The suicide rate in Australia has increased over recent years, indicating a need for novel suicide prevention interventions. A large proportion of people at risk of suicide do not seek professional help, but rather express their suicidal thoughts to close friends and family members. However, expressions of suicidal thoughts to family and friends can be indirect and ambiguous, and consequently can be missed or misunderstand, leading to dismissive response and missed opportunities for suicide prevention. A suicide prevention media campaign aimed at family members and friends may be one useful population-level suicide prevention strategy to encourage greater recognition of suicide risk and appropriate helping actions in response to suicidal communications. Current evidence for what messages would be acceptable and appropriate for inclusion in such as campaign, however, is limited. The research described in this thesis was undertaken to develop suicide prevention messages to include in an Australian suicide prevention campaign aimed at family and friends of adults at risk of suicide. To understand what messages would be most important to include, five studies were undertaken and are described in this thesis. Study 1 was an expert consensus study involving suicide prevention professionals and people with lived experience of suicide risk. This study established which suicide prevention actions these experts believe are the most important to encourage in a suicide prevention campaign aimed at family and friends. Study 2 was an online survey study involving people with lived experience of suicide risk that aimed to assess the most and least helpful actions taken by others in response to their suicidal communications. Studies 3 to 5 used data from a nationally representative telephone survey conducted with Australian adults. Study 3 examined Australian adults’ confidence and intentions to help a person close to them at risk of suicide. Study 4 examined helping actions given and received in response to suicidal communications. Study 5 assessed the relationships between beliefs in suicide ‘myths’ and helping intentions and actions. Overall, the findings from these studies show that Australian adults are confident and willing to assist a person close to them at risk of suicide. They intend to undertake, and indeed do undertake, a number of appropriate helping actions toward people close to them who are at risk of suicide. These appropriate actions include listening and talking to the person at risk and encouraging them to seek professional help. However, Australian adults also largely fail to ask important risk assessment questions, and commonly undertake actions that do not conform to best practice in suicide prevention. Such non-recommended actions include telling the person at risk ‘what they have going for them’ and telling them that their suicide would hurt their family and friends. A substantial minority of Australian adults also believe in suicide myths, including those related to encouraging suicidal thoughts by talking about suicide. The combined results of these studies have been utilised to make recommendations regarding the most useful messages to include in an Australian suicide prevention campaign aimed at family members and friends at people at risk of suicide.

Background The use of Community Treatment Orders (CTOs) in the treatment and care of those with severe mental illness continues to attract debate and remains controversial despite increasing use in many countries. Mental health legislation is driven by the policy directions of the government of the day, and also reflects community attitudes and expectations. In recent years there has been an increasing focus on individual rights as reflected in the expectation that use of restrictive interventions and compulsion will be kept to a minimum. These changes in policy and law have occurred in parallel with a reduction in bed-based services, and in Victoria, a reduction in service availability overall on a per capita basis. The objective of this thesis was to examine the processes involved in civil commitment under mental health legislation and factors that influence this, and to seek the views of those who provide the most direct care and support in the community. The thesis sought to determine whether subsequent care is influenced by the manner of discharge from a compulsory order, and the impact of new legislation with a greater emphasis on human rights. The thesis considers the changes in rates of compulsory care following reform of mental health legislation in the context of ratification of the United Nations Convention on the Rights of Persons with Disabilities. The thesis also considers the influence of substance use on whether compulsory care is instituted, and the experience of those who care for people with severe mental illness. Methods This thesis incorporates four studies. Studies one to three use the Victorian state-wide data system of mental health service utilisation. Two separate cohorts of people who had been on a CTO for at least three months were followed for two years. The first covered service utilisation for two years following the end of a CTO during 2008 – 