Background When a genetic cause is suspected in a person with dementia, it creates unique diagnostic and management challenges to the clinicians. Even in specialised neurogenetics clinics, these patients and their families face significant delays in diagnosis. The main cause for this delay is the incremental nature of the genetic tests performed in these clinics. This is due to the lack of a single, comprehensive genetic test that could identify the different types of genetic variants associated with dementia. Another problem in dementia genetics is the inability to predict the risk of Alzheimer disease (AD) in people with the APOE e4/e4 genotype, despite it being a well-known high-risk genotype. The diversity in cognition among APOE e4 homozygotes can range from early-onset AD to a lifetime with no symptoms. Identifying the modifiers of such variation will enable further refinement of risk prediction for people with this genotype. This thesis describes two main studies. First is a study that explores the clinical impact of whole genome sequencing (WGS) in patients with early onset dementia (EOD). The second study investigates the modifying effect of genomic factors, including a polygenic risk score (PRS), between APOE e4/e4 participants with EOD (onset <65 years) and elderly (aged >75 years) who have escaped AD despite their high-risk genotype and advanced age. Methods WGS analysis was performed in 50 patients with EOD, exploring point mutations, small insertions and deletions using a targeted panel of 117 genes, followed by a “Mendeliome” approach with a broad range of Mendelian disease genes, not included in the targeted panel. Structural variants (SV) as well as short tandem repeats (STR) were also analysed. A polygenic hazard score (PHS) was calculated in patients with AD. In order to examine the APOE e4/e4 genomic modifiers, first, a systematic review was conducted to search the literature for evidence of any genetic modifiers already published. As no clear answers could be deduced from the literature, a novel phenotypic extremes study was designed to examine the modifying effect of a PRS between cognitively healthy APOE e4 homozygotes aged >/=75 years (n=213) and early-onset APOE e4 homozygote AD cases aged </=65 years (n=223). Results Through the WGS study of EOD patients, clinically diagnostic pathogenic variants were identified in 7/50 (14%) of the patients, with a further eight patients (16%) found to have established risk factors which may have contributed to their EOD. All relevant variants were in the targeted set of genes with no additional findings through the Mendeliome approach. Two of the clinically diagnostic variants were identified through SV analysis. No pathogenic expanded STRs were found in this study cohort, but a blinded analysis with a positive control identified a C9orf72 expansion accurately. PHS results in AD patients indicated that ~47% (9/19) had a PHS equivalent to >90th percentile risk. Although fifteen studies met the inclusion criteria for the systematic review of APOE e4/e4 modifiers, only two SNPs (CASP7 rs10553596 and SERPINA3 rs4934-A/A) had sufficient evidence towards risk modification/resilience in APOE e4 homozygote participants. The CASP7 rs10553596 deletion variant was not captured in the genotyping platforms used in the phenotypic extremes participants and could not be imputed reliably and therefore was not analysed. The frequency of the SERPINA3 rs4934-A/A variant was not different between the extreme phenotypic cases and controls. The phenotypic extremes PRS study demonstrated that the PRS for AD was significantly higher in APOE e4 homozygote AD cases compared with older cognitively healthy APOE e4/e4 controls (OR 8.39; CI 2.0 to 35.2; p=0.003). Conclusions The clinical whole genome study showed that WGS can act as a single genetic test to identify different types of clinically relevant genetic variations in patients with EOD. WGS, if used as a first-line clinical diagnostic test, has the potential to increase the diagnostic yield and reduce time to diagnosis for EOD. The APOE e4/e4 extremes phenotyping study has identified a risk modifying PRS, comprised of common AD associated variants, that could explain the variability in clinical presentation in people with APOE e4/e4. Larger studies are needed to further validate both the WGS and modifying PRS results. Moreover, qualitative studies and cost-effectiveness analysis are needed to understand how these results would impact patients with dementia in the “real-world” setting.

This thesis primarily constitutes a compendium of published manuscripts encapsulating a substantial volume of work relating to prion diseases, undertaken from 1999 to 2020 through the Australian National Creutzfeldt-Jakob Disease Registry based at the University of Melbourne, frequently in collaboration with pre-eminent domestic and international scholars in the field of prion diseases. The thesis is divided into two principal parts. The first comprises 78 peer-reviewed papers submitted in full, representing a diverse array of acclaimed, often landmark or pivotal, studies published in prestigious general medical and specialist journals where I have been either senior and corresponding author, primary author or leader of the Australian contribution in major multi-national projects. This body of work is divided into seven sections spanning: normal prion protein cellular biology; prion pathogenesis; human prion disease epidemiological studies; human prion disease bio-marker studies; clinical aspects of human prion disease; prion disease treatment studies; and studies of human prion-like diseases. A short review at the start of the thesis and a preamble before each of these sections dealing with my principal publications are offered to place these published studies in the context of extant clinical and scientific literature. The second part of the thesis is submitted for completeness, comprising four sub-sections encompassing 61 publications and reports, listed by title and publication details, dealing with human prion disease for which I was a significant contributor.

