An Interdisciplinary Study of the Developmental Trajectories of AtRisk Children

1
An Interdisciplinary Study of the Developmental
Trajectories of At-Risk Children Dr. Hillel
Goelman Dr. Anne Synnes Dr. Jill Houbé Dr. Anne
Klassen Mari Pighini, Ph.D Student Edudata
Forum May 5, 2006 Morris J Wosk Centre for
Dialogue, Vancouver BC
2
The Design and Implementation of An
Interdisciplinary Study of the Developmental
Trajectories of At-Risk Children Dr. Hillel
Goelman Dr. Anne Synnes Dr. Jill Houbé Dr. Anne
Klassen Mari Pighini, Ph.D Student Edudata
Forum May 5, 2006 Morris J Wosk Centre for
Dialogue, Vancouver BC
3
Definition of At-Risk in this study

• At risk children are those, who as a result of
  medical, biological or environmental factors are
  more likely than typical children to have
  developmental delays and school difficulties

4
Outline for Todays Presentation

• Background What do we know about the development
  at risk children from the research in
• Neonatology (Dr. Anne Synnes)
• Early child development (Dr. Hillel Goelman)
• Developmental pediatrics (Dr. Jill Houbé)
• Quality of life studies (Dr. Anne Klassen)
• Rationale for study (Dr. Hillel Goelman)
• Objectives and research questions of the
  exploratory study

5
Outline for Todays Presentation

• Methods Identifying children and using linked
  data bases.
• Phase 1 Identifying the children
• Phase 2 Linking the Education and Health
  databases
• Phase 3 Analyzing the developmental trajectories
  of the children
• Current status and future directions

6
Background
7
Neonatal Intensive Care UnitsThe Neonatal
Followup ProgramAnne Synnes
8

• Level II and III regional neonatal-perinatal
  centers (NICUs) in BC provide care for high risk
  pregnancies and intensive care for severely ill
  infants.
• Babies born in BC with extremely low birth weight
  (ELBW, under 750 grams), are followed by the
  provincial Neonatal Follow-up Programme (NFUP) at
  the Womens and Childrens Health Centre of BC

9
Research in neonatology

• Children treated in Neonatal Intensive Care Units
  (NICUS) are at risk for developmental problems
  because of anomalies they are born with,
  complications of treatments required in the
  newborn period, effects of their newborn
  condition on their family/ environment and / or
  concomitant risk factors

10
The Early Years Early Child DevelopmentHillel
Goelman
11
Research in Early Child Development

• The early years are critical for childrens
  physical, cognitive, language, social and
  emotional development.
• Pre-term babies admitted in the NICUs are at risk
  for neurodevelopmental, behavioral and social
  dysfunction

12

• Some studies following up at-risk babies and
  children have shown higher rates of
• learning disabilities and
• behavioural difficulties
• Early identification of at-risk children can have
  a significant positive effects upon the
  childrens development because early intervention
  services help diminish some sequelae of at-risk
  factors such as preterm/low birthweight birth

13
The Early Years Research in Developmental
Paediatrics Jill Houbé
14

• There are few population-based studies of preterm
  infant outcomes that permit interpretation of the
  impact of environmental factors/multiple
  determinants of long-term outcomes (family,
  community and health policy)
• Limited number of population based longitudinal
  research on at-risk children who are NICU
  survivors is also due to many methodological
  challenges.

15

• Neonatal Follow-Up Clinics provide
  multidisciplinary diagnostic services for former
  preterm infants
• Community intervention services provide on-going
  support for children and families
• Almost all intervention services terminate upon
  school entry
• Education policy, budgets and staffing prohibit
  on-going provision of remedial services to all
  but the most disabled school-age former preterm
  infants

16

• Quality of Life for at-risk infants, toddlers and
  preschoolers
• Anne Klassen

17
Infants born into NICUs and their caregivers a
study of the relationship between their health
related quality of life, NICU practices and
outcomes, and post-NICU health care utilization

• Investigators Shoo Lee, Anne Klassen, Parminder
  Raina, Morris Barer
• Operating Grant Hospital for Sick Children
  Foundation (Toronto)
• Postdoctoral Funding Killiam CIHR

18
Background

• There are an increasing number of generic
  multidimensional questionnaires (i.e., Health
  Related Quality of Life questionnaires) that can
  be used to measure child health comprehensively,
  but these have not been widely used in NICU
  follow-up studies.

