Chronic hepatitis B surveillance in Victoria, 1998–2008: instituting a 21st Century approach to an old disease
Communicable Disease Prevention and Control Unit, Department of Health, Victoria; National Centre for Epidemiology and Population Health, Australian National University, Australian Capital Territory; Victorian Infectious Diseases Reference Laboratory
Victorian Infectious Diseases Reference Laboratory, Melbourne; Victorian Infectious Diseases Service, The Royal Melbourne Hospital & Faculty of Medicine, Dentistry and Health Sciences, The University of Melbourne, Victoria
Correspondence to: Dr Benjamin Cowie, Victorian Infectious Diseases Reference Laboratory, 10 Wreckyn St, North Melbourne, Victoria 3051. Fax: (03) 9342 2666; e-mail: email@example.com
Objective: To describe the epidemiology and assess the quality of the Victorian chronic hepatitis B (CHB) surveillance data in light of revised recommendations for the management of people with CHB infection published by the Centers for Disease Control (CDC).
Methods: Records of confirmed non-acute hepatitis B cases notified from 1998–2008 were extracted from the Victorian Notifiable Infectious Diseases Surveillance database. Age-standardised notification rates were calculated using population data from the Australian Bureau of Statistics. Descriptive epidemiological analyses were conducted on demographic fields and notifications were assessed for identity of notifier and completeness, especially for country of birth data as prioritised by the CDC.
Results: There were 19,024 notifications of CHB from 1998–2008. The annual age-standardised notification rates ranged from 28–38 per 100,000 population, with an upward trend from 2006. Country of birth information reflected the heterogeneity of HBV prevalence worldwide. A decreasing trend of doctor notifications resulted in poor completeness of risk factor fields such as Indigenous status (37%) and country of birth (27%).
Conclusion: This analysis highlighted the burden of CHB in Victorians born in high prevalence countries; however a true estimate of this burden was limited by the high proportion of incomplete notifications.
Implications: Improving the information collected from notifications will improve surveillance for CHB and ability to target appropriate responses. We suggest a pilot of enhanced surveillance to meet revised standards from CDC.
More than 350 million people are chronically infected with hepatitis B virus (HBV) worldwide, causing one million HBV-related deaths each year.1 In Australia, the prevalence of chronic hepatitis B (CHB) has increased in recent years, predominantly related to increased immigration from highly endemic regions.2 There are an estimated 160,000 people with CHB infection in Australia, more than half of whom were born in the Asia-Pacific region.2 About 55,000 of those infected are unaware of their diagnosis.3
CHB is associated with significant morbidity and mortality. Approximately 25% of people with CHB develop cirrhosis of the liver and 5–10% develop hepatocellular carcinoma (HCC).1 Over the past two decades, HCC incidence has risen in Australia. This has been most extensively studied in NSW, where HCC incidence is rising faster than any other internal cancer.4,5
Antiviral treatment for CHB and targeted screening for HCC can significantly reduce morbidity and mortality.6,7 Access to antiviral treatment in Australia requires referral to specialist physicians, but recent estimates suggest that fewer than 2% of people with CHB are on treatment.8 A recent study of children in NSW comparing notifications of CHB with the number of children seen at specialist clinics found that 851 of 930 notified infected children were not referred for specialist care, indicative of missed opportunities for early intervention.9 A similarly concerning disparity between case detection and appropriate management was reported for women diagnosed on antenatal screening at two Sydney tertiary hospitals.10 This study found no evidence of any hepatitis service involvement in 93% of women diagnosed with CHB during pregnancy.
In September 2008 the Centers for Disease Control (CDC) published revised recommendations for the identification and public health management of people diagnosed with CHB.11 The general recommendation is to scale up all aspects of the public health response to CHB. Improving the identification and public health management of persons with CHB can prevent the serious sequelae of chronic liver disease and complement immunisation strategies to eliminate HBV transmission.11 The guidelines also call for the improvement in information collected about CHB cases to facilitate development of registries of persons with chronic HBV infection. In contrast to the current case-counting approach to surveillance, the CDC recommend surveillance registries designed to facilitate the notification, counselling and medical management of persons with CHB.11 Victoria has a passive surveillance system for CHB that relies on notifications from laboratories and doctors, however the majority of cases are notified by laboratories only. Despite the legal requirement to notify CHB in accordance with the Victorian Health (Infectious Diseases) Regulations 2001, doctors are submitting notification forms for a decreasing proportion of laboratory-notified cases. Important information only available from doctor notification forms includes clinical, epidemiological and ethnographic data. As a result of low doctor notification rates, we are unable to examine the relative contribution of key risk factors. This limits the estimation of the true drivers of CHB in Victoria and subsequently our capacity to target appropriate interventions.