ABSTRACT Introduction This is the first Australian longitudinal retrospective study on the patients with traumatic Spinal Cord Injury (SCI), over a 14-year period, showing the trends of disease as well as health service utilization. This study involved linking Intensive Care Unit (ICU) datasets with Victorian Admitted Episodes Dataset (VAED) and Victorian Emergency Minimum Dataset (VEMD) from three major hospitals to work out determinants for disability, as well as co-morbidities. From the collected data, the study provides frameworks for chronic disease such as SCI and developed a prototype that can be used to bring the information together to be utilized by clinicians, patients and carers to improve health outcomes. Methods Health administrative datasets with the International Classification of Diseases, 10th revision, Australian Modification (ICD-10-AM) were used to conduct a pilot data linkage study using a unique identifier. A deterministic linkage method was used to link internal datasets from each of three major hospitals to build patient disease profiles and identify comorbidities. An extended Elixhauser comorbidity index (ECI) was used to study risk factors and comorbidities. Results The study identified a lack of coordination between clinical, administrative and statutory data custodians, and issues with coding quality. From the linked datasets from three major hospitals (Alfred Hospital, Royal Melbourne Hospital and Austin Hospital) data on almost 2,629 patients were extracted. The female: male ratio in this cohort was 1:2.9 and the largest proportion of patients was aged between 16 and 30 years. An increase in female admissions was apparent in the last decade. Sixty-six percent of readmissions were for patients from the Melbourne metropolitan area, with much lower proportions in the Gippsland and Hume regions. The three most frequent principal diagnoses were functional level of cervical spinal cord injury (C1-C8), concussion and oedema of thoracic spinal cord and concussion and oedema of cervical spinal cord. The main reasons for readmission were urinary tract infection (UTI), pressure ulcer, mental disorders and respiratory infections. The study of risk factors (alcohol, tobacco and illicit drugs) showed a significant association with overall length of stay in ICU and that males had twice the risk of death than females. Using the linked datasets as backend, a prototype of clinical decision support system (CDSS) was developed. This has scalability and can be improved with the latest technologies such as an ‘alert system’, as well as built-in artificial intelligence (AI) such as an Artificial Neural Network, which can assist clinicians, patients and carers. Conclusion This is the first longitudinal study of SCI, following the e-journey of patients with SCI over 14 years in three major hospitals in Victoria. The lack of coordination between clinical, administrative and statutory data custodians, and issues with coding quality have implications for resource allocation, decision making and planning by health administrators and clinicians. Although the total number of people with spinal cord injury is small, they have prolonged health utilization. This Clinical Data Linkage study has provided unique information about these patients, including the enormous number of readmissions, the reasons for readmission, the exact cost of care at the major Victorian hospitals rather than estimates, and the area of the residence of patients where ongoing care is needed. Ultimately, stratified patient profiles can be used as a backbone for eHealth and a as framework for clinical decision support systems that are known to support self-efficacy for patients with chronic conditions and to improve health outcomes. They may also be used to build a conceptual model for other chronic conditions that have a high number of medical interventions.

Atrial fibrillation (AF) is an extremely common clinical problem with increasing global prevalence. Besides its frequent association with stroke, heart failure and increased mortality, recent data have shown the significant quality of life impairment and psychological distress that results from this arrhythmia. The clinical spectrum of AF can manifest as brief episodic paroxysmal AF or as a sustained arrhythmia in persistent AF in patients with progressive atrial disease. Mechanistically, paroxysmal AF is often triggered by rapidly firing impulses that originate in the pulmonary veins, allowing catheter-mediated elimination of sources, with clinical success rates of 70–85%. However, in persistent AF, the mechanism that sustains the arrhythmia remains incompletely understood and is a topic of ongoing debate. Recent advances in cardiac mapping and computational methods have suggested localised drivers and non-pulmonary vein triggers particularly from the left atrial appendage but there is also accumulating evidence that the atrium frequently functions electrically as a 3-dimensional structure and there exists endocardial-epicardial dissociation in activation during AF. The aims of this thesis are three-fold: Firstly, we review the evidence for the current and expanding indications for catheter ablation in AF and highlight some of the novel tools and technological advancements that have emerged in the recent years for achieving durable pulmonary vein isolation (PVI). Secondly, we investigate the role of some of the novel mechanisms that potentially sustain persistent AF. We addressed this first by systematically reviewing the evidence for a computational mapping technique thought to identify localised sources and then performed a series of cardiac mapping studies in humans. These high-density mapping studies characterised atrial endocardial-epicardial electrical dissociation in the presence of structural heart disease and explored the mechanistic role of localised sources in the left atrium and the left atrial appendage (LAA) in persistent AF. Finally, given the emerging evidence for the mental health effects from AF, we discuss the rationale and methodology of a randomised controlled study comparing catheter ablation with medical therapy on psychological distress and neurocognitive function in patients with AF. Chapter 1 summarizes several aspects of AF including the epidemiology, our current understanding of the classic and novel mechanisms of AF the mechanistic role of risk factors and their implications in remodelling and atrial cardiomyopathy. Chapter 2 reviews the role of catheter ablation in AF and highlights the recent technological advancements. Catheter ablation is a safe and effective rhythm control strategy for symptomatic patients who failed medical therapy or who prefer not to take medications and there is emerging evidence for its role in mortality reduction in AF patients with heart failure. Pulmonary vein isolation is the cornerstone approach and both radiofrequency and cryoablation have similar efficacy. Mounting evidence demonstrates the importance of risk factor management for improving ablation outcomes. In the era of high-density cardiac mapping, FIRM (focal impulse and rotor modulation) emerged as a novel computational mapping technique to identify rotors and focal sources that could potentially be targeted during catheter ablation. Despite early promising results, many studies that followed showed mixed outcomes and indeed, some others showed a pro-arrhythmic consequence. Chapter 3 presents a systematic review and meta-analysis of 11 observational studies and demonstrates the wide variability in the medium-term outcomes of FIRM guided ablation and explores the significant heterogeneity between published studies. Chapters 4 and 5 examine the characteristics of endocardial-epicardial dissociation (EED) in patients with structural heart disease. We performed high-density simultaneous endocardial-epicardial phase mapping of the right atrium in patients undergoing cardiac surgery to study this. In Chapter 3 we report the for the first time, functional EED based on observations of synchronous activation during sinus rhythm and EED in activation timing and wavefront propagation during pacing drive and premature extra-stimulation providing compelling evidence for the functional nature of the atrial substrate. Chapter 4 presents data of phase mapping prolonged persistent AF recordings. The results provide novel evidence for endocardial-epicardial wave front mismatch in AF along with marked EED with temporal heterogeneity. In Chapter 6 we sought to characterise the preferential 3-dimensional nature of sinoatrial conduction in humans using simultaneous endocardial-epicardial mapping of the sinus node region. In intact hearts of patients with structural heart disease, data confirmed the presence of multiple differential endocardial and epicardial sino-atrial exits and hence the redundant structure of the pacemaker complex. This is consistent with data from optical mapping of ex-vivo human hearts and demonstrates that clinical sinus node dysfunction only occurs in the setting of advanced atrial structural remodelling. Recently, data from cohort studies and a randomised controlled trial have shown that LAA isolation improves ablation outcomes in patients undergoing redo ablation for persistent AF. However, there have been concerning reports that such empirical ablation is associated with a heightened risk of LAA thrombus, even in patients who are anticoagulated. Furthermore, data from mapping studies have also shown mixed results on the potential role of LAA as a driver in persistent AF. In Chapter 7 we examine the role of localised sources in the left atrium particularly in the LAA by performing regional high-density mapping of persistent AF. In addition to finding infrequent drivers in the left atrium, this project also showed paucity of triggers from the LAA providing further evidence of its passive role rather than an active driver in persistent AF. Besides physical symptoms that patients with AF experience, recent data has shown the enormity of mental health effects that AF can have that is often underappreciated by clinicians. More importantly, preliminary data from observational studies show the benefits of catheter ablation to improve mental health. Chapters 8 and 9 present the methodology and rationale of a multicentre, randomised controlled trial that assesses the impact of catheter ablation on psychological distress and neurocognitive function in patients with AF. The study has completed recruitment and successfully enrolled 100 participants across the Royal Melbourne and Alfred Hospitals and completion of analysis is expected by March 2022. Chapters 10 and 11 conclude by summarizing the key findings of the studies and their clinical implications. Further, it paves the way for future work that might progress our understanding of AF, especially in light of novel mechanisms

Stroke is one of the most important causes of acquired epilepsy in adults. Patients with seizures after stroke have higher mortality and disability than those without seizures. There is a building body of evidence to suggest that post-stroke epilepsy may be mediated by dysregulation of glutamate homeostasis, given the central role of glutamate in the pathogenesis of both stroke and seizures. There is currently no evidence to support the use of antiseizure medications as primary prevention of epilepsy in stroke patients. Ideally, anti-epileptogenic treatment would be targeted at those stroke patients at highest risk of epilepsy according to established biomarkers – studying glutamate as one such biomarker has been limited by a reliance on invasive procedures including lumbar puncture to quantify concentrations of glutamate in the brain. This thesis focuses on the role of 7T MRI as a method of brain glutamate quantification in the setting of stroke. The research has been conducted in the Department of Neurology, Royal Melbourne Hospital, the Department of Medicine (The Royal Melbourne Hospital), University of Melbourne, and the Melbourne Node of the National Imaging Facility, Department of Radiology, University of Melbourne. 