Hello,
Our study aimed to estimate the stress prevalence among health professionals who works on pre-hospital services.
The dependent variable was the presence or absence of stress (dichotomic variable: Y/N).
The independent variables were the social and professional characteristics.
We performed a logist regression, using the comand "logistic" on Stata in order to obtain the Odds ratio for all variables.
Currently, we are trying to calculate the prevalence ratio.
Could you please help us?
Thank you.
William
Use the binomial model with a log-link function. The exponent of the beta coefficients will be the prevalence ratios.
you can use cs command to reveal prevalence ratio. "cs dep.var indep.var". The formula will be similar to risk ratio, but different in term of type of study and strength association
Hello Fajar Awalia and Md. Belal Hossain ,
Thank you so much!
I used the Poisson Logistic regression with robust variance that, according to these authors, are better alternatives than logistic regression.
COUTINHO, Leticia M S; SCAZUFCA, Marcia; MENEZES, Paulo R. Methods for estimating prevalence ratios in cross-sectional studies. Revista de Saúde Pública, v. 42, n. 6, p. 992–998, 2008. Disponível em: < https://www.scielo.br/pdf/rsp/v42n6/en_7118.pdf >.
The binomial model with a log-link function suffers from convergence issues with many covariates. So you can use the Poisson regression with log-link function and robust variance. These two are equivalent.
- Badges
- Science topic
- Similar topics
- Biological Science
- Physiology
- Physiological Processes
- Stress