2010, and the second from 2014 – 2016. This enabled consideration of the role of external review such as by the Mental Health Review Board in discharge and continuity of care planning, a comparison between two different legislative regimes, and the impact of clinical factors such as having substance use disorder in addition to severe mental illness. Study four incorporated a mixed qualitative/quantitative survey posted to carers who were members of two carer support and advocacy groups. This study sought to understand the views and experience of people who cared for those with severe mental illness such as schizophrenia. Through these studies the candidate sought to determine: whether the mode of termination of a CTO was related to subsequent service utilisation; whether the intention of legislative reform to reduce the use of compulsory treatment was supported by comparison between service utilisation after a CTO under the Mental Health Act (1986) and Mental Health Act (2014); and whether co-morbid substance use resulted in greater likelihood of being made a compulsory patient. Main Findings Study 1 found that the way a CTO ended influenced subsequent treatment as a compulsory patient. If the CTO ended suddenly the person was more likely to require use of the mental health act in the future. Study 2 showed that there had been a reduction in the use and duration of orders under the Mental Health Act (2014) compared to the Mental Health Act (1986). There was also an increase in the number of CTOs varied to an inpatient order by the authorised psychiatrist. Study 3 found that having a substance use disorder increased the use and duration of compulsory orders. This related to those with schizophrenia and mood disorders, with the latter more likely to return to inpatient care. The study raised the question of how effectively or otherwise mental health services and drug and alcohol treatment services work together. Study 4 showed that most of the carers who responded supported the use of CTOs. They reported considerable negative experience of care-giving. Conclusions The Victorian mental health service system has changed greatly over the past decades. Legislation has been reformed to meet contemporary expectations of oversight and regulation, including the introduction of CTOs. However, the intent of legislative reform may have been limited by the system failing to keep pace with population growth. The research raises questions about how we care for those with severe mental illness in the community and the impact the current service system has on carers and other support services. Reforming mental health legislation does not, of itself, result in improved patient outcomes or experience of care. Under the current system of care, the thesis supports continuing use of CTOs, but with greater emphasis on clinical engagement and care planning with other relevant care providers, including family or carers. This research occurred in the context of growing disquiet regarding the mental health service system and the subsequent commissioning of a Royal Commission.

The purpose of this thesis is to review why and how Australian immigration levels were sharply increased from around the year 2000, largely but not solely in response to widespread concerns about the projected rate of population ageing. The literature on the economics of immigration and on the economic and budget impact of population ageing is extensive. However, there is relatively little research on the use of immigration to slow the rate of population ageing. Australia has long used immigration as a tool of economic policy. This thesis examines how change in the Howard Government’s approach to immigration from 2000 led to a sharp rise in the population growth rate and a slowing in the rate of population ageing. This was done primarily for economic reasons but was not projected in any scenario used by the Australian Bureau of Statistics (ABS) in its 1999 base year population projections. The thesis examines the impact of this from a range of perspectives but focuses particularly on economic and budget performance during the demographic dividend phase (ie when the working age to population ratio is rising towards a peak) compared with the demographic burden phase (ie when the working age to population ratio is falling after reaching a peak). The thesis also compares the economic and budget performance of the world’s most aged nation (ie Japan) with Australia, one of the developed world’s youngest nations. The research suggests economic growth should be weaker in the demographic burden phase – but is that borne out? This is not a thesis in demographic technique but in public policy on immigration and population. It seeks to highlight the importance of demographers developing a deeper understanding of immigration policy if they are to move away from the traditional approach of using a single measure of net overseas migration based on an average of past years (with a couple of high/low scenarios) to analyse population directions and to begin using a more granular breakdown of net overseas migration for this analysis. In that way, demographers can make a stronger contribution to public policy.

