Organization: Liverpool School of Tropical Medicine
Registration deadline: 31 Mar 2017
Starting date: 03 Apr 2017
Ending date: 13 Apr 2017
http://www.lstmed.ac.uk/study/courses/statistical-methods-for-epidemiological-and-clinical-research
3 good reasons to study
- Statistics lies at the heart of most epidemiological and medical research, enabling us to identify disease risk factors, effective drug treatments and make generalisations with accuracy and confidence.
- In this course, you will be given access to a large dataset that mimics the inhabitants of a fictitious community in a tropical country, and will be asked to design projects to address a series of research questions relating to tropical health issues.
- Our emphasis on developing skills and practical experience, from designing analyses, performing tests, and looking critically at statistics, will help build your confidence in dealing with numeric data to enhance your employability.
Programme Introduction
This course aims to provide students with knowledge and critical understanding of standard quantitative statistical methods used to evaluate epidemiological studies and clinical trials involving humans. It develops the ability of students to design, critically analyse, interpret and report the findings of a research project in a health related topic, primarily using methodologies appropriate for health surveys and observational cohort/case-control studies, but will also examine methodologies relevant for (randomised controlled) clinical trials.
Learning Outcomes - Upon successful completion of the course, you will be able to:
- Summarise the results of statistical analyses in a report format suitable for a non-mathematical readership
- Identify appropriate experimental design methodologies for carrying out health surveys, observational epidemiological studies (primarily cohort and case-control design studies) and interventional clinical trials involving human participants
- Critically evaluate and interpret the statistical analyses of data from health surveys, epidemiological studies and interventional clinical trials
- Identify and execute appropriate statistical methods for summarising data collected in health surveys, epidemiological studies and interventional clinical trials, using both conventional significance testing approaches and more contemporary effect size/confidence intervals concepts
Training Methodology
We will concentrate on the practical application of different experimental design strategies and on the interpretation of the results of statistical analysis methodologies, rather than on any detailed description of the mathematical derivations of these methods. Most emphasis will be given to methodologies relevant to epidemiological research.
You will be given access to a large data set, mimicing the inhabitants of a large (fictitious) community in a tropical country. Students will be required (either individually or in pairs depending on the number and background of the students who attend) to:
(a) design projects to address a series of research questions relating to (tropical) health issues in humans (b) select an appropriate number of individuals from the data set to mimic the conduct of each study design (c) apply appropriate statistical methods to evaluate the data sets drawn using standard statistical computer packages such as EpiInfo, SPSS and R.
The course will be taught intensively over 10 teaching sessions. Two further sessions are offered and will be dedicated for the students to produce the written assignment that constitutes the assessment. The number of contact hours is high relative to other LSTM modules so that students have access to staff support whilst completing the exercises.
Course Outline
Day 1
- Overview of design issues for the conduct of health surveys, including the development of sampling frames, selection of (random / representative) samples, stratification, bias, confounding, data management and survery size estimation.
- You will then be required to design a survey to address a set health-related research question, and then to select a (unique to them) sample of individuals from the large (imaginary community) data set mimicking the conduct of the survey they have designed.
Day 2
- Overview of the basic statistical methods commonly used to evaluate health surveys, concentrating on data summary methods and evaluation of error / uncertainty; statistical methods used to adjust paramater estimates for confounding factors and clustering / stratification effects will be considered.
- You will then be required to carry out a simple statistical analysis of the data selected - teaching staff will be available at all times to help with practical problems encountered. The session will end with a feedback session to compare the results obtained from each student's sample - this feedback will concentrate primarily on the issues relating to the fact that each student will obtain a (slightly) different result even though they will all be addressing the same question, and how this apparent paradox can be resolved.
Day 3
- Overview of design issues for the conduct of observational epidemiological studies, concentrating primarily on cohort and case-control study designs. Students will be provided with a research question unrelated to the module data set and, as a class exercise, asked to discuss the advantages and disadvantages of addressing this question using a cohort design and then using a case-control design. This discussion will be directed at practical issues such as subject selection, bias, confounding and data management.
- You will then be given a second research question linked to the module data set and will be required to write a short essay (not assessed) outlining how they would use the data set to address this research question using (a) a cohort study design and (b) a case-control study design, again outlining the advantages and disadvantages of each approach.
Day 4
- Overview of the statistical methods used to estimate the sample size / statistical power of cohort and case-control studies.
- You will then be required to calculate the sample sizes needed for the cohort and case-control studies they designed based on the module data set in session 3 and to draw appropriate samples from the data set. The session will conclude with an overview of the statistical methods used to analyse the data collected in observational cohort and case-control studies.
Days 5 & 6
- You will perform a statistical analysis of your cohort study data using the methods discussed in Day 4. There will be an extended feedback at the end of Day 6 in which the results obtained by each student will be collated as a class exercise and the range of findings explored to identify reasons for the different results obtained.
Days 7 & 8
- You will perform a statistical analysis of your case-control study data using the methods discussed in Day 4. There will be an extended feedback at the end of Day 8 in which the results obtained by each student will be collated as a class exercise and the range of findings explored to identify reasons for the different results obtained.
Day 9
- Overview of design issues for the conduct of interventional randomised controlled trials (RCTs), covering selection of subjects, randomisation, blinding, ethics, sample size / power estimation, etc.
- You will then be given a research question and will be required to design a RCT to address this question. You will then draw an appropriate sized sample of individuals from the module data set and to appropriately allocate each person selected to receive either a new or a standard intervention. The session will close with a feedback session around the practical issues that would be encountered conducting this study in a real-life tropical situation.
Day 10
- Overview of the statistical methods used to analyse the data collected in a RCT. Particular attention will be given to concepts such as intention-to-treat and per-protocol analysis methodologies.
- You will then analyse the primary outcome measure from the data set generated in Day 9. The session will end with a feedback in which the results obtained by each student will be collated as a class exercise and the range of findings explored.
How to register:
Please visit our course page, click the APPLY button and follow the onscreen instructions
http://www.lstmed.ac.uk/study/courses/statistical-methods-for-epidemiological-and-clinical-research