Harvard School of Public Health

Telephone: (617) 432-1056

Department of Biostatistics

FAX: (617) 739-1781


Internet: dept@hsph.harvard.edu



Biostat Connections

Editorial Staff: Sudeshna Adak, Paul Catalano, Victor DeGruttola, Bhaswati Ganguli, Erin Kammann,

Cyrus Mehta, Evelyn Ophir, John Rogus, and Florin Vaida

The PDF Version

August 1, 2000

No. 10

The Evolving State of Statistical Genetics

by John Rogus and Victor DeGruttola

Statistical analysis of genetic data is a growth industry. Opportunities abound in genetic linkage analysis, genetic association analysis, and most recently gene expression profiling. In linkage analysis, statistical inference based on inheritance patterns makes it possible to identify chromosomal regions containing disease predisposing genes. Because many diseases under current investigation are multifactorial, nonparametric linkage analysis (e.g., sib pair analysis) is now enjoying center stage. In association analysis, an attempt is made to correlate specific genetic variants with disease. To guard against spurious results due to population admixture, much of the recent work has focused on family-based analysis. Some of the most exciting work in this area, relaxing the criteria needed to qualify as a "family" and developing a unified framework for statistical analysis, is being carried out in our department under the direction of Nan Laird.

In addition to the opportunities afforded by linkage and association analysis, there is now another rapidly emerging field ripe for contributions by statisticians - gene expression profiling. Not surprisingly, what distinguishes this area is the focus on the expression of genes rather than on the differences in inherited DNA. Notably, differential expression of genes occurs not only across different tissue types for a given individual but also across individuals for a given tissue type. This is especially apparent when samples from many individuals are partitioned by some meaningful clinical measurement such as presence/absence of disease.

The goals of gene expression profiling include: 1) to improve diagnosis of diseases, 2) to develop disease subclassification, 3) to generate targets for drug development and 4) to customize drug therapies to particular genetic profiles. Despite these lofty ambitions, it is becoming apparent that gene expression profiling is not just science fiction. In 1999, the FDA approved a Genentech drug for the treatment of metastatic breast cancer in women with overexpression of a certain gene. Recognizing this emerging trend, members of the Biostatistics faculty have gotten involved in the statistical aspects of gene expression profiling. Particularly active areas include cancer and AIDS.

Cancer is perhaps one of the ideal diseases for gene expression profiling. For starters, gathering tissue (from tumors) is relatively straightforward compared to many other diseases. Moreover, uncontrolled cell growth is a process under a large degree of genetic, although not necessarily heritable, control and cancer is typically accompanied by a host of changes in gene expression. Among those studying gene expression in cancer is a group at Dana-Farber including Donna Neuberg (interviewed for this article), Robert Gentleman, Bob Gray, Joe Ibrahim, LJ Wei, and Marsha Wilcox.

Gene expression is currently assessed by a series of probes contained on chips, membranes, or glass slides. According to Donna, one of the fundamental areas in need of attention is the standardization of the resulting measurements. Such standardization is necessary not only to facilitate the communication of results but more critically, at the current time, to transform highly noisy data into a workable form. Currently, substantial variation exists not only among chips (membranes, etc.) but also among repeated measurements on a given chip. There can also be variation among different cells of a tumor and among different samples of cells especially if some normal cells find there way into the sample. Unfortunately, much of this nuisance variation can be relatively large in comparison to the variation of interest. Further complicating the situation is the dearth of studies with any meaningful degree of replication data built in. Nevertheless, with the Department's successful recruitment of Professor Wing Wong, this area is likely to be an active area of research in the coming months and years.

Once expression has been adequately quantified, the next step is to evaluate correlations with clinical outcomes. Typically, hundreds or thousands of genes are considered simultaneously, so one of the main challenges, according to Donna, is the development of clustering algorithms. For example, patterns in expression can be used in an attempt to develop previously unrecognized clinical distinctions (unsupervised learning approach). Alternatively, if clinical distinctions are available, it may be desirable to search for the genes that are differentially expressed depending on disease subtype (supervised learning approach). Researchers at the Whitehead Institute in Cambridge have recently used this approach to differentiate expression patterns between two types of leukemia. Such information can be useful in early diagnosis of disease or for customizing treatment regimens.

Donna also describes how statisticians can fill a void in the area of modeling. Instead of simply assessing whether low/high gene expression is associated with, say, survival, well-developed statistical methodology can be adapted to accommodate covariates such as gender, pathology, and age. There is also plenty of room for statistical modeling to investigate correlation patterns among thousands of genes and to explore various environmental interactions.

Statisticians with expertise in longitudinal data analysis will also have a role to play in gene expression profiling. This need, according to Donna, stems from the desire to watch the unfolding of the cascade of events associated with a disease process. Current work is focused on how expression levels in cell lines change as experimental conditions vary, but future experiments are likely to involve repeated biopsy over time.

In addition to genetic research on cancer, there has been considerable interest in the Department of Biostatistics in investigation of the relationship between the genetic variation of the human immunodeficiency virus (HIV) that causes AIDS and resistance to antiviral therapies. Resistance to therapies for HIV infection arises from both naturally occuring genetic variants or "polymorphisms" and from genetic mutations associated with treatment. Understanding the relationship between the features of the viral genome and resistance to treatment is essential for making appropriate treatment decisions at the individual patient level, for developing appropriate policy regarding the use of treatments at a population level, and for developing new drugs. While the viral genome is smaller than the human genome by many orders of magnitude, the statistical issues arising from such research still relate to the high dimensionality of genetic data. As a result, new statistical approaches are needed for making efficient use of available information on genetic sequences of the HIV found in patient samples and clinical response to antiviral therapy.

Gene expression profiling is a hot area of molecular biology with plenty of room for contributions by statisticians. In fact, many of the most basic statistical issues are currently unresolved. How can power of a study be estimated? How should p-values be assigned in light of massive multiplicity? (L.J. Wei is spearheading some interesting work in this area). How should outliers be defined and what should be done with them? However, the promise of this technology is also great. Diagnostic procedures can be improved dramatically. Drug targets can be more effectively generated. Drug treatments can be tailored to meet individual needs. The hope for this field is echoed by management guru Tom Peters, who recently proposed in Time Magazine a list of the "10 hottest jobs of the future". Coming in at number 2 was "gene programming", which involves scanning a patient's DNA and creating customized prescriptions. Also on the list was "data mining", which is needed to sift through mountains of data (such as gene expression data). Clearly the future looks bright in this field and statisticians will undoubtedly have a large role to play.

What We Are Up To These Days in Biostats!

by Stephen W. Lagakos, Department Chair

Looking back at this past year-my first as Chair of the Department-I am very pleased with many of the changes that have taken place. Last year at this time we were looking forward with excitement to the arrival of our incoming class (n=19), and they have proven to be an outstanding group. Our Curriculum Committee, headed by Dr. Zelen, re-examined our entire program and has recommended several important changes. Two new courses-Operational Mathematics and Probability II-were approved and will be taught for the first time in the Fall of 2000 along with a new Special Topics course in Computational Biology (more below).

We launched a partnership with Kitasato University, which is located in Tokyo and recently formed the first graduate degree program in Biostatistics in Japan. The newly-formed Department of Biostatistics at Kitasato is being chaired by Masahiro Takeuchi, one of our graduates, and a main goal of the partnership will be to help Masahiro's fledgling department prosper. Our partnerships with Schering-Plough, Pfizer, and the Genetics Institute continue to thrive, and this year's HSPH-Schering Plough Workshop focused on quality of life issues and, by all accounts, was very successful.

We had several junior faculty positions to fill and were very fortunate to have recruited 4 outstanding individuals to join us. Drs. Marco Bonnetti and Brent Coull, who had been postdoctoral fellows in the department, began their terms as Assistant Professors on July 1, 2000 and will join our activities in cancer and environmental studies, respectively. Dr. Robert Gentleman, currently in New Zealand, will join us as an Associate Professor. Dr. Gentleman, who is interested in statistical computing, will be based at the Dana-Farber Cancer Institute. Finally, Dr. Hongyu Jiang, from the University of Wisconsin, will join our AIDS projects this Fall as a postdoctoral fellow and become an Assistant Professor in July 2001.

