Lies, Damned Lies, and Statistics
Would you be nervous if I told you that using a phone doubles your risk of contracting a particular illness? What if I told you that the chance of getting this condition is one in a gazillion? To most, it might seem absurd to give up your phone just to cut your risk to two in a gazillion, but certainly there are others who would call for a federal ban on phone usage.
Patient access to the Internet continues to confuse a well-intentioned but often unlearned public. One of the biggest misunderstandings results from ignorance of absolute versus relative risk. Patients frequently are fearful of a particular treatment, “if it increases my chance of developing some rare condition by 5%.” Reported side effects of medications greatly worry my patients, who unfortunately have little knowledge of comparative placebo-controlled trials.
In fact, daily activities that we hardly think of day to day confer much higher risk. Yet dramatically better odds, when referring to medical risks, have an amazingly different effect on decision making.
Let me give you an example. Did you know that statistically speaking, walking a mile is 19 times or 1,900 percent more dangerous than driving a mile in a recalled Toyota?
Now let me show how statistics, when expressed differently, can have a profound effect on perception of health and disease.
Several years ago, a 62-year-old male patient of mine came in with the chief complaint of “something personal.” Since we’re dealing with statistics today, 90% of the time this translates into, “I’m here to get a prescription for Viagra”. Sure enough, this gentleman was highly distressed that the frequency of sexual intercourse with his wife had declined “by 50%” in the past several months.
Seeking additional history, I asked him what this translated to in encounters per week. “Well you know how it is doc, like everyone, the Mrs. and I usually had sex every night before we went to sleep and of course again every morning when we got up”. Dejectedly he added, “Now we just do at each night”.
I’m not joking when I tell you that he actually waved his hand in front of my speechless and disbelieving stare. “Doc, doc… are you listening to me? Did you hear what I said?” Regaining my professionalism, I replied, “I’m sorry, for a minute there I thought you said that you’re sex life has gone from 14 times per week to seven.” “Exactly!” he replied.
I explained to him that half of men over age 60 reported having sex once per month (admittedly a statistic out of context). He boldly strode out of my exam room with a new sense of youth and virility.
I’ve taken the liberty to assemble some (not so) related statistics. I leave it to you to draw your own conclusions from the data.
- 1: 4 – incidence of orthopedic surgeons who will cut the wrong limb at some point in their career
- 1:1000 – incidence of general pediatricians who will operate on the wrong appendage
- 1:2 – ratio of average salary of pediatricians compared to orthopedic surgeons
- 278 – number of bacteria colonies that are exchanged during a kiss
- 45 – percentage of tongue piercings that show signs of infection
- 2 and 63 – percentage of staph infections resistant to methicillin (MRSA) reported in 1973 and 2004, respectively
- 8 – percent increase in average hospital cost for unhelmeted motorcyclists involved in crashes compared to helmeted riders
- 8 – average percent savings on six pack of Pabst Blue Ribbon during ‘Bike Week’ promotional event
- 100 – number of cases of diphalia (two penises) reported, ever
- 50,000 – number of women in the US with didelphic uterus (two uteri and often two vaginas)
The power of analytics that will result from increasing codifiable EMR data will no doubt markedly transform medicine. Physicians themselves will have increasing difficulty comprehending this information without new approaches to reporting and representation.
It is our patients, however, who will be poorly prepared for this new (tidal) wave of information. I shudder to think how Personal Health Records will compound the problem.
Joel Diamond, MD is chief medical officer at dbMotion, adjunct associate professor at the Department of Biomedical Informatics at the University of Pittsburgh, and a practicing physician at UPMC and of the Handelsman Family Practice in Pittsburgh, PA. He also blogs on interoperability.