79
Lifecourse Epidemiology, Longitudinal Studies and Knowledge Transla:on John Lynch Professor of Public Health and Epidemiology University of Adelaide NHMRC Australia Fellow

Lifecourse*Epidemiology,*Longitudinal*Studies** and ... · Birth* ChildhoodAdolescence* Middle*Age* OldAge* Perinatal** Epidemiology* Birth** Outcomes AdultRisk*Factor** Epidemiology*

  • Upload
    others

  • View
    5

  • Download
    0

Embed Size (px)

Citation preview

Page 1: Lifecourse*Epidemiology,*Longitudinal*Studies** and ... · Birth* ChildhoodAdolescence* Middle*Age* OldAge* Perinatal** Epidemiology* Birth** Outcomes AdultRisk*Factor** Epidemiology*

Lifecourse  Epidemiology,  Longitudinal  Studies    

and    

Knowledge  Transla:on  

John  Lynch  Professor  of  Public  Health  and  Epidemiology    

University  of  Adelaide  

NHMRC  Australia  Fellow  

Page 2: Lifecourse*Epidemiology,*Longitudinal*Studies** and ... · Birth* ChildhoodAdolescence* Middle*Age* OldAge* Perinatal** Epidemiology* Birth** Outcomes AdultRisk*Factor** Epidemiology*

hHp://www.abc.net.au/tv/life/about_the_series/life_at_5.htm    

Life  at  5  

Page 3: Lifecourse*Epidemiology,*Longitudinal*Studies** and ... · Birth* ChildhoodAdolescence* Middle*Age* OldAge* Perinatal** Epidemiology* Birth** Outcomes AdultRisk*Factor** Epidemiology*

Lifecourse  Epidemiology  

Page 4: Lifecourse*Epidemiology,*Longitudinal*Studies** and ... · Birth* ChildhoodAdolescence* Middle*Age* OldAge* Perinatal** Epidemiology* Birth** Outcomes AdultRisk*Factor** Epidemiology*

Birth   Old  Age  Middle  Age  Childhood    Adolescence  

Perinatal    Epidemiology  

Birth    Outcomes  

Adult  Risk  Factor    Epidemiology  

Adult  Chronic  Diseases  

Lifecourse  Epidemiology  

Page 5: Lifecourse*Epidemiology,*Longitudinal*Studies** and ... · Birth* ChildhoodAdolescence* Middle*Age* OldAge* Perinatal** Epidemiology* Birth** Outcomes AdultRisk*Factor** Epidemiology*

The  childhood  shows  the  man,  

As  the  morning  shows  the  day.  

                                                                                               Milton  “Paradise  Lost”  (1667)  

                                                                                                                               lines  220–21  

Not  a  new  idea  

Page 6: Lifecourse*Epidemiology,*Longitudinal*Studies** and ... · Birth* ChildhoodAdolescence* Middle*Age* OldAge* Perinatal** Epidemiology* Birth** Outcomes AdultRisk*Factor** Epidemiology*

Died ~ 1920

Died ~ 1900

Page 7: Lifecourse*Epidemiology,*Longitudinal*Studies** and ... · Birth* ChildhoodAdolescence* Middle*Age* OldAge* Perinatal** Epidemiology* Birth** Outcomes AdultRisk*Factor** Epidemiology*

Widdowson  and  McCance  (1963)  showed  that  rat  pups  with  limited  nutri:on  

during  the  3  weeks  of  lacta:on  put  on  weight  more  slowly  over  their  en:re  

life:mes,  even  though  they  had  unlimited  food  post-­‐weaning.  

In  contrast,  in  another  groups  of  rats,  the  same  dura:on  of  limited  nutri:on  at  

9-­‐12  weeks  of  age  had  only  short  term  effects  on  their  weight.  

Widdowson  and  McCance.  Proc  R  Soc  Lond  B  Biol  Sci  (1963)  

Evidence  for  “cri:cal  periods”  of  “biological  programming”  

Page 8: Lifecourse*Epidemiology,*Longitudinal*Studies** and ... · Birth* ChildhoodAdolescence* Middle*Age* OldAge* Perinatal** Epidemiology* Birth** Outcomes AdultRisk*Factor** Epidemiology*

Rediscovering  Lifecourse  

Dubos.  Pediatrics  (1966)  Dubos,  et  al.  Pediatrics  (1966)  

Page 9: Lifecourse*Epidemiology,*Longitudinal*Studies** and ... · Birth* ChildhoodAdolescence* Middle*Age* OldAge* Perinatal** Epidemiology* Birth** Outcomes AdultRisk*Factor** Epidemiology*

Atherosclerosis,  most  frequently  manifested  as  coronary  heart  disease,  is  the  major  

health  concern  in  the  United  States.  This  disease  of  adulthood  has  its  origins  in  

infancy  and  childhood.  

Kannel  WB,  Dawber  TR.  Atherosclerosis  as  a  pediatric  problem.    

Journal  of  Pediatrics  1972:  80;  544-­‐554    

Page 10: Lifecourse*Epidemiology,*Longitudinal*Studies** and ... · Birth* ChildhoodAdolescence* Middle*Age* OldAge* Perinatal** Epidemiology* Birth** Outcomes AdultRisk*Factor** Epidemiology*
Page 11: Lifecourse*Epidemiology,*Longitudinal*Studies** and ... · Birth* ChildhoodAdolescence* Middle*Age* OldAge* Perinatal** Epidemiology* Birth** Outcomes AdultRisk*Factor** Epidemiology*

September  2010  

Page 12: Lifecourse*Epidemiology,*Longitudinal*Studies** and ... · Birth* ChildhoodAdolescence* Middle*Age* OldAge* Perinatal** Epidemiology* Birth** Outcomes AdultRisk*Factor** Epidemiology*

