Page 1

            Lloyd    Hughes    

Can We  Rely  On  Science?  

[Type the  abstract  of  the  document  here.  The  abstract  is  typically  a  short  summary  of  the               of     the  document.]              Scottish  Universities  Medical  Journal   contents  

                                                                                     Special  Issue  Supplement                                                

C o r r e s p o n d e n c e t o :   S U M J @ d u n d e e . a c . u k            


Editorial Board   Special  Edition  Editor:    Lloyd  Hughes  [5th  Year  Medical  Student;  University  of  Dundee   Medical  School;  Editor-­‐in-­‐Chief  2011-­‐13]     Editor  in  Chief:  Laura  Fraser  [5th  year  medical  student,  University  of  Dundee]     Deputy  Editor:  Kevin  Barr  [3rd  year  mental  health  nursing  student,  University  of  Dundee]     IT  Manager:  Naomi  McIlvenny  [5th  year  medical  student,  University  of  Dundee]     Head  Reviewer:  Rebecca  Grant  [1st  year  medical  student,  University  of  St  Andrews]     Features  Commissioner:  Lauren  Copeland  [BMSc  year  medical  student,  University  of   Dundee]     National  PR  Representative:  James  Millar  [Academic  Foundation  Doctor,  NHS  Grampian]    

Associate Editors  and  Consultant  Reviewers  for  Dundee  SUMJ   Dr  David  Booth  -­‐  [General  Practitioner  &  Doctors  Patients  and  Communities  Facilitator,   University  of  Dundee]     Prof  Jeremy  Hughes  -­‐  [Consultant  Nephrologist  NHS  Lothian,  Reader  in  Nephrology   University  of  Edinburgh]     Dr  Hannah  Lord  -­‐  [Consultant  Oncologist,  Ninewells  Hospital  and  Medical  School]     Lloyd  Hughes  -­‐  [5th  Year  Medical  Student;  University  of  Dundee  Medical  School,  Editor-­‐in-­‐ Chief  2011-­‐13]     John  Jungpa  Park  -­‐  [Editor-­‐in-­‐Chief  of  Res  Medica  2012/13,  Medical  Student,  University  of   Edinburgh]     Dhairya  Lakhani  -­‐  [Vice  President,  Medical  Students'  Association  of  India,  3rd  year  MBBS   Sumandeep  Vidyapeeth  University]  


Table of  Contents   Can  We  Rely  on  Science?  [Editorial]............................................................................................p.  s4   LD  Hughes     The  State  of  Science  and  Unreliable  Research.........................................................................p.  s6-­‐11   LD  Hughes     Has  the  pharmaceutical  industry  commandeered  evidence-­‐based  medicine?  Problems…….  p.s12-­‐s18   D  Christmas     Has  the  pharmaceutical  industry  commandeered  evidence-­‐based  medicine?  Solutions…...  p.s19-­‐s25     D  Christmas     The  Human  Genome  Project:  Undervalued  Ingenuity..........................................................p.  s26-­‐s28   IJ  Hyndman    

                       


Can we  rely  on  science?            Special  Issue  Editorial

Lloyd  D  Hughes  (5th  year  MBChB,  University  of  Dundee;  Associate  Editor,  SUMJ)   Correspondence  to:  Lloyd  Hughes  :  L.D.Hughes@dundee.ac.uk               EDITORIAL  

Science commands  an  enormous  amount  of  respect  in  society  at  large.  Indeed,  scientists  and   their   endeavours   have   helped   enlighten   the   world   over   the   last   several   hundred   years.   However,   in   the   past   decade   or   so   there   have   been   numerous   concerns   about   the   quality   of   the  science  papers  that  are  driving  policy  in  numerous  different  settings.     In   2010   a   paper   by   two   world   leading   economists   in   an   eminent   journal   reported   that   economic  growth  of  a  nation  slows  dramatically  when  a  government's  debt  exceeds  90%  of   a   country's   annual   economic   output.   The   study   was   cited   by   politicians   and   policymakers   worldwide,   helping   to   justify   that   reducing   public   spending   would   help   drive   economic   growth.  However,  in  2013  a  Harvard  student  found  simple  errors  in  the  paper  when  trying  to   verify   the   findings,   and   the   paper   was   later   retracted   with   the   authors   fervently   denying   they  were  selective  in  their  selection  of  data,  and  admitted  a  coding  error1.  In  the  realm  of   healthcare   there   have   been   unfounded   concerns   voiced   about   the   Measles   Mumps   and   Rubella   vaccine   (MMR)2   and   difficulties   in   highlighting   genuine   concerns   with   an   experimental   drug3   due   to   the   influence   of   a   drug   companies.   Indeed,   in   a   recent   Economist   article   it   was   suggested   that,   “the   false   trails   laid   down   by   shoddy   research   are   an   unforgivable  barrier  to  understanding”4.     This   Scottish   Universities   Medical   Journal   supplemental   edition   ‘Can   we   rely   on   science?’   focuses  upon  some  of  these  issues.  Firstly,  in  the  article,  ‘The  State  of  Science  and  Unreliable   Research’5,   your   Associate   Editor   will   look   at   the   current   challenges   to   high   quality   clinical   and  scientific  research,  including  the  competitiveness  of  science  and  academia,  publication   bias,  the  weaknesses  of  peer-­‐review  and  ‘hidden’  data-­‐sets.     Secondly,   Dr   David   Christmas   (Consultant   Psychiatrist)   will   report   about   the   influence   that     drug   companies   and   the   pharmaceutical   industry   has   upon   medical   research6-­‐7.   In   his   two   article   series,   Dr   Christmas   will   consider   both   the   problems   that   the   current   level   of   involvement   the   pharmaceutical   industry   causes   and   will   suggest   solutions   to   improve   the   relationship   between   scientists,   clinicians   and   industry.   Finally,   Iain   Hyndman   explores   the   Human   Genome   Project   (HGP)   in   a   personal   view/commentary   article.   The   HGP   was   the   largest  collaborative  research  project  carried  out  in  human  history,  whose  long-­‐term  effects   have  incredible  potential  to  aid  the  lives  of  millions  of  patients  worldwide8.   This   supplement   hopes   to   inform   our   readers   about   some   of   the   important   non-­‐clinical   factors   that   drive   evidence   based   practice.   As   always   we   welcome   comments   and   suggestions   about   our   publication   and   welcome   submissions   for   future   main   issues,   online   platform  and  supplements.  

References

All references  available  on  the  online  edition  of  this  paper    

 


The State  of  Science  and  Unreliable  Research

Lloyd D  Hughes  (5th  year  MBChB,  University  of  Dundee;  Associate  Editor,  SUMJ)   Correspondence  to:  Lloyd  Hughes  :  L.D.Hughes@dundee.ac.uk              

ABSTRACT

Scientific endeavours  in  all  fields  have  helped  develop  and  shape  society  across  the  globe.  In   the  past  decade  the  sheer  quantity  of  research  and  data  has  been  staggering,  with  90%  of  all   data  being  obtained  since  2011.  In  the  face  of  this  sheer  quantity  of  research  and  data  there   are   legitimate   concerns   that   quality   may   be   being   compromised   for   a   number   of   different   reasons.  There  are  concerns  that  the  peer-­‐review  process  is  not  as  effective  as  it  should  be,     that   papers   reporting   negative   results   only   account   for   a   small   proportion   of   main   stream   journal   articles,   and   that   the   competitiveness   of   science   and   academia   is   discouraging   verification   studies   particularly   in   basic   sciences   and   promotes   exaggeration   and   cherry-­‐ picking  of  results.     Such   concerns   are   legitimised   by   the   tenfold   increase   in   retractions   from   mainstream   journals  in  the  past  decade.  In  addition  to  wasting  time  and  resources  such  flawed  research   may  place  patient  lives  in  jeopardy.  Indeed,  between  2000  and  2010,  80,000  patients  took   part   in   clinical   trials   based   upon   research   that   was   later   retracted   (either   due   to   error   or   improprieties).  This  paper  aims  to  discuss  some  of  the  areas  including  the  competitiveness   of  science  and  academia,  publication  bias,  the  weaknesses  of  peer-­‐review  and  ‘hidden’  data-­‐ sets.   Key  Words:    scientific  research;  peer-­‐review;  clinical  research  

Introduction

Scientific endeavours  in  all  fields  have  helped  develop  and  shape  society  across  the  globe.  In   the  past  decade  the  sheer  quantity  of  research  and  data  has  been  staggering,  with  90%  of  all   data  being  obtained  since  20111.  In  the  face  of  this  sheer  quantity  of  research  and  data  there   are   legitimate   concerns   that   quality   may   be   being   compromised   for   a   number   of   different   reasons2.  There  are  concerns  that  the  peer-­‐review  process  is  not  as  effective  as  it  should  be3-­‐ 4 ,  that  papers  reporting  negative  results  only  account  for  a  small  proportion  of  main  stream   journal   articles5,   and   that   the   competitiveness   of   science   and   academia   is   discouraging   verification   studies   particularly   in   basic   sciences   and   promotes   exaggeration   and   cherry-­‐ picking  of  results5-­‐6.     Such   concerns   are   legitimised   by   the   number   of   article   retractions   in   mainstream   journals.   Retractions   have   steadily   increased   since   the   first   retraction   in   19777,   with   a   ten-­‐fold   increase  in  the  past  decade5.  In  addition  to  wasting  time  and  resources  such  flawed  research   may  places  patient  lives  in  jeopardy.  Indeed,  between  2000  and  2010,  80,000  patients  took   part   in   clinical   trials   based   upon   research   that   was   later   retracted   (either   due   to   error   or   improprieties)3.  This  paper  aims  to  discuss  some  of  the  areas  including  the  competitiveness   of  science  and  academia,  publication  bias,  the  weaknesses  of  peer-­‐review  and  ‘hidden’  data-­‐ sets.  

“Trust but  Verify”  

The concept  that  identical  experiments  always  get  the  same  result,  no  matter  who  performs   them,   is   a   cornerstone   to   scientific   research   and   aims   to   prove   a   hypothesis   as   fact.   Achieving   perfect   verification   of   all   studies   is   clearly   not   achievable   due   to   limitations   of   basic  science  research  describing  new  findings,  but  it  should  be  possible  for  the  majority  of  


science experiments   (in   all   scientific   fields)   to   be   replicated   and   verified   based   upon   the   methods  described  by  authors.  However,  this  appears  not  to  be  the  case.     In  2012,  Amgen  (an  American  drug  company)  attempted  to  replicate  53  pre-­‐clinical  oncology   studies   they   considered   land-­‐mark   papers   in   the   basic   science   of   cancer8.   Even   when   liaising   closely  with  authors  of  the  relevant  papers,  only  11%  of  the  papers   findings  were  replicated,   a   result   described   by   the   authors   as   “shocking”8.   In   2011   Bayer   (a   German   drug   company)   only   managed   to   replicate   a   quarter   of   67   seminal   studies   in   an   analysis   of   early   in-­‐house   projects   in   the   research   fields   of   oncology,   women's   health   and   cardiovascular   diseases9.   Such  work  underlines  the  importance  of  confirmatory  validation  studies,  which  should  also   aim  to  complement  the  knowledge  on  a  particular  therapeutic  target  and  hone  researchers   techniques.   Indeed,   the   difficulties   in   the   reproducibility   of   early   stage   scientific   research   may  be  part  of  the  reason  for  the  fall  in  the  success  in  newly  developed  Phase  II  clinical  trials   in  recent  years9-­‐10.     It  should  be  noted  that  a  difficulty  in  reproducibility  does  not  amount  to  fraud,  but  merely   shows   that   particular   novel   techniques   assessing   unexplored   scientific   avenues   in   basic   science   may   be   prone   to   a   greater   degree   of   variability   and   error   than   expected.   Another   potentially   contributing   factor   may   be   that   papers   do   not   adequately   report   all   research   resources   (including   antibodies,   model   organisms   etc.),   without   which   it   is   challenging   to   reproduce  experiments  even  when  the  underlying  science  is  sound.  Indeed,  a  recent  study   reported  that  more  than  half  of  238  biomedical  papers  in  84  journals  failed  to  identify  all  the   resources  required  to  verify  their  work11.  Another  review  of  351  randomly  selected  papers  in   high-­‐impact   journals   found   that   only   143   papers   were   subject   to   any   data   availability   policies,  and  adhered  to  the  data  availability  instructions  in  their  respective  journal12.     Thus,  it  is  important  to  promote  verification.  Adequately  verifying  scientific  techniques  will   be   of   benefit   for   drug   companies   as   their   investments   in   early   stage   therapeutic   products   would  be  more  likely  to  succeed  at  Phase  II  trials9  and  for  scientists  as  scientific  techniques   will  develop  over-­‐time  and  provide  greater  weight  to  research  findings.  However,  promoting   verification   is   not   easy.   Many   academic   researchers   would   prefer   to   embark   on   novel   research  that  is  more  likely  to  advance  their  own  career  and  junior  researchers  may  feel  that   replication  is  a  challenge  against  a  papers  authors.  More  importantly,  in  the  current  climate   the   costs   associated   with   verification   studies   alongside   a   desire   for   novel   research   by   funding  bodies  may  prevent  some  work  being  replicated.       Despite   these   barriers,   in   the   discipline   of   psychology   the   journal   Perspectives   on   Psychological   Sciences   will   shortly   have   a   section   devoted   to   replication   and   verification   studies.  There  is  also  the  newly  formed  Reproducibility  Initiative  through  which  life  scientists   can   pay   to   have   their   work   verified   by   an   independent   laboratory,   giving   the   papers   methods   and   findings   more   weight   thus   incentivising   replication.   The   Initiative   is   currently   reviewing  the  findings  of  the  50  highest-­‐impact  cancer  papers  published  between  2010  and   2012,  following  a  grant  from  the  Laura  and  John  Arnold  Foundation13.  Dr  Elizabeth  Iorns,  the   co-­‐director   of   the   Reproducibility   Initiative,   noted   “The   lack   of   reproducibility   in   cancer   studies   is   a   major   obstacle   in   the   development   of   viable   therapies   to   cure   cancer”13.   It   is   hoped  that  this  initiative  will  help  improve  the  reproducibility  of  work  in  other  science  and   healthcare  fields.    