22 patients with acute ischaemic or haemorrhagic stroke were recruited from the inpatient stroke unit at Royal Melbourne Hospital, with 7T MRI scans performed at the Melbourne Brain Centre Imaging Unit, University of Melbourne. In addition, peripheral blood samples were collected and underwent metabolomics analysis for plasma glutamate quantification at the Monash Institute of Pharmaceutical Sciences, Parkville. Across the patient cohort, glutamate concentration was lower in the region of infarction than in the corresponding hemisphere, when measured by Magnetic Resonance Spectroscopy. When measured by glutamate weighted chemical exchange saturation transfer imaging (GluCEST), the results were more heterogeneous, ranging from decreased to increased, ipsilateral to infarction. One patient in the cohort developed post-stroke epilepsy, with a GluCEST profile similar to the population overall. A single haemorrhagic stroke patient suffered a seizure prior to scan acquisition, with a pattern of increased cortical GluCEST contrast consistent with a post-seizure effect. The second part of the thesis focuses on a study protocol examining the anti-epileptogenic potential of the glutamate receptor antagonist and antiseizure medication perampanel in a population at high risk of post-stroke epilepsy. This involves a collaboration of clinicians at four Melbourne Hospitals (Alfred Hospital, Royal Melbourne Hospital, Monash Medical Centre, Austin Hospital), led by the candidate. Finally, there is evidence that patients with epilepsy have an increased risk of developing cerebrovascular or cardiovascular disease, although it is unclear whether this is due primarily to the epilepsy itself, or non-epilepsy factors such as antiseizure medications. The third part of the thesis comprises a data linkage study in the Department of Neurology, Royal Melbourne Hospital, based on medical records from a database of admitted video EEG monitoring patients from 1995 to 2015. It was found that the incidence of new-onset cerebrovascular disease was higher in epilepsy patients compared with the general Victorian population, although there was no difference in the composite incidence of cerebrovascular disease, cardiovascular disease and peripheral vascular between epilepsy and non-epilepsy patients in the cohort. Furthermore, patients taking treatment with valproic acid were at lower risk than both those taking enzyme-inducing antiseizure medications and those taking neither valproic acid nor enzyme-inducing antiseizure medications. Collectively, these results emphasise the role of non-epilepsy factors such as social determinants of health, medical comorbidity, and epilepsy treatment, in influencing vascular risk.

Ventricular arrhythmias (VA) and sudden cardiac death (SCD) are responsible for one-half of all deaths in patients with heart disease. It is estimated that up to 40% of SCD is attributed to VA, highlighting a clear need for additional strategies to reduce this figure. Despite the systematic and escalated use of antiarrhythmic drugs (AADs) and implantable cardioverter defibrillators (ICD), VA incidence continues to rise. Catheter ablation (CA) has been shown to be a superior treatment for VA compared to AAD therapy in regards to reducing ICD shocks and ventricular tachycardia (VT) storm. However, a mortality benefit has not been demonstrated in randomized controlled trials. Advancements in catheter and electroanatomical mapping technology has led to new frontiers in delineating VT mechanisms and substrate characterisation translating to improvements to targeted ablation strategies. The aim of this thesis was to expand on the diagnostic, mechanistic, substrate characterisation and ablative techniques used for VT in the presence and absence of structural heart disease (SHD). Initially, we evaluate pre-procedural prediction of idiopathic outflow tract VAs using the surface electrocardiogram (ECG). We then perform studies involving scar-related VT. These include assessing the current evidence for the role of CA in VT ablation including assessment of Australian procedural trends, exploration of VT mechanisms and the effect of wavefront directionality on delineating scar substrate in an ovine-infarct model. Finally, we assess procedural characteristics and outcomes of implementing simultaneous high output pacing during substrate-based CA as an expedited lesion endpoint. Chapter 1 outlines our understanding of VT and ventricular fibrillation (VF) mechanisms. We discuss VT in the absence of the SHD (idiopathic VT) with particular emphasis on morphological ECG ‘signatures’ and prediction algorithms used differentiate outflow tract premature ventricular complexes (PVC) and VT (OTVT) sites of origin (SOO). We then explore VT in the setting of SHD providing a summary of current techniques used to characterise arrhythmogenic substrate, high density mapping and use of CA in both ischaemic and non-ischaemic cardiomyopathies and use of novel ablative techniques. Chapter 2 prospectively evaluates a modified high-precordial ECG configuration compared to standard ECG positions in 50 patients with OTVT. The study is divided into 3 phases assessing the anatomical relationship of the modified leads to the outflow tracts using multiplanar computer tomography (CT) followed by an algorithm development phase and validation review. We develop an algorithm that differentiated the right ventricular outflow tract and left ventricular outflow tract PVCs and VT with high accuracy – the modified lead R-wave deflection interval (RWDI). The RWDI was superior to all previously developed algorithms and anatomically closer to the outflow tracts compared to standard precordial lead positions. Chapter 3 assess the impact of CA as compared to medical therapy on outcomes in patients with VT and SHD. We perform a systematic review and meta-analysis of 8 randomized controlled trials (RCTs) enrolling 797 patients with predominantly ischaemic scar-related VT reporting VT recurrence and mortality. We compared RCTs to 4 contemporary, large observational studies including 3065 patients. We conclude CA is superior to medical therapy with significant reductions in VT recurrence and electrical VT storm, however, there was no difference in terms of all-cause or cardiac-specific mortality. The proportion of patients with non-ischaemic cardiomyopathy (NICM) and electrical storm (ES) was much higher in the observational studies; however, despite this, there was further reductions in VT recurrence and mortality compared to RCT meta-analysis. Chapter 4 analyses procedural trends of