Health Column: Why doctors make mistakes
MOHLER’S MEDICATION MAXIMS
Free Press Health Columnist
It happened almost 25 years ago, but the memory still evokes a chill. We had just sat down to supper and the phone rang — the ER.
“Dr. Mohler, your patient is headed on to the operating room,” said the voice on the other end of the line.
I had seen Jim, a patient for 10 years, earlier in the day. He had flank pain, blood in his urine, and vomiting. I diagnosed a kidney stone, just as I had twice previously with Jim. On that day, I gave Jim pain and nausea medication and a strainer to catch the stone. He was feeling much better when he left the office. Jim’s pain recurred and he went to the ER where a CAT scan revealed a ruptured aneurysm. Four hours and 12 units of blood later, Jim left the operating room. He survived.
In 1999, the Institute of Medicine published a paper — “To Err is Human.” It estimated that medical errors result in 49,000 to 98,000 deaths in America annually. Among more than 2,300 patients randomly discharged from 10 North Carolina hospitals, there were 25 likely medical harms per 100 admissions. Of that, 8.5 percent of the issues were life threatening and 2.4 percent of the harms contributed to the patient’s death.
Most of the diagnostic errors we physicians make are not due to incompetence or inadequate knowledge, but to frailty of human thinking under complex, time-pressured situations. In medical school and residency, physicians are taught to use an analytic approach to diagnosis: slow, conscious, deliberate and resource intensive. As physicians mature, their approach to diagnosis becomes more intuitive; it is often subconscious and fast. This latter approach involves mental shortcuts. These shortcuts are indispensable for physicians in getting through the day. When they are effective, the shortcuts save lives and are efficient. When they fail, they may lead to grave errors as in Jim‘s missed diagnosis.
There were at least two shortcuts that led to my mistake with Jim. The first shortcut that failed me was “anchoring.” Anchoring means jumping to a conclusion too quickly. I fixated on specific features of Jim’s illness too early in the diagnostic process; it was a textbook kidney stone! Even more deadly was the “availability” shortcut. Jim had experienced two previous episodes of CAT-scan documented kidney stones. His kidney stone diagnosis was available, too available, and I quit thinking about the diagnosis.
The other factor that played a role in Jim’s misdiagnosis, as it does in many clinical mistakes, was overconfidence. If you had asked me when Jim walked out of the office that day, I would have claimed certainty of my diagnosis. Overconfidence is a trait of human nature. In a 1997 survey of academic professors, 94 percent rated themselves in the top half of their profession. Only 1 percent of automobile drivers rate their skills below that of the average driver.
Other heuristics abound! A particularly potent shortcut that is common to any problem solving is confirmation bias. This is the tendency to look for data to support our diagnosis, rather than to refute it. We have a tendency to ignore history and lab that does not fit the “picture” that we are painting. Confirming evidence feels good. Refuting evidence undermines the diagnosis. We have to start over, which leads to more work.
This column today may arouse thoughts of “doctor bashing” or a call to abandon traditional medicine. My intent is neither. The concepts outlined here are presented in a book written for the public several years ago by Dr. Jerome Groopman, “How Doctors Think.” I recommend it highly.
Dr. Groopman’s goal, as mine, is to empower patients to get the very best from their physicians. Here are four questions you can ask to help your physician make the correct diagnosis:
1) When your doctor is uncertain about the diagnosis, and even more importantly when he/she is very certain, it is fair game to ask, “ What else could this be?” This question brings closer to the surface the reality of the uncertainty in medicine.
2) “Could I have more than one problem?”
3) “Have we looked hard enough?”
4) “Is there anything in my history or exam or lab that seems at odds with the diagnosis?”
Free Press health columnist Dr. Mohler has practiced family medicine in Grand Junction for 39 years. He has a particular interest in pharmaceutical education. Phil works part-time for Rocky Mountain Health Plans. Email him at firstname.lastname@example.org.
Start a dialogue, stay on topic and be civil.
If you don't follow the rules, your comment may be deleted.