Updated dementia data.Rmd

---
title: "Update data dementia"
output: html_document
date: "2023-12-07"
---

```{r setup, include=FALSE}
knitr::opts_chunk$set(echo = TRUE)
```

```{r merge the data}

data.update<-read_dta("Updated dementia data.dta")

write.csv(data.update,"data_update.csv",row.names = F)  # need to be csv file, otherwise it fails
data.updatenew<-read.csv("data_update.csv")  #import data again

colnames(data.updatenew)<-c("idpicc","originalid","datevisitbl_new","dementia_new","datedementia_new","enddatedementia_new")

data1.up.new<-merge(data1.up,data.updatenew, by= c("idpicc","originalid"),all.x = T)

#exclude ICICLE

data1.up.new<-data1.up.new%>%
              filter(study!="ICICLE")


```

```{r check the difference}

data.check.dementia<-data1.up.new%>%
  select("idpicc","originalid","datevisitbl","datevisitbl_new","dementia_new","cens","datedementia_new",
         "datedementia","enddatedementia_new","t")



data.check.dementia[data.check.dementia$dementia_new!=data.check.dementia$cens,]$idpicc

data.check.dementia[data.check.dementia$datedementia_new!=data.check.dementia$datedementia,]$idpicc


data.check.dementia[data.check.dementia$enddatedementia_new!=data.check.dementia$t,]$idpicc

#Only difference is in the censoring date

```

```{r change the censoring data to check}

data1.up.new$tt<-as.Date(as.character(data1.up.new$enddatedementia_new), format="%Y-%m-%d")-
                as.Date(as.character(data1.up.new$datevisitbl), format="%Y-%m-%d")

data1.up.new$tt<-as.numeric(data1.up.new$tt)

data1.up.new$years<-data1.up.new$tt/365.25  #955 rows

#Details of the reason of removing can be found in Dementia pre1.Rmd (the data only MMSE longtidinal)

data1.up.new<-data1.up.new%>%
  filter(data1.up.new$years>0) #908


#----up to 10 years----

data1.temp10.up.new<-survSplit(Surv(years, dementia_new) ~ ., data = data1.up.new, cut = 10,
                  episode="timegroup")


data1.10.up.new<-subset(data1.temp10.up.new, timegroup == 1) #only the first 10 year

```

```{r KM plot check}

KM.new<-survfit(Surv(years,dementia_new)~study,data=data1.10.up.new)


ggsurv1<-ggsurvplot(KM.new,
           ylab="Cumulative probability of dementia",
           xlab="Time to dementia",
           legend="top",
           surv.median.line = "hv",
           fun="event",
           risk.table = T,
           risk.table.y.text.col = F,
           risk.table.y.text = T,
           risk.table.col = "strata",
           linetype = "strata",
           ggtheme = theme_bw(),
           legend.labs=c("CamPalGN","NYPUM","ParkWest","PICNICS","PINE"))

ggsurv1<-ggpar(ggsurv1, 
      font.main = c(14, "bold"),
      font.x = c(14, "bold"),
      font.y = c(14, "bold"),
      font.caption = c(14, "bold"), 
      font.legend = c(14, "bold"))

print(ggsurv1, 
      surv.plot.height = 0.6,
      risk.table.height = 0.4)

png("KM1.dementia.new.png",width = 4000,height =3000,res = 400)

print(ggsurv1, 
      surv.plot.height = 0.7,
      risk.table.height = 0.3)

dev.off()
```