I am also pleased to announce that Dr. Andrea Rotnitzky has been selected by a HSPH search committtee for a tenured professorship, filling an opening caused by Butch Tsiatis' departure two years ago. We are very fortunate to have been granted two additional tenured positions this past year. One has just been filled by Dr. Wing Wong, who joins us from UCLA. Dr. Wong has a brilliant background in several areas of statistics, and for the last decade has been interested in genetics and computational biology (sometimes referred to as Medical Informatics). This is an area of great interest to HSPH and Harvard in general, and we are very excited about the directions in which Wing will lead us. Wing has relocated his computational laboratory from UCLA to our department and already has established an active program, with three doctoral students coming with him from UCLA, and three postdoctoral fellows and two visiting faculty joining him (and us) this academic year. The second tenured faculty position is to provide us with much-needed leadership role in our HIV/AIDS activities. This search is underway.

Last but not least, I am very pleased with our progress in fostering diversity. As usual, Professor Louise Ryan has continued to spearhead our efforts in attracting students from under-represented minority groups in the U.S. However, more faculty have become involved in our diversity efforts. Professor L.J. Wei, who co-chairs our Diversity Committee with Louise, has focused on increasing our representation from outside the U.S. To this end, I am very happy to report that this Fall we will have new students from China, Greece, Korea, Singapore, Tanzania, and Thailand in addition to an outstanding group of students from within the U.S.

One organizational change that has helped in our student recruiting and advising activities is the appointment of Dr. David Wypij as Director of Student Admissions and Advising. With this appointment we aim to be more proactive in our recruitment efforts and to provide a better educational environment for our students. David and Louise will benefit enormously by the appointment of Melissa Sanchez (formerly from our Center for Biostatistics in AIDS Research-CBAR) as Coordinator of Admissions and our Summer Program. Two other administrative changes are noteworthy. First, we are saddened by the departure of Shari Millen, who is expecting her first child. We are pleased to have found an ideal replacement in Michael Toon. Finally, the department has benefited in many ways by the addition of Carolyn Dueck, as Director of Administration. Carolyn has continued to serve as Deputy Director for Administration in CBAR and has done an excellent job in fostering a positive attitude within the department and with other components of HSPH.

In the coming year, we are looking forward to exploring new ways of improving the effectiveness of our teaching, our student advising, our mentoring of junior faculty, and our curriculum. With respect to the latter, the Curriculum Committee will be examining our offerings in the area of computing and making suggestions for ways to improve the curriculum. Also, Dr. Nan Laird is heading a departmental committee to investigate whether we should offer the Ph.D. degree instead of or in addition to the D.Sc. degree, and Dr. Marcello Pagano is heading a new committee to explore a continuing education program for our department. It is shaping up to be another exciting year!

HSPH Mentoring Award recipient Louise Ryan with students

Zelen Leadership Award Recognizes Leaders in Our Field

by Florin Vaida

Each year, the Marvin Zelen Leadership Award in Statistical Science recognizes an individual for excellence in leadership and for "creating an environment in which statistical science and its applications have flourished". The award was established in 1997 in honor of Dr. Zelen, with support of colleagues, friends and family.

The 1999 award was conferred upon John Tukey. On June 4, 1999, before the Marvin Zelen award ceremony, Professor Tukey presented a talk on "A Smorgasbord of Handy Techniques That Can Help in Analyzing Data".

We join the statistical community in sadness and regret with the news of Professor Tukey's death on July 26, 2000. Professor Tukey reached legendary status as one of the most influential statisticians of the last century. His pioneering and original work on graphical and numerical exploratory data analysis, spanning from his great interest and respect for "real data", has been lasting and far-reaching: the boxplot and the stem-and-leaf diagram come to mind. All self-respecting statistical libraries have a copy of his orange EDA book, much of whose content is included in today's statistical packages. Tukey did breakthrough research on paired comparisons, robust statistics and time series and was well-known outside statistics for the Fast Fourier Transform (1965, with J. Cooley). He is credited with coining the words "software" and "bit" for Binary digIT. CRC Press has published an eight-volume edition of Tukey's Collected Works.

John Tukey was born in 1915 in New Bedford, Massachusetts. After receiving a Master's degree from Brown University in 1937, he went on to study mathematics at Princeton University. In 1939 he received his doctorate in Topology. During WWII, he conducted war-related work in the ambiguously named Fire Control Research Office. During this time he was a member of the influential nucleus of statisticians that had joined the war effort at Princeton, which included Wilks, Mosteller and Cochran. In 1945 he joined both the Mathematics Department at Princeton and the AT&T Bell Laboratories at nearby Murray Hill. He was instrumental in the creation of the Statistics Department at Princeton in 1966 where he was chairman from 1966-1969. He retired from Bell Labs in 1992. John Tukey performed extensive and prominent public service work, as chairman on the President's Science Advisory Committee on Chemicals and Health and consultant to the 1990 Census among others. His numerous awards include the National Academy of Sciences Award in 1961 and the National Medal of Science from President Nixon in 1973.

The recipient of the Zelen Leadership Award for the year 2000 was Lincoln Moses of Stanford University. The awarding ceremony took place on June 2, 2000 following the Schering-Plough workshop and a lecture by Professor Moses on "Deciding Whether Large Clinical Trials and Meta-Analyses Agree or Not".

Lincoln Moses has had a long and fruitful career, spanning the whole second half of the last century. He enjoys much prestige both within the academic world and outside it. He is responsible for founding and building up the Biostatistics group at Stanford University and has held numerous leadership appointments including Dean of Graduate Studies at Stanford University (1969-1975) and Administrator in the U.S. Department of Energy (1978-1980) during President Carter's administration. His academic achievements include numerous articles in statistical and medical journals and several books (perhaps the best known being Elementary Decision Theory (1959) with Herman Chernoff). One of his key contributions has been the highly influential National Halothane Study, which stimulated the development of much new statistical research on a major public health issue.

Dr. Moses was born in 1921 in Kansas City, Missouri. After earning a degree in Social Sciences from Stanford University in 1941 and serving in the American Navy during WWII, Professor Moses returned to Stanford, where he received a Ph.D. in Statistics in 1950. He then spent two years as an assistant professor of education at Columbia University. In 1952, he became an assistant professor in the Department of Statistics and the Department of Preventive Medicine at Stanford University, and an associate professor in these departments in 1955. From 1959 until his retirement in 1992, Dr. Moses was a Professor of Statistics in the Department of Statistics and the Department of Research and Health Policy.

The many facets of Dr. Moses' personality and his achievements are expressed in an interview appearing in the August 1999 issue of Statistical Science.

Dr. Moses is the fourth recipient of the award which, in addition to John Tukey, was previously conferred on Frederick Mosteller and Sir David Cox.

Professors Lincoln Moses, Richard Gelber and Herman Chernoff at this year's Zelen Lecture

Our Department Has Been Enriched This Year
by New Research Fellows

by Evelyn Ophir

This past year the department welcomed nine new research fellows who bring with them a variety of research interests and come from diverse backgrounds.

Heejung Bang joined our department as a research fellow in November 1999 after receiving her Ph.D. from the Department of Statistics, North Carolina State University. She became interested in Biostatistics after having had a chance to work with and take a course from Butch Tsiatis, who made biostatistics interesting for her! So far her experience has been very positive in the department. She considers her work a new and fresh opportunity and has learned a lot by being able to apply statistics to real data. Areas of research Heejung is currently working on are epidemiology, data analysis from clinical trials in Tanzania, missing data analysis, and semiparametrics. Heejung enjoys spending her free time watching TV, shopping, and meeting with friends.