In this paper we examine Victorian CHB surveillance data from 1998 to 2008 and assess the usefulness and quality of the data collected. We aim to highlight areas for achievable improvement in notifications data that could enhance CHB surveillance.
The case definition for confirmed non-acute hepatitis B (herein referred to as CHB) required the detection of hepatitis B surface antigen (HBsAg), or hepatitis B virus by nucleic acid testing, in a patient with no prior evidence of hepatitis B virus infection.12 De-identified notifications of laboratory-confirmed non-acute hepatitis B infection in Victoria from 1998 to 2008 were extracted from the Victorian Notifiable Infectious Disease Surveillance (NIDS) database.
Age-standardised notification rates were calculated using Victorian population data from the Australian Bureau of Statistics (ABS). The Australian population from the 2001 census was defined as the standard population. The Victorian data were assessed for completeness, including identity of notifiers and relevant demographic fields, including country of birth (COB) and year of arrival (YOA) in Australia. The data were extracted into Microsoft® Excel and analysed with Stata™ Version 9.
Human research ethics approval was granted by the Australian National University.
From 1998 to 2008, there were 19,024 notifications of confirmed CHB, with between 1,497 and 1,938 notifications annually. The annual age-standardised notification rates ranged from 28 to 38 per 100,000 population over the 11-year period examined, with a mean annual age-standardised notification rate of 33 per 100,000 population. A downward trend from 1998 reached the minimum notification rate of 28 per 100,000 in 2004. After a rise to 30.7 per 100,000 in 2005, the notification rates returned to 28.5 per 100,000 in 2006, but increased again to 33.5 per 100,000 in 2008 (Figure 1).
Age at notification peaked in 25–34 year olds, with 30% of notifications falling in this age range (Figure 2). The median age at notification was 35 years (interquartile range 26–45). Males accounted for 10,291 (54%) and females for 8,166 (46%) of all CHB notifications. Although most age groups showed a higher proportion of males notified, particularly between 35 and 59 years of age, females accounted for the majority of cases in the 20–24 year and 25–29 year age-groups (Figure 2).
Notification rates for CHB in metropolitan Melbourne were at least four times higher than non-metropolitan regions. Metropolitan areas contained a higher proportion of overseas-born residents compared to non-metropolitan regions.13
Risk factors for CHB
From 1998 to 2008, the number of notifications with an overseas country of birth (COB) recorded increased from 46 to 254 per year, but this represented an increase from only 2% to 13% of all CHB notifications with this field complete. Of total notifications specifying COB, South-East and North-East Asian countries represented 39% and 28% of these notifications, respectively. The most marked change in the regional representation of overseas COB notifications was from Africa. In 2008, African-born CHB cases were 36% of the overseas COB notifications compared to 1% in 1998 (Figure 3).
China and Vietnam were the most commonly recorded countries of birth on CHB notifications, followed by Sudan, Hong Kong and Burma (Table 1). The annual number of notifications with Vietnam recorded as COB has remained consistent at about 20 per year and 9–14% of notifications with overseas COB recorded. In contrast, the number of notifications with China as the COB has increased markedly from two in 1998 (4%) to 77 in 2007 (28%). Sudan and Burma also showed recent increases in cases from 2002 onwards (Figure 4). In that year, Sudan and Burma as COB were recorded on one (1%) notification each, rising to 18 (8%) and 38 (13%) in 2006, respectively.
Table 1. Top 12 countries of birth recorded on CHB notifications, 1998–2008.
Country of birth
Note: a) percentage of notifications with overseas COB recorded
Only 504 of the 1,775 (28%) notifications recording an overseas COB included the year of arrival. Years of arrival on notifications ranged from 1950 to 2008, with the majority (70%) arriving in the past eight years. The median time from arrival year to notification date was one year (interquartile range 0–5.5 years). However, when analysed according to country of birth there were differences in the mean time to notification between the 12 most commonly recorded countries of birth. Cases born in Vietnam, Cambodia and the Philippines had the longest mean time lag to notification (5–10 years), with the shortest times for cases from Burma, Sudan and Taiwan (<1.5 years) (Table 2).
Table 2. Mean years to notification by country of birth, 1998–2008.
