State of the Na­tion

The Hu­man Fac­tor

AQ: Australian Quarterly - - CONTENTS - DR ALAN FINKEL

Let’s imag­ine for a mo­ment that we’re meet­ing as guests, at the an­nual fam­ily bar­be­cue, in Un­cle Bob’s back­yard.

The topic will turn, as it al­ways does, to fam­ily his­tory. A few of the re­tirees will men­tion, as they al­ways do, that they’ve looked into it. And we’ll be sur­prised, as we al­ways are, to dis­cover that we’re prob­a­bly in line to the throne, and cer­tainly de­scended from roy­alty.

Then some­one will re­veal that she’s just re­ceived the re­sults of her ge­netic an­ces­try pro­file… a gift from the kids, for her birth­day… and sud­denly we’re all com­par­ing notes on what per­cent­age of our fam­ily genome is Ne­an­derthal DNA.

Yes, it's the Brave New World of di­rect-to­con­sumer ge­netic test­ing – and it's com­ing to a bar­be­cue near you. Al­ready, the com­pa­nies that of­fer the ser­vice turn an es­ti­mated $100 mil­lion in an­nual global sales. On some es­ti­mates, the mar­ket will climb to at least $300 mil­lion in just five years' time.

That growth has been kicked along by some very de­ter­mined mar­ket­ing. My staff even spot­ted at least one

com­pany spruik­ing a test­ing kit as the ideal Fa­ther's Day gift. It could cer­tainly add some in­ter­est­ing com­pli­ca­tions to the fam­ily's cel­e­bra­tions of Fa­ther's Days forever­more.

The same com­pany has now en­tered into a part­ner­ship with the mu­sic stream­ing ser­vice Spo­tify, so that your re­sults will come pack­aged with a cus­tomised playlist of your an­ces­tral mu­sic.

But it's not just the al­lure of dis­tant Vik­ing an­ces­tors and un­sus­pected sib­lings that draws us in.

If you're try­ing to lose weight, you can go to the chemist and pick up a diet spit kit. Just send it off for a diet plan for your DNA. If you're look­ing for ro­mance, you can ac­cess ‘ge­netic match' dat­ing ser­vices. If you've run out of hu­mans in your fam­ily to test, you can pay for a ge­netic pro­file of your pet.

To me, it all begs the ques­tions: how is di­rect-to-con­sumer ge­netic test­ing chang­ing the way we think about the pos­si­bil­i­ties of ge­netic re­search? And how pre­pared are we, as in­di­vid­u­als and as a com­mu­nity, to make re­spon­si­ble calls?

Science For Sale

As some­one with the good luck to be im­mersed in the med­i­cal re­search com­mu­nity in Aus­tralia, I mar­vel at the fu­tures un­fold­ing be­fore my eyes. The sci­en­tists drawn to this field know how good you have to be, and how hard you have to work, to push the lim­its of pos­si­bil­ity from within the stric­tures of ethics.

Like all sci­en­tists, they would prob­a­bly like to as­sume that ev­ery­one can eas­ily tell the dif­fer­ence be­tween – for ex­am­ple – the ge­netic re­search done at a glob­ally-cel­e­brated lab­o­ra­tory like the Wal­ter and El­iza Hall In­sti­tute in Mel­bourne, and the ‘DNA match' ser­vice of­fered by a com­pany called ‘DNA Ro­mance'.

But for non-sci­en­tists, I sus­pect it's not so clear-cut at all, par­tic­u­larly given that the or­gan­i­sa­tions they're likely to hear from first are those with the loud­est mar­ket­ing.

Try it. Type ‘ge­netic test’ into Google. You’ll be di­rected to com­pa­nies: an­ces­try tests, diet ad­vice and yes, DNA dat­ing.

So we have to as­sume that there are peo­ple form­ing their per­cep­tions and mak­ing de­ci­sions right now, on the ba­sis of in­for­ma­tion gleaned from the in­ter­net that is in­ac­cu­rate, un­help­ful, or in­com­plete

Try it. Type ‘ge­netic test' into Google. You'll be di­rected to com­pa­nies: an­ces­try tests, diet ad­vice and yes, DNA dat­ing.

If you click on a link, you'll go to a web­site with lists of papers ap­par­ently pub­lished in ac­tual sci­en­tific jour­nals. You'll also find tabs for ‘Fre­quently Asked Ques­tions', defin­ing terms like ‘vari­ant' and ‘genome se­quenc­ing'.

So the science will be ex­plained to you by com­pa­nies. And it may well be the only user-friendly ex­pla­na­tion of your bi­o­log­i­cal build­ing blocks that you've ever re­ceived.

