Part A1 – Critical appraisal

Task: Using the checklist on the following page, write a structured critical appraisal of the paper below by Smulders et al (available on Canvas).

Paper:

Smulders E, et al, Efficacy of a Short Multidisciplinary Falls Prevention Program for Elderly Persons With Osteoporosis and a Fall History: A Randomized Controlled Trial. Archives of Physical Medicine and Rehabilitation 2010;91:1705-11. DOI: 10.1016/j.apmr.2010.08.004

Reporting structure and allocation of marks

Your answer to Part A1 should be structured using the sections provided in the checklist on the following page (i.e. study question, study design, results, implications). Under each of

these sections, you need to address the items listed in the checklist. The appraisal should be written in the style of an essay, using the sections from the checklist as subheadings.

Marks for part A1 will be awarded in the following proportions:

.     Study question: 10%

.     Study design: 40%

.     Results: 30%

.     Implications: 10%

.     Presentation, including logical flow, correct use of grammar/spelling 10%

Maximum marks will be given based on the following criteria:

(1) Clarity and accuracy of the descriptions of the study aspects.  For example, what was the focus of the question, how were participants recruited, what was the exposure and how was it measured, what were the study results etc.

(2) Critical appraisal of the methods used (e.g. was the study design appropriate, to

what extent have the methods chosen minimised bias, confounding and chance)

(3) Correct description and interpretation of the study results

(4) Appropriateness of considerations regarding implications of the results

Critical appraisal checklist

This checklist contains the key elements that need to be included in your critical appraisal of the paper. The checklist is drawn from the Critical Appraisal Skills Programme (CASP)

checklists created by University of Oxford.

1. Study question (10% of marks)

a.   What is the study question? Did the study address a clearly focused issue? b.   What are the PICO/PECO elements?

2. Study design (40% of marks)

c.   What type of study design did the authors employ and was this appropriate for answering the study question?

d.   How were participants recruited, what were the inclusion/exclusion criteria and could this method have introduced bias?

e.   What methods were used for randomisation and allocation concealment? Could these have introduced bias?

f.    Were patients, researchers and other study staff blind to the intervention assignment? Why is blinding important?

g.   Aside from the experimental intervention, were the groups treated equally?

h.   How were the outcomes measured? Could this method have introduced bias? Was there a power calculation?

i.    The authors say that analyses were performed according to the intention-to-treat principle. What does this mean, and what is the strength of this type of analysis?

3. Results (30% of marks)

j.    Were all participants accounted for at the end of the trial (in terms of those included in the analysis for the primary outcome only)? Was there any missing or incomplete data? Could bias have been introduced?

k.   What were the baseline characteristics, and did these differ by group? Could any differences have impacted the findings?

l.    What were the results in terms of the primary outcome? Discuss the precision and statistical significance of the results. Very briefly summarise the findings for

secondary outcomes.

4. Implications (10% of marks)

m.  How strong is the evidence, and what are the implications for public health in the UK

and globally? Consider generalisability/external validity in your answer.

(The remaining 10% of marks is for presentation and clarity.)

Part A2 – Research study design

Childhood obesity has been linked to an increase in the risk of a number of different health

issues and diseases later in life, including type II diabetes. It has been suggested that

childhood obesity increases insulin demand, which could place greater stress on the β-cells and make them more vulnerable to autoimmune attack – leading to the destruction of these cells and development of type I diabetes. However, more evidence is needed to confirm a

link between childhood obesity and risk of type I diabetes. Type I diabetes is most commonly diagnosed at about age 14.

Research question

“Does obesity in childhood increase the risk of developing type I diabetes?”

Task: As a first step in designing a study to answer the research question above, answer the following questions and justify your choices:

(1) What study design will you use, and why?

(2) Who will the study population be, including inclusion/exclusion criteria? How will participants be recruited?

(3) What would you measure in order to investigate this study question and how would you collect data about these study measures? Think about your

exposure/intervention, comparator/control, outcome(s), and any relevant

confounders/covariates. Also think about study duration and follow-up period. Describe any important definitions you will use, for instance, in defining diagnoses.

Note: You do not need to describe how you would analyse the data collected.

Marks for part A2 will be awarded in the following proportions:

.     Study design 20%

.    Population 20%

.     Study measures 60%

Part B – Data analysis

Assessment submission

Section B (data analysis) is worth 40% of the overall assessment mark for the Practical

Epidemiology and Statistics module. This assignment requires you to produce a poster

which answers the questions posed on the following page. The poster should contain a brief description of the study, details of the methods used to analyse the data, a summary of the    results of the analysis, and a discussion of the results. You should include tables and figures (as appropriate) to illustrate your analysis.

