Addressing issues in sparseness, ecological bias and formulation of the adjacency matrix in Bayesian spatio-temporal analysis of disease counts

The main objective of this PhD was to further develop Bayesian spatio-temporal models (specifically the Conditional Autoregressive (CAR) class of models), for the analysis of sparse disease outcomes such as birth defects. The motivation for the thesis arose from problems encountered when analyzing a large birth defect registry in New South Wales. The specific components and related research objectives of the thesis were developed from gaps in the literature on current formulations of the CAR model, and health service planning requirements. Data from a large probabilistically-linked database from 1990 to 2004, consisting of fields from two separate registries: the Birth Defect Registry (BDR) and Midwives Data Collection (MDC) were used in the analyses in this thesis. The main objective was split into smaller goals. The first goal was to determine how the specification of the neighbourhood weight matrix will affect the smoothing properties of the CAR model, and this is the focus of chapter 6. Secondly, I hoped to evaluate the usefulness of incorporating a zero-inflated Poisson (ZIP) component as well as a shared-component model in terms of modeling a sparse outcome, and this is carried out in chapter 7. The third goal was to identify optimal sampling and sample size schemes designed to select individual level data for a hybrid ecological spatial model, and this is done in chapter 8. Finally, I wanted to put together the earlier improvements to the CAR model, and along with demographic projections, provide forecasts for birth defects at the SLA level. Chapter 9 describes how this is done. For the first objective, I examined a series of neighbourhood weight matrices, and showed how smoothing the relative risk estimates according to similarity by an important covariate (i.e. maternal age) helped improve the model’s ability to recover the underlying risk, as compared to the traditional adjacency (specifically the Queen) method of applying weights. Next, to address the sparseness and excess zeros commonly encountered in the analysis of rare outcomes such as birth defects, I compared a few models, including an extension of the usual Poisson model to encompass excess zeros in the data. This was achieved via a mixture model, which also encompassed the shared component model to improve on the estimation of sparse counts through borrowing strength across a shared component (e.g. latent risk factor/s) with the referent outcome (caesarean section was used in this example). Using the Deviance Information Criteria (DIC), I showed how the proposed model performed better than the usual models, but only when both outcomes shared a strong spatial correlation. The next objective involved identifying the optimal sampling and sample size strategy for incorporating individual-level data with areal covariates in a hybrid study design. I performed extensive simulation studies, evaluating thirteen different sampling schemes along with variations in sample size. This was done in the context of an ecological regression model that incorporated spatial correlation in the outcomes, as well as accommodating both individual and areal measures of covariates. Using the Average Mean Squared Error (AMSE), I showed how a simple random sample of 20% of the SLAs, followed by selecting all cases in the SLAs chosen, along with an equal number of controls, provided the lowest AMSE. The final objective involved combining the improved spatio-temporal CAR model with population (i.e. women) forecasts, to provide 30-year annual estimates of birth defects at the Statistical Local Area (SLA) level in New South Wales, Australia. The projections were illustrated using sixteen different SLAs, representing the various areal measures of socio-economic status and remoteness. A sensitivity analysis of the assumptions used in the projection was also undertaken. By the end of the thesis, I will show how challenges in the spatial analysis of rare diseases such as birth defects can be addressed, by specifically formulating the neighbourhood weight matrix to smooth according to a key covariate (i.e. maternal age), incorporating a ZIP component to model excess zeros in outcomes and borrowing strength from a referent outcome (i.e. caesarean counts). An efficient strategy to sample individual-level data and sample size considerations for rare disease will also be presented. Finally, projections in birth defect categories at the SLA level will be made.

