Zeng and Lin read paper to Royal Statistical Society

 
March 13, 2007
Photograph of Dr. Donglin Zeng

Photograph of Dr. Donglin Zeng
Photograph of Dr. Danyu Lin

Photograph of Dr. Danyu Lin

Drs. Donglin Zeng and Danyu Lin, professors in the School’s Department of Biostatistics, read a paper to the Royal Statistical Society (RSS) at its Ordinary Meeting in London on Jan. 31, 2007. The paper provides a new statistical methodology that can be used to solve difficult public health problems such as reliable assessments of treatment effectiveness, precise descriptions of the effects of environmental and genetic factors on the development of disease, and accurate predictions of patient outcomes.

In a RSS Ordinary Meeting, a single paper is presented and discussed live. Drs. Zeng and Lin’s paper, “Maximum Likelihood Estimation in Semiparametric Regression Models with Censored Data,” will be published in Series B (Statistical Methodology) of the Journal of the Royal Statistical Society (JRSS), along with the discussants’ contributions and the authors’ replies.

“Reading a paper to the RSS is one of the highest honors for statisticians,” said Dr. Barbara K. Rimer, dean of the UNC School of Public Health. “Discussion papers published in Series B of the journal have had tremendous impact on the theory and practice of statistics, and it is rare for a non-Briton to read a paper.”

Zeng and Lin’s methodology can help public health investigators better analyze data that are potentially censored. Many studies in public health and medicine are concerned with the developments of certain types of events, such as disease and death. In a cancer clinical trial, investigators are often interested in the time to cancer recurrence or death. In a genetic epidemiological study, investigators may be interested in the age at the onset of schizophrenia. In such studies, censoring is inevitable because some subjects are not followed long enough for their event times to be fully observed.

The standard method for analyzing censored data was proposed by renowned English statistician Sir David Cox in a RSS Ordinary Meeting in 1972, but his model is often violated in practice. Zeng and Lin’s paper extends the Cox model in several important directions. Their work deals with complicated data structures, which can arise when each study subject can experience several events, when multiple family members are enrolled in a study, or when different types of outcomes are simultaneously studied.

“Donglin and Danyu’s work has generated a great deal of interest,” said Dr. Michael Kosorok, chair of the Department of Biostatistics. “Their reading was well attended. Fourteen experts, including Sir David Cox, discussed their paper in person. Nearly twenty more experts provided written comments. This number of discussants is most likely a record in the long history of the RSS Ordinary Meetings.”

Dr. Lin is the Dennis Gillings Distinguished Professor of Biostatistics. The endowed professorship was created through a generous gift from Dr. Dennis Gillings and attracted Dr. Lin to the School in 2001. Dr. Zeng also joined the Carolina Biostatistics faculty in 2001.

To learn more about Drs. Zeng and Lin’s research, visit their websites at http://www.bios.unc.edu/~dzeng and http://www.bios.unc.edu/~lin. Additional information about RSS Ordinary Meetings, including a preprint of Drs. Zeng and Lin’s paper, can be found at the RSS website at http://www.rss.org.uk.

# # #

For more information, contact Ramona DuBose, director of communications for the UNC School of Public Health, by telephone at 919-966-7467 or by email at ramona_dubose@unc.edu.