Tianxi Cai defended her thesis in our department last summer, and we are pleased that she began working as a research fellow with us in September. Tianxi came to the United States from China in 1995 to study applied mathematics at MIT and in 1996 transferred to our program. Tianxi always wanted to work in the field of statistics and decided on a career in biostatistics so that she could use statistics to make a contribution in a health-related environment. Tianxi's experience as a research fellow has been very good, and it has been a great opportunity for her to develop her interest in statistical problems such as survival analysis and statistical genetics. Tianxi has worked with L.J. Wei on the analysis of correlated failure time data and statistical genetics and with Robert Glynn on a collaborative study in cardiovascular diseases. We just received word that Tianxi has accepted a position as assistant professor in the Department of Biostatistics, University of Washington and will be leaving us in August. We are sorry she will be leaving us but wish her the very best as she relocates to Seattle and begins her new life there. As Tianxi enjoys travelling, we hope that she will be able to take good advantage of the Pacific Northwest beauty!

Lieven Declerck defended his doctoral thesis in May of last year at the Limburgs Universitair Centrum in Diepenbeek, Belgium and started working in the Department of Biostatistical Science at the Dana-Farber as a research fellow last September. Until now, Lieven has learned a lot about all kinds of practical aspects related to clinical trials and is involved in a variety of statistical aspects of clinical trials. These include designing such studies, monitoring clinical trials which are open to patient accrual, performing interim and final analyses and providing statistical parts of papers dealing with these studies. More specifically, he is dealing with thoracic studies performed by Eastern Cooperative Oncology Group and with leukemia studies done by the Dana-Farber and other consortium institutions. Lieven likes cycling and jogging along the Charles River and enjoys the atmosphere around Harvard Square in Cambridge.

Andrea Foulkes graduated from our doctoral program in February of this year and we are glad that she decided to stay on as a research fellow in our department. Andrea has always enjoyed mathematics but became interested in its application to public health research after working at a refugee camp in Kenya during a semester abroad program during her junior year of college. Like others of our new research fellows, Andrea's experience so far in her current position has been great. At present, she is working on developing methods for understanding the relationship between mutations in the AIDS virus and resistance to anti-retroviral therapies. Lately Andrea and husband have been spending much of their time with their newborn girl, Sophie (their sunbeam!).

Anna Legedza is another of our graduates of the doctoral program and we are glad that she also decided to stay with us as a research fellow. Anna first heard of biostatistics while on a pharmaceutical industry project in New York. Biostatistics sounded more interesting to Anna than the management consulting she was doing at the time. She also missed using her quantitative side, so she took the plunge and decided to go to graduate school to learn more about the field. Anna's experience so far as a research fellow has been great. She works on a variety of interesting projects for Pediatric ACTG under the direction of her supervisor, Michael Hughes. Specifically, she's working on Phase I/II clinical trial design research with Michael as well as with L.J. Wei. Anna also recently collaborated with Bob Lansing and Bob Banzett in the Physiology Program on a respiratory study. Anna enjoys playing the violin with the Longwood Symphony, the orchestra of Boston's medical community. They'll be playing at the Hatch Shell on Wednesday evening, August 16, and Anna invites us to come bring a picnic and enjoy the free concert!

Yi Li became a research fellow in our department in August 1999 after receiving his Ph.D. from the Department of Biostatistics at the University of Michigan, in which his main field of interest was survival analysis. Yi discovered an interest in Biostatistics upon learning that it is a perfect combination of applied mathematics and medical sciences. This department has been a tremendously exciting place for Yi to work in so far, since he finds there are many outstanding professors to learn from and energetic colleagues to work with! Yi is interested in multivariate survival data analysis and its applications in cancer research and environmental health. He does some sports and also watches a lot of college football.

Maura Mazzetti began her position as a research fellow in the department last September. She received her doctorate in methodological statistics from the University of Trento, Italy in 1997 and after that worked as a researcher in biostatistics at the European Institute of Oncology in Milano. Maura decided to come to work with us after meeting Joe Ibrahim through an Association of Italian Cancer Research grant. Maura's experience so far in the department has been very positive. It is her first time working in the United States, and she is really enjoying it. Joe has been her advisor, and Maura finds Joe very helpful and makes the research very interesting for her. Maura's main area of research is in Bayesian survival analysis. She likes hanging out with friends, going to movies, and listening to music.

John Ritz joined us as a research fellow in September 1999 after receiving his doctorate in Statistics from the University of Rochester in 1998. John became interested in Biostatistics through lectures and workshops at the University of Rochester and from statistical conferences. Since his arrival here John has learned quite a bit on statistical methods in epidemiology through his research and project work. Currently, John is working on multivariate pooling methods and mixed effects models in epidemiology which he finds very interesting. Like Lieven, John enjoys bike riding along the Charles River.

Qiong Yang received her doctorate in Statistics from Columbia University this past September and joined our department as a research fellow after that. Qiong's interest in Biostatistics began because she was fascinated by various biostatistical models. Her experience so far in the department has been great. Qiong is currently working in the area of statistical genetics to develop statistical methods for gene mapping and study the effects of genes and gene-environmental interactions. Qiong enjoys hiking, biking, and swimming in her free time.

Department Visitors in 1999-2000

by Paul Catalano

Every year the department benefits from visiting scholars on leave from other universities and research institutions. This year was no exception. We have been very fortunate to welcome and work with five visitors who represent a broad array of interests.

Michael Escobar is from the Departments of Public Health Sciences and Statistics at the University of Toronto where he is associate professor and associate director of the Biostatistics program. During his visit (August 1999-June 2000) Michael worked in his research area of nonparametric Bayesian methods, specifically hierarchical models, Bayesian computation methods and longitudinal models. During his stay he collaborated with Nan Laird on methods for longitudinal data analysis and was also the director of the HSPH Biostatistics consulting laboratory.

Els Goetghebeur is from the Statistics Unit, Department of Applied Mathematics and Computer Sciences at Ghent University in Belgium where she is Associate Professor. Els visited during the first half of 2000 on a grant from the Belgian government that was also able to fund visits to Boston for a postdoc and a couple of her graduate students. Her research areas are causal inference and noncompliance in clinical trials, missing data and sensitivity analysis. During her stay Els worked on interval censoring with Louise Ryan, causal proportional hazards models with Dave Harrington and Sudeshna Adak and began work with Jamie Robins and L.J. Wei. Els is no stranger to the department, having been a visiting student with Louise Ryan more than 10 years ago. She says of her recent visit, "It was great to be back here for a longer visit after all those years."

Yutaka Matsuyama is assistant professor in the Department of Biostatistics, School of Health Sciences and Nursing at the University of Tokyo where there are undergraduate and graduate programs in Biostatistics. Yutaka's research is in the areas of non-compliance and missing data. Throughout August 1999-June 2000 Dr. Matsuyama collaborated with Nan Laird and Jamie Robins on various research projects.

Gita Mishra visited the department from January to June 2000 from the School of Mathematical and Physical Sciences, University of Newcastle, Callaghan, NSW, Australia where she is currently Lecturer in Statistics and an investigator in the Australian Longitudinal Study of Women's Health. Gita's line of research is in methods for longitudinal data analysis and missing value imputations. During her stay she was involved in the Department's longitudinal working group and worked with Nan Laird, David Wypij and Garrett Fitzmaurice on several research projects.

Chris Olola came to HSPH from the Severe Malaria in African Children (SMAC) Project based in Kenya where he is the data coordinator for five African countries. Chris visited during the first half of 2000 during which time he collaborated with David Wypij, attended some courses and worked with STATA. Chris says of his stay here that the program offered "significant assistance to me in my areas of responsibility such as malarial data management and scientific data analysis for the SMAC Project."

Giovanni Parmigiani Delivers 1999 Lefkopoulou Lecture

by Florin Vaida

Dr. Giovanni Parmigiani, now Associate Professor in the Division of Oncology Biostatistics, Johns Hopkins University and previously at the Institute of Statistics and Decision Sciences, Duke University, delivered the 7th annual Lefkopoulou Distinguished Lecture at HSPH on September 23, 1999. Dr. Parmigiani spoke on "Breast Cancer Genes: Modeling and Medical Care". His presentation focused on the important statistical, medical, public health, and ethical issues related to understanding genetic susceptibility to breast cancer. Two genes, BRCA1 and BRCA2, have been identified as associated with susceptibility to breast cancer; Dr. Parmigiani discussed statistical models that aim at estimating the probability of carrying a genetic susceptibility mutation, based on a given family pedigree with history of breast and related cancers. Using a Bayesian paradigm, the model incorporates biological evidence about the prevalence, penetrance and inheritance mechanism of the targeted genes. The analysis is incorporated in a statistical software package, named BRCAPRO, that may be used by practitioners and genetic counselors as a "black box" for individual predictions.

