Dataset Information

Too many covariates and too few cases? - a comparative study.

ABSTRACT: Prior research indicates that 10-15 cases or controls, whichever fewer, are required per parameter to reliably estimate regression coefficients in multivariable logistic regression models. This condition may be difficult to meet even in a well-designed study when the number of potential confounders is large, the outcome is rare, and/or interactions are of interest. Various propensity score approaches have been implemented when the exposure is binary. Recent work on shrinkage approaches like lasso were motivated by the critical need to develop methods for the p >> n situation, where p is the number of parameters and n is the sample size. Those methods, however, have been less frequently used when p?n, and in this situation, there is no guidance on choosing among regular logistic regression models, propensity score methods, and shrinkage approaches. To fill this gap, we conducted extensive simulations mimicking our motivating clinical data, estimating vaccine effectiveness for preventing influenza hospitalizations in the 2011-2012 influenza season. Ridge regression and penalized logistic regression models that penalize all but the coefficient of the exposure may be considered in these types of studies. Copyright © 2016 John Wiley & Sons, Ltd.

SUBMITTER: Chen Q

PROVIDER: S-EPMC5050102 | biostudies-literature | 2016 Nov

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

Too many covariates and too few cases? - a comparative study.

Chen Qingxia Q Nian Hui H Zhu Yuwei Y Talbot H Keipp HK Griffin Marie R MR Harrell Frank E FE

Statistics in medicine 20160630 25

Prior research indicates that 10-15 cases or controls, whichever fewer, are required per parameter to reliably estimate regression coefficients in multivariable logistic regression models. This condition may be difficult to meet even in a well-designed study when the number of potential confounders is large, the outcome is rare, and/or interactions are of interest. Various propensity score approaches have been implemented when the exposure is binary. Recent work on shrinkage approaches like lass ...[more]

PMID: 27357163

Similar Datasets

Project description:Most reintroduction and restocking programs consist of releasing captive-raised juveniles. The usefulness of these programs has been questioned, and therefore, quality control is advisable. However, evaluating restocking effectiveness is challenging because mortality estimation is required. Most methods for estimating mortality are based on tag recovery. In the case of fish, juveniles are tagged before release, and fishermen typically recover tags when fish are captured. The statistical models currently available for analyzing these data assume either constant mortality rates, fixed tag non-reporting rates, or both. Here, instead, we proposed a method that considers the mortality rate variability as a function of age/size of the released juveniles. Furthermore, the proposed method can disentangle natural from fishing mortality, analyzing the temporal distribution of the captures reported by fishermen from multiple release events. This method is demonstrated with a restocking program of a top-predator marine fish, the meagre (Argyrosomus regius), in the Balearic Islands. The estimated natural mortality just after release was very high for young fish (m0 = 0.126 day-1 for fish 180 days old), but it was close to zero for large/old fish. These large/old fish were more resilient to wild conditions, although a long time was needed to achieve a relevant reduction in natural mortality. Conversely, these large/old fish were more vulnerable to fishing, creating a trade-off in survival. The release age that maximizes the number of survivors after, for example, one year at liberty was estimated to be 1,173 days. However, the production cost of relatively old fish is high, and only a few fish can be produced and released within a realistic budget. Therefore, in the case of the meagre, increasing the number of released fish will have no or scarce effects on restocking success. Conversely, it is advisable implement measures to reduce the high natural mortality of young juveniles and/or the length of time needed to improve fish resilience.

Dataset Information

Too many covariates and too few cases? - a comparative study.

Publications

Too many covariates and too few cases? - a comparative study.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets