In one classic joke a doctor gives a patient bad news: "You have three months to live." "I want a second opinion!" snarls the patient. "OK. I also think you're ugly." This is the bedside manner of Dr. Nocebo, and it's probably not good for you.

Most people have heard of the placebo effect, but relatively few -- including physicians -- have heard of nocebo, which was only named in 1961. This is starting to change, with an increasing number of academic articles about nocebo, with a nocebo conference being given and with an odd array of non-medical groups getting interested.

The nocebo effect is the little-known evil twin of the placebo effect. In the placebo effect a sick person feels better because he believes he's being treated, often with a sugar pill or some other inert substance or intrinsically meaningless therapy. Placebo effects can be quite powerful, and a surprisingly large part of medicine, both ancient and modern, consists of placebo effects, whether physicians and patients realize it or not.

The nocebo effect comes into play in several situations: the one- to-one interaction of doctor and patient, each with his or her expectations; people's general beliefs and expectations outside the therapeutic situation; and expectations created in groups of people.

In a sickeningly simple example of doctor-to-patient nocebo effect, hospital patients were given sugar water and were told that it was an emetic. Eighty percent of them vomited. Thanks, Doc! This may provide one hint as to why there aren't more nocebo experiments. In another experiment I'm glad I didn't volunteer for, asthmatics inhaled a nebulized saline solution, which would have been inert except that they were told it was an irritant solution. Accordingly, they experienced breathing problems, and some had full-blown asthma attacks. They were then given the same saline spray, but this time they were told it was a helpful medicine -- and they recovered. Note that the patients didn't just think their airways were constricting -- they really were.

The extreme case of doctor-to-patient nocebo effect is voodoo death, in which, as we have all read, the shaman or witch doctor of the tribe pronounces the curse of death on some hapless person, who then dies. Sometimes they are said to die of fear, sometimes out of sheer conviction.

Unfortunately for the impressiveness of this evidence, it probably isn't true. Most accounts of voodoo death turn out to be the fantasies of excitable explorers and anthropologists. Others turn out to be retrospective. Say I notice that I am really, really sick and suspect that I'm going to die. I try to figure out why and it occurs to me that someone must have cursed me. (Perhaps that nasty editor at a certain alternative weekly who was so snippy about paying me! I'm sure he wants me dead, especially after I threatened to take him to tiny claims court.) Or perhaps someone else notices that I'm unwell and decides to claim the credit: "I cursed her! She crossed me and now she'll die. And the rest of you uppity freelancers better watch out."

One anthropologist has argued that some cases of voodoo death among the Yolngu people of Australia's East Arnhem Land are really cases of dehydration -- the patient and the patient's family figure it's pointless to waste water on someone who's doomed, so they don't, and eventually the person dries up and dies. Other anthropologists deny it hotly, saying that, 1) the Yolngu people actually take excellent care of sick people, not omitting the annoying insistence on just having a little sip of tea, come on, drink it for me; and, 2) there aren't any voodoo deaths in East Arnhem Land anyway, except for a few that get diagnosed in retrospect. I think this must be true, because I have crossed some pretty scary editors and publishers, yet I am still alive.

Your general expectations about your health can also affect your health, without anyone pointing the finger, burying a black cockerel under your doorstep or adjusting his white coat and shaking his head doubtfully. People who believe in the predictions of traditional Chinese astrology about unlucky combinations of birth years and disease are apt to die several years sooner than people who don't believe in or don't know about those predictions, according to a study done by sociologist David Phillips and colleagues. (Phillips and associates have hit a motherlode of data in California birth certificates, producing a torrent of studies on what kills people.) Traditional Chinese belief has it that people born in "fire years" are more apt to die of heart disease, and people born in "earth years" are more apt to die of cancer. And the more you believe it's true, the truer it is. It's not that the belief gives you the illness, but if you have the illness, it seems to kill you off somewhere between one and five years quicker.

A possible example of nocebo effect on coronary disease has been teased out of the famous Framingham study (a massive longitudinal study that began in 1948). Elaine Eaker and her colleagues found that women who said they were more likely than other women their age to develop heart disease were in fact twice as likely (over a 20-year period) to experience myocardial infarction or coronary death, even when the results were controlled for variables like smoking, high blood pressure and high cholesterol. Sometimes it's no comfort to be right. It also appears that women who felt they had little control over their lives and who were lonely and who didn't get to take many vacations had a higher rate of coronary disease, a discouraging finding. The meek will inherit the earth, but they may be too sick to have fun with it.

