writing up papers for publication, preparing grant proposals, and even getting an award and making the dream project a reality.

At present, Liz and I are working on preliminary data and writing up results for several projects that have been or will be turned into grant proposals. We're doing this pro bono, which gives us the freedom to pick our spots and keep our priorities straight.

Of course I've had lots of projects over the years that didn't have anything much to do science: co-raising children, managing finances, house and yard maintenance, travel, pointless tinkering with electronics and software, etc.

Being a cancer patient is also a project--make no mistake: there are daily, weekly, monthly, and annual cycles of medications, blood tests, X-rays, and consultations. Every month requires a new prescription, a telephone order complete with nurse consultation, a check of the insurance status for said order, and a consideration of any travel plans for the next month, since a street address and delivery signature are required. The process is not especially hard or complicated, but it is always front and center and cannot be ignored.

And then, sometimes, I write about my experiences and associated thoughts. At Christmas I took all the posts from the beginning of this blog and formatted them into a book, to give to my mother and others who weren't going to be reading blog posts. The result was a slender volume of about 100 pages--not monumental, to be sure, but a satisfyingly tactile indicator of having done something.

Dog raising is another project. I'm pretty sure that it becomes less of a project as your dog becomes better trained. Suffice it to say that Sunny is still a small dog, but a big project.

And we've added a new project/family member: a tiny teardrop trailer ready for road trips. Mostly it won't require much work on our part, but I will need to improve my trailer-backing skills.

N of one

It turns out that N-of-one trials are a thing. Back in the early days of this blog, I described the basic idea and naming principle--that my experience with treatment would be its own experiment, controlled to some degree by my pre- and post-treatment experiences.

I have been carefully recording many variables, but not quite with the rigor that would be required to do a formal trial. I've been focused on labs, and thinking about exposures, but haven't been good about recording exposures (diet, alcohol, probiotics, dog, travel, weather, etc.). I expect that, soon, N-of-one trials will answer lots of questions about how individual variation (in genetics, epigenetics, microbiome) interacts with exposures to make people sick (or not).

But the labs are good! Here's a picture, very much like the last (many) pictures, but produced by a website, all4cure.com, started by an oncologist who found himself diagnosed with multiple myeloma. The data transfers are definitely not as easy as they could be, but it's working, to what effect I'm not quite sure. The graphs are fairly nice, though.

I don't know how long this particular drug regimen will continue to be effective--no one does, because I was an early adopter. If Medicare and marrow continue to support it, I am not going to complain.