Ethics  in  the  Genomic  Age  

Luce  Skrabanek  ICB,  WMC  

May  30,  2013  

History  and  EvoluEon  of  Research  Ethics  

Timeline  •  1906  Pure  Food  and  Drug  Act  •  1948  Nuremberg  Code    –  (from  the  1946  Nuremberg  Trials)  

•  1962  Kefauver  Harris  Amendment  to  the  Food,  Drug  and  CosmeEc  Act  –  (in  response  to  the  Thalidomide  scandal)  

•  1964  DeclaraEon  of  Helsinki  (updated  1975,  1983,  1989,  1996,  2000,  2008)  

•  1974  NaEonal  Research  Act  –  (in  reacEon  to  the  1932-­‐1973  Tuskegee  Syphilis  Study)  

•  1979  Belmont  Report  

1906  Pure  Food  and  Drug  Act  •  Generally  considered  to  be  the  “birth”  of  the  FDA,  although  

not  named  as  such  unEl  1930  •  Up  unEl  then,  many  patent  medicines  had  high  alcohol  

contents,  infant  syrups  contained  opiates  •  Another  contribuEng  factor  was  the  study  of  common  food  

preservaEves  such  as  borax,  benzoate,  formaldehyde,  sulfites,  and  salicylates  •  Young  men  recruited  to  eat  meals  with  increasing  doses  of  these  

common  preservaEves  were  called  “The  Poison  Squad”  •  Truth  in  adverEsing  

•  Required  that  certain  special  drugs,  including  alcohol,  opium,  cocaine,  heroin,  morphine,  and  cannabis,  be  accurately  labeled  with  contents  and  dosage  

•  1938  Replaced  by  the  Food,  Drug  and  CosmeEc  Act  

1948  Nuremberg  Code  •  1946  Nuremberg  Trials  – Medical  experiments  conducted  on  thousands  of  concentraEon  camp  prisoners  without  their  consent,  resulEng  in  the  death  or  permanent  crippling  of  many  of  them  

•  Includes  10  points:  – Voluntary  consent  of  parEcipants  is  essenEal  – Benefits  of  research  must  outweigh  the  risks  – Avoid  all  unnecessary  harm  and  suffering  

•  First  advocacy  of  voluntary  parEcipaEon  and  informed  consent  

Nuremberg  Code  •  1        The  voluntary  consent  of  the  human  subject  is  absolutely  essenEal.  •  2        The  experiment  should  be  such  as  to  yield  fruiful  results  for  the  good  of  society,  unprocurable  

by  other  methods  or  means  of  study,  and  not  random  and  unnecessary  in  nature.  •  3        The  experiment  should  be  so  designed  and  based  on  the  results  of  animal  experimentaEon  and  

a  knowledge  of  the  natural  history  of  the  disease  or  other  problem  under  study  that  the  anEcipated  results  will  jusEfy  the  performance  of  the  experiment.  

•  4        The  experiment  should  be  so  conducted  as  to  avoid  all  unnecessary  physical  and  mental  suffering  and  injury.  

•  5        No  experiment  should  be  conducted  where  there  is  a  prior  reason  to  believe  that  death  or  disabling  injury  will  occur;  except,  perhaps,  in  those  experiments  where  the  experimental  physicians  also  serve  as  subjects.  

•  6        The  degree  of  risk  to  be  taken  should  never  exceed  that  determined  by  the  humanitarian  importance  of  the  problem  to  be  solved  by  the  experiment.  

•  7        Proper  preparaEons  should  be  made  and  adequate  faciliEes  provided  to  protect  the  experimental  subject  against  even  remote  possibiliEes  of  injury,  disability,  or  death.  

•  8        The  experiment  should  be  conducted  only  by  scienEfically  qualified  persons.  The  highest  degree  of  skill  and  care  should  be  required  through  all  stages  of  the  experiment  of  those  who  conduct  or  engage  in  the  experiment.  

•  9        During  the  course  of  the  experiment  the  human  subject  should  be  at  liberty  to  bring  the  experiment  to  an  end  if  he  has  reached  the  physical  or  mental  state  where  conEnuaEon  of  the  experiment  seems  to  him  to  be  impossible.  

•  10        During  the  course  of  the  experiment  the  scienEst  in  charge  must  be  prepared  to  terminate  the  experiment  at  any  stage,  if  he  has  probable  cause  to  believe,  in  the  exercise  of  the  good  faith,  superior  skill  and  careful  judgment  required  of  him  that  a  conEnuaEon  of  the  experiment  is  likely  to  result  in  injury,  disability,  or  death  to  the  experimental  subject.  

1962  Kefauver  Harris  Amendment  •  “Drug  Efficacy  Amendment”  •  In  response  to  the  Thalidomide  scandal  – Used  as  a  sedaEve  and  an  anE-­‐emeEc  for  morning  sickness  

– Approved  for  used  in  Europe;  no  FDA  approval  in  the  US,  although  many  US  women  took  it  anyway  

–  Caused  severe  deformiEes  in  fetuses  •  Drug  manufacturers  required  to  prove  to  the  FDA  that  their  products  were  effecEve  and  safe,  and  to  disclose  accurate  informaEon  about  side  effects  –  Requires  clinical  trials  with  informed  consent  

1964  DeclaraEon  of  Helsinki  •  Set  of  ethical  principles  regarding  medical  research  developed  by  the  World  Medical  AssociaEon  

•  First  significant  effort  of  the  medical  community  to  regulate  research  itself  

•  Incorporated  Nuremberg  Code  with  DeclaraEon  of  Geneva  (updated  HippocraEc  Oath)  

•  Not  a  legally  binding  instrument  in  internaEonal  law  

•  Updated  1975,  1983,  1989,  1996,  2000,  2008  

DeclaraEon  of  Helsinki  •  Basic  principles  – Research  with  humans  should  be  based  on  the  results  from  lab  and  animal  experimentaEon  

– Research  protocols  should  be  appropriately  examined  

–  Informed  consent  – Research  should  be  undertaken  by  scienEfically/medically  qualified  people  

