272 AUSTRALIAN AND NEW ZEALAND JOURNAL OF PUBLIC HEALTH 2008 vol. 32 no. 3© 2008 The Authors. Journal Compilation © 2008 Public Health Association of Australia

Submitted: May 2007 Revision requested: December 2007 Accepted: March 2008Correspondence to:Dr Nicholas Wood, Clinical Fellow, National Centre for Immunisation Research and Surveillance of Vaccine Preventable Diseases (NCIRS), Locked bag 4001, Westmead NSW 2145Fax: 02 9845 1418; e-mail: Nicholw3@chw.edu.au

Nicholas Wood, Margaret Warlow, Helen QuinnNational Centre for Immunisation Research and Surveillance of Vaccine Preventable Diseases (NCIRS)

Christine SelveyFormerly Northern Territory Department of Health and Community Services

Gary LumNorthern Territory Government Pathology Service

Peter McIntyreNational Centre for Immunisation Research and Surveillance of Vaccine Preventable Diseases (NCIRS)

John KaldorNational Centre for HIV Epidemiology and Clinical Research

Abstract

Objective: To examine how routine

hepatitis B surface antigen (HBsAg) testing

of antenatal women (as identified on the

NT Midwifes Data Collection System) can

be used to track the impact of hepatitis

B (HBV) vaccination on the prevalence

of chronic HBV infection in the Northern

Territory (NT).

Methods: Women who gave birth between

01 July 2002 and 30 June 2004 were

identified from the NT Midwives Data

Collection System (MDCS). For each

woman, the unique hospital record number

(HRN) was linked to the information

system of the NT Government pathology

service to obtain the results of serological

tests for hepatitis B. The prevalence of

HBsAg was calculated by age, Indigenous

status, and maternal country of birth.

Results: During the study period, 1061

records of women from the NT MDCS

could be linked to HBsAg results. Overall,

33 (3.1%) were positive for HBsAg, of

whom 29 were recorded as Indigenous

and the remaining four were born outside

Australia.

Conclusions: Linking data from the

NT MDCS and HBsAg results from

government pathology service is a feasible

means to monitor the impact of HBV

vaccination policy.

Implications: Routine inclusion of HBsAg

results in all state and territory midwives

data collections should be pursued.

Keywords: hepatitis B, vaccination, Abo-

riginal and Torres Strait Islander, Epidemi-

ology, midwives, midwives data collection

systems.

Aust N Z Public Health. 2008; 32:272-5

doi: 10.1111/j.1753-6405.2008.00228.x

Establishment of a surveillance system (utilising Midwifes Data Collection Systems) for monitoring the impact of hepatitis B vaccination on the population prevalence of chronic hepatitis B virus infection in Australia

Hepatitis B virus (HBV) infection

is a major global health concern

with more than 350 million people

chronically infected.1 Most of the morbidity

and mortality associated with chronic HBV

infection is due to liver cirrhosis and cancer

that occurs decades after initial infection. As

the risk of becoming chronically infected is

far higher if exposure occurs early in life, the

World Health Organization recommends that

routine vaccination of infants as an integral

part of national immunisation programs

should be given a high priority and is current

practice in many countries.2

While Australia is generally regarded as

being a country of low risk for hepatitis B, the

prevalence of chronic infection has been found

to be elevated in some Aboriginal communities,

as well as among migrants from countries with

a high level of HBV infection, people with a

history of injecting drug use and amongst

homosexual men.3 Recent estimates of the

overall number of persons with chronic

hepatitis B virus (HBV) infection in the

Australian population, based on the first

national serosurvey in 1996–1999 range from

91,500 (0.49%) to 163,500 persons (0.87%).4

Several studies have shown a high prevalence

of HBV infection in Indigenous communities

in Australia prior to the commencement of

HBV immunisation. In Brewarrina, New

South Wales, in 1985, 69% of 0–19 year olds

had been infected and there was a chronic

carriage (HBsAg positive) rate of 23%. In

particular, there was a high infection (87%)

and carriage (22%) rate among 15–19 year

olds.5 In the Northern Territory (NT) in 1989,

the overall carriage rate among Indigenous

children aged 9–17 years was 8.2%, with

higher rates in more remote rural areas.6 A

Data Article

2008 vol. 32 no. 3 AUSTRALIAN AND NEW ZEALAND JOURNAL OF PUBLIC HEALTH 273© 2008 The Authors. Journal Compilation © 2008 Public Health Association of Australia

recent report showed that HBV notification and hospitalisation rates

in Australia are at least four times higher in Aboriginal and Torres

Strait Islander people.7 People with exposure to community groups

with high prevalence of HBV infection, including the above, are

termed to be ‘at risk’.