19
Background

• Definitions of HRQL vary, but most view HRQL as
  being
• Subjective -- assessed from the patients
  perspective whenever possible
• Multidimensional -- e.g., the WHOs definition of
  health, which is a state of complete physical,
  mental, and social well-being and not merely the
  absence of disease

20
Modified WHO Definition for Children, Bradlyn et.
al 1996

• QOL includes, but is not limited to the social,
  physical and emotional functioning of the child
  and adolescent, and when indicated, his or her
  family, and it must be sensitive to the changes
  that occur throughout development.

21
Preschool Measures of HRQL

• Fekkes et al. (The Netherlands) -- TAPQOL, a 12
  domain generic pre-school measure of health
  status, and used this instrument in a study of
  preterm infants.
• Varni et al. (USA) -- Pediatric Quality of Life
  Inventory, which measures physical, mental and
  social health in children and adolescents aged 2
  to 18 years.
• Landgraf (USA) -- Infant and Toddler Quality of
  Life Questionnaire, which measures 8 child and 5
  parental domains for children aged 2 months to 5.
• Saigal (Canada) -- Health Status Classification
  System Preschool Version, which measures 12
  health domains.

22
Aims

• To validate the new ITQOL questionnaire
• To measure HRQL and HS of children admitted to
  NICUs in BC
• To measure the HRQL of the informal family
  caregivers
• To link HRQL and HS data with the CNN and BCLHD
  to examine the independent and interactive
  effects of different risk factors on long-term
  health outcomes

23
Method

• Retrospective cross-sectional survey of BC-born
  cohort included in the Canadian Neonatal Network
  (CNN) Study
• CNN study included 19,507 infants admitted to 17
  NICUs from Jan 96 to Oct 97. Data were
  collected to look at variation in practice and
  outcome

24
Sample

• A total of 1140 of 2221 surviving infants
  admitted for gt24 hrs to the 3 tertiary care NICUs
  in BC (March 1996 to June 1997)
• Comparison group of 393 of 718 healthy full-term
  children recruited from 2 of these hospitals

25
Method

• A questionnaire booklet and consent letter was
  sent to the mother as child turned 3.5 years.
• Data for consenting families linked with hospital
  birth record data from the CNN database and with
  BCLHD.

26
Infant and Toddler Quality of Life
Questionnaire
27
Health Status Classification System Preschool
Version
28
Measures

• Child Behavior Checklist 1.55 (CBCL)
• Externalizing symptoms
• Internalizing symptoms
• Total problem score

29
Results

• Overall response rate was 55
• Response rate for located families was 67.1.
• 72 provided consent to link questionnaire and
  health data

30
(No Transcript)
31
Findings

• The ITQOL questionnaire demonstrated acceptable
  reliability and construct validity in a sample of
  children requiring NICU care and a sample of
  healthy children.

32
(No Transcript)
33
Findings

• NICU children had poorer reported HRQL and HS in
  a range of domains, and more behaviour problems
  on the CBCL, compared with healthy children.
• Poorer HRQL and HS were reported for infants who
  were born lt27 weeks gestation and for children
  who experienced gt 1 major morbidity during their
  NICU stay.

34
(No Transcript)
35
Findings

• Psychosocial health of NICU parents did not
  differ from parents of healthy children
• Predictor of lower psychosocial health of parents
  included
• parental age
• poorer family function
• higher caregiver strain
• More child internalizing and externalizing
  behavioural symptoms

36
Klassen A, Lee SK, Barer M, Raina P. (2005)
Linking Survey Data with a Administrative Health
Information Characteristics Associated with
Consent from a NICU Follow-up Study. Canadian
Journal of Public Health 2005 96 (2)151-154.
37

