Disturbances in arterial blood flow and endothelial shear stress (ESS) are associated with pathological processes underlying atherosclerosis and complications after stent placement. Image-based computational fluid dynamic (CFD) simulations allow in vivo estimation of ESS and other indices of flow disturbances. While low ESS is a relatively sensitive marker of future plaque progression, it remains too non-specific for clinical application. Possible improvements include incorporation of more realistic simulation methodologies. Although there are many possible ways to improve or change simulation accuracy, such as the choice of imaging, reconstruction techniques, boundary conditions, incorporation of arterial wall compliance, the primary aim of this thesis was to evaluate the effect of non-Newtonian rheology on ESS calculation. Although blood is a non-Newtonian fluid, most CFD studies assume blood to be a Newtonian fluid with constant viscosity. As opposed to the Newtonian model, non-Newtonian rheological models treat blood viscosity as a variable solved during CFD simulation. As a result, the non-Newtonian assumption offers two hypothetical advantages over the Newtonian model: 1) improved accuracy in calculation of traditional haemodynamic indices; 2) novel viscosity-based indices of blood flow disturbances are made available, and which may correlate with atherosclerosis. These primary hypotheses are investigated in this thesis by using CFD in combination with high-resolution optical coherence tomographic (OCT) imaging of atherosclerotic plaques and stents in patients with coronary artery disease. This work began with a randomised controlled trial of 60 patients comparing two second-generation drug-eluting stents using OCT imaging immediately after implantation and at 6 months (Chapter 8). Although there were no significant differences in the primary endpoint of stent malapposition, the platinum-chromium stent demonstrated a significantly higher incidence of late longitudinal deformation without concurrent events during 12-month clinical follow up. Next, non-Newtonian CFD simulation was performed in 7 patients who received a fully bioabsorbable coronary scaffold and OCT imaging immediately after implantation and 5 years later (Chapter 9). Low ESS between scaffold struts after implantation significantly improved by 5 years, and the overall ESS distribution narrowed to more normal physiologic levels associated with vascular quiescence. Up to 10-fold increases in blood viscosity were identified near scaffold struts, but peak viscosity in the scaffolded segment significantly decreased by 5 years. Comparison of CFD results using Newtonian versus non-Newtonian rheological models was then undertaken in 16 patients who had non-culprit plaques completely imaged in baseline and 6-month OCT imaging. By purely quantitative comparison of rheological models, the Newtonian model significantly underestimates ESS, resulting in up to a 40% higher estimate of vessel areas exposed to atherogenic low ESS (Chapter 10). While the Newtonian and non-Newtonian models can lead to different conclusions about the relationship of ESS with underlying plaque composition, non-Newtonian indices local blood viscosity (LBV) and local Reynolds number (ReL) are significantly and independently associated with underlying calcium and lipid, respectively (Chapter 11). Further, vessel areas exposed to high ESS along with both high and low ReL demonstrate increases in lipid over 6 months, indicating the role of high inertial and viscous forces in lipid accumulation (Chapter 12). Finally, blood flow disturbances were evaluated in 18 patients with acute plaque erosion and thrombus (Chapter 13). High gradient of ESS and high ReL were significantly associated with the presence of thrombus, implying their role in acute coronary syndrome due to plaque erosion.

