Nonprofit, nonpartisan journalism. Supported by readers.


UCare generously supports MinnPost’s Second Opinion coverage; learn why.

What cyberchondriacs and gamblers have in common

Gamblers and cyberchondriacs alike find patterns in random events.

A solid majority of us in the United States — 61 percent, according to a 2009 Pew survey — turn to the web when we’re feeling under the weather and are trying to decide whether or not to see a doctor. A half hour or so later, we may reach the reassuring conclusion that we have nothing to worry about.

Or not. For online searches lead many of us to immediately jump to the most dire conclusion possible — that a headache means we have a brain tumor, for example, or that a muscle cramp means we’re in the early stages of amyotrophic lateral sclerosis (ALS).

Such dark leaps in logic have led researchers (and physicians) to identify a special kind of patient:  the cyberchondriac.

Why are we inclined to interpret symptom lists on the web in the worst possible way? A new study suggests an answer. And it has to do with something very basic: the order in which the symptoms are presented.

Two experiments

The study, led by University of Arizona psychologist Virginia Kwan and published in the April issue of Psychological Science, involved two separate experiments. In one, healthy volunteers were presented with lists of six symptoms (three general, three specific) of a fictional type of thyroid cancer and asked to check off which of the symptoms they’d experienced during the previous six weeks. They were also asked to rate what they thought was their personal risk of having such a disease. (The volunteers did not know the disease was fictional.) The order of the symptoms varied. Some volunteers received a list that cited the milder and most general symptoms (such as feeling easily fatigued) first. Others were given a list with the least common and most specific symptoms (a lump in the throat) at the top. A third list mixed up the common and uncommon symptoms.

The study found that when the symptoms alternated between general and specific, people were much less likely to think they were at risk of having the disease.

The mind of a gambler

What this finding suggests, according to Kwan and her colleagues, is that, like gamblers, health consumers tend to see patterns in what are actually random events. Gamblers often irrationally believe, for example, that if they’ve just experienced a brief streak of winning rolls of the dice, the next roll will also come up positive for them.

In the same way, when we read through a list of disease symptoms, if we check off two or three symptoms consecutively on the list, we tend to unconsciously see a “streak,” or pattern, which we interpret as meaning we probably have the next symptom on the list — and perhaps the disease itself.

Just like a gambler believes in “hot hands,” people searching the web for medical information believe in “hot symptoms.”

Long vs. short lists

The study’s second experiment found a way of dampening the “streak” effect. It gave volunteers lists of symptoms of a real disease (a rare type of brain cancer called meningioma). This time, however, the lists contained 12 symptoms (six general and six specific).

The study found that people were less likely to believe they were at risk of having meningioma with this longer list of symptoms, even when the common and the specific symptoms were grouped separately. This is probably the result of the “dilution effect,” say the study’s authors. The more symptoms we see that we haven’t checked off, the less concerned we are.

What the findings mean

Both experiments, Kwan and her co-authors conclude, “demonstrates that something as arbitrary as the order in which symptoms are presented in a checklist can affect perceived risk of disease.”

This finding could prove useful to public health officials, they add. For example, if you want people to seek medical help at the earliest signs of symptoms — such as during an epidemic — then grouping common, general symptoms together may help. But if the goal is the opposite — if you don’t want people to panic and jump to the conclusion that they have some rare disease — then alternating general with specific symptoms would be the way to go.

The finding also suggests, of course, that all of us should use a bit more logic when seeking out medical information on the web.

Cyberchrondriacs, take note.

You can also learn about all our free newsletter options.

No comments yet

Leave a Reply