– Benefits  must  outweigh  risks  

DeclaraEon  of  Helsinki  •  The  US  FDA  rejected  the  2000  and  subsequent  revisions,  only  recognizing  the  third  (1989)  revision,  and  in  2006  announced  it  would  eliminate  all  reference  to  the  DeclaraEon  – Disagreement  over  whether  placebos  were  an  acceptable  control,  even  when  proven  treatments  were  known  

•  As  of  October  2008,  the  DeclaraEon  of  Helsinki  is  replaced  by  Good  Clinical  PracEce  in  the  US  

1974  NaEonal  Research  Act  •  In  reacEon  to  the  Tuskegee  Syphilis  Study  (1932-­‐1973)  –  ProspecEve  longitudinal  study  on  the  natural  progression  of  untreated  syphilis,  even  aker  penicillin  shown  to  be  effecEve  

•  CreaEon  of  the  NaEonal  Commission  for  the  ProtecEon  of  Human  Subjects  of  Biomedical  and  Behavioral  Research  –  Tasked  with  idenEfying  basic  ethical  principles  that  should  underlie  biomedical  research,  and  to  develop  guidelines  on  their  incorporaEon  

Ø Belmont  Report  

1979  Belmont  Report  1.  Respect  for  persons  –  Individuals  should  be  treated  as  autonomous  agents  –  People  with  diminished  autonomy  should  be  protected  

Ø Informed  consent  –  ParEcipants  must  be  given  the  opportunity  to  choose  what  shall  and  shall  not  happen  to  them  

–  Consent  process  includes  •  InformaEon  •  Comprehension  •  Voluntary  parEcipaEon  (no  coercion)  

hlp://www.hhs.gov/ohrp/humansubjects/guidance/belmont.html  

Belmont  Report:  principle  2  2.  Beneficence  and  non-­‐maleficence  – Human  parEcipants  shall  not  be  harmed  unnecessarily  

– Research  should  maximize  possible  benefit  and  minimize  possible  risk  

Ø Assess  risks  and  benefits  in  a  systemaEc  manner  Ø ProtecEon  of  confidenEality  

Belmont  Report:  principle  3  3.  JusEce  – Benefits  and  risks  must  be  distributed  fairly  Ø Research  parEcipants  must  be  selected  using  fair  procedures  

Ø PaEents  have  a  right  to  access  and  receive  care  that  is  non-­‐discriminatory,  suitable  and  fair  

Ø Healthcare  resources  should  be  allocated  in  a  fair  and  equitable  manner  

Other  principles  •  Many  other  factors  also  in  play  here  •  When  we  talk  about  ethics,  we  also  implicitly  include  ELSI:  ethics,  legal  and  social  implicaEons  –  Introduced  by  the  Human  Genome  Project  

•  This  implies  that  we  not  only  take  into  consideraEon  the  medical  ethics  from  the  Helsinki  DeclaraEon,  and  the  research  ethics  put  forward  in  the  Belmont  Report,  but  also  various  legal  rights  (e.g.,  privacy),  as  well  as  implicaEons  of  informaEon  for  the  individual,  the  individual’s  family,  and  society  in  general.  

Right  to  privacy  •  First  implicit  declaraEon  of  a  U.S.  right  to  privacy  

–  1890,  Warren  and  Brandeis,  Harvard  Law  Review  •  Many  people  argue  that  no  such  right  exists  or  that  the  right  to  

privacy  ‘should  not  be  defined  as  a  separate  legal  right’  •  Universal  DeclaraEon  of  Human  Rights,  ArEcle  12:  

–  No  one  shall  be  subjected  to  arbitrary  interference  with  his  privacy,  family,  home  or  correspondence,  nor  to  alacks  upon  his  honor  and  reputaEon.  Everyone  has  the  right  to  the  protecEon  of  the  law  against  such  interference  or  alacks  

•  The  U.S.  Supreme  Court  finds:  –  The  ConsEtuEon  implicitly  grants  a  right  to  privacy  against  

governmental  intrusion  –  Also,  many  states  grant  the  right  to  privacy  and  recognize  the  

following  torts:  •         Intrusion  upon  seclusion  or  solitude,  or  into  private  affairs  •         Public  disclosure  of  embarrassing  private  facts  •         Publicity  which  places  a  person  in  a  false  light  in  the  public  eye  •         AppropriaEon  of  name  or  likeness  

Privacy  •  How  does  health  data  fit  into  privacy  concerns?  •  It  can  be  harmful  to  the  subject  for  certain  health  informaEon  to  be  made  public  –  Imagine  that  a  poliEcian  has  been  diagnosed  with  a  predisposiEon  to  Alzheimer’s.  Public  knowledge  of  this  informaEon  may  result  in  his  not  geqng  re-­‐elected.  

•  But  privacy  is  not  absolute…  –  Imagine  you  have  been  diagnosed  with  SARS.  Do  you  have  a  right  to  keep  that  private?    

•  HIPAA  ensures  some  protecEon  of  healthcare  informaEon  

HIPAA  1996  •  Health  InformaEon  Portability  and  Accountability  Act  – Requires  that  health  care  providers  and  others  protect  the  privacy  of  health  informaEon  

– Sets  boundaries  on  the  use  and  release  of  health  records  

– Empowers  individuals  to  control  certain  uses  and  disclosures  of  their  health-­‐related  informaEon  

– GeneEc  informaEon  is  considered  health  informaEon,  and  is  covered  by  HIPAA    

Genomic  data  

Genomics  data  •  SNPs,  CNVs,  genotype  data,  complete  sequences  

•  Trying  to  decipher  the  geneEc  basis  for  diseases  – Many  of  the  single-­‐gene  Mendelian  diseases  already  known  (OMIM)  

–  Complex  diseases  will  be  affected  by  many  genes,  some  of  which  will  have  very  small  effects  

– Need  large  amounts  of  data  to  get  enough  staEsEcal  power  to  be  able  to  alribute  causaEon  to  variants  

Benefits  •  Ideally,  knowing  the  geneEc  basis  of  an  illness  will  allow  for    –  Earlier  detecEon  of  disease  (perhaps  while  sEll  asymptomaEc)  

–  Knowing  risk  can  lead  to  steps  to  reduce  chances  of  developing  a  parEcular  disease  

– More  targeted  drug  administraEon  (pharmacogenomics:  variaEon  in  drug  response  related  to  geneEc  variaEon)  

