Statistical Analysis of Long-Term Therapies for Body Weight Management

Background and Introduction

The main aim of the experiment was to investigate any differences among the long-term effectiveness of each therapy between the treatments. It was also important to see how much variation was associated with differences among the patients and among the clinic centers. Perform an informal examination (plots, descriptive statistics) and a formal statistical analysis (appropriate statistical testing) of this data and produce a clear and concise report for the doctor to summarise the findings. The important questions to address in this analysis are

(i) Is there any evidence at the 5% significance level that the effect of the treatments on the body weights differs over time?

(ii) What percentage of variation is associated with variation of body weights between the patients?

(iii) Is there any evidence at the 5% significance level that the long-term effect differs among the treatments? You might consider to use the percentage of weight loss at the end of the therapy compared to the initial weight to quantify the long-term effect.

(iv) What percentage of variation is associated with the long-term effect between the clinic centres?

(v) Are you able to conclude that one treatment is better than the others?

 

Report Guidelines

Your report should explain the results of your analysis in a way that non-statisticians would find simple to read and to understand. The report should be concise and include only relevant information and output. It should be produced in the format of a standard scientific report (maximum 8 pages excluding R code; font size 12; single spacing). i.e.

Background and Introduction (5 marks)

This should include brief details on the study, why it was conducted, the background health context and relevance of the research, supported by proper references. Students are expected to conduct some research into the background beyond what is given in the project description.

Materials and Methods (10 marks)

Include information on the study design used to collect the data and brief details of the statistical methods and models employed (including model assumptions and constraints), as well as the statistical software being used and the significance level set for any statistical tests. You should also include a discussion of why you treat some of the factors as fixed or random effects. It is also where the marks for correctly reading in the data and selecting your own data set are awarded.

Results (20 marks)

Give a clear but brief interpretation of the output which would make sense to anyone reading the report with a limited knowledge of statistics. Discuss the assumptions of your statistical model, the validity of each and any implications this has on the results.