Sample Size Calculation Made Easy - Case Control Study Design -HeDaL

The session begins with an introduction to case control studies, distinguishing them from other study designs like cross-sectional and cohort studies. The focus is on understanding the sample size calculations and issues specific to case control studies. An interactive approach is encouraged, with participants asked to respond to questions in the chat box. The speaker uses a hypothetical study about neonatal jaundice and its potential impact on childhood IQ as an example to illustrate the concepts of exposure and outcome in study design.

This paragraph delves deeper into the specifics of case control studies, emphasizing their retrospective nature and starting with cases to look back at exposure levels. It contrasts this with other study designs like cohort and randomized control trials, which are prospective. The speaker discusses scenarios where case control studies are particularly useful, such as when diseases are rare or when ethical concerns prevent other study designs. The importance of selecting appropriate controls to mitigate biases is highlighted, along with the potential for case control studies to provide valuable insights despite their inherent limitations.

The speaker introduces the concept of the odds ratio, a measure of association used in case control studies, and explains how to calculate it using a two-by-two contingency table. The explanation includes the definition of odds, the difference between odds ratio and risk ratio, and the importance of understanding these concepts when interpreting study results. The paragraph also addresses potential biases like recall bias and the importance of having accurate records to minimize them.

This section explores the relationship between odds ratios and risk ratios, demonstrating how they can be numerically identical in certain conditions, such as when the disease is not very common. The speaker uses examples with different disease prevalences to illustrate how the odds ratio remains constant even when the number of controls increases, whereas the risk ratio can change. The importance of selecting a representative control group to ensure the validity of the odds ratio is emphasized.

The paragraph discusses the determinants of sample size in case control studies, including the risk factor or intervention being studied, its distribution in the population, the desired power and statistical significance of the study, and the case to control ratio. It also touches on the importance of considering confounding factors and adjustments that may be necessary for factors like cluster sampling or response rates.

The speaker explains how the odds ratio influences the required sample size, showing that a smaller odds ratio necessitates a larger sample size to detect a significant effect. The paragraph provides a graphical representation from a WHO reference to illustrate this point, emphasizing that for odds ratios less than 2.5, the sample size requirement increases dramatically. The examples highlight the importance of considering the odds ratio when planning a study to ensure adequate power.

This paragraph provides a detailed explanation of how to calculate sample size for binary exposure in case control studies. The speaker presents the underlying formula used for calculation and explains the significance of each component in the formula, such as the ratio of controls to cases, the average proportion, and the effect size. The paragraph also discusses the importance of understanding the odds ratio and how it can be calculated from proportions.

The speaker provides practical examples of sample size calculation for case control studies, using real-world scenarios such as diabetic retinopathy and hepatitis C infection. The examples demonstrate how to apply the formulas and concepts discussed earlier in the session, showing the inputs required for sample size calculation and the resulting sample size needed for each study.

This paragraph addresses a common issue in case control studies where the number of cases is limited. The speaker explains how to adjust the sample size by increasing the number of controls to maintain study power. The explanation includes a formula for calculating the adjustment factor based on the number of controls per case, highlighting that there is a maximum benefit to increasing controls beyond a certain point.

While the session primarily focuses on binary outcomes, this paragraph briefly touches on situations where exposure is continuous, such as measuring carbon monoxide levels in blood. The speaker explains that in such cases, sample size should be calculated based on the mean to avoid losing the dose-response relationship and overestimating the sample size.

The session concludes with a summary of key points and a reminder of the importance of understanding case control study design and sample size calculation. The speaker also addresses questions from participants, providing clarifications on topics such as control group purposes, defining controls, and the use of matched versus unmatched studies. The session ends with a poll and a reminder of the next session in March.