China 
Africa 
Explore chapters and articles related to this topic
The use of Industrial Hygiene Data in Occupational Epidemiology
Published in Frederick C. Kopfler, Gunther F. Craun, Environmental Epidemiology, 2019
Robert F. Herrick, Larry J. Elliott
In addition to the standardized mortality ratio analysis just described, a matched case control analysis was also performed. Conditional logistic regression was used to compare exposure histories of workers known to have died of leukemia with controls, who were workers known to have died of other causes. Using the exposure estimates previously developed, the cases and controls were compared by their cumulative (lifetime) benzene exposure, duration of exposure, and exposure rate, which was calculated by dividing cumulative exposure by duration of exposure. Logistic regression models of the form OR=exp(B1X1+B2X2+…….+BnXn)
Riding behavior and electric bike traffic crashes: A Chinese case-control study
Published in Traffic Injury Prevention, 2020
Yining Qian, Qiannan Sun, Gaoqiang Fei, Xinyu Li, Lorann Stallones, Henry Xiang, Xujun Zhang
Data were entered into EpiData 3.0 and the statistical analysis was performed with the statistical package SPSS 11.0 (SPSS Inc., Chicago, IL). First, demographics of study participants and characteristics of e-bikes were described, including age, gender, school education level, marital status, prior crash history, type of e-bikes, and equipped with bicycle pedals. Univariate logistic models were used to identify whether there were significant differences between cases and controls in e-bike-related traffic risk factors. The dependent variable was an e-bike-related traffic crash (yes/no). The independent variables were 14 potential risk riding behaviors including: running a red light; riding after drinking; speeding; making or receiving a call while riding; listening to loud music with headphones while riding; carrying children while riding; carrying adults while riding; no helmet use; turning without signaling; riding in the motor vehicle lane; riding in the wrong direction; not slowing down in bad weather; having a history of crashes, and riding scooter-style e-bikes. Finally, variables that were statistically significant (P value < .10) in univariate analysis were included in the multivariate conditional logistic regression analysis. Conditional logistic regression model was performed based on the 1:2 paired case-control study to control confounding. P < .05 was considered as statistically significant.
A matched case-control method to model car-following safety
Published in Transportmetrica A: Transport Science, 2023
Qianwen Li, Handong Yao, Xiaopeng Li
The matched case–control method has been well established in epidemiology to investigate risk factors to disease since decades ago (Breslow 1996; Breslow et al. 1978; Ingram et al. 1997; Wey et al. 1989). It has also been used in transportation to study the contributing factors to traffic crashes (Gjerde et al. 2011; Gross 2013; Híjar et al. 2000). The following three steps are required when conducting a matched case–control study (Schlesselman 1982). Defining cases and controls. Cases are unsafe car-following segments that experience crash risks and controls are safe car-following segments that do not experience crash risks.Randomly matching cases and controls. Each case is randomly matched to several controls to form a stratum based on confounding variables (i.e. unobserved driver characteristics) that are related to both risk factors of interest (i.e. observed driving characteristics) and the outcome (i.e. unsafe car-following). After matching, the impacts of confounding variables on the outcome are eliminated and thus unbiased associations between risk factors of interest and the outcome can be estimated. As the matching ratio (i.e. control to case) increases, the analysis power of the matched case–control study increases. Yet, the number of resulting strata decreases. Thus, the matching ratio shall be carefully chosen by balancing the two aspects.Conducting analysis. Conditional logistic regression was developed to fit matched data in case–control studies. It is an extension of logistic regression (Breslow et al. 1978).
A matched case–control approach to identify the risk factors of fatal pedestrian crashes on a six-lane rural highway in India
Published in International Journal of Injury Control and Safety Promotion, 2023
Laxman Singh Bisht, Geetam Tiwari
The conditional logistic regression model is an extension of logistic regression (Breslow et al., 1978; Li et al., 2021). This study evaluated the effect of the considered risk factors on fatal pedestrian crashes. The model is conditional cases matched to controls on the risk factors. This model can also account for multi-level risk factors. In this study, the confounding has been addressed during the matching stage of the study design. The conditional probability of a fatal pedestrian crash () to be observation in stratum calculated as (Li et al., 2021): where is the stratum-specific interpretation term reflecting the different combination effects of confounding variables for different strata ); is a vector of k explanatory variables included in the model, and is the estimated coefficient for . The conditional likelihood of each stratum is calculated with the help of the following expression: where is the case segment of the stratum; for are the matched control segments of the stratum; is the vector of risk factors associated with For strata, the conditional log-likelihood function was calculated using the following equation (Schlesselman, 1982):