CODEWITHSUNDEEP
rmanualjournalr-forestplot
Juan Pablo
Juan Pablo

Reputation: 11

Forest plot with data introduced manually in R

I have just recently received some data from my university. I have been ordered to create a forest plot with Odds ratios but the main problem is that I do not have access to the dataset except for the final values. I only have an excel with the ORs, the 95% Confidence interval, the population, the p-values and the variables.

I would like to create a forest plot similar to this one on the NEJM but I can't figure out how I can do this. If anyone could help me I would really appreciate it. I have seen there are similar questions in the platform but none of them are oriented to information introduced manually without having to make calculations or regressions.

enter image description here

I leave you a link to the image I have in mind.

Thank you for your help.

Upvotes: 1

Views: 1945

Answers (1)

user13069688
user13069688

Reputation: 353

I was in a similar situation as you. I had only the OR and CI. I have created the following forest plot inputing my data into vectors. You should check the package called forestplot and the documentation provided along with it.

test_data <- data.frame(coef1 =c(1.85, 1.86, 1.88, 1.45, 1.64, 1.54, 1.45, 1.64, 1.54, 1.45, 1.64),
                    coef2 =c(1.54, 1.45, 1.64, 1.54, 1.85, 1.86, 1.88, 5.85, 5.86, 5.88, 1.45),
                    coef3 =c(1.64, 1.54, 1.45, 1.64, 1.54, 1.45, 1.64, 1.54, 1.45, 1.64, 1.54),
                    low1  =c(1.45, 1.25, 1.30, 1.10, 1.20, 1.25, 1.15, 1.30, 1.15, 1.15, 1.30),
                    low2  =c(1.15, 1.15, 1.30, 1.15, 1.45, 1.25, 1.30, 3.45, 3.25, 3.30, 1.10),
                    low3  =c(1.20, 1.25, 1.10, 1.20, 1.25, 1.10, 1.20, 1.25, 1.10, 1.20, 1.25),
                    high1 =c(2.52, 2.10, 2.30, 2.54, 2.52, 2.25, 2.54, 2.52, 2.25, 2.54, 2.52),
                    high2 =c(2.25, 3.54, 3.52, 3.25, 2.52, 2.10, 2.30, 7.52, 7.10, 7.30, 2.54),
                    high3 =c(2.52, 2.25, 2.54, 2.52, 2.25, 2.54, 2.52, 2.25, 2.54, 2.52, 2.25))

tabletext<- c("Diabetes", "Heart disease", "Stroke", "COPD", "Asthma", 
"Cancer", "Arthtitis", "Osteoporosis", "Depression", "Severe mental diseases", 
"Dementia")


coef <- with(test_data, cbind(coef1, coef2, coef3))
low <- with(test_data, cbind(low1, low2, low3))
high <- with(test_data, cbind(high1, high2, high3))

library(forestplot)
forestplot(tabletext, coef, low, high,
       title="Comorbidities",
       zero= 1,
       lwd.zero = gpar(lwd=1),
       boxsize=0.18,
       line.margin = .1,
       xticks = c(0:8),
       clip=c(0, 8),
       fn.ci_norm = c(fpDrawNormalCI, fpDrawDiamondCI, fpDrawCircleCI),
       col=fpColors(box=c("darkblue", "darkred", "darkgreen"),
                    line=c("darkblue", "darkred", "darkgreen"),
                    zero = "black"),
       xlab="Odds ratio (95% CI)",
       new_page = TRUE,
       txt_gp = fpTxtGp(label = list(gpar(fontfamily = "Times New Roman", cex=1),
                                     xlab  = gpar(fontfamily = "Times New Roman", cex=1),
                                     ticks = gpar(fontfamily = "Times New Roman", cex=0.8),
                                     title = gpar(fontfamily = "Times New Roman", cex=1.2))),
       legend=c("Unadjusted", "Adjusted for Age and Sex", "Adjusted for all"),
       legend_args = fpLegend(title="Model Adjustment",
                              pos = "top",
                              r=unit(.2, "snpc"),
                              gp = gpar(col="black", fill= "white", lwd=0.7)))

Upvotes: 2

Related Questions