The  Barker  Hypothesis  

The  “Developmental  Origins”  Hypothesis  

hHp://www.barker.org/    

Page 13: Lifecourse*Epidemiology,*Longitudinal*Studies** and ... · Birth* ChildhoodAdolescence* Middle*Age* OldAge* Perinatal** Epidemiology* Birth** Outcomes AdultRisk*Factor** Epidemiology*
Page 14: Lifecourse*Epidemiology,*Longitudinal*Studies** and ... · Birth* ChildhoodAdolescence* Middle*Age* OldAge* Perinatal** Epidemiology* Birth** Outcomes AdultRisk*Factor** Epidemiology*

Men    Women  Birthweight  <5.5                                                            1.0                                                                  1.0  6  -­‐                                                                0.81                                                              0.87  7  -­‐                                                                0.80                                                              0.81  8  -­‐                                                                0.74                                                              0.71  9  -­‐                                                                0.55                                                              0.52  10  +                                                          0.65                                                              0.59  

 Osmond,  et  al    BMJ  (1993)  

Rate  Ra:os  for  CHD  in  Rela:on  to  Birthweight  

Weight  at  1  year  <18                                                                    1.0                                                                  1.0  19  -­‐                                                                    0.79                                                              0.59  21  -­‐                                                                    0.81                                                              0.75  23  -­‐                                                                    0.62                                                              0.52  25    -­‐                                                                  0.62                                                              0.84  27  +                                                                  0.40                                                              0.84  

Page 15: Lifecourse*Epidemiology,*Longitudinal*Studies** and ... · Birth* ChildhoodAdolescence* Middle*Age* OldAge* Perinatal** Epidemiology* Birth** Outcomes AdultRisk*Factor** Epidemiology*

Barker  et  al.  NEJM  (2005)  

Trajectories  of  Growth  Among  Children  Who  Have  Coronary  Events  as  Adults  

3  “cri:cal”  phases:  

Lower  birthweight    Sub-­‐op:mal  growth  to  age  2    Catch  up  growth  to  adolescence  

Girls  Boys  

Page 16: Lifecourse*Epidemiology,*Longitudinal*Studies** and ... · Birth* ChildhoodAdolescence* Middle*Age* OldAge* Perinatal** Epidemiology* Birth** Outcomes AdultRisk*Factor** Epidemiology*

Di  Kuh  and  Yoav  Ben  Shlomo  (2004)  

Page 17: Lifecourse*Epidemiology,*Longitudinal*Studies** and ... · Birth* ChildhoodAdolescence* Middle*Age* OldAge* Perinatal** Epidemiology* Birth** Outcomes AdultRisk*Factor** Epidemiology*

Developmental  Plas:city  and  Human  Health  

Bateson,  Barker,  et  al.  Nature  (2004)  

•   In-­‐utero  responsiveness  to  their  mothers’  

condi:on   before   birth   may   generally  

prepare   individuals   so   that   they   are   best  

suited  to  the  environment  forecast  by  cues  

available  in  early  life.    

•    Rapid   improvements   in   nutri:on   and  

other   environmental   condi:ons   may   have  

damaging   effects   on   the   health   of   those  

people   whose   parents   and   grandparents  

lived  in  impoverished  condi:ons.  

•   A   fuller  understanding  of  paHerns  of  human  plas:city   in   response   to  early  nutri:on  and  

other  environmental  factors  will  have  implica:ons  for  the  administra:on  of  public  health.  

Page 18: Lifecourse*Epidemiology,*Longitudinal*Studies** and ... · Birth* ChildhoodAdolescence* Middle*Age* OldAge* Perinatal** Epidemiology* Birth** Outcomes AdultRisk*Factor** Epidemiology*

2006  

JE  Harding  

The  nutri:onal  basis  of  the  fetal  origins  of  adult  disease  

 Int.  J.  Epidemiol.  2001  30(1):  15-­‐23  

2001  

Gluckman  and  Hanson  

Page 19: Lifecourse*Epidemiology,*Longitudinal*Studies** and ... · Birth* ChildhoodAdolescence* Middle*Age* OldAge* Perinatal** Epidemiology* Birth** Outcomes AdultRisk*Factor** Epidemiology*

Sundström,  et  al.  BMJ  (2011)  

Page 20: Lifecourse*Epidemiology,*Longitudinal*Studies** and ... · Birth* ChildhoodAdolescence* Middle*Age* OldAge* Perinatal** Epidemiology* Birth** Outcomes AdultRisk*Factor** Epidemiology*

Birth  weight  and  Cogni:ve  Func:on  

Richards,  et  al.  BMJ  (2001)  

 BW  =    Cogni:ve  Func:on  @  ages  8,  11,  15,  26  

Page 21: Lifecourse*Epidemiology,*Longitudinal*Studies** and ... · Birth* ChildhoodAdolescence* Middle*Age* OldAge* Perinatal** Epidemiology* Birth** Outcomes AdultRisk*Factor** Epidemiology*

Survival  at  ages  12-­‐76  years  by  IQ  score  at  age  11  in  UK  

Whalley  &  Deary,  BMJ  2001  

Female   Male  

Page 22: Lifecourse*Epidemiology,*Longitudinal*Studies** and ... · Birth* ChildhoodAdolescence* Middle*Age* OldAge* Perinatal** Epidemiology* Birth** Outcomes AdultRisk*Factor** Epidemiology*

Early Life Conditions

Health in Adulthood

Adult Socioeconomic

Position

Adult Exposures

Biological Processes

“Health Capital”

Economic Resources D  

Educational Experiences

C  

Exposures    in-­‐utero    

and  childhood  

A   Health Behaviours

B  

Kuh,  Power,  Blane,  Bartley.  (2005)  

Child and Adolescent Health

E  

Page 23: Lifecourse*Epidemiology,*Longitudinal*Studies** and ... · Birth* ChildhoodAdolescence* Middle*Age* OldAge* Perinatal** Epidemiology* Birth** Outcomes AdultRisk*Factor** Epidemiology*

•   1  kg  difference  in  birthweight    4mm  Hg  difference  in  systolic  BP  

•   Difference  in  BW  between  smoking/non-­‐smoking  mothers  ~  200  gm  

•   Protein  supplementa:on  among  under-­‐nourished  women    50  gm  increase  BW  

•   More  realis:c  change  is  100  gm  BW  improvement    0.4  mmHg    

Page 24: Lifecourse*Epidemiology,*Longitudinal*Studies** and ... · Birth* ChildhoodAdolescence* Middle*Age* OldAge* Perinatal** Epidemiology* Birth** Outcomes AdultRisk*Factor** Epidemiology*

“Indeed,  for  most  of  the  epidemiological  associa:ons  described  to  date,  the  

extent  to  which  interven:ons  that  are  intended  to  modify  risk  can  improve  

long-­‐term  health  is  not  yet  clear.”  

