In Defense of Food

Home > Nonfiction > In Defense of Food > Page 8
In Defense of Food Page 8

by Michael Pollan


  Even a cur­sory exa­mi­na­ti­on of the study’s met­hods ma­kes you won­der what, if anyt­hing, it pro­ved, eit­her abo­ut di­etary fat or me­at eating. You co­uld ar­gue that, li­ke the Nur­ses’ He­althy Study, all any such tri­als pro­ve is that chan­ging one com­po­nent in the di­et at a ti­me, and not by much, do­es not con­fer a sig­ni­fi­cant he­alth be­ne­fit. But per­haps the stron­gest conc­lu­si­on that can be drawn from an analy­sis of the Wo­men’s He­alth Ini­ti­ati­ve is abo­ut the in­he­rent li­mi­ta­ti­ons of this kind of nut­ri­ent-by-nut­ri­ent nut­ri­ti­on re­se­arch.

  Even the be­gin­ning stu­dent of nut­ri­ti­onism will im­me­di­ately spot se­ve­ral flaws: The fo­cus was on di­etary fat rat­her than on any par­ti­cu­lar fo­od, such as me­at or da­iry. So wo­men co­uld re­ach the­ir go­al simply by switc­hing to lo­wer-fat ani­mal pro­ducts. Al­so, no dis­tinc­ti­ons we­re ma­de bet­we­en dif­fe­rent types of fat: Wo­men get­ting the­ir al­lo­wab­le por­ti­on of fat from oli­ve oil or fish we­re lum­ped to­get­her with wo­men get­ting the­ir fat from low-fat che­ese or chic­ken bre­asts or mar­ga­ri­ne. Why? Be­ca­use when the study was de­sig­ned six­te­en ye­ars ago, the who­le no­ti­on of “go­od fats” was not yet on the ma­inst­re­am sci­en­ti­fic sco­pe. Sci­en­tists study what sci­en­tists can see.

  Anot­her prob­lem with the tri­al was that the low-fat gro­up fa­iled to hit the tar­get of re­du­cing the­ir fat in­ta­ke to 20 per­cent of to­tal ca­lo­ri­es. The best they co­uld ma­na­ge was 24 per­cent in the first ye­ar, but by the end of the study they’d drif­ted back to 29 per­cent, only a few per­cen­ta­ge po­ints lo­wer than the cont­rol gro­up’s fat in­ta­ke. Which was it­self drif­ting down­ward as the wo­men al­lo­wed to eat as much fat as they wan­ted pre­su­mably re­ad the news­pa­pers and the fo­od pro­duct la­bels and ab­sor­bed the cul­tu­re’s ent­hu­si­asm for all things low fat. (This cor­rup­ti­on of a cont­rol gro­up by po­pu­lar di­etary ad­vi­ce is cal­led the tre­at­ment ef­fect.) So it’s hardly surp­ri­sing that the he­alth out­co­mes of the two gro­ups wo­uld not gre­atly dif­fer-by the end, they might ha­ve be­en con­su­ming pretty much the sa­me di­et.