Optimal mission path planning (MPP) for an air sampling unmanned aerial system

This paper presents advanced optimization techniques for Mission Path Planning (MPP) of a UAS fitted with a spore trap to detect and monitor spores and plant pathogens. The UAV MPP aims to optimise the mission path planning search and monitoring of spores and plant pathogens that may allow the agricultural sector to be more competitive and more reliable. The UAV will be fitted with an air sampling or spore trap to detect and monitor spores and plant pathogens in remote areas not accessible to current stationary monitor methods. The optimal paths are computed using a Multi-Objective Evolutionary Algorithms (MOEAs). Two types of multi-objective optimisers are compared; the MOEA Non-dominated Sorting Genetic Algorithms II (NSGA-II) and Hybrid Game are implemented to produce a set of optimal collision-free trajectories in three-dimensional environment. The trajectories on a three-dimension terrain, which are generated off-line, are collision-free and are represented by using Bézier spline curves from start position to target and then target to start position or different position with altitude constraints. The efficiency of the two optimization methods is compared in terms of computational cost and design quality. Numerical results show the benefits of coupling a Hybrid-Game strategy to a MOEA for MPP tasks. The reduction of numerical cost is an important point as the faster the algorithm converges the better the algorithms is for an off-line design and for future on-line decisions of the UAV.

Clustering with random indexing K-tree and XML structure

This paper describes the approach taken to the clustering task at INEX 2009 by a group at the Queensland University of Technology. The Random Indexing (RI) K-tree has been used with a representation that is based on the semantic markup available in the INEX 2009 Wikipedia collection. The RI K-tree is a scalable approach to clustering large document collections. This approach has produced quality clustering when evaluated using two different methodologies.

QoS-oriented sesource allocation and scheduling of multiple composite web services in a hybrid cloud using a random-key genetic algorithm

In cloud computing resource allocation and scheduling of multiple composite web services is an important challenge. This is especially so in a hybrid cloud where there may be some free resources available from private clouds but some fee-paying resources from public clouds. Meeting this challenge involves two classical computational problems. One is assigning resources to each of the tasks in the composite web service. The other is scheduling the allocated resources when each resource may be used by more than one task and may be needed at different points of time. In addition, we must consider Quality-of-Service issues, such as execution time and running costs. Existing approaches to resource allocation and scheduling in public clouds and grid computing are not applicable to this new problem. This paper presents a random-key genetic algorithm that solves new resource allocation and scheduling problem. Experimental results demonstrate the effectiveness and scalability of the algorithm.

Hierarchical models for 2D presence/absence data having ambiguous zeroes: With a biogeographical case study on dingo behaviour

This dissertation is primarily an applied statistical modelling investigation, motivated by a case study comprising real data and real questions. Theoretical questions on modelling and computation of normalization constants arose from pursuit of these data analytic questions. The essence of the thesis can be described as follows. Consider binary data observed on a two-dimensional lattice. A common problem with such data is the ambiguity of zeroes recorded. These may represent zero response given some threshold (presence) or that the threshold has not been triggered (absence). Suppose that the researcher wishes to estimate the effects of covariates on the binary responses, whilst taking into account underlying spatial variation, which is itself of some interest. This situation arises in many contexts and the dingo, cypress and toad case studies described in the motivation chapter are examples of this. Two main approaches to modelling and inference are investigated in this thesis. The first is frequentist and based on generalized linear models, with spatial variation modelled by using a block structure or by smoothing the residuals spatially. The EM algorithm can be used to obtain point estimates, coupled with bootstrapping or asymptotic MLE estimates for standard errors. The second approach is Bayesian and based on a three- or four-tier hierarchical model, comprising a logistic regression with covariates for the data layer, a binary Markov Random field (MRF) for the underlying spatial process, and suitable priors for parameters in these main models. The three-parameter autologistic model is a particular MRF of interest. Markov chain Monte Carlo (MCMC) methods comprising hybrid Metropolis/Gibbs samplers is suitable for computation in this situation. Model performance can be gauged by MCMC diagnostics. Model choice can be assessed by incorporating another tier in the modelling hierarchy. This requires evaluation of a normalization constant, a notoriously difficult problem. Difficulty with estimating the normalization constant for the MRF can be overcome by using a path integral approach, although this is a highly computationally intensive method. Different methods of estimating ratios of normalization constants (N Cs) are investigated, including importance sampling Monte Carlo (ISMC), dependent Monte Carlo based on MCMC simulations (MCMC), and reverse logistic regression (RLR). I develop an idea present though not fully developed in the literature, and propose the Integrated mean canonical statistic (IMCS) method for estimating log NC ratios for binary MRFs. The IMCS method falls within the framework of the newly identified path sampling methods of Gelman & Meng (1998) and outperforms ISMC, MCMC and RLR. It also does not rely on simplifying assumptions, such as ignoring spatio-temporal dependence in the process. A thorough investigation is made of the application of IMCS to the three-parameter Autologistic model. This work introduces background computations required for the full implementation of the four-tier model in Chapter 7. Two different extensions of the three-tier model to a four-tier version are investigated. The first extension incorporates temporal dependence in the underlying spatio-temporal process. The second extensions allows the successes and failures in the data layer to depend on time. The MCMC computational method is extended to incorporate the extra layer. A major contribution of the thesis is the development of a fully Bayesian approach to inference for these hierarchical models for the first time. Note: The author of this thesis has agreed to make it open access but invites people downloading the thesis to send her an email via the 'Contact Author' function.

