Question

2. Which two study designs are best suited for establishing temporality? | 3. In the first of two steps to determining causal inference, epidemiologists must determine is an association is free of bias, confounding, and random error. a) What is confounding and explain which epidemiological study design is best for controlling confounding, b) In studies in which random assignment is not possible, how might blas due to a known confounder be minimized. Describe four methods. c) What is selection bias and one important way to avoid selection bias? d) In a cohort study that aims to determine causal inference between two factors. there is significant loss to follow-up. When will loss to follow-up result in selection bias in a cohort study? 4. Specificity is a criterion often used in causal inference. Explain why this criterion may be supportive or not supportive of causality.

Answer 1

2. Prospective cohort and experimental study designs are best suited for temporality

prospective cohort study is a longitudinal Cohort study that follows over time a group of similar individuals (cohorts) who differ with respect to certain factors under study, to determine how these factors affect rates of a certain outcome. Experimental study design is a systematic and scientific approach to research in which the researcher manipulates one or more variables, and controls and measures any change in other variables.

3.a. Confounding variable is a Extraneous influence that changes the effect of a dependent and independent variable. It is best controlled in randomised control study where Randomized controlled where the study population is divided randomly in order to mitigate the chances of self-selection by participants or bias by the study designers

3. b. Methods to reduce confoundance

• randomization is the process of assigning clinical trial participants to treatment groups such that each participation has a known (usually equal) chance of being assigned to any of the groups. successful randomisation requires that group assignment cannot be predicted in advance. By the use of this strategy ,random distribution of confounders is done between study groups.
• restriction i.e. limiting the entrance to the study on a restricted number of subjects, on the basis of a possible confounder. It restrict entry to study of individuals with confounding factors - risks bias in itself.
• matching is comparing the treated and the non-treated units in an observational study or quasi-experiment (i.e. when the treatment is not randomly assigned)(of individuals or groups, aim for equal distribution of confounders)
• stratification is partitioning of subjects and results by a factor other than the treatment given. (confounders are distributed evenly within each stratum)

3. C. selection bias is the bias that is resulted by the selection of individuals, groups or data for analysis in such a way that proper randomization is not achieved, thereby ensuring that the sample obtained is not representative of the population intended to be analyzed.

to reduce selection bias, study sample should be clearly defined.

3.d. even a follow up loss of 20% will result in selection bias

4. Specificity is supportive for causal inferences as Causation is likely if there is a very specific population at a specific site and disease with no other likely explanation. The more specific an association between a factor and an effect is, the bigger the probability of a causal relationship.