2012 vol. 36 no. 3 AUSTRALIAN AND NEW ZEALAND JOURNAL OF PUBLIC HEALTH 281© 2012 The Authors. ANZJPH © 2012 Public Health Association of Australia

Author: The Editors have requested active titles that reflect findings. Is the amended title acceptable?

Births data are vital for understanding

demographic profiles, including

changing population size and infant

mortality rates; for evaluation and monitoring

purposes; and for informing the prioritising

and planning of maternal and infant health

services.1,2 There are two main sources of

State and Territory data: birth registrations

collected by Registrars of Births, Deaths and

Marriages3 and perinatal data collected by

midwives.4

In contrast to other jurisdictions, Northern

Territory birth registrations are consistently

higher than counts from the Territory’s

Midwives Collection (midwives data).3 Some

of the discrepancies are due to varying data

collection and reporting processes (Table

1),5,6 but differences persist even when these

are considered.3

When using routinely collected data,

it is critical to understand its quality.7 A

key consideration for birth registrations

and midwives data is the accuracy of birth

counts. Making comparisons between data

sources is an ideal way to elucidate this.8

This paper compares livebirth counts for two

remote Aboriginal communities from three

data sources to examine data accuracy. Birth

registrations, midwives data and local birth

counts collected by each community’s health

centre are used to explore the reasons for and

implications of differences found.

MethodsThe communities were the field sites for a

large National Health and Medical Research

Council-funded project.9 They are in the Top

End, similar in population size (estimated

2,200-3,000); distance from Darwin (>500

km); age profiles (young); accessibility (no

road access during the wet season) and place

of birth (majority at Royal Darwin Hospital).

For 2004-06, we purchased from the

Northern Territory Registrar of Births,

Deaths and Marriages the total number of

livebirths registered to mothers who were

usually resident in the two communities.

The Northern Territory Department of

Health provided midwives’ livebirth counts

by Aboriginal status and by place of usual

residence as derived from patient records.

Local livebirth counts were obtained by the

third author during 2007-08 from community

birth records held within each remote

community health centre. In Community 1,

infant names, date of birth and their mother’s

name were obtained. Missing data and the

Can we count? Enumerating births in two remote

Aboriginal communities in the Northern Territory

Malinda SteenkampUniversity Centre for Rural Health North Coast, School of Public Health, The University of Sydney, New South Wales

Kim JohnstoneAustralian Demographic and Social Research Institute, Australian National University, Northern Territory

Sarah Bar-ZeevUniversity Centre for Rural Health North Coast, School of Public Health, The University of Sydney, New South Wales

Abstract

Objective: To examine the accuracy of

birth counts for two remote Aboriginal

communities in the Top End of the

Northern Territory.

Methods: We compared livebirth counts

from community birth records with birth

registration numbers and perinatal counts.

Results: For 2004-06, for Community 1,

there were 204 recorded local livebirths,

190 birth registrations and 172 livebirths in

perinatal data. In Community 2, the counts

were 244, 222 and 208, respectively. The

mean annual number of babies, indicating

service requirements for babies and their

mothers, ranged from 57 to 68 (depending

on source) in Community 1, and from 69

to 81 in Community 2. Most differences

were for births to Aboriginal mothers.

Births to ‘visitors’ accounted for 16 births in

Community 1 and 30 cases in Community 2.

Conclusion: Birth registration and

perinatal data apparently underestimate

community birth counts at a local level.

Mobility of Aboriginal women seems to

partly explain this.

Implications: The differences in birth

counts have important implications for local

planning in relation to demand on housing,

health and education services. The number

of births is also a critical data requirement

for measuring infant health status,

including mortality rates, with measures

of disadvantage strongly influenced by

the number of births. Aboriginal mobility is

not a ‘data problem’, but an integral part

of Aboriginal life that needs to be catered

for in administrative data collections in the

Northern Territory.