Development and validation of a novel approach to work sampling : a study of nurse practitioner work patterns

Objectives: This methodological paper reports on the development and validation of a work sampling instrument and data collection processes to conduct a national study of nurse practitioners’ work patterns. ---------- Design: Published work sampling instruments provided the basis for development and validation of a tool for use in a national study of nurse practitioner work activities across diverse contextual and clinical service models. Steps taken in the approach included design of a nurse practitioner-specific data collection tool and development of an innovative web-based program to train and establish inter rater reliability of a team of data collectors who were geographically dispersed across metropolitan, rural and remote health care settings. ---------- Setting: The study is part of a large funded study into nurse practitioner service. The Australian Nurse Practitioner Study is a national study phased over three years and was designed to provide essential information for Australian health service planners, regulators and consumer groups on the profile, process and outcome of nurse practitioner service. ---------- Results: The outcome if this phase of the study is empirically tested instruments, process and training materials for use in an international context by investigators interested in conducting a national study of nurse practitioner work practices. ---------- Conclusion: Development and preparation of a new approach to describing nurse practitioner practices using work sampling methods provides the groundwork for international collaboration in evaluation of nurse practitioner service.

Gaseous air pollution and emergency hospital visits for hypertension in Beijing, China : a time-stratified case-crossover study

Background: A number of epidemiological studies have been conducted to research the adverse effects of air pollution on mortality and morbidity. Hypertension is the most important risk factor for cardiovascular mortality. However, few previous studies have examined the relationship between gaseous air pollution and morbidity for hypertension. ---------- Methods: Daily data on emergency hospital visits (EHVs) for hypertension were collected from the Peking University Third Hospital. Daily data on gaseous air pollutants (sulfur dioxide (SO2) and nitrogen dioxide (NO2)) and particulate matter less than 10 μm in aerodynamic diameter (PM10) were collected from the Beijing Municipal Environmental Monitoring Center. A time-stratified case-crossover design was conducted to evaluate the relationship between urban gaseous air pollution and EHVs for hypertension. Temperature and relative humidity were controlled for. ---------- Results: In the single air pollutant models, a 10 μg/m3 increase in SO2 and NO2 were significantly associated with EHVs for hypertension. The odds ratios (ORs) were 1.037 (95% confidence interval (CI): 1.004-1.071) for SO2 at lag 0 day, and 1.101 (95% CI: 1.038-1.168) for NO2 at lag 3 day. After controlling for PM10, the ORs associated with SO2 and NO2 were 1.025 (95% CI: 0.987-1.065) and 1.114 (95% CI: 1.037-1.195), respectively.---------- Conclusion: Elevated urban gaseous air pollution was associated with increased EHVs for hypertension in Beijing, China.