Background Globally, colorectal cancer is one of the leading causes of cancer related death. Asia has had relatively low incidence rates of colorectal cancer when compared with western countries. Over the last decade, however, occurrences have increased in many Asian countries. In East Asia, colorectal cancer incidence has been increasing over the last few decades; however, many countries are still lack of strategies on colorectal cancer screening. Colorectal cancer screening guidelines in many Western countries use family history information when assigning people into different risk categories to recommend for frequency and intensity of screening strategies. However, the relationship between family history and colorectal cancer and adenoma risk has not been well studied in Asian countries. Given that family structure and environmental exposures are likely to be different in Asian countries it is assumed, the association may also prove to be different. Aims This thesis aimed to investigate the role of family history of colorectal cancer in colorectal cancer screening in East Asian countries. Method This thesis includes the following five studies addressing the gaps in knowledge of the role of family history in colorectal cancer screening strategies. The main database for this thesis is the prospective data of 175,000 men and women from three countries (Japan, Korea, and China) who have been recruited into the Asia Cohort Consortium (ACC, https://www.asiacohort.org/). 1. A literature review of current colorectal cancer screening guidelines in Asian countries. 2. A systematic review of studies on the association between family history of colorectal cancer and the risk of colorectal cancer in Asian countries. 3. A pooled analysis of six prospective cohorts from Asia Cohort Consortium to investigate the association between family history of colorectal cancer and the incidence of colorectal cancer in Asian countries 4. A pooled analysis of six prospective cohorts from Asia Cohort Consortium to investigate the association between family history of colorectal cancer and the mortality of colorectal cancer in Asian countries 5. External validation of the Asian-Pacific Colorectal Cancer Screening (APCS) Score using six prospective cohorts from Asia Cohort Consortium. Results Firstly, from the meta-analysis, the association of colorectal neoplasia with having at least one first-degree relative family history of colorectal cancer differed by Asian region: being 1.69 (1.46, 1.96) for East Asia; 3.44 (2.32, 5.10) for Western Asia and 5.14 (2.44, 10.8) for South-Eastern Asia. The strength of associations between family history of colorectal cancer and the risk of colorectal cancer and adenoma for East Asia was similar to that of Western populations but higher for Western and South-Eastern Asia. This information will be useful to triage the populations in Asian countries for colorectal cancer screening by the strength of family history in different regions of Asia. Secondly, with the pooled analysis of exist data from the Asia cohort consortium, the family history of colorectal cancer associated with a higher risk of incidence of colorectal cancer in Asian countries. Thirdly, the evidence from family history and mortality, this is no evidence for an association between family history of colorectal cancer and overall survival or colorectal cancer-specific mortality. However, specific mechanisms underlying family history may have a prognostic impact and merit further study. Future studies need to collect data on possible confounders of the effect, such as tumor stage at diagnosis, presence of multiple cancers, mode of presentation to hospital, and surgical or clinical management in order to produce more conclusive results. Last, the APCS scores were efficient to identify individuals with a high risk of colorectal cancer in the East Asia population. Conclusions This thesis has generated new knowledge about family history and the risk of colorectal cancer in East Asia. The systematic review and meta-analysis demonstrate that there is large heterogeneity in the association with colorectal neoplasia with having at least one first-degree relative family history of colorectal cancer differed by Asian regions. Family history could be a useful tool for colorectal cancer screening program but might not be associated with the mortality of colorectal cancer. APCS scores can effectively identify individuals with a high risk of colorectal cancer in East Asia. The feasibility and acceptability of the APCS score to the general population in Asia will need to be further examined.

Introduction The rapidly rising burden of non-communicable diseases (NCDs) in China affects the demand for healthcare services among households and threatens to create an increased financial burden on households, at a time when economic growth has been slowing down and traditional forms of inter-generational support are in decline. However, the design of effective policies to address household impacts of NCDs in China is hindered by a lack of high-quality work on the economic roots of NCDs, the impact of NCDs on households and economic gains that might result from policy action to address NCDs. This thesis aims to fill this gap by investigating the association between socio-economic characteristics and behavioural risk factors and NCD onset, by estimating the economic consequences of NCDs on current and future household welfare associated with coping strategies and by estimating the economic consequences of delayed treatment for NCDs. Methods This thesis consists of three quantitative analyses, all of which use longitudinal datasets. Study One uses three waves of longitudinal data from the China Family Panel Studies survey (CFPS, 2010-2014) and a dynamic model conditional on not having an NCD in the initial period to estimate the effects of behavioural and socio-economic risk factors on the onset of NCDs among individuals in China. Study Two uses the same dataset from CFPS over 2010 and 2014 and fixed-effect models to examine whether households in China can maintain consumption expenditures in the face of a new NCD diagnosis and how the households cope with the diseases and the consequences of the coping mechanisms. Study Three draws data from the China Health and Retirement Longitudinal Study (CHARLS, 2011-2015) and applies difference-in-differences (DID) methods on a matched dataset of individuals to estimate the effects of variation in treatment timing (i.e. early versus late treatment) on healthcare use and health spending among individuals newly diagnosed with hypertension. Results Study One shows that being obese, using solid cooking fuels, a history of frequent drinking, and a higher level of household total consumption expenditure during the preceding period were positively associated with NCDs. More importantly, through disaggregating the effects by age subgroups, it shows that some risk behaviours one established decades ago (such as frequent drinking) have long-term and accumulative effects on health, even people have changed behaviour at older ages. Study Two reveals that (non-poor) Chinese households that experienced a substantial increase in out-of-pocket (OOP) health expenditures following an NCD shock can smooth consumption expenditure, through the depletion of housing assets and receipt of public transfers. For poor households, there was no significant increase in health expenditures and no effects on household consumption and assets, suggesting that they might abandon treatment. Both the non-poor and poor were paying the price, the first group in terms of future wealth and the second in terms of health. Study Three suggests that patients who started treatment late spent more on health services and were more likely to incur catastrophic health expenditures (CHE) once they initiated treatment, relative to counterparts who began hypertension treatment earlier. Hypertensive patients having started treatment earlier experienced declines in the probability of incurring health expenditures in the subsequent period. Implications In conclusion, NCDs have a long-term impact on household welfare in China. Ignoring the economic cost associated with coping with NCDs and with delayed NCD treatment, can underestimate the costs in the future. Public intervention policies to address the emerging risks of NCDs at an early stage, and insurance policies to protect individuals and households from potentially catastrophic expenses of treating NCDs are needed.