                                                                     Gillman  M.  Developmental  Origins  of  Health  and  Disease.  NEJM  (2005)    

Page 25: Lifecourse*Epidemiology,*Longitudinal*Studies** and ... · Birth* ChildhoodAdolescence* Middle*Age* OldAge* Perinatal** Epidemiology* Birth** Outcomes AdultRisk*Factor** Epidemiology*

Death  Rates  from  CHD  and  Stroke,  US,  1950-­‐2002    

Stroke  

CHD  

Rate  Per  100,000  

70%  

70%  

Page 26: Lifecourse*Epidemiology,*Longitudinal*Studies** and ... · Birth* ChildhoodAdolescence* Middle*Age* OldAge* Perinatal** Epidemiology* Birth** Outcomes AdultRisk*Factor** Epidemiology*

US  Birthweight  Distribu:on  1950-­‐2000  

Page 27: Lifecourse*Epidemiology,*Longitudinal*Studies** and ... · Birth* ChildhoodAdolescence* Middle*Age* OldAge* Perinatal** Epidemiology* Birth** Outcomes AdultRisk*Factor** Epidemiology*

Longitudinal  Studies  

Page 28: Lifecourse*Epidemiology,*Longitudinal*Studies** and ... · Birth* ChildhoodAdolescence* Middle*Age* OldAge* Perinatal** Epidemiology* Birth** Outcomes AdultRisk*Factor** Epidemiology*

1937-­‐1939  

“In   1926-­‐1927   an   experiment   with   1500   children   done   by   the  

RoweH   Ins:tute  showed  that   the  addi:on  of  milk  made  a  definite  

improvement  in  the  rate  of  growth  and  health  in  ordinary  children  

…  milk  made  good  the  dietary  deficiencies  in  their  home  diet  which  

was  worst  in  the  poorest  homes.”    

                                                                                                                                           Lord  Boyd  Orr  

Dr  John  Boyd  Orr  

Page 29: Lifecourse*Epidemiology,*Longitudinal*Studies** and ... · Birth* ChildhoodAdolescence* Middle*Age* OldAge* Perinatal** Epidemiology* Birth** Outcomes AdultRisk*Factor** Epidemiology*

Nature  March  2011  

Page 30: Lifecourse*Epidemiology,*Longitudinal*Studies** and ... · Birth* ChildhoodAdolescence* Middle*Age* OldAge* Perinatal** Epidemiology* Birth** Outcomes AdultRisk*Factor** Epidemiology*

The  first  few  decades  of  the  twen:eth  century  found  Britain  acutely  concerned  about  its  

falling  birth  rate  and  stagnant  infant  mortality  –    “how  are  we  going  to  maintain  Britain  

and  its  empire?”    

A  Popula:on  Inves:ga:on  CommiHee  recommended  a  maternity  survey  to  explore  

whether  the  social  and  economic  costs  of  childbearing  were  discouraging  prospec:ve  

parents.  James  Douglas  was  appointed  to  head  it.  

Mike  Wadsworth,  a  social  epidemiologist  joined  Douglas’s  team  in  1968  –  the  study  was  

just  gewng  going.  “I  thought  the  changing  paHern  of  health  of  these  people  would  be  

interes:ng  over  life,”  he  says.  

Page 31: Lifecourse*Epidemiology,*Longitudinal*Studies** and ... · Birth* ChildhoodAdolescence* Middle*Age* OldAge* Perinatal** Epidemiology* Birth** Outcomes AdultRisk*Factor** Epidemiology*

Di  Kuh  —  who  had  trained  in  economics  —  wanted  to  build  up  the  biomedical  data  

that  Wadsworth  had  been  collec:ng.  Un:l  that  :me,  all  the  examina:ons  had  been  

performed  at  the  study  members’  homes,  but  by  this  stage  the  nurses  were  

staggering  under  all  the  equipment.  

To  really  understand  the  par:cipants’  physiology  and  biology,  Kuh  argued,  the  study  

needed  to  get  them  to  a  clinic.  “People  appreciate  a  free  bone  scan,”  she  said.  By  

2008  she  had  convinced  the  MRC  to  pay  for  every  willing  cohort  member  to  visit  

one  of  a  number  of  clinics  around  the  country.  

Page 32: Lifecourse*Epidemiology,*Longitudinal*Studies** and ... · Birth* ChildhoodAdolescence* Middle*Age* OldAge* Perinatal** Epidemiology* Birth** Outcomes AdultRisk*Factor** Epidemiology*

The  UK  Birth  Cohorts  

•   1946  

•   1958  •   1970  

•   Millenium  Cohort  

•   2012  Cohort  

Perinatal  death  and  health  care  

Infant  feeding  

Educa:on  and  schooling  transi:ons  

Secular  socioeconomic  change  –  employment,  family  structure  

Equity  and  geographic  varia:on  

Trajectories  linking  early  life  and  later  health    

1959  –  PlaH  Comm.  on  Welfare  of  Children  in  Hospital  

1966  –  Finer  Comm.  On  Single  Parent  Families  

1967  –  Plowden  Comm.  on  Primary  Educa:on  

1978  –  Warnock  Comm.  On  Special  Needs  

1981  –  Black  Report  

1998  –  Acheson  Report  on  Social  Inequality  

1999  –  Moser  Comm.  On  Basic  Skills  

Page 33: Lifecourse*Epidemiology,*Longitudinal*Studies** and ... · Birth* ChildhoodAdolescence* Middle*Age* OldAge* Perinatal** Epidemiology* Birth** Outcomes AdultRisk*Factor** Epidemiology*