How do we reach them? Comparing random samples from mobile and landline phones

Quantitative studies of nascent entrepreneurs such as GEM and PSED are required to generate their samples by screening the adult population, usually by phone in developed economies. Phone survey research has recently been challenged by shifting patterns of ownership and response rates of landline versus mobile (cell) phones, particularly for younger respondents. This challenge is acutely intense for entrepreneurship which is a strongly age-dependent phenomenon. Although shifting ownership rates have received some attention, shifting response rates have remained largely unexplored. For the Australian GEM 2010 adult population study we conducted a dual-frame approach that allows comparison between samples of mobile and landline phones. We find a substantial response bias towards younger, male and metropolitan respondents for mobile phones – far greater than explained by ownership rates. We also found these response rate differences significantly biases the estimates of the prevalence of early stage entrepreneurship by both samples, even when each sample is weighted to match the Australian population.

QoS-oriented resource allocation and scheduling of multiple composite Web services in a hybrid cloud using a random-key genetic algorithm

In cloud computing resource allocation and scheduling of multiple composite web services is an important challenge. This is especially so in a hybrid cloud where there may be some free resources available from private clouds but some fee-paying resources from public clouds. Meeting this challenge involves two classical computational problems. One is assigning resources to each of the tasks in the composite web service. The other is scheduling the allocated resources when each resource may be used by more than one task and may be needed at different points of time. In addition, we must consider Quality-of-Service issues, such as execution time and running costs. Existing approaches to resource allocation and scheduling in public clouds and grid computing are not applicable to this new problem. This paper presents a random-key genetic algorithm that solves new resource allocation and scheduling problem. Experimental results demonstrate the effectiveness and scalability of the algorithm.

Deterring drug drivers : a study into the initial impact of oral random roadside drug testing in Queensland

Objective: The global implementation of oral random roadside drug testing is relatively limited, and correspondingly, the literature that focuses on the effectiveness of this intervention is scant. This study aims to provide a preliminary indication of the impact of roadside drug testing in Queensland. Methods: A sample of Queensland motorists’ (N= 922) completed a self-report questionnaire to investigate their drug driving behaviour, as well as examine the perceived affect of legal sanctions (certainty, severity and swiftness) and knowledge of the countermeasure on their subsequent offending behaviour. Results: Analysis of the collected data revealed that approximately 20% of participants reported drug driving at least once in the last six months. Overall, there was considerable variability in respondent’s perceptions regarding the certainty, severity and swiftness of legal sanctions associated with the testing regime and a considerable proportion remained unaware of testing practices. In regards to predicting those who intended to drug driving again in the future, perceptions of apprehension certainty, more specifically low certainty of apprehension, were significantly associated with self-reported intentions to offend. Additionally, self-reported recent drug driving activity and frequent drug consumption were also identified as significant predictors, which indicates that in the current context, past behaviour is a prominent predictor of future behaviour. To a lesser extent, awareness of testing practices was a significant predictor of intending not to drug drive in the future. Conclusion: The results indicate that drug driving is relatively prevalent on Queensland roads, and a number of factors may influence such behaviour. Additionally, while the roadside testing initiative is beginning to have a deterrent impact, its success will likely be linked with targeted intelligence-led implementation in order to increase apprehension levels as well as the general deterrent effect.

Attitudes to living and working in pandemic conditions among emergency prehospital medical care personnel