Now a cynic might con­clude that many of th­ese com­pa­nies are more in­ter­ested in look­ing like science than ac­tu­ally do­ing it.

But the same cynic would also have to be mind­ful that at least some of the more cred­i­ble com­pa­nies do have an ob­vi­ous fi­nan­cial in­ter­est in con­tribut­ing to le­git­i­mate and im­por­tant re­search.

We got some idea of just how sig­nif­i­cant that fi­nan­cial in­ter­est might be in July 2018, when the phar­ma­ceu­ti­cal gi­ant Glax­o­smithk­line ac­quired a $300 mil­lion stake in the ge­netic test­ing com­pany 23andme. In ex­change, 23andme handed over ex­clu­sive rights to its cus­tomer data­base for Glax­o­smithk­line to use for drug dis­cov­ery. That in­cludes the ge­netic in­for­ma­tion of the Aus­tralians who willingly parted with a sam­ple of their saliva, and paid 23andme for the priv­i­lege.

Other ge­netic test­ing com­pa­nies are look­ing to cut sim­i­lar deals, and the phar­ma­ceu­ti­cal in­dus­try is ac­tively court­ing them.

23andme is the poster-child of the di­rect-to-con­sumer ge­netic test­ing in­dus­try. Its prac­tices and poli­cies are closely scru­ti­nised, of­ten, as a proxy for a big­ger de­bate we're strug­gling to grasp, around big data, on­line pri­vacy, and Sil­i­con Val­ley.

To the com­pany's credit, its data pro­to­cols are made clear on its web­site, and in its con­sent form. Cus­tomers are ad­vised that “on av­er­age, a cus­tomer who chooses to opt in to re­search con­trib­utes to over 230 stud­ies on top­ics that range from Parkin­son's dis­ease to lu­pus to asthma and more.”

But 23andme is not the only com­pany you might en­counter via Google – and not ev­ery­one in this busi­ness is go­ing to be scrupu­lous, or trans­par­ent. Not ev­ery­one has an in­ter­est in safely han­dling your data. And not ev­ery­one par­tic­u­larly cares about the qual­ity of the ad­vice they pro­vide.

So we have to as­sume that there are peo­ple form­ing their per­cep­tions and mak­ing de­ci­sions right now, on the ba­sis of in­for­ma­tion gleaned from the in­ter­net that is in­ac­cu­rate, un­help­ful, or in­com­plete. I worry about what hap­pens when some of those peo­ple

are in­evitably burned.

To­day, we can say that most Aus­tralians are very trust­ing. We have the ev­i­dence of Re­search Aus­tralia's re­cently pub­lished sur­vey.

75% of re­spon­dents said they would be will­ing to use ge­netic test­ing to iden­tify the most ef­fec­tive drug to treat their con­di­tion. 22% would need more in­for­ma­tion to make a de­ci­sion. Only 3% gave a flat out no. Of that 75% who would be will­ing, 95% said they would be will­ing for their re­sults to be used to im­prove the treat­ments for fu­ture pa­tients. 95%. Just 1% said no.

That says some­thing deeply en­cour­ag­ing about hu­man be­ings. We re­ally do care. And it sig­nals some­thing else that's en­cour­ag­ing, as well: Aus­tralians have learned to trust re­searchers.

They see a sec­tor that con­ducts it­self as it does in this coun­try: with in­tegrity, with trans­parency, and most im­por­tantly, to bril­liant ef­fect.

But that hard-earned trust is eas­ily abused.

And we've seen it so many times be­fore. A sci­en­tific field be­comes gen­uinely ex­cit­ing – and in­stantly, the le­git­i­mate re­search is jostling for space with snake-oil.

And we've seen it so many times be­fore. A sci­en­tific field be­comes gen­uinely ex­cit­ing – and in­stantly, the le­git­i­mate re­search is jostling for space with snake-oil.

I'm think­ing of stem cell re­search – and the num­ber of clin­ics that leapt at the chance to take peo­ple's life sav­ings for treat­ments that were at best un­help­ful… and of­ten, much worse.

As an old neu­ro­science PHD, I'm think­ing of brain stim­u­la­tion. It's gen­uinely in­ter­est­ing for the clin­i­cal treat­ment of Parkin­son's. It's not par­tic­u­larly use­ful to buy what's ba­si­cally a ten­nis head­band with bat­ter­ies and elec­trodes, from a dodgy com­pany that tells you that zap­ping your brain will mag­i­cally boost your in­tel­li­gence.

But peo­ple buy it. And le­git­i­mate re­search suf­fers the col­lat­eral dam­age.

The next big scan­dal might not be a com­pany sell­ing non­sense. It might be a start-up that's care­less with data se­cu­rity. Or it might be a con­sumer who takes a ge­netic test in good faith, pos­si­bly from an overseas provider, with­out aware­ness of the pos­si­ble im­pli­ca­tions un­der Aus­tralian law.