The poster should be created as an A0 size (1189mm x 841mm) and it can be produced

using commonly available software e.g. Microsoft PowerPoint. You can choose whether to

present the poster in portrait style or landscape style. A template can be found on Canvas, in the Practical Epidemiology and Statistics module page (in the assignment section). Like

other assessments, you should not put your name on your poster, but only your student ID   number. Please put your student ID in the top right-hand corner. The minimum font size is 12pt. References and abstract are optional – however, please note that if you do make

use of other sources, these must be referenced. There is no need for pictures or diagrams, and only include figures and tables that support your results. You should attempt all stages of the analyses that are described below.

Guidance on general rules for production of this document is given in the MPH assignment  submission guidelines (see the Handbook, or Canvas), which you must follow. In particular, you are reminded of the Department’s guidelines on Plagiarism, which can be found in the   Handbook. Check to ensure that your submission meets the regulations.

A PDF version of the poster should be submitted via Canvas by 12th  December 2023.

Task: What patient characteristics impact the time to cardiovascular disease (CVD) occurrence among adults

Study details

The dataset you are provided with (peas_dataset_2023.dta) is from a cohort study to

examine the relationship between patient characteristics and the time to cardiovascular

disease (CVD) occurrence among adults. The study included a large population-based

cohort of adults aged 40-65 years without a history of CVD at baseline. Participants were

followed for an extended period, during which incident cases of CVD were identified. You are provided with information on various patient characteristics, including age, gender, smoking   status, body mass index (BMI), air pollution exposure, and socio-economic status.

The task

Using the dataset provided, your task is to investigate the relationship between participant characteristics and cardiovascular disease. You should also investigate whether the relationship between participant characteristics and cardiovascular disease differs by smoking status.

The dataset you have been given includes, for each participant, the outcome, cardiovascular disease (cvd_occurrence), and participants’ demographic information. The variables are summarised in the table below.

The stages of your analysis should be:

Summary statistics table

1.   Create a table of descriptive summary statistics to describe all the characteristics of the study population, stratifying as appropriate. You should briefly discuss how the   characteristics compare across the stratification variable.

Analysis

1.   Perform univariable analysis. Examine and describe the impact of the participant

characteristics on the outcome. Place particular emphasis on the point estimate and the confidence interval.

2.   Perform multivariable analysis. Examine and describe the impact of the participant   characteristics on the outcome. Place particular emphasis on the point estimate and the confidence interval.

3.   Perform multivariable analyses separately for smokers and non-smokers. Discuss whether you think there is any evidence of a differential effect of the participant

characteristics in smokers and non-smokers.

4.   Summarise the limitations of the study (or the information you have been told about it) and what, if any, the clinical value might be from your findings.

Table of variables in the dataset

Accessing the data file

Data for this assessment is available on Canvas, in the assignments section of the Practical Epidemiology and Statistics module page. The file is called peas_dataset_2023.dta. In the event   of    any    problems   with    accessing    the   data,    you    can   email    James    Martin (j[email protected]) for a copy.

Note: you should ensure that you have a working copy of this dataset as soon as possible. Requests for extensions due to last minute problems with file access are unlikely to be viewed favourably.

Reporting structure and allocation of marks

The task for your assessment is to analyse the dataset appropriately to answer the questions posed as clearly and fully as you can.

Your work should be submitted as an academic poster, with sections (such as: Introduction, Methods,  Results,  and  Discussion),  tables  and  figures  (as  appropriate),  to  illustrate  your investigations. Statistical output should be presented carefully, and interpreted and discussed where necessary. Statistical output should not come directly from Stata. You may wish to look at the level of presentation in published papers and other academic posters as guidance. The focus  of  marking  for  this  assignment  is  on answering the question  posed,  including the reporting of your methods and statistical analysis, the interpretation of your results, and the structure of the poster. The introduction is not a key focus for our marking, so we only expect students to briefly summarise the study based on the information provided in the question.

You are not expected to find outside sources for your introduction.

You may wish to use the following sections:

.    An  Introduction  section.  Here  you  should  very  briefly  summarise  the  study  and describe the objectives of your investigations.  There is no need to include a review of literature.  This  brief  summary  should  come  from  the  information  provided  in the question.

.     A Methods section detailing all data and analysis-related methods that you have used, justifying and explaining their selection and any assumptions made.

.     A Results section. Here you should present your findings. Tables and Figures should be used where appropriate, with an explanation of their context. Be sure that the results presented answer the assignment question.

.     A Discussion section. Here you should explain what the results mean in the context of the assignment question. You should also detail any limitations of the results.