Introduction: Little is known about the risk perceptions and attitudes of healthcare personnel, especially of emergency prehospital medical care personnel, regarding the possibility of an outbreak or epidemic event. Problem: This study was designed to investigate pre-event knowledge and attitudes of a national sample of the emergency prehospital medical care providers in relation to a potential human influenza pandemic, and to determine predictors of these attitudes. Methods: Surveys were distributed to a random, cross-sectional sample of 20% of the Australian emergency prehospital medical care workforce (n = 2,929), stratified by the nine services operating in Australia, as well as by gender and location. The surveys included: (1) demographic information; (2) knowledge of influenza; and (3) attitudes and perceptions related to working during influenza pandemic conditions. Multiple logistic regression models were constructed to identify predictors of pandemic-related risk perceptions. Results: Among the 725 Australian emergency prehospital medical care personnel who responded, 89% were very anxious about working during pandemic conditions, and 85% perceived a high personal risk associated with working in such conditions. In general, respondents demonstrated poor knowledge in relation to avian influenza, influenza generally, and infection transmission methods. Less than 5% of respondents perceived that they had adequate education/training about avian influenza. Logistic regression analyses indicate that, in managing the attitudes and risk perceptions of emergency prehospital medical care staff, particular attention should be directed toward the paid, male workforce (as opposed to volunteers), and on personnel whose relationship partners do not work in the health industry. Conclusions: These results highlight the potentially crucial role of education and training in pandemic preparedness. Organizations that provide emergency prehospital medical care must address this apparent lack of knowledge regarding infection transmission, and procedures for protection and decontamination. Careful management of the perceptions of emergency prehospital medical care personnel during a pandemic is likely to be critical in achieving an effective response to a widespread outbreak of infectious disease.

Tissue engineering of stratified articular cartilage from chondrocyte subpopulations

Objective: To test if subpopulations of chondrocytes from different cartilage zones could be used to engineer cartilage constructs with features of normal stratification. Design: Chondrocytes from the superficial and middle zones of immature bovine cartilage were cultured in alginate, released, and seeded either separately or sequentially to form cartilage constructs. Constructs were cultured for 1 or 2 weeks and were assessed for growth, compressive properties, and deposition, and localization of matrix molecules and superficial zone protein (SZP). Results: The cartilaginous constructs formed from superficial zone chondrocytes exhibited less matrix growth and lower compressive properties than constructs from middle zone chondrocytes, with the stratified superficial-middle constructs exhibiting intermediate properties. Expression of SZP was highest at the construct surfaces, with the localization of SZP in superficial-middle constructs being concentrated at the superficial surface. Conclusions: Manipulation of subpopulations of chondrocytes can be useful in engineering cartilage tissue with a biomimetic approach, and in fabricating constructs that exhibit stratified features of normal articular cartilage. (C) 2003 OsteoArthritis Research Society International. Published by Elsevier Ltd. All rights reserved.

Short-term retention of labeled chondrocyte subpopulations in stratified tissue-engineered cartilaginous constructs implanted in vivo in mini-pigs

It is likely that effective application of cell-laden implants for cartilage defects depends on retention of implanted cells and interaction between implanted and host cells. The objectives of this study were to characterize stratified cartilaginous constructs seeded sequentially with superficial (S) and middle (M) chondrocyte subpopulations labelled with fluorescent cell tracking dye PKH26 (*) and determine the degree to which these stratified cartilaginous constructs maintain their architecture in vivo after implantation in mini-pigs for 1 week. Alginate-recovered cells were seeded sequentially to form stratified S*/M (only S cells labelled) and S*/M* (both S and M cells labelled) constructs. Full-thickness defects (4 mm diameter) were created in the patellofemoral groove of adult Yucatan mini-pigs and filled with portions of constructs or left empty. Constructs were characterized biochemically, histologically, and biomechanically, and stratification visualized and quantified, before and after implant. After 1 week, animals were sacrificed and implants retrieved. After 1 week in vivo, glycosaminoglycan and collagen content of constructs remained similar to that at implant, whereas DNA content increased. Histological analyses revealed features of an early repair response, with defects filled with tissues containing little matrix and abundant cells. Some implanted (PKH26-labeled) cells persisted in the defects, although constructs did not maintain a stratified organization. Of the labelled cells, 126 +/- 38% and 32 +/- 8% in S*/M and S*/M* constructs, respectively, were recovered. Distribution of labelled cells indicated interactions between implanted and host cells. Longer-term in vivo studies will be useful in determining whether implanted cells are sufficient to have a positive effect in repair.