–  Beler  therapies  (although  individual  variaEon  complicates  malers)  [precision  medicine]  •  Different  molecular  signatures  will  respond  to  different  treatments  (e.g.,  Cancer  Genome  Atlas  Network  Nature  2012  paper:  4  geneEcally  disEnct  types  of  breast  cancer;  significant  molecular  heterogeneity  within  each  subtype)  

Pharmcogenomics:  CombaEng  adverse  drug  reacEons  

•  Fikh  leading  cause  of  death  in  the  US  •  About  50%  of  drugs  are  metabolized  by  the  cytochrome  P450  family  (~30  genes)  

•  Many  adverse  reacEons  are  caused  by  variaEons  in  enzyme-­‐coding  genes  that  can  increase  or  decrease  drug  metabolism  rates  

•  Increased  metabolism  rates  →  the  drug  is  metabolized  too  quickly  to  have  an  effect  

•  Decreased  metabolism  rates  →  the  levels  of  drugs  in  the  body  quickly  reach  toxic  levels  

Caveats  •  Genomics  is  not  the  end-­‐all  and  be-­‐all  of  scienEfic  explanaEon  – Also  environmental,  social,  cultural,  economic  factors  

•  Fostering  a  gene-­‐centric  view  of  health  may  be  detrimental  by  making  everybody  a  “paEent”  – May  be  appropriate  for  people  with  clear  high  risks  for  diseases  with  appropriate  treatments  (e.g.,  staEns  for  familial  hypercholesterolemia)  

– Most  probably  not  appropriate  for  the  vast  majority  of  the  populaEon  

ProtecEng  research  parEcipants  

Specific  genomic  issues  •  Issues  specific  to  genomic  data  include:  –  Large  amounts  of  data  produced  –  Future  (unknown)  uses  of  that  data  –  ImplicaEons  for  family  members  –  Ability  and  expectaEon  of  public  release  

•  Main  challenges  are  respecEng  autonomy  and  protecEng  privacy  

•  Need  robust  governance  and  oversight  schemes  –  Ensure  research  integrity  –  Balance  protecEon  of  interests  of  all  involved    

•  ParEcipants,  family  members,  communiEes,  public,  research  community  

Caulfield  et  al,  PLoS  Biol,  2008  

Sample  ethical  quesEons  •  Who  has  access  to  personal  geneEc  informaEon?  •  If  geneEc  tesEng  becomes  standard  pracEce,  who  should  

have  access?  •  Should  parents  have  their  children  tested  for  adult-­‐onset  

diseases?  •  Should  genes  be  tested  for  associaEon  with  behavioral  

alributes?  –  If  the  goal  is  to  treat  “unacceptable”  behaviors,  who  defines  what  

is  “unacceptable”?  

•  Should  only  people  with  health  insurance  have  access  to  precision  medicine?  

More  sample  quesEons  •  Is  it  okay  to  use  paEent’s  data  to  develop  risk  profiling  and  

diagnosEc  tools?  Is  it  okay  to  offer  free  profiling,  in  exchange  for  allowing  use  of  that  data  in  research?  

•  Should  we  profile  paEents  for  diseases  even  if  there  is  no  known  treatment?  

•  Will  pharmaceuEcal  companies  focus  more  on  secEons  of  a  populaEon  who  are  “geneEcally  right”  for  the  drugs  they  want  to  sell?  –  3rd  world  countries  would  be  less  profitable    –  TargeEng  smaller  paEent  populaEons  may  make  the  cost  of  developing  a  drug  (~$500m)  economically  unviable  

Genomic  data:  challenges  to  consent  •  Future  unknown  uses  of  sample  (re-­‐consent?)  •  Public  release  of  data  •  Return  of  results  (including  far  into  future)  •  CommercializaEon  of  results  •  How  do  implicaEons  for  family  and  community  members  impact  individual  consent?  

•  PracEcal  limitaEons  (rapid  and  widespread  disseminaEon  of  data)  on  the  right  to  withdraw  

Right  to  withdraw  •  Central  component  of  exisEng  research  statements  –  Right  to  withdraw,  at  any  Eme,  for  any  reason  

•  MinimizaEon  of  this  right  may  impact  on  public  trust  •  DestrucEon  of  biosamples  and  wrilen  informaEon  feasible,  but  destrucEon  of  widely  disseminated  electronic  data  may  not  be  –  LimitaEons  must  be  clearly  communicated  –  RecogniEon  (and  communicaEon)  of  how  far  along  in  the  research  process  withdrawal  is  sEll  possible  

•  Genomics  projects  should  design  collecEon,  storage  and  disseminaEon  so  as  to  maximize  right  of  withdrawal  

Return  of  results  •  Right  not  to  know  is  implied  in  the  principle  of  autonomy  •  Projects  may  result  in  informaEon  ranging  from:  

–  Clinically  relevant  –  Genealogically  relevant  –  RecreaEonally  interesEng  

•  Do  researchers  have  a  moral  duty  to  return  such  results  to  parEcipants?  –  What  about  associaEons  found  by  researchers  using  that  data  for  secondary  

analyses?  Do  the  secondary  researchers  have  a  duty  to  noEfy  parEcipants,  that  they  may  never  have  had  dealings  with?    

–  What  if  the  finding  has  no  treatment?  What  if  a  treatment  is  developed  years  later?  

–  What  if  minors  are  used  in  research?  Should  acEonable  data  be  returned  to  them  or  their  parents?  

•  IniEal  consent  forms  should  specify  whether  results  (incidental  or  otherwise)  may  be  returned,  and  if  the  parEcipant  wishes  to  be  noEfied  

•  If  results  are  to  be  returned,  they  should  only  be  returned  –  If  they  are  valid  and  confirmed  (and  have  health  implicaEons?)  –  Done  so  by  professionals  who  can  clearly  explain  the  implicaEons  

Public  release  of  data  •  Strongly  encouraged  by  internaEonal  policy  

–  Provides  significant  scienEfic  research  uElity  •  But  major  implicaEons  with  respect  to  risk  to  privacy  •  Once  data  are  public,  cannot  be  withdrawn  

•  What  are  the  actual  harms  and  benefits  that  arise  from  public  access  to  data?  

•  What  are  the  implicaEons  for  family  and  community  members?  

•  What  is  the  appropriate  balance  between  public  access  and  individual  privacy  interests?  

