Dis­ease of flawed in­sights

In­de­pen­dent verification of med­i­cal re­search is a must to en­sure that the find­ings are cor­rect

Down to Earth - - COLUMN - DOWN TO EARTH

Oand minds are pris­on­ers of mod­ern UR BOD­IES science. Not a day goes by with­out news­pa­pers, TV chan­nels and the In­ter­net ply­ing us with the lat­est med­i­cal “in­sights” into the work­ings of mal­adies such as di­a­betes, or the promis­ing re­sults of a clin­i­cal trial of a drug for Alzheimer’s.

As most such stud­ies are pub­lished in peer-re­viewed jour­nals of re­pute, even peers and prac­tis­ing physi­cians trust them,let alone the lay public.

How­ever,the truth is shock­ingly con­tra­dic­tory. In the last one decade, a se­ries of metas­tud­ies (study of stud­ies) has re­vealed that a large num­ber of stud­ies, es­pe­cially in med­i­cal science and psy­chol­ogy, are ei­ther shoddy or sta­tis­ti­cally ma­nip­u­lated, and hence false.

John Ioan­ni­dis, an epi­demi­ol­o­gist at Stan­ford Uni­ver­sity, was the first to blow the whis­tle in his provoca­tively ti­tled 2005 pa­per, Why Most Pub­lished Re­search Find­ings are False. He demon­strated that in stud­ies characterised by re­searcher prej­u­dice, greater flex­i­bil­ity in de­sign and tech­nique, high glam­our quo­tient, and high fi­nan­cial stakes (all of which is true of most stud­ies), the re­searcher is likely to come up with an er­ro­neous re­sult. In 2012, re­searchers at a US biotech firm, Am­gen,re­ported that they could repli­cate no more than six of the 53 “land­mark” stud­ies in can­cer re­search.Like­wise, in 2011, re­searchers at Bayer found they could not ver­ify more than 20-25 per cent of the re­search pa­pers on can­cer, car­dio­vas­cu­lar and women’s health stud­ies. Omi­nously, be­tween 2000 and 2010,no fewer than 80,000 pa­tients en­rolled in clin­i­cal tri­als based on re­search that was later re­tracted be­cause of

er­rors or fraud.

Ar­guably,that is only the tip of the ice­berg.Most drug com­pa­nies are ex­tremely se­cre­tive about data on clin­i­cal tri­als. Ben Goldacre, a Bri­tish doc­tor and au­thor of the best­seller Bad Science, is cam­paign­ing for pe­nal­is­ing them if they don’t make full dis­clo­sure.

To ver­ify, and not sim­ply trust, a re­search find­ing is the lynch­pin of mod­ern science’s pre­rog­a­tive to ob­jec­tive truth. That sci­en­tists are not in­ter­ested in repli­cat­ing the find­ings of their peers is be­hind the cur­rent cri­sis of fraud­u­lent re­search.

Sta­tis­ti­cal er­rors, in­ad­ver­tent or oth­er­wise, is just one of many fac­tors that per­pet­u­ate bad science.Peer re­view­ers have been found guilty of over­sight when it comes to spot­ting er­rors. To make mat­ters worse, pro­fes­sional ri­valry cou­pled with com­pe­ti­tion for a shrink­ing pool of funds and jobs en­cour­ages a pub­lish-or-per­ish cul­ture, which clearly af­fects the qual­ity of pa­pers pub­lished.

There is dan­ger that some of th­ese shoddy sci­en­tific claims might get re­flected in med­i­cal prac­tice. Be­sides, it also means that pre­cious funds, a sig­nif­i­cant amount of which comes from the tax­pay­ers’ pocket, are squan­dered on bad or use­less re­search.

Alarmed by the scale of the prob­lem, some or­gan­i­sa­tions have be­gun to put repli­ca­tion back on the ta­ble. No­tably, in 2012, plos one, an open ac­cess jour­nal, launched the Re­pro­ducibil­ity Ini­tia­tive,a ser­vice through which life sci­en­tists can get their work ver­i­fied by an in­de­pen­dent lab for a fee.

Un­for­tu­nately, even as In­dia ex­pands its med­i­cal re­search foot­print, there are no signs of alarm over this wor­ri­some is­sue.How much of In­dian med­i­cal re­search is ver­i­fied is any­body’s guess.But it’s high time the sci­en­tific fra­ter­nity started a public de­bate on this sub­ject that has bear­ing not only on the qual­ity of re­search but also on the lives of mil­lions of In­di­ans.

Ioan­ni­dis is all for mak­ing it a public af­fair but he is not very san­guine that this will hap­pen quickly, though. He told The At­lantic mag­a­zine:“It’s dif­fi­cult to change the way that ev­ery­day doc­tors, pa­tients, and healthy peo­ple think and be­have.”

Hope he’s wrong.

TARIQUE AZIZ / CSE

Newspapers in English

Newspapers from India

© PressReader. All rights reserved.