---
title: "R Notebook"
output:
  html_document:
    df_print: paged
---
#data(1) is rawest version
#data(2) has some consolidation (ie. between categories like cMITT and MITT)
#data(3) has consolidation and removal of studies in infections which QIDP drugs havent been approved for
```{r}
setwd("~/Desktop/SPROJ")
raw_data<-read.csv("data(4).csv", stringsAsFactors = TRUE)

#making cure/success rates integers instead of characters
raw_data$X.8<-as.integer(raw_data$X.8)

library(dplyr)
library(ggplot2)
library(ggpubr)

#checking number of QIDP and non-QIDP studies
table(raw_data$X.14)





```

```{r}
#filtering raw data by endpoint
raw_data.clinicalcure<-filter(raw_data, X.5=="clinical cure")
raw_data.clinicalsuccess<-filter(raw_data, X.5=="clinical success")
raw_data.clinicalcuresuccess<-filter(raw_data, X.5=="clinical success"|X.5=="clinical cure")
raw_data.microberadication<-filter(raw_data, X.5=="microbial eradication")
raw_data.allcause<-filter(raw_data, X.5=="all cause mortality")

#looking at filtered data
table(raw_data.clinicalcuresuccess$X.6)
table(raw_data$X.14)
table(raw_data.microberadication$X.14)
table(raw_data.allcause$X.14)


```

```{r}
#filtering data by subpopulation studied
raw_data.MITT<-filter(raw_data,X.7=="MITT")
raw_data.ITT<-filter(raw_data,X.7=="ITT")
raw_data.CE<-filter(raw_data, X.7=="CE")
raw_data.ME<-filter(raw_data, X.7=="ME")
```

```{r}
#general ANOVA comparing QIDP and non-QIDP
biganova<-aov(X.8~X.14,raw_data.clinicalcuresuccess)
summary(biganova)
#not enough data
biganova2<-aov(X.8~X.14,raw_data.microberadication)
#not enough data
biganova3<-aov(X.8~X.14,raw_data.allcause)

#data cleaning/checking results of cleaning
typeof(raw_data$X.14)
raw_data$X.14<-as.character(raw_data$X.14)
raw_data$X.14[raw_data$X.14 == ""]<-NA
table(raw_data$X.5)
table(raw_data$X.14)
table(cIAI_cure.success$X.7)

summary(biganova3)

```


```{r}
#cIAI cure/success anova
cIAI_cure.success<-filter(raw_data.clinicalcuresuccess,X.6=="cIAI")
summary(aov(X.8~X.14,cIAI_cure.success))

```

```{r}
#cUTI cure/success anova
cUTI_cure.success<-filter(raw_data.clinicalcuresuccess,X.6=="cUTI")
summary(aov(X.8~X.14,cUTI_cure.success))

```

```{r}
#cSSSI cure/success anova
cSSSI_cure.success<-filter(raw_data.clinicalcuresuccess,X.6=="cSSSI")
table(raw_data.clinicalcuresuccess$X.6)

#dont have enough data (no QIDP antibiotics)
sum(raw_data.clinicalcuresuccess$X.6=="cSSSI"&raw_data.clinicalcuresuccess$X.14=="No")
summary(aov(X.8~X.14,cSSSI_cure.success))


```

```{r}
#ABSSSI cure/success anova
ABSSSI_cure.success<-filter(raw_data.clinicalcuresuccess,X.6=="ABSSSI")

summary(aov(X.8~X.14,ABSSSI_cure.success))

```

```{r}
#pneumonia cure/success anova
HABP.VABP.CABP_cure.success<-filter(raw_data.clinicalcuresuccess,X.6=="HABP"|X.6=="CABP"|X.6=="VABP")

table(HABP.VABP.CABP_cure.success$X.7)

summary(aov(X.8~X.14,HABP.VABP.CABP_cure.success))