•  Should  there  be  a  mechanism  for  compensaEon  for  harms  or  losses  associated  with  making  their  data  publicly  available?  

Restricted-­‐access  databases  •  May  be  appropriate  to  restrict  access  to  data  to  researchers  only  – AuthenEcaEon  requirement  

•  Databases  with  increasingly  sensiEve  or  idenEfying  data  (e.g.,  linking    genotypic  to  phenotypic  records,  such  as  medical  health  records)  may  benefit  from  access  restricEons  – Balance  between  uElity  and  risk  to  privacy  

•  Stringency  and  pracEcality  of  security  and  oversight  will  vary  

Re-­‐idenEficaEon  •  Anonymized  parEcipants  in  large  genomics  projects  can  be  idenEfied  by  cross-­‐referencing  their  genomes  with  genealogy  databases  [Gymrek  Science  2013]  –  Surnames  can  be  recovered  by  profiling  short  tandem  repeats  on  the  Y  chromosome  (Y-­‐STRs)  and  querying  recreaEonal  geneEc  genealogy  databases.  

•  A  systemaEc  review  of  re-­‐idenEficaEon  studies  [El  Emam  PLoS  One  2011]  showed  that  in  the  14  papers  that  they  studied,  on  average,  25%  of  people  could  be  re-­‐idenEfied  – Wide  variaEon;  most  of  the  re-­‐idenEfied  people  were  in  small-­‐scale  studies  that  did  not  adhere  to  exisEng  standards  

Public  release  implicaEons  •  IniEal  consent  form  should  clearly  explain  – Any  data  security  mechanisms  – Possibility  of  idenEficaEon  – Possibility  of  geneEc  profiling  – Loss  of  right  to  control  access  to  individual’s  data  (i.e.,  finality  of  release  process)  

– Any  issues  associated  with  family  or  community  members  •  CollecEve  consent  may  be  applicable  in  small  communiEes  

Case  study:  23andMe  2010  

•  Web-­‐based,  self-­‐reporEng  paradigm,  9126  parEcipants,  535076  SNPs  

•  Looked  at  22  traits  (none  of  them  disease-­‐related)  

•  Hair  morphology  (curl  and  color)  •  Eye  color    •  Freckles  •  Methanethiol  (asparagus)  •  PhoEc  sneezing  •  Handedness  •  Footedness  •  Ocular  dominance  •  Hand  clasp  

•  Alached  earlobes  •  AsEgmaEsm  •  Glasses  •  CaviEes  •  Braces  •  Wisdom  teeth  out  •  MoEon  sickness  •  OpEmism  •  Sweet  vs  salty  •  Morning  vs  night  

Eriksson  et  al,  PLoS  Genet  2010;  Gibson  et  al,  PLoS  Genet  2010  

23andMe  study  2010  •  Concerns:  – No  IRB  review  for  human  subjects  research  

•  Exempted  by  commercial  IRB  

– Concerns  over  the  consent  document  •  Technical  jargon  an  obstacle  to  understanding  •  Ambiguity  over  what  data  will  be  published  •  Agreement  to  parEcipate  in  research  not  separate  from  the  rest  of  the  document  

– No  access  to  raw  data  •  Contrary  to  open  access  policies  

IRB  review    •  Human  subjects  research  as  defined  by  the  Office  for  Human  Research  ProtecEons  (OHRP)  – Will  the  invesEgators  obtain  the  data  through  intervenEon  or  interacEon  with  the  parEcipant?  

– Will  the  idenEty  of  the  subjects  be  readily  ascertained  by  the  invesEgator  or  associated  with  the  data?  

•  In  both  cases,  commercial  IRB  said  “no”  

Consent  document  •  All  23andMe  parEcipants  required  to  sign  extensive  

consent  and  legal  agreement  document  –  Services  not  intended  to  diagnose  disease  or  be  used  to  provide  medical  advice  

–  Acknowledges  risks  associated  with  obtaining  unanEcipated  self-­‐knowledge  

–  Samples  will  be  used  to  further  the  field  of  geneEcs  and  human  health  

–  Before  embarking  on  collaboraEons,  further  consent  will  be  sought  

•  Consent  for  research  bundled  into  an  agreement  to  buy  a  service,  not  a  freestanding  decision  

•  One  author  clearly  from  Columbia  –  is  he  a  “collaborator”  for  whom  addiEonal  consent  should  be  sought?  

Data  access  •  23andMe  consent  form  clearly  precludes  open  disseminaEon  of  parEcipants’  geneEc  data  

•  Limited  aggregate  data  not  enough  to  idenEfy  parEcipants  

•  PLoS  GeneEcs  editors  had  to  balance:  – Public  good  of  open  access  to  research  – Public  good  of  disseminaEng  valuable  science  performed  by  a  commercial  company  

•  Decided  on  the  laler,  given  that  data  was  alainable  via  re-­‐consent  

Centralized  databases  and  biobanks  

Centralized  databases  •  Enable  the  decipherment  of  geneEc  basis  of  complex  diseases  

–  StaEsEcal  models  require  very  large  N  •  Enable  examinaEon  of  gene-­‐environment  interacEons  •  Reduced  reliance  on  parEcular  populaEons  

•  Chance  of  idenEficaEon  (even  if  data  de-­‐idenEfied)  •  Linking  of  genotype  data  to  medical  data  may  erode  some  of  the  privacy  

safeguards  •  UlEmate  lack  of  control  

–  Unreasonable  to  expect  that  in  the  long  term  you  can  control  who  has  access  to  the  data.  Somebody  at  some  point  will  make  a  copy  

–  Possibility  of  unintended  unauthorized  access  to  data  –  Push  to  make  genomes  publicly  available  

•  Uncertainty  of  future  applicaEons    

•  Does  the  individual's  right  to  privacy  outweigh  the  societal  benefits  that  would  arise  from  such  a  database?  