VT ablation procedures in Australia using two major data sources (Australian Institute of Health, Wealth and Aging and Medicare Australia) over a 10-year period. We calculated population-adjusted procedures per year and applied a regression model to calculate percentage change per year. We report growth of VT ablations have surpassed atrial fibrillation (AF) and percutaneous coronary intervention (PCI) procedures and discuss possible explanations for VT ablation growth and health care implications. Chapter 5 examines procedural characteristics and outcomes in patients undergoing ventricular assist device (VAD)-related VT ablation. We performed a systematic review of 18 studies including 110 patients demonstrating that scar-related re-entrant VT was the predominant mechanism in >90% cases followed by cannula-adjacent VT in approximately 20% of VTs. Furthermore, the haemodynamic tolerability of the VAD allowed an activation and entrainment mapping strategy as compared to a substrate-based approach. Perceived procedural difficulties during CA of VT were rare and complications comparable to a non-VAD cohort; however, need for cardiac transplantation and mortality were expectedly high underscoring the severe underlying disease in this group. Chapter 6 focuses on the prevalence of a primary or co-existent focal mechanism in patients with SHD who historically are exclusively thought to have scar-related reentry. Using an extended VT induction protocol incorporating catecholamine stimulation and right ventricle (RV) burst pacing in 112 patients over a 2-year period, we show 16% had a focal mechanism elucidated. Focal VTs commonly co-existed with re-entrant VTs and typically localised to scar borderzone or remote regions usually attributed to “idiopathic” locations. We found a high rate of failure of device therapies including repetitive initiation which was unique to this group with CA a highly effective treatment for durable outcomes. Chapter 7 examines the effect of directional influence of substrate characterisation using a multielectrode catheter (HD Grid) in a post infarct ovine model using proprietary software which selects the highest voltage from orthogonal bipolar pairs (HD wave solution). We compare this catheter to a validated linear duodecapolar catheter. We conclude that varying the wavefront of activation results in significant differences in the percentage of scar and abnormal electrograms (EGM) using the HD grid catheter, despite the HD wave solution. The HD grid demonstrated a higher proportion of normal voltage compared to the duodecapolar catheter, across all activation wavefronts. Compared to the duodecapolar catheter, we show abnormal EGMs, specifically local abnormal ventricular activations (LAVA) were significantly less using the HD grid catheter. Chapter 8 evaluates simultaneous high-output pacing and ablation adopting a substrate-based scar homogenisation strategy to achieve scar inexcitability for lesion endpoints. We assess the procedural characteristics and ablation parameters in 10 patients and show this technique is safe and effective with high acute procedural success for VT non-inducibility with no recurrence or repeat procedure in short-term (6 month) follow-up.

Abstract/Overview Australia has benefited from a long history of migration, including accepting refugees and asylum seekers (referred to as refugees within this discussion) from a wide range of countries experiencing conflict and hardship. Multiple health care challenges exist for refugees, including language difficulties, low health literacy, poor mental health and exposure to infectious diseases that may be unfamiliar to Australian health care practitioners. This is exacerbated by the growing numbers of immigrants and refugees settling in regional areas, where there are poorer health outcomes and provision of quality health care is more difficult. The published papers included in this thesis outline a number of the challenges in the provision of high-quality health care to refugees who settle in regional Australia, and some solutions that have been successfully utilised to improve health care provision to this population. The demonstrated solutions used for refugee populations are also valid for other immigrants and locally born Australians living in regional areas. The literature review provides the context and rationale for each of these papers with more specific references included within each paper. The literature review provides an overview of the topic and outlines the importance of the papers in addressing issues previously poorly understood or not considered. Paper one investigates the challenges of identifying appropriate screening for newly arrived immigrants and focuses on the cost effectiveness of screening new arrivals for H pylori. Modelling outlines the situations where a screening program could be considered and highlights the uncertainties that exist within the assumptions used for this modelling. Expanding on specific infectious disease challenges in immigrant populations, paper two examines the genotypes of hepatitis B that are encountered in immigrants from Burma and paper three describes genotypes amongst African immigrants. Papers two and three address the clinical relevance of genotype and country of birth in the management of hepatitis B. The subsequent three papers examine the provision of health care to refugees and immigrants via telehealth. Paper four outlines the development of a tertiary hospital based Infectious Diseases telehealth program. Paper five measures and compares the effectiveness of healthcare delivered to regional areas via telehealth, with a focus on care for individuals with hepatitis C virus. Paper six then describes the extension of the use of telehealth to include the provision of interpreters. The discussion outlines the conclusions that have been drawn from these papers, and then suggests further improvements for the care of refugees and immigrants who settle in regional areas. Reducing the gap in health care outcomes between those who live in large urban centres and those who live in regional areas is relevant to the whole population. Drawing on these findings, recommendations are outlined to reduce this gap.