Tobacco use is the leading risk factor for poor health outcomes in Australia. Aboriginal and Torres Strait Islander people (Aboriginal people) are disproportionately affected due to the historical legacies that have embedded social inequities associated with higher smoking prevalence in all population groups. However, smoking prevalence has declined over recent decades for all Australians, including Aboriginal people, although a persisting disparity remains. While encouraging existing smokers to quit is a key part of reducing smoking prevalence overall, as most smoking initiation occurs during adolescence and young adulthood, preventing young people from becoming established, long term smokers is also critically important to this effort. This thesis presents the work from a multi-component PhD project exploring trends in smoking initiation; the experiences of non-smokers; and individual, social, cultural and environmental risk and protective factors influencing Aboriginal adolescents. Using a multi-methods research design the project has synthesised findings from a systematic review, cross-sectional analysis of survey data from several datasets, as well as qualitative research. A systematic review of international evidence indicates that young Indigenous peoples globally experience similar smoking influences to other youth populations, such as smoking among family, friends and the broader community, however few studies have explored the unique historical and cultural factors. In Australia, Aboriginal youth smoking prevalence has declined steadily with the majority of secondary students identifying as never smokers. However, these changes in smoking prevalence are occurring in parallel to the total population and remain significantly higher for Aboriginal youth. While initiation of regular daily smoking mostly occurs before aged 18, it can occur over a wide age range and has been increasingly delayed until early adulthood for some. Factors which may protect against smoking initiation among young Aboriginal people are good social and emotional wellbeing, connections to community and school, not being sexually active, using alcohol or having contact with the justice system. Sustained investment in tobacco control is required to accelerate declines in prevalence and not reverse earlier gains. This includes whole of population approaches as well as targeted programs. Community organisations should be resourced to deliver programs promoting positive engagement with family, school and community as these may be protective against smoking initiation. Programs should aim to reach young people in early adolescence and continue to reinforce smoke-free intentions into young adulthood. However, some young people may have already developed nicotine dependency and need support to quit. Community priorities are for broader, holistic programs that build on cultural strengths and empower young people, reducing smoking as part of a process that improves health and wellbeing overall.

Malaria remains a major public health threat with pregnant women and young children the most vulnerable populations. Approximately 125 million pregnant women are at risk of malaria infection every year. Immunity to malaria can significantly reduce the severity of malaria symptoms. This has long motivated efforts to better understand the immune response to malaria in order to develop an efficacious antimalarial vaccine. However, the complex behaviour of the immune system, particularly in pregnant women, has hindered identifying the underpinning mechanisms of the immune response to malaria infection. This project investigates antibody-mediated immunity in pregnant women who attended antenatal clinics located at the Thai-Myanmar border, where malaria transmission is low and P. falciparum and P. vivax (the most prevalent malaria species globally) coexist. Antibody responses during pregnancy to six parasite antigens were measured for 250 pregnant women in a median of 7 samples per woman (range 2 to 13) over the gestation period. A multivariate mixture linear mixed model was fitted to longitudinal antibody data of 250 pregnant women to characterise the highly dynamic antibody responses using a Bayesian approach. The posterior distribution of the parameters estimated via Markov chain Monte Carlo (MCMC) simulations were used to classify the antibody responses. The results show that the infectious status of a woman during follow-up is a key factor influencing the classification of the joint behaviour of the antibody responses. Hence, the two malaria exposure categories (exposed to infection during pregnancy defined as a case and those non-exposed defined as controls) were represented by the two antibody profile clusters. Using a manually developed code, entropy values were computed for each antibody, with which, the contribution made by each antibody on the classification was assessed. The antibodies PfMSP3 and PvAMA1 which maintained less dynamic antibody profiles significantly influenced the classification. However, these antibodies identified controls exceptionally well but did not perform well for the cases. For sero-surveillance, antibodies which best identify the cases are required, hence, the study was extended in performing classification based on all possible univariate and multivariate combinations of the six antibodies. The bivariate combination PfAMA1 and PfVAR2CSA resulted in identifying the majority of cases by contributing towards identification of potential biomarker(s) for sero-surveillance of recent exposure to malaria during pregnancy. Therefore, the bivariate combination of antibodies, PfAMA1 and PfVAR2CSA should be used in the field, particularly for accurate malaria surveillance of pregnant women living in low malaria transmission settings. This could lead to the early detection and treatment of malaria infections in pregnant women, reducing transmission and thereby progressing towards elimination of malaria.

