Master of Biostatistics

This subject is a core subject within the Master of Public Health, the Master of Epidemiology and the Master of Science (Epidemiology). Students should enrol in this subject early in their program of study.

Epidemiology is the discipline of studying the distribution and determinants of disease in populations and is a fundamental science of public health.

The subject covers the role of epidemiology in public health and ethical conduct of quantitative research. Within this subject the measures of population health and disease frequency, measures of association and measures of the impact of specific risk factors are studied. The subject includes descriptive epidemiology using routinely collected data. The common experimental and observational study designs, and systematic reviews, and their relative strengths and weaknesses are discussed. The implications of common types of bias (selection bias, information bias, and confounding) are discussed, as are methods to minimise them. Causal inference is considered within a framework of critical appraisal of epidemiological evidence. The validity and performance of screening and diagnostic tests are considered. Current infectious diseases will also be examined by considering the principles of infectious disease transmission and surveillance systems used for health protection. The cultural considerations in undertaking research within indigenous populations, and epidemiological measures in the context of indigenous health will be considered in an online module.

The aim of this subject is to provide students with the knowledge and skills required to undertake moderate to high level data manipulation and management in preparation for statistical analysis of data typically arising in health and medical research. In particular, students gain experience in data manipulation and management using two major statistical software packages (Stata and R) and acquire fundamental programme skills for efficient use of each of these software packages.

This subject begins with the study of probability, random variables, discrete and continuous distributions, and the use of calculus to obtain expressions for parameters of these distributions such as the mean and variance. Joint distributions for multiple random variables are introduced together with the important concepts of independence, correlation and covariance, and marginal and conditional distributions. Techniques for determining distributions of transformations of random variables are discussed. The concept of the sampling distribution and standard error of an estimator of a parameter is presented, together with key properties of estimators. Large sample results concerning the properties of estimators are presented with emphasis on the central role of the normal distribution in these results. General approaches to obtaining estimators of parameters are introduced. Numerical simulation and graphing with Stata are used throughout to demonstrate key concepts.

This subject provides the foundation theory and methods needed for biostatisticians to apply and critically interpret statistical inference, the science of drawing conclusions from data that are subject to variability. Major topics include review of the key concepts of estimation including sampling variability and construction of confidence intervals; null hypothesis testing; methods of inference based on likelihood theory (Fisher and observed information, likelihood ratio, Wald and score tests); and an introduction to the Bayesian approach to inference. The approach will emphasise a critical understand¬ing of the role of statistical inference in health research.

This subject provides the foundation for regression modelling. Topics covered include: the method of least squares; regression models and related statistical inference; flexible nonparametric regression; analysis of covariance to adjust for confounding; multiple regression with matrix algebra; model construction and interpretation (use of indicator variables, parameterisation, interaction and transformations); model checking and diagnostics; regression to the mean; handling of baseline values; the analysis of variance; variance components and random effects.

This subject begins with the study of probability, random variables, discrete and continuous distributions, and the use of calculus to obtain expressions for parameters of these distributions such as the mean and variance. Joint distributions for multiple random variables are introduced together with the important concepts of independence, correlation and covariance, and marginal and conditional distributions. Techniques for determining distributions of transformations of random variables are discussed. The concept of the sampling distribution and standard error of an estimator of a parameter is presented, together with key properties of estimators. Large sample results concerning the properties of estimators are presented with emphasis on the central role of the normal distribution in these results. General approaches to obtaining estimators of parameters are introduced. Numerical simulation and graphing with Stata are used throughout to demonstrate key concepts.

Introduction to and revision of conventional methods for contingency tables especially in epidemiology:

• Odds ratios and relative risks;
• Chi-squared tests for independence;
• Mantel Haenszel methods for stratified tables;
• Methods for paired data.

The exponential family of distributions;

• Generalized Linear Models (GLMs);
• Parameter estimation for GLMs;
• Inference for GLMs, including the use of score, Wald and deviance statistics (including residuals) for confidence intervals and hypothesis tests.

Binary variables and logistic regression models:

• Methods for assessing model adequacy;
• Nominal and ordinal logistic regression for categorical response variables with more than two categories;

Count data and Poisson regression models:

• Log-linear models.

Software:

• Fitting GLMs in Stata and R.

Clinical trials (equivalence trials, cross-over trials); Clinical agreement (Bland-Altman methods, kappa statistics, intraclass correlation); Statistical process control (special and common causes of variation; quality control charts); Diagnostic tests (sensitivity, specificity, ROC curves); Meta-analysis (systematic reviews, assessing heterogeneity, publication bias, estimating effects from randomised controlled trials, diagnostic tests and observational studies).

Topics include: routinely collected health-related data; quantitative methods in demography, including standardisation and life tables; health differentials; design and analysis of population health surveys, including the role of stratification, clustering and weighting.

Topics covered: Paired data; the effect of non-independence on comparisons within and between clusters of observations; methods for continuous outcomes: normal mixed effects (hierarchical or multilevel) models and generalised estimating equations (GEE); role and limitations of repeated measures ANOVA; methods for discrete data: GEE and generalized linear mixed models (GLMM); methods for count data.

Topics include: Kaplan-Meier life tables; logrank test to compare two or more groups; Cox's proportional hazards regression model; checking the proportional hazards assumption; time-dependent covariates; sample size calculations for survival studies.

