Why the whole-genome test fails

Los Angeles Times - - OP-ED - By H. Gil­bert Welch and Wylie Burke H. Gil­bert Welch is a pro­fes­sor of medicine at Dart­mouth Med­i­cal School and the au­thor of “Less Medicine, More Health: 7 As­sump­tions that Drive Too Much Med­i­cal Care.” Wylie Burke is a med­i­cal ge­neti­cist and pro­fes­sor

Pres­i­dent Obama pro­poses to plunk down $215 mil­lion on “pre­ci­sion medicine,” and the Na­tional In­sti­tutes of Health and its Na­tional Can­cer In­sti­tute will spend it by se­quenc­ing the whole genome of a mil­lion or more Amer­i­cans.

Is whole-genome testing the path to health? The short an­swer is no.

The main prob­lem with the pro­posal is that the re­search is bound to pro­duce more noise than sig­nal. The is­sue isn’t ge­net­ics but “big” data. The ba­sic idea of pre­ci­sion medicine is to look for pat­terns in the genome that seem to travel with prob­lems we all care about: di­a­betes, heart dis­ease, can­cer and de­men­tia. But there are a lot of pos­si­ble pat­terns to look for.

Imag­ine there were only 10 data points in a per­son’s genome and that each point could only take on one of two val­ues: red or green. The first could be red or green, the sec­ond red or green and so forth. The num­ber of pos­si­ble pat­terns that could emerge from those 10 data points is 2 to the 10th power — 1,024 pat­terns. It goes up fast from there: 20 data points would pro­duce more than 1 mil­lion pat­terns.

Among the 3 bil­lion data points in the genome, ge­neti­cists be­lieve that each of us has at least hun­dreds — prob­a­bly thou­sands — that vary in ways that could af­fect our health. So the num­ber of po­ten­tial pat­terns is as­tro­nom­i­cal.

With all those pat­terns to ex­plore in a mil­lion or more Amer­i­cans — who have hun­dreds of di­ag­noses — the ini­tia­tive is bound to iden­tify a whole bunch of ap­par­ent re­la­tion­ships.

A few years ago statis­ti­cians in Canada high­lighted the big-data prob­lem. They looked at the re­la­tion­ship be­tween as­tro­log­i­cal signs and hos­pi­tal di­ag­noses in 5 mil­lion ad­mis­sions in On­tario. They found that Leos were sig­nif­i­cantly more likely to be ad­mit­ted for di­ges­tive tract bleed­ing and that Sagit­tar­i­ans were sig­nif­i­cantly more likely to be ad­mit­ted for up­per-arm frac­tures.

Not that any­one ac­tu­ally be­lieves that as­tro­log­i­cal signs are re­spon­si­ble for spe­cific hos­pi­tal di­ag­noses. (OK, maybe some­one does, but we don’t be­lieve him.)

Too much noise is only one part of the big-data prob­lem. What­ever sig­nals are found are bound to be weak.

There are some strong sig­nals in the genome, such as rare mu­ta­tions that in­crease the risk of colon can­cer some ten­fold. And there may be value in se­lec­tively look­ing for th­ese sig­nals. But we don’t need to test the whole genome to find them.

For the most part, th­ese sig­nals are al­ready known; they didn’t re­quire big data. If your mother, your mother’s mother, two of your three sis­ters and half of your cousins all had a dis­ease — a ge­netic ba­sis was ob­vi­ous. Breast can­cer mu­ta­tions, for ex­am­ple, were iden­ti­fied in such fam­i­lies.

So with much of the low-hang­ing fruit gone, what’s left for pre­ci­sion medicine? Ge­netic pat­terns that are as­so­ci­ated with changes in risk of a few per­cent­age points — for ex­am­ple, in­creas­ing your chance of de­vel­op­ing Dis­ease X from 4% to 6%.

Make no mis­take about it: Th­ese data will scare peo­ple — par­tic­u­larly since they are likely to be framed as a 50% in­crease in your risk of Dis­ease X. But it’s just as likely they won’t make a dif­fer­ence in your health.

That’s be­cause weak sig­nals can’t tell us what to do. For ex­am­ple, what should doc­tors do dif­fer­ently for some­one with a slightly higher risk for colon can­cer and Alzheimer’s dis­ease and a slightly lower risk of lung can­cer and di­a­betes? In­crease the fre­quency of screen­ing and Su­doku puzzles? Tell him it’s fine to smoke and gain weight?

And yet there will be sub­stan­tial pres­sure for doc­tors to do some­thing, which typ­i­cally trans­lates into a hap­haz­ard process of more testing and more in­ter­ven­tion. Some in­ter­ven­tions may help a lit­tle; oth­ers may harm peo­ple; still oth­ers may have no im­pact at all.

The re­al­ity is that it’s hard to pre­dict the fu­ture, and even harder to know what to do about it.

And don’t make the mis­take of think­ing whole-genome data are go­ing to save the health­care sys­tem money, no mat­ter how cheap se­quenc­ing be­comes. Pa­tients will waste doc­tors’ time chas­ing the weak sig­nals; doc­tors will then waste med­i­cal re­sources “do­ing some­thing.” Will your in­sur­ance cover all of this? Don’t count on it.

But cost isn’t the only is­sue. No amount of ge­netic data will ever change the fact that most of us need to eat bet­ter and move more. Yet ex­plain­ing “pre­dic­tive” data in a 10- to 15-minute of­fice visit will di­vert at­ten­tion from more im­por­tant top­ics. Equally im­por­tant is the at­ten­tion of the med­i­cal re­search en­ter­prise, where all this em­pha­sis on pre­dict­ing dis­ease will di­vert re­searchers from more fun­da­men­tal ques­tions: How do we best treat clin­i­cally ev­i­dent dis­ease? How do we pre­vent it?

Fur­ther­more, the fo­cus on molec­u­lar pathol­ogy only re­in­forces the as­sump­tion that health is a set of spec­i­fied phys­i­cal mea­sure­ments and the ab­sence of any de­tectable ab­nor­mal­ity. Be­cause all of us har­bor ab­nor­mal­i­ties, and be­cause our tech­nolo­gies are in­creas­ingly able to find them, the dis­turb­ing truth is that this is a recipe for mak­ing all of us feel sick.

Of course there is a lot of en­thu­si­asm for the ef­fort. Aca­demic med­i­cal cen­ters and biotech­nol­ogy firms, in par­tic­u­lar, love this stuff. For the med­i­cal-industrial com­plex, it is po­ten­tially a huge rev­enue stream. For most peo­ple, how­ever, whole-genome se­quenc­ing is an ab­surd med­i­cal test. If you get en­ter­tain­ment value from a for­tune-teller read­ing — or from a whole genome — that’s one thing. But if you value your health, di­vert your re­sources to some­thing more mean­ing­ful — maybe whole foods.

Newspapers in English

Newspapers from USA

© PressReader. All rights reserved.