Using linked birth and infant death data to describe disparities in birth outcomes across various de

1
Using linked birth and infant death data to
describe disparities in birth outcomes across
various demographic groupings in British
Columbia from 1981 through 2000

• Russell Wilkins (Statistique Canada)
• William Kierans (BC Vital Statistics Agency)
• Michael Kramer (McGill University)
• Canadian Population Society, Congress 2004
• Winnipeg, Manitoba, 2-5 June 2004

2
Data sources

• Live births 1981-2000
• Infant deaths 1981-2001
• Stillbirths 1981-2000
• Census income data for EAs
• 1981, 1986, 1991, 1996
• 2001 not available when analysed

3
Methods data preparation

• Link births and deaths (probabilistic)
• Determine First Nation status of births
• Generate 1996 EA from postal codes
• Translate from 1996 EA to
• 1981 EA, 1986 EA, 1991 EA, and 2001 DA
• Calculate family size-adjusted IPPE
• Define IPPE quintiles for each census year

4
Methods analysis

• Define events and rates of interest
• Low birth weight (LBW), prematurity, SGA
• Stillbirth, neonatal, post-neonatal mortality
• Birth weight percentiles by gestational age
• Neonatal mortality by birth weight gest age
• Compile results by groupings of interest
• Sex, plurality, parity, marital status, period
• First Nation status, community size, region
• Chinese, South Asian, European immigrants
• Neighbourhood income quintile

5
Results

• 880,283 live births
• 5,506 stillbirths (20 weeks or 500 g)
• 5,698 infant deaths (lt 1 year)
• 3,565 neonatal deaths (lt28 days)
• 2,107 post-neonatal deaths (28-364 days)
• 44,407 low birth weight (lt2500 g)
• 55,637 preterm (lt37 weeks)
• 89,423 small for gestational age (lt10 ptile)

6
Birth weight percentiles by gestational age for
the full population
7
Neonatal mortality () by birth weight and
gestational age
8
Neonatal mortality av bwtby gest-age bwt
risk group

• Risk group N1000 Neo bwt(g)
• Very low 360 0.1 3638
• Low 182 0.2 3015
• Mod high 35 0.8 2442
• High 11 3.7 1807
• Very high 3 17.2 1048
• Extremely high 1 68.0 645

9
Rates by gestational age

• Weeks Inf Neo Post Stil LBW
• 20-27 10.8 5.8 5.0 39.5 100.0
• 28-36 6.9 5.3 1.6 3.0 53.0
• 37-41 3.1 1.2 1.9 1.9 1.9
• 42-44 3.6 1.7 1.9 0.2 0.4
• Infant, neonatal and post-neonatal mortality per
  1,000 live births
• Stillbirth rate per 1,000 total births (live
  births stillbirths)
• Low birth weight rate per 100 live births

10
Rates by secular period

• Inf Neo Post Stil LBW Pre
• 1981-84 8.6 5.3 3.3 7.2 5.1 6.7
• 1985-88 8.0 4.7 3.3 6.8 4.9 5.9
• 1988-92 6.7 4.1 2.5 6.3 5.0 6.1
• 1993-96 5.3 3.5 1.8 5.5 5.1 6.3
• 1997-00 4.0 2.8 1.2 5.5 5.1 6.6
• RR 2.2 1.9 2.8 1.3 1.0 1.0
• RD 4.6 2.5 2.1 1.7 0.0 0.1
• rates per 1000 except LBW Pre per 100

11
Rates by income quintile

• Q Inf Neo Post Stil LBW Pre SGA
• 1 6.8 4.1 2.7 6.1 5.7 6.9 10.9
• 2 5.8 3.7 2.0 5.9 5.2 6.2 10.5
• 3 5.5 3.5 1.9 5.8 4.9 6.0 9.6
• 4 5.4 3.6 1.8 5.4 4.7 6.0 9.0
• 5 5.2 3.2 2.0 5.5 4.4 5.8 8.3
• RR 1.3 1.3 1.4 1.1 1.3 1.2 1.3
• RD 1.6 0.9 0.7 0.6 1.3 1.1 2.6
• rates per 1000 except LBW Pre SGA per 100

12
Rates for immigrants by origin for First
Nations vs BC total

• Inf Neo Post Stil LBW Pre
• Chinese 3.2 2.3 1.0 3.1 4.7 5.6
• South Asian 4.7 3.5 1.2 4.4 7.1 7.4
• European 4.1 3.0 1.1 4.3 4.1 5.8
• BC total 6.5 4.1 2.4 6.2 5.0 6.3
• First Nations 13.0 5.3 7.6 7.5 5.9 9.6
• rates per 1000 except LBW Pre per 100

13
Discussion

• Mortality grids have only rarely been calculated
  for large populations (firsts for Canada and so
  many sub-populations)
• Mortality grids and birth weight percentile
  charts are needed by obstetricians and
  perinatalogists, as well as by families of
  premature and/or growth-retarded or macrosomic
  infants.

14
Work ahead ...

• Articles for peer-reviewed journals
• appropriateness of single vs multiple standards
  for specific subpopulations such as immigrants
  and First Nations (physiologic/pathologic?)
• trends and differentials in birth outcomes by
  neighbourhood income (incl causes of death)
• body length and head circumference measurements
  for recent years
• urban-rural and regional differences (access?)

15
Note

• Conventional denominators were used for this
  descriptive report
• The fetuses-at-risk (KS Joseph) approach will
  be used for future analytical studies based on
  these data

16
For copies of the report

• Kierans W et al.,
• Charting birth outcome in British Columbia
• BC Vital Statistics Agency, Victoria BC
• April 2004
• www.vs.gov.bc.ca/stats/features/index.htm
• (report plus 2 appendices)

17
Authors

• Kierans W, Uh SH, Mohamed J (BCVS)
• Kramer M (McGill Epidemiol, Pediatrics)
• Foster L (BC Min Children Fam Dev)
• Liston R (UBC Obstetrics)
• Wilkins R (Statistics Canada, U Ottawa)