Science  –  A  Competitive  Industry  

Science is   a   competitive   industry.   Full   professors   in   America   earned   on   average   $135,000,   and   with   6   applications   from   new   PhDs   for   every   academic   post   in   the   US,   the   stakes   are  


high3. The   pressure   to   publish   work   in   high-­‐impact   journals   is   crucial,   as   high-­‐impact   publications   are   important   for   being   successful   in   the   academic   job   market   and   subsequently   obtaining   and   maintaining   funding   for   research.   The   pressure   to   ‘publish   or   perish’  appears  to  have  an  effect  upon  some  researchers  overall  scientific  conduct.  A  2009   meta-­‐analysis  covering  18  fraud  surveys  [six  targeting  biomedical  scientists  directly],  found   that   2%   of   those   participating   admitted   to   having   falsified,   fabricated   or   modified   data   at   least  once  themselves14.  Although  this  figure  is  reasonably  low,  an  alarming  14%  reported  to   have  noticed  this  behaviour  in  colleagues14.  The  author  also  proposed  that  as  these  surveys   ask   sensitive   questions   and   have   other   limitations,   these   figures   are   likely   to   be   a   conservative   estimate   of   the   true   prevalence   of   scientific   misconduct14.   Cherry-­‐picking,   selectively   reporting   data   that   supports   a   desired   outcome,   may   be   much   more   common   that  deliberate  fabrication.  A  fine  balance  must  always  be  found  about  reporting  data  that  is   relevant   and   important   to   the   narrative   of   a   paper,   and   data   that   is   not.   Massaging   the   data   so  they  favour  a  particular  finding/hypothesis  walks  a  tightrope,  between  sloppy  science  and   scientific  misconduct  depending  on  the  circumstances.       Such   cases   have   the   potential   to   have   a   hugely   detrimental   impact   for   all   involved   in   scientific   research.   Firstly,   upon   honest   researchers   by   preventing   those   researchers   obtaining   funding   for   their   work   and   tainting   institutions   with   a   fraudulent   tag.   Secondly,   scandals  in  sensitive  scientific  areas  including  stem-­‐cell  work  and  oncology  may  erode  public   confidence   in   science   and   their   research   more   widely6.   Finally,   clinical   work   involving   patients  based  upon  flawed  research  may  be  placing  lives  at  risk.     There  are  several  approaches  that  are  currently  attempting  to  address  some  of  the  concerns   associated   with   research   integrity.   Firstly,   it   is   becoming   increasingly   stressed   that   the   responsibility  of  maintaining  academic  integrity  should  fall  with  the  senior  investigator.  In  a   recent   Nature   editorial   it   was   noted   that   this   individual   has   a   unique   position   of   trying   to   develop,   “a   laboratory   environment   that   provides   a   strong   foundation   in   best   practice   in   research   and   research   ethics,   and   to   be   a   compelling   role   model   for   trainees”6.   Secondly,   there   has   been   recent   work   at   both   a   national   level   and   international   level   to   support   research  integrity.  In  the  UK,  the  UK  Research  Integrity  Office  provides  support  and  advice   for  individuals  and  institutions  throughout  the  research  process  and  offers  advice  for  those   who   feel   that   they   may   have   witnessed   research   malpractice.   Furthermore,   international   guidelines   have   been   developed   from   the   2nd   Worldwide   Research   Integrity   Conference,   aiming   to   raise   awareness   of   science   ethics   and   to   lobby   governments   and   institutions   to   design   and   implement   policies   to   promote   research   integrity.   The   guidelines   note   the   importance   of   reporting   and   raising   irresponsible   research   practice.   Finally,   the   guidelines   stress  to  researchers  that  despite  the  huge  pressure  they  can  be  placed  under  to  cherry-­‐pick   or  fabricate  results,  irreducible  findings  will  be  discredited  in  the  long-­‐run  and  the  potential   implications  of  fabrication  can  be  career  ending.  Finally,  journals  can  promote  integrity  and   transparency  in  the  practise  and  presentation  of  research  by  developing  and  implementing   rigorous   standards   for   data   presentation,   author   contributions   and   statements   of   authors'   conflict  of  interests.    Although   there  is  more  work  to  do  in  this  area,  the  fact  that  there  is   increased  awareness  and  interest  in  this  area  hopefully  will  improve  the  overall  climate  for   researchers  working  in  the  ‘publish  or  perish’  climate.  

Publication  Bias  

Yet another  challenge  facing  science  is  publication  bias.  Journals  yearn  for  papers  that  prove   a  hypothesis  or  describe  a  novel  intervention  or  treatment.  Indeed,  negative  results,  where   a  hypothesis/treatment  is  shown  to  be  incorrect,  account  for  only  14%  of  published  papers,   down  from  30%  in  the  1990s3.  However,  publishing  negative  results,  where  a  hypothesis  is  


shown to   be   incorrect,   remains   crucial.   Failing   to   publish   these   results   creates   a   false   impression   and   bias   in   the   literature.   Other   scientists   may   waste   valuable   time   and   resources   on   treatments   that   have   already   been   shown   to   be   ineffective.   The   importance   of   publishing   negative   studies   is   even   more   important   in   the   area   of   clinical   research,   as   patients   may   be   treated   with   medications/therapies   that   have   already   been   shown   in   unpublished   data   to   be   of   no   benefit   and/or   harm.   Furthermore,   not   publishing   negative   data  undermines  all  of  the  efforts  made  to  minimise  bias  in  individual  trials.15     There  have  been  numerous  efforts  to  try  and  address  this.  Firstly,  in  the  US  there  are  moves   to  enforce  legislation  requiring  all  results  of  clinical  trials  funded  by  the  state  to  be  posted   on   clinicaltrials.gov   within   12   months   of   completion16.   To   date,   this   legislation   appears   to   have   been   widely   ignored16   and   up   to   a   third   of   clinical   trials   financed   by   the   National   Institute   for   Health   remain   unpublished   after   51   months5.   Secondly,   closer   to   home,   the   AllTrials.net   campaign,   started   12   months   ago,   has   called   for   all   trials   on   all   uses   of   all   currently   prescribed   treatments   to   be   registered,   with   their   full   methods   and   results   reported17.   The   campaign   is   widely   supported   by   medical   and   academic   organisations,   and   has   obtained   the   support   of   some   in   the   pharmaceutical   industry   including   GlaxoSmithKleine.   This   unfortunately   will   only   include   research   carried   out   from   2014   and   beyond,   meaning   that   previous   trial   data   will   continue   to   skew   the   evidence   base.   However,   efforts  are  continuing  to  include  historical  data.  Thirdly,  the  British  Medical  Association  has   passed   a   motion   noting   that   withholding   trial   results   constitutes   research   misconduct18   with   the   General   Medical   Council   reviewing   their   position15.   Finally,   for   individual   researchers   many  journals  now  have  online  only  publication  meaning  they  have  more  space  to  publish   papers  with  negative  results.    

The  Challenges  of  Peer-­‐Review  

Peer-­‐review is   often   reported   to   be   the   ‘gold-­‐standard’   of   vetting   the   quality   of   scientific   work4.   However,   there   are   numerous   examples   of   research   that   passed   this   standard   subsequently  being  retracted  due  to  clear  fraud  or  unethical  conduct.  The  most  commonly   quoted   examples   include   Hwang   Woo-­‐suk   who   carried   out   fraudulent   work   on   stem-­‐cells   published   in   Science19   and   Andrew   Wakefield   whose   work   in   the   1998   Lancet   led   to   the   MMR  vaccine  scandal20.  Importantly,  these  cases  extend  the  highest  impact  journals,  which   have  rejection  rates  in  excess  of  90%.     Work   by   John   Bohannon,   a   Harvard   biologist,   found   that   a   pseudonymous   paper   on   the   effects  of  a  chemical  derived  from  lichen  to  treat  cancer  cells  was  accepted  by  157  of  304   journals   the   work   was   submitted   to21.   This   was   despite   the   fact   that   the   paper   had   numerous  deliberate  errors  in  study  design,  analysis  and  the  interpretation  of  results20.  This   study   only   focused   upon   open   access   journals,   which   generally   charge   a   fee   to   publish.   Although   this   study   only   focused   upon   lower   impact   journals,   similar   problems   have   been   found  in  well-­‐established  and  respected  journals  such  as  the  BMJ21.  Indeed,  an  internal  BMJ   study   took   a   paper   about   to   be   published   in   the   BMJ,   and   inserted   eight   deliberate   errors   before  sending  it  to  420  potential  BMJ  reviewers:  221  (53%)  responded22.  The  results  were   stark,  the  median  number  of  errors  spotted  was  two,  nobody  spotted  more  than  five,  and   16%   didn’t   spot   any22.   Another   concern   voiced   refers   to   referees   who   use   anonymity   to   prevent  publications  in  their  field  of  research23.       There  have  been  several  suggestions  as  to  how  the  process  of  peer-­‐review  can  be  improved.   Firstly,  various  short-­‐term  peer-­‐review  training  programmes  have  been  shown  to  be  of  some   limited   benefit24-­‐25,   although   the   value   of   longer   interventions   needs   to   be   assessed24.   Secondly,   the   BMJ   has   implemented   a   process   of   open   review   where   anonymity   is  


removed26 ..  In  addition  to  removing  concerns  about  reviewers  being  deliberately  obtrusive4,   the   BMJ   editorial   board   felt   this   was   ultimately   ethical   –   placing   authors   and   reviewers   in   equal   positions   and   increasing   accountability22.   A   randomized   controlled   trial   found   that   open  review  made  no  significant  difference  to  the  quality  of  reviews  and  it  was  acceptable   by   both   authors   and   reviewers.   Although   there   was   a   higher   tendency   to   recommend   acceptance,  the  fact  that  the  editors  retain  total  control  as  whether  to  publish  means  that   this   increase   is   less   important.   Clearly   open   review   may   not   be   suitable   for   all   journals,   particularly   in   specialist   areas   where   clinicians   and   scientists   may   know   each   other   very   well4.   An   intermediate   solution   to   accountability   has   been   tried   in   other   journals,   where   reviews   although   anonymous   are   published   online4,22.   This   approach   has   been   shown   to   affect   reviewers   as   any   comments   they   make   on   a   specific   paper   is   available   to   view,   meaning   that   openly   hostile   or   inappropriate   reviews   become   less   common   and   authors   receiving  such  reviews  are  more  supported  by  other  researchers4.     Electronic   pre-­‐publication   prior   to   final   publication   is   another   approach   used   by   some   journals.   Articles   are   uploaded   online   a   few   weeks   before   becoming   incorporated   into   the   journals  database  and  this  provides  a  window  where  general  readers  can  make  comments.   This   approach   allows   journals   to   provide   researchers   an   additional   setting   to   receive   further   feedback   to   develop   and   improve   their   article.   It   should   be   noted   that   the   Editorial   Board   have   the   final   say   whether   comments   require   to   be   addressed   not   by   the   authors.   This   approach  has  worked  well  for  the  Cochrane  Collaboration  as  protocols  have  been  changed   and  meta-­‐analysis  adjusted  following  comments  for  its  clinical  readership.     It  is  unclear  if  such  approaches  would  have  prevented  some  of  the  fraudulent  papers  of  the   past   being   published,   but   will   certainly   help   to   improve   peer-­‐reviews   standing   in   the   scientific  communioty22.  