Key words: livebirth counts, accuracy,

Aboriginal mobility, remote community

Aust NZ J Public Health. 2012; 36:281-4

doi: 10.1111/j.1753-6405.2012.00871.x

Submitted: May 2011 Revision requested: October 2011 Accepted: January 2012Correspondence to: Ms Malinda Steenkamp, University Centre for Rural Health North Coast, School of Public Health, The University of Sydney, PO Box 3074, Lismore, NSW 2480; e-mail: malinda.steenkamp@sydney.edu.au

Article Indigenous Health

282 AUSTRALIAN AND NEW ZEALAND JOURNAL OF PUBLIC HEALTH 2012 vol. 36 no. 3© 2012 The Authors. ANZJPH © 2012 Public Health Association of Australia

identity of stillborn infants were confirmed by the health centre

manager as birth outcomes were not routinely recorded in the

community records. In Community 2, the local midwife’s birth

record was used to identify the mother’s name, her infant’s date

of birth and the birth outcome. Data were cross-checked against

a record of infant names, their dates of birth and community of

residence created by the child health staff.

Staff members who had worked in each community during

2004-06 were consulted to better understand local recording and to

identify ‘local’ residents. In Community 1, interviewees were three

administration staff, two local Aboriginal Health Workers (AHWs)

and the health centre manager. In Community 2, they were the

child health nurse, two former AHWs and the receptionist. Both

communities’ midwives were consulted about recording processes,

but did not assist with infant identification.

The Department of Health Human Research Ethics Committee

and Menzies School of Health Research provided ethics approval.

ResultsFor both communities, there were differences between all three

data sources, with midwives data providing the lowest counts and

community records the highest (Table 2). Most differences were

for births to Aboriginal mothers. Staff from both communities

indicated that community records included infants who were born

to visitors of the communities. For Community 1, 16 infants were

born to non-locals; of whom eight were non-Aboriginal mothers.

In Community 2, staff identified 30 visitors. Of these, 18 were

non-Aboriginal. These mothers were typically working in the

communities, had partners working in the community, or had ‘local’

Aboriginal partners. Non-local Aboriginal mothers were those who

were visiting the community. Even so, a number of these Aboriginal

mothers were deemed to ‘belong’ to the community because of

marriage and family relationships. These families were absent

from the community at different times for varying periods, but their

infants were considered to have a connection with the community,

were recorded in the local records and used local health services.

Other infants whose parents were born in the community but lived

in Darwin or elsewhere and visited the community infrequently were

also considered to have a primary connection with the community.

DiscussionBirth registrations and midwives data are the main sources for

Northern Territory birth counts. Both apparently underestimate

community-level birth counts. For the communities investigated

Table 1: Information about Northern Territory birth registrations and midwives data.Aspect Birth registrations3 Northern Territory Midwives Collection4

Inclusion criteria All livebirths All births of ≥400 grams birthweight or ≥20 weeks gestation

Year of reporting Year of birth registration Year of birth occurrence

Aboriginal Status Mother, father and child’s Aboriginal status are recorded. Mother’s Aboriginal status is recorded.

Data collection • Birth registration is a legal requirement and prerequisite for obtaining a birth certificate.

• Northern Territory hospitals/health centres provide a ‘Notification of Birth’ to the Registrar of Births, Deaths and Marriages within 10 days of a birth.

• The parents of the infant need to register the birth within 60 days of the birth using a ‘Registration of Birth’ form. (This is also required for births in non-hospital settings.)

• The Office of the Registrar matches the two forms.

• If no birth registration form is received, notification details are used for registration.

• The collection is a population-based census of all Northern Territory births. It includes births occurring in public and private hospitals, home births, and community health centre births.

• In the public hospitals, midwives enter data shortly after the birth of a baby into the patient information system.

• Births in Darwin Private Hospital and planned home births are entered via the intranet site of the midwives collection.

• Other births are submitted in paper form and entered by the perinatal data manager.