Country of birth
Mean years (95% CI)
From 1998 to 2008, the identity of the notifier was recorded in 99% of notifications. Only 37% of CHB notified cases included information from the doctors requesting the hepatitis B tests. Furthermore, the percentage of total notified cases including information from doctors appeared to be following a decreasing trend over time from 55% in 1998 to 27% in 2008. The fields of gender, date of birth (DOB), age, postcode of residence, and Aboriginal and Torres Strait Islander (ATSI) status were used to assess data completeness. The completeness was high for basic demographic fields (gender 97%/ DOB 96%/ age 99%/ postcode 96%) but ATSI status was only 37% complete.
The completeness of COB and YOA data were assessed for the 7,048 notifications that included information from doctors. These fields were not mandatory for doctors to complete until 2008, however it has been possible to provide COB information prior to last year. Country of birth, including Australia, was completed for 1,912 records (27%) and of the 1,775 notifications that indicated an overseas country of birth, the year of arrival field was completed in 504 notifications (28%).
In this study we describe the epidemiology of CHB in Victoria based on notification data, highlight the limitations in this data and consider priorities for improved CHB surveillance. Following a gradual decline in the age-standardised notification rate of CHB in Victoria from 1998 to 2004, rates have followed an increasing trend over the past three years. This recent trend has been predicted to continue for at least another five years based on modelling of population, migration and notifications data14 which underscores the relevance of improving the public health management of this disease in Victoria.
A unimodal age distribution with 30% of notifications occurring in the age-group 25–34 years was observed and this is consistent with prior Australian and international studies.15–17 The high rate among this age group is most likely due to biological factors. Hepatitis B e-antigen (HBeAg) serves as a marker for active viral replication and seroconversion, marked by the development of antibodies to the HBeAg and is a process associated with hepatic inflammation.18 The inflammation causes abnormalities in liver function tests and sometimes clinical symptoms.15,18 If detected, these abnormalities can lead to a consideration of causes of liver disease and subsequent diagnosis and notification.19 The median age at notification in Victorian data is consistent with findings from a large population study in Chinese adults with CHB that found the median age at HBeAg seroconversion was 34.5 years, although symptoms of hepatitis occurred in only 14.6% of the 1,393 patients in this study who experienced a hepatitis flare.15
Once infected with hepatitis B virus, males are more likely to remain chronically infected than women.20 The Victorian notification data showed that males were more commonly notified than females in most age-groups. The predominance of females notified in the age groups 20–24 and 25–29 years was unexpected and likely reflects the detection of HBsAg as a result of antenatal screening. The fact that screening could lead to an excess of Victorian female cases when compared to male cases in 20–29 age-group – contradicting the expectation for a male predominance of CHB – suggests that targeted screening can improve the detection of disease.10 A study of antenatal women from central Sydney between 1996 and 1999 resulted in an overall HBsAg seroprevalence of 1.5%.2 There was, however, a much higher seroprevalence among mothers born in South-East Asia (5.4%) and North-East Asia (4.9%) compared to non-Indigenous Australian-born mothers (0.2%). More recent data from antenatal screening at two large Sydney teaching hospitals revealed an overall HBsAg prevalence of 2%, with 98% of these women having been born overseas and more than half of them from China or Vietnam.10 There has not been a published study of the Victorian antenatal population since 1984, when analysis of antenatal testing of women at the Royal Women's Hospital demonstrated an overall prevalence of HBsAg of 0.8%, with a prevalence of 0.5% in Australian-born women compared to 11% in women born in South-East Asia.21
The regional age-standardised notification rates were highest in the metropolitan areas of Melbourne with high proportions of overseas-born residents. The country of birth information reflected the heterogeneity of HBV prevalence worldwide, with the majority of notified cases from China and Vietnam. Those born in Sudan and Burma represented an increasing proportion of CHB cases notified in the past six years. This is most likely related to migration patterns of refugees and the targeted screening that occurs in these groups leading to increased case ascertainment for CHB. The recent increase in the number of notifications specifying China as the country of birth is difficult to explain without comprehensive year of arrival information. According to the 2006 Census, the number of people living in Australia born in China has increased considerably in the past 10 years (by 96,000 people).10,13 This increase was second only to the increase in the number of Australian residents born in New Zealand over this same period (98,000).13
The year of arrival information was completed for less than a third of cases with an overseas country of birth. In 2008, the country of birth and year of arrival became mandatory fields on the notification form, therefore the proportion of notifications recording this information may increase, however it is too early to comment on the effect of this change on the data. For cases with recorded year of arrival, the median time between year of arrival and notification date was only one year. When time to notification was analysed for the 12 most commonly recorded countries of birth, there were clear differences between countries. Cases born in Vietnam took an average of nine years to be notified, compared to 1.3 years for cases born in Sudan. This probably reflects the increased screening offered to humanitarian arrivals in recent years. A population modelling study completed in 2009 has suggested a median time lag of 10 years from arrival to notification that was significantly correlated with migration patterns in the 10 years prior to the notification. This was supported by the median age of 25 years at migration and then 35 years at notification.14 This finding could explain the longer time to notification in Vietnamese and Chinese-born cases whose notification is most probably related to clinical presentation or antenatal screening as opposed to screening on arrival for humanitarian entrants. The lack of available year of arrival information may have affected the representativeness of the results and, therefore, limited our ability to draw firm conclusions.