  I say “might ha­ve be­en” be­ca­use we ac­tu­al­ly ha­ve lit­tle idea what the­se wo­men we­re re­al­ly eating. Li­ke most pe­op­le as­ked abo­ut the­ir di­et, they li­ed abo­ut it-which brings us to what is per­haps the sing­le big­gest prob­lem in do­ing nut­ri­ti­on sci­en­ce. Even the sci­en­tists who con­duct this sort of re­se­arch con­duct it in the know­led­ge that pe­op­le un­de­res­ti­ma­te (let’s be ge­ne­ro­us) the­ir fo­od in­ta­ke all the ti­me. They ha­ve even de­ve­lo­ped sci­en­ti­fic fi­gu­res for the mag­ni­tu­de of the er­ror. “Va­li­da­ti­on stu­di­es” of di­etary tri­als li­ke the Wo­men’s He­alth Ini­ti­ati­ve or the Nur­ses’ Study, which rely on “fo­od-fre­qu­ency qu­es­ti­on­na­ires” fil­led out by su­bj­ects se­ve­ral ti­mes a ye­ar, in­di­ca­te that pe­op­le on ave­ra­ge eat bet­we­en a fifth and a third mo­re than they say they do on qu­es­ti­on­na­ires.* How do the re­se­arc­hers know that? By com­pa­ring what pe­op­le re­port on the­ir fo­od-fre­qu­ency qu­es­ti­on­na­ires with in­ter­vi­ews abo­ut the­ir di­etary in­ta­ke over the pre­vi­o­us twenty-fo­ur ho­urs, tho­ught to be so­mew­hat mo­re re­li­ab­le. So­mew­hat. Be­ca­use as you might ex­pect, the­se “twenty-fo­ur-ho­ur re­call” da­ta ha­ve the­ir own ac­cu­racy prob­lems: How typi­cal of yo­ur ove­rall di­et is what you ate du­ring any sing­le twenty-fo­ur-ho­ur pe­ri­od?

  To try to fill out the fo­od-fre­qu­ency qu­es­ti­on­na­ire used by the Wo­men’s He­alth Ini­ti­ati­ve, as I re­cently did, is to re­ali­ze just how shaky the da­ta on which all such di­etary stu­di­es rely re­al­ly are. The sur­vey, which ta­kes abo­ut forty-fi­ve mi­nu­tes to comp­le­te, starts off with so­me re­la­ti­vely easy qu­es­ti­ons. “Did you eat chic­ken or tur­key du­ring the last three months?” Ha­ving ans­we­red yes, I then was as­ked, “When you ate chic­ken or tur­key, how of­ten did you eat the skin?” And, “Did you usu­al­ly cho­ose light me­at, dark me­at, both?” But the sur­vey so­on be­ca­me har­der, as when it as­ked me to think back over the past three months to re­call whet­her when I ate ok­ra, squ­ash, or yams we­re they fri­ed, and if so, we­re they fri­ed in stick mar­ga­ri­ne, tub mar­ga­ri­ne, but­ter, shor­te­ning (in which ca­te­gory they inexp­li­cably lum­ped to­get­her hydro­ge­na­ted ve­ge­tab­le oil and lard), oli­ve or ca­no­la oil, or nons­tick spray? I wo­uld ho­pe they’d ta­ke my ans­wers with a gra­in of salt be­ca­use I ho­nestly didn’t re­mem­ber and in the ca­se of any ok­ra eaten in a res­ta­urant, even a hypno­tist or CIA in­ter­ro­ga­tor co­uld not ext­ract from me what sort of fat it was fri­ed in. Now that we spend half of our fo­od dol­lars on me­als pre­pa­red out­si­de of the ho­me, how can res­pon­dents pos­sibly know what type of fats they’re con­su­ming?

  Mat­ters got even sketc­hi­er in the se­cond sec­ti­on of the sur­vey, when I was as­ked to spe­cify how many ti­mes in the last three months I’d eaten a half-cup ser­ving of broc­co­li, among a diz­zying ar­ray of ot­her fru­its and ve­ge­tab­les I was as­ked to tally for the di­etary qu­ar­ter. I’m not su­re Mar­cel Pro­ust him­self co­uld re­call his di­etary in­ta­ke over the last ni­nety days with the sort of pre­ci­si­on de­man­ded by the FFQ.

  When you get to the me­at sec­ti­on, the por­ti­on si­zes spe­ci­fi­ed ha­ven’t be­en se­en in Ame­ri­ca sin­ce the Ho­over ad­mi­nist­ra­ti­on. If a fo­ur-oun­ce por­ti­on of ste­ak is con­si­de­red “me­di­um,” was I re­al­ly go­ing to ad­mit that the ste­ak I enj­oyed on an un­re­cal­lab­le num­ber of oc­ca­si­ons du­ring the past three months was pro­bably the equ­iva­lent of two or three (or in the ca­se of a ste­ak ho­use ste­ak, no fe­wer than fo­ur) of the­se por­ti­ons? I think not. In fact, most of the “me­di­um ser­ving si­zes” to which I was as­ked to com­pa­re my own con­sump­ti­on ma­de me fe­el li­ke such a pig that I badly wan­ted to sha­ve a few oun­ces he­re, a few the­re. (I me­an, I wasn’t un­der oath or anyt­hing.)