I am think­ing in par­tic­u­lar of the pos­si­bil­ity that ge­netic data could be re­quested by life in­sur­ers. As I un­der­stand the cur­rent po­si­tion, health in­sur­ers can­not ad­just their pre­mi­ums on the ba­sis of ge­netic in­for­ma­tion. But life in­sur­ers can re­quest and re­quire dis­clo­sure, and they can dis­crim­i­nate by rais­ing the pre­mium or deny­ing cov­er­age al­to­gether, as long as there is a rea­son­able jus­ti­fi­ca­tion.

Is this a prob­lem?

We know that life in­sur­ance is not

a triv­ial con­sid­er­a­tion for peo­ple who might make the choice to be tested. We have the ev­i­dence of a Vic­to­rian study that looked at peo­ple's will­ing­ness to be tested for a marker of Lynch syn­drome, or col­orec­tal can­cer. When life in­sur­ance was men­tioned on the con­sent form, the num­ber of peo­ple de­clin­ing the test in­creased from fewer than a fifth, to nearly half.

As a re­search com­mu­nity, my colleagues ob­vi­ously have to think about what in­for­ma­tion is owed to any­one who par­tic­i­pates in a ge­netic study as part of their duty of care.

But they, and I, also have to think very hard about the larger le­gal frame­work, be­cause if oth­ers aren't eth­i­cal, and peo­ple suf­fer loss, then the rep­utable op­er­a­tors are just as ex­posed to the blow­back.

The UK has had a mora­to­rium on the use of pre­dic­tive ge­netic test re­sults by life in­sur­ers for close to two decades. Canada has gone a step fur­ther, pro­hibit­ing in­sur­ers from re­quest­ing or re­quir­ing dis­clo­sure of past or fu­ture test re­sults, by leg­is­la­tion.

In Aus­tralia, the Par­lia­men­tary Joint Com­mit­tee set up to in­quire into the life in­sur­ance in­dus­try has called for the same ap­proach in Aus­tralia – look­ing first at the in­dus­try Code of Prac­tice, and then, po­ten­tially, the changes that might be nec­es­sary in law.

The re­search com­mu­nity has led this dis­cus­sion from the out­set, con­scious that what­ever de­ci­sions we come to, what mat­ters is to be on the front foot – not forced to re­spond in a cri­sis.

But where, for sci­en­tists, to be­gin?

Ethics and Op­por­tu­ni­ties

Let me sug­gest that be­fore we even

think about a spe­cific model, we start with a ba­sic ques­tion.

What do we need to do to earn the com­mu­nity's trust?

I want to em­pha­sise those words, “earn trust”.

I've read many doc­u­ments called

If oth­ers aren’t eth­i­cal, and peo­ple suf­fer loss, then the rep­utable op­er­a­tors are just as ex­posed to the blow­back.

‘out­reach strate­gies' over the years. I of­ten get the im­pres­sion that their aim is to ‘per­suade peo­ple' – which isn't the same as earn­ing trust. ‘Per­suad­ing' im­plies that there's a per­fect set of words that will sud­denly make ev­ery­thing clear for all those peo­ple who just think the wrong way.

But there's never a per­fect set of words. That's be­cause it's not about elo­quence. It's about peo­ple's lived ex­pe­ri­ence. We earn trust when peo­ple can see that we're de­liv­er­ing the out­comes we prom­ise – and they are ben­e­fit­ting.

There are three ba­sic prin­ci­ples that I think are help­ful to keep in mind.

The first and most im­por­tant is – don't be greedy. Don't ask more from peo­ple than they're com­fort­able to ac­cept.

That was the les­son that Ice­land learned with ge­netic data in the early 2000s. At the time, my com­pany Axon In­stru­ments had just gone to mar­ket with a DNA mi­cro-ar­ray scan­ner called Genepix, for scan­ning and in­ter­pret­ing gene ex­pres­sion ar­rays. In that con­text, my aware­ness was high when buzz started build­ing around a promis­ing Ice­landic start-up, called DE­CODE.

Its CEO had worked out that Ice­land could be ex­tremely at­trac­tive as a hunt­ing ground for ge­netic fac­tors.

Ice­landers are a ‘founder pop­u­la­tion': ge­o­graph­i­cally iso­lated, and ge­net­i­cally ho­moge­nous. The peo­ple know a lot about their an­ces­try. They're fa­mous for it. And the pub­lic sec­tor was sit­ting on ex­tremely de­tailed health records go­ing back to the nine­teenth cen­tury.

A data gold­mine!