AIMS

ICT is an important component to ensuring quality, safety, access and efficiency in healthcare. This subject introduces current approaches and future directions in eHealth and the use of ICT in healthcare generally as well as key concepts and tools from the underlying discipline of health informatics.

INDICATIVE CONTENT

Topics include electronic health records (EHRs); hospital and primary care and public health information systems; supporting clinical decision-making for health professionals through ICT; eHealth in the community for preventive healthcare and for patient and carer support; regulatory influences on eHealth including management and governance, privacy, security, and confidentiality; the role of data standards, vocabularies, and nomenclatures in eHealth; research and development in eHealth.

Computing techniques and data mining methods are indispensable in modern statistical research and data science applications, where “Big Data” problems are often involved. This subject will introduce a number of recently developed methods and applications in computational statistics and data science that are scalable to large datasets and high-performance computing. The data mining methods to be introduced include general model diagnostic and assessment techniques, kernel and local polynomial nonparametric regression, basis expansion and nonparametric spline regression, generalised additive models, classification and regression trees, forward stagewise and gradient boosting models. Important statistical computing algorithms and techniques used in data science will be explained in detail. These include the bootstrap resampling and inference, cross-validation, the EM algorithm and Louis method, and Markov chain Monte Carlo methods including adaptive rejection and squeeze sampling, sequential importance sampling, slice sampling, Gibbs sampler and Metropolis-Hastings algorithm.

Topics include: ethical considerations; principles and methods of randomisation in controlled trials; treatment allocation, blocking, stratification and allocation concealment; parallel, factorial and crossover designs including n-of-1 studies; practical issues in sample size determination; intention-to-treat principle; phase I dose finding studies; phase II safety and efficacy studies; interim analysis and early stopping ; multiple outcomes/endpoints, including surrogate outcomes, multiple tests and subgroup analyses, including adjustment of significance levels and P-values; missing data; reporting trial results and use of the CONSORT statement.

Faced with the rising cost of vaccines and increasing drug resistance, public health decision makers increasingly rely on epidemiological models of infectious disease transmission to predict the impact, and define optimal implementation of, intervention strategies. Such considerations are particularly critical in resource-constrained settings.

This subject introduces students to the concepts of infectious diseases modeling required to interpret modeling papers relevant to the public health context. By considering real world examples of the use of models to support practice, they will learn to distinguish between different types of modeling frameworks, and understand their relevance to alternative questions and settings. Building on their strengths in infectious diseases epidemiology, students will develop confidence in assessing whether model frameworks incorporate all relevant knowledge and are ‘fit for purpose’ to support decision making.

This subject builds on methods and techniques learned in theoretical subjects by studying the application of statistics in real contexts. Emphasis is on the skills needed for a practising statistician, including the development of mature statistical thinking, organizing the structure of a statistical problem, the contribution to the design of research from a statistical point of view, measurement issues and data processing. The subject deals with thinking about data in a broad context, and skills required in statistical consulting.

This subject is a core subject within the Master of Epidemiology and the Master of Science (Epidemiology) and an elective within the Master of Public Health and Master of Environment.

It covers the main experimental and observational study designs used in epidemiological research: These will include randomised controlled trials including the variants of trials, cohort studies, case-control studies including the variants of nested case-control studies, case-cohort studies and case-crossover studies and ecological studies.

Causal diagrams are introduced as a unifying means for identifying confounding and selection bias and interpreting associations. Other topics include: measurement error, effect modification and validity of the findings.

Skills in critically appraising findings from research will build on the base of Epidemiology one. Students will apply their knowledge to designing studies with the aim of investigating topical problems in public health.

AIMS

The subject introduces key topics in modern information organization, particularly with regard to structured databases. The well-founded relational theory behind modern structured query language (SQL) engines, has given them as much a place behind the web site of an organization and on the desktop, as they traditionally enjoyed on corporate mainframes. Topics covered may include: the managerial view of data, information and knowledge; conceptual, logical and physical data modelling; normalization and de-normalization; the SQL language; data integrity; transaction processing, data warehousing, web services and organizational memory technologies. This is a core foundation subject for both the Master of Information Systems and Master of Information Technology.

INDICATIVE CONTENT

This subject serves as an introduction to databases and data modelling from a data management perspective. Database design, from conceptual design through to physical implementation will be covered. This will include Entity Relationship modelling, normalisation and de-normalisation and SQL. Additionally the use of databases in various contexts will be explored (web based databases, connecting programs to databases, data warehousing, health contexts, geospatial databases).

AIMS

The aims for this subject is for students to develop an understanding of approaches to solving moderately complex problems with computers, and to be able to demonstrate proficiency in designing and writing programs. The programming language used is Java.

INDICATIVE CONTENT

Topics covered will include:

• Java basics
• Console input/output
• Control flow
• Defining classes
• Using object references
• Programming with arrays
• Inheritance
• Polymorphism and abstract classes
• Exception handling
• UML basics
• Interfaces
• Generics.

Topics include: simple one-parameter models with conjugate prior distributions; standard models containing two or more parameters, including specifics for the normal location-scale model; the role of non-informative prior distributions; the relationship between Bayesian methods and standard ‘classical’ approaches to statistics, especially those based on likelihood methods; computational techniques for use in Bayesian analysis, especially the use of simulation from posterior distributions, with emphasis on the WinBUGS package as a practical tool; application of Bayesian methods for fitting hierarchical models to complex data structures.