We should acknowledge the role that Dr. Colton has played in these symposia since their inauguration in 1988. It was through his vision that the "rst of these symposia was held. The proceedings from this conference will mark the 10th anniversary of the collaboration of CDC and Statistics in Medicine, published by John Wiley & Sons. His support, encouragement, and constant urging to quality and timely publication have greatly a!ected the creation and dissemination of the results of these symposia.
Secondly, I thank Ken Falter and his Deputy Chair, Don Betts, for a most commendable symposium. They, along with their excellent sta!, have produced a stimulating and valuable conference. Donald and Ken, I welcome you now to the distinguished society of past chairs of CDC/ATSDR statistical symposia.
To summarize the conference, I used an exploratory data analytic approach. A pattern emerged, and in honour of Dr. Colton, who could not be with us for the "rst time ever in this series, we call these the &C' concepts. The "rst concept that I saw emerging from the talks and the posters is completeness. The "rst symposium in 1988 was a symposium on statistics and surveillance. It was held at CDC and approximately 100 people came together and presented about 25 papers. The topic of surveillance was chosen since it was and remains a core public health function. The purpose of the "rst symposium was to bring to one of the core public health functions the science of statistics and begin to illustrate how statistics can bring an important piece of evidence to that activity. Still in this conference we see statisticians dealing with issues relating to surveillance. We heard about analysis of aberrations, quality control charts from industry, cluster analyses, and Bayesian methods of surveillance analysis.
We have learned much since that "rst symposium. In 1997, we have approximately three times the number of attendees from academia, industry and international countries. We have representation from 28 states and nine international countries. As a colleague of mine said recently &So, 300 of your closest friends are coming to town tomorrow, right?' The second symposium had to do with cluster analyses, and we heard about clusters again at this symposium (for example, Dr. Chen's paper on q-statistics and the last parallel session on geographical smoothing and spatial analysis).
The topic of the third symposium on the evaluation of intervention and prevention strategies was revisited here when we heard about the evaluation of hepatitis B immunization programmes and the evaluation of control measures for a measles outbreak. The fourth symposium on the combination of data from multiple sources was recalled by a meta-analysis of educational and intervention activities in schools. And "nally, the last symposium on geographical information systems concerned methods used in the plenary session on spatial statistics.