• Families of healthy children were as likely to
  provide consent as families of NICU children.
• Higher rates of consent were associated with
• being the biological parent
• not requiring survey reminders
• involvement in a parent support group
• not working full-time
• having less healthy children
• multiple births
• higher income

38
Future Publications

• Schiariti V, et al. Caregiver-Reported Health
  Outcomes of preschool Children Born at 28 to 32
  Weeks Gestation
• Houbé J, et al. Can neonatal severity of illness
  scores predict health status and health services
  use at 42 months of age?
• Houbé J, et al. Influences on health care
  utilization in a population-based sample of NICU
  graduates at 42 months of age
• Houbé J et al. Validity of Maternal Report of
  Health Care Use for Former Preterm Children at 42
  Months of Age

39
Rationale for the proposed study on the
Developmental Trajectories of At-risk Children
Hillel Goelman
40
To date, there is a dearth of

• longitudinal research from birth to school years
  
• interdisciplinary research to provide the
  multiple perspectives needed to fully understand
  the childs development across various domains of
  development
• ecologically valid research on the social
  determinants of optimal child development

41
The challenges to studying the developmental
trajectories of these children

• There is no universal system (such as Health or
  Education) for monitoring and documenting
  childrens development from the ages of 0 to 5.

42

• The data that do exist on these children in BC
  are stored in (at least) three different
  databases and in different formats
• The BC Health Linked Database (BCHLD)
• Neonatal Follow-up Programme (NFUP) at BC
  Childrens Hospital
• Edudata Canada

43
Objectives and Research Questions
44
Objectives of the study

• To better understand the health and developmental
  trajectories of at-risk children from birth to
  age 9
• To better understand the social determinants that
  impact on the developmental health trajectories
  of at-risk children from birth to age 9 and
• To conduct an interdisciplinary study of at-risk
  infants that draws on the strengths and
  complementary perspectives of psychology,
  neonatology, and developmental pediatrics and
  population health

45
Research QuestionsIn comparing a cohort of
at-risk children admitted to Level II/III NICUs
with a matched comparison group of non-risk
children

• What differences are there in the medical
  histories of the two cohorts?
• What differences are there in the educational
  progress of the two cohorts in their elementary
  school years in such areas as
• Receiving learning assistance services
• Scores on standardized outcome measures

46
Methods
47
Phase 1

• To identify at-risk infants born in BC in 1996/7
  admitted to the BC Level II/III NICUs (Cohort 1)
  and a matched comparison group (Cohort 2) of
  non-risk infants born in BC in 1996/7 through the
  BC Linked Health Data Base (BCLHD).

48
Phase 2

• To link the health information on Cohort 1 and
  Cohort 2 obtained through the BCHLD in Phase I
  with their respective educational outcome
  measures in Grade 4 through Edudata Canada, e.g.
  FSA scores.

49
Phase 3

• To identify and to link health information on
  specific sub-groups of at-risk children in BC who
  were admitted to the BC Level II/III NICUs and
  matched comparison groups.

50
(No Transcript)
51
Phase 1 Finding the Children!

• Identify at-risk children (and their PHNs) in the
  1996/7 birth cohort who were admitted to the
  three BC Level II/III NICUs. (Cohort 1)
• (Note The researchers will receive data with
  encrypted study numbers assigned for cohorts 1
  and 2 with no personal identifiers.)

52
Phase 1 Finding the Children!

• Identify a matched comparison group of non-risk
  children (and their PHNs) based on DOB, sex and
  (three-digit) home postal code (Cohort 2)
• (Note The researchers will receive data with
  encrypted study numbers assigned for cohorts 1
  and 2 with no personal identifiers.)

53
Phase 1 Finding the Children!

• Conduct descriptive analyses for the case (Cohort
  1) and comparison cohorts (Cohort 2) and
  subgroups , e.g. by gestational age and
  birth-weight groups

54
Phase 1 Finding the data on the children in the
British Columbia Health Linked Database

• There are many different files in the BCHLD with
  coded information, for example,
• MSP Registration, Premium Billing (R PB),
  Hospital Separation Files
• In this study, we are particularly interested in
  the
• Vital Statistics Clinical Data, that include
  information on pregnancy, gestational age, birth
  weight, infant being singleton/twins, and number
  of live births to mother and marital status of
  mother at the time of giving birth.