Original  Hypotheses  (1948)  

1.  CVD  increases  with  age.  It  occurs  earlier  and  more  frequently  in  males.  

2.  Persons  with  hypertension  develop  CVD  at  a  greater  rate  than  those  who  are  not  hypertensive.  

3.  Elevated  blood  cholesterol  level  is  associated  with  an  increased  risk  of  CVD.  

4.  Tobacco  smoking  is  associated  with  an  increased  occurrence  of  CVD.  

5.  Habitual  use  of  alcohol  is  associated  with  increased  incidence  of  CVD.  

6.  Increased  physical  ac:vity  is  associated  with  a  decrease  in  the  development  of  CVD.  

7.  An  increase  in  thyroid  func:on  is  associated  with  a  decrease  in  the  development  of  CVD.  

8.  A  high  blood  hemoglobin  are  associated  with  an  increased  rate  of  the  development  of  CVD.  

9.  An  increase  in  body  weight  predisposes  to  CVD.  

10.  There  is  an  increased  rate  of  the  development  of  CVD  in  people  with  diabetes  mellitus.  

11.  There  is  a  higher  incidence  of  CVD  in  people  with  gout.  

Dr.  Thomas  Dawber  

Page 34: Lifecourse*Epidemiology,*Longitudinal*Studies** and ... · Birth* ChildhoodAdolescence* Middle*Age* OldAge* Perinatal** Epidemiology* Birth** Outcomes AdultRisk*Factor** Epidemiology*

American  J  Public  Health.  1951;14:279-­‐286  

The   study   is   focused   on   arteriosclero:c   and   hypertensive   heart   disease  

because   these  are   the  most   important  of   the   cardiovascular  diseases  and  

the  least  is  known  about  their  epidemiology.      

As  a  working  hypothesis  it  is  assumed  that  these  diseases  do  not  each  have  

a  single  cause  (as  is  the  case  in  most  infec:ous  diseases),  but  that  they  are  

the  result  of  mul:ple  causes  which  work  slowly  within  the  individual.  

Page 35: Lifecourse*Epidemiology,*Longitudinal*Studies** and ... · Birth* ChildhoodAdolescence* Middle*Age* OldAge* Perinatal** Epidemiology* Birth** Outcomes AdultRisk*Factor** Epidemiology*

Terman  Lifecycle  Study  (1922-­‐1986)  began  by  comparing  a  group  of  children  with  high  IQ  

(n=1470)  with  groups  of  children  typical  of  the  general  popula:on,  to  discover  similari:es  

and  differences.  Research  was  con:nued  over  60  years.  

Wisconsin  Longitudinal  Study  (1957  -­‐  )  is  a  long-­‐term  study  of  a  random  sample  of  10,317  

men  and  women  who  graduated  from  Wisconsin  high  schools  in  1957.  The  WLS  provides  an  

opportunity  to  study  the  life  course,  intergenera:onal  transfers  and  rela:onships,  family  

func:oning,  physical  and  mental  health  and  well-­‐being,  and  morbidity  and  mortality  from  

late  adolescence  through  2008.    

Panel  Study  of  Income  Dynamics  (1968  -­‐  )  examines  income,  behaviour  and  demographic  

change  of  about  18,000  individuals  in  4,800  families.  Now  includes  inter-­‐genera:onal  

component.  

Longitudinal  Studies  with  Roots  in  Social  Science  

Page 36: Lifecourse*Epidemiology,*Longitudinal*Studies** and ... · Birth* ChildhoodAdolescence* Middle*Age* OldAge* Perinatal** Epidemiology* Birth** Outcomes AdultRisk*Factor** Epidemiology*

Dunedin  (1973  -­‐  )  

The  Dunedin  Mul:disciplinary  Health  and  Development  Study  involves  1037  babies  born  

in  Dunedin  between  April  1972  and  March  1973  at  the  Queen  Mary  Maternity  Hospital.    

Of  those  1037  babies,  1014  of  the  original  cohort  are  s:ll  alive  today.  

The  babies  were  first  followed  up  at  the  age  of  3,  and  then  at  5,  7,  9,  11,  13,  15,  18,  21,  26  

and  32.  Future  assessments  are  scheduled  for  age  38  (2010-­‐2012),  44  and  on  into  the  

future  as  study  members  have  their  own  families,  age,  and  re:re.  

Christchurch  (1977  -­‐  )  

The  Christchurch  Health  and  Development  Study  began  in  1977  and  followed  the  health,  

educa:on  and  life  progress  of  a  group  of  1,265  children  born  in  Christchurch.  The  data  

gathered  over  the  course  of  the  study  now  comprises  some  50  million  characters  of  

informa:on  describing  the  life  history  of  this  cohort.    

Both  these  studies  have  achieved  remarkable  levels  of  follow-­‐up  of  parLcipants  

New  Zealand  Longitudinal  Studies  

Page 37: Lifecourse*Epidemiology,*Longitudinal*Studies** and ... · Birth* ChildhoodAdolescence* Middle*Age* OldAge* Perinatal** Epidemiology* Birth** Outcomes AdultRisk*Factor** Epidemiology*

Na:onal  Longitudinal  Survey  of  Children  and  Youth  (1994-­‐2009)  

The  Na:onal  Longitudinal  Survey  of  Children  and  Youth  (NLSCY)  was  a  long-­‐term  study  

of  Canadian  children  that  followed  their  development  and  well-­‐being  from  birth  to  

early  adulthood.    

The  NLSCY  began  in  1994  with  children  aged  0-­‐11  (n=  ~16,900)  and  was  jointly  

conducted  by  Sta:s:cs  Canada  and  Human  Resources  Development  Canada  (HRDC).  

Cycle  8  was  completed  in  2009.  AHri:on  rates  are  ~  25-­‐30%.  

It  is  no  longer  ac:ve.  