```

```{r}
#anova's about subpopulation QIDP 
MITTanov<-aov(X.8~X.14,raw_data.MITT)
summary(MITTanov)

summary(aov(X.8~X.14, raw_data.ITT))

CEanov<-aov(X.8~X.14, raw_data.CE)
summary(CEanov)

MEanov<-aov(X.8~X.14, raw_data.ME)
summary(MEanov)
```

```{r}
#comparing QIDP and non-QIDP while controlling for disease and subpopulation

cUTI_cure.success.<-filter(cUTI_cure.success, X.7=="CE")
#summary(aov(X.8~X.14, cUTI_cure.success.MITT))
#cUTI_cure.success.MITT

cSSSI_cure.success.MITT<-filter(cSSSI_cure.success, X.7=="ITT"|X.7=="MITT")
#summary(aov(X.8~X.14,cSSSI_cure.success.MITT))
table(cSSSI_cure.success.MITT$X.14)

ABSSSI_cure.success.MITT<-filter(ABSSSI_cure.success, X.7=="ITT")
summary(aov(X.8~X.14, ABSSSI_cure.success.MITT))

HABP.MITT<-filter(HABP.VABP.CABP_cure.success, X.7=="MITT")
HABP.ITT<-filter(HABP.VABP.CABP_cure.success,X.7=="ITT")

HABP.MITT.ITT<-filter(HABP.VABP.CABP_cure.success, X.7=="MITT"|X.7=="ITT")
summary(aov(X.8~X.14, HABP.MITT.ITT))


HABP.CE<-filter(HABP.VABP.CABP_cure.success,X.7=="CE")
summary(aov(X.8~X.14,HABP.ITT))
table(HABP.VABP.CABP_cure.success$X.7)

table(HABP.CE$X.14)

cIAI.MITT.ITT<-filter(cIAI_cure.success, X.7=="MITT"|X.7=="ITT")
cIAI.MITT<-filter(cIAI_cure.success, X.7=="MITT")
cIAI.CE<-filter(cIAI_cure.success, X.7=="CE")
cIAI.ITT<-filter(cIAI_cure.success, X.7=="ITT")
summary(aov(X.8~X.14,cIAI.CE))
table(cIAI.CE$X.14)