Dr. Parmigiani obtained his Ph.D. in Statistics from Carnegie-Mellon University in 1990, under Jay Kadane, after which he joined Duke University as an assistant professor. He visited our department and Dana-Farber Cancer Institute in the Fall of 1994. At the end of the lecture, David Harrington conferred the Myrto Lefkopoulou award to Dr. Parmigiani. The lecture was preceded by a piano recital of the Greek pianist Lola Totsiou, childhood friend and classmate of Myrto's.

The Award was established in memory of Dr. Lefkopoulou, a faculty member and graduate of the department who died of cancer at the age of 34. Each year the lectureship is awarded to a promising statistician within 15 years of receiving an earned doctorate for contributions to collaborative or methodological research in statistical applications to biology or medicine, and/or for excellence in the teaching of biostatistics.

Diversity Initiatives

by Louise Ryan

This has been a good year for the various diversity and Summer Minority Programs in the Department.

Established in 1994, the Summer Program in Biostatistics is a 4-week intensive program targeted toward underrepresented minority students in the undergraduate mathematics, statistics, computer science, and related fields. The program seeks to expose students in these fields to public health issues and to raise awareness of the impact biostatistics has in public health research.

The program runs in June of each year and involves an introductory course in biostatistics, general lectures on public health research and a research project. The program provides a great opportunity to recruit some outstanding students to our doctoral programs (Mahlet Tadesse, Kevin Roberts and Cassandra Arroyo were all graduates from the program). A former summer program attendee, Ronnie Sebro, will start in the doctoral program this September.

Two years ago, our Summer Minority Program expanded with the award of a large NIH grant which supports predoctoral trainees as well as summer research interns. This award allowed for the Summer Research Internship Program to be run jointly with the Department of Health and Social Behavior at the School. The internship program recruits undergraduate students with pre-existing research experience and provides them with 10-week intensive training in community-based research with a scientific focus. To learn more about our Summer Minority programs visit our website at www.biostat.harvard.edu/courses.

Since the fall, we have been lucky to have Melissa Sanchez on board as our new program coordinator. Our previous coordinator, Sharmon Davis, has decided to head off to grad school in psychology. Melissa brings a lot of great skills and energy to the program. She spent a number of years with CBAR (see article by Steve Lagakos on page 2), working primarily as Victor DeGruttola's assistant. In addition to coordinating our minority programs, Melissa works closely with David Wypij with the Department's recruiting and admissions.

Student and faculty connected with our programs have been active over the past two years at a newly initiated "Diversity Workshop" at ENAR. This year's workshop, held during the spring meetings in Chicago, was a great success. Organized this year by Dr. Amita Manatunga from Emory University, the workshop aimed to create a forum to discuss how our profession can do a better job of attracting students from underrepresented minority groups to consider careers in biostatistics. Many past and current students from our Department and summer programs attended including Cassandra Arroyo, Renee Moore, Kevin Roberts, and Cheryl Jones.

Among our major goals for the next year are improving our diversity web site and preparing for the spring 2001 resubmission of our grant.

Mentors, Interns, and Students in the Summer Program

Our New Practical Training Program

by Michael Hughes

Recognizing the importance of real-life practical experience in the application of biostatistics, the Department of Biostatistics has initiated the Practical Training Option as part of the Doctoral Program. The Program provides students with the opportunity to be involved in one of the many collaborative projects with which the faculty work and which are often at the frontline of national and international research efforts. Examples include cancer research through the Dana- Farber Cancer Institute including the Eastern Oncology Cooperative Group, HIV research through the Center for Biostatistics in AIDS Research including the Adult and Pediatric AIDS Clinical Trials Groups, and a variety of research projects in environmental health.

The Option combines training while working as a biostatistician on collaborative projects with an academic component that includes a seminar series at which students present and discuss aspects of their work. The seminar series will include topics such as study design, data management, computational issues, report writing, as well as providing students with an opportunity to present results from the research in which they are involved. It aims to complement other components of the doctoral program which offer applied biostatistical training, such as participation in the consulting lab, and the summer project that students complete at the end of their first year.

Students will generally start the option in their third year upon completion of most of their coursework, and will work part-time for about two days a week for two years, in parallel with undertaking thesis research. Students with relevant previous training, such as a prior Master's degree in statistics, may start the Option during their first or second year of coursework. The extended period of training allows students to be fully involved in the realities of study design, conduct, analysis and reporting, including the interactions with clinical investigators which are so critical to the success of collaborative research. The Option therefore provides a complement to thesis research and coursework in both theoretical and applied statistics, and aims to provide a broad and thorough foundation on which to build a career in biostatistics in academic and government settings as well as in the pharmaceutical industry.

Hooray to Our Graduating Students!!

by Evelyn Ophir

This year eleven of our students graduated from the department's doctoral program: Tianxi Cai, Debbie Cheng, Chris Corcoran, Tony Fitzgerald, Aaron Foster, Andrea Foulkes, Amy Herring, Steve Horvath, Anna Legedza, David Shera, and Peter Slasor.

We are fortunate to be able to still enjoy the company of Andrea Foulkes and Anna Legedza, who upon completion of their thesis defense decided to remain in the department as research fellows. Tianxi Cai also stayed on with us as a research fellow but is leaving in August to take a position in the Department of Biostatistics, University of Washington as assistant professor. Amy Herring and David Shera worked as research assistants at the Dana-Farber this year and are leaving us this summer as well. Amy has just started as an assistant professor in the Department of Biostatistics, University of North Carolina. As of August 1st David is joining the University of Pennsylvania Medical School, Department of Pediatrics as an assistant professor with a joint appointment in the Department of Biostatistics.

Debbie Cheng, Chris Corcoran, Tony Fitzgerald, and Steve Horvath directly headed for academic institutions outside Harvard. Debbie accepted a position as a Research Assistant Professor in the Department of Environmental Health Sciences at New York University. Chris is working as an Assistant Professor in the Department of Mathematics and Statistics at Utah State University. Tony is a Statistician in the Department of Epidemiology and Preventive Medicine at the Royal College of Surgeons in Dublin, Ireland. Steve is currently a postdoctoral fellow at the Institute of Medical Biometry, University of Bonn and accepted a position starting September 1st as assistant professor at the University of California, Los Angeles with a joint appointment between the Department of Human Genetics and the Department of Biostatistics.

Aaron Foster is working as a Senior Consultant for KPMG LLP in New York, a professional advisory firm that provides consulting, tax and legal, financial advisory and assurance services worldwide. Peter Slasor is working as a Senior Biostatistician with Biogen, Inc. in Cambridge.

In addition to our doctoral graduates, six of our students graduated from the department's Master's program: Cassandra Arrroyo, Meredith Goldwasser, Peter Park, Kevin Roberts, Amy Stubbendick, and Mahlet Tadesse. Peter, who has a Ph.D. in Applied Mathematics, will be remaining in the department as a research fellow. Cassandra, Meredith, Kevin, Amy, and Mahlet will continue their studies in our department as doctoral students.

Harvard/Schering-Plough Workshop High on Quality

by Richard D. Gelber

The year-end workshop sponsored jointly by the Department of Biostatistics and Schering-Plough Research Institute is an outstanding example of an academia/industry collaboration that enriches our profession. The annual event is now in its eighth year. The workshop held at the School of Public Health on June 1-2, 2000 was entitled "Measurement and Analysis of Quality-of-Life Outcomes for Drug Development". The conference provides an open forum for interchange of ideas among representatives from the pharmaceutical industry, academia, and the Food and Drug Administration. Over 170 participants were treated to lectures and discussions on a variety of issues concerning quality-of-life assessment from the statistical and regulatory perspectives. Each year a large number of former students return to the department to participate and catch up with friends and colleagues.

