The Epidemiology of Skin Cancer
Henry W. Lim, Herbert Hönigsmann, John L. M. Hawk in Photodermatology, 2007
A more refined analysis of the effect of candidate etiologic factors on the disease occurrence is offered by analytical epidemiology methods, that is, cohort and case-control studies. A cohort study involves following-up over time subjects with different levels of exposure to a candidate etiologic factor comparing the incidence of diseases of interest in these subjects. A case-control study involves comparing previous exposure to etiologic factors in a group of people diagnosed with a disease of interest (cases) and in a group of people, otherwise comparable, without the disease (controls). The measure adopted to express the link between the exposure and the disease is the “relative risk.” This is the ratio of the incidence of a disease among the exposed to the risk among the unexposed. Odds ratios, that is, the ratio of the odds in favor of getting disease, if exposed to the odds in favor of getting the disease, if otherwise, can be calculated from case-control studies as an estimate of the relative risk. The approximation works well for rare disorders. Multivariate models can be used to simultaneously control the effects of variables other than the one of interest, when calculating relative risks or odds ratios.
Cancer Epidemiology
Peter G. Shields in Cancer Risk Assessment, 2005
The advantages of the case–control study design include the ability to study multiple etiological factors, the ability to study exposures over a broad period of time (to better identify latency periods), and the efficiency in studying rare diseases. The two most common concerns in the conduct of a case–control study are selection bias and recall bias. Selection bias occurs when the case or control group is not representative of cases or controls in the underlying study base. Recall bias occurs when cases or controls differentially recall past exposures, leading to biased associations. A recent example of concern about differential recall bias is in the evaluation of induced abortion and breast cancer risk. There is evidence that while cases fairly accurately report induced abortions, healthy controls are likely to systematically underreport this procedure leading to a bias that suggests that there is an association (35). While case-control studies are susceptible to these types of biases, there are clear approaches to minimize their effect (36–38).
Introduction to the portfolio station
Sukhpreet Singh Dubb in Core Surgical Training Interviews, 2020
In contrast to cohort studies, a case-control study takes patients already suffering from a given state of interest, such as disease. These patients are then matched to similar candidates without this condition and then investigated to identify a potential exposure that may be contributory. This may involve ascertaining the past medical and social history, or even having patients recount previous exposures. Case control studies are most useful for investigating rare conditions. As with any study, case control series are subject to bias and weaknesses. Poor definition of what the disease or condition is, with subsequent misallocation of participants, can cause disastrous impacts on the outcomes of the study. Similarly, case control studies cannot establish any causal relationships – only risk associations.
A real-world disproportionality analysis of FDA Adverse Event Reporting System (FAERS) events for baricitinib
Published in Expert Opinion on Drug Safety, 2020
Ling Peng, Kui Xiao, Silvia Ottaviani, Justin Stebbing, Ying-Jie Wang
Our study is a case/non-case one, which can be viewed as a case–control study. We analyzed AEs caused by investigating drugs, but not by disease state. A disproportionality analysis was performed using the ROR to detect spontaneous signals, which were calculated using the case/non-case method, indicating whether there is a signal for a potentially increased risk of drug-related AE [6]. ‘Cases’ were defined as patients treated with a drug who reported a specific AE, while all other possible pairs were ‘non-cases.’ The calculation for ROR using two-by-two contingency tables of reported event counts for specific drug and other drugs. ROR represents the odds that an outcome will occur given a particular exposure and is a measure of the degree of association between an exposure to a drug and the odds of a specific outcome [7].
Association between Psoriasis Disease and IFN-λ Gene Polymorphisms
Published in Immunological Investigations, 2022
Büşra Yilmaz, Güneş Çakmak Genç, Sevim Karakaş Çelik, Nilgün Solak Tekin, Murat Can, Ahmet Dursun
Statistical analysis was performed using SPSS (version. 19.0; SPSS Inc., Chicago, IL, USA). A post-hoc power calculation was performed using the G-Power software to determine the sample size. A case-control study was conducted. The allelic and genotypic frequencies of the polymorphisms were calculated for both for cases and for controls. Analyses were performed using dominant, additive, and recessive models. Dominance was defined in terms of allele 2 (minor allele) effects. In the dominant allele 2 models, homozygous individuals for allele 1 were compared with carriers of allele 2. In the recessive allele 2 models, homozygous individuals for allele 2 were compared with carriers of allele 1. The χ2 test was used to compare the genotype and allele frequencies of each gene polymorphism in psoriasis patients and controls. The odds ratio (OR) and 95% confidence interval (CI) were calculated to compare psoriasis risk for the alleles and genotypes. The Hardy–Weinberg equilibrium (HWE) test was conducted using Excel (Microsoft Office Excel, Microsoft Corp., Redmond, WA). Data distribution was determined with the Shapiro–Wilk test. Continuous variables were expressed as mean ± standard deviation or median (minimum-maximum). The categorical variables were frequency and percent. Continuous variables were compared using the independent-sample t-test or the Mann–Whitney U-test. P values < .05 were considered statistically significant.
Regression analysis of case-cohort studies in the presence of dependent interval censoring
Published in Journal of Applied Statistics, 2021
Mingyue Du, Qingning Zhou, Shishun Zhao, Jianguo Sun
As discussed in Sections 4 and 5, a type of studies that is similar to case-cohort studies is the case–control study and the key difference between the two is the generation of the subcohort. With the case-cohort design, the subcohort is sampled from all study subjects, while the case–control design samples the subcohort only from the subjects who do not experience the failure event of interest during the follow-up. It is apparent that the data structures under the two designs are different but on the other hand, the simulation study suggested that the proposed estimation approach seems to be valid too for the case–control design. A possible explanation for this is that the resulting data may carry similar information about the model and the regression parameters of interest given the low percentage of the event rate.
Related Knowledge Centers
- Cohort Study
- Observational Study
- Power of A Test
- Randomized Controlled Trial
- Rate Ratio
- Odds Ratio
- Relative Risk
- Power of A Test
- Confounding
- Hierarchy of Evidence
- Rare Disease Assumption