```

```{r}


#Plot of cure rate by condition
conditionplot<-filter(raw_data.clinicalcuresuccess, X.6!="CDI", X.6!="CAP", X.6!="HABP/VABP")
ggplot(conditionplot, aes(x=X.6, y=X.8, fill=X.6))+geom_boxplot()+ylab("Positive Clinical Response (% of Patients)")+xlab(element_blank())+theme_classic()+theme(legend.position="none")+scale_fill_manual(values=c("#000000", "#E69F00", "#56B4E9", "#009E73", "#F0E442", "#0072B2", "#D55E00", "#CC79A7"))




#Plot of Cure Rate by condition by QIDP
ggplot(conditionplot, aes(x=X.6, y=X.8, fill=X.14))+geom_boxplot(outlier.shape=NA)+facet_grid(~X.6)+theme(legend.position = "none")+ylim(70,100)

conditionplot<-filter(raw_data.clinicalcuresuccess, X.6=="ABSSSI"|X.6=="CABP"|X.6=="cIAI"|X.6=="cUTI")

#creating panels for condition plot
ABSSSI<-ggplot(ABSSSI_cure.success,aes(x=X.14, y=X.8, fill=X.14))+geom_boxplot(outlier.shape=NA)+theme_classic()+theme(legend.position ="none")+ylim(70,100)+ylab("Succesful Clinical Response (% of Patients)") +scale_fill_manual(values=c("orange","blue"))+xlab(element_blank())
ABSSSI

#FOR GRAPH I COMBINED HABP, VABP, CABP INTO ONE GROUP
HABP.VABP.CABP_cure.success$X.6[HABP.VABP.CABP_cure.success$X.6 == "VABP"]<-"CABP"

raw_data$X.6[raw_da$X.14 == ""]<-NA

HABP.CABP.VABP<-ggplot(HABP.VABP.CABP_cure.success,aes(x=X.14, y=X.8, fill=X.14))+geom_boxplot(outlier.shape=NA)+theme_classic()+theme(legend.position = "none", axis.ticks.y=element_blank(), axis.text.y=element_blank(), axis.line.y=element_blank())+ylim(70,100)+xlab(element_blank())+ylab(element_blank())+scale_fill_manual(values=c("orange","blue"))
HABP.CABP.VABP

cIAI<-ggplot(cIAI_cure.success,aes(x=X.14, y=X.8, fill=X.14))+geom_boxplot(outlier.shape=NA)+theme_classic()+theme(legend.position = "none", axis.ticks.y = element_blank(), axis.text.y = element_blank(), axis.line.y=element_blank())+ylim(70,100)+xlab(element_blank())+ylab(element_blank())+scale_fill_manual(values=c("orange","blue"))

cIAI

cUTI<-ggplot(cUTI_cure.success, aes(x=X.14, y=X.8, fill=X.14))+geom_boxplot(outlier.shape=NA)+theme_classic()+theme(legend.position = "none", axis.text.y=element_blank(), axis.line.y=element_blank(), axis.ticks.y=element_blank())+ylim(70,100)+xlab(element_blank())+ylab(element_blank())+scale_fill_manual(values=c("orange","blue"))
cUTI

condition.qidp<-ggarrange(ABSSSI, HABP.CABP.VABP, cIAI, cUTI, ncol=4, labels=c("ABSSSI","BP*","cIAI*","cUTI"),legend = "none",common.legend=TRUE,hjust = c(-1,-2,-2,-2))
condition.qidp



ggplot(cIAI_cure.success.MITT, aes(x=X.14, y=X.8, fill=c(X.14)))+geom_boxplot()+theme(legend.position = "None")+xlab(element_blank())+ylab("Clinical Cure/Success Rate")+ggtitle("Clinical Response Rates of QIDP vs Non-QIDP Antibiotics in cIAI infections")

#plot of overall QIDP vs non-QIDP 
ggplot(raw_data.clinicalcuresuccess, aes(x=X.14, y=X.8, fill=X.14))+geom_boxplot()+theme(legend.position="none")+xlab(element_blank())+ylab("Clinical Cure/Success Rate")+ggtitle("Clinical Response in QIDP vs. Non-QIDP Antibiotics")
raw_data.clinicalcuresuccess$X.14<-as.character(raw_data.clinicalcuresuccess$X.14)
raw_data.clinicalcuresuccess$X.14[is.na(raw_data.clinicalcuresuccess$X.14)]<-"QIDP"
raw_data.clinicalcuresuccess$X.14[raw_data.clinicalcuresuccess$X.14=="No"]<-"non-QIDP"

#ggplot(HABP.VABP.CABP_cure.success)

#raw_data.clinicalcuresuccess
#condition.qidp
#l<-get_legend(condition.qidp)
#l<-l+scale_fill_discrete(c("non-QIDP", "QIDP"))
#boxplot.stats(raw_data.clinicalcuresuccess$X.8)$X.5

#making panels for subpopulation graph
MITTplot<-ggplot(raw_data.MITT, aes(x=X.14, y=X.8, fill=X.14))+geom_boxplot(outlier.size=0.5)+xlab(element_blank())+ylab("Clinical Cure/Success Rate")+ylim(60,100)+theme_classic()+theme(legend.position = "none", axis.text.y=element_blank(), axis.line.y=element_blank(), axis.ticks.y=element_blank())+ylim(70,100)+xlab(element_blank())+ylab(element_blank())+scale_fill_manual(values=c("orange","blue"))

