The demographic and clinical characteristics of primary liver cancer patients in Tasmania

Primary liver cancer (PLC) is the sixth most common cancer globally, and is the fourth leading cause of cancer mortality, accounting for 8.2% of all cancer deaths in 2018. Despite advances of new diagnostics and treatments, survival time of PLC patients remains poor, particularly in the absence of targeted screening programs to increase rates of early diagnosis. In Australia, both incidence and mortality rates of PLC have increased substantially, in contrast to the improved outcomes that have been observed for almost all other cancers. The risk factors associated with PLC include chronic hepatitis B (CHB), chronic hepatitis C (CHC), diabetes, alcohol related liver disease, and non-alcoholic fatty liver disease (NAFLD). Demographic and epidemiological changes are impacting incidence of PLC in Australia, for example, the migration of CHB patients from high hepatitis B prevalence countries, and the increasing rates of obesity and diabetes. Understanding these changing risk factors and the impact on survival is important because it will help to inform healthcare policy for appropriate allocation of resources for prevention, targeted screening/surveillance and treatment. The aims of this Masters' research program are to describe the demographic characteristics, risk factors and survival time for all PLC cases in Tasmania (Study 1: Chapter 3) and to investigate the level of agreement for cause of death data between the Australian Bureau of Statistics (ABS) and Tasmanian Cancer Registry (TCR), along with medical specialist opinions and its impact on cause-specific survival (Study 2: Chapter 4). Over a nine year period between 2007 and 2015, there were 293 PLC cases identified by using linked administrative data between the datasets. Hepatocellular carcinoma (HCC) (51.9%) and cholangiocarcinoma (20.5%) were the main types reported. Three-quarters of all PLC cases were male, and the average age was 70 years. For cases with a public hospital admission, 43% did not have a risk factor for PLC identified. Of those who did, the most common were cirrhosis (37%), chronic viral hepatitis (35%), diabetes (27%), and alcohol-related liver disease (23%). The mean age at diagnosis for all cases was 69.6 years, with a median survival time of 6.2 months. The 1-,3- and 5-year relative survival rates were 38.3%, 12.8%, and 6.7% respectively. The linked PLC cases provided the opportunity to compare specific causes of death between the TCR and ABS. Conflicting records of cause of death were recorded for almost half of all PLC, with 20 cases had non-PLC underlying causes of death from the TCR dataset and 42 cases from the ABS dataset. These cases were independently reviewed by medical specialists with expertise in PLC. Concordance of cause of death data was estimated using Kappa statistics, and the impact on cause-specific survival time was explored using a competing risk framework. The overall concordance regarding causes of death data was minimal between the ABS and TCR (Kappa=0.35), moderate between the ABS and medical practitioners (Kappa=0.61), and strong agreement (Kappa=0.87) between the medical practitioners were observed. These results reflect a greater level of agreement between medical practitioners than between coding agencies. Overall, cause-specific survival time was similar across the TCR, ABS and medical practitioners by sex, place of residence and country of birth, however, a small difference was observed for the type of liver cancer between the ABS, TCR and medical practitioners. As liver cancer is a low survival cancer, such results may be different to cancers with better survival such as breast cancer. This thesis has made many distinct contributions including reporting the epidemiology of PLC in Tasmania and characterising the risk factors specific to this setting. These results support the need for surveillance to increase the rate of early detection of HCC in high risk patients. In turn, this will improve the very poor survival of PLC cases in Australia. This is the first study that has evaluated the impact of different coded causes of death on estimates of cause-specific survival for liver cancer cases.