55
What data on these children are there in the
BCHLD?

• The BC Linked Health Database (BCLHD) includes
  data on medical services that have been billed to
  and paid for by provincial health insurance
• Fields include date of birth, sex, admittance to
  hospital and to specific units NICUs,
  hospitalization, length of stay, diagnosis,
  services provided, professionals involved, among
  many others

56

• From these data we can learn
• how many at-risk children were admitted to the
  Level II/III Neonatal Intensive Care Units in
  B.C.,
• the reasons for their referral, and
• the history of their medical treatments,
  intervention and assessments

57
What are some data we can find in the BCHLD files?

• What is the health history of each child?
• Doctors visits, illnesses e.g chronic ashtma?
• Hospitalizations surgery eg- cleft palate?
  (When? For how long?)
• Interventions, therapies eg, physiotherapy,
  speech and language therapy

58
Phase 2 Linking Health and Educational
Information

• In Phase 2, these childrens anonymized Personal
  Health Numbers (PHNs) will be linked with their
  Permanent Education Numbers (PENs) so as to
• Obtain and compare anonymous and confidential
  records of educational achievement and status
  (for example, standardized test scores i.e., FSA
  scores special education designation and school
  support services, i.e., the student being
  assigned with a special support worker, among
  others

59
Linking the databases Edudata Canada

• Edudata Canada is a database which includes
  information on children in the public school
  system starting when the child enters
  kindergarten, usually at age 5.
• Edudata Canada includes information on the
  childs developmental and academic progress, as
  well as demographics information, for example

60
What relevant data are there for this study in
Edudata Canada?

• The Early Development Instrument (EDI)
• A rating instrument completed by BCs
  kindergarten teachers
• A tool for measuring children's readiness (in 5
  developmental domains) to participate in and
  benefit from school activities (ECD mapping
  project)

61
What relevant data are there for this study in
Edudata Canada?

• The Student Level Data Collection (SLDC)
• Information on all BC students attending
  public/independent schools
• Includes DOB, gender, grade level, primary
  language spoken in the home, the school they are
  attending and the program and number of courses
  they are enrolled , and

62
What relevant data are there for this study in
Edudata Canada?

• The Foundation Skills Assessment (FSA)
• Standardized tests in three subject areas
  Reading, Writing and Numeracy
• Measures critical skills embedded in BCs
  curriculum (language and mathematics prescribed
  outcomes) for students in Grades 4, 7, and 10
• among other databases

63
What are some answers we can find from BCHLD
/Edudata Canada crosswalk?

• The crosswalk will enable us to
• Link neonatal medical information on a given
  child with information on that childs school
  performance on standardized assessment measures
  in Grade 4, i.e., the Foundations Skills
  Assessment or FSA.

64
Phase 2 Linking health and educational
information with socio-demographic information

• Collect socio-demographic data in the communities
  in which the children live.
• In addition, using social geography mapping
  techniques, develop a better understanding of the
  interplay between child development in the
  multiple contexts of family, neighborhood,
  community and government health and education
  policy, e.g., the ECD Mapping Project

65
Proportion of Students Vulnerable on the Language
and Cognitive Development Subscale of the Early
Development Instrument, 2000-2003
66
Phase 2 Initial descriptive data analyses

• Descriptive analyses for the matched case (Cohort
  1) and comparison cohorts (Cohort 2) with linked
  PHNs and PENs to obtain frequency counts for the
  children who can now be followed up in the school
  system (Kindergarten through Grade 4)

67
Phase 2 Inferential and predictive data analyses

• Inferential and predictive statistical techniques
  will be used to identify the relative weight
  contributed by different health and developmental
  variables to the school outcome variables from
  Edudata Canada

68
Phase 3 Studying Specific sub-groups in the NICU
Cohort

• To identify and to link health, developmental and
  educational information on children with
  extremely low birthweight (e. g. followed by the
  Neonatal Follow Up Programme), and on children
  with identified special needs (e. g. followed up
  by IDP, AIDP and SCDP) and other specific
  subgroups of children
• To identify and to link health, developmental and
  educational information on at-risk infants from
  1997 to the present in BC, admitted to the BC
  Level II/III NICUs with their educational
  performance and outcomes in Grade 4