Hepatitis B virus immunisation of ‘at risk’ infants commenced

in 1987 in Australia and was implemented at different times in

each jurisdiction. In 1985, the NT, with the highest proportion of

Aboriginal people in the country, introduced HBV screening for

all pregnant women and vaccination of newborns born to mothers

found to have chronic infection. In 1988, NT implemented hepatitis

B vaccination for all Aboriginal infants and others considered to

be at increased risk. In 1990, universal infant immunisation was

recommended and phased in over the next three years, with a

catch-up program targeting 6–16 year olds introduced in 1998.

As a result of these initiatives, there is now a cohort of children

in the NT, aged 10–19 years, who are the first to have received

infant HBV vaccination in Australia. Universal neonatal HBV

vaccination was expanded Australia wide in 2000.

To monitor the impact of the vaccination programs, it is necessary

to measure the prevalence of chronic hepatitis B infection in

a systematic way. Population-based surveys are expensive and

logistically challenging, but the routine testing of antenatal women

provides a unique opportunity to track the prevalence of infection

in a diverse cross-section of the female population.

In order to measure age-specific rates of chronic HBV infection,

with the ultimate goal of assessing the impact of the Hepatitis

B immunisation program in the NT, we carried out an analysis

of the prevalence of HBsAg among women tested at antenatal

visits, by linking records from routine obstetrics and pathology

laboratory databases.

MethodsData collection

All women giving birth in the NT are required to be registered on

the Midwives Data Collection System (MDCS). For this study, the

MDCS was used to produce a list of all women who delivered from

1 July 2002 to 30 June 2004. For each woman on the list, records

were obtained of the hospital record number (HRN), date of birth,

country of birth, location at time of birth and Indigenous status.

Laboratory records for hepatitis B testing were available from

the Northern Territory Government Pathology Service, which

conducts antenatal tests for a proportion of antenatal women in the

NT. The HRN and date of birth recorded on the MDCS were used

to identify which women had the result of a test for HBsAg stored

on the pathology service database and recorded as having taken

place during the nine month period prior to the delivery date.

Data analysisFor women recorded on the MDCS with a linked HBsAg test

result from the NT Government Pathology Service, prevalence

of infection was calculated by age group, Indigenous status and

location of delivery. The 15-40 year old population prevalence was

then calculated by weighting the age-specific prevalence estimates

by the age distribution of the 2002 NT female population. The

Australian Bureau of Statistics NT mid-year estimated resident

population 2002 and NT projected Indigenous resident population

2003 were used. Data from women aged 10–14 years (n=8) and

women over 40 years (n=13) were not included in the population

prevalence estimate due to the small number of linked records.

The location of delivery was classified as Top End (Royal

Darwin, Gove and Katherine Hospitals), Central Australia (Alice

Springs and Tennant Creek Hospitals) or other (in transit, home

birth, community health centre).

The chi-square test was used to assess the statistical significance

of associations between women’s characteristics and the prevalence

of HBsAg. A p value less than 0.05 was considered statistically

significant, and 95% confidence intervals were calculated where

appropriate. Analysis of data was performed using Epi Info.

This study was approved by the Human Research Ethics

Committee of NT Department of Health and Community Services

and Menzies School of Health Research.

ResultsOf a total of 5788 women who were recorded on the MDCS

as having given birth during the study period, 1061 were found

to have an eligible linked record of HBsAg serology. For 4131

records there was no linked record of testing at the NT Government

Pathology Service, the remaining 596 records were excluded

because testing was performed outside the time interval, no

collection time was available or they were duplicate records. Half

of the women with linked records were recorded as Indigenous

and 45% gave birth in Royal Darwin Hospital (RDH). There was

no significant age difference between women with and without

linked HBsAg results, but Indigenous women were slightly more

likely to have linked records than non-Indigenous women (50.3%

vs 46.8%; p=0.04), as were women who did not deliver in RDH

(Table 1).