So DE­CODE said to the Prime Min­is­ter of Ice­land: wouldn't it be a great idea to build a na­tional health sec­tor data­base, and sell the ac­cess rights to com­pa­nies and in­sti­tutes all over the world?

And sweeten the deal for in­vestors by in­te­grat­ing all the health records, all the ge­netic data, and all the ge­nealog­i­cal in­for­ma­tion?

And run it on a model of ‘pre­sumed con­sent' – with all the records of dead peo­ple au­to­mat­i­cally in­cluded, be­cause by def­i­ni­tion, at that point they couldn't refuse?

Sci­en­tists know that our Aus­tralian re­searchers over­whelm­ingly do the right thing: they will pull them­selves up short, even when the rules might tech­ni­cally al­low them.

It only takes one breach to make a big re­form un­touch­able – for years, if not decades. And it’s that much harder to rally the troops and try again. DON’T BE GREEDY. DON’T ASK MORE FROM PEO­PLE THAN THEY’RE COM­FORT­ABLE TO AC­CEPT.

And hand over the ex­clu­sive con­tract for build­ing it and op­er­at­ing it to DE­CODE?

Now Ice­landers had grown very com­fort­able over the years with the idea of the gov­ern­ment hold­ing their health records. But this was a model that many just wouldn't stom­ach – and the project fell apart when cit­i­zens took the mat­ter to court.

It was even­tu­ally re­assem­bled in a more mod­est form – but by then, the dam­age was done. And it wasn't con­fined to Ice­land. To this day, health bu­reau­crats point to Ice­land when any­one pro­poses to con­sol­i­date health records. Here Be Dragons. The bet­ter les­son to take is that it's im­por­tant to let trust build up with fa­mil­iar­ity, over time. Some­times that means scal­ing back our am­bi­tions, and ac­cept­ing that we can't take full ad­van­tage of the ex­pan­sive tools at our fin­ger­tips.

Sci­en­tists know that our Aus­tralian re­searchers over­whelm­ingly do the right thing: they will pull them­selves up short, even when the rules might tech­ni­cally al­low them to do some­thing that they know is in­con­sis­tent with their fun­da­men­tal duty of care.

But we can't as­sume that ev­ery­one will be as scrupu­lous. And we can't ask the com­mu­nity to take our as­sur­ances on trust.

So if the pro­tec­tions in the sys­tem to­day are im­plicit, then we have to go the ex­tra mile to make them ex­plicit.

We also have to be hum­ble enough to give our­selves time: time to un­der­stand just what new risks we might cre­ate when we tran­si­tion to a new model.

So many times I've seen project teams ex­cited about the po­ten­tial of cen­tral­is­ing data, and mak­ing it search­able, and en­sur­ing it's easy to ac­cess, and max­imis­ing the cov­er­age. All good things – but they come with dif­fer­ent chal­lenges that need to be an­tic­i­pated and man­aged.

That in­cludes the mag­ni­fy­ing ef­fect of AI – ar­ti­fi­cial in­tel­li­gence.

With AI, you don't just have a myr­iad of data points on a given per­son. Now you've got the ca­pac­ity to join up the dots, at speed, and at scale, in ways that the peo­ple who com­piled the data sim­ply didn't an­tic­i­pate.

It only takes one breach to make a big re­form un­touch­able – for years, if not decades. And it's that much harder to rally the troops and try again.

So we need to be pre­pared to rein in the ex­cite­ment, to find a level of pro­tec­tion that gives con­fi­dence to the com­mu­nity, and clar­ity to real sci­en­tists do­ing eth­i­cal re­search.

The sec­ond prin­ci­ple fol­lows on from the first: make the ben­e­fits vis­i­ble.

Ice­land broke this rule, too: they de­signed the data­base to be as ap­peal­ing as pos­si­ble to for­eign ven­ture cap­i­tal­ists.

They suc­ceeded – but the for­eign ven­ture cap­i­tal­ists weren't the prob­lem. It's a pat­tern we've seen many times be­fore. Again and again, we lose the ar­gu­ment be­cause peo­ple don't see ben­e­fits for them­selves.

Take GM foods.

Who ben­e­fits from a to­mato that lasts on the shelf for a month – in­stead of a week?

It's the grow­ers, and the re­tail­ers – not the con­sumer.

And you can bet that no su­per­mar­ket on Earth will tell its cus­tomers that – thanks to science – they can now en­joy month-old toma­toes.

So cus­tomers may be 99% sure that GM foods are safe… but why would you ac­cept any risk, if you car­ried the

IM­AGE: © 23andme.com

IM­AGE: © thierry ehrmann-flickr

Newspapers in English

Newspapers from Australia

© PressReader. All rights reserved.