Canada  

Page 38: Lifecourse*Epidemiology,*Longitudinal*Studies** and ... · Birth* ChildhoodAdolescence* Middle*Age* OldAge* Perinatal** Epidemiology* Birth** Outcomes AdultRisk*Factor** Epidemiology*

Growing  Up  in  Australia  -­‐  Longitudinal  Study  of  Australian  Children  (2003  -­‐  )  

The  study  addresses  a  range  of  research  ques:ons  about  paren:ng,  family  rela:onships,  

childhood  educa:on,  non-­‐parental  child  care  and  health.  The  study  will  further  

understanding  of  child  and  adolescent  development,  inform  social  policy  debate,  and  

will  be  used  to  iden:fy  opportuni:es  for  early  interven:on  and  preven:on  strategies  in  

policy  areas  concerning  children  and  families.  

LSAC  involves  data  collected  from  two  cohorts  every  two  years.  The  first  cohort  of  5000  

children  was  aged  0-­‐1  years  in  2003-­‐2004,  and  the  second  cohort  of  5000  children  was  

aged  4-­‐5  years  in  2003-­‐2004.  Study  informants  include  the  child  (when  of  an  appropriate  

age)  and  parents  (both  resident  and  non-­‐resident),  carers  and  teachers.  

Ini:al  response  rate  50-­‐55%  

Australia  

Page 39: Lifecourse*Epidemiology,*Longitudinal*Studies** and ... · Birth* ChildhoodAdolescence* Middle*Age* OldAge* Perinatal** Epidemiology* Birth** Outcomes AdultRisk*Factor** Epidemiology*

Na:onal  Children’s  Study  (USA)  

The  Na:onal  Children’s  Study  is  a  mul:-­‐centre  cohort  that  will  aHempt  to  recruit  in  pre-­‐

pregnancy  and  result  in  100,000  live  births.  Planning  began  in  1999.    It  will  examine  the  

effects  of  the  environment  -­‐  broadly  defined  -­‐  air,  water,  diet,  sound,  family  dynamics,  

community  and  cultural  influences,  and  gene:cs  on  the  growth,  development,  and  health  

of  children  across  the  United  States,  following  them  from  before  birth  un:l  age  21  years.  

Hundreds  of  scien:sts  and  representa:ves  from  community  groups  and  professional  

organiza:ons  have  contributed  to  the  iden:fica:on  of  key  children’s  environmental  

health  ques:ons  for  the  Study.    

The  Study  Design  Working  Group  cons:tuted  in  1999  proposed  the  development  of  core  

hypotheses  encompassing  exposures  and  child  health  outcomes  of  great  public  health  

significance  requiring  long-­‐term  follow-­‐up  and  which  cannot  be  reasonably  studied  with  

fewer  children  or  a  different  study  design.  

Page 40: Lifecourse*Epidemiology,*Longitudinal*Studies** and ... · Birth* ChildhoodAdolescence* Middle*Age* OldAge* Perinatal** Epidemiology* Birth** Outcomes AdultRisk*Factor** Epidemiology*

★  ★  

★  

★  

★  ★  

★  

★  ★  

★  

★  

★  ★  

★  

★  

★  

★  

★  

★  ★  

★  

★  

★  

★  

★  

★  

★  

★  

★  

★  

★  

★  

★  

★  

★  ★  

★  

39  Birth  Cohorts  

Page 41: Lifecourse*Epidemiology,*Longitudinal*Studies** and ... · Birth* ChildhoodAdolescence* Middle*Age* OldAge* Perinatal** Epidemiology* Birth** Outcomes AdultRisk*Factor** Epidemiology*

Growing  Up  in  New  Zealand  (2009  -­‐  )  

N  =  6822  mothers  and  4,404  partners  

Page 42: Lifecourse*Epidemiology,*Longitudinal*Studies** and ... · Birth* ChildhoodAdolescence* Middle*Age* OldAge* Perinatal** Epidemiology* Birth** Outcomes AdultRisk*Factor** Epidemiology*

“   …   Growing   Up   in   New   Zealand   is   designed   to   give   us   a   more   complete   picture   of   the  

pathways   that   lead   to  differen:al  outcomes,   and   to  provide  much  beHer  evidence   for   the  

development  of  strategies  to  reduce  inequali:es  and  to  improve  outcomes  for  all  children.”  

1.  What  determines  developmental  trajectories  across  mul:ple  levels  of  influence  (poli:cal,  social,  cultural,   intergenera:onal,   familial,  and   individual)  through  the   life  course   in  the  major  domains  of  interest?  

2.  How  are  New  Zealand  children  faring  developmentally  across  mul:ple  domains  at    discrete  points  in  the  lifecourse?  

3.  How  are  these  developmental  trajectories  and  outcomes  associated  with  ethnicity    across  the  lifecourse?  

4.  What  factors  and  trajectories,  par:cularly  across  mul:ple  levels  of  influence,  confer    resilience  and  op:mise  development  for  New  Zealand  children?  

Page 43: Lifecourse*Epidemiology,*Longitudinal*Studies** and ... · Birth* ChildhoodAdolescence* Middle*Age* OldAge* Perinatal** Epidemiology* Birth** Outcomes AdultRisk*Factor** Epidemiology*
Page 44: Lifecourse*Epidemiology,*Longitudinal*Studies** and ... · Birth* ChildhoodAdolescence* Middle*Age* OldAge* Perinatal** Epidemiology* Birth** Outcomes AdultRisk*Factor** Epidemiology*
Page 45: Lifecourse*Epidemiology,*Longitudinal*Studies** and ... · Birth* ChildhoodAdolescence* Middle*Age* OldAge* Perinatal** Epidemiology* Birth** Outcomes AdultRisk*Factor** Epidemiology*

•   1  in  3  of  our  children  is  born  to  at  least  one  parent  who  did  not  grow  up  in  New  Zealand    

•   for  1  in  10  children  the  rela:onship  between  their  parents  will  change  significantly  between  the  beginning  of  the  pregnancy  and  their  birth  

•   increasingly  our  children  are  being  born  into  varied  family  structures  including  mothers  

living  with  extended  families  (nearly  24%  overall  and  more  than  half  of  all  Pacific  families),  

with  non-­‐kin  (3%  overall  but  greater  than  10%  for  Asian  mothers)  or  without  other  adults  

(3%  overall  but  7%  for  Māori  mothers).  

