Epi-Bio Small Group Session 3: Evaluating a Cohort Study

Objectives

- To introduce students to a method to evaluate a prospective cohort study
- To enhance the students’ understanding of a prospective cohort study

# Outside Preparation: Due in Small Group # 3

## Assignments:

- There are four medical schools in Massachusetts. The deans would like to know the average age of all medical students in Massachusetts. Each of these school reports the age of all their students to a statistician who reports that 24.5 is the mean age. Should the statistician report a confidence interval on the 24.5 age? Explain your answer.
- A researcher gathers a group of 100 subjects, of which half are male. She'd like to know if these subjects lose weight on a new diet fad. Baseline weights are obtained in both the males and females. They are then put on the new fad diet for six months and both the males and females are weighed again. The data are normally distributed. What statistical test should the research use?
- The distribution of grades on an exam is normal with a mean of 80% and a standard deviation of 4%. A teacher will randomly select one of the students who took the exam for a discussion about the fairness of the exam. What is the probability that the teacher will randomly select a student who scored higher than 84%? Show your work.
- Pediatricians conduct a prospective study about the efficacy of the varicella vaccine. By age 25, 2 of 100 subjects who were vaccinated as a child developed varicella while 8 of 80 subjects who were not vaccinated as a child developed varicella. What statistical tests should the pediatricians use to determine if the difference in the incidence of varicella between the two groups was statistically significant?
- A student obtains the white blood cell count in 20 randomly selected students from a class of 160. The average from this sample was 7,600. The student calculates a 95% confidence interval from 7,100 to 8,100. A teacher asks the student to interpret the confidence interval and the student replies as follows: "If I were to randomly select 20 sample groups from all students in the class of 160, each sample group consisting of 20 students, I would expect that 19 of the 95% confidence intervals calculated on each sample mean would contain the true white blood cell count of all students in the class of 160." Is the student's interpretation correct?

# Approximate Class Schedule:

30 minutes | Instructor review of key concepts from lectures |

15 minutes | Review of homework assignment; distribute paper critique |

1 hour | The class will be divided into small groups. Each group will answer the questions attached to this lab. |

15 minutes | Class discuss of answers |

# To Be Completed in Small Group Session

- Was it likely that misclassification of exposure (or non-exposure) status occurred? If not, explain your answer. If yes, explain your answer and indicate how this might have led to an erroneous study conclusion.
- Were the cohorts comparable? If yes, explain your answer. If not, explain your answer and indicate how this might have led to an erroneous study conclusion?
- What was the loss to follow-up? Was it a problem?
- Was the study blinded? If yes, explain your answer. If not, do you think it lead to an erroneous study conclusion? Explain your answer.
- Do you think there could have been misclassification of outcome status? If not, explain your answer. If yes, indicate how this might have lead to an erroneous study conclusion.
- If the study results were not statistically significant, did the study report the power?
- Comment on the external validity of this paper.