Conclusion  

The status  of  science  and  research  is  based  upon  the  idea  that  findings  from  research  will  be   right   most   of   the   time   and   will   correct   its   errors   as   and   when   they   arise.   However,   with   recent  high  profile  scandals  in  the  world  rocking  public  confidence  in  science  it  is  important   to   review   the   status   quo.   It   is   hoped   that   attempts   to   improve   the   quality   of   peer-­‐review,   address  publication  bias,  promote  verification  and  prevent  fraud  will  over  the  coming  years   strengthen  trust  once  again  in  scientific  endeavours.    

References   1. 2. 3.

4. 5.

6. 7.

SINTEF (2013,  May  22).  Big  Data  -­‐  for  better  or  worse.  90%  of  world's  data  generated  over   the   last   two   years.   Available   from:   http://www.sintef.no/home/Press-­‐Room/Research-­‐ News/Big-­‐Data-­‐-­‐for-­‐better-­‐or-­‐worse/    Last  Accessed:  23/02/14   Arndt  KA  (1992).  Information  excess  in  medicine.  Overview,  relevance  to  dermatology,  and   strategies  for  coping.  Arch  Dermatol.  128(9):1249-­‐56.   The  Economist  (2013).  Problems  with  Scientific  Research  –  How  science  goes  wrong.   Available  from:  http://www.economist.com/news/leaders/21588069-­‐scientific-­‐research-­‐has-­‐ changed-­‐world-­‐now-­‐it-­‐needs-­‐change-­‐itself-­‐how-­‐science-­‐goes-­‐wrong  Last  Accessed:  21/02/14     Henderson  M  (2010).  Problems  with  peer  review.  BMJ  2010;340:c1409   The  Economist  (2013).  Unreliable  research  –  Trouble  at  the  Lab.  Available  from:   http://www.economist.com/news/briefing/21588057-­‐scientists-­‐think-­‐science-­‐self-­‐ correcting-­‐alarming-­‐degree-­‐it-­‐not-­‐trouble  Last  Accessed:  21/02/14     Combating  scientific  misconduct.  Nat  Cell  Biol.  2011  Jan;13(1):1.   Cokol   M,   Ozbay   F,   Rodriguez-­‐Esteban   R   (2008).   Retraction   rates   are   on   the   rise.   EMBO   Reports  9(1):2  


8. 9. 10. 11.

12. 13.

14. 15. 16. 17. 18.

19. 20.

21. 22. 23. 24.

25. 26.

           

Begley GC   &   Ellis   LM   (2012).   Drug   development:   Raise   standards   for   preclinical   cancer   research.  Nature  483.  531–533.   Prinz  F,  Schlange  T  &  Khusru  A  (2011).  Believe  it  or  not:  how  much  can  we  rely  on  published   data  on  potential  drug  targets?  Nature  Reviews  Drug  Discovery  10,  712.   Arrowsmith   J   (2011).   Trial   watch:   Phase   II   failures:   2008–2010.   Nature   Reviews   Drug   Discovery  10,  328-­‐329   Vasilevsky   NA,   Brush   MH,   Paddock   H,   Ponting   L,   Tripathy   SJ   et   al.   (2013)   On   the   reproducibility   of   science:   unique   identification   of   research   resources   in   the   biomedical   literature.  PeerJ  1:e148   Alsheikh-­‐Ali   AA,   Qureshi   W,   Al-­‐Mallah   MH,   Ioannidis   JPA   (2011)   Public   Availability   of   Published  Research  Data  in  High-­‐Impact  Journals.  PLoS  ONE  6(9):  e24357   Center  for  Open  Science  Website.  Reproducibility  Initiative  Receives  $1.3M  Grant  to  Validate   50   Landmark   Cancer   Studies.   Available   from:   http://centerforopenscience.org/pr/2013-­‐10-­‐ 16/  Last  Accessed:  21/02/14   Fanelli   D   (2009)   How   Many   Scientists   Fabricate   and   Falsify   Research?   A   Systematic   Review   and  Meta-­‐Analysis  of  Survey  Data.  PLoS  ONE  4(5):  e5738.   Goldacre  B  (2014).  Improving,  and  auditing,  access  to  clinical  trial  results.  BMJ  348:g213   Prayle  AP,  Hurley  MN,  Smyth  AR  (2012).  Compliance  with  mandatory  reporting  of  clinical  trial   results  on  ClinicalTrials.gov:  cross  sectional  study.  BMJ.  344:d7373.   AllTrials  (2013).  What  does  all  trials  registered  and  reported  mean?  All  Trials.  Available  from:   www.alltrials.net/all-­‐trials/  Last  Accessed:  06/03/14   BMA.   Non-­‐publication   of   trials   results   is   research   misconduct.   Available   from:     http://ecancer.org/journal/news/4200-­‐bma-­‐-­‐-­‐-­‐-­‐non-­‐publication-­‐of-­‐trials-­‐results-­‐is-­‐research-­‐ misconduct.php  Last  Accessed:  06/03/14   Hwang  WS,  Roh  SI,  Lee  BC,  Kang  SK,  Kwon  DK,  Kim  S,  et  al.  Patient-­‐specific  embryonic  stem   cells   derived   from   human   SCNT   blastocysts   [retracted   in   Science   2006;311:335]   Science   2005;308:1777-­‐83.   Wakefield   AJ,   Murch   SH,   Anthony   A,   Linnell   J,   Casson   DM,   Malik   M,   et   al.   Ileal-­‐lymphoid-­‐ nodular   hyperplasia,   non-­‐specific   colitis,   and   pervasive   developmental   disorder   in   children   [retracted  in  Lancet  2010  Feb  2.]  Lancet  1998;351:637-­‐41.   The  Scientist  (2013).  Fake  paper  exposes  failed  peer  review.  Available  from:  http://www.the-­‐ scientist.com/?articles.view/articleNo/37798/title/Fake-­‐Paper-­‐Exposes-­‐Failed-­‐Peer-­‐Review/   Last  Accessed:  21/02/14   Smith  R  (1997).  Peer  review:  reform  or  revolution?  BMJ  315:759   Ghosh   P   (2010).   Journal   stem   cell   work   “blocked.”   BBC   News.   Available   from:   http://news.bbc.co.uk/1/hi/8490291.stm.  Last  Accessed:  21/02/14   Schroter   S,   Black   N,   Evans   S,   Godlee   F,   Osorio   L,   Smith   R   (2008).   What   errors   do   peer   reviewers   detect,   and   does   training   improve   their   ability   to   detect   them?   J   R   Soc   Med.   101(10):507-­‐14.   Schroter   S,   Black   N,   Evans   S,   Carpenter   J,   Godlee   F,   Smith   R   (2004).   Effects   of   training   on   quality  of  peer  review:  randomised  controlled  trial.  BMJ  328:673   Van   Rooyen   S,   Godlee   F,   Evans   S,   Black   N,   Smith   R   (1999).   Effect   of   open   peer   review   on   quality  of  reviews  and  on  reviewers’  recommendations:  a  randomised  trial.  BMJ  318:23-­‐7.  


Has the  pharmaceutical  industry  commandeered  evidence-­‐ based  medicine?  1)  Problems                   Dr  David  Christmas  (Consultant  Psychiatrist,  Advanced  Interventions  Service,  NHS  Tayside)   Correspondence  to:  David  Christmas:  david.christmas@nhs.net                       Please  note  the  super-­‐script  numbers  refer  to  links  at  the  bottom  of  each  page  (in  the  form   of  weblinks  and  definitions),  and  numbers  in  brackets  refer  to  main  reference  list  

ABSTRACT The  pharmaceutical  industry  has  come  under  increasing  scrutiny  in  recent  years  because  of   its   practices.   For   example,   financial   penalties   imposed   by   the   US   Government   and   other   agencies   on   pharmaceutical   companies   between   1991   and   July   2012   exceed   $30   billion.   More  worryingly,  in  recent  years  there  has  been  increasing  attention  on  some  of  the  failures   of   multiple   parts   of   the   wider   system   including   pharmaceutical   development,   safety   monitoring  and  regulation.     This   paper   will   examine   some   of   the   techniques   used   by   the   pharmaceutical   industry   to   market   their   drugs,   and   examine   the   reliability   of   some   of   the   claims   and   counter-­‐claims   being  made  regarding  the  impact  that  such  practises  have  upon  evidence  based  medicine.   Key  Words:    pharmaceutical  company;  research  malpractice;  Big  Pharma  

Introduction The  pharmaceutical  industry  has  come  under  increasing  scrutiny  in  recent  years  because  of   its   practices.   For   example,   financial   penalties   imposed   by   the   US   Government   and   other   agencies  on  pharmaceutical  companies  between  1991  and  July  2012  exceed  $30  billion  (1).   It  is  often  argued  that  because  cases  take  many  years  to  court,  these  reports  are  describing   out-­‐dated   and   corrected   practices.   However,   settlements   between   November   2010   and   July   2012   account   for   approximately   one-­‐third   of   this   ($10.2   billion)   and   the   recent   bribery   scandals  in  China  admitted  by  GlaxoSmithKline1  suggest  that  problematic  practices  have  not   been  completely  eradicated.  

Are the  pipelines  drying  up?   Of  concern  is  the  announcement  in  the  last  few  years  that  many  companies  are  withdrawing   funding   into   research   and   development   into   brain   disorders,   citing   the   high   risk   of   experimental   drugs   failing   to   reach   market   (2).   However,   many   commentators   have   called   into  question  the  current  process  of  drug  development;  highlighting  the  fact  that  most  drug   classes   have   come   about   because   of   serendipitous   discovery   rather   than   planned   development  (3).   The   current   approaches   to   drug-­‐discovery   for   CNS   disorders   have   arguably   struggled  to  deliver  on  their  promises  over  the  last  two  decades.  

Failures of  regulation   More  worryingly,  in  recent  years  there  has  been  increasing  attention  on  some  of  the  failures   of   multiple   parts   of   the   system:   pharmaceutical   development;   safety   monitoring;   and   regulation.   For   example,   Rofecoxib   (Vioxx)   was   withdrawn   from   the   market   by   the   FDA   after   concerns  about  cardiac  safety,  but  the  manufacturer  (Merck)  had  played  down  such   risks  for  

                                                                                                              1

http://www.bbc.co.uk/news/business-­‐23402154    


a long  time  beforehand.  In  fact,  it  is  likely  that  the  use  of  Vioxx  resulted  in  over  2,000  excess   deaths  in  the  USA  alone  (4).   More  recently,  after  a  reanalysis  of  efficacy  data  by  the  Institute  for  Quality  and  Efficiency  in   Health   Care,   IQWIG   (the   German   equivalent   of   NICE),   the   antidepressant   Reboxetine   was   reported  to  be  not  only  ineffective  but  harmful  (5).  There  are  still  withheld  data  regarding   Tamiflu   (Oseltamivir)   which   leave   considerable   doubt   over   its   effectiveness   and   safety,   despite   most   countries   continuing   to   stockpile   the   drug   in   an   attempt   to   militate   against   pandemic   flu. 2  Recently,   substantial   concerns   over   publication   bias   have   led   to   a   reconsideration  of  the  evidence  for  efficacy  of  the  melatonergic  antidepressant  Agomelatine   (6).     The   need   for   transparency   in   trial   reporting   has,   hopefully,   reached   a   crescendo   with   the   British  Medical  Journal  announcing  that  they  would  not  be  publishing  trials  unless  the  raw   data  were  made  available  to  researchers.3   This   paper   will   examine   some   of   the   techniques   used   by   the   pharmaceutical   industry   to   market   their   drugs,   and   examine   the   reliability   of   some   of   the   claims   and   counter-­‐claims   being  made  regarding  the  impact  that  such  practises  have  upon  evidence  based  medicine.  