•   our  families  are  highly  mobile  with  over  half  of  all  families  moving  more  than  twice  in  the  

last  five  years  

•   many  parents  in  the  most  deprived  areas  were  not  aware  of  either  Paid  Parental  Leave  or  

Working  for  Families  tax  credits.  

Page 46: Lifecourse*Epidemiology,*Longitudinal*Studies** and ... · Birth* ChildhoodAdolescence* Middle*Age* OldAge* Perinatal** Epidemiology* Birth** Outcomes AdultRisk*Factor** Epidemiology*
Page 47: Lifecourse*Epidemiology,*Longitudinal*Studies** and ... · Birth* ChildhoodAdolescence* Middle*Age* OldAge* Perinatal** Epidemiology* Birth** Outcomes AdultRisk*Factor** Epidemiology*

Shulruf  et  al  (2007)  

NZ  

Oz  

Page 48: Lifecourse*Epidemiology,*Longitudinal*Studies** and ... · Birth* ChildhoodAdolescence* Middle*Age* OldAge* Perinatal** Epidemiology* Birth** Outcomes AdultRisk*Factor** Epidemiology*

Longitudinal  Studies  –  Poten:al  and  Challenges  

Page 49: Lifecourse*Epidemiology,*Longitudinal*Studies** and ... · Birth* ChildhoodAdolescence* Middle*Age* OldAge* Perinatal** Epidemiology* Birth** Outcomes AdultRisk*Factor** Epidemiology*

Lancet  (2008)  

Page 50: Lifecourse*Epidemiology,*Longitudinal*Studies** and ... · Birth* ChildhoodAdolescence* Middle*Age* OldAge* Perinatal** Epidemiology* Birth** Outcomes AdultRisk*Factor** Epidemiology*

“The  record  of  achievement  toward  the  goals  of  these  public  policy  ini:a:ves  

requires  con:nuous  replenishing  of  the  evidence  base  from  a  new  birth  cohort  as  

the  circumstances  of  birth  and  infancy  change.”  

Longview  Report  for  the  ESRC.  The  scien:fic  case  for  a  new  birth  cohort  (2007)  

Page 51: Lifecourse*Epidemiology,*Longitudinal*Studies** and ... · Birth* ChildhoodAdolescence* Middle*Age* OldAge* Perinatal** Epidemiology* Birth** Outcomes AdultRisk*Factor** Epidemiology*

Marmot  Review  (2009)  

UK  Health  Inequali:es  

Based  on  Feinstein  (2006)  

Age  5   Age  10  

Page 52: Lifecourse*Epidemiology,*Longitudinal*Studies** and ... · Birth* ChildhoodAdolescence* Middle*Age* OldAge* Perinatal** Epidemiology* Birth** Outcomes AdultRisk*Factor** Epidemiology*

Challenges:  Policy  and  Prac:ce  Relevant  

•   Close  rela:onships  with  policy/prac:ce  partners  

•   For  researchers  -­‐  framing  policy  relevant  ques:ons  

•   For  policy  makers  -­‐  framing  scien:fically  “answerable”  ques:ons  

•   Efficacy  (RCTs)  and  effec:veness  (pragma:c)  trials  –  embedded  in  cohorts  

•   Asking  ques:ons  that  can  help  inform  prac:ce  

Page 53: Lifecourse*Epidemiology,*Longitudinal*Studies** and ... · Birth* ChildhoodAdolescence* Middle*Age* OldAge* Perinatal** Epidemiology* Birth** Outcomes AdultRisk*Factor** Epidemiology*
Page 54: Lifecourse*Epidemiology,*Longitudinal*Studies** and ... · Birth* ChildhoodAdolescence* Middle*Age* OldAge* Perinatal** Epidemiology* Birth** Outcomes AdultRisk*Factor** Epidemiology*

Young  maternal  age  and  poor  child  development  outcomes  by  age  5:  findings  on  

predicJve  validity  from  a  longitudinal  birth  cohort  

 Catherine  ChiHleborough,  Debbie  Lawlor,  John  Lynch.  Pediatrics  (in  press)  

0

10

20

30

40

50

60

70

80

Age <20 years No partner or not cohabitating

Financial difficulties

Depression Smoking Education less than O level

At least 1 of the six predictors

At least 2 of the six predictors

ALSPAC  Dev  Scale  @  18  mths  

SDQ  @  47  mths  

Std  Educa:onal  Assessment  @  60  mths  school  entry  

%  Cases  in  BoH

om  10t

h    

51%   23%  

Page 55: Lifecourse*Epidemiology,*Longitudinal*Studies** and ... · Birth* ChildhoodAdolescence* Middle*Age* OldAge* Perinatal** Epidemiology* Birth** Outcomes AdultRisk*Factor** Epidemiology*

ALSPAC  (2011)  

AHri:on  and  Missing  Data  

•   Sample  replenishment  

•   Complete  case  analysis?  

•   Mul:ple  imputa:on  

Page 56: Lifecourse*Epidemiology,*Longitudinal*Studies** and ... · Birth* ChildhoodAdolescence* Middle*Age* OldAge* Perinatal** Epidemiology* Birth** Outcomes AdultRisk*Factor** Epidemiology*

Individuals   Q1  T1   Q2  T1   Q3  T1   Q4  T1   Q5  T1   Q1  T2   Q2  T2   Q3  T2   Q4  T2   Q5  T2  

DNR  

R  T1  DNR  Ts  

DNR  T1  R  T2  

Complete  

C  T1  M  T2  

M  T1  M  T2  

1  

2  

3  

4  

5  

Complex  PaHerns  of  Missing  Data  

Page 57: Lifecourse*Epidemiology,*Longitudinal*Studies** and ... · Birth* ChildhoodAdolescence* Middle*Age* OldAge* Perinatal** Epidemiology* Birth** Outcomes AdultRisk*Factor** Epidemiology*