PopulaEon  effects  •  PopulaEon  effects  could  be  miEgated  by  use  of  large  databases  –  Not  applicable  if  using  a  homogeneous  populaEon  

•  BRCA1  discovered  by  Skolnick  using  Mormon  records  – Mormon  populaEon  considered  an  excellent  model  for  geneEc  studies  •  Homogeneous  •  Strong  founder  effect  •  Good  genealogy  records  •  Large  families  •  High  rate  of  consent  

–  Some  argue  that  the  results  of  the  study  are  highly  specific  to  that  populaEon  and  have  been  over-­‐generalized  

Biobanks  •  AmalgamaEon  of  bio-­‐samples  and  databases  •  Samples  may  come  from  clinical  seqngs,  research  projects,  judiciary  domains  

•  Oken  naEonal  resources  

•  Many  issues  similar  to  research  databases:  –  Impact  on  rights  of  individuals  –  Future  uses  –  CommercializaEon  of  knowledge  

•  Some  specific  – How  to  define  the  priority  rights  of  researchers  and  companies  over  samples  and  data  

Previous  samples  •  How  do  we  deal  with  consent  for  previously  collected  samples  (e.g.,  millions  of  Essues  stored  in  pathology  labs)?  –  Can  we  even  do  genomic  research  on  samples  that  were  not  consented  for  this  type  of  research?  

–  Possible  pracEcal  problems  with  geqng  re-­‐consent  •  Permiled  if  samples  were  anonymized,  or  research  is  ‘minimal-­‐risk’  (i.e.,  few  implicaEons  of  research  for  the  individual)  

–  RecogniEon  that  loss  of  individual’s  control  over  their  samples  might  be  okay  if  for  the  common  good?  

–  IRB  approval  necessary  –  If  it  is  possible  to  re-­‐idenEfy  a  sample,  and  a  life-­‐threatening  condiEon  is  discovered,  is  there  a  duty  to  track  down  that  person  and  tell  them?  Or  is  there  a  duty  not  to?  

PopulaEon  Biobanks  •  Iceland  HSD  

–  First  large  populaEon-­‐based  naEonal  biobank  •  deCode  GeneEcs  granted  a  private  license  to  build  the  HSD  

(Health  Sector  Database),  1998  –  AmalgamaEon  of  health  records,  genealogy  and  genotypic  database  

–  Sold  to  the  government  as  an  instrument  to  improve  health  services  in  the  country  (in  large  part  via  pharmacogenomics)  

•  Intended  to  find  correlaEons  between  genes  and  diseases;  the  findings  were  to  be  sold  to  pharmaceuEcals  and  HMOs  –  $200  million  deal  struck  with  Hoffman-­‐LaRoche  to  idenEfy  genes  associated  with  common  diseases  like  stroke,  heart  disease,  Alzheimer's  disease,  and  emphysema  

–  Icelanders  would  get  access  to  drugs  resulEng  from  this  for  free  

Concerns  with  deCode  database  •  Fears  that  personal  health  informaEon  could  be  accessed  by  

unauthorized  persons,  violaEng  individual’s  privacy  •  “Presumed  consent”  to  have  medical  records  entered  

–  Changed  to  an  opt-­‐out  scheme  –  Only  six  months  noEce  from  Eme  of  database  construcEon  to  deny  

inclusion  of  records  –  140k/260k  eventually  gave  bio  samples  (blood  and  Essue)  and  access  

to  medical  records  •  Danger  of  naEonal  geneEc  stereotyping  •  deCode  had  monopoly  control  of  data  

–  Exclusive  commercial  rights  for  12  years  –  Sole  arbiter  of  who  else  could  use  the  data  –  Exclusive  arrangement  with  Hoffman-­‐LaRoche  prevented  others  from  

studying  those  diseases  in  the  Icelandic  populaEon  •  deCode  could  make  data  available  to  pharmaceuEcal  and  insurance  

companies  for  a  price  

deCode  and  Icelandic  government  •  Much  internaEonal  opposiEon  to  the  database  act  

•  ConstrucEon  of  database  ceased  aker  Gudmundsdóqr  v.  Iceland  (2003)  –  Ruled  that    ‘the  1998  law  governing  the  creaEon  of  the  database  is  unconsEtuEonal  because  it  fails  to  protect  personal  privacy  adequately’  

•  DeCode  focused  on  gene  hunEng  instead  •  Introduced  deCodeMe  •  IPO  in  2000,  went  bankrupt  in  2009,  delisted  from  NASDAQ  – Had  never  made  money,  and  was  increasingly  indebted  to  its  creditors  

Who  owns  biodata  aker  bankruptcy?  •  EDH  set  up  as  a  subsidiary  of  deCode,  with  access  to  database  and  

biosamples,  Stefansson  as  execuEve  chair  and  head  of  research  •  Bought  by  Saga  Investments,  but  public  facing  company  sEll  called  

deCode  •  Promised  to  do  whole  genome  sequencing  of  2500  Icelandic  

genomes  –  According  to  Stefansson,  no  need  for  re-­‐consent  

•  Amgen  bought  deCode  in  December  2012,  primarily  for  use  of  its  database  –  Stopped  the  DTC  arm  of  the  company  

Conclusions  •  Much  peer-­‐reviewed  science  accomplished  

–  E.g.,  2012  Linked  age  of  father  to  incidence  of  auEsm,  discovered  gene  predisposing  to  Alzheimer’s,  and  a  gene  protecEve  against  Alzheimer’s  

•  But,  all  genomic  data  now  in  the  hands  of  a  for-­‐profit  commercial  company,  with  no  direct  benefit  to  Icelanders  

UK  BioBank  •  Will  follow  500,000  volunteers  (aged  45-­‐69  at  Eme  of  enrollment)  for  decades,  trying  to  correlate  genes,  lifestyle  and  disease    

•  Lengthy  and  rigorous  public  debate  period  •  Set  up  as  a  charity  (not  for-­‐profit)  –  Any  profits  will  be  re-­‐invested  in  the  BioBank  

•  Well-­‐structured,  clearly  wrilen  ethical  guidelines,  including  anonymizaEon  of  data  and  right  to  withdrawal;  re-­‐consent  required;  no  return  of  results  

•  Open  to  scienEfic  community  •  PosiEons  itself  as  custodian  of  the  data  •  In  the  event  of  closure:  –  “A  detailed  strategy  is  being  developed…”  

Direct-­‐to-­‐Consumer  genomics  

Direct-­‐to-­‐Consumer  geneEc  tests  •  An  enabling  tool  for  exercising  one’s  autonomous  quest  for  personal  health  informaEon  

•  Has  value  in  raising  awareness  of  geneEc  tesEng  and  screening,  as  well  as  prevenEve  medicine  

•  Concerns:  – Quality  – UElity  – Harm    – Unclear  privacy  policies  –  False  adverEsing  

DTC:  Possible  harms  •  Unique  seqng  of  DTC  is  highly  accessible,  potenEally  anonymous,  oken  free  from  involvement  of  professional  third  parEes  

•  Anonymity  –  DTC  companies  oken  have  no  effecEve  way  of  authenEcaEng  personal  idenEEes  

•  Non-­‐consensual  use  of  other  people’s  samples    –  CollecEon  of  geneEc  material  (saliva,  hair,  etc.),  leading  to  unauthorized  usage  of  personal  informaEon    

–  ConsEtutes  an  intrusion  on  basic  human  dignity  and  autonomy,  violaEon  of  source’s  bodily  integrity,  breach  of  informaEon  privacy,  deprives  the  source  of  opportunity  to  exercise  right  not  to  know  

–  UK  “DNA  thek”  offence  under  the  2004  Human  Tissue  Act  

Non-­‐consensual  DNA  thek  as  “art”  

•  Heather  Dewey-­‐Hagborg  

•  Uses  markers  for  sex,  ancestry,  eye  color,  hair  color,  freckles,  lighter  or  darker  skin,  and  certain  facial  features  like  nose  width  and  distance  between  eyes  

•  No  informaEon  on  age,  so  makes  everybody  around  25  

MtDNA  Haplogroup:  H2a2a1  (Eastern  European);  SRY  Gene:  present;  Gender:  Male;  HERC2  Gene:  AA;  Eye  Color:  Brown  

MtDNA  Haplogroup:  D1  (NaEve  American,  South  American);  SRY  Gene:  present;  Gender:  Male;  HERC2  Gene:  AA;  Eye  Color:  Brown  

GeneEc  tesEng  of  minors  •  SensiEve  issue,  protected  under  principle  1  •  Opponents:  

–  Can  cause  emoEonal  distress  (psychological  harm)  –  Adverse  effects  on  self-­‐esteem  and  self-­‐percepEon  –  Loss  of  future  adult  autonomy,  privacy  and  confidenEality  

•  Proponents:  –  Benefits  of  certainty  outweigh  the  harms  of  ambiguity  –  Learning  geneEc  informaEon  can  be  liberaEng  (can  provide  valuable  

preparaEon  acceptance  Eme)  •  Consensus  that  predicEve  tesEng  is  jusEfied  if  there  are  evident  medical  

benefits,  but  inappropriate  for  adult-­‐onset  disorders  or  carrier-­‐status  (relevant  for  reproducEve  decisions)  [not  necessarily  relevant  wrt  DTC]  

•  Other  factors  to  take  into  account:  –  Growing  emancipaEon  of  minors  –  RelaEve  maturity  of  some  minors  

•  DTC  offers  unrestrained  access  (although  all  require  the  signatory  of  the  consent  form  to  say  that  are  over  13  or  18)  and  is  not  supposed  to  be  used  for  medical  purposes  

GeneEc  counseling  •  Absence  of  adequate  geneEc  counseling  –  Consequences  of  unanEcipated  knowledge  –  Impose  professional  decision-­‐making  on  naïve  laypeople  –  Lead  to  adverse  clinical  and  personal  decisions  such  as  unnecessary  tesEng  or  harmful  reliance  on  misinterpreted  test  results  •  ProphylacEc  mastectomies,  pregnancy  terminaEons,  cessaEon  of  prescribed  medicaEon  usage  

•  Healthcare  professionals  are  not  necessarily  trained  in  geneEc  counseling  –  How  to  convey  genomic  informaEon  to  paEents  – Whether  to  withhold  informaEon  from  paEents  – Whether  to  share  informaEon  with  close  relaEves  –  How  to  allow  paEents  and  relaEves  the  right  not  to  know  

Ethics  challenges  wrt  DTC  •  Autonomy:  –  Promotes  availability  of  geneEc  tesEng  – Allows  anonymous  tesEng  (decrease  concerns  about  discriminaEon)  

–  But  fully  informed  consent  is  difficult  to  guarantee  •  No  (or  perfunctory)  geneEc  counseling  offered  •  MulEple  geneEc  tests  compound  complexity  of  pre-­‐test  counseling  

–  Tenet  of  non-­‐direcEve  decision  making  unduly  influenced  by  commercial  adverEsement  

–  Ease  of  misrepresentaEon  (sending  somebody  else’s  sample,  or  pretending  to  be  older  than  you  are)  breaches  autonomy  and  non-­‐maleficence  

DTC  ethics  •  Beneficence/non-­‐maleficence:  – Enables  beler  health  and  lifestyle  decisions  

•  Dependent  on  actual  uElity  of  DTC  tests  •  MarkeEng  may  exaggerate  medical  uElity  

– Lack  of  follow-­‐up  counseling  may  lead  to  harm  from  inappropriate  medical  or  lifestyle  decisions  

– No  strategy  in  place  to  either  inform  or  educate  ‘at  risk’  relaEves  

DTC  ethics  •  JusEce:  – Low  cost  of  geneEc  tesEng  makes  it  available  to  all  – But,  follow-­‐up  worries  may  put  addiEonal  strain  on  the  healthcare  system  and  may  not  be  available  to  the  uninsured  •  2009  survey:  78%  of  potenEal  DTC  users  said  that  they  would  consult  the  family  physician  for  help  interpreEng  the  results  •  May  lead  to  expensive  follow-­‐up  screens  (especially  due  to  pracEce  of  defensive  medicine)  

German  oversight  •  GeneEc  tesEng  only  carried  out  on  doctors’  orders  and  following  detailed  counseling  

•  Outlaws  non-­‐consensual  tesEng  by  requiring  receipt  of  explicit  wrilen  consent  both  to  geneEc  tesEng  and  sampling  

•  Seen  as  overly  restricEve  “policy  of  geneEc  excepEonalism  enforced  by  paternalisEc  regulaEon”  [Clark,  Genomics  Law  Report,  2009]  