I suppose this means that you can hearten depressed people by telling them that they must look on the bright side or risk damaging their health, maybe even fatally. "Cheer up or die!" you can tell them. No doubt they will thank you some day.

Other much-written-about forms of the nocebo effect are sociogenic illness, psychogenic illness or mass hysteria. "Knowledge of sickness in others fosters an expectation that one may also be subject to the same condition," writes Robert Hahn, perhaps the foremost scholar of nocebo. In cases like these, groups of people complain of symptoms like nausea, dizziness, shortness of breath, numbness, and coughing. Typically they have been made aware that someone else has gotten inexplicably sick, that there is a strange smell in their workplace or that something else worrisome is going on. Gradually more and more people display symptoms, making it seem all the more certain that something is terribly wrong.

Doctors may tread warily in linking the nocebo effect with terms like "mass hysteria," because almost all doctors have experienced this themselves in the form of MSD -- medical student's disease. The medical student, assumed to be a reasonably intelligent, level-headed person, reads about various unpleasant conditions with diffuse or common symptoms and becomes convinced that he or she has the disease. "Look at this checklist! Headaches: yes, I have headaches. Dizziness: I felt dizzy yesterday in the revolving door. Insomnia: I hardly got any sleep last night. Nervousness: totally! Oh my God, I have a brain tumor. Probably inoperable. It is both tragic and ironic. Wait, read these symptoms of tetanus! Headaches, dizziness, agitation ..."

Why, I understand that it is not unusual for medical students to read about rabies, look in the mirror and see themselves beginning to foam at the mouth, and rush out into the street barking and biting people. Most doctors remember this period in their career with some embarrassment, naturally, but perhaps it gives them a little sympathy for other victims of the nocebo effect.

Sociogenic factors like these are used by some to explain phenomena like Gulf War syndrome or illness following breast implants. Naturally such explanations are extremely controversial.

In cases of sociogenic illness, anthropologist Dan Moerman says, "The kinds of symptoms are always pretty much the same. They're sleepless, and they have achy joints and there's nothing ever very specific."

Moerman describes an incident in Michigan in the mid-1970s in which a large amount of animal feed became contaminated with compounds meant for fireproofing pajamas -- PBBs (polybrominated biphenyls). Contaminated feed sickened and killed a number of cattle. (However, they did not catch fire in their beds.) Because the effects of PBBs were little known (but people knew that the related polychlorinated biphenyls, or PCBs, are dangerous), and because it was unclear how many farms had gotten contaminated feed, there was widespread alarm on Michigan farms. Many farmers were frightened to find themselves experiencing mysterious symptoms.

"Guess what their symptoms were?" asks Moerman. "They had sore knees, and they couldn't sleep and they were distracted." Now, nearly 25 years later, epidemiological studies have looked at the health of those farmers, and compared them to Wisconsin farmers without PBB exposure. "They studied these people from pillar to post," says Moerman. "There was no difference. Zero. No mortality difference, no cancer difference. But the Michigan farmers who say they were afflicted, were afflicted. They were sick! They were not sick with anything that anyone could measure, but they were sick."

To Moerman, the most intriguing part of the PBB episode may come from the cows. He recently charted milk production in Michigan cows in pounds of milk per cow per year ("I had never really realized how wonderful agricultural statistics are") and found a sharp drop in milk production in 1974. "It drops by 10 to 12 percent. It just goes 'kerchunk' and it drops. Ninety-nine percent of the cows never got any of this PBB, but their production of milk dropped anyway. Well, you know cows don't read the newspaper. But what do cows do?" Moerman asks. "Hang out with farmers?" I guess. "That's right. They hang out with farmers. I think we have a nocebo effect in domestic animals."

Moerman intends to research this intriguing area further before publishing, but the working hypothesis is that the farmers' concern about their cows, who for all they knew had been gobbling pure poison, was somehow communicated to the cows, who responded by slacking off on the milk production. He also points out a difference between nocebo and placebo.

Rats are apparently as vulnerable as cows. In one experiment, which didn't set out to be about nocebo, rats were given sweetened water to drink and then given shots of cyclophosphamide, which causes nausea and also induces sometimes fatal immunosuppression. To the experimenters' surprise, rats who only got one shot of cyclophosphamide but kept on being fed the sweetened water continued dying at a high rate. The rats associated feeling dreadful with the water, and so as long as they got the water they felt dreadful -- even unto death.

