Working in the emergency department (ED) recently, I was approached by a frustrated patient. She didn’t feel that her symptoms were being adequately addressed. Holding her lower back with one hand, she asked when she could have her pain medication topped up. Judging by her posture and the expression on her face, I could tell she was really struggling.

“I know it’s probably just radiculopathy,” she said, “but I can’t take it anymore.”

She had pronounced the term radiculopathy with the expertise of a colleague – or so I thought.

“Are you in the medical field?” I asked.

“No,” she replied, “I’ve got Google.”

Not so long ago, someone with a medical concern really had only two options: they could either take their doctor’s word as gospel or hope that their grandmother’s folk remedy would do the job. Now patients can browse online chat forums, WebMD, and Wikipedia long before they ever decide to go to the hospital. In the age of Google, our role as doctors has changed in a profound way: we’re no longer in the business of simply meting out bits of information; we now have a responsibility to engage in a dialogue with our patients and to interpret their latest Internet findings as responsibly as we can.

I got up from my chair and walked over to the patient. She could barely stand up with the pain. I helped her hobble a short distance back to her room, and then supported her as she managed to get back onto the stretcher. Her shoulders were tense, and when she grabbed my arm for support, her grip was white-knuckled.

“So what makes you think it’s a radiculopathy?” I asked.

“I read about it online. It’s shooting down the back of my leg like a lightning bolt,” she told me. “It’s probably one of those pinched nerves.”

Self-diagnosis on this level is a fairly recent phenomenon. Sometimes patients even convince themselves that they’re heading for an early grave. We’ve all seen it. “I’m here for a full body MRI,” one patient told me recently. “That’s the only way we’re going to find this cancer.” With a proper workup and a bit of reassurance, these patients can get some clarity and go home feeling significantly better. But self-diagnosis also has a much more sinister side: patients can “anchor” themselves – and us – to a benign diagnosis, when in reality there’s something much more dangerous going on.

Back in the ED, my patient tells me that her pain is similar to her prior episodes. I can’t identify any red flag features in her history, and her neurological and musculoskeletal exams are all reassuring. I walk back to my computer and bring up her chart on our system. There’s an MRI report from just a few months ago.

“Bingo,” I thought.

The report notes a bulging lumbar disc impinging on a nearby nerve root just as it exits the spinal canal.

Every day, with almost every patient we see, we anchor ourselves to one diagnosis over another. It happens naturally. We tell ourselves: common things are common; if we don’t make quick decisions, the whole system will grind to a halt. We make excuses. Anchoring and confirmation bias are dangerous, sure – but whether or not we like to admit it, most of the time, we rely on them to keep us efficient.

“You were right,” I said as I walked back into the patient’s room. “It is radiculopathy, and you do have a pinched nerve. We’ll be able to get you home today, with a prescription for some new pain meds and a follow-up appointment at our spine clinic.”

She smiled and shook my hand.

“Just one more question,” I said. “What website were you using when you came across the term radiculopathy?”

“Oh I’m not sure,” she said. “It was some website about spinal problems. I just searched for ‘back pain, leg pain, and chills.’”

I stopped in my tracks.

“Chills?” I asked.

“Yeah, this morning I had some chills,” she said. “I was shaking a bit. But it went away.”

Earlier I had asked her about fevers, intravenous drug use, chemotherapy, immunodeficiency, and she had said “no” to all of them. Her temperature at triage was 36.8°C, so I hadn’t given it much more thought. But I guess I hadn’t asked about chills.

I checked her temperature again at the bedside: 38.7°C.

We’re forced to come up with a diagnosis. We cinch ourselves to the most parsimonious one, the most plausible one, the one we feel explains the whole clinical picture with as small of a remainder as possible. We don’t like messiness. We don’t like uncertainty. We try to listen to our patients, but when they tell us their stories, we pick out what’s valuable and what’s not. We try to separate fantasy from reality.

I asked my patient to stand up and walk with me like she did before. Her feet fumbled their way to the ground, seeking purchase in the air and looking markedly different than they did earlier.

“It’s weird,” she said. “Now my legs feel funny. It’s like they’re not really following commands.”

Later that night, I sat at home in front of my laptop. It was 3 AM. A Google search bar sat in the centre of my screen, empty. I thought about my patient, whom I had almost sent home with a diagnosis of radiculopathy instead of the epidural abscess that she really had. A new MRI confirmed the diagnosis, and she was headed to the operating room just as I was leaving the hospital.

Sitting at my kitchen table, I typed in “back pain leg pain chills” and hit “Enter.” A deluge of information appeared on my screen, with links to online symptom-checkers and patient forums. I clicked on a link near the bottom and saw the word radiculopathy in bold.

The Internet can be a scary place for patients. It’s up to us to protect them from it.