November 20, 2009

Risky Business

I have been thinking about risk of late, in a variety of contexts.  It seems like we are surrounded by the concept.  Listen to the news (or read it, if you're so quaint) and you will hear numerous stories on the government trying to reduce the risk to its citizens.  We invaded Iraq to ostensibly reduce the risk posed to U.S. citizens by that country, and by the terrorists that were supposed to be scheming there.  We are trying to establish various and sundry new financial measures to reduce the risk posed to the economy by rogue institutions and individuals.  We are inundated with advertisements for cars with more and more safety features, not to mention larger and heavier cars by the year.  This has contributed, in part, to the fact that fuel efficiency hasn't improved much in decades (Julia's '89 Corolla got 40+ mpg on the highway, comparable to the Prius we own now).  The latest risk that must be eliminated is being pitched to me by our church.  We may spend millions to build a new childrens'/education wing in part because the old one has security flaws (as well as a perceived lack of space, aesthetic appeal, etc).  My struggle to see the necessity in this is a post unto itself, but I found it odd that one of the main points was that we need greater security in this one part of our church.  Maybe I shouldn't have been surprised as a post 9/11 mentality permeates all aspects of life now.  But call me na├»ve, I was.

And it got me thinking, how safe can we really be?  How much risk can we really avoid?  I don't know exactly what security measures the church thinks we need in place, but I'm guessing there will be technological solutions involved.  Badge access, cameras, etc, etc.  I'm a doctor.  We're pretty obsessed with evidence to support promised outcomes these days.  Is there evidence that risk-reducing measures work as advertised?  Yes and no.  If you evaluate populations already at risk, measures can be effective in reducing that risk.  Condoms used by people already engaging in risky sexual behavior do in fact reduce their risk of acquiring HIV.  But what is often not discussed are the unintended consequences of enforcing such measures. Is there a behavioral compensation when we know, or at least believe, we are safer?

This is of course a loaded question.  There are always unintended consequences of broad policies, and most of the examples I'm talking about affect large groups of people.  Unintended consequences are essentially unavoidable in complex systems when you don't fully understand every aspect of the system. This is why it's so infuriating to me to hear politicians try and oversimplify issues and assure people that their policy will absolutely work, or alternatively that there is no way the opposition's policy could possibly have any benefits.  But I digress.

Are we safer because our cars have airbags, or because seat belt laws have been passed, or because we're using condoms?  Not definitively.  There is no clear evidence that condom use promotion has decreased HIV transmission.  Countries with high uptake of condoms (Thailand for example) do show a drop in the rate of transmission.  But countries with a low rate of condom uptake (Uganda) show a very similar decline in the rate of transmission.  The argument is that all epidemics will peak and decline as vulnerable populations become saturated, and avoidance measures are undertaken by those not most at risk.  At some point a steady-state should be expected.  This is in no way to suggest that condom use doesn't work.  But the benefits of mass campaigns to promote condom use are difficult to prove.  Furthermore, there is some evidence that people compensate their behavior when they feel safer in regards to HIV.  Otten et al showed that among patients that had a positive HIV test and received counseling there was a decrease in further sexually transmitted infections.  However, in those who tested negative, the rate of STI doubled when you compared the six months before their negative test to the six months after. 

In the '70s and '80s when seat belt laws were being passed in countries around the world, people were extolling the thousands of lives that would be saved when everyone was buckling up.  However, the data show that such predictions never really came to bear.  Consider this figure from an opinion piece in the Lancet in 2000 (ref here):

Of the 17 countries with the most cars, death rates fell in all of them, whether they enacted a law or not.  In fact, the four countries that didn't enact a law had a more pronounced drop in death rates.  The authors speculate that what was really driving the death rate down was the economic downturn and energy crisis (people driving less, and driving at slower speeds to conserve fuel presumably) and a movement to reduce drinking and driving that was afoot at the same time.  What's perhaps even more interesting, is that death rates among those not in cars rose.  The death rates among pedestrians, bicyclists, etc went up.  "In the 23 months that followed the introduction of the UK seat-belt law, the number of deaths among pedestrians, cyclists, and unbelted rear seat passengers rose by 8%, 13%, and 25%, respectively." Why?  People wearing seat belts feel safer, and their sense of individual risk is reduced.  In turn they may be inclined to drive faster, drive more, or be more daring on the road.  Peterson and colleagues (article here) studied the effects of airbags in the early '90s as they were becoming more standard on cars.  They found that people in airbag equipped cars tend to drive more aggressively, and multi-car crashes were more likely to be initiated by the car with an airbag.  In single car accidents people in airbag equipped cars were no less likely to die than people in cars that lacked them.  Passengers were more likely to die if the car had an airbag (remember that this is before passenger-side airbags were common).  Again, if the driver feels safer, everyone else had best watch out. 

James Hedlund suggests in a lecture (which is fascinating and I encourage anyone who finds this even remotely interesting to read it) that if you want to reduce risk without creating compensatory behavior, you need to aim for measures that "are invisible to people, or that do not affect their actions or attitudes, or for which they have no motivation or freedom to change behavior."

I have to wonder what actions of my own are compensatory in nature because I do not perceive risk.  What industrial or governmental policies are we just waiting to see the compensation for as everyone tries to make our lives safer?  And in the end, just how "safe" do I want my life to be?

 "Playing it safe will always end in disaster"