Table 2: Livebirth counts from three sources for two remote Aboriginal communities in the Northern Territory, 2004-06.Data Source Community 1 Community 2

2004 2005 2006 Total 2004 2005 2006 TotalCommunity birth records

Births to Aboriginal mothers

Births to non-Aboriginal mothers

60

59

1

77

73

4

67

64

3

204

196

8

74

67

7

80

75

5

90

84

6

244

226

18

Birth registrations* 60 72 58 190 71 70 81 222

Northern Territory Midwives Collection (Midwives data)

Births to Aboriginal mothers

Births to non-Aboriginal mothers

47

46

1

63

58

5

62

59

3

172

163

9

68

59

9

63

59

4

77

70

7

208

188

20

Difference Birth registrations and Midwives data

Difference Community records and Birth registrations

Difference Community records and Midwives data (All mothers)

Difference Community records and Midwives data (Aboriginal mothers)

11

2

13

11

9

5

14

17

3

9

5

6

18

14

32

33

3

5

8

9

5

10

15

15

4

9

13

14

14

22

36

38* By year of registration

Steenkamp, Johnstone and Bar-Zeev Article

2012 vol. 36 no. 3 AUSTRALIAN AND NEW ZEALAND JOURNAL OF PUBLIC HEALTH 283© 2012 The Authors. ANZJPH © 2012 Public Health Association of Australia

here, birth registrations were 9% less and midwives’ counts 18%

less than community records. Most of the differences concerned

livebirths to Aboriginal mothers. Non-Aboriginal mothers were

identified as ‘non-local’ in the community records. Community

records identified some births to Aboriginal mothers as ‘non-local’,

but, they also included births to mothers who may have spent little

time in the community and who may have been recorded under a

different locality in birth registration and/or midwives data.

Community records seem to better reflect the Aboriginal

perspective of ‘community’ based on interrelationship and

belonging10 and are indicative of service population size because of

the inclusion of visitors. Birth registrations and midwives data define

‘community’ in geographical terms and assume usual residency in

one place most of the time.11 However, Aboriginal people often have

multiple residences across different locations12 and may ‘belong’ to

multiple communities.10 Mobility is a norm for many Aboriginal

women2,13-18 and is influenced by personal circumstances, cultural

obligations, ceremonial practices/duties, seasonal variations, and

access to mainstream services.13

Clearly, there are junctures where misclassification of ‘usual

place of residence’ for Aboriginal mothers in administrative data

can occur. It is Territory practice for women to birth in hospital

and to travel to Darwin, Alice Springs, Katherine or Nhulunbuy,

where birthing services are located, at about 38 weeks gestation.1

Therefore, women may be recorded as ‘usually resident’ in the

particular regional centre where they give birth.1,3,19 The smaller

difference between birth registrations and community records may

reflect parents identifying usual place of residence more ‘accurately’

when registering a birth compared to how health system data track

patient movements from a community to hospital and back again.

The issue of mobility in the Northern Territory and its impact on

administrative data is well known to Department of Health.20,21 A

significant body of work using Census data describes permanent and

temporary Aboriginal mobility and its implications for planning,

service delivery and data interpretation.2,13-18 However, there is little

published literature regarding birth count accuracy in the Territory,

and the influence and implications of Aboriginal mobility. Our study

illustrates that temporary mobility is poorly catered for in birth

registration and midwives data and that Aboriginal mobility appears

to have a significant impact on the accuracy of community-level

counts. The issues outlined here are not understood well enough to

determine the level of misclassification and how it affects midwives

and birth registration data at a regional/Territory level. Most

mobility in the Northern Territory involves movement of smaller

groups from out-stations/homelands to larger communities in the

same region12 with much temporary mobility contained within

small, specific regions that are delimited by intertwined settings

of ancestral country and family networks. The impact of mobility

may ‘even out’ at higher levels of aggregation. A data audit at Royal

Darwin Hospital found 93% agreement between patients’ health

district of residence as recorded in inpatient data and the place

people identified as their usual health district of residence when

asked in an interview.20

Although mobility may explain the differences in birth counts to

a large extent, Aboriginal women sometimes choose not to engage

with the formal health services or want to avoid the ‘displacement’

of birthing in regional centres by choosing to birth ‘on country’.