The use of health-related quality of life (HRQOL) as an endpoint for assessing treatment benefit within clinical trials has increased dramatically during the past decade. Statistical issues that complicate the use of HRQOL for drug development include multiplicity of domains and assessment time points (p-value adjustment) and missing data problems. The conference faculty discussed the topic from the FDA advisory committee perspective, from the regulatory perspective, and from the industry perspective. Consensus emerged that the "soft" HRQOL endpoint could inform the drug development process, but that definitions of substantial evidence should evolve to the same standards as those applied to other endpoints.

New approaches to the measurement of HRQOL including Rasch techniques and structural equation modeling were presented. Computer-based dynamic testing procedures that can substantially reduce the number of questions that a patient needs to answer to classify his/her HRQOL status were also demonstrated.

The lectures on issues in data analysis focused on a variety of approaches to the missing data problem. Generalized linear mixed models with nonignorable missing data mechanism, sensitivity analysis for pattern-mixture models, and models that combine quality and quantity of life were described. A lively floor discussion followed, during which data analysts were urged to be certain that the question being answered by the analysis was really the question of interest.

The workshop concluded with a practical application of HRQOL in clinical trials. An actual clinical trial involving HRQOL measurement in the treatment of brain cancer was presented. The data set had been previously provided to several of the workshop speakers who analyzed the data using their particular approach and discussed their findings. These presentations stimulated lively discussion. Clearly there is much progress yet to be made to define the role of HRQOL in drug development and many exciting opportunities for statistical scientists to contribute to this progress.

Workshop faculty included Laurie Burke, Joseph Cappelleri, Bernard Cole, Ralph D'Agostino Sr., James Dewey, Diane Fairclough, Richard Gelber, Joseph Ibrahim, Nan Laird, Geert Molenberghs, Stephen Sun, Marcia Testa, Ralph Turner, and Sara Zaknoen. The workshop was followed by the presentation of the year 2000 Marvin Zelen Leadership Award in Statistical Science to Professor Lincoln E. Moses.

Congratulations to Our Faculty and Students For
Outstanding Academic Excellence and Service to Our Profession!!

by Evelyn Ophir

Once again our faculty and students have been recognized this past year by many institutions including our own HSPH for contributions and service to the field of biostatistics. Dr. Joseph Ibrahim, Associate Professor of Biostatistics, was elected a fellow of the Institute of Mathematical Statistics in March of this year "for innovative and productive research on Bayesian methods and missing data problems targeted at applications in the medical sciences; for outstanding teaching; and for effective collaborative research on the treatment of cancer." Joe was also elected a member of the International Statistical Institute this past March. Dr. Mei-Ling Ting Lee, Associate Professor, HMS and HSPH, was elected ASA Fellow at the Joint Statistical Meetings held in August 1999 in Baltimore. Mei-Ling was recognized "for influential contributions in statistical applications in microbiology and medical research; for pioneering editorial work; for contributions to the theory and application of association of multivariate distributions; and for service to the profession."

Dr. Cyrus Mehta, Adjunct Associate Professor of Biostatistics, received the "Mosteller Statistician of the Year" Award this past February from the Boston Chapter of the American Statistical Association. Dr. Donna Neuberg, Lecturer on Biostatistics, received a HSPH junior faculty sabbatical award for the coming year to study statistical issues in gene expression profiling and other areas of functional genomics.

This year Dr. Louise Ryan received the HSPH Mentoring Award. Louise was nominated by a number of students, including those on her NIH grant that supports underrepresented minority students. Louise's award not only honors her, but also the Minority Training Program that the Department has worked so hard to develop over the past several years.

Andy Housman and Ming-Chih Jeffrey Kao, both first-year doctoral students, received Howard Hughes Medical Institute Predoctoral Fellowship awards this past year.

Cassandra Arroyo, a second-year Master's student, was awarded an NIH Minority Predoctoral Fellowship through the NIDDK. The mission of her grant is to explore statistical methodology in diabetes type II research in minority populations, specifically in community-based research.

Lu Tian was awarded the Robert B. Reed Prize at the HSPH Awards Dinner which took place in May. This prize is presented each year to one or two Biostatistics students for academic excellence.

Additionally, the Department presented tuition awards to several of its students in the 1999-00 academic year: Denise Scholtens and Michael Szarek were named Schering Plough scholars; Aparna Keshaviah, Andy Houseman, and Min-Chih Kao were Pfizer scholars.

Biostat Department Fosters Partnerships with Industry

by Cyrus R. Mehta

Over the past eight years the Harvard Department of Biostatistics has developed valuable partnerships with three major U.S. biopharmaceutical corporations. Through this mechanism many new ideas are exchanged between academic and industry statisticians for future methodological research and for consulting. Each partnership has its own special flavor, however, and brings different benefits to the partners involved.

Schering Plough Partnership. The oldest partnership is with the Schering Plough Corporation and was started in 1992. The highlight of this partnership is the annual Harvard-Schering Plough workshop. Towards the close of each academic year our department hosts this workshop for over a hundred prominent biostatisticians from industry, government and universities. The workshop lasts for one and a half days and presentations on a theme of major importance are made from methodological, practical and regulatory perspectives by experts on the problem. Previous workshop themes have included interim monitoring, longitudinal data, handling missing values, Bayesian inference, multiple endpoints, and quality of life. The organizers of this workshop have an extraordinary ability for identifying at an early stage problems that are likely to influence future biostatistical practice, research and software development in major ways. Sandy Heft, Senior Director of Biostatistics at Schering Plough says "while the one-and-a-half-day workshop is the highlight of our partnership with Harvard, that is not all. The real collaboration goes on behind the scenes all year with forming a program committee, exploring workshop themes, selecting the speakers and creating a workshop study group. The opportunity for our statisticians to collaborate with the Harvard statisticians in this way on a long term basis is simply wonderful and only gets better with time." We asked a long time participant, "why do you come to this workshop year after year, and how is it different from going to any other professional meeting like ENAR or IBC?" She volunteered the following comment: "This is the only meeting I know of where one can gain in-depth knowledge of an important scientific topic quickly and in a concentrated dose."

Pfizer Partnership. The initial goal of the Pfizer partnership, established in 1993, was for a Harvard faculty member to visit the Pfizer site at Groton, CT once a year and give a one-day workshop on some important topic. Under this format many interesting workshops were conducted on topics that included genomics, informative drop-outs, repeated measures and longitudinal data, non-proportional hazards models, and exact inference. This partnership is now on the threshold of evolving into a substantially deeper collaboration thanks to the efforts of Lianng Yuh, Senior Director of Biometrics and Gordon Lan, a senior research scientist at Pfizer. The old format, according to Yuh and Lan did not take full advantage of what Harvard has to offer, primarily because there was very little follow-up contact between the Pfizer and Harvard sites once the workshop ended. "What is really needed", said Yuh, "is an exchange of ideas, continued collaboration and practical guidance by Harvard experts on some of the really difficult problems faced by Pfizer biostatisticians." Thus a new format is being worked out for next year under which Pfizer will survey its statisticians for design and analysis problems that plague them and for which there are no obvious answers. A list of topics will thus be created and submitted in advance to the Harvard partnership coordinator ahead of time. Some examples of topics include strategies for conducting Phase I and Phase II multi-arm trials, and informative missing value problems. The Harvard faculty member coming down to Groton for the day is thus prepared to carry on a discussion on a topic of burning importance at Pfizer. The interaction that takes place that day will benefit both sides of the partnership. For the Harvard statistician it is an opportunity to work on an important practical problem with possibilities of grant support for methodological research. For the Pfizer statisticians it is an opportunity to get a second opinion on the line of approach they have already taken, to get fresh insights, and to build confidence in their solution to the problem. It is also hoped that the discussions will not simply terminate with the faculty member's return to Harvard, but that that the collaboration thus set in motion will continue throughout the year.