MITTplot

ITTplot<-ggplot(raw_data.ITT, aes(x=X.14, y=X.8, fill=X.14))+geom_boxplot(outlier.size=0.5)+xlab(element_blank())+ylab("Clinical Cure/Success Rate")+ylim(60,100)+theme_classic()+theme(legend.position = "none")+ylim(70,100)+xlab(element_blank())+ylab("Positive Clinical Response (% of Patients)")+scale_fill_manual(values=c("orange","blue"))
ITTplot

MEplot<-ggplot(raw_data.ME, aes(x=X.14, y=X.8, fill=X.14))+geom_boxplot(outlier.size=0.5)+xlab(element_blank())+ylab(element_blank())+ylim(60,100)+theme_classic()+theme(legend.position = "none", axis.text.y=element_blank(), axis.line.y=element_blank(), axis.ticks.y=element_blank())+ylim(70,100)+xlab(element_blank())+ylab(element_blank())+scale_fill_manual(values=c("orange","blue"))

CEplot<-ggplot(raw_data.CE, aes(x=X.14, y=X.8, fill=X.14))+geom_boxplot(outlier.size=0.5)+xlab(element_blank())+ylab(element_blank())+ylim(60,100)+theme_classic()+theme(legend.position = "none", axis.text.y=element_blank(), axis.line.y=element_blank(), axis.ticks.y=element_blank())+ylim(70,100)+xlab(element_blank())+ylab(element_blank())+scale_fill_manual(values=c("orange","blue"))

ggarrange(ITTplot, MITTplot, CEplot,MEplot, ncol = 4,legend="none",labels = c("ITT","MITT", "CE","ME*"),vjust = 1,heights=3, hjust = c(-2.4,-1,-2,-2))

raw_data.MITT

#making panels for pneumonia-subpopulation plot
BP1<-ggplot(HABP.MITT.ITT, aes(x=X.14, y=X.8,fill=X.14))+geom_boxplot(outlier.size=0.5)+xlab(element_blank())+ylab("Positive Clinical Response (% of Patients)")+ylim(40,90)+scale_fill_manual(values=c("orange","blue"))+theme_classic()

BP2<-ggplot(HABP.ITT,aes(x=X.14, y=X.8,fill=X.14))+geom_boxplot(outlier.size=0.5)+xlab(element_blank())+ylab(element_blank())+ylim(40,90)+scale_fill_manual(values=c("orange","blue"))+theme_classic()+theme(axis.line.y=element_blank(), axis.ticks.y=element_blank(), axis.text.y=element_blank())

BP3<-ggplot(HABP.CE,aes(x=X.14, y=X.8, fill=X.14))+geom_boxplot()+xlab(element_blank())+ylab(element_blank())+ylim(40,90)+scale_fill_manual(values=c("orange","blue"))+theme_classic()+theme(axis.line.y=element_blank(), axis.ticks.y=element_blank(), axis.text.y=element_blank())

ggarrange(BP1,BP2,BP3,ncol=3,legend="none",labels=c("MITT/ITT*","ITT*","CE"),heights=0.5,hjust=c(-1,-3.5,-3.5),vjust=1)

#making panels for cIAI subpopulation plot
cIAI1<-ggplot(cIAI.ITT, aes(x=X.14,y=X.8, fill=X.14))+geom_boxplot()+xlab(element_blank())+ylab(element_blank())+ylim(70,95)+scale_fill_manual(values=c("orange","blue"))+theme_classic()+theme(axis.line.y=element_blank(), axis.ticks.y=element_blank(), axis.text.y=element_blank())

cIAI2<-ggplot(cIAI.MITT, aes(x=X.14, y=X.8, fill=X.14))+geom_boxplot(outlier.size=0.5)+xlab(element_blank())+ylab(element_blank())+ylim(70,95)+scale_fill_manual(values=c("orange","blue"))+theme_classic()+theme(axis.line.y=element_blank(), axis.ticks.y=element_blank(), axis.text.y=element_blank())

cIAI3<-ggplot(cIAI.CE,aes(x=X.14,y=X.8,fill=X.14))+geom_boxplot()+xlab(element_blank())+ylab(element_blank())+ylim(70,95)+scale_fill_manual(values=c("orange","blue"))+theme_classic()+theme(axis.line.y=element_blank(), axis.ticks.y=element_blank(), axis.text.y = element_blank())

cIAI4<-ggplot(cIAI.MITT.ITT,aes(x=X.14,y=X.8,fill=X.14))+geom_boxplot()+xlab(element_blank())+ylab(element_blank())+ylim(70,95)+scale_fill_manual(values=c("orange","blue"))+theme_classic()+ylab("Clinical Success Rate (% of patients)")

ggarrange(cIAI4, cIAI1, cIAI2, cIAI3, ncol=4, legend="none", labels=c("MITT/ITT*","ITT","MITT", "CE"),heights=0.5,vjust=1, hjust=-1)

#general QIDP vs non-QIDP plot
ggplot(raw_data.clinicalcuresuccess, aes(x=X.14, y=X.8,fill=X.14))+geom_boxplot()+scale_fill_manual(values=c("orange","blue"))+theme_classic()+theme(legend.position = "none")+ylab("Positive Clinical Response (% of Patients)")+xlab(element_blank())


```