June 17, 2014 — Victor De Gruttola, chair of the Department of Biostatistics and Henry Pickering Walcott Professor of Biostatistics at Harvard School of Public Health (HSPH), recently answered three questions about the role of biostatistics in public health.
What does a biostatistician do?
A biostatistician’s work is driven by questions relating to the health of people—as individuals or members of population. For example, how might the benefit of a treatment vary based on an individual’s characteristics, such as genotype or exposures? Every research question poses a unique challenge. Biostatisticians consider the nature of information available from sources such as genomic studies or large medical discharge records databases, and then how the information was collected and what populations it represents. They also must consider whether the question can be answered with currently available methods or if new analytical methods are required, for example, to account for missing data or a complex interaction among the genes being observed. In close consultation with subject experts such as cancer biologists and infectious disease specialists, as well as those responsible for collecting the data, biostatisticians develop a study designs, advise on study contact, and apply quantitative methods to analyze the resulting data. Not only will research results be reported to the scientific community, but also new methods or software may be developed for application to future studies.
How is biostatistics used in public health?
Biostatistics can help identify the best way to deploy resources to treat populations. To control an epidemic, the goal is not only finding the best way to treat an infected person, but also to control spread in the population. In both infectious diseases and behavioral research, interventions provided to individuals may well impact others in the community. This requires research methods that permit investigation of the relationship between responses at individual patient and population levels.
Such research can now be done on a microbiological level. For example, it is possible to monitor not only the prevalence of HIV infection, but also to ascertain the strains of the virus that are infecting people and how they spread among various susceptible groups. Questions like this require new analytical tools because the data and processes involved are very complex.
The work by HSPH biostatisticians Marvin Zelen and Stephen Lagakos documenting a link between chemically tainted well water and cases of childhood leukemia in Woburn, Massachusetts in the 1970s was made famous by the book and movie A Civil Action. What are some other notable accomplishments by faculty members in the department?
Nan Laird, Harvey V. Fineberg Professor of Public Health and Professor of Biostatistics, and colleagues developed the EM algorithm, a technique that can be used to account for data that is missing by happenstance or by design. It uses the totality of the observed data in ways that take into account the fact that some data are missing. It is one of the most widely used advances in methods in the last 40 years and can be applied in a wide range of settings in biomedical research, as well as in genetics imaging reconstruction.
Professors Francesca Dominici and Brent Coull and Assistant Professor Corwin Zigler have developed new quantitative methods used in analyses that form the basis for air quality policy regulation for particulate matter and ozone. Their innovative methods account for the misaligned nature of the data (air pollution is measured at monitoring stations, whereas health data is often aggregated by zip codes) and also permit causal conclusions to be drawn from messy data. The impact of such work is very high. Since the estimated benefits of particulate matter reductions play such a central role in regulatory policy, these estimates must be based on the most rigorous possible science.
The Center for Biostatistics in AIDS Research (CBAR), headed by Professor Michael Hughes, developed study designs and statistical methods that proved vital in the development of highly effective strategies for the treatment of HIV infection and for the prevention of mother-to-child transmission of HIV. Contributions by CBAR statisticians were essential in helping change HIV infection from an almost certainly fatal condition to a manageable disease, leading to dramatic and permanent reductions in mortality associated with AIDS.