Aspira:onal  Measurement  and  Measurement  Error  

•   Desire  to  measure  the  dynamic  complexity  

•   Measure  some  things  poorly  

•   Measurement  error  may  be  the  largest  source  of  effect  es:mate  uncertainty  

•   InaHen:on  to  confounding  –  poorer  causal  inference  

•   Richly  phenotyped  sub-­‐groups  -­‐  ALSPAC  and  asthma  

•   Biological  data  including  gene:cs  

•   New  measurement  technologies  –  web,  IT  –  iPhone  apps,  data  linkage  

Page 58: Lifecourse*Epidemiology,*Longitudinal*Studies** and ... · Birth* ChildhoodAdolescence* Middle*Age* OldAge* Perinatal** Epidemiology* Birth** Outcomes AdultRisk*Factor** Epidemiology*

Inter-­‐disciplinarity  

Lynch  Int  J  Epidemiol  (2006)  

“It  isn’t  easy  being  inter-­‐disciplinary”  

Page 59: Lifecourse*Epidemiology,*Longitudinal*Studies** and ... · Birth* ChildhoodAdolescence* Middle*Age* OldAge* Perinatal** Epidemiology* Birth** Outcomes AdultRisk*Factor** Epidemiology*

2010  

Page 60: Lifecourse*Epidemiology,*Longitudinal*Studies** and ... · Birth* ChildhoodAdolescence* Middle*Age* OldAge* Perinatal** Epidemiology* Birth** Outcomes AdultRisk*Factor** Epidemiology*
Page 61: Lifecourse*Epidemiology,*Longitudinal*Studies** and ... · Birth* ChildhoodAdolescence* Middle*Age* OldAge* Perinatal** Epidemiology* Birth** Outcomes AdultRisk*Factor** Epidemiology*

Produc:vity  

1  per  week  

1  per  day  

Produc:vity  –  resourcing  data  collec:on  +  data  management  +  use  

Page 62: Lifecourse*Epidemiology,*Longitudinal*Studies** and ... · Birth* ChildhoodAdolescence* Middle*Age* OldAge* Perinatal** Epidemiology* Birth** Outcomes AdultRisk*Factor** Epidemiology*

Methodological  Complexity  and  Causal  Inference  

•   How  do  we  fully  exploit  the  rich  data  we  collect  over  :me?  

•   Time  dependent  confounding  

•   Complex  causal  structures  

•   Innova:ve  sta:s:cal  methods  

•   Longitudinal  studies  are  not  just  repeated      cross-­‐sec:ons  

Page 63: Lifecourse*Epidemiology,*Longitudinal*Studies** and ... · Birth* ChildhoodAdolescence* Middle*Age* OldAge* Perinatal** Epidemiology* Birth** Outcomes AdultRisk*Factor** Epidemiology*

Intact (3590)

Single

Step

Single (306)

Step (84)

Intact (3200)

Intact (2744)

Single (130)

Step (38)

Step

Single

Step

Step

Single

Single (213)

Step

Step

Single

Single

Single (175) Step (38)

Step

Single (161)

Single

Intact

Intact (2912)

Intact

Intact

Age nine Age 11/12 Age five Birth

NLSCY  (n  =  ~4000)  

76%  

Page 64: Lifecourse*Epidemiology,*Longitudinal*Studies** and ... · Birth* ChildhoodAdolescence* Middle*Age* OldAge* Perinatal** Epidemiology* Birth** Outcomes AdultRisk*Factor** Epidemiology*

Non-Intact (198)

Single

Step

Single (119)

Step (36)

Intact (43)

Intact

Single

Step

Step

Step

Single

Step

Step

Single

Single (74)

Step (18)

Step

Single

Single

Intact

Age 11/12 Age five Birth

Page 65: Lifecourse*Epidemiology,*Longitudinal*Studies** and ... · Birth* ChildhoodAdolescence* Middle*Age* OldAge* Perinatal** Epidemiology* Birth** Outcomes AdultRisk*Factor** Epidemiology*

“Would  more  sophis:cated  sta:s:cal  methods  greatly  clarify  things?  Somehow  I  

doubt  it;  however  we  should  clearly  try  and  evaluate  whether  they  would  …  

Much  of  lifecourse  epidemiology  is  focused  on  trajectories  of  the  lives  of  individuals,  

in  which  stochas:c  processes  that  we  may  never  be  able  to  measure  or  understand  

play  an  important  role.  Epidemiologists  may  be  faced  with  a  similar  ‘gloomy  prospect’  

to  the  one  iden:fied  by  Eric  Turkheimer  in  the  context  of  the  inability  of  behavioural  

gene:c  studies  to  iden:fy  important  environmental  factors  shared  between  siblings  

within  families.    

Lifecourse  trajectory-­‐influencing  events  may  o�en  be  of  chance  or  idiosyncra:c  origin,  

and  thus  not  tractable  by  current  methods.”  Davey  Smith.  Int  J  Epidemiol  (2007)  

Page 66: Lifecourse*Epidemiology,*Longitudinal*Studies** and ... · Birth* ChildhoodAdolescence* Middle*Age* OldAge* Perinatal** Epidemiology* Birth** Outcomes AdultRisk*Factor** Epidemiology*
Page 67: Lifecourse*Epidemiology,*Longitudinal*Studies** and ... · Birth* ChildhoodAdolescence* Middle*Age* OldAge* Perinatal** Epidemiology* Birth** Outcomes AdultRisk*Factor** Epidemiology*

“…   an   exhausLve   causal   inves:ga:on   of   any   concrete  

phenomenon   in   its   full   reality   is   not   only   prac:cally  

impossible    -­‐  it  is  simply  nonsense  ….  The  more  “general”,  i.e.,  

the  more   abstract   the   laws,   the   less   they   can   contribute   to  

the  causal  imputa:on  of  individual  phenomena.”    