Multiple sclerosis is a heterogenous neurological disorder characterised by an interplay between neuroinflammation and neurodegeneration. Treatment decisions are frequently prompted by the presence of disease activity presenting as relapses or progression-of-disability events. Treatment initiation and switch are therefore two critical periods in multiple sclerosis management. There is a delay between treatment initiation and the onset of full clinically manifest effect of therapy, referred to as ‘therapeutic lag’. It has been suggested that not accounting for delayed treatment effects may obscure treatment response in clinical trials of progressive multiple sclerosis. Further exploration of the influence of therapeutic lag and its determinants, however depends on development of a robust method to detect when treatments attain full clinical effect. Similar to therapeutic lag, there is a delay between treatment stop and the return of disease activity. Disease reactivation after treatment cessation has been identified as a risk factor for disability accrual. Understanding of the optimal timing of the start of the next treatment however remains limited. All studies used data from the two largest multiple sclerosis registries: MSBase (multinational) and OFSEP (French). The first study developed, and externally validated, an objective differential-calculus based method to estimate the duration of therapeutic lag for clinical measures of treatment effect. Time to full clinical treatment effect was estimated as 12-30 weeks for relapses and 30-70 weeks for disability progression. The second study applied this method to groups of patients with different clinical and disease characteristics to investigate the determinants of lag duration. Pre-treatment level of disability and annualised rate of relapse were identified as the most important influencers of therapeutic lag. The third study investigated the effect of high and low-efficacy therapy in secondary progressive multiple sclerosis, after accounting for therapeutic lag. High-efficacy therapy was found to be superior to low-efficacy therapy in reducing relapses in patients with episodic inflammatory activity in the two years before starting therapy. High- and low-efficacy therapies however had comparable effectiveness on disability progression. In the fourth study, the focus was shifted from treatment start to the period after treatment cessation. The rate of disease reactivation after the cessation of disease-modifying therapies was explored. Risks of disease reactivation were highest after cessation of anti-trafficking therapies and were reduced by starting a new treatment. On-treatment rates of relapse, age, sex, and level of disability further influenced the risk of disease reactivation. These findings are highly relevant to clinical decision-making in a number of contexts: (i) patients who experience early on-treatment disease activity should be re-evaluated after the lapse of therapeutic lag, (ii) the selection of immunotherapy in patients with active secondary progressive multiple sclerosis should carefully consider the amount of preceding relapsing activity and the risk-vs-benefit ratio, (iii) decisions regarding wash-out durations after discontinuation of disease-modifying therapies should be individualised based on the discontinued therapy and a patient’s clinical profile. Moreover, treatment outcomes in cohorts enriched with patients with higher disability scores should be interpreted with the expected duration of therapeutic lag in sight.

