A number of diverse applications of statistical methods will be described in outline and the implicit challenges for the future of statistics discussed.

Models for the analysis of multivariate spatial data are receiving increased attention these days. In many applications it will be preferable to work with multivariate spatial processes to specify such models. A critical specification in providing these models is the cross covariance function. Constructive approaches for developing valid cross-covariance functions offer the most practical strategy for doing this. These approaches include separability, kernel convolution or moving average methods, and convolution of covariance functions. We review these approaches but take as our main focus the computationally manageable class referred to as the linear model of coregionalization (LMC). We introduce a fully Bayesian development of the LMC. We offer clarification of the connection between joint and conditional approaches to fitting such models including prior specifications. However, to substantially enhance the usefulness of such modelling we propose the notion of a spatially varying LMC (SVLMC) providing a very rich class of multivariate nonstationary processes with simple interpretation.

We illustrate the use of our proposed SVLMC with application to more than 600 commercial property transactions in three quite different real estate markets, Chicago, Dallas and San Diego. Bivariate nonstationary process models are developed for income from and selling price of the property.

Dose-response studies are an extremely important part of the drug development process as knowledge of the relationship between response and dose is an essential requirement for making informed decisions about dosage.

One potential difficulty in the use of a limited number of doses in a parallel group design to investigate dose- response is the danger that the steep part of the dose-response curve may fall between two doses and little is learnt. The use of a large number of doses in order to circumvent this problem is potential wasteful in its use of patients because a large number will either be receiving doses which are little different from placebo or doses which have a greater potential for causing side effects. Ideally the vast majority of patients should receive doses in the steepest part of the dose response function. One design that achieves this is based on Bayesian Adaptive principles in which knowledge of the dose-response curve is updated on an on-going basis in order to inform two decisions. First, what dose should be allocated to the next patient; second should the study continue or should it be stopped.

There have been at least two major hindrances to the use of Bayesian methods in pharmaceutical R&D, one of which is practical, the other more philosophical. The practical constraint has been the lack of availability of methods and software for their implementation; the philosophical constraint has been a perceived antipathy from regulators to the use of priors. We discuss aspects of these issues in practically implementing the proposed design including some, or all, of :

• choice of dose-response function
• solutions to the decision problems
• simulation of adaptive clinical trials to sample size the study and to provide details of the operating characteristic of the design
• use of prior information
• software validation
• interaction with regulatory authorities.

We know that statistics are essential to good government, despite the familiar quotation "lies, damned lies and statistics." As political and social life require clear information on people, agriculture, industry and economy, statistics is necessary to deciding which actions or dispositions are ethical.

The UK Chancellor, Gordon Brown, provides an example of the contribution to statistics in the international debate on the ethics of trade and poverty. He hopes Britain can secure a new system of independent reports on the true state of government finances around the world to ensure aid money was spent wisely. He also states that Europe should publish the full consequences of trade tariffs and protectionism on agriculture and other goods.

The paper shall consider the development of ethics, and debate which particular ethical theories are most appropriate to statisticians. Early discussion of ethics were firmly social. However, 2.5 millenia later, individual automony is very popular. The place of statistics within ethics which focus on a single person is more subtle.

The controversial 'star-ratings' approach for monitoring NHS performance is due for a major overhaul by the Healthcare Commission, and statistical analysis of performance indicators is bound to form a component of any future monitoring and assessment procedure. Adaptations of classical industrial quality-control procedures, such as over-dispersed funnel plots and risk-adjusted CUSUMS, seem to be appropriate tools to help in comparing performance and guiding inspections. Their application to routinely collected data will be illustrated.