August 11, 2015
In the two months since the 2015 Duke Commencement, the Biostatistics and Bioinformatics Department is pleased to report 100% employment by the graduating class in their field. Biostats alumni have obtained positions with businesses such as Seattle Genetics, comScore, Green Key, PPD, and Tiny Co. Our graduates are also working at the Department of Veterans Affairs and also in several Duke organizations such as the Duke Cancer Institute, the Duke Biostatistics Core, the Duke Clinical Research Institute and the Duke Aging Center. Four alumni have been accepted into Computational Biology and Biostatistics PhD programs. You are off to an impressive start!
July 29, 2015
Fan Li has received a student travel award for the 11th International Conference on Health Policy Statistics (ICHPS), held in October at Providence, Rhode Island.
He will be presenting his work on the ‘spatiotemporal quantile regression model for emergency department expenditures’ in a contributed paper session. Health costs have been rising steadily in recent years and community-oriented efforts to reduce these costs while maintaining the quality of patient care are desired. Thus it is important to uncover the geographic variation of such expenditures and identify geographic areas that can be targeted for interventions to reduce a number of preventable costs, possibly associated with non-urgent ED use. Quantile regression models are often useful in the analysis of medical expenditures, as there is often little spatial and temporal variation in mean expenditures but more pronounced variation at the extremes. By modeling random spatial and temporal components, the proposed model helps to visualize the change in expenditures over space and time and to identify high-expenditure regions in Durham County, North Carolina. This work is under the supervision of a previous department faculty member, Dr. Brian Neelon.
July 20, 2015
One of the newest members of our Department, Anna Maria Masci, PhD, is unique at Duke in that she is not only an immunologist, but her primary research is in the field of biomedical ontology development. Ontology is a controlled, structured vocabulary intended to represent knowledge within a particular domain. Ontology terms have logical relationships to each other and to terms in other ontologies to allow for reasoning and inference. Biomedical ontologies allow for annotation and integration of scientific data within particular domains of science and medicine, as their logical structure facilitates data analysis.
Anna Maria’s work in biomedical ontology is complimented by extensive experience in laboratory research in immunology, molecular and cell biology. Her prolonged wet lab experience blended with ontology development expertise makes her a very rare combination of skills in the Duke environment.
Dr. Masci’s previous research has focused on ontologies for both basic and clinical applications, in collaboration with researchers both at Duke and other institutions. She directed efforts to revise the method for representing cells by developing an improved ontology of dendritic cells. This innovative method has been more widely used for the ontological representation of cells of hematopoietic lineage and has been used as a model for the revised upper level reference ontology: Cell Ontology. Other work has focused on developing an ontological representation of Liver Immunology, a Toll‐like Receptors signal pathways, a MethodO ontology formal representation of laboratory assays and statistical methods, and more recently on Tumor Microenvironment Ontology.
Current Duke collaborations include: LungMap project Dr. Scott Palmer, Vice Chair for Research, Department of Medicine; Biobank Ontology with Helena Ellis, Director of the Duke Biobanking Program; and developing ontological approach for a description of Bipolar and Generalized Anxiety Disorder with Meredith Zozus and Cancer with Susan Halabi, of the Biostatistics Department. Welcome to the Department Dr. Masci!
June 14, 2015
A University of Pennsylvania research team lead by Mitchell Lazar recently completed a study of a leading diabetes drug that has important implications for personalized medicine. The UPenn study focused on a class of anti-diabetes drugs known as thiazolidinediones (TZDs). TZDs can be effective - in some people. But for 20 percent to 30 percent of patients, they are useless and can even cause serious side effects. The Penn team provided new insights into how genetic changes in the regulatory regions of the genome contribute to those differences in effectiveness between individuals.
Biostatistics & Bioinformatics faculty member Tim Reddy provided a preview of the article. Click here to read the article.
In an interview with the Philadelphia Inquirer, Reddy said, “Finding the mutations that matter is the hardest part. However, projects like this, which identify the genetic predictors of people who respond to medications vs. people who don't, are the first steps toward making personalized medicine a reality.
And, Reddy said, like any good study, "this opens up a lot more questions and a lot more opportunities. Nuclear receptors are the targets of countless drugs for other diseases. I'm really excited to see how the principles that were revealed here can be generalized for other drugs that target" these molecules.
July 10, 2015
In his recent Blog post, Keith Bradnam, an Associate Project Scientist at the UC Davis Genome Center, discussed the gender bias that disproportionally overrepresents male speakers at genomics/bioinformatics conferences. Studying further, Bradnam tracked the number of men and women in senior research roles. It was gratifying (although not surprising) that he noted that the most equitable result for any one academic institute with at least 15 senior research scientists was the Duke Department of Biostatistics and Bioinformatics.
