Placebo Effects
New! Improved! And Still 100 Per­cent Fake

By Jay Dixit
Sun­day, May 19, 2002; Page B01

On the win­dowsill of his Colum­bia Uni­ver­sity office, psy­chi­a­trist B. Tim­o­thy Walsh has a lit­tle golden bot­tle labeled “placebo.” The con­tainer is filled with sugar pills — breath mints, to be exact — and pur­ports to treat every­thing from “bad hair” to “can’t take a joke” to “no rhythm” to “the blahs.” About 50 con­di­tions are listed with a box next to each. The label directs patients: “Select symp­tom. Check box. Take 2 mints. Bingo… all better!”

Walsh’s “cure” is rem­i­nis­cent of the snake oil, ton­ics and mag­i­cal elixirs that used to be pop­u­lar in this coun­try — the kind of car­ni­val quack­ery that no one falls for any­more. Right?

Well, not quite. Accord­ing to a recent report by Seat­tle psy­chi­a­trist Arif Khan, who con­ducted analy­ses of the placebo effect in 96 clin­i­cal tri­als sub­mit­ted to the FDA between 1979 and 1996, sugar pills were often as effec­tive as anti­de­pres­sants. And judg­ing by the media’s reac­tion to the pub­li­ca­tion of Khan’s find­ings, you might think that peo­ple are get­ting ready to throw away their lit­tle green-and-white Prozac pills. It all leaves our col­lec­tive faith in mod­ern med­i­cine a lit­tle shaken.

But it shouldn’t. Stud­ies like Khan’s, which was first reported in The Post, are con­fus­ing, and and some peo­ple have assumed that his find­ings mean that sugar pills work just as well as Prozac, Paxil and Zoloft. But in fact, as the orig­i­nal arti­cle pointed out, these stud­ies don’t mean that anti­de­pres­sants don’t work, or that sugar pills work just as well or even that anti­de­pres­sants don’t work as well as we pre­vi­ously thought. Quite the oppo­site. We’re see­ing these results because anti­de­pres­sants do work — and because of our increased faith in the power of mod­ern med­i­cine. What has hap­pened is that our con­fi­dence in anti­de­pres­sants has grown — partly because of mod­ern phar­ma­ceu­ti­cal mar­ket­ing tech­niques. Our belief in their power causes place­bos to work bet­ter. When peo­ple vol­un­teer to take part in clin­i­cal tri­als, they expect the med­ica­tion they are given to work — and it does work, even if it is just sugar.

Con­fu­sion abounds about the “placebo effect” — the term that doc­tors use to describe the phe­nom­e­non where patients get bet­ter because they expect the treat­ment to work, even though they’re actu­ally ingest­ing a phar­ma­co­log­i­cally inert sub­stance. And that con­fu­sion has been com­pounded by a vari­ety of recent stud­ies that seem to con­tra­dict each other. Last year, an arti­cle in the New Eng­land Jour­nal of Med­i­cine declared the placebo effect to be a sta­tis­ti­cal chimera, argu­ing that sick peo­ple tend to get bet­ter over time, regard­less of whether they take sugar pills, and that belief plays no part in that process. Khan’s study seemed to refute that: Prozac had five clin­i­cal tri­als, and in only two of them did it work sig­nif­i­cantly bet­ter than sugar pills. For Zoloft and Paxil, it took even more tri­als to obtain a pos­i­tive result.

There is an answer to these appar­ent con­tra­dic­tions — and it’s not to stop tak­ing anti­de­pres­sants. The first step is to ask why the placebo effect is so strong rel­a­tive to the actual effect of the drug.

A major rea­son for the placebo’s effi­cacy stems from the care and con­cern shown to patients dur­ing clin­i­cal tri­als, which has a big impact on their well-being — espe­cially in the case of depres­sion. “Care, con­cern, and gen­eral atten­tion are cru­cial fac­tors in our patients’ recov­er­ies from ill­ness — not only from depres­sion, but from all med­ical ill­nesses,” says Andrew Leuchter, a pro­fes­sor of psy­chi­a­try at UCLA.

Another fac­tor is the time scale. Clin­i­cal tri­als are fairly short (usu­ally eight weeks) and often that’s not enough time to observe the long-term effects of a depres­sion treat­ment. A placebo group might show equal improve­ment to a med­ica­tion group at first. But if a placebo study were done over a period of sev­eral years, the placebo group would almost cer­tainly fall behind.

But the main fac­tor that explains why anti­de­pres­sants often fare no bet­ter than sugar pills is more mys­te­ri­ous, and it was sug­gested by a sep­a­rate study: The placebo effect is grow­ing more powerful.

In a sur­vey of anti­de­pres­sant tri­als from 1981 to 2000, Walsh found that the placebo effect grew more pro­nounced over the years. (As Walsh jokes, “They’re mak­ing place­bos bet­ter and bet­ter.”) But why? The answer to that ques­tion is cul­tural, not med­ical. After all, those tri­als were con­ducted in basi­cally the same way 20 years ago as they are today. The thing that’s changed is us.

It’s well estab­lished that con­fi­dence in a treat­ment can have a mind-over-body effect on how well the treat­ment works. And our culture’s faith in tech­nol­ogy gen­er­ally, and in med­i­cine in par­tic­u­lar, has grown in that time period.

