China
Africa
Explore chapters and articles related to this topic
Basic Research Design:
Published in Lynne M. Bianchi, Research during Medical Residency, 2022
Lynne M. Bianchi, Luke J. Rosielle, Justin Puller, Kristin Juhasz
With observational methods, the investigator observes what naturally occurs and provides a summary of findings, typically in a narrative format with quantified information included where appropriate. Analytical observational studies can also test hypotheses. Experimental studies involve investigator-controlled interventions, modifications that are manipulated and measured to determine their effects. With experimental studies, hypotheses are tested.
Introduction and Brief History of Structural Equation Modeling for Health and Medical Research
Published in Douglas D. Gunzler, Adam T. Perzynski, Adam C. Carle, Structural Equation Modeling for Health and Medicine, 2021
Douglas D. Gunzler, Adam T. Perzynski, Adam C. Carle
In observational research (and quasi-experiments), multiple strategies are available that attempt to mitigate the bias of the non-random sample in a comparison study. For example, propensity matching [22], an approach to balance treatment and control group observations across covariates, can be used for this purpose. Observational studies, due to the high possibility of confounding, often have poor internal validity. Similarly, due to possible confounding and lack of control over the study conditions, researchers conducting observational studies must confront challenges in establishing causality and poor external validity.
Bayesian Methods for Evaluating Drug Safety with Real-World Evidence
Published in Harry Yang, Binbing Yu, Real-World Evidence in Drug Development and Evaluation, 2021
Unlike randomized clinical trials, investigators do not have control over the treatment assignment in observational studies. Without randomization, the treatment and control groups may have widely different distributions of relevant covariates. Even though bias due to observed confounders may be corrected using the regression analysis or matching techniques, the analyses may still be subject to potential bias arising from unobserved confounders. Sensitivity analysis is a technique for assessing whether the inference drawn from an observational study could be altered by a moderate “imbalance” between the distribution of the covariates in the treatment and control groups. There has been extensive research on how to perform sensitivity analysis for observational data in epidemiologic studies, for example, Cornfield et al. [7], Gastwirth [8], Rosenbaum [9], and others. Most of the literature considers models for the imbalance or association among the unobserved confounder, exposure, and response, and then recomputes the test statistics and p-values for a range of plausible association between the confounder and the exposure. If a moderate degree of confounding could change the estimate of the treatment effect, then the validity of the study findings is not robust.
Gaps in evidence on treatment of male osteoporosis: a Research Agenda
Published in The Aging Male, 2023
Adam J. Rose, Susan L. Greenspan, Guneet K. Jasuja
It is therefore also important to consider how high-quality observational studies could inform clinical practice in our management of male osteoporosis. Observational studies can have important advantages – they can be conducted relatively quickly, they can have relatively low cost, they can involve large populations and thus have the potential to detect rare events, and they may in some ways be more generalizable than randomized trials – because their populations are more representative [74]. The advantages of randomized trials are also well-known, especially their ability to control for both observed and unobserved confounders, which has led to the justified emphasis on using them as the mainstay of clinical evidence to guide practice [75]. However, we think that high-quality observational studies of men with osteoporosis may be preferable to continuing to rely on evidence from clinical trials among women.
The prognostic value of the interaction between ASXL1 and TET2 gene mutations in patients with chronic myelomonocytic leukemia: a meta-analysis
Published in Hematology, 2022
Wenxia Zhao, Conghui Zhang, Yiming Li, Yang Li, Yang Liu, Xiaoyu Sun, Mingyan Liu, Rongguang Shao
Although we included the previous research as fully as possible into our analysis, the results still have their limitations. First of all, due to lack of data, we cannot evaluate other clinical parameters except OS. There are only three studies on the two genes of TET2 and ASXL1. Some other statistical methods can only play a supporting role. It causes certain limitations to our analysis. Secondly, all selected studies are observational studies, not prospective randomized controlled studies. Finally, the number of included studies is relatively small, especially for TET2 studies. When the number of studies included in the meta-analysis is less than 10, the Egger’s and Begg’s test are relatively low. So, Egger’s and Begg’s test may not be able to detect publication bias. Therefore, more epidemiological data, many in vitro experimental studies and prospective randomized controlled studies, are still needed for more subgroup analysis and further analysis to verify our conclusions.
Prehospital Airway Management: A Systematic Review
Published in Prehospital Emergency Care, 2021
Nancy Carney, Annette M. Totten, Tamara Cheney, Rebecca Jungbauer, Matthew R. Neth, Chandler Weeks, Cynthia Davis-O'Reilly, Rongwei Fu, Yun Yu, Roger Chou, Mohamud Daya
The most serious limitations of this review were related to the comparatively weak study designs used to compare airway management approaches and the risk of biases that are common challenges in prehospital and emergency care research. While the body of evidence did include randomized clinical trials, the majority of included studies were retrospective observational studies based on analyses of data from national or regional registries or administrative data from a single health system or EMS agency. Observational studies are more susceptible to bias. Indication bias, classifying patients by the treatment received, and survival bias, including only patients who survive a treatment, are variants of selection bias that are likely to occur in observational studies of prehospital care. Furthermore, data on important confounding variables are often limited in large databases, and retrospective analyses may not account for all relevant potential confounders, even with matched propensity score analyses.