ProfiEng  from  your  genes  

HeLa  cell  line  •  Controversial  topic:  the  original  cell  sample  cultured  to  grow  them  was  

taken  without  consent  from  a  paEent,  Henriela  Lacks  •  IdenEty  protecEon  and  consent  

–  In  the  1950s,  the  pracEce  of  taking  rouEne  Essue  samples  without  consent  from  paEents  was  not  uncommon,  and  anonymizing  the  source  wasn’t  a  primary  concern.  A  1976  McKusick  paper  on  HeLa  geneEcs  specifically  idenEfies  family  members  

–  Other  Lacks  family  members  later  asked  to  contribute  blood  for  research  projects  •  Unclear  if  their  consent  was  “informed”;  some  thought  they  were  receiving  ‘cancer  tests’  

•  Can  we  predict  future  uses  of  biospecimens  and  do  we  have  to  be  explicit  in  the  iniEal  consent  form?  –  Possible  that  Henriela  Lacks  could  have  been  told  that  her  cells  would  be  

used  for  cervical  cancer  research,  but  what  about  research  on  the  effects  of  zero  gravity  in  outer  space,  or  for  the  study  of  leukemia  or  lactose  intolerance  or  longevity  or  the  maEng  of  mosquitoes?  Would  specific  consent  have  hamstrung  future  research?  

–  Recent  court  case  brought  by  Havasupai  Indians  whose  original  consent  was  for  diabetes,  but  also  used  in  schizophrenia  and  migraEon  studies.  Offensive  due  to  cultural  ancestral  origin  stories.  Sued  and  won  

Profits  from  HeLa  cell  line  •  Should  companies  be  able  to  profit  from  cultured  cells?  

–  George  Gey,  the  doctor  who  treated  Henriela  Lacks,  didn't  profit  from  the  cells  

–  PharmaceuEcal  businesses  who  culture  HeLa  cells  do  •  Ethicists  support  a  “common  good  model”  for  Essue  donors:  

‘compensated’  by  knowing  that  research  on  their  biomaterials  may  contribute  to  science  –  Issue  of  JusEce  and  shared  benefits  –  Historically,  those  on  whom  the  research  has  been  done  are  oken  

from  the  poorest  segments  of  society  and  in  the  least  posiEon  to  benefit  

–  The  Lacks  family  are  unable  to  afford  healthcare.  Henriela  Lacks  did  receive  free  healthcare  treatment  for  her  cervical  cancer  (such  as  it  was)  

•  But  could  argue  that  the  scienEfic  accomplishment  of  creaEng  the  immortal  cell  line  is  what  is  actually  being  compensated  

Moore  v.  Regents  of  the  University  of  California,  1990  

•  Dealt  with  the  issue  of  property  rights  in  one's  own  body  parts  

•  Ruled  that  a  person's  discarded  Essue  and  cells  are  not  their  property  and  can  be  commercialized  

Johns  Hopkins  consent  form  “Johns  Hopkins  may  dispose  of  any  Essues  or  parts  that  are  removed  during  the  procedure.  Johns  Hopkins  may  retain,  preserve,  or  use  these  Essues  or  parts  for  internal  teaching  or  other  educaEonal  purposes  without  my  permission,  even  if  these  Essues  or  parts  idenEfy  me.  However,  Johns  Hopkins  may  only  use  or  disclose  Essues  or  parts  that  idenEfy  me  for  research  with  my  permission  or  with  approval  of  a  review  board  governed  by  federal  laws  protecEng  these  acEviEes.  If  the  Essues  or  parts  do  not  idenEfy  me,  Johns  Hopkins  may  use  them  for  scienEfic  (research)  purposes  without  my  permission  or  acEon  by  a  review  board.”  

Myriad  GeneEcs  – BRCA1/2  genes  specifically  patented  by  Myriad  GeneEcs  

– Screen  for  mutaEons  costs  $3000  – High-­‐profile  case:  Angelina  Jolie  

•  Public  announcement  of  bilateral  prophylacEc  mastectomy  

– Media  tended  to  emphasize  the  “naEonal  conversaEon”  without  leading  people  to  helpful  sites  like  cancer.gov  

– Missed  opportunity  to  highlight  the  absurdity  of  patenEng  genes?  

– Gene  patenEng  is  a  very  complex  debate  

Overview  of  patenEng  arguments  •  Commercial  enEEes  argue  that  they  would  not  engage  in  research  where  there  is  no  hope  of  financial  gain    

•  ScienEsts  argue  that  patenEng  of  genes  is  sEfling  their  ability  to  do  research  –  Assert  that  companies  could  patent  the  tests  or  treatments  that  arise  out  of  research  on  genes,  but  that  patenEng  the  geneEc  material  itself  is  over-­‐reaching  

•  Maler  currently  before  the  Supreme  Court,  ruling  expected  later  this  year  

•  Is  it  ethical  to  create  financial  benefits  from  free  donaEons  and  who  has  the  right  to  a  share  in  these  profits?  

Societal  issues:  geneEc  profiling  

GeneEc  tesEng  •  TesEng  for  Mendelian  traits  – E.g.,  Tay  Sachs  or  Canavan  disease  

•  TesEng  for  risk  of  disease  – E.g.,  HBOC  

•  TesEng  for  desirable  traits  – E.g.,  Intelligence  

•  TesEng  for  undesirable  traits  – E.g.,  Mental  disease  

Worst  case  scenarios  •  DiscriminaEon  [employer,  insurance]  •  Eugenics  [personal,  state]  – MiscegenaEon  

•  Pace  v  Alabama  1883;  Loving  v  Virginia  1967  •  Last  state  to  repeal:  Alabama  2001  

– Forced  sterilizaEon  – “Designer”  babies  

•  Galaca  – “Valids”  qualify  for  professional  employment  while  “invalids”  are  relegated  to  menial  jobs  

Group  discriminaEon  •  Many  geneEc  condiEons  may  be  associated  with  a  parEcular  racial  or  ethnic  group,  leading  to  discriminaEon  against  that  group  – 1970s:  state  law  mandated  geneEc  screening  of  all  African  Americans  for  sickle  cell  anemia  

GeneEc  discriminaEon  at  LBL  •  Norman-­‐Bloodsaw  v.  Lawrence  Berkeley  Laboratory  1998  –  Blood  and  urine  samples  tested,  without  knowledge  or  consent,  for  sickle  cell  trait,  syphilis,  and  pregnancy  

–  District  court  iniEally  dismissed  the  claims  •  General  consent  to  medical  exams  covered  the  tests  

–  Dictrict  court  decision  reversed  by  the  Ninth  Circuit  Court  of  Appeals  on  February  3,  1998  

–  Upheld  the  right  of  an  employee  to  be  free  from  nonconsensual  geneEc  tesEng  during  a  pre-­‐employment  medical  examinaEon  •  "[I]t  goes  without  saying  that  the  most  basic  violaEon  possible  involves  the  performance  of  unauthorized  tests  -­‐-­‐  that  is,  the  non-­‐consensual  retrieval  of  previously  unrevealed  medical  informaEon  that  may  be  unknown  even  to  plainEffs....  One  can  think  of  few  subject  areas  more  likely  to  implicate  privacy  interests  than  that  of  one's  health  or  geneEc  make-­‐up.”  