Out of the 1061 women with linked results, a total of 33 (3.1%)

were found to be positive for HBsAg. Of those 33 women, 29

(88%) were recorded as Indigenous and four were born outside

Table 1: Characteristics of those women recorded on the NT Midwives Data Collection System July 02-June 04, according to whether or not their records could be linked to a test result for HBsAg.

Linked Not linked p value Age group (n) tested % (n) tested % 10–14 8 0.8 30 0.6 0.66

15–19 201 18.9 810 17.1 0.16

20–24 271 25.5 1199 25.4 0.9

25–29 274 25.8 1194 25.3 0.65

30–34 201 18.9 988 20.9 0.15

35–39 93 8.8 440 9.3 0.58

Over 40 13 1.2 66 1.4 0.66

Indigenous 534 50.3 2212 46.8 0.04Delivery in RDH 480 45.2 2491 53.0 <0.01

Data Surveillance system, hepatitus B vaccination

274 AUSTRALIAN AND NEW ZEALAND JOURNAL OF PUBLIC HEALTH 2008 vol. 32 no. 3© 2008 The Authors. Journal Compilation © 2008 Public Health Association of Australia

Australia in countries of the Asia-Pacific Region. Overall the

prevalence among Indigenous women aged 15–40 years was higher

than in non-Indigenous women (Table 2).

HBsAg rates were not significantly higher in those who gave

birth in Central Australian hospitals compared to those who gave

birth in Top End hospitals (3.3% vs 2.7%; p= 0.57).

DiscussionAntenatal data from the NT between 2002 -2004 indicate an

overall population HBsAg prevalence of 3.1% and an Indigenous

population HBsAg prevalence of 5.5% in women aged 15-40 years.

These estimates are lower than those from previous studies5,6 and

are similar to those obtained in a recent audit by Romanes,8 who

examined antenatal HBsAg results for women who gave birth

in 2003 at Royal Darwin Hospital. In this audit, of 1,515 births,

results were available for 1,407 women, 32 (2.3%) were HBsAg

positive, with a prevalence in Indigenous women 3.5 times higher

than in non-Indigenous women (4.1% vs 1.2%). These data were

obtained after individual record review and include results from

both public and private pathology laboratories. Of the 32 positive

women in the audit, only one-third had results available in the RDH

pathology database (a subset of the data source for our study) and

17/32 results were from private pathology laboratories. Importantly,

Romanes found that the standard ‘Examination of Newborn’

discharge documentation was very well completed and all infants

born to carrier mothers received appropriate and timely Hepatitis B

immunoglobulin at birth and more than 90% received a three dose

HBV vaccine schedule by nine months of age. Our study only had

access to results from the main public pathology laboratory, and

included results from women who gave birth in NT hospitals other

than RDH. The prevalence in women found in the Romanes study

are consistent with the 95% confidence limits for the estimates

of population prevalence in both Indigenous and non-Indigenous

women in our study.8 However, our results are higher than another

recent estimate of HBV prevalence obtained from additional testing

of available NT sera used in the national serosurvey, 1996-1999,9

where the population prevalence of HBsAg was 0.8% (0-1.7%).

However, this study had significant limitations in that the number

of sera tested for HBsAg (n=32) was small and individual clinical

data, including Indigenous status, were not available.

In general, antenatal HBsAg testing is performed on women

who are well and undergoing routine screening. Hence, our

results should be a valid measure of community prevalence

of HBV infection in NT women. This is supported by the age

profile of excluded data in our study not being significantly

different, although a greater proportion of results were available

for Indigenous than non Indigenous women. It is likely that for

the majority of women where no results were found in the NT

Government Pathology database, results are available within

private pathology laboratory databases. This is supported by

Romanes8 who found an overall HBV antenatal screening uptake

of 94%.

In the NT, women aged 15–19 years make up approximately

15% of the total number giving birth in 2002-04. Some of these

women may have received HBV vaccination as infants or in the

catch up campaign in 1998 that targeted 6–16 year olds. The oldest

members of this targeted group would have been 22 years of age in

2004. HBV carriage in Indigenous women is lower in 15–19 year

olds (4.8%) than 20–30 years olds (6.4-6.7%) and this may be in

part due to HBV vaccine introduction in 1990. It is expected that

over time the age specific seroprevalence rate in the younger age

groups will decrease as a result of the HBV vaccination program.

These data will serve as useful baseline prevalence data in five-year

age cohorts and allow assessment over time of changes in age-

specific rates as a measure of the progressive impact of universal

HBV vaccination program.