Does all this mean I make a space for my own illness? Should I take responsibility for my pain, sickness and plantar warts? I think not. OK, a little. But just as placebo has limits, just as flu viruses laugh scornfully when I tell them I took my vitamins, so they should go away, there are limits to nocebo as well.

If a person could give himself cancer by sheer mental effort, one relative of mine would have managed it, by intensive brooding on the cancer deaths and diagnoses of family members, by careful examination of all the carcinogenic activities he had engaged in and by horribly protracted discussion of the inevitability of his own cancer. He ought to have been riddled with tumors. He ought to have had tumors with tumors of their own. Yet, despite being a strong-minded man of considerable talents, he was unable to give himself cancer and was forced to die in old age of completely unrelated causes.

The nocebo effect that can occur in doctor-patient interactions is perhaps the least studied form of nocebo. Most doctors are surprisingly unenthusiastic about the placebo effect, which is at least on their side. ("They don't want to talk about it," Moerman says.) Perhaps it is because no matter how many studies are done, it just doesn't feel like science. Doctors don't want to cure people with sugar pills and bedside manner if they can cure them with effective medications and therapies. Indeed, as a patient, if I had my choice, I'd prefer not to rely on the hearty handshake and the sugar pill. But medicine has not reached the point where these things are unnecessary.

If doctors don't want to hear about placebo, imagine how they feel about nocebo, about the idea that things they say could make their patients sicker.

Neither the placebo nor the nocebo effect has been much studied -- medical discomfort with such squishy phenomena aside, there's no money in it. As Dan Moerman recently commented on NPR, drug companies are unlikely to study something so impossible to patent.

However little the average physician may wish to consider the nocebo effect, it's becoming profoundly interesting to that segment of the legal profession that defends corporations in product liability cases. A recent conference on nocebo was attended by attorneys, to the surprise of the academics who presented papers there. "It was weird," says Moerman.

At a Web site for the law firm of Venable, Baetjer and Howard, an article by James Archibald and Scott Zebrak tells defense attorneys that "expert testimony on the nocebo effect has the potential to rebut a claim that a defendant's tortious conduct or product caused a plaintiff's injuries." So far, "in no case has the nocebo effect per se been articulated as a cause of a plaintiff's injuries," but it's inevitable that the argument will be made eventually. "I have a certain sympathy with that perspective," says Moerman, thinking back to the cows. The people at Medical Litigation News are also thinking hard: "Placebo and nocebo reactions are a part of the human condition and must be considered in every Personal Injury and Medical Negligence action."

Another focus of interest in nocebo comes from certain alternative healers, who feel the nocebo effect may explain cases where homeopathy appears to fall flat: It was the fault of doctors wielding nocebo. One creative thinker writes that if you give a "normally 'effective'" drug to a patient and it doesn't do any good, it may be "because the doctor has convinced the patient the drug wouldn't work or was just water." (This site has a number of intriguing publications for sale covering everything from lucid dreaming, weather control and the curative powers of chicken feet.)

As a nonbeliever in Chinese astrology who hasn't noticed my cattle acting funny, my principle personal interest in the nocebo effect is in the doctor-patient relationship. A few years ago I had a go-round with a grim diagnosis. When I asked several physicians whether there was a chance I might get better without treatment, thus skipping some really unpleasant therapies, one said there was every chance in the world that I'd get better, one said anything is possible and one snorted and said, "Are you kidding?"

Through no particular virtue of mine, I had a spontaneous remission, but I'm still mad at Dr. Are-You-Kidding-Me-Doomed-Fool. I feel that he needs to take remedial bedside manner. Over and over, until he passes.

In one hospital study, patients awaiting surgery were randomly divided into two groups. The control group got a "cursory" visit from their anesthesiologist the night before, whereas the experimental group got a longer, more sympathetic, chatty visit. The experimental group ended up requiring only half as much anesthesia, and they left the hospital two to three days sooner. (I know what you're thinking: They left to get away from those gabby anesthesiologists. You're a very cynical person and there is no place for you in medical science.)

It's apparently helpful to feel that your doctor cares whether you live or die and in fact has a preference that you live.

The moral for the physician would seem to be that a good doctor is also a spin doctor. The moral for the patient? If your doctor tells you you're ugly, get a new doctor.