  The­se are is the sort of da­ta on which the lar­gest qu­es­ti­ons of di­et and he­alth are be­ing de­ci­ded to­day. “The most in­tel­lec­tu­al­ly de­man­ding chal­len­ge in the fi­eld of nut­ri­ti­on,” as Ma­ri­on Nest­le wri­tes in Fo­od Po­li­tics, “is to de­ter­mi­ne di­etary in­ta­ke.” The un­com­for­tab­le fact is that the en­ti­re fi­eld of nut­ri­ti­onal sci­en­ce rests on a fo­un­da­ti­on of ig­no­ran­ce and li­es abo­ut the most ba­sic qu­es­ti­on of nut­ri­ti­on: What are pe­op­le eating? Over lunch, I as­ked Nest­le if I was per­haps be­ing too harsh. She smi­led.

  “To re­al­ly know what a per­son is eating you’d ha­ve to ha­ve a se­cond in­vi­sib­le per­son fol­lo­wing them aro­und, ta­king pho­tog­raphs, lo­oking at ing­re­di­ents, and con­sul­ting ac­cu­ra­te fo­od com­po­si­ti­on tab­les, which we don’t ha­ve.” When you re­port on an FFQ that you ate a car­rot, the ta­bu­la­tor con­sults a U.S. De­part­ment of Ag­ri­cul­tu­re da­ta­ba­se to de­ter­mi­ne exactly how much cal­ci­um or be­ta-ca­ro­te­ne that car­rot con­ta­ined. But be­ca­use all car­rots are not cre­ated equ­al, the­ir nut­ri­ent con­tent var­ying with everyt­hing from the va­ri­ety plan­ted and type of so­il it was plan­ted in to the ag­ri­cul­tu­re system used (orga­nic? con­ven­ti­onal?) and the car­rot’s fresh­ness, the­se tab­les suf­fer from the­ir own inac­cu­ra­ci­es.

  I was be­gin­ning to re­ali­ze just how much sus­pen­si­on of dis­be­li­ef it ta­kes to be a nut­ri­ti­on sci­en­tist.

  “It’s im­pos­sib­le,” Nest­le con­ti­nu­ed. “Are pe­op­le un­cons­ci­o­usly un­de­res­ti­ma­ti
ng con­sump­ti­on of things they think the re­se­arc­her thinks are bad or ove­res­ti­ma­ting con­sump­ti­on of things they think the re­se­arc­her thinks are go­od? We don’t know. Pro­bably both. The is­sue of re­por­ting is ext­ra­or­di­na­rily se­ri­o­us. We ha­ve to ask, How ac­cu­ra­te are the da­ta?”

  It’s not as tho­ugh the epi­de­mi­olo­gists who de­ve­lop and dep­loy FFQs are una­wa­re of the­ir li­mi­ta­ti­ons. So­me of them, li­ke Wal­ter Wil­lett, stri­ve he­ro­ical­ly to re­pa­ir the fa­ulty da­ta, de­ve­lo­ping “energy adj­ust­ment” fac­tors to cor­rect for the fact that the ca­lo­ri­es re­por­ted on sur­veys are in­va­ri­ably wrong and comp­li­ca­ted “me­asu­re­ment er­ror” al­go­rithms to fix the er­rors in the twenty-fo­ur-ho­ur re­call sur­veys used to fix the er­rors in the FFQ.