Collaboration or  covert  marketing?   In   2012,   the   Ethical   Standards   in   Health   &   Life   Sciences   Group   published   ‘Guidance   on   collaboration   between   healthcare   professionals   and   the   pharmaceutical   industry’,   a   document   endorsed   by   most   of   the   UK   Royal   Colleges  (7)   at  the  time.  Some  of  the  problems   with  this  document  have  been  highlighted  (8)  with  The  Lancet  withdrawing  their  support  for   the  guidance  shortly  after  publication;  closely  followed  by  the  BMA4.   The   guidance,   however,   provides   a   useful   framework   for   critically   exploring   some   of   the   claims   made,   and   also   affords   opportunities   to   reflect   on   the   compelling   evidence   that   contradicts  most  of  the  claims.   1. “Industry   is   responsible   for   the   vast   majority   of   medicines   research   and   development   (R&D)  in  the  UK…”   This   may   be   true   for   the   UK,   but   in   the   USA   more   than   half   of   new   compounds   were   discovered   by   Universities   and/   or   biotechnology   companies,   and   later   transferred   to   pharmaceutical   companies   (9).   Further,   most   drugs   are   marketed   internationally   so   the   ‘origin’  of  a  particular  drug  is  less  relevant.  The  importance  of  UK-­‐based  R&D  is  an  important   symbol   for   the   pharmaceutical   industry,   so   it’s   a   shame   that   Novartis   and   Roche   have   threatened  to  pull  out  of  the  UK5  and  Pfizer  has  closed  its  research  plant  in  the  UK6.   2. “It   takes   10   to   15   years   to   develop   a   new   medicine   and   typically   costs   £550   million   to   do   all  the  work  necessary  before  a  medicine  can  be  licensed  for  use.”   This   often-­‐repeated   figure   comes   from   research   (sponsored   by   the   pharmaceutical   industry)   conducted  in  2003  (10),  but  these  costs  have  been  challenged,  along  with  the  assumptions   used,   resulting   in   protracted   debates   about   how   to   determine   the   costs   of   drug   discovery   (11-­‐14).  More  conservative  estimates  have  suggested  much  lower  figures  (15).  

                                                                                                              2

http://dx.doi.org/10.1136/bmj.e7305      http://dx.doi.org/10.1136/bmj.e7304     4  http://bma.org.uk/news-­‐views-­‐analysis/news/2013/march/bma-­‐withdraws-­‐from-­‐pharma-­‐standards-­‐group     5  http://www.theguardian.com/business/2010/apr/11/novartis-­‐roche-­‐threaten-­‐quit-­‐uk     6  http://www.bbc.co.uk/news/business-­‐12335801     3


Whatever the  real  costs,  a  significant  proportion  of  industry  spend  in  recent  years  has  been   on   incremental   improvements   to   existing   drugs,   rather   than   new   ‘breakthrough’   compounds.   Only   about   10%   of   new   drugs   offer   significant   benefits   over   existing   drugs  (16).   For  these  ‘me-­‐too’  drugs  (a  drug  that  is  structurally  very  similar  to  already  known  drugs)  and   ‘evergreening’ 7 ,   the   development   process   is   much   shorter,   and   consequently,   much   cheaper.  Evergreening  also  has  the  important  consequence  of  delaying  the  entry  of  generic   drugs   into   the   market   (17).   Other   examples   of   patent-­‐extension   include   the   licensing   of   unusual  doses  –  for  example,  a  23mg  dose  of  Donepezil  which  offered  no  additional  benefits   but  could  not  be  made  up  of  generic  20mg  and  5  mg  tablets  (18).   Finally,   the   high   costs   of   drug   discovery   are   often   used   to   justify   the   high   prices   of   new   drugs.  Industry  claims  that  more  money  is  spent  on  drug  development  than  promotion  are   easily   deconstructed,   suggesting   that   promotion   is   the   largest   component   of   the   costs   of   drugs  (19).   3. “…the  results  of  controlled  clinical  trials  are  made  available  in  the  public  domain  through   clinical   trial   registries   and   portals,   peer   reviewed   publications,   medical   meetings   and   company  websites.”   This   claim   is   spurious.   As   discussed   above   in   the   case   of   Reboxetine,   unfavourable   clinical   trial  data  are  often  withheld  by  pharmaceutical  companies  because  they  are  not  obliged  to   produce   it   when   submitting   applications   to   regulatory   bodies.   In   addition,   despite   the   existence   of   trial   registries   and   mandatory   reporting,   approximately   50%   of   trials   used   in   support   of   marketing   approval   are   never   published   (20).   Further,   there   is   compelling   evidence   of   publication   bias   for   antidepressants   and   antipsychotic   drugs   (21,   22).   There   is   convergent   evidence   that   studies   sponsored   by   industry   are   more   likely   to   report   positive   results  (23,  24).  Despite  apparent  mandatory  reporting  of  data  from  trials  in  registries  within   one   year   of   completion,   most   triaIs   fail   to   comply   (25,   26)   and   fewer   than   half   of   such   studies  are  ever  published  in  peer-­‐reviewed  journals  (27).  In  summary,  anywhere  between   25%  and  50%  of  trial  data  involving  patients  never  reaches  the  light  of  day.   4. “Industry  plays  a  valid  and  important  role  in  the  provision  of  medical  education.”   This   suggestion   is   not   supported   by   the   evidence.   Unfortunately,   industry-­‐sponsored   medical  education  creates  clear  conflicts  of  interest  (28)  and  has  been  shown  to  be  biased   (29).   In   2005,   the   House   of   Commons   Health   Committee   concluded   that:   “The   pharmaceutical   industry’s   promotional   efforts   are   relentless   and   pervasive.   The   evidence   presented   showed   the   lengths   to   which   the   industry   goes   to   ensure   that   promotional   messages   reach   their   targets,   and   that   these   targets   include   not   only   prescribing   groups,   but   patients  and  the  general  public.”  (17).   Many   doctors   fall   foul   of   the   ‘third-­‐person   effect’,   i.e.   they   believe   that   others   are   more   susceptible  to  influence  than  themselves  (30).  It  is  commonly  assumed  by  people  that  they   are  immune  from  the  influence  of  promotional  strategies,  but  it  is  not  the  case  –  we  are  all   vulnerable  to  such  tactics  (31).   5. “Whilst   medical   representatives   are   employed   to   promote   medicines,   they   can   be   a   useful  source  of  information  for  healthcare  professionals  and  are  another  vital  feedback   mechanism  to  the  companies  they  represent.”  

                                                                                                              7

“Evergreening  involves  extending  the  patented  life  of  a  branded  product,  typically  by  reformulating  the  drug,  for   instance  by  using  a  different  drug  delivery  system,  changing  a  dosage  form,  or  presentation  (e.g.  from  tablet  to   capsule).”(17)      


There is   no   evidence   to   support   the   claim   that   information   from   industry   is   consistently   reliable,   accurate,   or   free   from   bias.   Information   from   pharmaceutical   companies   is   associated   with   higher   prescribing   frequency,   higher   costs,   and   lower   prescribing   quality   (32).   Large   conferences   organised   by   independent   groups   provide   ample   opportunity   for   industry   to   gain   contact   with   doctors   and   influence   the   scientific   programme   through   satellite  meetings  (33).   The   main   reason   for   having   representatives   is   to   influence   the   prescribing   patterns   of   doctors.   The   influence   of   medical   reps   on   prescribing   has   been   demonstrated   repeatedly   (34).   In   addition,   the   ‘gifts’   commonly-­‐received   (such   as   pens,   notepads,   etc.)   and   free   samples  all  impact  upon  behaviour  (35-­‐38).   With   regards   to   feedback,   there   are   already   well-­‐established   mechanisms   by   which   health   professionals  can  supply  post-­‐marketing  data  on  the  drugs  they  use;  for  example,  the  Yellow   Card   Scheme 8 ,   managed   by   the   Medicines   and   Healthcare   products   Regulatory   Agency   (MHRA).   Giving   feedback   to   representatives   from   pharmaceutical   companies   is   unlikely   to   result   in   safer   or   better   drugs   since   the   information   is   not   shared   between   companies   or   reported  centrally.   6.  “Industry   is   able   to   provide   factual   information   to   patients   about   the   medicines   they   have  been  prescribed  including  patient  information  leaflets  and  websites.”   Direct  to  Consumer  Marketing  (DTCA)  is  not  permitted  in  the  UK  and  many  other  countries.   It   has   been   shown   to   have   no   clear   benefits   and   results   in   increased   prescribing   of   advertised  drugs  (39).  As  discussed  above,  the  information  from  drug  companies  is  often  of   low  quality.  Such  schemes  are  typically  used  to  generate  demand  for  a  product  (40)  and/or   galvanise   support   among   patients   groups.   Such   a   process   is   often   called   ‘astroturfing’   because  of  its  ability  to  generate  an  artificial  ‘grass-­‐roots’  community.  The  funding  of  patient   groups  by  the  pharmaceutical  industry  is  concerning  to  many  (41).   7. “Strict  rules  govern  adverse  event  reporting.  Companies  have  pharmacovigilance  teams   dedicated  to  post-­‐marketing  surveillance…”   This   would   be   reassuring   if   it   wasn’t   for   a   number   of   high-­‐profile   examples   where   either   information   about   safety   was   withheld   by   the   company,   e.g.   Vioxx   (42),   or   Rosiglitazone   (43);   or   marketing   was   initiated   to   downplay   potential   risks   of   the   drug,   in   the   case   of   Olanzapine   (44).   This   claim   is   perhaps   most   concerning   given   that   it   is   offering   false   reassurance   of   safety   when   the   track   record   of   the   pharmaceutical   industry   suggests   that   active   attempts   at   hiding   or   downplaying   risks   are   often   employed.   Further   examples   of   drugs  where  risks  outweigh  benefits  include  Reboxetine,  which  turned  out  to  be  ineffective   and  harmful,  largely  due  to  the  fact  that  74%  of  the  patient  outcome  data  held  by  Pfizer  had   never  been  published  (5).   Regulatory  bodies,  historically,  have  not  been  quick  at  removing  potentially  dangerous  drugs   from  the  market.  For  example,  Benfluorex  (Mediator)  is  estimated  to  have  resulted  in  over   1,000  excess  deaths  before  it  was  removed  from  the  market  (45).  

Conclusions The  pharmaceutical  business  is  just  that  –  a  business.  Whilst  its  products  have  the  potential   to  improve  health  and  benefit  patients,  the  magnitude  of  benefit  to  public  health  is  difficult   to   quantify.   Whilst   some   studies   have   suggested   that   many   of   the   recent   gains   in   life  

                                                                                                              8

https://yellowcard.mhra.gov.uk/  


expectancy can   be   attributed   to   new   drugs   (46),   these   assumptions   have   been   challenged   (47).   Critics  will  argue  that  the  business  element  of  the  industry  has  led  to  some  of  the  practices   described   above,   with   a   focus   on   profit,   sales,   and   marketing   over   genuine   improvements   in   patient  benefits.  Further,  the  need  to  maximise  profit  over  a  relatively  short  period  of  time   leads  to  the  suppression  of  concerns  over  risks  associated  with  novel  compounds.   The   number   and   size   of   fines   imposed   on   the   pharmaceutical   industry   in   recent   years   would   suggest  that  many  of  the  more  questionable  practices  are  not  products  of  a  bygone  age.  The   wider   implications   include   the   overpromotion   of   new   drugs   that   offer   little   benefit   over   existing   drugs,   and   the   risk   of   patient   harm   that   is   downplayed   for   as   long   as   possible,   in   order  to  maximise  the  return  to  shareholders.   Evidence-­‐based   medicine,   rather   than   providing   a   mechanism   to   challenge   questionable   claims   has   become   another   tool   through   which   covert   advertising   and   promotion   furthers   the   industry’s   aims,   rather   than   benefiting   patients.   We   will   look   at   some   of   the   other   techniques,  and  discuss  solutions  in  the  second  paper.  