Genetics Institute Partnership. The partnership between Harvard and the Genetics Institute in Cambridge started in 1996 with informal discussions between L.J. Wei, Professor of Biostatistics at Harvard, and Jerry Schindler, Director of Biostatistics at GI, on how the two institutions could work together. It has now stabilized into a long term relationship whereby the Genetics Institute gets the benefit of consulting advice from Harvard biostatisticians on an as-needed basis. Clinicians and statisticians at the Genetics Institute have taken advantage of this consulting opportunity to get help with their regulatory submissions. Prior to an important submission the Harvard statisticians offer advice on how to present the material and act as a mock FDA advisory committee, anticipating difficult questions, verifying that the statistical concerns have been handled correctly and clearing up any potential for misunderstanding at the actual presentation. The value of these prior rehearsals is evident in the fact that so far every single submission that had the benefit of these prior rehearsals has sailed smoothly through the actual advisory committee meeting. All the tough questions that could be asked had already been anticipated in advance. In exchange for this consulting advice the Genetics Institute has awarded the Department of Biostatistics a grant that can be used to support faculty research and fund advanced graduate students. "Our partnership has worked out very positively and helped us a lot", says Jerry Schindler. "In the future we might seek consulting advice on more sophisticated rules for early stopping of clinical trials, Bayesian methods and genomics. Statisticians are currently only on the outskirts of genomic research. But now that the human genome has been mapped, statistics might play a major role in understanding what the genes do. Additional new statistical problems may also come our way because we are now a part of Wyeth Ayerst, a major pharmaceutical company."

Environmental Statistics in the Biostatistics Department

by Brent Coull

The past year has been one of growth for the environmental statistics group in the Department of Biostatistics. Currently, eight pre-doctoral students are supported by the NIEHS-funded environmental training grant, and two new postdoctoral fellows will join the group this summer. Two research associates (Sally Thurston and Edie Weller) and ten faculty members (Paul Catalano, Brent Coull, Wenzheng Huang, Joe Ibrahim, Louise Ryan, Joel Schwartz, Donna Spiegelman, Matt Wand, Paige Williams, and David Wypij) focus on environmental applications.

Current collaborations involve research on topics such as occupational adult-onset asthma, indoor-air induced "Sick Building Syndrome", outdoor allergens and asthma, cancer effects in asphalt workers, physiological mechanisms of morbidity and mortality from air particulate inhalation, spatial distributions of cancer and reproductive outcomes on upper Cape Cod, and the effects of in utero exposure to antiepileptic drugs. Over the past year, our activities have expanded to encompass a broad definition of the environment, namely social and behavioral determinants of health and their interaction with classical environmental exposures. For example, one such project involves the role of chronic stress, particularly with respect to violence, in childhood asthma.

More generally, our group is becoming active in community-based research. One example is a project we conducted in Dudley Square last summer looking at spatial patterms of diesel emissions. These collaborative projects have led to methodological efforts that include the incorporation of biomarker data into risk assessments, the incorporation of historical data into the analysis of toxicological data, small-sample methods for the analysis of neurotoxicological data, Bayesian model averaging for environmental risk assessment, geoadditive models for disease mapping, Bayesian approaches to covariate measurement error problems, and the incorporation of missing data methods into additive models. The group conducts an Environmental Statistics working group and a Measurement Error working group for seminar presentation and discussion of current projects.

Research fellows and faculty members in the group also constitute the Biostatistics core for two interdisplinary centers at the Harvard School of Public Health, the Kresge Center for Environmental Health Science and the Superfund Basic Research Program. The Kresge Center conducts research in the areas of physiology, pharmacology, pathology, cell biology, molecular biology, and epidemiology. Its primary goal is the development of new therapies and prevention strategies for environmental disease.

Initiated by the U. S. Environmental Protection Agency in 1980, Superfund is a federal program established to detoxify hazardous waste sites and to set aside funds for related environmental health research. At Harvard, Superfund research focuses on the risk assessment of exposure to heavy metals such as lead, arsenic, and vanadium, and other toxins, such as PCBs and pesticides, often found at such hazardous waste sites. In particular, these research projects relate to exposures of concern at a Superfund site in New Bedford, Massachusetts. The Superfund Biostatistics core provides statistical and data management support to environmental health researchers in areas such as aquatic toxicology, developmental toxicology, environmental microbiology, environmental medicine, inhalation toxicology, and occupational health. Further details on the Kresge Center and Superfund at HSPH can be found at www.hsph.harvard. edu/kresge and www.hsph.harvard.edu/superfund.

Each fall, our group joins colleagues from Cornell University to conduct an Environmental Statistics symposium in Western Massachusetts. Further details on this symposium and other aspects of our group can be found at our environmental statistics web page, biowww.dfci.harvard. edu/~corrigan/Envirostats. You will also find a link to this homepage from the departmental webpage.

The Newly Created
Dana-Farber/Harvard Cancer Center

by Sudeshna Adak

The Dana-Farber Cancer Institute (DFCI) has joined forces with Harvard affiliates to create the Dana-Farber/Harvard Cancer Center (DF/HCC). The existing Cancer Center at Dana-Farber has expanded to now include Harvard Medical School and four of its affiliated hospitals (Beth Israel Deaconess Medical Center, Brigham and Women's Hospital, Children's Hospital, and Massachusetts General Hospital) and Harvard School of Public Health.

The primary goals of this Center will be to create opportunities and incentives for multidisciplinary research and facilitate the growing need for collaborations among basic, clinical and population researchers. The essential component of DF/HCC will be its disease-based programs which will interact with all the discipline-based programs. Five disease-based programs (breast, gynecologic, leukemia, lymphoma, and prostate cancer programs) have been created. These programs were an amalgamation of the research focus of various inter-institutional groups of clinical, basic, and population scientists. Currently, there are ten discipline-based programs which include biostatistics, cancer epidemiology, and cancer genetics. Biostatistics is also one of seventeen core research facilities supported by DF/HCC. The Core facilities provide shared resources for members of DF/HCC and external researchers. Together, these Programs and Core Facilities support a multi-modal approach to cancer clinical research and care and a unified system for cancer clinical trials.

In the new newly created DF/HCC, Biostatistics is both a research program and a core facility under the leadership of Dr. David Harrington. The Biostatistics Core facility also integrates biostatisticians at the different member institutions of DF/HCC with senior leadership being provided by Dr. Rebecca Gelman from the Dana-Farber, Dr. Dianne Finkelstein from Massachusetts General Hospital, Dr. Bernard Rosner from Brigham and Women's Hospital, Dr. Donna Spiegelman from Harvard School of Public Health and Dr. Roger Davis from Beth Israel Deaconess Medical Center.

In addition to their past activities of providing expertise for the planning, conduct, analysis and reporting of clinical trials and epidemiological studies, the Biostatistics core will also provide advice and consultation to the DFCI Quality Control Center and the DFCI Protocol Administration Office in efficient and accurate database design and management of clinical research data; education for members of the DF/HCC in the areas of study design, data collection, computerization, and statistical methods for laboratory, clinical and population based studies.

While the Biostatistics Program and the Biostatistics Core will continue in its current activities of interdisciplinary research and collaborative projects, they face new challenges in this integrated Center. The challenges come not only due to the increase in the number of institutions and clinical trials handled by the biostatisticians, but also by the new research frontiers being created by this merger-in particular, the multi-disciplinary backbone of DF/HCC is fostering new research in statistical methods pertaining to genomics, molecular biology, and population science.

