[R] How to calculate relative risk from GAM model in mgcv package?

Eren Yalın y@||neren42 @end|ng |rom gm@||@com
Wed Dec 6 14:52:03 CET 2023

Hi R users,I am a beginner in the use of R. I need urgent help for my
thesis study.

I have daily air pollution parameters PM10, PM2.5 CO, NO2, SO2, and O3. I
also have daily hospital admission numbers. Taking into account the effect
of weekends and holidays, I would like to used generalised additive model
(GAM) to explore the relationship between daily patients admissions, and
air pollution parameters. I would like tu use mgcv package. How to get
overall relative risk and 95%CI for every pollutant?

I don't know if it's correct but here are the codes I used:




data$date <- as.Date(data$date, format="%d.%m.%Y")

data$weekend <- factor(data$weekend)

data$holiday <- factor(data$holiday)

model <- gam(adm ~ s(PM10, k = 5) + s(PM2.5, k = 5) + s(CO, k = 5) + s(NO2,
k = 5) + s(SO2, k = 5) + s(O3, k = 5) + weekend + holiday, family =
quasipoisson(link = "log"), data = data, method = "REML")

pred <- predict.gam (model, type = "response")

relative_risk <- exp(pred$fit)

However, when I look at the results, it calculates RR for 365 days

How can I get a result like the table 4 in this article (
https://pubmed.ncbi.nlm.nih.gov/36161569/)? There is only one RR
calculation for each pollutant.

I would be very grateful if you could help me. Thank you.

Eren YALIN M.D., Research Assistant

 . University, Medical Faculty,  Department of Public Health

	[[alternative HTML version deleted]]

More information about the R-help mailing list