References 1.   Almashat   S,   Wolfe   S.   Pharmaceutical   Industry   Criminal   and   Civil   Penalties:   An   Update.   Washington,  D.C.:  Public  Citizen’s  Health  Research  Group;  2012  September  27,  2012.   2.   Abbott  A.  Novartis  to  shut  brain  research  facility.  Nature.  2011;480(7376):161-­‐2.   3.   Fibiger   HC.   Psychiatry,   The   Pharmaceutical   Industry,   and   The   Road   to   Better   Therapeutics.   Schizophr  Bull.  2012;38(4):649-­‐50.   4.   Graham   DJ,   Campen   D,   Hui   R,   Spence   M,   Cheetham   C,   Levy   G,   et   al.   Risk   of   acute   myocardial   infarction   and   sudden   cardiac   death   in   patients   treated   with   cyclo-­‐oxygenase   2   selective   and   non-­‐ selective   non-­‐steroidal   anti-­‐inflammatory   drugs:   nested   case-­‐control   study.   Lancet.   2005;365(9458):475-­‐81.   5.   Eyding  D,  Lelgemann  M,  Grouven  U,  Härter  M,  Kromp  M,  Kaiser  T,  et  al.  Reboxetine  for  acute   treatment   of   major   depression:   systematic   review   and   meta-­‐analysis   of   published   and   unpublished   placebo  and  selective  serotonin  reuptake  inhibitor  controlled  trials.  BMJ.  2010;341:c4737.   6.   Koesters  M,  Guaiana  G,  Cipriani  A,  Becker  T,  Barbui  C.  Agomelatine  efficacy  and  acceptability   revisited:  systematic  review  and  meta-­‐analysis  of  published  and  unpublished  randomised  trials.  Br  J   Psychiat.  2013;203(3):179-­‐87.   7.   Ethical   Standards   in   Health   and   Life   Sciences   Group.   Guidance   on   collaboration   between   healthcare   professionals   and   the   pharmaceutical   industry.   London:   Ethical   Standards   in   Health   and   Life  Sciences  Group;  2012.   8.   Goldacre   B,   Carroll   D,   Hall   E.   Guidelines   for   collaboration   with   industry   should   be   transparent.  BMJ.  2013;347:f6100.   9.   Kneller  R.  The  importance  of  new  companies  for  drug  discovery:  origins  of  a  decade  of  new   drugs.  Nat  Rev  Drug  Discov.  2010;9(11):867-­‐82.   10.   DiMasi   JA,   Hansen   RW,   Grabowski   HG.   The   price   of   innovation:   new   estimates   of   drug   development  costs.  Journal  of  Health  Economics.  2003;22(2):151-­‐85.   11.   DiMasi   JA,   Hansen   RW,   Grabowski   HG.   Setting   the   record   straight   on   setting   the   record   straight:   Response   to   the   Light   and   Warburton   rejoinder.   Journal   of   Health   Economics.   2005;24(5):1049-­‐53.   12.   DiMasi  JA,  Hansen  RW,  Grabowski  HG.  Extraordinary  claims  require  extraordinary  evidence.   Journal  of  Health  Economics.  2005;24(5):1034-­‐44.   13.   Light   DW,   Warburton   RN.   Extraordinary   claims   require   extraordinary   evidence.   Journal   of   Health  Economics.  2005;24(5):1030-­‐3.   14.   Light   DW,   Warburton   RN.   Setting   the   record   straight   in   the   reply   by   DiMasi,   Hansen   and   Grabowski.  Journal  of  Health  Economics.  2005;24(5):1045-­‐8.   15.   Light   DW,   Warburton   R.   Demythologizing   the   high   costs   of   pharmaceutical   research.   BioSocieties.  2011;6(1):34-­‐50.  


16. van   Luijn   JC,   Gribnau   FW,   Leufkens   HG.   Superior   efficacy   of   new   medicines?   Eur   J   Clin   Pharmacol.  2010;66(5):445-­‐58.   17.   House   of   Commons   Health   Committee.   The   Influence   of   the   Pharmaceutical   Industry.   London:  The  House  of  Commons;  2005  22  March  2005.   18.   Schwartz   LM,   Woloshin   S.   How   the   FDA   forgot   the   evidence:   the   case   of   donepezil   23   mg.   BMJ.  2012;344:e1086.   19.   Gagnon   M-­‐A,   Lexchin   J.   The   Cost   of   Pushing   Pills:   A   New   Estimate   of   Pharmaceutical   Promotion  Expenditures  in  the  United  States.  PLoS  Med.  2008;5(1):e1.   20.   Lee   K,   Bacchetti   P,   Sim   I.   Publication   of   Clinical   Trials   Supporting   Successful   New   Drug   Applications:  A  Literature  Analysis.  PLoS  Med.  2008;5(9):e191.   21.   Turner  EH,  Knoepflmacher  D,  Shapley  L.  Publication  Bias  in  Antipsychotic  Trials:  An  Analysis   of   Efficacy   Comparing   the   Published   Literature   to   the   US   Food   and   Drug   Administration   Database.   PLoS  Medicine.  2012;9(3):e1001189.   22.   Turner   EH,   Matthews   AM,   Linardatos   E,   Tell   RA,   Rosenthal   R.   Selective   publication   of   antidepressant  trials  and  its  influence  on  apparent  efficacy.  N  Engl  J  Med.  2008;358(3):252-­‐60.   23.   Lexchin   J,   Bero   LA,   Djulbegovic   B,   Clark   O.   Pharmaceutical   industry   sponsorship   and   research   outcome  and  quality:  systematic  review.  BMJ.  2003;326(7400):1167-­‐70.   24.   Montgomery  JH,  Byerly  M,  Carmody  T,  Li  B,  Miller  DR,  Varghese  F,  et  al.  An  analysis  of  the   effect   of   funding   source   in   randomized   clinical   trials   of   second   generation   antipsychotics   for   the   treatment  of  schizophrenia.  Controlled  Clinical  Trials.  2004;25(6):598-­‐612.   25.   Prayle   AP,   Hurley   MN,   Smyth   AR.   Compliance   with   mandatory   reporting   of   clinical   trial   results  on  ClinicalTrials.gov:  cross  sectional  study.  BMJ.  2012;344:d7373.   26.   Huser  V,  Cimino  JJ.  Linking  ClinicalTrials.gov  and  PubMed  to  Track  Results  of  Interventional   Human  Clinical  Trials.  PLoS  ONE.  2013;8(7):e68409.   27.   Ross   JS,   Tse   T,   Zarin   DA,   Xu   H,   Zhou   L,   Krumholz   HM.   Publication   of   NIH   funded   trials   registered  in  ClinicalTrials.gov:  cross  sectional  analysis.  BMJ.  2012;344:d7292.   28.   Institute   of   Medicine.   Conflict   of   Interest   in   Medical   Research,   Education,   and   Practice   Washington,  DC:  Institute  of  Medicine;  2009  April  28,  2009.   29.   Brodkey   AC.   The   Role   of   the   Pharmaceutical   Industry   in   Teaching   Psychopharmacology:   A   Growing  Problem.  Acad  Psychiatry.  2005;29(2):222-­‐9.   30.   Davison   WP.   The   Third-­‐Person   Effect   in   Communication.   Public   Opinion   Quarterly.   1983;47(1):1-­‐15.   31.   Rutledge   P,   Crookes   D,   McKinstry   B,   Maxwell   SR.   Do   doctors   rely   on   pharmaceutical   industry   funding  to  attend  conferences  and  do  they  perceive  that  this  creates  a  bias  in  their  drug  selection?   Results  from  a  questionnaire  survey.  Pharmacoepidemiology  and  Drug  Safety.  2003;12(8):663-­‐7.   32.   Spurling   GK,   Mansfield   PR,   Montgomery   BD,   Lexchin   J,   Doust   J,   Othman   N,   et   al.   Information   from   Pharmaceutical   Companies   and   the   Quality,   Quantity,   and   Cost   of   Physicians'   Prescribing:   A   Systematic  Review.  PLoS  Med.  2010;7(10):e1000352.   33.   Ioannidis  JA.  Are  Medical  Conferences  Useful?  And  for  Whom?  JAMA.  2012;307(12):1257-­‐8.   34.   Søndergaard   J,   Vach   K,   Kragstrup   J,   Andersen   M.   Impact   of   pharmaceutical   representative   visits  on  GPs'  drug  preferences.  Fam  Pract.  2009;26(3):204-­‐9.   35.   Wazana   A.   Physicians   and   the   Pharmaceutical   Industry:   Is   a   Gift   Ever   Just   a   Gift?   JAMA.   2000;283(3):373-­‐80.   36.   Fugh-­‐Berman   A,   Ahari   S.   Following   the   Script:   How   Drug   Reps   Make   Friends   and   Influence   Doctors.  PLoS  Medicine.  2007;4(4):0621-­‐5.   37.   Katz  D,  Caplan  AL,  Merz  JF.  All  Gifts  Large  and  Small:  Toward  an  Understanding  of  the  Ethics   of  Pharmaceutical  Industry  Gift-­‐Giving.  The  American  Journal  of  Bioethics.  2010;10(10):11-­‐7.   38.   Orlowski   JP,   Wateska   L.   The   effects   of   pharmaceutical   firm   enticements   on   physician   prescribing  patterns.  There's  no  such  thing  as  a  free  lunch.  Chest.  1992;102(1):270-­‐3.   39.   Gilbody   S,   Wilson   P,   Watt   I.   Benefits   and   harms   of   direct   to   consumer   advertising:   a   systematic  review.  Quality  and  Safety  in  Health  Care.  2005;14(4):246-­‐50.   40.   Gellad  ZF,  Lyles  KW.  Direct-­‐to-­‐Consumer  Advertising  of  Pharmaceuticals.  American  Journal  of   Medicine.  2007;120(6):475-­‐80.   41.   Mintzes  B.  Should  patient  groups  accept  money  from  drug  companies?  No.  BMJ.  2007;334.   42.   Berenson  A.  Evidence  in  Vioxx  Suits  Shows  Intervention  by  Merck  Officials.  New  York  Times.   2005  April  24;Sect.  A5.   43.   Cohen  D.  Rosiglitazone:  what  went  wrong?  BMJ.  2010;341:c4848.  


44. Applbaum   K.   Shadow   science:   Zyprexa,   Eli   Lilly   and   the   globalization   of   pharmaceutical   damage  control.  BioSocieties  2010;5(2):236–55.   45.   Fournier   A,   Zureik   M.   Estimate   of   deaths   due   to   valvular   insufficiency   attributable   to   the   use   of   benfluorex   in   France.   Pharmacoepidemiology   and   Drug   Safety   [In   Press:   doi:101002/pds3213].   2012.   46.   Lichtenberg  FR.  Why  Has  Longevity  Increased  More  in  Some  States  than  in  Others?  The  Role   of   Medical   Innovation   and   Other   Factors.   New   York,   NY:   Manhattan   Institute   for   Policy   Research;   2007.   47.   Baker  D,  Fugh-­‐Berman  A.  Do  New  Drugs  Increase  Life  Expectancy?  A  Critique  of  a  Manhattan   Institute  Paper.  J  Gen  Intern  Med.  2009;24(5):678-­‐82.                                                    


Has the   pharmaceutical   industry   commandeered   evidence-­‐ based  medicine?  2)  Solutions                   Dr  David  Christmas  (Consultant  Psychiatrist,  Advanced  Interventions  Service,  NHS  Tayside)   Correspondence  to:  David  Christmas:  david.christmas@nhs.net                       Please  note  the  super-­‐script  numbers  refer  to  links  at  the  bottom  of  each  page  (in  the  form   of  weblinks  and  definitions),  and  numbers  in  brackets  refer  to  main  reference  list  

ABSTRACT   Evidence-­‐based   medicine,   rather   than   providing   a   mechanism   to   challenge   questionable   claims   has   become   another   tool   through   which   covert   advertising   and   promotion   furthers   the   industry’s   aims,   rather   than   benefiting   patients.     This   paper   will   review   how   some   of   the   techniques   used   by   pharmaceutical   companies   affect   evidence   based   medicine,   before   suggesting  potential  solutions  to  some  of  these  challenges.       It   may   be   some   time   before   confidence   can   be   fully   restored   in   academic   medicine,   and   clinicians   and   patients   can   be   reassured   that   published   study   outcomes   are   free   from   influence.  There  are  changes  occurring  which  are  moving  towards  greater  transparency,  and   some   pharmaceutical   companies   are   speaking   the   right   language,   but   only   time   will   tell   whether   the   pharmaceutical   industry   are   prepared   to   shake   off   their   chequered   past   and   change  for  good.     Key  Words:    pharmaceutical  company;  research  malpractice;  Big  Pharma  

Introduction In  the  first  paper  we  saw  how  some  of  the  claims  made  by  the  pharmaceutical  industry  don’t   stand  up  to  scrutiny.  Some  additional  issues  have  the  potential  to  impact  significantly  on  the   safety   of   our   medicines.   First,   selective   publication   leads   to   an   exaggeration   of   benefits   whilst  harms  are  downplayed.  This  has  led  to  cases  where  licensed  drugs  were  only  found  to   be  harmful  long  after  deaths  were  occurring.  Second,  regulatory  authorities  are  not  given  (or   are   not     demanding)   enough   information   to   make   fully-­‐informed   decisions   on   the   medicines   they   are   approving   –   full,   detailed,   patient-­‐level   data   is   held   by   the   drug   companies   and   rarely  released.  This  makes  it  impossible  for  researchers  to  determine  the  true  benefits  and   risks  of  particular  treatments.   Before  discussing  some  solutions,  it  makes  sense  to  discuss  some  other  techniques  that  are   used  to  ‘muddy  the  waters’  when  trying  to  determine  the  safety  and  efficacy  of  medicines.  

