Lies, Damn Lies, and Statistics

Back when the dinosaurs roamed the earth and I was in college, I studied the field of political science.  In both undergraduate and graduate school my areas of specialization were public opinion, voting behavior and polling.  I worked with the legendarily brilliant Jack Citrin, still at UC Berkeley and now Executive Director of the Institute for Governmental Studies, of which I am currently Vice Chair of the Advisory Board.  It has been a very long time since I spent my days analyzing polling research in an office overlooking People’s Park in Berkeley while listening to the dulcet tones of very stoned people enjoying the Grateful Dead.  However, the thing I remember as vividly as any other thing I learned back then was how fluid statistics can be and, in particular, how question formulation and question order were as much to blame for survey research results than were the actual opinions being collected.

I bring this up because I can’t attend any party, board meeting or water cooler conversation these days without someone bringing up whether or not I or they think  that the Supreme Court will overturn the healthcare reform law known as the Patient Protection and Affordable Care Act (PPACA).  This is always accompanied by a discussion of what percent of Americans love or hate the law.  I even overheard some otherwise normal looking people speaking about this in the line at Safeway this morning as they unloaded their beer, ice cream and cookies onto the checkout counter (irony alert:  these may well be the people most likely to use our healthcare system in the future).  Everybody has an opinion on this topic, it seems, and most of those opinions are not particularly well-informed.  Even if they were, it doesn’t really matter as Mr. beer and cholesterol is unlikely to be called by Chief Justice Roberts to weigh in on the opinion.  Nevertheless, it brought me back to that whole survey research topic (the contents of that grocery cart would have gone particularly well with the herbal tonics no doubt still found in People’s Park).

Back in March 2012, the Pew Research Center published a review of a variety of recent polls, including their own, about public support for the (PPACA).  According to the article, entitled Public Remains Split on Healthcare Bill, Opposes Mandate:

“…a review of recent polling finds that basic assessments of the law are as divided as they were when the law was passed two years ago. In the latest Pew Research Center survey conducted March 7-11, 47% of Americans approved of the health care legislation passed by Barack Obama and Congress, while 45% disapproved.  This mirrors the balance of opinion in the weeks after the bill was passed. In April 2010, 40% approved and 44% disapproved of the law. Five other major surveys conducted over the past month find similar divisions of opinion, and with little change from two years ago.” 

Bottom line, this set of surveys found that people in the country are equally divided about the utility and value of the PPACA when they are asked very general questions about it.  However, when you start digging in and asking them more specific questions about what should be kept, what should be changed, whether the law should be amended or repealed, etc., the responses get particularly murky.  For instance:

“Fox News survey found 59% in favor of some kind of repeal, either repealing the law entirely (31%) or repealing parts of the law (28%). Yet the Kaiser survey found just 41% in favor of repeal; 23% said they favored repealing the bill and not replacing it, while 18% favored repealing it and replacing it with a Republican-sponsored alternative. Unlike the Fox News poll, the Kaiser survey asked respondents about what may replace the law if it is actually repealed.  Even the number of Americans who support making no changes to the Affordable Care act can vary substantially depending on the question. The Pew Research Center’s March survey found little support for the status quo: Just 20% favored leaving the health care law as it is, with most supporting either repealing (38%) or expanding (33%) the law.  Yet another survey conducted at the same time found more than half favoring the status quo when phrased differently. The March Bloomberg News poll found 46% saying it may need small modifications, but we should see how it works, and another 11% saying the law should be left alone. In this question, 37% favored repealing the law.”

As Mark Twain once said, “There are lies, damned lies and statistics.” Nothing could elucidate that thought better than the preceding survey research polling results showing that either 37% of Americans want to repeal PPACA or 59%, depending on how you ask the question and whether you include questions about what should replace it.  In a democracy where 50.1% of the ayes have it, that is a big swing.  In another great quote, humorist Evan Esar once said that the definition of statistics is, “The science of producing unreliable facts from reliable figures.”  He also was quoted as saying, “Statistics:  The only science that enables different experts using the same figures to draw different conclusions.” How true it is.

This is a critical issue, of course, because of the damning statistics we all know about healthcare costs in America.  If you take a good look at this recent Atlantic article, entitled 10 Ways to Visualize How Americans Spend Money on Healthcare, you can see very poignantly the issues that need to be addressed.  This is a very compelling article because it puts into pictures some of the critical public health and healthcare system issues that we face as a nation.  Among the stories these pictures tell are:

• The top 5% of spenders account for almost half of all health care spending.
• The top 1% of Americans spends $90,000 per person on health care, 381 times more than the bottom 50%.
• Hospital costs are rising more than twice as fast as wages … but those only make up about a third of total health expenditures.
• Due to advances in science, cardiovascular related deaths per capita have declined by 80 percent since 1950. (this is the token good news, but of course it means that people with heart disease will go on to be old and spend lots of money on healthcare rather than dying cheaply—again, life is full of irony)
• The top 1% of health care spenders make up 20% of all health care spending … and 2/3rds of them are older than 55.
• Our long-term budget crisis is nearly entirely a crisis of government health care spending, which is overwhelmingly in Medicare and Medicaid.

That last two are particular significant in my view.  Since over 10,000 Americans per day are turning 65, we should be very afraid about the manner in which costs will continue to rise in the U.S. if no change is made in the way we access and pay for healthcare.  As for the last bullet above, just a look at this chart should make you worry for your future, whatever your political proclivities.

THE LONG-TERM MEDICARE/MEDICAID CRISIS 

Personally I think there are both good and bad things about the PPACA, but if it is struck down we had better come up with something better than a Republican victory party in response.   And that’s not even a dig at Republicans.  The Democrats don’t have a lot to be proud of here either.

By the way, here is my favorite healthcare statistics quote from American political fiction writer Fletcher Knebel, “It is now proved beyond doubt that smoking is one of the leading causes of statistics.”  Knebel died of cancer–go figure.

It seems to me that the only people better than using statistics to their advantage than those in the healthcare industry are economists.  I love this article entitled The 29 Most Bizarre Economic Indicators In the World, which shows different statistical methods for determining the health of various world economies.

My favorite indicators in this article are these:

• The Super Bowl Indicator, which statistically demonstrates how financial markets will post gains for the year if a team from the old National Football League (NFC division) wins the Super Bowl.
• The Sports Illustrated Swimsuit Cover Indicator, which measures that correlations between an American being chosen to grace the Sports Illustrated Swimsuit cover and the odds that the S&P 500 will outperform historical returns.
• The Consumer Consumption of Beer Index, which statistically demonstrates that when consumers spend less time and money drinking at pubs there is an impending drops in jobs.
• The Baby Diaper Rash Indicator, which shows that parents worried about the economy will change their babies’ diapers less often to save money, sending diaper sales down and diaper rash cream sales up.
• The Men’s Underwear Index, which shows the statistical correlation between men’s decision to forgo purchasing new underwear and a weakening economy.

There are some other good ones too—funny reading.   I would love to see the healthcare industry come up with some similar indices for predicting healthcare inflation trends.  God knows our field needs a sense of humor.