Participation in decent work is recognised as central to supporting individuals, with and without disabilities, to attain socio-economic opportunities, and, health and well-being. Through their ratification of the United Nations Convention on the Rights of Persons with Disabilities (UNCRPD), Australia has committed to supporting people with disabilities realise their right to work. Yet, Australians with disabilities continue to experience significantly poorer employment outcomes when compared to Australians without disabilities. Among those with disabilities, people with psychosocial disability experience some of the lowest levels of employment. Challenging life circumstances, intertwined with vocational and non-vocational barriers to employment, are recognised as both causes and consequences of poorer employment outcomes. The Australian Disability Employment Services (DES) program, is the federal government’s specialised welfare program for people whose disability is assessed as their main barrier to employment. Over recent decades, continual reform has seen the DES program transition from a publicly-funded and delivered program to a quasi-market of government contracted for-profit and non-profit businesses who support and monitor the efforts of people with disabilities in receipt of income support (and a small number of voluntary participants) to actively promote their employability and participation in work. Under the current reforms, implemented in July 2018, DES participants have been given more choice and control to determine which provider they use and to change providers if they are not satisfied. It is assumed that these reforms will incentivise DES providers to be more responsive to participants’ needs, ultimately leading to improved employment outcomes. Drawing on survey data from 369 respondents with disabilities engaged with government funded employment services, and, qualitative interviews conducted at two time points with 30 DES participants with psychosocial disability (total 56 interviews), this thesis explored the perspectives of these participants on how their life circumstances influence and/or are influenced by their engagement with the DES program, in the context of the current reforms. The findings suggest that despite considerable investment and reform, the DES program struggles to address the significant and complex barriers to employment experienced by DES participants with psychosocial disability. Addressing these barriers requires broader social policy investment. Until this is achieved, the effectiveness of the DES program will continue to be undermined.

Background: Globally, there has been an extraordinary growth in refugees and asylum-seekers in the past two decades. Australia experienced a dramatic increase in boat-arrivals in 2008–2013, with many being Afghan Hazara including unaccompanied asylum-seeker minors (UAMs). Global and Australian studies highlight the significance of providing UAMs with appropriate settlement services during their adjustment period, yet few ethno-specific studies have to date sought service provider perspectives on support provided to male Hazara UAMs pre-18 years and former UAMs (fUAMs) post-18 years. Methods: This qualitative study addressed this research gap using ecological systems theory and social networks theory in a dual-sector case study design to explore the perceptions of professional case workers (n=6) and voluntary English language teachers (n=6). Based on key informant views, this study focused on the expectations and experiences of male Hazara UAMs’ compared to fUAMs’ select settlement services during their early adjustment period (2015–2016) in Melbourne, Australia and how this support influences their settlement and wellbeing. Results: The findings indicate UAMs’ three-part migration trajectory (from Afghanistan, via transit countries to Australia) was perceived to be driven by classic push and pull factors and shaped by family separation. Following arrival, Australia’s immigration and settlement policies determined the type and level of support provision, with UAMs in community detention receiving maximum services while fUAMs in mainstream community being offered highly reduced services and temporary visas. English language was perceived to have strong influence on their early adjustment, and different forms of social networks could promote their socioeconomic participation. Despite UAMs experiencing potential risks pre-arrival, protective factors post-arrival, especially support services, could rebuild their resilience. Conclusion: This study has implications for asylum-seeker and settlement policies, casework practice, and researchers to leverage settlement process and wellbeing outcomes of this group of unaccompanied youths in Australia.