In 1981, the first year exam­ined in Walsh’s study, AIDS was just appear­ing on our radar screens.Doctors were diag­nos­ing a mys­te­ri­ous and fright­en­ing new dis­ease and no one knew how to fight it. Now, peo­ple expect to live long and pro­duc­tive lives with HIV. That’s just one of the exam­ples of the ways the med­ical and tech­no­log­i­cal break­throughs of the past 20 years have changed our expec­ta­tions of med­i­cine and sci­ence. We are map­ping the genome, can explore the brain using mag­netic res­o­nance imag­ing, trans­plant organs and build arti­fi­cial hearts.

Today, we also have a bet­ter under­stand­ing of men­tal ill­ness. We know that many peo­ple expe­ri­ence depres­sion, we view depres­sion as an ill­ness, and the stigma about seek­ing treat­ment for it is fad­ing. A study pub­lished in Jan­u­ary in the Jour­nal of the Amer­i­can Med­ical Asso­ci­a­tion found that the num­ber of Amer­i­cans being treated for depres­sion had tripled between 1987 and 1997, and the num­ber tak­ing anti­de­pres­sants had dou­bled. In fact, accord­ing to an ABC News poll, one in every eight Amer­i­cans has been treated with anti­de­pres­sants at some stage in their lives.

Phar­ma­ceu­ti­cal mar­ket­ing also has bol­stered our feel­ings of faith in drugs. With a mil­lion anti­de­pres­sant pre­scrip­tions a week in this coun­try, drug giants can afford to spend bil­lions on adver­tis­ing to con­vince us their prod­ucts work. Com­mer­cials fea­ture ani­ma­tions of sad neu­rons weep­ing, pre­sum­ably because they don’t have enough sero­tonin to go around. Once they get Zoloft, they become smil­ing, fly­ing, ani­mated neurons.

That mar­ket­ing has an effect on our psy­che. We have become more will­ing to believe that depres­sion can be caused by a chem­i­cal imbal­ance in the brain — and that it can be treated chem­i­cally. “It has to do with the pro­foundly cog­ni­tive nature of depres­sion,” says Jack Glaser, a social psy­chol­o­gist and pro­fes­sor of pub­lic pol­icy at the Uni­ver­sity of Cal­i­for­nia at Berke­ley. Since depres­sion can be treated with­out med­ica­tion, using cog­ni­tive ther­apy, tak­ing a pill you think will make you bet­ter can break the cycle and help you cope. “The most likely rea­son placebo effects are get­ting big­ger is the increas­ingly pre­vail­ing belief that pills work for treat­ing depres­sion — thus the placebo has more credibility.”

Walsh’s study found that, for the most part, the effec­tive­ness of a drug increased along with the effec­tive­ness of the placebo. But he found that the placebo effect is not only stronger today, but more vari­able. This vari­abil­ity accounts for why place­bos some­times score bet­ter than real drugs. “The drugs are more effec­tive, but there’s so much noise in the sys­tem, and peo­ple are get­ting bet­ter for so many dif­fer­ent rea­sons, that it can be hard to show every time that the drug is bet­ter than placebo,” says Walsh. “But it is.”

Doc­tors have long known that the power of sug­ges­tion can have real med­ical effects. And expec­ta­tion of harm — the so-called nocebo effect — can be as pow­er­ful as the placebo effect. In one nocebo study, patients were given sugar water and told it was an emetic. Eighty per­cent vom­ited. In another, asth­matic patients inhaled saline spray think­ing it was an irri­tant. Many had breath­ing prob­lems and asthma attacks. Many recov­ered when researchers gave them the same spray again, this time telling them it was a bronchiodilator.

Other stud­ies have demon­strated that place­bos improve blood pres­sure, cho­les­terol lev­els, heart rate and aller­gies, and can make warts van­ish. For some rea­son, red sugar pills kill pain bet­ter than green, blue, or yel­low ones; blue sugar pills work bet­ter as seda­tives than pink ones. Patients don’t just think they get bet­ter on place­bos: They really do get bet­ter, and the changes are reflected in brain scans and blood tests.

Because peo­ple are sug­gestible, doc­tors know that the hype (often, true hype) sur­round­ing a med­ica­tion can make it more effec­tive. One old doc­tors’ cyn­i­cal joke goes, “Use new drugs quickly, while they still work.”

What does all of this research imply for phar­ma­ceu­ti­cal com­pa­nies and for the pub­lic? The first con­clu­sion is that if the kind of care and atten­tion patients received in the clin­i­cal tri­als can have such a strong impact on an individual’s recov­ery, health insur­ance com­pa­nies should real­ize that encour­ag­ing longer doctor’s vis­its will save money in the long run, and we should make sure that depressed peo­ple receive ther­apy or coun­sel­ing. The ABC News poll that found one in eight Amer­i­cans had taken anti­de­pres­sants also showed that 59 per­cent had not received ther­apy in con­junc­tion with the med­ica­tion. The Jour­nal of the Amer­i­can Med­ical Asso­ci­a­tion, which reported that anti­de­pres­sant use has gone up, also reported that face-to-face ther­apy has gone down: From 1987 to 1997, the per­cent­age of patients in coun­sel­ing dropped from 71 per­cent to 60 percent.

The key to under­stand­ing the placebo effect is to dis­en­tan­gle the effect of the drug from the effect of the care. But in the real world, we don’t have to worry about which is help­ing more. The mes­sage we should take from Khan’s new research is not that anti­de­pres­sants don’t work, but that they work best in con­junc­tion with increased care and con­cern — in the form of ther­apy. That’s the com­bi­na­tion of treat­ment that the med­ical pro­fes­sion should aim to provide.

Jay Dixit is a writer liv­ing in New York City.

© 2002 The Wash­ing­ton Post Company

The Post’s opin­ion and com­men­tary sec­tion runs every Sunday.

• More in Out­look