Subversion of  Evidence-­‐Based  Medicine   Off-­‐Label  Marketing   Off-­‐label   marketing,   the   promotion   of   drugs   for   conditions   that   the   drug   doesn’t   have   a   license  for,  is  relatively  common.  Indeed,  a  significant  proportion  of  drug  company  financial   penalties   relates   to   off-­‐label   marketing   and   some   of   the   largest   corporate   fines   include   penalties  for  off-­‐label  promotion  (1).   Although   it   is   not-­‐permitted,   off-­‐label   marketing   can   generate   significant   additional   revenue   for   the   company.   Whilst   a   company   cannot   easily   promote   study   findings   outside   of   its  


marketing license,  conference  presentations  and  posters  provide  a  platform  for  the  display   of  data  relating  to  non-­‐licensed  indications  and  can  still  be  cited  in  advertising  for  the  drug   (2).  Continuing  Professional  Development,  and  educational  seminars,  provides  another  way   in  which  companies  can  propagate  off-­‐label  data  by  asking  key-­‐opinion  leaders  (KOLs)  to  talk   about  their  experience  of  the  drug  to  fellow  professionals  (3).   The  release  of  company-­‐confidential  material  as  part  of  litigation  proceedings  in  the  US  has   highlighted   the   role   that   off-­‐label   marketing   plays   in   the   overall   release   of   a   new   drug;   in   many   cases,   it   will   be   an   active   part   of   the   promotional   strategy.   For   example,   off-­‐label   promotion   formed   part   of   the   advertising   strategy   for   Quetiapine19  and   companies   have   been  fined  for  the  off-­‐label  selling  of  Paroxetine  (Paxil®)  (4),  Olanzapine  (Zyprexa®)  (5),  and   Semisodium  Valproate  (Depakote®)  (6).  

Ghost-­‐Writing Ghost-­‐writing   is   the   practice   whereby   a   company   will   typically   ask   an   external   medical   writing   company   to   draft   papers   that   are   favourable   to   the   company’s   product.   They   will   then  invite  prominent  researchers  or  key-­‐opinion  leaders  to  become  authors,  in  an  attempt   to   increase   the   credibility   of   the   paper.   The   original   author(s)   are   rarely   acknowledged,   hence   their   description   as   ‘ghosts’.   A   variation   is   ‘guest’   authorship,   when   a   physician   or   researcher  who  has  made  little  contribution  to  the  scientific  work  is  credited  as  an  author.   Ghost-­‐writing   is   common,   but   can   be   difficult   to   spot.   A   common   clue   is   the   use   of   the   term   ‘editorial  assistance’  in  the  acknowledgements  section.  Very  often,  this  person  will  have  had   a   much   more   prominent   role   in   drafting   the   paper   and   will   commonly   work   for   a   medical   writing   firm.   The   large   depression   study   STAR*D   resulted   in   approximately   124   papers,   at   least   45   of   which   acknowledge   Jon   Kilner   (who   is   also   a   science   fiction   author2 10 )   as   providing   ‘editorial   assistance’.   Standard   rules   about  authorship   do   not   necessarily   protect   against  ghostwriting  (7).   A   number   of   studies   have   been   examined   in   connection   with   ghost-­‐writing.   One   of   the   most   prominent   is   Study   329,   a   placebo-­‐controlled   study   of   Paroxetine   (Paxil)   for   depression   in   adolescents  which  was  later  determined  to  have  been  written  by  Sally  Laden,  who  worked   for  the  medical  writing  firm  Scientific  Therapeutics  Information  (STI)  (8).  

Fixing the  problem   There   are   numerous   possible   solutions   to   the   challenges   currently   facing   evidence-­‐based   medicine   which   hopefully   will   begin   to   re-­‐establish   trust   in   academic   medicine   and   clinical   research.  

Improved disclosures  of  conflicts  of  interest   Potential   conflicts   of   interest   have   been   described   in   the   drafting   of   the   Diagnostic   and   Statistical   Manual   of   Mental   Disorders   in   both   the   recent   versions:   DSM-­‐IV   and   DSM-­‐5   (9,   10).  They  have  also  been  identified  in  the  groups  producing  clinical  practice  guidelines  (11-­‐ 14).   It   is   not   just   professionals   that   have   ties   with   the   pharmaceutical   industry.   Third   sector   organisations  also  have  significant  amounts  of  their  funding  coming  from  industry  (15).  The   overall   effect   is   that   it   is   sometimes   difficult   to   determine   where   particular   interests   lie,   and  

                                                                                                             

1

http://psychrights.org/research/Digest/NLPs/Seroquel/090520UnsealedSeroquelExhibits/Off-­‐ label/Schwartz  7  [Redacted]  -­‐  Seroquel  Strategy  Summary  (12-­‐18-­‐00).pdf       2 http://kilnerwriter.net/bookspage.htm    


how those   interests   are   influencing   the   scientific   literature.   Undisclosed   interests   are   particular   concerning   (16).   For   example,   a   group   of   scientists   published   a   review   of   research   on   vagus   nerve   stimulation   (VNS),   a   controversial   treatment   for   depression   (16).   However,   the  authors  omitted  they  are  paid  advisers  to  the  company  that  manufactures  a  device  for   VNS  that  was  approved  in  2007  by  the  U.S.  Food  and  Drug  Administration  (16).   In   a   response   to   the   problem,   the   US   has   introduced   what   is   known   as   the   ‘Sunshine   Act’   where   pharmaceutical   companies   have   to   disclose   all   payments   to   doctors.   These   payments   will   be   made   available   on   publicly-­‐accessible   websites311.   However,   it   is   worth   bearing   in   mind  that  disclosure  doesn’t  automatically  remove  a  conflict  of  interest.  Furthermore,  it  is   possible   that   biases   may   be   magnified   by   disclosure   (17),   through   mechanisms   such   as   “strategic  exaggeration”  and  “moral  licensing”  (18).    

Comparative trials  should  be  the  norm   Pharmaceutical  companies  do  not  have  to  demonstrate  that  their  new  drugs  are  better  than   existing  treatments,  they  only  have  to  demonstrate  that  they  are  better  than  placebo.  This   marker  of  success  is  often  much  easier  to  achieve.  This  is  particularly  true  in  the  USA  where   placebo-­‐controlled  trials  are  mandated  by  the  FDA.  Such  an  approach,  whilst  ensuring  that   drugs  demonstrate  efficacy,  results  in  a  number  of  consequences.   First,  where  it  is  known  that  a  particular  type  of  intervention  is  better  than  placebo  it  raises   ethical   issues   about   whether   it   is   appropriate   to   offer   people   dummy   medication   (19).   Second,   placebo-­‐controlled   trials   are   usually   conducted  across   a   number   of   different   sites   in   different   populations.   This   can   result   in   differences   between   subgroups   being   obscured   in   the   overall   need   to   demonstrate   efficacy   for   the   average   patient.   In   extreme   cases,   drugs   that  appear  efficacious  for  one  gender  may  offer  little  benefit  for  the  other.  

Better evaluations  of  the  benefits  of  new  medications   We  saw  in  the  first  paper  that  the  benefits  of  new  medications  often  seem  small.  Over  time,   we   are   seeing   fewer   and   fewer   completely   novel   drugs,   or   drugs   that   offer   major   gains   over   existing   agents.   In   recent   years,   the   French   industry   journal   Prescrire   has   found   it   increasingly   difficult   to   award   drugs   that   offer   genuine   incremental   advances   in   care   (20).   Only   1-­‐in-­‐10   new   drugs   offers   “a   statistically   significant   difference   in   primary   clinical   endpoints”  (21).   Consequently,   as   drug   costs   remain   high   but   gains   become   smaller,   there   is   an   increasing   need   for   health   services   to   ensure   that   public   funds   are   spent   in   the   most   effective   and   efficient  way  possible.  

Value-­‐based pricing   Value-­‐based   pricing   for   branded   drugs   will   come   into   effect   in   2014.   NICE   has   always   evaluated  treatments  on  the  basis  of  gains  in  utility  –  i.e.  quality  of  life  and  functioning  was   an   important   measure   when   assessing   the   benefits   of   an   intervention.   Under   the   new   rules,   when  considering  new  treatments,  NICE  will  look  at  costs  that  extend  beyond  those  borne   by  the  NHS  to  include  costs  to  carers  and  costs  relating  to  employment  (22).  This  means  that   if   a   treatment   improves   quality   of   life   to   a   similar   degree   to   other   drugs,   but   improves   other   outcomes  more,  it  may  be  considered  more  favourably  by  NICE.  

                                                                                                              3  http://www.ama-­‐assn.org/ama/pub/advocacy/topics/sunshine-­‐act-­‐and-­‐physician-­‐financial-­‐

transparency-­‐reports.page  


Some have   voiced   concerns   that   drugs   that   are   second   to   market   (i.e.   ‘me-­‐toos’)   may   be   disadvantaged,  but  value-­‐based  pricing  means  that  they  will  be  unable  to  charge  more  than   existing  drugs  unless  they  can  demonstrate  additional  benefit  (23).  

Full disclosure  of  data  from  Clinical  Study  Reports  (CSRs)   Since   November   2010,   the   European   Medicines   Agency   (EMA)   has   released   almost   two   million  pages  of  clinical  trial  information  in  response  to  requests  from  research  groups  and   other  agencies,  as  part  of  its  2010  access-­‐to-­‐documents  policy.  The  policy  aims  to  give  access   to   non-­‐clinical   and   clinical   information   (including   clinical   study   reports)   that   have   been   submitted   by   pharmaceutical   companies   as   part   of   their   applications   for   marketing-­‐ authorisation.   Such   apparent   transparency   has   been   welcomed   by   medical   researchers,   particularly   given   EMA’s   history   of   being   unwilling   to   disclose   data   on   the   grounds   of   commercial   sensitivity   (24).   However,   the   moves   are   being   actively   resisted   by   some   pharmaceutical  companies  (see  below).  

The AllTrials  campaign   The   AllTrials   campaign   was   established   in   early   2013   as   a   collaboration   between:   Bad   Science   (led   by   Ben   Goldacre),   the   British   Medical   Journal,   the   Centre   for   Evidence-­‐based   Medicine,   the   Cochrane   Collaboration,   the   James   Lind   Initiative,   Public   Library   of   Science,   and  Sense  About  Science.  The  campaign  calls  for  all  clinical  trials  to  be  registered  and  their   results  reported,  and  it  has  produced  detailed  plans  for  this  aspiration.412   The   aspiration   to   ensure   full   disclosure   of   clinical   study   reports   is   an   important   step   in   ensuring  that  decisions  can  be  made  about  healthcare  interventions.  It  has  been  estimated   that  less  than  50%  of  clinical  trial  outcomes  are  in  the  public  domain;  i.e.  published  in  peer-­‐ reviewed   literature   (25).   Importantly,   clinical   study   reports   (CSRs)   contain   much   more   detailed  information  about  harms  of  medication.   We   have   seen   in   the   first   paper   that   selective   publication   and   withholding   of   clinically-­‐ relevant   information   leads   to   poor   decision-­‐making.   Reboxetine   was   considered   to   be   an   effective  and  safe  antidepressant  on  the  basis  of  published  studies,  but  it  was  only  when  the   German   equivalent   of   NICE   obtained   unpublished   data   did   it   emerge   that   reboxetine   was   ineffective   and   potentially   harmful   (26).   The   problem   was   that   75%   of   patient   data   had   never  been  published.  

Pharmaceutical company  sign-­‐up  to  the  AllTrials  campaign   A  few  drug  companies  have  publicly  announced  their  support  for  the  AllTrials  campaign  for   clinical   data   transparency.   GlaxoSmithKline   was   one   of   the   first   to   announce   their   commitment   to   providing   trial   data513  in   early   2013,   but   it   has   gone   on   to   be   caught   up   in   bribery   scandals   in   China614.   Roche   has   also   declared   a   greater   willingness   to   share   clinical   trial   data   (27),   although   there   remains   some   scepticism   given   Roche’s   unwillingness   to   disclose  data  regarding  Oseltamavir  (28,  29).  