Changes in the Department's Master's Program

by Joseph Ibrahim

The Master's Degree program committee this year has recommended some mild changes to the department's standard two year master's program. Some students with more diverse backgrounds may benefit from such an alternative program. The proposed alternative program would consist of the following courses:

First Year (Fall)

EPI 201a: Introduction to Epidemiology
BIO 222ab: Basics of Statistical Inference
BIO 271ab: Statistical Computing Environments
     BIO 113b: Introduction to SAS
     EPI 202b: Elements of Epidemiologic Research
     Additional courses

First Year (Spring)

BIO 211cd: Regression and Analysis of Variance in
                    Experimental Research
BIO 214c: Principles of Clinical Trials
BIO 223cd: Applied Survival Analysis and Discrete Data
     EPI 203c: Design of Case-Control and Cohort Studies
     EPI 204d: Analysis of Case-Control and Cohort Studies
     Additional courses

Second Year (Fall)

BIO 230ab: Probability Theory and Applications I
BIO 232ab: Methods I
BIO 226ab: Applied Longitudinal Analysis

Second Year (Spring)

BIO 231cd: Statistical Inference I
BIO 233cd: Methods II
BIO 235cd: Regression and Analysis of Variance
BIO 312: Consultation

This program has several advantages. First, it accommodates diversity of students' backgrounds and is suitable for candidates who enter our program without as much background in mathematics and/or statistics as we would like. The new program does not impose any changes or new courses in the current two-year Master's program and gives our Master's students more applied training and provides better preparation for jobs in industry. Last, it allows for a smoother transition into the doctoral program.

What Do Our Students Think about Their Education in the Department?

by Erin Kammann and Bhaswati Ganguli

The Department of Biostatistics is currently in its 6th decade of educating students. As the department has grown in numbers of students, it has simultaneously adopted curriculum changes and facilitated activity in broader research areas to complement a rapidly evolving field, and to accommodate the goals and backgrounds of its students.

A delicate balance between the needs of the department's students, changes in the field, and the many facets of the education this department provides in preparing us for careers in biostatistics is necessary for the success of all involved. We surveyed current students, recent graduates, and current faculty in the department who received degrees here, asking them to offer their opinions and perceptions on three primary areas: career goals and reasons for entering the field of biostatistics, curriculum, and activity in the field. The response rates for current students in year 1, 2, 3, 4, or 5+ of the program were 70%, 75%, 100%, 70%, and 30%, respectively. Approximately 70% of the recent grads and 35% of the faculty graduates replied. Copies of the survey instrument are available online at www.biostat.harvard.edu/ ~erin/newsletter.

What inspired you to study Biostatistics?

Responses to this question were inspirational and varied from an article in the American Statistician to teaching math in high school, wanting to see applications in medicine and learning in a job at the EPA that "public health studies really help people"! Many students desired to combine a background in math with epidemiology, medicine, and/or public health. Two students credit the Summer Program in Biostatistics organized by Louise Ryan for their decisions to enter the field (and our department!). Several workplace self-taught statisticians felt the need for more formal education in the field. Three students employed were inspired by the ongoing research around them and wanted to pursue the design and statistical aspects of such studies. Many reported liking the security of a myriad of post-graduation job possibilities and the opportunity to apply mathematics and statistics to many worldly causes.

Where Do We See Ourselves After Graduation?

The table below ranks the favored career paths for current students, positions of the year 2000's doctoral graduates, and positions of all graduates between 1984-1994. From these data it appears that biostatistics/statistics faculty positions remain consistently the most favored, whereas industrial positions are less favored by current students than in previous years.

Year 2000
1984 - 1994
Faculty in Biostatistics
      or Statistics Department
Faculty in Medical School or
      School of Public Health in
      department other than Biostatistics
Pharmaceutical Company634
Non-Pharmaceutical Industry433
Medical Center or Hospital2-7
Research Scientist with CBAR5-6
Other (not Public Health or Statistics)--5

Curriculum Changes

The courses Generalized Linear Models, Probablity Theory and Applications II, and Operational Mathematics will be added to the course list for 2000-2001. Advanced Statistical Computing and Bayesian Methods in Biostatistics (previously offered alternating years) will now be offered every year. All surveyed reported to be very pleased with these changes; most wish they had been implemented sooner (especially 3rd-5th year students). All students reported having ample opportunity to provide input regarding the curriculum changes throughout the year. Marvin Zelen, current chair of the Curriculum Committee, appointed a student representative, Fan-fan Yu, who served as a liaison between students and faculty, collecting student input via several email surveys. Recent graduates reported their appreciation of a meeting held with Steve Lagakos who interviewed them about changes they felt were necessary, having completed the program.

Many advantages were noted for the curriculum changes. The availability of the Advanced Computing course in the 2nd year prior to commencement of thesis work and not having to look for such courses in Cambridge or at MIT is important to students. The availability of more theoretical coursework was also cited which some felt will make our students more competitive with students from other programs across the country. The Operational Math course will decrease variability in math backgrounds of incoming students as well as place useful mathematics in the context of statistics which is not often done in math or undergraduate programs.

Some disadvantages were mentioned as well. If course requirements are added, the degree program will be lengthened which seems unpopular to current and possibly to prospective students. More coursework will take time away from thesis work. As many students and several faculty graduates reported learning the most through reading, computing, and collaboration during dissertation work, a balance needs to be found. In addition, more coursework may leave less time in the program to spend on the substantive minor.

Suggested future course offerings include missing data, generalized additive models, bioinformatics, time series

analysis, and applied regression. Many students and several faculty graduates responded that Experimental Design and a clinical trials course should be offered every year and perhaps be made a requirement. On the other hand, several students emphasized the need for more time for thesis work and independent study, and more experience consulting rather than time spent in topics courses.

General Curriculum

Two-thirds responded positively to the usefulness of the substantive minor. They reported the imperative for biostatisticians to be able to converse with other scientists, that a substantive minor justifies our department's place in a school of public health, and that such courses provide broader scope to the purpose of one's statistical research. Several students raised the point that time in substantive courses may be better spent after one's thesis topic is chosen since often the two end up being very different. This would require more flexibility in the current training grant requirements.

Nearly every student replied to having TA'ed at least one semester here. Eighty-five percent reported having a positive experience as well as having gained insight into teaching and confidence in public speaking. The school-wide service courses were particularly popular for these reasons.

Activity in the Field

A greater proportion of 3rd-5th year students use their allocated travel funds on a yearly basis to attend statistical meetings than 1st and 2nd year students, with the 1st year students rarely using any travel money. This is to be expected given that the goals of 1st year students (passing the qualifying exam) are quite different than those of more senior students. Similarly, senior students tend to give more talks at the meetings they attend.

Eighty percent of the students regularly attend departmental working group seminars. Most senior students find the seminars to be helpful to their own research and generally informative. Most first year students reported their backgrounds to be not yet sufficient to follow many of the seminar topics. Several students praised the department colloquium series organized by Rebecca Betensky.

Thank you to Peter Gilbert and Mary Schaefer for providing the career data on 1984-1994 graduates and to Marvin Zelen and Fan-fan Yu for providing information on curriculum changes.

Doctoral graduates at Commencement: 1st row-Amy Herring, Anna Legedza; 2nd row-Debbie Cheng, Andrea Foulkes; 3rd row-Aaron Foster, Chris Corcoran, David Shera

Students Liam O'Brien, Pat Stephenson, Erin Kammann, and Max Su together in the Computer Room

(all events held at HSPH)

  1. Myrto Lefkopoulou Distinguished Lecture:
    Dr. Steven N. Goodman, Associate Professor of Oncology,
    Epidemiology and Biostatistics at the Johns Hopkins University
    School of Medicine and Public Health will speak on "When
    Worlds Collide: Statistics, Ethics and Clinical Research" on
    Thursday, September 21, 2000. Reception to follow lecture.

  2. Harvard/Schering-Plough Workshop: May 31-June 1, 2001

  3. Zelen Leadership Award Lecture: June 1, 2001

Grapevine Rhythms...

by Liz "FF" Smith

Psst! Seems like several of our statisticians are moonlighting under...a different profession? Tip to Steve: better check some of those grants. BC sent out our most reliable undercover yellow journalists to get the scoop.

Overheard: Longwood Symphony Orchestra principal second violinist Anna Legedza. "My mom encouraged me to start playing when I was eight," Anna claims. "Fortunately, she had the double blessing of loving music as well as being slightly tone deaf." BC reporters couldn't find her mom, but did find that this post-doc's all the rage with the local orchestra scene: she's also played with the MIT Summer Philharmonic and Boston Bar Association Orchestra; in the past, the Princeton University Orchestra and New Jersey Youth Symphony. Anna's not a trained singer, but sources reveal that her drinking buds, who sang in a NY City-based Ukrainian folk group, let her join provided she sang "quietly, in the alto section away from the microphones." Sounds like Anna's got a sense of humor AND talent to boot.

