Speaker 1: Welcome back to Epi in a Minute. In our first video, we compared each of the study designs and talked a little about the pros and cons of each. In this series, we'll progress from the simplest of study designs to the more complex. Our first two videos look at case reports and cross-sectional studies, which can help us form a study question or hypothesis. Then, we'll move to case control studies that test the association of exposures and health outcomes. Finally, we'll talk about cohort studies and controlled trials, which take the longest and have the largest price tag, but can give a much stronger causal evidence about the relationship between exposure and health outcome. Our second video went over case reports and what they're useful for. In this, our third video, we're going to be taking a deeper look at cross-sectional studies. Just like for each of the five study designs, the five C's, we're going to ask four questions. Number one, so what is a cross-sectional study? Cross-sectional studies measure health outcomes and exposures in a population at a single point in time or over a period of time. Cross-sectional studies are like a snapshot and give you the prevalence of a health outcome at a specific point in time and place, called point prevalence. They also describe demographics of the population, for example, age, gender, education, and income level, the conditions in which the health outcome occurs, and what exposures are near the outcome at the time that the snapshot was taken. Number two, what are cross-sectional studies good for? Cross-sectional studies are great for many things. They're relatively quick and easy to do. You can study multiple diseases and multiple exposures at the same time in your snapshot study. Cross-sectional studies help you to estimate the burden of a disease in a population, and you can use cross-sectional studies to help you determine the priority of diseases to address within that population. They can be conducted at a single point in time or at several points, called a serial cross-sectional study, where you can estimate a trend over time towards a particular health outcome. Number three, what are cross-sectional studies not good for? Because cross-sectional studies take a snapshot in time, it's difficult, if not impossible, to determine if the exposure being measured happened before the health outcome to be studied. This means that temporality, assuring that the exposure happened before the health outcome, a critical requirement in assessing possible associations, cannot be established. You can't prove that smoking causes lung cancer if the cancer occurred before the patient started smoking. In addition, cross-sectional studies often use convenient samples, which select participants based on their ready availability, instead of randomly selecting participants. For example, asking people at an NFL football game, rather than taxpayers as a whole, about funding for a new football stadium. Studies that use convenient samples are more prone to erroneous results. Finally, cross-sectional studies are also not useful for rare diseases, because cross-sectional studies measure a whole population and then associate the exposure and health outcome. But what if, after measuring the population, nobody has the disease? For all these reasons, cross-sectional studies are generally considered less reliable than cohort and case control studies. They cannot establish cause-and-effect relationships. They are generally viewed as hypothesis-generating, and reported associations between exposure and health outcomes must be cautiously considered. Number four, how do we measure data in a cross-sectional study? Although there are many available techniques, the most basic analysis tool for a cross-sectional study is the odds ratio. Odds ratios, or OR, are simple to calculate, and they measure the strength of the association between the exposure and the health outcome variable. To calculate an odds ratio, you create a two-by-two table and count up the number of exposed and unexposed, diseased and non-diseased, and fill in the table. The formula for an odds ratio is A times B over C times D. An odds ratio of 1 suggests that there is no difference, meaning the exposure neither increases nor decreases the risk of the health outcome. An odds ratio greater than 1 suggests that the exposure may increase the risk. An odds ratio of less than 1 suggests the exposure may reduce the risk. The larger the odds ratio, the greater the estimate of increased risk from the exposure. Odds ratios greater than 2 are generally considered meaningful. Odds ratios greater than 4 are considered very strong. And that's the basic idea behind the snapshot study design known as cross-sectional. Next up, we're talking about case control studies.