Experience is not the Answer

Bradley Flansbaum

First posted 8/25/11 on The Hospitalist Leader

I assume, incorrectly perhaps, that mechanics have a basic knowledge of their craft such that routine auto repairs require little effort.  The tasks do not supersede the expected competency of the repairperson, and the customer can expect a car that operates at the time of pick up.  A small percentage of jobs may stretch that assumption, but that is okay by me.  Just like medicine, some mends are complicated. You need assistance from another mechanic or you refer the auto to a specialty garage.  No one is superman.

How does this relate to the practice of medicine?

I frequently notice pharmaceutical ads on Sunday AM television broadcasts, as well as newspaper articles that advertise a medical product, or report on a new device, surgery, or therapy—usually of the latest and greatest vintage.  As the data for these interventions is incomplete, or the costs unknown, the story concludes with a riposte conveying that the reader need not concern themselves with the alien facts—just “consult your health professional” and all will be well.

I also observe that politicians object to “meddling” when EBM-based policies from expert committees passively (or actively) affect the doctor-patient relationship, especially as it relates to decision-making and the counsel we provide.  Just watch the nightly news—sound bites abound.  This relationship is sacrosanct after all, and our advice is authoritative and 98.7% correct.  Who would question a physician after all?

Numerous articles document the poor health literacy of patients; their ability to understand prescription instructions, comprehend insurance policies, or reliably complete an informed consent form.  However, the literacy of physicians resides on a loftier peak.  Society assumes we know our business.

Nevertheless, consider the capacity of our profession to interpret basic clinical studies or unravel slanted industry messages.  Is society overestimating our ability?  Is our erudition reliable enough when patients seek our guidance for care?  More importantly, for those who contemplate this role, how it needs expanding in a patient directed system, and its authenticity, can we deliver?  After all, the strength of our health system depends on whether we perform this task—assuming congress requires us to execute it.

A new study in The Journal of General Internal Medicine examines both patient and physician ability to discern potential differences in how each interprets survival odds in a hypothetical case vignette.

Outcomes were expressed as follows:  [%] absolute survival, [%] absolute mortality, [%] relative mortality reduction, or all three simultaneously.

As children, teachers ask us if a pound of feathers or bricks weigh more.  No difference obviously, and likewise, the study tests whether framing, how we cast equivalent data, influences how caregivers and patients interpret information.

The following table, with notable highlights, answers the question.

Of note, the variance between physicians and patients in categorizing treatment effects (better to worse) is notably similar—regardless of how they viewed the information.  Of greater concern however, is the loss of physician’s contextual ability when presented with a relative risk calculation.

The doctors rate the treatment as clearly better two-thirds of the time when the investigators present a 2% absolute improvement in this way (RRR=33%).  The error is disturbing.  Also, while it was a Swiss study, I am comfortable extrapolating the results to our shores.  You might be a skeptic, but I expect identical findings here.

As an illustrative example, associate the findings above and apply that same physician health “illiteracy” to the interpretation of a seminal study recently released this summer.  The findings were available earlier in November 2010.

A journalist’s take:

The hospitals – including the University of Pittsburgh Medical Center, Swedish Medical Center in Denver, Abbott Northwestern Hospital in Minneapolis, Rhode Island Hospital in Providence and Pomona Valley Hospital Medical Center in California – say they are responding to the study by the National Cancer Institute. It  found annual low-dose CT – or computerized tomography – screening of asymptomatic current or former smokers could cut the death rate from lung cancer by 20 percent by identifying the disease earlier than X-rays.  The resultswere so striking that federal officials last November ended the study early to announce their findings.

Caution however, as that is not the final word:

How would you perceive the publication?  Would have reflexively ordered the screen based on the RRR of 20%, or rather, would you have noticed the ARR and high false (+) rate? Most importantly, do you have the ability to overcome the framing bias, or the capacity to defend against it when patients request a screen?

Not all that glitters is gold.  The numbers require analysis and thinking, and not of the passive sort (read the study, if for no other reason than it is influential, and patients will inquire about its relevance to their care).

On a personal note, I cared for an 83-year-old woman this week. Regrettably, I diagnosed her with metastatic lung cancer.  She had partial decision-making capacity and poor functional status.

Hospice was my first inclination, where we ultimately moved her.  However, while I do not regret the oncology consult, I opened a therapeutic door I was unable to close.  It was the system gateway I often (altruistically) rail against, and sometimes loathe.  Guilty as charged.

My reluctance to accept a recommendation, which I assumed was fruitless, produced thispaper from the pocket of the oncology fellow.  This study was my patient.

Tolerability.  Survival.  Cost.  Weigh the balance.  I was in the dark, and my first thought was the internal sounding board we all utilize: if this was my mom…

Politicians, society, the WSJ op-ed page, they all assume the verdict of treatment certitude is between the doctor and patient only–and the course we navigate is always correct.  However, take it from one who knows.  We need help.  We do not have the answers.  We are not that black box that will cure our system’s ills, and the bond between physician and patient needs more than just our expertise.

Did I give the drug by the way?  No, and it was a difficult decision.  That is the hat I hate wearing, and the one I feel so conflicted in.  It does not suit me.

Brad Flansbaum is Director, Hospitalist Services at Lenox Hill Hospital in New York City. He is a founding member of the Society of Hospital Medicine and has served as a board member and officer.