Word has it that Aparna Keshaviah was both crazy and cool in her recent role as Shark girl Francisca in the Longwood Players' rendition of West Side Story. Our connections in North Carolina have released tapes showing Aparna singing with the Duke Chorale, a 50-voice choir with a variegated repetoire of folk, classical, spiritual, secular, and non-secular music. Aparna's been spotted with TWO different Duke a-capella groups and even co-directed an all-female one in high school. BC sources dug deeper and found that she's been at the singing since kindergarten, and enjoys classical piano and singing in the shower on the sly.

Apparently, assistant prof Florin Vaida also enjoys singing in the shower, but notes that the need has decreased since he joined the basses of the Longwood Chorus in November. "It keeps me off the streets," he explains, "and keeps my neighbors happier." BC undercover posed as census takers and visited his neighbors, who verified that this was indeed the case. "Seems like the singing's been displaced," says one. "After his two performances with the chorus, he's definitely toned it down a bit." Meanwhile, our international sources in Europe were tracing his singing career to his native Romania and on to the University of Chicago. He's also been spotted caroling in harmony with Nan Laird's soprano at the department's hoppin' holiday fest 1999.

Speaking of parties, saxophonist and assistant professor Peter Gilbert was spied playing jazz standards with the reception trio at his own wedding-most notably, "My Funny Valentine" while his wife Kristi danced the first one with her dad. Hidden cameras at the ceremony also showed him playing a duet with his pianist dad during the seating of the mothers. We all know it don't mean a thing if it ain't got that swing... and prior to his nuptials, Peter was swinging in everything at his former University of Washington haunts. He's made rounds from the Chamber Ensemble, Saxophone Quartet, Jazz Band and Jazz Combo, to a worship band in Seattle.

Looks as though statistics couldn't keep oboist Carlos Brain away from music, either. In fact, this grad student did double duty majoring in both stats and music performance at his alma mater University of Minnesota. He's definitely been around the classical scene, having performed in full and chamber orchestras, and in chamber groups such as piano quintets and trios. He prefers orchestras but also enjoys playing with stringed instruments in chamber groups where there's more freedom. "Besides," he says, "the woodwind quintet was ignored by many a great composer, so there's hardly any good music for it." Although Carlos has been on a brief hiatus since moving to Beantown and surviving the program's notorious first year, he plans on getting back in the groove this summer on his "ahem! qual study breaks."

Patricia Stephenson's involvement with prophetic arts ministry Consecration is an example par excellence that you don't have to be a veteran of childhood music lessons to be musically involved as an adult. Patricia, who has never had any formal dance or musical training, joined the group a year ago and has been actively involved ever since. "Its main goal," she says, "is to make God known on the earth using the arts." The 12-person group incorporates dance, drama, and song, set to music of all venues. "We minister through the movement in dance and in song," notes the supa-cool grad student. BC followed the members of Consecration to hospitals, nursing homes, schools, girl scout meetings, and impoverished neighborhoods. Artistic expression to address spiritual and physical needs? Simply divine!

But seriously - why do these individuals enjoy playing music? Anna cites the amusement in her conductor's "seemingly inexhaustible supply of mildly dirty jokes." And Peter notes the utmost importance of trying to imitate Charlie Parker. But the significance of music lies deeper than simply an outlet for social fun and emotional expression. These artists mention the sophisticated, unspoken communication and connection with fellow musicians: the synchronization of rhythm and tone; the control of creativity; the spontaneity of improvisation. "The understanding of music that comes from playing," says Carlos, "is something that cannot be had by the most attentive listener...It's the difference between listening to a foreign language and actually speaking it yourself. It really changes the way you hear." Well, hear this: the activities mentioned are voluntary, with little monetary but immense personal rewards-looks like Steve won't have to review those grants, after all.

Alumni and Friends Connections

by Evelyn Ophir

The following is from some of our alums, who are all quite busy!

Rob Strawderman (ScD '92) is an Associate Professor in the Biostatistics department at the University of Michigan in Ann Arbor, where he has been since graduation. His wife Myla continues to work as a biostatistician at the UM Cancer Center. Tommy (born 11/8/97) and Emma (born 10/22/99) are keeping Rob and Myla very happy and incredibly busy. (strawder@umich.edu)

Julie (ScD '97) and Daniel (ScD '96) Scharfstein are living in Baltimore, where Daniel is an Assistant Professor of Biostatistics at Johns Hopkins School of Hygiene and Public Health and Julie is a stay-at-home mom and a part-time decision analysis consultant. They just returned from a three-month sabbatical at UC Berkeley, where they would love to move if housing were even reasonably affordable. Their daughter Kayla, who turned two at the end of has learned to count to twenty (although she skips 7, 8, and 11-19). (dscharf@jhsph.edu)

On March 31, 1996 Alicia Toledano (ScD '94) married Lee Shekter in Chicago. Marian (Pugh) Ewell (ScD '94) was one of their bridesmaids. Their sons, Jacob Samuel and Dylan Harry, were born on January 8, 1998 and September 10, 1999, respectively. For those who remember her, they still have Sunny, who will be 11 on May 7! The family moved to Providence this winter - Alicia is now Assistant Professor in the Center for Statistical Sciences at Brown University. She spends 50% of her time on her methodological research, "Multi-variate Methods for Evaluating Diagnostic Tests," funded by NCI, 40% of her time with ACRIN (American College of Radiology Imaging Network) projects, and teaches one course a year. Alicia is really glad to work with Constantine Gatsonis (Faculty '89-'98) and Joe Hogan (ScD '96) again, and the rest of the people she works with are also great! Lee also accepted a position at Brown, only a couple of buildings over from Alicia's office. In their free time, Alicia and Lee enjoy taking the boys to play in the park, and visiting with family and friends. (toledano@stat.brown.edu)

Last year Qi Zeng (ScD, '97) left Merck Inc. and joined the quantitative strategy group in Morgan Stanley Dean Witter in New York City. So now instead of working with drug efficacy and doing adverse event analyses, she is involved with interest rates, bond yield, and stock options, quite a change from her work at Merck! This career redirection is exciting for Qi, and she is learning a lot in her new position.


Andrea Troxel (ScD '96) is an Assistant Professor in the Division of Biostatistics at the J.L. Mailman School of Public Health of Columbia University, where she collaborates with researchers in cancer and complementary medicine, and is continuing research in missing data, quality of life, and clinical trials. She and her husband Adam Kosto are having a great time living in Manhattan, and they are expecting their first child in October. (atroxel@biostat.columbia.edu)

Mary Putt (ScD '98) and family (John, Ari, and Maia) have been living in the Philadelphia area for over a year. Mary is an Assistant Professor in the Department of Biostatistics and Epidemiology in the School of Medicine, University of Pennsylvania. (mputt@cceb.upenn.edu)

We would love to hear from you with recent news! Please email your news to ophir@hsph.harvard.edu.


Little Known Music Facts: See if you can match the person with the appropriate, true description.

1. Louise Ryana) singer in college band which played REM-inspired music.
2. Jonathan Frenchb) cast as a gingerbread boy in grade-school production of Hansel and Gretel; converted to music hater via NYC school-system mandatory singing of "Welcome to Canaaaal Street, Welcome to Canaaaal Street" to the tune of William Tell Overture.
3. Marvin Zelenc) revised Elvis lyrics of "I Can't Help Falling in Love With You" to "I Can't Help Proving Some Theorems, Too"; clearly did not attend same municipal school system as person matching b).
4. Rebecca Gelmand) studied piano for 10 years and performed recital at hometown Bega Town Hall.
5. Fan-fan Yue) violinist who unsuccessfully tried convincing thesis advisor (and this year's Newsletter editor) to count this year's music article as 3rd thesis paper ("it's a biostatistical publication, after all").

BIOSTAT Connections (2000)

Department of Biostatistics
Harvard School of Public Health
655 Huntington Avenue
Boston, MA 02115