  I trac­ked down Gladys Block, the pro­mi­nent epi­de­mi­olo­gist who de­ve­lo­ped the FFQ on which the Wo­men’s He­alth Ini­ti­ati­ve ba­sed its own qu­es­ti­on­na­ire. We met for cof­fee in Ber­ke­ley, whe­re she is a pro­fes­sor in the Scho­ol of Pub­lic He­alth. Ne­aring re­ti­re­ment, Block is unu­su­al­ly tho­ught­ful abo­ut the li­mits of her fi­eld and di­sar­mingly can­did. “It’s a mess,” she sa­id, spe­aking not of the FFQ it­self but of the va­ri­o­us for­mu­lae and al­go­rithms be­ing used to cor­rect er­rors in the da­ta. “Be­ca­use if the energy [i.e., the re­por­ted ca­lo­rie con­sump­ti­on] is off, then the nut­ri­ents are off too. So if you’re go­ing to cor­rect for ca­lo­ri­es, do you then al­so cor­rect for…” She pa­used and then sig­hed. “No, it’s a mess.”

  Block thinks the prob­lem with nut­ri­ti­on sci­en­ce, which she fe­els “has led us ast­ray,” is not the FFQ it­self but mis-and ove­rin­terp­re­ta­ti­on of the da­ta de­ri­ved from the FFQ, a to­ol for which she ma­kes re­alis­tic but stri­kingly mo­dest cla­ims: “The re­al pur­po­se of the FFQ is to rank pe­op­le” on the­ir re­la­ti­ve con­sump­ti­on of, say, fru­its and ve­ge­tab­les or to­tal ca­lo­ri­es. “If so­me­one re­ports con­su­ming fi­ve hund­red ca­lo­ri­es a day, that’s not true, ob­vi­o­usly, but you can say they’re pro­bably at the low end of the spect­rum. Pe­op­le over­wor­ry abo­ut ac­cu­racy.”

  This was not the sort of thing I ex­pec­ted to he­ar from an epi­de­mi­olo­gist. But then ne­it­her was this: “I don’t be­li­eve anyt­hing I re­ad in nut­ri­ti­onal epi­de­mi­ology any­mo­re. I’m so skep­ti­cal at this po­int.”

  TEN - NUTRITIONISM’S CHILDREN

  S o whe­re do­es this le­ave us eaters? Mo­re con­fu­sed abo­ut how to eat than any pe­op­le in his­tory, wo­uld be my strictly uns­ci­en­ti­fic conc­lu­si­on. Ac­tu­al­ly, the­re is so­me sci­en­ce, ad­mit­tedly a lit­tle soft, which has cap­tu­red a bit of the con­fu­si­on that the sup­po­sedly har­der sci­en­ce of nut­ri­ti­on has sown in the Ame­ri­can mind. Pa­ul Ro­zin is a psycho­lo­gist at the Uni­ver­sity of Pen­nsyl­va­nia who has dre­amed up so­me of the mo­re ima­gi­na­ti­ve sur­vey qu­es­ti­ons ever as­ked of Ame­ri­can eaters; the ans­wers he’s col­lec­ted of­fer a pretty go­od in­dex to our cur­rent be­fud­dle­ment and an­xi­ety abo­ut eating. He has fo­und, for examp­le, that half of us be­li­eve high-ca­lo­rie fo­ods eaten in small amo­unts con­ta­in mo­re ca­lo­ri­es than low-ca­lo­rie fo­ods eaten in much lar­ger amo­unts. And that a third of us be­li­eve that a di­et ab­so­lu­tely free of fat-a nut­ri­ent, lest you for­get, es­sen­ti­al to our sur­vi­val-wo­uld be bet­ter for us than a di­et con­ta­ining even just “a pinch” of it. In one ex­pe­ri­ment, he sho­wed the words “cho­co­la­te ca­ke” to a gro­up of Ame­ri­cans and re­cor­ded the­ir word as­so­ci­ati­ons. “Gu­ilt” was the top res­pon­se. If that stri­kes you as unex­cep­ti­onal, con­si­der the res­pon­se of the French eaters to the sa­me prompt: “ce­leb­ra­ti­on.” (Oh, ye­ah.) I think of Ro­zin as a kind of psycho­analyst of nut­ri­ti­onism.