Last year Bradnam conducted a survey to look at gender bias in bioinformatics spanning career stages and found essentially no gender bias at all in the stages prior to the level of faculty. This indicates that there are plenty of women in academia but there are barriers when it comes to attaining senior research positions. To combat this, he suggests going forward that all conferences aim for at least a third of all speakers to be female. Click here to read his blog post and check out the data.
June 20, 2015 by Karen Butler
Susan Halabi, PhD, Professor of Biostatistics and Bioinformatics, was named in June a Fellow of the American Statistical Association (ASA), the nation’s preeminent professional statistical society.
“I congratulate Dr. Halabi on being honored as a new ASA Fellow,” said David R. Morganstein, president of ASA. “Her accomplishments have contributed greatly to the advancement of statistical science and have rightfully earned her the respect and admiration her ASA peers.” To be recognized as an ASA Fellow, each honoree must make outstanding professional contributions to the field of statistical science. Among other accomplishments, Halabi is being recognized for her outstanding statistical leadership, innovative designs and analyses in cancer clinical trials and exceptional development and validation of prognostic models of clinical outcomes.
“I am extremely honored to receive the ASA award and I am also grateful for the recognition of my contribution to the statistical science and cancer scientific communities,” said Halabi, a resident of Durham. “Being honored as an ASA fellow would not have been possible without the support, inspiration and encouragement that I have received from my mentors, colleagues and family, for whom I have the deepest gratitude. I look forward to continuing to make important contributions to both statistical research methodology and to clinical research as it will permit me to draw on my strongest abilities and to provide the greatest contribution to my profession and, most importantly, to cancer patients.”
The American Statistical Association is the world’s largest community of statisticians and the second oldest continuously operating professional society in the United States. Its members serve in industry, government and academia in more than 90 countries, advancing research and promoting sound statistical practice to inform public policy and improve human welfare.
June 4, 2015
Researchers from UNC, Duke and N.C. State University have been awarded a five-year, $10.4 million renewal grant from the National Cancer Institute. The Statistical Methods for Cancer Clinical Trials program project grant, initially funded in 2010, seeks to develop new statistical methods and computational tools to dramatically improve the efficiency of the cancer clinical trial process and ultimately to improve the health and longevity of cancer patients.
The renewal award will continue to support a major collaborative and multidisciplinary effort that takes advantage of the unrivaled concentration of leading experts in cancer biostatistics, genomics and clinical across the three institutions and the two highly-rated cancer research institutes at Duke and UNC. During the next five years, the grant investigators will build and expand upon the accomplishments of the inaugural cycle to develop cutting edge methods incorporating clinical, genomics and observational data to rigorously and efficiently address the pressing questions and problems in cancer research. Software, accompanied by documentation and tutorials, to implement the methods developed by the investigators are made available to the cancer research community through the program’s website.
Michael Kosorok, from UNC’s Gillings School of Global Public Health, serves as the contact principal investigator and is joined by Kouros Owzar, PhD from the Department of Biostatistics and Bioinformatics at Duke and Marie Davidian, PhD, from the Department of Statistics at North Carolina State University as principal investigators. Xiaofei Wang, PhD from Duke serves as a co-principal investigator and Andrew Allen, Kimberly Blackwell, Raluca Gordan, Terry Hyslop, Sin-Ho Jung, Zhiguo Li and Yuan Wu as co-investigators. In addition to the Department of Biostatistics and Biostatistics, the Duke School of Medicine and the Duke Cancer Institute have provided additional support for this research.
May 13, 2015 by William Alexander
In many ways, rocket science has nothing on cancer research. Rocket science is firmly based on the reproducible domain of physics. Rockets and fuels can be replicated en masse, and identical rockets will fly the same distance the first time as they do the thousandth.
May 12, 2015
Andrzej Kosinski, has been named the Principal Investigator for a new DCRI data analytic center for a study funded by the Agency for Healthcare Research and Quality (AHRQ) with the US Department of Health and Human Services.
This project will use data from the Society of Thoracic Surgery General Thoracic Surgery Database and Medicare Inpatient Claims data to provide the first longitudinal lung cancer therapy outcomes on a national level. Researchers will also compare various approaches to surgical therapy for lung cancer and will explore cost across these approaches.
The study will focus on patients who are 65 years and older and will include data on approximately 40,000 patients.