GINA  (GeneEc  InformaEon  NondiscriminaEon  Act)  2008  

•  Prohibits  employers  and  health  insurance  companies  from  discriminaEng  on  the  basis  of  geneEc  informaEon  –  Employers:  Prohibits  the  use  of  geneEc  informaEon  in  hiring,  terminaEon,  or  making  decisions  

related  to  compensaEon,  terms,  condiEons,  or  privileges  of  employment  –  Health  insurance  companies:  Prohibits  geneEc  informaEon  requests  or  requirements  before  

enrollment,  premium  increases  or  denial  of  coverage  based  on  geneEc  informaEon.  •  Difficult  to  segregate  geneEc  from  non-­‐geneEc  informaEon;  if  employers  require  

health  records,  they  may  inadvertently  end  up  geqng  geneEc  informaEon  too  [“Safe  harbor”  language]  

•  Does  not  include  other  types  of  insurance  such  as  life,  disability,  long  term  care  or  mortgage  insurance  

•  Does  not  cover  members  of  the  military  •  Insurance  protecEon  only  applies  to  as-­‐yet-­‐asymptomaEc  condiEons  

–  Target  of  healthcare  reform  (i.e.,  Affordable  Care  Act)  •  Designed  to  allay  fears  about  consequences  of  undergoing  geneEc  tesEng  •  Defines  “geneEc  informaEon”  with  respect  to  any  individual,  as    

–  The  individual’s  geneEc  tests  –  The  geneEc  tests  of  the  individual’s  family  members  –  The  manifestaEon  of  a  disease  or  disorder  in  family  members  of  the  individual  

California  SB  559  (2011)  •  Expands  protecEons  by  prohibiEng  discriminaEon  based  on  geneEc  informaEon  for:  – Housing  – Mortgage  lending  –  Public  accommodaEons  –  Emergency  medical  services  –  Licensing  exams  –  Programs  administered  or  funded  by  the  state  

•  Also  much  greater  potenEal  for  damages  (no  statutory  limit  on  compensatory  or  puniEve  damages)  

Eugenics,  really?  •  Early  geneEc  science  led  people  to  believe  that  they  could  

get  rid  of  undesirable  traits  by  sterilizaEon  –  Mental  diseases,  epilepsy,  blindness,  deafness  –  1907:  Indiana  state  sterilizaEon  law  –  1909:  California  –  30  states  ulEmately  had  some  kind  of  sterilizaEon  law  –  At  least  64000  people  (many  poor  young  women)  sterilized  

•  Most  state  laws  repealed  in  1963  or  earlier  –  Up  to  25%  of  NaEve  American  women  sterilized  in  the  1970s  –  Oregon  the  last  state  to  repeal  its  laws  in  1983  –  Virginia  sEll  allows  it  within  strict  guidelines  

•  Do  we  dare  risk  history  repeaEng  itself?    •  In  whose  hands  do  we  want  to  leave  these  decisions?  

Designer  babies?  •  At  least  2  papers  in  2012  that  can  non-­‐invasively  decipher  a  baby’s  genome  in  utero,  based  on  the  premise  that  fetal  DNA  circulates  in  the  mother’s  bloodstream  – How  many  prenatal  tests  will  deliver  “bad”  news?  – What  are  the  consequences  of  finding  something  “bad”?    

– Who  decides  what  “bad”  is?  Risk  vs.  certain  pathology  •  How  about  Tay-­‐Sachs?  Propensity  for  heart  disease?  

– What  if  the  “bad”  trait  has  no  known  treatment  or  cure?  What  if  it  does?  

– How  about  selecEng  for  “good”  traits  (e.g.,  intelligence,  height)?  

Kitzman  et  al  Sci  Transl  Med,  2012;  Fan  et  al  Nature  2012  

Human  height  variaEon  •  GWA  data  from  13665  individuals  •  39  variants  from  16482  individuals  •  20  variants  associated  with  human  height  •  Account  for  ~3%  of  height  variaEon  

•  Known  that  diet  and  environment  also  have  a  large  role  to  play  in  height  

Nature  2008  

PredicEng  intelligence  •  TesEng  for  intelligence  @BGI  (China)  •  Using  1600  super-­‐intelligent  people  with  excellent  maths  and  verbal  reasoning  skills  from  the  Study  of  MathemaEcally  Precocious  Youth  (SMPY)  

•  Raises  the  specter  of  eugenics  –  Plot  by  the  Chinese  government  to  breed  super-­‐intelligent  populaEon?  

•  Other  problems:  –  Sample  size  too  small  and  intelligence  is  too  complex  – Argue  that  if  intelligence  is  as  complex  as  say  schizophrenia,  then  they’ll  need  a  minimum  of  10000  samples  and  10000  controls  

Yong,  Nature,  2013  

Closing  remarks  •  As  scienEsts,  between  the  thrill  of  discovery,  and  the  need  to  publish,  we  can  someEmes  forget  the  people  by  whose  generosity  we  have  data  with  which  to  work  

Autonomy          Beneficence          Jus2ce  

•  Important  to  keep  these  ethical  principles  in  mind,  at  all  levels,  whether  during  the  grand  concept  and  design  phase,  or  in  the  decision  of  which  staEsEcal  test  to  use.  We  have  an  obligaEon  to  research  parEcipants  to  carry  out  science  ethically  and  with  integrity