The information collected on CHB cases relies on data quality of laboratory notifications as these constitute the majority of the notifications. Typically, laboratories provide high-quality basic demographic data, however they cannot provide information about Indigenous status, country of birth and year of arrival due to the absence of such fields from laboratory test request forms. Doctor notifications are needed to collect risk factor data, but the proportion of notifications that include information from doctors has been steadily decreasing since 1998 and now represents less than a third of notifications.
There are two important consequences of this reduction in doctor notifications. First, experience from Australia and overseas has found that in the absence of doctors’ notifications up to 30% of patients can be unaware of their diagnosis.16,19 One of the Victorian surveillance objectives for CHB is the reduction in further transmission by providing information to cases through treating doctors.22 If the majority of notifications are from laboratories only, some patients may remain unaware of their diagnosis. This could facilitate the spread of infection to susceptible contacts and prevent timely receipt of medical advice and specialist care. The second consequence of a continuing decline in notifications from doctors is further reduction in the representativeness of surveillance data. Birth in a high-prevalence country has been shown to be the most important risk factor for the diagnosis of CHB in Australia;16 however, this field is completed in only 28% of doctor notifications, which currently make up less than half of all CHB notifications. In 2008, low notification rates for all infectious diseases from doctors compared to laboratory notifications prompted the development of a new Communicable Disease Prevention and Control Unit (CDPCU) strategy to reverse this trend.23 The strategy includes improving technology to enable notifications and focus on educational activities that support notification, such as integrating key messages into communication from the divisions of General Practice Victoria and engaging with universities.23 Improving the quantity and quality of doctor notifications for CHB could not only make surveillance data more representative, it could significantly increase its policy utility when considering targeted CHB interventions.
There were limitations of our descriptive analysis of CHB. First, we used only Victorian surveillance notifications, and it is well established that under-notification affects the ascertainment of all notifiable infectious diseases24 and, therefore, our data would not represent all cases of CHB diagnosed in the past 11 years. However, since it is unlikely that under-ascertainment rates would have altered significantly over the period of this analysis, we are confident that the trends observed in this study reflect general disease trends in Victoria. Nevertheless, our data were affected by the targeted screening that occurs in the community, such as antenatal and refugee screening, but we were unable to quantify this effect.
The revised CDC recommendations for the public health management of chronic hepatitis B were developed in response to the growing burden of this chronic infectious disease. The recommendations discussed the importance of registries of persons with chronic HBV infection and suggested a stepwise response to the establishment of registries within health departments. Developing registries begins with improving the completeness of data fields currently collected and then adding data fields where risk factor information is not routinely collected. In terms of public health management, gradually increasing the expected amount of follow-up of CHB cases on the part of public health authorities is vital public health collaboration.11 In keeping with the department's strategy to improve doctors’ notifications for all diseases, we recommend a short pilot of enhanced surveillance for CHB. One option for this surveillance could include a laboratory-led intervention whereby a link to the web-based notification form is embedded in electronic laboratory results positive for CHB when results are sent to doctors. An alternative option is to send notification forms to all doctors whose patients’ positive CHB tests have already been notified by laboratories to the department. In an attempt to improve response rates, notification forms could be accompanied by a summary of up-to-date clinical or public health information about hepatitis B to make the notification process become more of an information exchange with doctors.
In conclusion, in Australia there are no systematic guidelines to improve the public health management of chronic hepatitis B infection although the recent release of Australia's first National Hepatitis B Strategy provides an ideal opportunity to address this deficiency.25 Surveillance is a key priority as measuring the extent of the hepatitis B epidemic is integral to evaluation of the current public health response.26 A priority action for hepatitis B surveillance should include an improvement in the reporting of country of birth and Indigenous status on routine chronic hepatitis B notifications to enable an improved estimation of these risk factors.26 Improving information collected from notifications is the first step in a tiered approach to the development of a CHB registry as recommended by the CDC.
Trevor Lauer (DH) for assistance with customised data extraction. Stephanie Williams is a Master of Applied Epidemiology scholar funded by the Australian Government Department of Health and Ageing.