Multiple sclerosis (MS) is a chronic inflammatory, demyelinating and neurodegenerative disease of the central nervous system. It is one of the most common causes of disability in young adults and represents a significant personal and societal burden. The aetiology of MS is unknown but involves both environmental and genetic risk factors. Women are three times more likely to develop relapsing-remitting MS (RRMS) than men, and sex hormones are hypothesised to be protective in MS. Further understanding the role of these risk factors will advance our knowledge of the aetiology of MS and lead to improved counselling. MS is also a heterogeneous disease and there is currently no good measure of predicting which patients are at highest risk of disability accrual. Brain atrophy is one potential biomarker of neurodegeneration but there are a paucity of studies exploring its application in the real-world clinical setting. This thesis has two parts. In the first part, we selected one modifiable risk factor, pregnancy, to further explore its association with MS. In chapter 1, we reported on the epidemiology of pregnancy in women with MS over the last 15 years using the international MSBase registry. Of 9098 women with RRMS, we found a low but increasing incidence of pregnancy as well as a greater proportion of women who conceived on disease-modifying therapy (DMT) over time. We did not find any differences in pregnancy outcomes between those pregnancies conceived on or off therapy. A higher disability was associated with a reduced pregnancy incidence. In chapter 2, we investigated the association of pregnancy with time to onset of clinically isolated syndrome (CIS), the first clinical demyelinating episode of MS. We recruited 2557 women across 4 sites and found that women with previous pregnancies and childbirths had a later onset of CIS compared to women without pregnancies and childbirths. A higher number of pregnancies and childbirths was not associated with a delayed CIS onset. In the second part, we selected one quantifiable outcome, MRI brain volumetry, and modelled its association with MS disability in a real-world setting. We used an automated volumetry software, icobrain, for centralised analysis of clinical MRIs. In chapter 3, we assessed the variability of percentage brain volume change (PBVC) measurements in a multicentre clinical setting. 6829 brain MRIs acquired on different scanners across 4 sites were analysed by icobrain. We found that after applying strict selection criteria for MRI pairs, which included filtering by image quality and an alignment similarity above the threshold for same-scanner scan-rescan images, only 42% of the original number of consecutive MRI pairs remained. The final MRI pairs had a mean PBVC comparable to single-site centres (-0.26% +/- 0.34). In chapter 4, we determined the association between various MRI metrics and MS disability using the subgroup of MRI pairs with the lowest variance from chapter 3. Our cohort of 260 relapse-onset MS patients were almost all treated, had a mean PBVC in the range for healthy individuals (-0.26% +/- 0.52), and a low disability accrue. We found no association between PBVC and future disability, however, cross-sectional whole brain volume was associated with disability. Our study findings highlight that pregnancy plays an important immunomodulatory role in women predisposed to MS, and whole brain volume may be the best performing MRI outcome to predict future disability in patients with minimal brain atrophy.

This thesis advances the use of clinician-performed ultrasound for two common respiratory presentations; acute respiratory failure, and chronic obstructive pulmonary disease. Ultrasound in acute respiratory failure (ARF) To begin, a narrative review of ultrasound in ARF was undertaken and identified a paucity of evidence for the efficacy of ultrasound in ARF beyond diagnostic accuracy, and consequently a need for studies to examine its impact on clinical management efficacy, and patient and societal outcome efficacy. A prospective study of ultrasound in ARF was undertaken to establish clinical management efficacy. Ultrasound yielded new or additional diagnoses in 34% and enabled multiple clinical diagnoses in individual patients to be rationalised to a single diagnosis in 69%. Clinician diagnostic confidence was increased in 44%. Following ultrasound, patient management was altered in 30%, most frequently in patients with multiple diagnoses on admission. Additionally, a protocol was developed for a randomised controlled trial of ultrasound in ARF. Set in a high-dependency respiratory unit, primary study outcomes will be (i) time to resolution of respiratory failure, representing patient outcome efficacy, and (ii) time to readiness for unit discharge, representing societal outcome efficacy. Ultrasound in chronic obstructive pulmonary disease (COPD) A systematic review of diaphragm and intercostal muscle imaging in COPD was undertaken and included 52 studies. In COPD alterations in diaphragm morphology and function were detected across different imaging modalities, with reduced diaphragm excursion correlated with greater physiological impairment, and diaphragm dysfunction in ICU settings associated with poor outcomes. CT-intercostal muscle area correlated to worse airflow obstruction. Lung volume reduction and physiotherapy were associated with improvements in diaphragm parameters. A prospective study was undertaken in COPD to examine parasternal intercostal muscle thickness and echogenicity in the 2nd and 3rd intercostal spaces by ultrasound. Inter- and intra-rater reliability for thickness was moderate-to-high. Inter-rater measurement of echogenicity was moderate, with moderate-to-high intra-rater reliability. Reduced parasternal intercostal thickness and increased echogenicity (indicative of higher fat content) were both correlated to worse airflow obstruction as measured by FEV1. Finally, parasternal intercostal muscle thickness and echogenicity was studied in COPD patients before and after endobronchial valve (EBV) insertion. There was no change in echogenicity, but there was a strong correlation between reductions in residual volume and increases in parasternal intercostal muscle thickness in the treated hemi-thorax, suggesting length-tension relationships are a key determinant of muscle thickness. Ultrasound is a useful technique in patients presenting with acute respiratory failure and assists clinician decision-making and confidence. Ultrasound measured changes in parasternal intercostal muscles correlate with severity of airflow obstruction in COPD, with strong correlations between change in residual volume and thickness following relief of hyperinflation.