                                                                                                                                                   Max  Weber,  1904  

Page 68: Lifecourse*Epidemiology,*Longitudinal*Studies** and ... · Birth* ChildhoodAdolescence* Middle*Age* OldAge* Perinatal** Epidemiology* Birth** Outcomes AdultRisk*Factor** Epidemiology*

Marmot  Review  (2009)  

UK  Health  Inequali:es  

Based  on  Feinstein  (2006)  

Age  5   Age  10  

•   Popula:on  average  processes  

Page 69: Lifecourse*Epidemiology,*Longitudinal*Studies** and ... · Birth* ChildhoodAdolescence* Middle*Age* OldAge* Perinatal** Epidemiology* Birth** Outcomes AdultRisk*Factor** Epidemiology*

Knowledge  Transla:on  

Page 70: Lifecourse*Epidemiology,*Longitudinal*Studies** and ... · Birth* ChildhoodAdolescence* Middle*Age* OldAge* Perinatal** Epidemiology* Birth** Outcomes AdultRisk*Factor** Epidemiology*

My  quesJon  is:      “Are  we  making  an  impact?”  

Page 71: Lifecourse*Epidemiology,*Longitudinal*Studies** and ... · Birth* ChildhoodAdolescence* Middle*Age* OldAge* Perinatal** Epidemiology* Birth** Outcomes AdultRisk*Factor** Epidemiology*

Paré  1575  

Trea:ng  wounds  with  boiling  oil  

did  more  harm  than  good.  

Ligatures  were  more  effec:ve  but  

100  years  before  it  was  widely  

accepted  

Slow  Transla:on?  

Page 72: Lifecourse*Epidemiology,*Longitudinal*Studies** and ... · Birth* ChildhoodAdolescence* Middle*Age* OldAge* Perinatal** Epidemiology* Birth** Outcomes AdultRisk*Factor** Epidemiology*

•     Scurvy  Known  since  the  early  1600s,  Lind  showed  in  1747  how  oranges  and  lemons  eliminated  scurvy  among  sailors  but  1864  before  Bri:sh  Board  of  trade  used  citrus  for  all  its  sailors  

•     Smoking  In  1950  Doll  showed  smoking  and  lung  cancer  linked  but    adver:zing  not  completely  banned  in  UK  (2002),  US  (2003),  and  Australia  (1998)  

•   Asbestos    In  1955  Doll  demonstrated  rela:onship  between  asbestos  and  lung  cancer  but  asbestos  manufacture  not  banned  in  Australia  un:l  1987  

•   Childhood  obesity  The  first  data  about  increases  in  childhood  obesity  appeared  in  the  1960s,  certainly  by  the  1980s  but  yet  liHle  systema:c  policy  response  to  quell  the  ‘obesogenic  environment’  

Page 73: Lifecourse*Epidemiology,*Longitudinal*Studies** and ... · Birth* ChildhoodAdolescence* Middle*Age* OldAge* Perinatal** Epidemiology* Birth** Outcomes AdultRisk*Factor** Epidemiology*

“lifecourse”  or    “life  course”    in  the  :tle  

Publica:ons  N=1523  

Cita:ons  N=18,217  

H  Index  =  59  

Page 74: Lifecourse*Epidemiology,*Longitudinal*Studies** and ... · Birth* ChildhoodAdolescence* Middle*Age* OldAge* Perinatal** Epidemiology* Birth** Outcomes AdultRisk*Factor** Epidemiology*

“returns  on  investment”  

Page 75: Lifecourse*Epidemiology,*Longitudinal*Studies** and ... · Birth* ChildhoodAdolescence* Middle*Age* OldAge* Perinatal** Epidemiology* Birth** Outcomes AdultRisk*Factor** Epidemiology*
Page 76: Lifecourse*Epidemiology,*Longitudinal*Studies** and ... · Birth* ChildhoodAdolescence* Middle*Age* OldAge* Perinatal** Epidemiology* Birth** Outcomes AdultRisk*Factor** Epidemiology*

Research  Produc:on  Observa:onal  selec:on  bias  confounding  

measurement  error  

Experimental  RCTs  –  not  ethical  ,  not  prac:cal,  not  externally  valid,  efficacy  focus  

Pragma:c  Trials  –  effec:veness  focus,  but  who’s  ques:on?  

Natural  experiments  

Research  Consump:on  

Policy  

Implementa:on  in  Prac:ce  

Transla:on  

The  Gap  

Co-­‐crea:on  

Informa:on  Systems  

Page 77: Lifecourse*Epidemiology,*Longitudinal*Studies** and ... · Birth* ChildhoodAdolescence* Middle*Age* OldAge* Perinatal** Epidemiology* Birth** Outcomes AdultRisk*Factor** Epidemiology*

Research  Challenges  

•   Ask  relevant  research  ques:ons  

•   Best  evidence  from  high  quality  observa:onal  studies  

•   Appropriate  designs  for  valid  /  “good  enough”  evidence  

•   Funding  –  pragma:c  trials  unlikely  to  get  perfect  score  on  NHMRC  criteria  

•   Reward  structures  –  “of  sausages  and  salami”  (Konrad  Jamrozik,  MJA)  

•   Managed  por�olios  of  ECD  research  vs  Inves:gator  driven  

Policy  and  PracJce  Challenges  

•   “Health  in  all  policies”  implica:ons  for  cross-­‐sector  integrated  collabora:ons  

•   Integrated  finance  and  management  of  inter-­‐disciplinary  teams  

•   Workforce  training  and  workforce  intransigence  

•   Openness  to  change  

•   A  real  desire  to  use  informa:on  to  inform  prac:ce  

Page 78: Lifecourse*Epidemiology,*Longitudinal*Studies** and ... · Birth* ChildhoodAdolescence* Middle*Age* OldAge* Perinatal** Epidemiology* Birth** Outcomes AdultRisk*Factor** Epidemiology*
Page 79: Lifecourse*Epidemiology,*Longitudinal*Studies** and ... · Birth* ChildhoodAdolescence* Middle*Age* OldAge* Perinatal** Epidemiology* Birth** Outcomes AdultRisk*Factor** Epidemiology*

“Life  is  to  be  understood  backwards,  but  it  is  lived  forwards”  

                                                                                                 Søren  Kierkegaard  1813-­‐1855