Frailty is a loss of physiological reserve that leaves individuals at risk of poor recovery when exposed to a stressor. Frailty has been identified as a risk factor for poor outcomes for older adults when they are admitted to hospital, although there are still some barriers to implementation of measurement tools. Some frail older adults will still make a good recovery. Health assets are factors that are associated with health and recovery and are also desirable in their own right. Inclusion of health assets in models of illness and recovery may improve prognostication and identify patient centred strategies to facilitate recovery. Aims of the PhD 1. Determine whether is feasible to measure frailty based on routine clinical assessment 2. Examine whether health assets can be identified in hospitalised older adults 3. Investigate whether individual health assets improve outcomes for hospitalised older adults 4. Develop a health assets index 5. Validate the health assets index in hospitalised older adults Methods In addressing these aims, five phases of research were undertaken: Phase 1: A systematic review of the literature was undertaken to identify health assets in the hospital setting. MEDLINE, EMBASE, CINAHL and PsycINFO were searched to identify studies examining outcomes for hospitalised older adults. Included studies examined at least one potential individual health asset, which was a psychosocial characteristic or health characteristic. Study quality was assessed, and findings are narratively described. Phase 2: A prospective cohort study was conducted in an acute general medical unit to determine whether frailty could be measured based on routine clinical information by junior medical staff. All patients aged 65 and over admitted to a general medical unit during August and September 2013 were eligible for the study. CFS score at baseline was documented by a member of the treating medical team. Demographic information and outcomes were obtained from medical records. The primary outcomes were functional decline and death within three months. Phase 3: A secondary analysis of an existing data set was conducted to examine the interaction between health assets and frailty. Patients of 1418 aged ≥ 70 years admitted to 11 hospitals in Australia were evaluated at admission using the interRAI assessment system for Acute Care, which surveys a large number of domains, including cognition, communication, mood and behaviour, activities of daily living, continence, nutrition, skin condition, falls and medical diagnosis. The data set was interrogated for potential health assets and a multiple logistic regression adjusted for frailty index, age and gender as covariates was performed for the outcomes mortality, length of stay, readmission and new need for residential care. Phase 4: Based on phases 2 and 3, a heath assets index was created. A pilot study was conducted to determine the feasibility of collecting this information in hospitalised older adults. Phase 5: A prospective cohort study was conducted to determine whether the health assets index had predictive validity for older inpatients. Adults aged 70 and older with unplanned admission to hospital were eligible to participate. Frailty and other co-variates were measured. The primary outcomes were mortality at 30 days and functional decline at discharge. Results Phase 1: Nine prospective cohort and two retrospective cohort studies were identified. subjective, functional and biological health assets were identified. Health assets were associated with decreased risk of post-hospital mortality, functional decline, new need for residential care and readmission. Phase 2: Frailty was assessed in 95 % of 179 eligible patients. 45 % of patients experienced functional decline and 11 % died within three months. 40 % of patients were classified as vulnerable/mildly frail, and 41 % were moderately to severely frail. When patients in residential care were excluded, increasing frailty was associated with functional decline (p = 0.011). Increasing frailty was associated with increasing mortality within three months (p = 0.012). Phase 3: Inpatient mortality was 3% and 4.5% of patients died within 28 days of discharge. Median length of stay was 7 days (IQR 4-11). In multivariate analysis that includes frailty, being able to walk further [OR 0.08 (0.01-0.63)], ability to leave the house [OR 0.35 (0.17-0.74)] and living alone [OR 0.28 (0.10-0.79)] were protective against mortality. The presence of a support person was associated with a decreased length of stay [OR 0.14 (0.08-0.25)]. Phase 4: It was feasible to measure health assets in older adults admitted to hospital. The time taken of 2-3 minutes indicated that it was not too onerous. Some questions were adjusted to make the wording clearer to participants. Phase 5: There were 298 participants, with an average age of 84.7 and 66% were women. 80.1% had a frailty score of greater than 0.25, with a population mean score of 0.38 (SD 0.12). The mean HAI score was 10.86 (SD 2.87) with a minimum of 5.5 and a maximum of 15. 56.4% of participants had functional decline on discharge from hospital and there was 5.7% 30 day mortality. There was an inverse relationship between frailty and health assets. In a multivariate analysis that accounted for interaction, for those who were not frail, a higher number of health assets was associated with lower mortality. This relationship was reversed at higher levels of frailty. Conclusions It is possible to measure frailty using routine clinical information, but the time taken to enter data is likely to present an ongoing barrier to frailty measurement, which can be overcome with the use of an electronic medical record. Health assets can be identified in older adults who have been admitted to hospital. A higher number of health assets is associated with a decreased level of frailty. Health assets may confer protection against mortality in more robust older adults. Further research could help to elucidate strategies that older adults identify as important and how these can be applied in the hospital setting.