I learned two important lessons in a statistics course I took in college: 1) don’t play the lottery; 2) be skeptical of statistics, especially nice-looking charts and graphs.
Yes, I did grow up to be a major purveyor of statistics, but I think it’s a good thing that my first instinct is to be skeptical of the power of numbers when someone is telling a story.
That’s why I was disappointed to miss a recent presentation entitled, “Consumer Interest in Health & Genomic Information,” by Steve Bodhaine, a Yankelovich researcher, who appeared at an HHS-sponsored conference . (For more on the event, read Jen McCabe Gorman’s summary of a panel that included Esther Dyson, Matthew Holt, Linda Avey, and Rebecca Fisher).
Happily, I ran into Esther the next day as I walked down Connecticut Avenue to work. She dug into her bag and handed me the Yankelovich slides, which include the finding that consumers are “interested but skeptical” about how genetic testing is used.
Since I missed the presentation, I asked Esther for her review of the findings. Here is what she wrote:
My data, my self
Yes, people are woefully unprepared to understand this kind of data, just as they are unprepared to file taxes, handle a sudden request for a divorce, deal with a non-genetic diagnosis of cancer, respond to an SAT score that indicates a 20-to-1 chance of getting into Harvard and a 3-to-1 chance of getting into the local state university, etc. etc. But we don’t require that people hire professionals – even though mistakes are made. In health care, people are not required to get (nor does the system usually pay for) second opinions, even though they would reduce error and heartbreak. Life just *is* tough, and medical care is riddled with errors and dangers.
The best way for people to understand genomic data is to explore their own – ideally with competent, lucid explanatory information that is somewhat personalized. And ideally, they could easily find experienced genetic counselors to help them interpret the data. But even if they can’t, those who want to explore their own data (and pay for it) should have the opportunity to do so. It’s amazing how people can understand complex batting averages, race-track betting odds or stock-market statistics, according to their interests. I expect the same to happen with genomic information. Just give people who are interested a chance to explore (with credible information and context), and it’s amazing what they will figure out. Over time, people will understand exactly what you can and cannot glean from genetic data – and as more of it becomes available for responsible research, the data itself will become more meaningful. The first adopters are benefactors (well beyond the money they spend); later users will be beneficiaries. See “What You Can Do for 23andMe (and Future Generations)”
As for regulation, it’s absolutely appropriate to regulate labs and to vet sites for misleading information, spurious promises and fraudulent sales pitches. But I’d like to control my own data, please, and make my own decisions about whom to share it with, whether that’s no one, a doctor, a relative…or the whole world.
[Disclosure: I’m on the board of 23andMe, and I’m a research subject for the Personal Genome Project, for which I will publish my full genome and my health records, personally identified, on the Web.]
What is an e-patient to do when faced with questions about genomic data? One answer is to stay tuned to the statisticians who, as Sharon Begley writes in Newsweek, “are the new sheriffs in town.” Another is to read a memoir like the one a friend handed to me last week, “Pretty Is What Changes: Impossible Choices, The Breast Cancer Gene, and How I Defied My Destiny,” by Jessica Queller. What are some other possibilities (besides playing the lottery)?
Okay, I confess:
1. For some reason I’m not willing to have this post go by unfathomed.
2. But I’m not fathoming it. It’s over my head. Too deep.
I speculate that since you’ve been at this several years longer than I have, your thinking is way ahead of mine. Got a flashlight to share?
I mean, here I sit, thinking most days about merely changing healthcare to be more patient-centered, and you ask “What’s an e-patient to do when faced with questions about genomic data?” I think you skipped a grade and I didn’t.
Thanks, Dave, but I assure you that I haven’t skipped a grade. I’m just trying to shed light on an issue that is increasingly in the news.
For number nerds, the statistics provide one way to look at this issue.
A personal story is another way to put something complicated in context, so I’ll link to some more material about Jessica Queller:
“Unfortunately the test for the BRCA gene is just a decade old, and doctors can offer no definitive guidance to women diagnosed with a genetic predisposition to cancer. In the case of BRCA mutations, science has outpaced our understanding of what to do with the data. Because the test is unaccompanied by any clear medical recommendations, it doesn’t provide solace so much as open a Pandora’s box.”
I think more people are going to be presented with this Pandora’s box, as she puts it, so I wanted to highlight the issues for discussion.
Susannah,
The genetic conundrum hit home for me last week when my college roommate’s sister was diagnosed with Stage III ovarian cancer. These are women whose mother carried the Ashkenazi mutation, and died of breast and ovarian cancer. They are watched like hawks, so it was a terrible shock when the docs discovered that what they were sure were benign polyps were cancerous. “Mary” the patient, had never been able to bring herself to have genetic testing, so maybe she was a carrier all along. “Lisa,” my friend, did do the testing, and is not a carrier, but now wonders if she should have prophylactic surgery to avoid her sister’s fate. Tough, touch choices.
On another front, I was at my college reunion, and saw a good friend who is deeply into genetic research at NIH. I started to tell him about my experience in online patient communities, using as an example the scores of messages that get traded on my BC community about Xeloda dosing. For many e-patients, it’s only the knowledge they’re exchanging online that helps them persuade their docs that one size doesn’t fit all when it comes to dosages, at least of this drug. My friend looked bemused. From his vantage point, there’s no such thing as a single disease, such as “breast cancer.” There are hundreds of thousands of instantiations of the disease, depending on one’s genetic makeup. Of course one size doesn’t fit all — we’re headed for a world in which everyone may have their own, personalized treatment regimen.
In the meantime, though, I’m grateful for the wisdom of patients, which is helping to bridge the gap between the package insert and the era of genetic medicine!
Judy’s remarks are as always right on the spot.
Cancer e-patients already deal daily with the understanding that they suffer from a disease characterized by a truly unique set of mutations. While every one else is trying to build companies to quantify & classify the personal risk of individuals, members of online communities deal with the reality of the incredible diversity of genetic variations.
And sometimes by communicating with others in a seemingly casual way, they can ferret out new genetic connections between them and even sometimes come up with sufficient data to convey to researchers.
The cancer e-patients don’t have time to wait until personalized medicine becomes a reality. They have been dealing with the “in the meantime” for now close to 20 years. IMO, those who have been for some time in an active online community know better that most doctors that the side effects of their treatments can vary widely and that hopefully they can obtain help and support from some other patients who have experienced similar side effects. I will never forget the discussions about Gleevec, many years ago, when Gleevec was called STI-571. Patients started talking about chronotherapy (without ever mentioning the term or even knowing anything about it), while there was not a single entry in medline about chronotherapy and Gleevec.