One study showed that 9% of women in Community 2 gave birth in

the community during the period for our study.22 Although there is

a process for adding these births to the midwives collection, some

births may be ‘lost’ to the official data collections even though they

are recorded in community records; which is something that needs

further investigation.

Differences in counts between the data sources imply that using

either of the two main data sources to inform local planning for

maternal and infant health services are likely to result in under-

resourcing for community health centres. In Community 1, this

would amount to 10-22 extra mothers and their infants if birth

registration or midwives data were used. For Community 2, the

apparent under-resourcing was for 14-24 patients. Although

these are small numbers, remote health workers deal with high

caseloads of Aboriginal patients who have complex health needs1,23

and come from diverse cultural backgrounds. These challenges

are compounded by the impact of social determinants of health

(overcrowding, lack of opportunities),24 unlinked patient information

systems and high remote staff turnover.1,9 Optimal maternal and

infant care is time-consuming and complex and providing care for

ten extra patients can have a large impact for remote staff.

ConclusionBirth statistics are relevant for many population-based health and

socioeconomic indicators. The two main Northern Territory birth

data collections appear to underestimate the number of livebirths

to remote-dwelling Aboriginal women at community level. Many

of these differences appear to result from Aboriginal mobility. It is

not clear whether the apparent inaccuracies in the counts reported

here are ‘evened out’ at the Territory level. This aspect will only be

clarified by further investigations. There are many current initiatives

for linking Northern Territory health data and an investigation into

the accuracy of birth counts would be a useful inclusion in this work.

This is important as we try to measure progress against ‘closing the

gap’ and respond to Australian and Northern Territory Government

initiatives to invest in ‘growth towns’.

ImplicationsDifferences in birth counts have implications for local planning

in relation to the number of children who will put demand on

housing, health and education services. Greater insight into local

health centre service populations, for these two communities, as

well as other communities, can be gained through investigations

using the Northern Territory Primary Care Information System.

The number of births is also a critical data requirement for

measuring infant health status, including mortality rates, with

measures of disadvantage strongly influenced by the number

of births. The Northern Territory midwives collection reflects

Indigenous Health Enumerating births in remote Aboriginal communities

284 AUSTRALIAN AND NEW ZEALAND JOURNAL OF PUBLIC HEALTH 2012 vol. 36 no. 3© 2012 The Authors. ANZJPH © 2012 Public Health Association of Australia

data elements appropriate to the reality of mainstream Australia,

i.e. the data elements collected are culturally embedded in a

‘Western’ paradigm.11 We illustrated that the variable ‘usual

residence’ is not culturally appropriate for capturing Aboriginal

residence. In jurisdictions with high proportions of Aboriginal

births (such as the Northern Territory), it is imperative to explore

ways that data collection can be made more culturally appropriate

to better reflect the reality of people’s mobile lives.

References1. Banscott Health Consulting. Maternity Services Review in the Northern

Territory. Darwin (AUST): Northern Territory Department of Health and Community Services; 2007.

2. Biddle N, Prout S. Indigenous Temporary Mobility: An Analysis of the 2006 Census. CAEPR Working Paper 55/2009. Canberra (AUST): Australian National University, Centre for Aboriginal Economic Policy Research; 2009.

3. Johnstone K. Indigenous Birth Rates − How reliable are they? People Place. 2009;17(4):29-39.

4. Laws P, Li Z, Sullivan E. Australia’s Mothers and Babies 2008. Perinatal Statistics Series No. 24. Canberra (AUST): Australian Institute of Health and Welfare; 2010.

5. Northern Territory Treasury. Birth Registrations in the Northern Territory. Darwin (AUST): Government of Northern Territory; 2005.