69
Current status and future directions
70
Population Health and Learning Observatory

• One-stop place for data crosswalk
• Advantage Researchers do not have to depend on
  probabilistic linkage of data (see PHLO chart).
• Crosswalk accuracy 99

71
PHLO Data Linkage between BCHLD and Edudata
Canada
PHLO Crosswalk
BCLHD Health data
Edudata Education data
PHN
PEN
4
The Population Health Learning Observatory
facilitates applying for data jointly BC
Ministry of Education ( through Edudata Canada)
and BC Ministry of Health (through BCHLD)
72
Who is the Population Health Learning
Observatory?(K Kinar, HELP, Fall 2004)
CYDTRU
PHLO
EDUDATA
CHSPR
73
Next step What differences make a difference?

• The challenge until now, though, is that these
  programs often do not record childrens PHNs and
  so linking them to the NICU, BCLHD and Edudata
  databases is very difficult.
• Current developments The KIDS-LINK database
  Project

74
Next step What differences make a difference?

• Tracking the children The NICU Study (Flowchart)
• Listening to parents voices The IDP Case
  Studies
• Linking information from preschool programs
• In BC there are a number of outstanding early
  intervention programs.
• For example

75
The Infant Development Programs of BC Early
intervention/ Aboriginal I.D.P. Preschool Suppo
rted Child Development Program Programs Alan
Cashmore Centre BC Centre for
Ability
76
Strategies to Improve School Readiness
Trajectories (source Halfon, 2006)
Family Discord Family stressors?
Social-emotional, Physical Cognitive, Language
function
Lack of health services Family chronic health
conditions? Genetic predispositions?
Poverty Parents SES? New immigrant family?
Lower trajectory With diminished function
Following up children (and their families) from
birth to school
77
Whats in the databases?

• BC Linked Health Database (BCLHD)
• http//www.chspr.ubc.ca/Bclhd/codetables/datainbcl
  hd.htm
• edudataCanada
• http//www.edudata.educ.ubc.ca/
• Populations Health and Learning Observatory
  (PHLO)

78
References

• Goelman , H, Brynelsen, D, Pighini, MJ Kysela,
  GM (in press). The infant development programs
  early assessment and early intervention model in
  British Columbia. In M. Guralnick (Ed.), A
  Developmental Systems Approach to Early
  Intervention National and International
  Perspectives. Baltimore Paul H. Brookes
• Klassen, AF, Landgraf, JM , Lee, SK , Barer, M,
  Raina, P, Chan, H WP, Matthew, D,
• Brabyn, D. (2003) Health related quality of life
  in 3 and 4 year old children and their parents
  Preliminary findings about a new questionnaire.
  Health and Quality of Life Outcomes, 181.
  Retrieved October 26, 2004 from
  http//www.hqlo.com/content/pdf/1477-7525-1-81.pdf
  
• Synnes AR, Ling EWY, Whitfield MF, MacKinnon M,
  Lopes L, Wong G, Effer SB (1994) Perinatal
  outcomes of a large cohort of extremely low
  gestational age infant (23-28 completed weeks).
  Journal of Pediatrics, 125952-60
• Whitfield MF, Eckstein Grunau RV, Holsti L.
  (1997). Extreme premature ( 800 g)
  schoolchildren Multiple areas of hidden
  disability. Archives of Disease in Childhood
  77F85-90

79
Research Team Members

• HELP/CHILD
• Dr Hillel Goelman
• Dr Clyde Hertzman
• Barry Forer
• Keely Kinar
• Mari Pighini
• William Warburton
• British Columbia Health Centre for Women and
  Children
• Dr Anne Synnes
• Dr Jill Houbé
• Dr Anne Klassen
• Herb WP Chan
• edudataCanada
• Jennifer Lloyd
• Centre for Health Services Policies and Research
  (CHSPR)
• Denise Morettin

80
Thank You!