April 29, 2015
Meredith Zozus has received nearly $1.4 million for a research project to expand involvement of patients in determining the accuracy of medical data used in research. The funding for this study, “Measuring and talking to Patients about the Accuracy of Data used in Patient-Centered Outcomes Research” comes from the Patient Centered Outcomes Research Institute (PCORI).
The study will compare self-reported data and electronic health record data to identify discrepancies. The researchers will then interview patients to identify the cause of the discrepancy and determine which data source is accurate and will examine the proportion of discrepancies that patients report to health systems and the likelihood that a reported discrepancy results in an actual change in the health record. Ultimately, the researchers want to know whether the changes affect guideline-based clinical decision-making.
“Eventually, we want to develop and test algorithms combining both self-reported and electronic health record data sources to see if we can develop algorithms that outperform either data source alone,” Zozus says. Click here to read more.
April 27, 2015
With the increasing use of high throughput experiments to derive scientific results, the Biostatistics and Bioinformatics Department will offer a new 6 week course for students, postdocs, research staff and junior faculty to cover guidance on proper procedures to ensure experimental and statistical validity. The course will be held From July 13 to August 21. For more information, visit our course page.
April 6, 2015
The Biostatistics and Bioinformatics Department is pleased to announce the upcoming Duke – Industry Statistical Symposium to be held October 22-23, 2015 on the Duke University Campus. The theme is “Challenges and Innovations for Pharmaceutical Products Development.”
The symposium will promote research and collaboration among colleagues from industry, academia, and regulatory agencies to discuss challenging issues and recent advances related to the clinical development of drugs and devices. Department Chair Liz DeLong and representatives from industry and regular agencies will give opening remarks. Dr Yi Tsong from the Center for Drug Evaluation and Research (CDER) of the Food and Drug Administration will give the keynote address on Duality of significance test and confidence interval in drug development.
This event is intended for researchers from pharmaceutical companies, contract research organizations (CROs), regulatory agencies and educational institutions. The Department recognizes the following sponsors for this symposium: Abbvie, Boehringer-Ingelheim, Parexel, Quintiles and SAS. For more information, please see our Symposium page.
March 10, 2015
The Prostate Cancer Foundation (PCF) announced 3 Challenge Awards to support discoveries for the treatment of lethal prostate cancer. Susan Halabi is part of a cross disciplinary team headed by Thomas Jefferson University that was granted the first 2015 PCF Challenge Award, titled “Optimizing First Line Treatment for Men with Castrate Resistant Prostate Cancer.” This study will examine whether the status of the retinoblastoma gene in tumors can function as a biomarker to determine whether patients with lethal prostate cancer will respond to abiraterone (Zytiga®) or chemotherapy.
The Challenge Awards have a shared goal of putting men with advanced prostate cancer back into longer remission when existing drugs no longer are working. Some men have significant increased survival from prostate cancer treated with chemotherapy; others do not. This study will move research forward on correctly predicting which patient is which.
PCF Challenge Awards are multi year awards composed of teams of scientists from 3 or more cancer centers. In order to conduct pioneering research, these entrepreneurial scientists require large investments in areas that fall outside the parameters of traditional funding organizations. These are the first major awards funded by PCF in 2015. Click here to read more.
Feb. 18, 2015
Feb. 11, 2015
Heart disease is the leading cause of death in the United States. A new study published in the American Journal of Transplantation titled "National decline in donor heart utilization with regional variability: 1995-2010” has found that while the need for heart transplants is rising, a growing number of surgeons and transplant centers are increasingly rejecting donor hearts. National trends in donor heart acceptance for transplantation were studied by analyzing data from the Organ Procurement and Transplantation Network for all potential adult cardiac organ donors between 1995 and 2010. There was a significant decrease in donor heart acceptance.
The analysis also demonstrated regional variability, showing some regions using a relatively higher proportion of donor hearts than others. This suggests that donor heart acceptance practices are not standardized and indicates a need for clinical guidelines for donor heart acceptance, along with more intense efforts to increase the use of donor hearts in areas with relatively low utilization rates. "Over time heart transplant doctors have gotten pickier about the hearts they transplant but there has been little concordant evidence that this has improved transplant recipient outcomes. Instead the result has been a larger waitlist with longer waiting times," said Ben Goldstein, a senior author of the study.
There are currently no standardized, evidence-based donor heart acceptance guidelines in the US. The study concludes that a more systematic manner of evaluating hearts based on scientific evidence could help increase the number of heart transplants performed. Click here to read more.
It has long been accepted that there are four enzymes, called hexokinases, that kick-start the body’s process of metabolizing sugar and getting energy from food. But according to ground breaking research by scientists at Duke and Northwestern universities, the hexokinase team actually has a fifth player. “This swims against the past 40 years of research and what we thought we knew, hexokinases are critical to basically all of our energy production. Finding a fifth one opens the door to more study into how we metabolize sugar, as well as genetic links to metabolic disorders.” said Tim Reddy, Ph.D., a senior author of the study.