Vaccines are a cornerstone of public health saving millions of lives each year, however, adverse events following immunisation (AEFI) are a reality. Comprehensive post-licensure vaccine safety surveillance is therefore essential to maintain confidence in immunisation, inform policy and ensure the ongoing safety of immunisation programs. Vaccine pharmacovigilance in Australia has been criticised and found inadequate due to its inability to inform on vaccine safety events in a timely manner. In 2010 this meant failure to prevent serious neurological sequalae arising from increased febrile reactions with one particular brand of influenza vaccine. Subsequent expert reviews of Australian AEFI surveillance proposed the Victorian state-level enhanced passive surveillance system “SAEFVIC” as a potential model system, with suggestion to expand to jurisdictions that did not have robust AEFI surveillance. However, despite this recommendation, no evaluation of SAEFVIC’s ability to provide effective vaccine pharmacovigilance had been undertaken. My doctoral research evaluates “SAEFVIC”, the Victorian AEFI surveillance system’s ability to detect and investigate emergent vaccine safety issues with the findings presented in a range of publications and discrete reviews. Within this thesis I have developed real-time investigation of vaccine safety issues, which have informed policy decisions on immunisation program delivery and ensured community confidence. Description of the system epidemiology provides a foundation for research integrity. Individual studies have identified opportunities for improvement and guided amendments to data collection to better inform on at-risk populations. I explored and applied a simple, feasible and replicable measure of disproportionality for signal detection enabling implementation of routine signal detection. The establishment of contemporary reporting and visualisation has contributed to national expansion of SAEFVIC as an integrated surveillance platform, markedly increasing national capability to provide “comprehensive and effective post-licensure AEFI surveillance critical to maintain confidence in immunisation and inform evidence-based policy for vaccine safety”. My research has provided the evidence-base to optimise the SAEFVIC passive AEFI surveillance system and—through translation of the findings into policy and practice—helped consolidate Victoria as a leader in Australian AEFI surveillance, a major southern hemisphere partner in global AEFI surveillance and ultimately made immunisations safer. The timeliness of my doctoral thesis has become prophetically pertinent in the midst of the 2020 global Coronavirus SARS-CoV-2 crisis, with calls for development and release of novel vaccines at “pandemic speed”, highlighting the need for robust and timely vaccine safety surveillance.

Background: Cancer is the second-highest cause of mortality worldwide, accounting for 9.6 million deaths in 2017, and its burden on individuals, economy and health system is a global problem. Although Australia has a universal health-care system, cancer survival varies by socio-economic disadvantage. The magnitude and age-related patterns of these inequalities are not clear in Victoria, Australia and little is known about whether the gap in cancer survival is narrowing or widening over time. Despite strong evidence that sex is associated with the incidence and prognosis of several diseases, few studies have assessed differences in cancer survival by sex, and no Australian study has comprehensively evaluated the magnitude and temporal and age-related patterns of differences in cancer survival by sex. This thesis aimed to investigate inequalities and trends in cancer patient survival by sex, area-level socio-economic disadvantage and remoteness of residence and identify the underlying causes of socio-economic inequality in colon cancer survival. Methods: Six studies were designed and conducted to address the aims of this thesis. The first study was a systematic review of peer-reviewed published articles since 2005 that attempted to identify the underlying reasons for socio-economic inequalities in cancer survival. The second study was a systematic review of peer-reviewed published articles that assessed the association of rural-urban residence with cancer survival in high-income countries and summarised the current evidence in relation to potential contributing factors to the observed differences in survival between rural and urban regions. Studies 3, 4 and 5 used the Victorian Cancer Registry data to assess differences in cancer patient survival by sex, area-level socio-economic disadvantage and remoteness of residence. Study 6 used population-based linked health data to identify factors explaining socio-economic inequalities in colon cancer survival. Net survival was estimated using the Pohar-Perme method. Inequalities in survival were modelled using Poisson regression to estimate excess mortality rate ratios. Interventional causal mediation analysis was used to assess the underlying reasons for socio-economic inequalities in colon cancer survival. Results: Findings from the first systematic review suggest that socio-economic inequalities in