The rising burden of chronic disease in the developed world has resulted in an increasing accumulation of patients requiring long-term specialist input in care, despite a relatively stagnant specialist capacity in tertiary hospital services. Newer models of care, incorporating specialist input whilst empowering and enabling community-based treatment in a more cost-effective primary care setting are desirable. This thesis details the adaptation, implementation and evaluation of a new model of care that is underpinned by digital technology, using the HealthElink system, in facilitating community and prison-based treatment of chronic hepatitis C virus (HCV). This approach involved establishing outcomes for current treatment models, preliminary system functionality testing and prospective implementation of the eHealth model of care in both community and prison settings throughout Australia. Chronic hepatitis C virus represents a model disease in which rapid treatment advances have allowed care to primarily shift from hospital to community-based treatment. The advent and availability of direct-acting antiviral (DAA) therapy for chronic hepatitis C virus (HCV) in Australia in March 2016 represented a treatment revolution. In effect, HCV became the first chronic viral infection for which a safe, efficacious and permanent cure exists, affording the opportunity to facilitate global eradication programs. The World Health Organisation therefore issued a position paper, calling for the elimination of HCV as a major public health threat by 2030 and ending onward transmission through ‘treatment as prevention’ programs. Owing to the efficacy, ease and safety of DAA therapy, Australia became one of the first jurisdictions to allow any prescriber to use these lifesaving medications. Despite an estimated 80% of patients deemed suitable for treatment by a primary care practitioner, only 18% of prescriptions were from a GP in the first twelve-months. This formed the ideal opportunity to deploy and examine the eHealth model of care. This thesis is built upon the largest prospective clinical study of a novel shared eHealth model of care in Australia. Baseline studies in this thesis established the real-world outcomes of DAA treatment in Australia in the new era of treatment. Failure to test for HCV cure and loss to follow-up were higher than previously reported in controlled trials, particularly in community and prison settings compared to tertiary-based treatment. The eHealth model of care was usable, acceptable and more accurate in treatment selection than the current standard of care in preliminary functionality testing. A quasi-experimental, hybrid implementation-effectiveness study of the eHealth model in both community settings and prisons was undertaken showing similar clinical outcomes to the standard of care, with an improvement in guideline-based quality of care and efficiency. Uptake of the eHealth model was low by general practitioners, likely influenced in part by low HCV screening particularly in regional areas. Integrated hepatitis nurses exhibited high uptake of the system, contrasting with lower usability scores. The electronic patient portal exhibited low utilisation and may hold limited value in the current iteration due to both clinician and patient factors. The clinical decision support system was the most useful component of the eHealth model of care. Integration into existing electronic systems is seen as crucial to future electronic shared care models amongst clinicians. When applied to chronic hepatitis C, the eHealth model is usable and acceptable and demonstrates similar clinical outcomes to the current standard of care. Additional benefits in adherence to guideline-based care and efficiency of care were found. An integrated eHealth model of care for chronic disease, although currently nascent, holds potential to profoundly improve the delivery and quality of care.

Stroke is the second leading cause of death, as well as the second leading cause of disability worldwide. In Australia, the estimated financial burden of stroke on direct healthcare costs and healthy life lost is upwards of $50 billion Australian dollars annually. Highly effective reperfusion therapies exist to improve outcomes for ischaemic stroke but must generally be administered within the first few hours of stroke onset to ensure salvageable tissue is still present. Similarly, most haematoma expansion in intracerebral haemorrhage occurs early and promising interventions are being investigated to potentially reduce growth in this hyperacute period. The journey of a patient from stroke onset through to paramedic management, transport to hospital and eventual treatment is complex, with many workflow steps both in the pre-hospital and in-hospital phases that are susceptible to delays. This means that effective acute interventions in stroke may be delayed or may not be available to a proportion of patients. Historically, much of the effort for improving time to treatment has focussed largely on optimisation of in-hospital workflows. However, the pre-hospital phase may account for up to 50% of the time from stroke recognition to treatment. This is even greater if patients require a secondary inter-hospital transfer, should the initial centre not have capability to provide specialised treatments like endovascular thrombectomy or neurosurgery. This thesis explores two complementary methods of improving pre-hospital care of patients suffering from stroke. The first method is optimising ambulance triage of patients to allow direct transport to specialised stroke centres with advanced treatments. This is achieved through paramedic-led assessment of a clinical severity tool to determine patients likely to benefit from endovascular thrombectomy or neurosurgical services. This thesis studies the feasibility of such a system in metropolitan Melbourne, the design of a new triage tool adapted to local needs and validation of the tool in a real-world pre-hospital cohort of stroke patients. The second method is the use of an Australian-first mobile stroke unit, a custom-built ambulance with on-board computed tomography scanner and multidisciplinary stroke team, that allows pre-hospital assessment, imaging, treatment and triage of stroke patients. This thesis explores the operationalisation of the Melbourne Mobile Stroke Unit, the clinical benefits achieved by the service for treatment of both ischaemic and haemorrhagic stroke and the provision of novel therapies in the pre-hospital setting. Future adoption of the two synergistic pre-hospital strategies on a larger scale is expected to deliver substantial benefits to patients with stroke. Those eligible for time-critical treatment will be able to receive such therapies significantly faster, with expected improvements in longer term health outcomes. Improved pre-hospital management minimises the disadvantage experienced by patients located further away from appropriate stroke services, allowing some mitigation of healthcare inequality. Successful implementation in Melbourne will provide a template for roll-out of the strategies nationally and internationally.