Biostatistics Course Evaluation: Fill & Download for Free

This is a tough question because there is some overlap between the work of an epidemiologist and a biostatistician, especially these days.I can only really speak about biostatistics and about where biostatistics and epidemiology overlap.BiostatisticsIn general, biostatistics is to epidemiology what econometrics is to economics. Or what psychometrics is to psychology. It is a specialized subfield of statistics devoted to developing *new* methods for handling inference problems that typically arise in epidemiological/biomedical data (e.g. survival analysis).Some subfields of biostatistic include: correlated data methods, causal inference, clinical trial design, and missing data methods.Theoretically focused biostatisticiansThey are concerned with things like: high-dimensional inference, valid confidence interval construction (e.g., will this confidence interval have the desired coverage?), construction of new estimators and evaluation of the properties (e.g. bias, efficiency) of those estimators - both finite-sample and asymptotic.Given this concerns, you don’t really need that much biology/epidemiology/biomedical knowledge to be an effective theoretical biostatistician. This is true especially if you’re in a top biostatistics PhD program with a heavy theory focus. Of course, there are exceptions…it’s difficult to be a biostatistician working in statistical genetics without a solid biology background.In contrast, epidemiologists are seldom concerned with these things. You would rarely find an epidemiologist trying to device a novel statistical method for getting around informatively censored patients in his data set, for example.Applied biostatisticiansThey generally do less of the theoretical work and instead work on many applied research projects where they are not developing new methods, but rather applying existing methods in new/interesting settings. On research teams, biostatisticians typically don’t have subject-matter expertise - and no one expects them to.Instead, biostatisticians in these settings are expected to have a minimally sufficient understanding of the epidemiological/biomedical issues to develop an appropriate and error-free data analysis plan. That is, to translate the research question into a testable hypothesis. They will also actually carry out the data analysis and handle complex coding and data management issues.For example, epidemiologists will generally not be coding complicated bootstrap procedures to estimate standard errors. This is something the biostatistician will generally be expected to do.The OverlapThere is significant overlap between epidemiology and biostatistics, particularly in the subfield of causal inference. In this subfield, all the general statements I made above break down and the lines get very blurry.Researchers working primarily with observational data have pioneered this field. This is because it is very difficult to make causal inferences with this type of data and it can typically only be done under strict assumptions.Some famous researchers in this field include Donald Rubin, James Robins, Miguel Hernan, and Judea Pearl. Looking at the background of these researchers gives you an idea of just how blurry the lines get.Robins has neither a biostatistics nor epidemiology PhD. He has an MD. Yet he has done pioneering theoretical work in biostatistics. He is both a professor of epidemiology and biostatistics.Rubin was also a pioneer in this field, but is neither a professor of biostatistics nor epidemiology. He is a professor in a statistics department.Hernan has both an MD and PhD in Epi, but is a professor in both a biostatistics and epi department.Judea Pearl has a PhD in electrical engineering and developed causal methods working in Artificial Intelligence. He is a professor in a computer science department but published alongside Robins and other epidemiologists and biostatisticians.All of this is to say that in some very specialized subfields, there is no effective difference between biostatisticians and epidemiologists!

Check out the course from Johns Hopkins University  Biostatistics for Medical Product Regulation : SyllabusDescription:Provides a broad understanding of the application of biostatistics in a regulatory context. Reviews the relevant regulations and guidance documents. Includes topics such as basic study design, target population, comparison groups, and endpoints. Addresses analysis issues with emphasis on the regulatory aspects, including issues of missing data and informative censoring. Discusses safety monitoring, interim analysis and early termination of trials with a focus on regulatory implications.SyllabusThis work is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike License . Your use of this material constitutes acceptance of that license and the conditions of use of materials on this site.Course DescriptionProvides a broad understanding of the application of biostatistics in a regulatory context. Reviews the relevant regulations and guidance documents. Includes topics such as basic study design, target population, comparison groups, and endpoints. Addresses analysis issues with emphasis on the regulatory aspects, including issues of missing data and informative censoring. Discusses safety monitoring, interim analysis and early termination of trials with a focus on regulatory implications.Course ObjectivesUnderstanding the relevance and application of statistics and of statistical thinking to the regulatory process; an understanding of the basic principles of clinical studies and clinical trials, and their importance to the regulatory evaluation of new drugs, biologics and devices; and an awareness of sources of regulatory guidance, requirements, and evaluation information.PrerequisitesIntended audience is graduate students in clinical medicine and public health interested in regulatory applications, or medical product development professionals, e.g., the biologics, medical device or pharmaceutical industry.Prerequisite courses in basic biostatistics or epidemiology such as:Principles of EpidemiologyStatistical Reasoning I or IIFundamentals of EpidemiologyReadingsSee Readings page for full list.Course RequirementsQuizzes: 15%Exam: 75%Participation/Discussion: 10%

I'm not a Ph.D. student there, but having been a undergrad majoring in the department and having done work with several professors, I can attest to some of its qualities.The most distinguishable would be its excellent research, particularly on the frontiers of probability theory and statistical learning. We have incredible people like David Aldous, Jim Pitman as well as Michael Jordan, Peter Bartlett, Bin Yu, etc. I can't speak much to other domains of statistics research at Berkeley since the faculty is quite large. We're of course fantastic in the physical and social sciences as well, particularly biostatistics.Berkeley also has a tendency more than other top universities towards theory and rigor. Even applied statistics courses can be incredibly theoretical, and the content is always taught in a way that applications can be easily followed upon. Assumptions, significance, and careful analysis is always considered.The community as a whole lends itself to an awesome experience. The classrooms grow from the clear enthusiasm and skill of the professors (didactically speaking---teaching skill is sorely under-appreciated when evaluating departments). The open collaborative culture leads to a very friendly atmosphere. Moreover, the department tends to be large enough to be diverse, collaborative, and non-exclusive, but not so large that it feels foreign and divided.