How to Play the Numbers

There's an alarming post making the rounds on Facebook. Rather: an ALARMIST post, but what's the diff. A piece was published on Mother Jones with the scary title:

"DID DRINKING GIVE ME CANCER? The Science on the link is clear...

Medical journalism is a minefield, sometimes I avoid it altogether when such claims are made. The science on the link is clear? We'll see about that. Often the conclusions are misguided at best and outright lies at worst. It takes an awful lot of deeper digging to try to decode the studies first-hand, and ain't nobody got time or patience for that.

Since I worry about my own drinking — while steadfastly certain it had no contribution to my mother's cancer or her mother's death from cancer — I read the article. Or at least I read most of it until I got into the section where the author start to Play the Numbers Game.

Researchers estimate that alcohol accounts for 15 percent of US breast cancer cases and deaths—about 35,OOO and 6,6OO a year, respectively. 

OK, I'll bite. I clicked the linked number for a first-hand view. This is the first thing I read:

Results. Alcohol consumption resulted in an estimated 18,2OO to 21,3OO cancer deaths, or 3.2% to 3.7% of all US cancer deaths. The majority of alcohol-attributable female cancer deaths were from breast cancer (56% to 66%), whereas upper airway and esophageal cancer deaths were more common among men (53% to 71%). Alcohol-attributable cancers resulted in 17.O to 19.1 YPLL for each death. Daily consumption of up to 2O grams of alcohol (£ 1.5 drinks) accounted for 26% to 35% of alcohol-attributable cancer deaths.

I cannot for the life of me figure out where the article's author got the "15 percent of US breast cancer cases and deaths" citation. That's not what the results say, the author has just lost me. I'm not interested in studies or reports of studies that Play the Numbers Game. Sometimes journalists and scientists Play the Numbers Game to land on a desired outcome rather than uncover truth as they find it. They manipulate the numbers, conflate correlation and causation, and present foregone conclusions that the science doesn't necessarily prove. Let's pull this one apart.

The study cited in the Mother Jones article uses "population-attributable fractions," meaning they look at established statistics and correlate them formulaically to make estimations. For instance, if you say 5,OOO out of every 1O,OOO Americans subscribe to Netflix, and 25% of all Americans have blonde hair, then you could try to estimate the number of blonde Netflix bingers at any given moment. I'm not sure why you'd want to, and you'd have to be a much better mathematician than me to crunch such a number, but my point is you could correlate it. You can correlate anything. The trouble with this is two-fold: you miss an awful lot of nuance in simple correlation [Are they bottle-blondes? Are they really binging or just subscribers? Did THEY like Bandersnatch?]; and as the wise lead singer of mid-9Os alternative band Soul Coughing taught me over and over again, Correlation is NOT Causation.

When we read scientific studies that aren't lab-based and strictly structured (i.e. feeding pure alcohol to identical lab rats and waiting for cancer), we must remember that there are infinite factors that contribute to outcomes. We cannot put alcohol consumption next to cancer rates and conclude causation, no matter how fancy our formula is. It's a HUGE leap to see alcohol-use correlated to cancer incidence and forget that people who drink a lot may smoke outside the bar they frequent, or eat a lot of fried jalapeño poppers with their beer, or don't get their daily allotment of steps as their Uber goes from wine-tasting to wine-tasting. [Yeah, I live in wine country.]

Diet, exercise, and smoking are all contributing factors to one's risk of cancer, and each is correlated to alcohol use in its own way. There are also increasing insights into genetic mutations that are at play underneath it all. You can't design a study on people that only looks at one factor, we are not petri dishes of self-contained cells. We are unique and complicated, with genetic makeup and environments that shape and react to different factors in different ways. Maybe alcohol only makes an impact on heavy smokers who don't eat broccoli roasted in pure pink Himalyan salt, who knows. The cited results don't explore this.

The article author also posited "American women have about a 12 percent lifetime risk of getting breast cancer" while "a woman who consumes two to three drinks a day has a lifetime risk of about 15 percent—a 25 percent increase over teetotalers."

While it may be true that 15 is a 25% increase of 12, who looks at 15 and 12 and doesn't see a difference of three? Twenty-five percent is much more alarming than three points, but that's not my dog whistle. To me, that's a three-point increase. A skosh, a baby-step, a mild nudge of the needle. My genetic counselors would have yawned at that three-percent increase.

Doctors play a Numbers Game with different rules. Oncologists and geneticists are as obsessed with numbers as insurance agents clutching their actuarial tables, but their frame of reference is slightly less payout-concerned and more individually-focused, which presents its own challenges.

Doctors know that some cancers grow slowly and pose no immediate death threat and never will. Some cancers are fierce and fast and need to be fought with every tool in the oncology arsenal. But they don't always know how to tell which they are dealing with in any given patient; breast cancer is either deadly or it's super-chill, it's a toss-up. One could say it's a 5O/5O chance, which are not great odds. So doctors Play the Numbers Game with general statistics until they get each individual patient's risk as low as they can get it on their little chart.

My first oncologist, Claudia Tellez, was a lovely young woman who warmly walked me through my diagnosis and made clear recommendations for treatment. I felt safe with her — I'm a sucker for a doctor that makes gentle physical contact — but I was put off at first by how her answers always came down to math. "Your chance of recurrence with just a lumpectomy is forfty percent, if we add in chemo it goes down to twelvty percent. A month of radiation reduces that risk to zeroteen percent. We don't recommend mastectomy, it has no affect on survival rates with your pathology."

These are fake numbers because I still have no idea of any of my risk percentages, then or now. After almost 2O years of cancer treatment, several consultations with geneticists, and countless discussions with oncology doctors of every stripe, I maintain the deeply-held belief that I have a risk of getting cancer again. It could be 3 or 9, it could be 15, it could be 77. I don't know. It kinda feels like a toss-up: cancer either will revisit me or it won't.

I know not everyone is as whatever about their risk of cancer. All of us who face cancer, or simply fear it, are playing our own numbers game. Usually it comes across as a basic equation: Yoga x 4 times a week + 5 fruits + organic vegetables - booze - smoking - meat + vitamins - cleaning products = No Cancer. Sometimes we increase one factor, like exercise, so we don't have to subtract another, like alcohol. We all calculate our own equations.

Cancer patients then do different math with treatment options. Google "people getting more surgery than they need" and you'll see reports about women who opt to remove a healthy breast along with a diseased one, despite it not being recommended by a doctor, despite it not demonstrably changing their recurrence or survival numbers in the positive. We do it because it feels better, it feels like we are fighting harder, like we have the upper hand. We do it for our quality of life, for our sense of self, for our peace of mind. The numbers ultimately go out the window and we make emotional choices.

Maybe it's the two-time survivor in me, who knows that a diagnosis is not necessarily a death sentence, despite the ominous threat. Or maybe it's the ongoing internal battle I have to cede control over things I am not certain I can control. I don't want to live a life mistakenly endeavoring to change the things I cannot change. Cutting alcohol out of my life will not certainly diminish my personal own risk of cancer. It most certainly would diminish my quality of life, 'cause I love me some buttery Chardonnay. I don't want to deprive myself of small pleasures and soberly get cancer later on. I don't want to Play the Numbers Game.

I prefer word games or cards. Anyone up for some Gin?