Paper assigned: Pi-Sunyer, Xavier, Arne Astrup, Ken Fujioka, Frank Greenway, Alfredo Halpern, Michel Krempf, David CW Lau et al. “A randomized, controlled trial of 3.0 mg of liraglutide in weight management.” New England Journal of Medicine 373, no. 1 (2015): 11-22.

Format: Typed in Arial font and 11 size font with one inch margins and no more than two pages. Please use the provided template. There is no minimum size limit but you should address all the relevant questions of interest (see below checklist) and there should be enough detail that the gradeers can see your intent. The template is attached.

Description: This is an individual assignment; this means that this must be written up individually. You are provided a paper from the health sciences literature and asked to describe the type of data used, the motivation behind the analysis, and identify the statistical techniques (and possible alternatives) used. You also have to judge the appropriateness of the techniques, identify hypotheses for tests, note any assumptions made (explicitly or implicitly) by the paper as relevant to our course material, and asked to interpret the results, make conclusions, etc. You should be able to recognize and articulate reasons for why a statistical test is appropriate, and evaluate the validity of conclusions generated based on the application of those tests to the data sets.

Relevant sub-questions: Focus ONLY on the following outcomes for your report:

1. Difference in mean body weight change between the treatment groups using the baseline and 56 weeks

timepoint.

2. Comparison of proportions of subjects who lost >5% body weight between the treatment groups using the

baseline and 56 weeks timepoint.

3. Comparison of proportions of subjects who lost >10% body weight between the treatment groups using the

baseline and 56 weeks timepoint.

Checklist: If a checklist item is not addressed in the publication, mention that in your report.

1. Describe data used.

How was the data collected? Observational or randomized controlled? If randomized, what was given to the comparator group: another

medication (active-controlled), or a placebo and what ratio was the randomization done in for the treatment groups?

Is it only a medicinal intervention or is there other, e.g., behavioral support as well?

What was the design of the trial? Superiority, equivalence, or non-inferiority?

Was it blinded or open-label? How was the drug administered?

What was the outcome or outcomes to be measured for the sub-questions of

interest? What are the type of variables relevant to the relevant sub-questions (the prespecified co-primary endpoints) above (continuous, categorical, ordinal, dichotomous?)

Shape of distribution of variables that form the basis of relevant sub-questions (was it mentioned descriptively or graphically? If not, what can you assume).

2. Describe question of interest.

Hypotheses of interest (see relevant sub-questions above). Provide the null and alternative hypotheses.

3. a)

Identify and explain rationale for statistical techniques used in the paper.

Regarding the relevant sub-questions, what hypothesis tests OR confidence intervals, if any, did they use?

Sample size estimation

Did they do any multiple comparisons correction (may be noted as multiple testing

correction or adjustment for multiplicity)? If so, what kind was it? If not, do the authors mention possible implications? What is your take on this?

b) Reflect on the assumptions

that the investigators made in the context of the technique used.

What assumptions did they need for the tests or confidence intervals the authors used to run as expected? What was it about the data and the parameter investigated that justifies them using these techniques?

c) Provide alternative

technique(s)

Given the sample size, name an alternative test or confidence interval they could have used for relevant sub-questions?

4. Interpret results obtained.

What do their results mean in terms of statistical significance?

What do their results mean in terms of clinical significance? Do the authors mention

clinical importance? If so, what is your take on their conclusion? If not, what is your

conclusion regarding clinical significance?

What do their results (check P-values and confidence intervals) mean in the clinical

context? Keep in mind the following:

Did the randomization work? Are the groups more-or-less balanced in terms of

characteristics? What table are you using to make this claim?

Blinding

Adjustment/accounting for covariates

5. Critically analyze results and

comment on validity of conclusions generated.

Share your thoughts on the validity of conclusions generated. Did the authors conduct a reasonable and justifiable an