Population-based seroprevalence estimates are likely to be the

best indicator of the impact of HBV vaccine, but are resource

intensive and therefore not affordable in many countries. Further

HBsAg prevalence studies in Australia, as the vaccinated cohorts

age, will assist in documenting the impact of our vaccination

programs. International experience indicates that rates of HBV

seroprevalence have dropped dramatically during the 10–15

years of follow-up after the introduction of HBV vaccination

programs.10,11 One such example is Taiwan, where following

universal HBV vaccination introduction in 1984, HBsAg carriage

fell from 10% to less than 1%12 and rates of hepatocellular

carcinoma in 6–14 year olds were halved.13 Other methods used

to measure the impact of HBV vaccination programs include

surveillance for acute infections due to HBV and morbidity

and mortality related to primary hepatocellular carcinoma.14

Table 2: Hepatitis B surface antigen by age group and Indigenous status for women aged 15 to 40 years.

HBsAg Total Indigenous non-indigenous p valueAge group n tested %pos (95% CI) n tested %pos (95% CI) n testeda %pos (95% CI)

15–19 201 3.5 (1.4–7.0) 145 4.8 (2.0–9.7) 56 0 (0.0–6.4) 0.09

20–24 271 3.7 (1.8–6.7) 156 6.4 (3.1–11.5) 115 0 (0.0–2.6) 0.02

25–29 274 3.3 (1.5–6.1) 120 6.7 (2.9–12.7) 153 0.7 (0.0–3.6) 0.01

30–34 201 1.0 (0.1–3.6) 67 1.5 (0.0–8.0) 134 0.8 (0.0–4.1) 0.62

35–39 93 4.3 (1.2–10.7) 34 8.8 (1.9–23.7) 58 1.7 0.0–9.2) 0.10

Total 1040b 3.1 (1.9–4.3)a 522 5.5 (3.4–7.7)b 516c 0.8 (-0.2–1.7)b

Notes:(a) Two repeat-reactive non-confirming results were considered as not-detected(b) weighted population estimate for women aged 15 to 40 years(c) Two records were not included as Indigenous status was not stated

Wood et al. Article

2008 vol. 32 no. 3 AUSTRALIAN AND NEW ZEALAND JOURNAL OF PUBLIC HEALTH 275© 2008 The Authors. Journal Compilation © 2008 Public Health Association of Australia

Problems exist with each of these surveillance methods; acute

infections in children are often clinically silent, and changes in

hepatocellular carcinoma rates may not be seen until over 20 years

post vaccination and thus require long-term monitoring probably

through dedicated cancer registries.11

This study involved linking data from two large datasets

and required manual extraction of HBsAg results. Although

novel, the method was time consuming and results from private

pathology laboratories were not available, consequently this

method is not likely to become part of routine HBV surveillance.

Similarly, individual record review, as conducted by Romanes,

is also time consuming and labour intensive, and not sustainable

for surveillance. One potential way to implement routine

surveillance of HBsAg prevalence in women is to include results

of HBsAg testing in the Midwives Data Collection system. Data

is entered onto this system by obstetric staff of all public and

private hospitals in the NT, with comparable systems throughout

Australia. Entered data currently includes maternal antenatal

history, such as Indigenous status, rubella immunity, intrapartum

and postnatal details. Inclusion of the HBsAg result as a field

at the time of data entry would mean that the data used in this

study would be collated in a single database and would not

need to be linked. Romanes’ record review found that 94% of

women had HBV screening antenatally, only 4% had no result

available, while 100% HBsAg positive results were included in the

medical records.8 Thus obstetric hospital staff could enter results

directly into the Midwives Data Collection system. The Northern

Territory Midwives Collection has one data manager and most

midwives are trained in the use of the system, supported by a

self-directed learning package and reference guide.15 Importantly,

data quality checks are made by database managers of midwives

data collections to ensure that all fields are complete and there

are no inconsistencies in the data. In addition, procedures exist

to correct incomplete records or records with errors so validity,

which depends on accurate data entry and coding,15–17 should be

well maintained. However, information on the antenatal period

collected at the time of confinement is less reliable, as shown

in a study using this system to report on hypertensive disorders

of pregnancy.18 Obstetric data collections cannot provide HBV

prevalence data for children, non pregnant women and men,

which could only come from specific serosurveys from states and

territories where most Indigenous people reside but is hampered

by difficulties in identifying the Indigenous status of sera.