  A few ye­ars ago, Ro­zin pre­sen­ted a gro­up of Ame­ri­cans with the fol­lo­wing sce­na­rio: “Assu­me you are alo­ne on a de­sert is­land for one ye­ar and you can ha­ve wa­ter and one ot­her fo­od. Pick the fo­od that you think wo­uld be best for yo­ur he­alth.”

  The cho­ices we­re corn, al­fal­fa spro­uts, hot dogs, spi­nach, pe­ac­hes, ba­na­nas, and milk cho­co­la­te. The most po­pu­lar cho­ice was ba­na­nas (42 per­cent), fol­lo­wed by spi­nach (27 per­cent), corn (12 per­cent), al­fal­fa spro­uts (7 per­cent), pe­ac­hes (5 per­cent), hot dogs (4 per­cent), and milk cho­co­la­te (3 per­cent). Only 7 per­cent of the par­ti­ci­pants cho­se one of the two fo­ods that wo­uld in fact best sup­port sur­vi­val: hot dogs and milk cho­co­la­te.

  Evi­dently so­me of the wrec­ka­ge of the li­pid hypot­he­sis has was­hed up on Ro­zin’s de­sert is­land.

  “Fat,” he wri­tes, “se­ems to ha­ve as­su­med, even at low le­vels, the ro­le of a to­xin” in our di­etary ima­gi­na­ti­ons. I won­der why. As Ro­zin po­ints out, “Wor­rying so much abo­ut fo­od can’t be very go­od for yo­ur he­alth.” In­de­ed. Ort­ho­re­xia ner­vo­sa is an eating di­sor­der not yet re­cog­ni­zed by the DSM-IV, but so­me psycho­lo­gists ha­ve re­cently sug­ges­ted that it’s ti­me it was. They’re se­e­ing mo­re and mo­re pa­ti­ents suf­fe­ring from “an un­he­althy ob­ses­si­on with he­althy eating.”

  So this is what put­ting sci­en­ce, and sci­en­tism, in char­ge of the Ame­ri­can di­et has got­ten us: an­xi­ety and con­fu­si­on abo­ut even the most ba­sic qu­es­ti­ons of fo­od and he­alth, and a ste­adily di­mi­nis­hing abi­lity to enj­oy one of the gre­at ple­asu­res of li­fe wit­ho­ut gu­ilt or ne­uro­sis.

  But whi­le nut­ri­ti­onism has its ro­ots in a sci­en­ti­fic ap­pro­ach to fo­od, it’s im­por­tant to re­mem­ber that it is not a sci­en­ce but an ide­ology, and that the fo­od in­dustry, jo­ur­na­lism, and go­vern­ment be­ar just as much res­pon­si­bi­lity for its con­qu­est of our minds and di­ets. All three hel­ped to amp­lify the sig­nal of nut­ri­ti­onism: jo­ur­na­lism by unc­ri­ti­cal­ly re­por­ting the la­test di­etary stu­di­es on its front pa­ges; the fo­od in­dustry by mar­ke­ting du­bi­o­us fo­od­li­ke pro­ducts on the ba­sis of te­nu­o­us he­alth cla­ims; and the go­vern­ment by ta­king it upon it­self to is­sue of­fi­ci­al di­etary ad­vi­ce ba­sed on sketchy sci­en­ce in the first pla­ce and cor­rup­ted by po­li­ti­cal pres­su­re in the se­cond. The no­vel fo­od pro­ducts the in­dustry de­sig­ned ac­cor­ding to the la­test nut­ri­ti­onist specs cer­ta­inly hel­ped push re­al fo­od off our pla­tes. But the in­dustry’s inf­lu­en­ce wo­uld not be ne­arly so gre­at had the ide­ology of nut­ri­ti­onism not al­re­ady un­der­mi­ned the inf­lu­en­ce of tra­di­ti­on and ha­bit and com­mon sen­se-and the trans­mit­ter of all tho­se va­lu­es, mom-on our eating.

