In this article we are sharing a personal statement from Mohammad Kaviul Anam Khan who is currently finishing up his PhD in Biostatistics at the University of Toronto. He has been appointed as a CLTA Assistant Professor at the Department of Statistical Sciences at the University of Toronto. Previously he received his MSc in Epidemiology from University of Alberta in 2014 and MSc in Biostatistics from the University of Toronto in 2016.
Sample Personal Statement:
I am hereto commence my career goal to pursue M.Sc. admission in public health science (Field: Biostatistics) in the Department of Public Health Sciences, Dalla Lana School of Public Health at the University of Toronto. I have graduated with M.Sc.-Epidemiology degree from the School of Public Health, University of Alberta in 2014. Previously, I received an M.Sc. and B.Sc. (Hons.) degree in Applied Statistics
from University of Dhaka, Bangladesh. I have worked as a Research Assistant in University of Alberta for two years and had been involved in multiple projects. My thesis project was \Population-based evaluation of disparities in survival of lung cancer patients in Alberta, Canada”, in which we conducted a survival analysis on the data of the lung cancer patients from cancer registry of Alberta. Currently I am appointed as a Statistical Analyst in the Population Health Intervention Research Unit (PHIRU), School of Public
Health, University of Alberta. My current job responsibilities are to conduct statistical analysis to investigate relationships between diet-quality, physical activity with obesity and other health indicators of the grade 5 children of Alberta and Nova Scotia and also communicating the findings through scientific manuscripts.
During my work and research I became very concerned that for the past few decades chronic diseases such as cancer (various types), type II diabetes, cardiovascular disease and childhood obesity have become major public health burden in Canada. The challenges to deal with these diseases are immense. Every year a large number of people become cases of these diseases. Appreciable number of researches have been conducted in these fields and substantial progresses have been made. However, still there are some unanswered questions in these areas of research, which are yet to be explored. For example, a number of interventions have been applied to control the prevalence of childhood obesity but the
improvements were little. Various risk factors for cancer, cardiovascular disease and type II diabetes have been identified but without much explanation of causality. My goal as health scientist is to investigate these issues and identify the solutions of these problems. Such contribution will help policy making and implementation of proper interventions.
In my M.Sc. in epidemiology program, I have learned about the differences between association and causation, criteria of causation and counterfactual. I became very interested in these topics during my program and became more interested when I started my working as a professional. Although I have learned/read about these topics from an epidemiologist’s perspective, being an applied statistician, I always wondered how could we tackle these issues from statistical point of view. While taking the Advanced Statistical Inference course, our instructor strongly suggested me to read the book titled Statistical Models and Causal Inference” by Dr. David A. Freedman after coming to know about
my interest in causal inference. In this book I learned in detail about the foundations of statistical inference and their limitations to determine causality. The concerns raised in the book made me interested to learn about more sophisticated statistical techniques to explain causality used in the field of causal inference.
Since reading Dr. Freedman’s book I have studied some of the sophisticated methods in the field of causal inference. I was very interested to learn about concepts such as dynamic treatment regimes and targeted maximum likelihoods, since, they can be very useful to identify causal pathways of various chronic diseases. Unfortunately, most of the substantive researches use many common statistical methods, which may not be the best tool to detect causality. Inappropriate use of the methods can produce wrong conclusions and thus produce adverse impact upon policy making. For example, Bayesian inference is used extensively to determine causality in directly acyclic graph (DAG) approach. This
technique has become very popular in modeling complex hierarchical structures of data. One of the limitations of Bayesian inference is that the prior distribution has to be prespecified. Results can be misleading if the prior specification is wrong. The most common way of dealing with this problem is specifying a non-informative prior. However, if noninformative prior is selected for a parameter, then the transformed version of the parameter will have an informative prior. For example, consider a simple Bernoulli trial with parameter p. Suppose, we set an Uniform[0; 1] as a non-informative prior for p. Now, if we are interested in – log(p), then the prior for this quantity would be exponential(1) distribution, which is surely an informative prior. This problem can be dealt with Data cloning” method, which can draw classical statistical inference using Markov Chain Monte Carlo (MCMC) approach. My intension is to apply this Data cloning” method in the field of causal inference, more specifically in DAG based approaches. However, to apply such a sophisticated analysis tool to some complex data structure could be very challenging without having an in-depth knowledge on statistical methodologies used in causal inference.
After reviewing the profiles of some of the faculties of University of Toronto, I found that they are very much involved in developing sophisticated methods on Bayesian data analysis. Opportunity to learn from them will help me to gather knowledge on Bayesian data analysis, MCMC algorithms; and guide me to fulfill my research objectives. In addition University of Toronto will give me the opportunity to work on substantive researches, which will help me to show my experience and skills as an Epidemiology or Biostatistics researcher.
I am a health scientist who has a goal to contribute to chronic disease research with expertise in Epidemiology and Biostatistics, and therefore, providing useful input in policy implementation. So far, I have gained three years of research experience in the field of public health. However, I feel that although I have had a very good experience of analyzing the data of multiple chronic diseases, I still need to understand their causal pathways in a more mechanistic way. To learn about the methods in causal inference I felt that University of Toronto is the best school in Canada with the best combination of
faculties. If I get admitted to University of Toronto and continue my investigation on the causal pathways of multiple chronic diseases, it will help me progress towards my goal of becoming a successful health scientist.