6. Tew K, Zhang X. Northern Territory Midwives’ Collection – Mothers and Babies 2006. Darwin (AUST): Northern Territory Department of Health and Families; 2010.

7. Zeni M, Kogan M. Existing population-based health databases: Useful resources for nursing research. Nurs Outlook. 2007;55:20-30.

8. Australian Institute of Health and Welfare. Aboriginal and Torres Strait Islander Health Labour Force Statistics and Data Quality Assessment. Canberra (AUST): AIHW; 2009.

9. Northern Rivers Department of Rural Health. 1+1=A Healthy Start to Life Project: Targeting the Year Before and the Year After Birth in Aboriginal Children, Research Update No. 7. Darwin (AUST): Charles Darwin University, Graduate School of Health Practice; 2009 Nov.

10. Victorian Department of Human Services. Building Better Partnerships: Working with Aboriginal Communities and Organisations: A Communication Guide for the Department of Human Services. Melbourne (AUST): State Government of Victoria; 2006.

11. Morphy F. Uncontained subjects: ‘population’ and ‘household’ in remote Aboriginal Australia. J Popul Res. 2007;24(2):163-84.

12. Bartlett B, Duncan P. Top End Aboriginal Health Planning Study. Darwin (AUST): Top End Regional Indigenous Health Planning Committee of the Northern Territory Aboriginal Health Forum; 2000 April.

13. Prout S. On the Move? Indigenous Temporary Mobility Practices in Australia. CAEPR Working Paper 48/2008. Canberra (AUST): Australian National University, Centre for Aboriginal Economic Policy Research; 2008.

14. Taylor J, Bell M. The Mobility Status of Indigenous Australians. Discussion Paper No.78/1994. Canberra (AUST): Australian National University, Centre for Aboriginal Economic Policy Research; 1994.

15. Taylor J, Bell M. The Relative Mobility Status of Indigenous Australians: Setting the Research Agenda. Discussion Paper No.77/1994. Canberra (AUST): Australian National University, Centre for Aboriginal Economic Policy Research; 1994.

16. Hunter B, Smith D. Surveying Mobile Populations: Lessons from Recent Longitudinal Surveys of Indigenous Australians. Discussion Paper No. 203/2000. Canberra (AUST): Australian National University, Centre for Aboriginal Economic Policy Research; 2000.

17. Biddle N. The Geography and Demography of Indigenous Migration: Insights for Policy and Planning. CAEPR Working Paper 58/2009. Canberra (AUST): Australian National University, Centre for Aboriginal Economic Policy Research; 2009.

18. Memmott P, Long S, Thomson L. Indigenous Mobility in Rural and Remote Australia. AHURI Final Report No. 90. St Lucia (AUST): Queensland Research Centre, Australian Housing and Urban Research Institute; 2006.

19. Kildea S. And the Women Said...Reporting on Birthing Services for Aboriginal Women from Remote Top End Communities. Darwin (AUST): Northern Territory Health Services; 1999.

20. Tew K, You J, Pircher S. Validation of Hospital Patient Demographic Data: Northern Territory Hospitals, 2008. Darwin (AUST): Northern Territory Department of Health and Families; 2008.

21. Zhao Y, Guthridge S, Li S, Connors C. Patterns of mortality in Indigenous adults in the Northern Territory, 1998-2003. Med J Aust. 2009;191(10):581-2.

22. Ireland S. Niyith Niyith Watmam The Quiet Story: Exploring the Experiences of Aboriginal Women Who Give Birth in a Remote Community in the Northern Territory. Darwin (AUST): Charles Darwin University; 2009.

23. Clucas D, Carville K, Connors C, Currie B, Carapetis J, Andrews R. Disease burden and health-care clinic attendances for young children in remote Aboriginal communities of northern Australia. Bull World Health Organ. 2008;86:275-81.

24. The Senate. Select Committee on Regional and Remote Indigenous Communities: First Report 2008. Canberra (AUST): Parliament of Australia; 2008.

Steenkamp, Johnstone and Bar-Zeev Article