cancer survival appear to be partly explained by differences in disease stage, health-related behaviours, co-morbidities and treatment modalities. The second review shows that rural cancer patients generally have worse survival compared with their urban counterparts. The underlying reasons for survival disadvantage for rural patients are not well known, but possible contenders include differences in lifestyle behaviour, stage of cancer at diagnosis, co-morbid conditions, and treatment modalities. Findings from analysing the Victorian Cancer Registry data suggest that cases living in more socio-economically disadvantaged areas or outside major cities generally had lower five-year survival than their counterparts living in less disadvantaged regions or major cities. In addition, men showed lower survival from most cancers than women. The observed socio-economic inequalities for colon cancer survival were not explained by variations in stage at diagnosis or surgery. Not receiving chemotherapy within 8 weeks after surgery, emergency admission prior to diagnosis and receiving surgery at a public hospital appeared to partly contribute to lower survival for men with stage III colon cancer living in the most disadvantaged areas, although this observation might be due to unmeasured characteristics of disadvantaged men receiving surgery in public hospitals rather than public/private hospitals themselves. Conclusions: This thesis has shown that inequalities in cancer patient survival by sex, area-level socio-economic disadvantage and remoteness of residence exist and have persisted over time, even though Australia has a universal health care system. Monitoring the magnitude of these inequalities and conducting high-quality research to identify the underlying reasons for inequalities in cancer survival is important to prioritise actionable factors to improve cancer outcomes for all patients.

Background Mother-to-child transmission (MTCT) accounts for 90% of HIV in children. Globally, HIV has resulted in the deaths of 120,000 children. African countries have reduced MTCT risk from 25-30% to less than 5% through prevention of MTCT (PMTCT) using antiretroviral therapy (ART), elective caesarean section, and replacement feeding. Indonesia implemented PMTCT in 2004, yet the MTCT risk (26.6%) remains the highest in the world. MTCT is the second cause of new HIV infections in Papua, Indonesia. Nearly half of (29,083) HIV-infected people in Papua are women within childbearing age of 15-49 years. Despite the implementation of Option B+ PMTCT in Papua since 2012, new HIV infections in children continue to rise, suggesting low PMTCT coverage or ineffective PMTCT programs. There was limited literature on PMTCT implementation and outcomes in Papua. This research aims to investigate the performance of Option B+ PMTCT in Papua and to identify factors that influenced women’s uptake, adherence, and retention in the program. Methods This research used a mixed-methods research design comprising of three studies: 1) Papua Study, 2) Indonesia Study, and 3) Global Study. Study 1 included field observations and in-depth interviews of 20 HIV-positive women and 20 PMTCT health workers in Papua. Study 2 involved analyses of PMTCT data in Indonesia in 2017, and Study 3 comprised a realist review of Option B+ for PMTCT in resource-limited settings. Results Using a quantitative approach, Study 2_Indonesia found high ART uptake (98%) but low ART retention (33%) in Papua. Health facility characteristics that predicted women’s retention in PMTCT program were hospitals, hospitals class A/B, generalized HIV epidemic, and high HIV program priority. Using a realist synthesis approach, Study 3_Global identified three program theories, consisting of Context-Mechanism-Outcome Configurations of Option B+ PMTCT in resource-limited settings. The program contexts identified are similar to the factors identified in Studies 1 and 2. Mechanisms identified in Study 3_Global included: women’s willingness, readiness, and motivation to participate in Option B+ PMTCT; and health workers and women’s confidence in the PMTCT program. These mechanisms are triggered by enabling program contexts to achieve the desired outcomes. Conclusions This research contributes to a greater understanding of Option B+ PMTCT in Papua by describing program performance using local data and identifying factors that influenced women’s uptake, adherence, and retention. Findings suggested an improvement of women’s participation through the creation of enabling program contexts. Recommendations include: provision of high-quality HIV counselling to increase women’s belief in ART efficacy or to gain partner support, reduction of community stigma and discrimination through campaigns, training to improve health workers’ skills to deliver the program, availability of clear guidelines to ensure high quality PMTCT services are delivered, and continuation of free HIV services to reduce financial burden. Future studies are warranted to investigate cultural factors related to pregnancy care in Papua, and to identify mechanisms that are likely to be more influential to increase women’s uptake, adherence, and retention in Option B+ PMTCT in Papua.