Is Big  Pharma  resisting  change?   Two  pharmaceutical  companies  (AbbVie  and  Intermune)  are  challenging  the  EMA’s  plans  to   provide   access   to   clinical   information   (including   CSRs)   that   were   submitted   as   part   of  

                                                                                                             

4

http://www.alltrials.net/all-­‐trials/      http://www.gsk.com/media/press-­‐releases/2013/GSK-­‐announces-­‐support-­‐forAll-­‐Trials-­‐campaign-­‐ for-­‐clinical-­‐data-­‐transparency.html     6  http://www.bbc.co.uk/news/business-­‐23402154     5


applications for  marketing-­‐authorisation.  Most,  it  should  be  noted,  have  not  launched  legal   proceedings.  At  a  meeting  hosted  by  the  European  Federation  of  Pharmaceutical  Industries   and  Associations  (EFPIA),  a  representative  from  AbbVie  (a  research-­‐based  company  that  split   from  Abbott  in  early  2013),  claimed  that  information  about  adverse  effects  “is  confidential   commercial   information   because   if   released   other   companies   could   use   it   to   help   them   get   products   approved.”   The   Head   of   the   Dutch   board   which   evaluates   medicines   went   on   to   question   whether   AbbVie   really   thought   that   adverse   events   are   commercially   sensitive   information   and   Hans   Georg   Eichler,   the   EMA’s   chief   medical   offer   said   “I   have   been   a   regulator  for  many  years  and  I  am  totally  flabbergasted.”715  

A willingness  to  resist  indication  creep  and  expanding  disease  definitions   Many   people   are   concerned   by   the   gradual   expansion   of   disease   definitions   so   that   previously-­‐normal   health   states   are   now   considered   to   be   disease   and   requiring   of   treatment.  As  a  general  rule,  medicine  tends  to  expand  disease  rather  than  contract  it  (13).   Some   examples   include   hypertension   and   hypercholesterolaemia   (where   the   benefits   of   treatment  for  those  in  low  risk  groups  have  been  questioned)  (30)  and  metabolic  syndrome   and   diabetes   where   the   number   of   potential   ‘customers’   might   be   huge   but   benefits   unknown  (31).     A   related   phenomena   is   indication   creep   –   the   use   or   promotion   of   a   drug   for   off-­‐label   indications   (32).   This   has   been   discussed   above,   and   while   not   necessarily   bad   medicine   (eg:   amitriptyline  for  neuropathic  pain)  it  may  be  related  to  drug  companies  trying  to  maximise   profits  by  widening  the  therapeutic  indications  for  their  drugs  (eg:  Levodopa  for  restless  legs   syndrome).  Despite  being  common,  it  may  be  very  difficult  to  change  current  practice  (1).     Other   related   approaches   include   the   use   of   online   and   self-­‐diagnosis   tools   to   reach   a   larger   range   of   consumers,   many   of   whom   would   never   had   known   that   they   were   unwell   until   they’d  filled  in  on  online  form,  often  on  a  drug  company’s  website  (33).  

An acknowledgement   that   in   many   cases   lifestyle   and   behavioural   interventions   can  be  just  as  effective  as  drugs   In   some   conditions   where   drugs   are   becoming   increasingly   used,   and   in   particular   those   conditions   where   boundaries   between   disease   states   and   normality   are   becoming   blurred   (e.g.   hypercholesterolaemia   and   hypertension),   there   is   very   good   evidence   that   lifestyle   changes   can   be   just   as   effective   as   drugs  (34).   This   requires   a   greater   effort   on   behalf   of   the   patient,  but  the  risks  associated  with  such  interventions  may  be  lower  than  those  associated   with  drug  treatment.  

Summary and  Conclusions   We  have  seen  that  evidence-­‐based  medicine  is  a  powerful  tool  for  improving  care  but  it  is   not   perfect   and   current   systems   do   not   offer   sufficient   protection   from   influence   from   companies   that   have   financial   interests   in   ensuring   that   their   message   is   promoted.   The   historical   record   does   not   support   the   view   that   pharmaceutical   companies   are   changing   quickly.  Systems  are  changing  in  response  to  concerns  from  academics  and  other  interested   parties,   but   we   are   still   not   in   a   position   whereby   we   can   be   confident   that   all   the   data   arising  from  clinical  trials  is  in  the  public  domain.   It  may  be  some  time  before  confidence  can  be  restored  in  academic  medicine,  and  clinicians   and  patients  can  be  reassured  that  published  study  outcomes  are  free  from  influence.  There  

                                                                                                             

7 http://www.alltrials.net/2013/2338/    


are changes   occurring   which   are   moving   towards   greater   transparency,   and   some   pharmaceutical  companies  are  speaking  the  right  language,  but  only  time  will  tell  whether   the  pharmaceutical  industry  are  prepared  to  shake  off  their  chequered  past  and  change  for   good.    

References 1.   Kesselheim   AS,   Mello   MM,   Studdert   DM.   Strategies   and   Practices   in   Off-­‐Label   Marketing   of   Pharmaceuticals:  A  Retrospective  Analysis  of  Whistleblower  Complaints.  PLoS  Med.  2011;8(4):e1000431.   2.   Fugh-­‐Berman  A,  Melnick  D.  Off-­‐Label  Promotion,  On-­‐Target  Sales.  PLoS  Medicine.  2008;5(10):e210.   3.   Moynihan   R.   Key   opinion   leaders:   independent   experts   or   drug   representatives   in   disguise?   BMJ.   2008;336(7658):1402-­‐3.   4.   U.S.   Department   of   Justice.   GlaxoSmithKline   to   Plead   Guilty   and   Pay   $3   Billion   to   Resolve   Fraud   Allegations   and   Failure   to   Report   Safety   Data.     2012   July   2,   2012   [cited   November   10,   2012];   Available   from:   http://www.justice.gov/opa/pr/2012/July/12-­‐civ-­‐842.html   5.   U.S.   Department   of   Justice.   Eli   Lilly   and   Company   Agrees   to   Pay   $1.415   Billion   to   Resolve   Allegations   of   Off-­‐label   Promotion   of   Zyprexa.     2009   January   15,   2009   [cited   November   10,   2012];   Available   from:   http://www.justice.gov/opa/pr/2009/January/09-­‐civ-­‐038.html   6.   U.S.  Department  of  Justice.  Abbott  Labs  to  Pay  $1.5  Billion  to  Resolve  Criminal  &  Civil  Investigations  of   Off-­‐label   Promotion   of   Depakote.     2012   May   7,   2012   [cited   November   10,   2012];   Available   from:   http://www.justice.gov/opa/pr/2012/May/12-­‐civ-­‐585.html   7.   Leo   J,   Lacasse   JR,   Cimino   AN.   Why   Does   Academic   Medicine   Allow   Ghostwriting?   A   Prescription   for   Reform.  Society.  2011;48(5):371-­‐5.   8.   McHenry   LB,   Jureidini   JN.   Industry-­‐Sponsored   Ghostwriting   in   Clinical   Trial   Reporting:   A   Case   Study.   Accountability  in  Research.  2008;15(3):152-­‐67.   9.   Cosgrove  L,  Krimsky  S.  A  Comparison  of  DSM-­‐IV  and  DSM-­‐5  Panel  Members'  Financial  Associations  with   Industry:  A  Pernicious  Problem  Persists.  PLoS  Med.  2012;9(3):e1001190.   10.   Cosgrove  L,  Krimsky  S,  Vijayaraghavan  M,  Schneider  L.  Financial  Ties  between  DSM-­‐IV  Panel  Members   and  the  Pharmaceutical  Industry.  Psychother  Psychosom.  2006;75(3):154-­‐60.   11.   Choudhry  NK,  Stelfox  HT,  Detsky  AS.  Relationships  Between  Authors  of  Clinical  Practice  Guidelines  and   the  Pharmaceutical  Industry.  JAMA.  2002;287(5):612-­‐7.   12.   Mendelson   TB,   Meltzer   M,   Campbell   EG,   Caplan   AL,   Kirkpatrick   JN.   Conflicts   of   Interest   in   Cardiovascular  Clinical  Practice  Guidelines.  Arch  Intern  Med.  2011;171(6):577-­‐84.   13.   Moynihan   RN,   Cooke   GPE,   Doust   JA,   Bero   L,   Hill   S,   Glasziou   PP.   Expanding   Disease   Definitions   in   Guidelines  and  Expert  Panel  Ties  to  Industry:  A  Cross-­‐sectional  Study  of  Common  Conditions  in  the  United  States.   PLoS  Med.  2013;10(8):e1001500.   14.   Neuman   J,   Korenstein   D,   Ross   JS,   Keyhani  S.  Prevalence  of  financial  conflicts  of  interest  among  panel   members   producing   clinical   practice   guidelines   in   Canada   and   United   States:   cross   sectional   study.   BMJ.   2011;343:d5621.   15.   Harris   G.   Drug   Makers   Are   Advocacy   Group’s   Biggest   Donors.   New   York   Times.   2009   October   22,   2009;Sect.  A23.   16.   Holden   C.   The   Undisclosed   Background   of   a   Paper   on   a   Depression   Treatment.   Science.   2006;313(5787):598-­‐9.   17.   The   PLoS   Medicine   Editors.   Does   Conflict   of   Interest   Disclosure   Worsen   Bias?   PLoS   Med.   2012;9(4):e1001210.   18.   Loewenstein   G,   Sah   S,   Cain   DM.   The   unintended   consequences   of   conflict   of   interest   disclosure.   JAMA.   2012;307(7):669-­‐70.   19.   Lewis   JA,   Jonsson   B,   Kreutz   G,   Sampaio   C,   van   Zwieten-­‐Boot   B.   Placebo-­‐controlled   trials   and   the   Declaration  of  Helsinki.  Lancet.  2002;359(9314):1337-­‐40.   20.   Anonymous.   1981-­‐2011:   31   years   of   Prescrire   Drug   Awards.   Prescrire   International.   2011;21(125):78-­‐ 82.   21.   van   Luijn   JC,   Gribnau   FW,   Leufkens   HG.   Superior   efficacy   of   new   medicines?   Eur   J   Clin   Pharmacol.   2010;66(5):445-­‐58.   22.   Raftery  J.  Value  based  pricing:  can  it  work?  BMJ.  2013;347:f5941.   23.   Claxton  K,  Briggs  A,  Buxton  MJ,  Culyer  AJ,  McCabe  C,  Walker  S,  et  al.  Value  based  pricing  for  NHS  drugs:   an  opportunity  not  to  be  missed?  BMJ.  2008;336:251.   24.   Gøtzsche   PC,   Jørgensen   AW.   Opening   up   data   at   the   European   Medicines   Agency.   BMJ.   2011;342:d2686.   25.   Wieseler   B,   Wolfram   N,   McGauran   N,   Kerekes   MF,   Vervölgyi   V,   Kohlepp   P,   et   al.   Completeness   of   Reporting   of   Patient-­‐Relevant   Clinical   Trial   Outcomes:   Comparison   of   Unpublished   Clinical   Study   Reports   with   Publicly  Available  Data.  PLoS  Med.  2013;10(10):e1001526.  


26. Eyding  D,  Lelgemann  M,  Grouven  U,  Härter  M,  Kromp  M,  Kaiser  T,  et  al.  Reboxetine  for  acute  treatment   of   major   depression:   systematic   review   and   meta-­‐analysis   of   published   and   unpublished   placebo   and   selective   serotonin  reuptake  inhibitor  controlled  trials.  BMJ.  2010;341:c4737.   27.   O’Dowd   A.   GSK   and   Roche   tell   MPs   that   more   clinical   trial   data   should   be   published.   BMJ.   2013;346:f2639.   28.   Cohen  D.  Complications:  tracking  down  the  data  on  oseltamivir.  BMJ.  2009;339(dec08_3):b5387.   29.   Doshi  P,  Jefferson  T,  Del  Mar  C.  The  Imperative  to  Share  Clinical  Study  Reports:  Recommendations  from   the  Tamiflu  Experience.  PLoS  Med.  2012;9(4):e1001201.   30.   Abramson  JD,  Rosenberg  HG,  Jewell  N,  Wright  JM.  Should  people  at  low  risk  of  cardiovascular  disease   take  a  statin?  BMJ.  2013;347:f6123.   31.   Hauber  A,  Gale  EA.  The  market  in  diabetes.  Diabetologia.  2006;49(2):247-­‐52.   32.   Djulbegovic  B,  Paul  A.  From  Efficacy  to  Effectiveness  in  the  Face  of  Uncertainty:  Indication  Creep  and   Prevention  Creep.  JAMA.  2011;305(19):2005-­‐6.   33.   Ebeling  M.  ‘Get  with  the  Program!’:  Pharmaceutical  marketing,  symptom  checklists  and  self-­‐diagnosis.   Social  Science  &  Medicine.  2011;73(6):825-­‐32.   34.   Naci   H,   Ioannidis   JPA.   Comparative   effectiveness   of   exercise   and   drug   interventions   on   mortality   outcomes:  metaepidemiological  study.  BMJ.  2013;347:f5577.  

