When I talk to other doctors about epilepsy, I try to offer a few bits of my own experience in diagnosing the condition and caring for patients. I find that my own shortcomings are a good source of "pearls," and examining them reminds me not to fall back on bad habits. Examining past mistakes is a good habit, one that doctors in general do not practice often enough, at least publicly. We seldom discuss our mistakes. They are denied, repressed, or whispered about in corners. We rarely publish accounts of our spectacular misdiagnoses or therapeutic failures. Published articles on new drugs or surgical procedures highlight success and safety. It is a struggle to find a single article that focuses on failures. That is a pity, because mistakes and failures are highly instructive.

One especially common and dangerous error that we doctors are prone to is trusting a prior diagnosis. For a doctor, a diagnosis is like a child. Once we create it, our ego gets involved and we feel pride and ownership. We think that it can do no wrong. Enormous forces prevent us from recognizing any error or limitation. These forces (failure to look, bias, rationalization) lead us to reject any evidence that suggests the diagnosis may be incorrect while we eagerly look for and magnify evidence that supports our view.

Even worse, once one doctor has made a diagnosis, other doctors nearly always accept it as correct. Initial errors in gathering information and putting it together are repeated, over and over. If the first doctor failed to ask an important question, disregarded something the patient said that seemed irrelevant, or overlooked the result of some laboratory test, watch out -- later doctors are even more likely to miss the same information.

Situations that are unusual and beyond our knowledge base lead us to err most gravely. Recently a colleague -- a neurologist -- asked me to see her uncle. For nearly 3 years, he had been feeling episodes of pins and needles in his arms and legs, attacks of weakness in his right hand, and migraine-like headaches. Lately he had had a few tonic-clonic seizures. Several internists and neurologists had examined him over the years. The first neurologist, at a major academic medical center, asked him for details about the pins and needles and found that their pattern and timing did not fit a diagnosis of a pinched nerve, minor stroke, multiple sclerosis, seizure, migraine, or any other neurologic disorder. Electrical studies of the nerves, MRI of the brain and spinal cord, and blood tests all seemed normal. The doctor concluded that the man had a psychological disorder (although he was never quite told this).

A few months before I saw him, the symptoms became worse. Another neurologist repeated some tests but found no answer. This supported the idea of a psychological cause. Then the man had a major seizure. He was treated with Dilantin, but when he missed a single dose, he had another seizure and was taken to an emergency room. A blood test was performed and showed an elevated platelet count.

When I saw this man, a short time later, I had not yet received his records from the other neurologists. Other doctors' notes can be very valuable because they give information that was obtained closer to the time that the symptoms actually occurred. Thus they may be more accurate than data collected through the foggy window of memory. But at this time the records were not available, so I asked him to tell me his story again in detail.

As we finished talking, there was a nagging sense that all of these things fit together -- the pins and needles, hand weakness, headaches, seizures, and maybe even the high platelet count. (An important principle in medicine is called Occam's razor. It says that if multiple symptoms [patient complaints] and signs [evidence from examination or laboratory studies] develop at about the same time, a single diagnosis that unites them is more likely to be correct than several separate diagnoses.) It turned out that all of this man's symptoms, including the seizures, resulted from a very elevated platelet count.

At first, I didn't consider this possibility, because even if the high platelet count was causing symptoms now, how could the symptoms 3 years ago have been caused by the same thing? The blood test results then were normal. My mistake -- a big one -- was to accept the reports and conclusions of the previous doctors at face value. When we got a copy of the earlier blood test results, they were not normal. The elevated platelet count was sitting there, always abnormal -- 3 years ago, 2 years ago, last year.

As a patient, what can you do to help your doctor avoid the trap of trusting prior reports and diagnoses? Provide your own summary of the history of the illness, from the beginning. Include details, but not so many that the doctor cannot read or record them all. Don't suggest a diagnosis. Just focus on describing events and experiences accurately. Then, don't be afraid to do some research and offer your thoughts.