In summary, this study provides another recent estimate of the

HBV carriage rate in women in the Northern Territory, which

will serve as useful baseline data to assess the impact of universal

HBV vaccination. For other regions in Australia, particularly

in areas where past prevalence was high, consideration should

be given to including antenatal HBsAg testing as a field in the

routinely collected Midwives Data Collection system. These

additional methods will complement current methods of national

serosurveillance and monitoring of HBV infection related

morbidity and mortality.

References1. Goldstein S, Fiore A. Towards the global elimination of hepatitis B virus

transmission. J Pediatr. 2001;139:343-5.2. Department of Vaccines and Biologicals [homepage on the Internet]. Geneva

(CHE): WHO; 2001 [cited 2007, November 24]. Introduction of Hepatitis B Vaccine into Childhood Immunization Services. Management Guidelines, Including Information for Health Workers and Parents. Catalogue No.: (WHO/V&B/01.31). Available from: http://www.who.int/vaccines-documents/DocsPDF01/www613.pdf

3. Kaldor J, Plant A, Thompson S, et al. The incidence of hepatitis B infection in Australia: an epidemiological review. Med J Aust. 1996;165:322-6.

4. O’Sullivan B, Gidding H, Law, Kaldor J, Gilbert G, Dore G. Estimates of chronic hepatitis B virus infection in Australia, 2000. Aust N Z J Public Health. 2004;28:212-16.

5. Campbell D, Sargent J, Plant A. The prevalence of markers of infection with hepatitis B virus in a mixed-race Australian community. Med J Aust. 1989;150:489-92.

6. Gardner I, Wan X, Simms P, Worswick D, Burrell C, Mathews J. Hepatitis B virus markers and staff in Northern Territory schools. Med J Aust. 1992;156:638-41.

7. Menzies R, McIntyre P, Beard F. Vaccine preventable diseases and vaccination coverage in Aboriginal and Torres Strait Islander people, Australia, 1999 to 2002. Commun Dis Intel. 2004;28 Suppl 1:1-45.

8. Romanes F. Retrospective audit of immunoglobulin and vaccine uptake in infants at risk of perinatal transmission of hepatitis B virus. Northern Territory Disease Control Bulletin. 2006;13:15-20

9. Wood N, Backhouse J, Gidding H, et al. Estimates of chronic hepatitis B virus infection in the Northern Territory. Commun Dis Intell. 2005;29:289-90.

10. Williams A. Reduction in hepatitis B related burden of disease – measuring the success of universal immunisation programs. Commun Dis Intell. 2002;26:458-60.

11. Clements C, Baoping Y, Crouch A. et al. Progress in the control of hepatitis B infection in the Western Pacific Region. Vaccine. 2006;24:1975-82.

12. Chen H, Chang M, Ni Y, et al. Seroepidemiology of hepatitis B virus infection in children. Ten years of mass vaccination in Taiwan. JAMA. 1996;276:906-8.

13. Kao J, Hsu H, Chang M, et al. Universal hepatitis B vaccination and the decreased mortality from fulminant hepatitis in infants in Taiwan. J Pediatr. 2001;139:349-52.

14. Assessing the Impact of Hepatitis B Vaccination Programme. Strategies, Overview of Issues and Recommendations. Field Test Version 1. Geneva (CHE): World Health Organization; 2004

15. Stewart M, Li S. Northern Territory Midwives Collection. Mothers and Babies 2000-2002. Acute Care Information Services. Health Gains Planning. Darwin (AUST): Northern Territory Department of Health and Community Services; 2005.

16. Demand and Performance Evaluation Branch. Midwives Data Collection (MDC) Reporting and Submission Requirements. New South Wales Department of Health; 2005

17. Victorian Government Health Information [perinatal homepage on the Internet]. Melbourne (AUST): Victorian Department of Human Services; 2008 [cited 2006 Dec 14]. Victorian Perintatal Data Collection Unit: Data Quality. Available from: www.health.vic.gov.au/perinatal/vcpdu/data.htm

18. Thornton C, Makris A, Ogle R et al. Generic obstetric database systems are unreliable for reporting the hypertensive disorders of pregnancy. Aust N Z J Obstet Gynaecol. 2004;44:505-9.

Data Surveillance system, hepatitus B vaccination