                                     


The Human  Genome  Project:  Undervalued  Ingenuity Personal  View    /  Commentary  Piece  

Iain  J  Hyndman  (3rd  year  MBChB,  University  of  Dundee)     Correspondence  to:  Iain  J  Hyndman  :    i.hyndman@dundee.ac.uk                

ABSTRACT

This article  aims  to  inform  medical  students  and  clinicians  about  the  Human  Genome  Project   (HGP).   The   article   discusses   the   barriers   that   have   been   broken   down   to   allow   for   wider   access   to   genetic   testing   and   the   potentially   negative   effects   that   an   increase   in   genetic   testing  may  have  on  patients.  It  is  hoped  that  by  contrasting  the  triumphs  of  the  HGP  (such   as   personalised   medicine)   with   the   potential   pitfalls   of   the   project   (such   as   genetic   discrimination),  readers  will  develop  an  enhanced  understanding  of  the  HGP.     Key  Words:    human  genome  project;  genetics  

Introduction  

If you  were  to  take  a  group  of  medical  students  and  ask  them  to  identify  the  single  greatest   breakthrough  in  medical  science,  I  am  quite  certain  that  the  Human  Genome  Project  (HGP)   would  not  feature  highly.       The  reason  is  quite  simple:  people  do  not  know  what  it  is.  The  HGP  was  a  scientific  project   that   boasted   the   aim   of   sequencing   all   3   billion   base   pairs   in   the   human   genome   –   the   blueprint   of   life.   It   began   formally   in   1990   and   was   declared   complete   in   2003.   The   HGP   remains   the   largest   collaborative   research   project   in   human   history   and   it   is   therefore   surprising   that   future   doctors   know   little   about   a   study   which   has   its   roots   in   all   walks   of   medicines   (genes   play   a   role   in   virtually   all   diseases).   The   scale   and   impact   of   the   HGP   is   made  even  more  impressive  when  you  consider  that  DNA  itself  is  a  relatively  new  discovery:   the  double  helix  structure  was  only  discovered  in  1953  by  James  Watson  and  Francis  Crick.1     Now,  60  years  on,  we  all  have  a  basic  understanding  of  DNA  and  the  world  is  a  better  place   for  it.  Or  is  it?  With  genetic  testing  becoming  more  widely  available,  how  can  we  ensure  that   this   information   is   used   appropriately   and   ethically   so   as   to   prevent   discrimination   on   a   genetic  basis?  In  this  article  I  consider  the  breakthroughs  and  drawbacks  of  the  HGP  and  the   challenges  that  lie  ahead.      

Angelina and  BRCA  

Since its   completion,   the   HGP   has   helped   us   to   identify   the   root   of   many   monogenic   diseases.     Monogenic   (or   Mendelian)   diseases   offer   the   best   chance   of   predicting   life-­‐time   risk  of  disease  or  providing  personalised  therapy  as  the  single  gene  mutation  involved  often   makes   for   an   easier   predictive   or   therapeutic   target.       Earlier   this   year   Ms   Jolie   elected   to   have   a   double   mastectomy   to   reduce   her   risk   of   developing  cancer  after  discovering  that  she  carries  the  BRCA1  gene.2  BRCA  1  &  2  are  good   at  predicting  life-­‐time  risk  of  breast  or  ovarian  cancer  (a  female  carrier  of  a  mutation  on  one   of  the  genes  has  a  50-­‐85%  risk  of  breast  cancer  and  a  15-­‐60%  risk  of  ovarian  cancer).3-­‐4       However,   as   with   everything   in   medicine,   predicting   disease   is   not   as   straightforward   as   this   and   indeed   Genome   Wide   Association   Studies   (GWAS)   have   shown   that   there   are   other   genetic   factors   which   have  a  role  in  determining  the  risk  of  developing  cancer,  and  that  only   a   small   number   of   genetic   variants   with   a   role   in   BRCA1/2-­‐related   cancer   have   been   identified   to   date.5   However   the   complexity   of   genetics   should   not   be   seen   by   medical  


students as  a  reason  to  run  for  the  hills  but  rather  as  an  exciting  area  of  research  in  which   there  is  still  much  more  work  to  be  done,  with  the  BRCA  genes  representing  a  piece  in  the   puzzle  of  targeted  drug  therapy.    

Monopoly: one  winner,  many  losers  

Indeed if  ever  there  was  a  time  for  BRCA  testing  to  become  more  popular,  it  is  now.  On  the   13th   of   June   2013   the   US   Supreme   Court   ruled   that   the   previously   patented   BRCA   genes   were  no  longer  eligible  for  patent,  as  they  are  a  naturally  occurring  biological  material.6  This   was   a   major   breakthrough   for   American   women   as   BRCA   testing   previously   cost   $3,340   when   Myriad   Genetics   held   their   legal   monopoly   on   the   gene7.   Just   hours   after   the   announcement,  competing  companies  stated  that  they  would  offer  the  test  at  a  cost  as  low   as   $995.8   It   is   hoped   that   the   dissolution   of   other   gene   patents   will   follow   this   historic   move   and  thus  genetic-­‐based  medicine  will  become  more  widely  available  by  way  of  competition   driving   prices   down.   The   counter   argument   is   that   by   allowing   these   molecules   to   remain   patentable,  we  provide  researchers  with  financial  incentive  to  discover  new  and  innovative   gene   therapies   that   will   ultimately   benefit   patients.     Whatever   your   own   opinion,   affordable   genetic   testing   is   the   only   way   that   patients   will   be   able   to   fully   reap   the   benefits   of   personalised  medicine.    

What exactly  is  ‘personalised  medicine’?  

Personalised medicine   is   the   art   of   tailoring   treatments   to   individuals   based   on   their   genetic   profile.  Personalised  medicine  has  the  potential  to  revolutionise  clinical  decision-­‐making  and   prescribing,  and  could  mean  the  development  of  new  drugs  targeted  at  a  particular  group  of   people,   such   as   a   specific   race   or   even   individuals   with   certain   blood   properties.9   It   could   even  mean  better  use  of  drugs  already  in  use  or  which  failed  in  trial  stage.  There  have  been   many  developments  in  the  field  of  personalised  medicine  in  recent  years.     However,   the   usefulness   of   personalised   medicine   is   limited   by   the   fact   that   we   still   have   so   much   to   learn   about   the   human   genome10,   and   indeed   knowledge   is   only   acquired   with   time.   We   need   only   look   at   cystic   fibrosis   to   see   that   identifying   disease-­‐causing   genes   is   only   part   of   the   solution   to   treating   the   disorder.   It   is   almost   25   years   since   the   most   common   mutation   in   cystic   fibrosis   -­‐   ΔF508   -­‐   was   identified   and   only   now   are   therapies   being   developed   that   can   treat   the   disease   by   targeting   the   faulty   gene   directly11.     Indeed   when   I   sat   in   on   an   asthma   clinic   recently   I   heard   the   doctor   tell   the   patient   that   the   reason  for  changing  her  medication  yet  again  was  “more  out  of  desperation,  than  scientific   rationale”   -­‐   she   had   not   responded   to   all   of   the   other   recommended   treatments   and   the   genetic   testing   that   she   underwent   for   specific   beta2-­‐adrenergic   receptor   mutations   had   come  back  negative.  It  would  seem  that  genetic  testing  is  not  as  clear-­‐cut  as  it  is  often  made   out  to  be  and  there  are  still  many  questions  to  be  answered.    

Opportunity  for  Discrimination  

There is   another   side   to   genetic   testing,   a   darker   side   where   the   threat   of   discrimination   looms.  Indeed  there  are  some  who  fear  that  genetic  testing  could  lead  to  discrimination  by   employers  who  may  wish  to  select  against  those  employees  with  a  genome  that  may  affect   their   ability   to   work.12  The   theoretical   potential   for   discrimination   is   huge,   however   it   is   very   unlikely  that  this  kind  of  discrimination  would  ever  be  legal.  There  is  however  a  real  risk  of   genetic   testing   affecting   health   insurance.   There   are   insurance   companies   who   would   just   love   the   chance   to   use   the   results   of   your   genetic   test   to   charge   you   a   higher   premium,   despite  the  fact  that  we  are  unsure  of  just  how  much  each  “risky”  gene  actually  elevates  the   risk   of   disease.   In   order   to   protect   individuals   from   being   charged   more,   there   are   a   number   of  regulatory  frameworks  in  place  although  there  is  debate  over  how  much  a  risky  genome  


would actually  increase  premiums,  with  some  suggesting  that  the  increase  in   cost  would  be   very  small  indeed12.       The   Council   of   Europe’s   Convention   on   Human   Rights   and   Biomedicine   sets   out   the   laws   relating  to  medical  research  and  states  that  it  “bans  all  forms  of  discrimination  based  on  the   grounds   of   a   person's   genetic   make-­‐up   and   allows   the   carrying   out   of   predictive   genetic   tests   only   for   medical   purposes.”  13   At   present   the   United   Kingdom   has   not   signed   the   treaty   however   it   is   recognised   that   the   United   Kingdom’s   regulation   of   medical   research   with   respect  to  genetic  testing  is  consistent  with  the  European  framework.14   There  is  currently  a   moratorium   in   the   UK   -­‐   a   temporary   ban   agreed   to   last   until   at   least   2017   -­‐   which   forbids   insurance   companies   from   requesting   that   a   person   undergoes   a   genetic   test   when   purchasing   life   or   critical   illness   insurance   and   also   allows   an   individual   the   right   to   withhold   the   results   of   previous   genetic   tests   when   buying   premiums   up   to   a   certain   value.15   But  what  will  happen  in  2017?  Is  genetic  testing  a  ticking  time-­‐bomb  waiting  to  blow  up  in   our   faces?   One   way   to   stop   the   mass   introduction   of   genetic   testing   would   be   to   follow   Germany’s  lead  and  to  only  allow  medical  doctors  to  carry  out  genetic  tests,  as  described  in   their  Genetic  Diagnosis  Act  (GenDg).16  This  may  serve  to  curb  the  recent  increase  in  whole   genome   sequencing   which   has   now   become   more   popular   on   account   of   its   ever-­‐decreasing   price-­‐tag.       There  are  number  of  other  concerns  also  relating  to  genetic  testing.  For  example,  whilst  an   individual  may  purchase  a  genetic  test  to  learn  of  their  susceptibility  to  one  condition  there   may   be   other   conditions   to   which   they   learn   they   are   susceptible,   despite   them   not   wanting   to  know  this  ‘additional’  information.17   There  are  also  likely  to  be  implications  for  all  family   members  if  one  member  is  found  to  be  susceptible  to  a  disease18  and  so  the  impact  of  the   result  should  be  fully  considered  before  the  test.  There  are  also  concerns  that  the  increase  in   genetic  testing  could  lead  to  a  greater  number  of  abortions,  with  many  prospective  parents   potentially   choosing   to   abort   children   with   profound   disabilities   (as   is   already   often   the   case   with  diseases  such  as  Down’s  Syndrome).  It  is  clear  that  genetic  testing  is  an  ethical  and  legal   minefield  with  many  considerations.    

Conclusion

Twenty years  after  its  completion,  the  HGP  remains  a  hot  topic  that  divides  opinion.  There   are  clearly  many  benefits  to  be  reaped  from  genetic  testing  but  at  what  cost  to  patients?  I   believe   that   if   genetic   testing   is   to   become   more   commonplace   then   it   must   be   tightly   regulated  to  protect  individuals  from  discrimination.       There   have   been   many   breakthroughs   in   medicine   as   a   result   of   the   HGP,   such   as   the   beginnings   of   personalised   medicine.   Indeed,   personalised   medicine   has   the   potential   to   prevent   100,000   deaths   caused   by   adverse   drug   reactions   each   year   in   the   US.19       Ultimately,  your  opinion  of  the  HGP  is  likely  to  be  as  individual  as  your  DNA,  but  ponder  this:   where  would  the  HGP  feature  on  your  list  of  the  greatest  breakthroughs  now?  

References  

All references  available  on  the  online  edition  of  this  paper    

Scottish Universities Medical Journal Volume 3 Supplement 1  

Volume 3 Supplement 1 Special Edition - Can we rely on science?

Advertisement