  Now, all this might be to­le­rab­le if eating by the light of nut­ri­ti­onism ma­de us, if not hap­pi­er, then at le­ast he­alt­hi­er. That it has fa­iled to do. Thirty ye­ars of nut­ri­ti­onal ad­vi­ce ha­ve left us fat­ter, sic­ker, and mo­re po­orly no­uris­hed. Which is why we find our­sel­ves in the pre­di­ca­ment we do: in ne­ed of a who­le new way to think abo­ut eating.

  Part II - THE WESTERN DIET AND THE DISEASES OF CIVILIZATION

  ONE - THE ABORIGINE IN ALL OF US

  I n the sum­mer of 1982, a gro­up of ten mid­dle-aged, over­we­ight, and di­abe­tic Abo­ri­gi­nes li­ving in set­tle­ments ne­ar the town of Derby, Wes­tern Aust­ra­lia, ag­re­ed to par­ti­ci­pa­te in an ex­pe­ri­ment to see if tem­po­ra­rily re­ver­sing the pro­cess of wes­ter­ni­za­ti­on they had un­der­go­ne might al­so re­ver­se the­ir he­alth prob­lems. Sin­ce le­aving the bush so­me ye­ars be­fo­re, all ten had de­ve­lo­ped type 2 di­abe­tes; they al­so sho­wed signs of in­s
u­lin re­sis­tan­ce (when the body’s cells lo­se the­ir sen­si­ti­vity to in­su­lin) and ele­va­ted le­vels of trigly­ce­ri­des in the blo­od-a risk fac­tor for he­art di­se­ase. “Me­ta­bo­lic syndro­me,” or “syndro­me X,” is the me­di­cal term for the comp­lex of he­alth prob­lems the­se Abo­ri­gi­nes had de­ve­lo­ped: Lar­ge amo­unts of re­fi­ned car­bohyd­ra­tes in the di­et com­bi­ned with a se­den­tary li­festy­le had di­sor­de­red the int­ri­ca­te (and still im­per­fectly un­ders­to­od) system by which the in­su­lin hor­mo­ne re­gu­la­tes the me­ta­bo­lism of car­bohyd­ra­tes and fats in the body. Me­ta­bo­lic syndro­me has be­en imp­li­ca­ted not only in the de­ve­lop­ment of type 2 di­abe­tes, but al­so in obe­sity, hyper­ten­si­on, he­art di­se­ase, and pos­sibly cer­ta­in can­cers. So­me re­se­arc­hers be­li­eve that me­ta­bo­lic syndro­me may be at the ro­ot of many of the “di­se­ases of ci­vi­li­za­ti­on” that typi­cal­ly fol­low a na­ti­ve po­pu­la­ti­on’s adop­ti­on of a Wes­tern li­festy­le and the nut­ri­ti­on tran­si­ti­on that typi­cal­ly en­ta­ils.

  The ten Abo­ri­gi­nes re­tur­ned to the­ir tra­di­ti­onal ho­me­land, an iso­la­ted re­gi­on of north­west Aust­ra­lia mo­re than a day’s dri­ve by off-ro­ad ve­hic­le from the ne­arest town. From the mo­ment they left ci­vi­li­za­ti­on, the men and wo­men in the gro­up had no ac­cess to sto­re fo­od or be­ve­ra­ges; the idea was for them to rely exc­lu­si­vely on fo­ods they hun­ted and gat­he­red them­sel­ves. (Even whi­le li­ving in town, they still oc­ca­si­onal­ly hun­ted tra­di­ti­onal fo­ods and so had pre­ser­ved the know­led­ge of how to do so.) Ke­rin O’Dea, the nut­ri­ti­on re­se­arc­her who de­sig­ned the ex­pe­ri­ment, ac­com­pa­ni­ed the gro­up to mo­ni­tor and re­cord its di­etary in­ta­ke and ke­ep tabs on the mem­bers’ he­alth.

 

‹ Prev