The new protein is called HKDC1, and the researchers report that this enzyme may be a genetic predictor for whether an expectant mother develops hyperglycemia during pregnancy. Hyperglycemia is a potentially harmful environment for a growing fetus and can contribute to obesity and diabetes later in the child’s life. Researchers hope the new findings could lead to a test for pregnant women that indicates their potential for developing hyperglycemia. “The discovery of this gene creates a path forward to better predicting a woman’s risk,” Reddy said. “Knowing that there is this new hexokinase at play could also give us more information on how to inhibit or activate it, and anything we can to do disrupt the cycle would be an important advance to stem the epidemic of diabetes we see today.”
The findings appear in the online journal Nature Communications. Click here to read more.
This week, Circulation published an article “Hyperlipidemia in Early Adulthood Increases Long-Term Risk of Coronary Heart Disease” that is getting wide circulation in the print, radio and online media. The article is based on research that suggests that even moderate elevation of cholesterol levels in adults between the ages of 35 and 55 can have long-term impact on their heart health, with every decade of high cholesterol increasing their chances of heart disease by 39 percent. The study, conducted by colleagues at Duke, Boston University and McGill University examined data on 1,478 adults who were free of heart disease at age 55 who were part of the Framingham Heart Study, which began in 1948. "Few, if any, studies have gathered the quality of the cardiovascular data that the Framingham study has," said biostatistician Michael Pencina, Ph.D., a senior author of the paper. "That wealth of data collected over time made it possible to analyze the long-term effects of cholesterol in young people -- a topic on which not enough is known because it requires decades of tracking."Researchers calculated the length of time each participant had high cholesterol by age 55 and they were followed for up to 20 years to see how cholesterol levels affected their risk of heart disease. “Dr. Robert Eckel, past president of the American Heart Association, said if these new results are confirmed in future studies, it could influence guidelines on the use of cholesterol-lowering statin drugs.” Click here to read more.
On September 30, 2014, more than 150 participants, including employees of GSK, members of the Duke Department of Biostatistics and Bioinformatics and Master’s and PhD biostatistics students from Duke, attended the second Joint GSK-Duke Workshop on Critical Statistical Issues for Drug Development, held at Duke in the Hock Plaza building, the home of the Duke Department of Biostatistics and Bioinformatics. For more information, please see our News and Updates Section.
Shuyen Ho (GSK), Margaret Polinkovsky (GSK), Sara Hughes (VP and Head, Clinical Statistics of GSK), Greg Campbell (Director, Division of Biostatistics, Center for Devices and Radiological Health, Food and Drug Administration) Liz DeLong (Chair, Duke Department of Biostatistics and Bioinformatics), Shein-Chung Chow (Duke), Xiaofei Wang (Duke)
July 28, 2014
Dr. DeLong and Dr. Chow were invited by Dr. Junzhi Wang, Deputy Director-General of National Institutes for Food and Drug Control (NIFDC) of China Food and Drug Administration (CFDA) to sign an MOU on biostatistics and bioinformatics collaboration between Duke B&B and ICDC (Institute for Chemical Drug Control) of NIFDC in Beijing, China. The MOU was signed by Dr. DeLong (Duke B&B) and Dr. Huaxin Yang (ICDC of NIFDC) under supervision of Dr. Junzhi Wang followed by group photo and seminars by Dr. DeLong. She spoke on the expanding role of biostatistics at Duke and Dr. Chow spoke on Good Statistics Practices in non-clinical drug research and development).
Dr. Dejiang Tan (Deputy Director of ICDC) and Dr. Chow have been identified as the primary contact people for this collaboration. Both of them will identify topics and research areas of common interest in pharmaceutical research and development and circulate within B&B and colleagues in NIFDC in the near future. The collaboration between Duke B&B and ICDC/NIFDC will provide the opportunity for our faculty to participate joint research with scientists and principal investigators at ICDC/NIFDC in pharmaceutical research and development.
July 1, 2014
A new compilation of the world's most cited scientists released by Thomson Reuters, includes Biostat faculty member Michael Pencina along with 31 other Duke researchers. The list includes 3215 most-cited scientists who are in the top one pecent of their fields. Most-cited means a paper has been named frequently in references by other papers in that field. And that "is a measure of gross influence that often correlates well with community perceptions of research leaders within a field," Thomson Reuters says. The Thomson Reuters analysis is based on their Web of Science database.