Our Unique Perspective to Biostatistics
Biostatistics comprises the reasoning and methods for using data as evidence to address public health and biomedical questions.
It is an approach and a set of tools for designing studies and for quantifying the resulting evidence, for quantifying what we believe, and for making decisions.
Biostatistics research here at the Johns Hopkins Bloomberg School of Public Health is characterized by a commitment to statistical science, its foundations and methods, as well as the application of statistical science to the solution of public health and biomedical problems. The two-way arrows in Figure 1 indicate that research on foundations, methods, and applications is mutually supportive. To be excellent, biostatistical research must be built on a foundation of first-rate public health and biomedical research, like that which occurs at Johns Hopkins.
Research on foundations has as its goal the development of better strategies, or ways of reasoning, for empirical research. For example, past chair William Cochran demonstrated how observational studies can be used to draw inferences about the causal effect of a treatment on a health outcome. Jerry Cornfield showed how case control studies can be used to draw valid inferences about parameters in prospective models. Richard Royall led a transition in statistical reasoning from decision methods (p-values, tests of hypotheses) toward likelihood methods, which quantify scientific evidence.
Research on statistical methodology has as its goal the creation of new tools for drawing inferences from data. To illustrate, Scott Zeger, together with former faculty member Kung-Yee Liang, and Mei-Cheng Wang developed methods for regression analysis with correlated responses. Dan Scharfstein and colleagues have developed graphical techniques for assessing the possible impact of missing data in clinical trials and observational studies. Constantine Frangakis and colleagues have developed principal stratification, a ground-breaking method to infer causal relationships. Michael Rosenblum has pioneered adaptive designs by which to more efficiently and effectively conduct clinical trial research. Hongkai Ji has led in the development of methods to delineate gene transcription pathways at both the bulk and single cell scales. Brian Caffo, Ciprian Crainiceanu, Martin Lindquist, Vadim Zipunnikov and their colleagues are advancing methods to interpet data of massive scope as arise in neurological images, accelerometers, and other advanced research technologies.
Biostatistics also includes research on important substantive questions. For example, faculty member Roger Peng and colleagues have used multiple national databases to determine the effects of air pollution on mortality across the 90 largest American cities. Marie Diener-West, Jim Tonascia, and others have led or collaborated in clinical trials of new therapeutic treatments. Nilanjan Chatterjee is creating polygenic scores by which to identify persons at risk for developing major diseases. Ingo Ruczinski and his colleagues are identifying genetic determinants of cancers, autism, cleft lip, and other diseases. Karen Bandeen-Roche leads in programs, educational initiatives and research to determine the causes and course, and ultimately to postpone the onset, of disability and frailty in older adults. Throughout its history, the Department of Biostatistics has embraced a broad definition of our discipline, including foundations, methodology, and applications. Our faculty's commitment to this inclusive perspective and the support of the School's administration and faculty are two of the intangible yet critical components of the Department's current and future success.