Supplementary file 4: R statistical software scripts prepared for analyzing calf morbidity and mortality event data

setwd("D:/desk top collected/Umar_data")
library(survival)
library(ggplot2)
library(ggpubr)
library(GGally)
library(ggpubr)
library(foreign)
library("ggfortify")
library(survminer)
library(prevalence)
library(epiR)
mydata<- read.csv("final_data.csv", header = T,",")
class(mydata$morbidity.interval)
mydata$morbidity.interval<- as.numeric(mydata$morbidity.interval)
class(mydata$mortality.interval)
mydata$mortality.interval<- as.numeric(mydata$mortality.interval)
class(mydata$mortality.interval)

###incidence rate calculation with its 95%CI

nmor<- 15  ## number of mortality during follow up
ntmor<- 107.41 ## calf 6 months at risk of mortality
tmor<- as.matrix(cbind(nmor,ntmor))  ## combine the two vectors
epi.conf(tmor,ctype = "inc.rate",method = "exact",N=1000,design = 1,
         conf.level = 0.95)*100
ncas<-53  ## number of morbidity in the follow up period
ntar<- 97.05 ## calf 6 months at risk of morbidity
temp<- as.matrix(cbind(ncas,ntar))  ## combine the two vectors
epi.conf(temp,ctype = "inc.rate",method = "exact",N=1000,design = 1,
         conf.level = 0.95)*100
ndia<- 16   ## no of diarric calf
nmatr<- 106.87 ## calf at 6 months at risk
dia<- as.matrix(cbind(ndia,nmatr)) ## combine two vectors
epi.conf(dia,ctype = "inc.rate",method = "exact",N=1000,design = 1,
         conf.level = 0.95)*100

npnu<- 10  ## no of pnumonia calf
nmatr<- 106.99 ## calf at 6 months at risk
pnu<- as.matrix(cbind(npnu,nmatr)) ## combine two vectors
epi.conf(pnu,ctype = "inc.rate",method = "exact",N=1000,design = 1,
         conf.level = 0.95)*100

nsep<- 7  ## no of sept. calf
nmatr<- 107.24 ## calf at 6 months at risk
sep<- as.matrix(cbind(nsep,nmatr)) ## combine two vectors
epi.conf(sep,ctype = "inc.rate",method = "exact",N=1000,design = 1,
         conf.level = 0.95)*100

nomp<- 3  ## no of omphalities calf
nmatr<- 107.18 ## calf at 6 months at risk
omp<- as.matrix(cbind(nomp,nmatr)) ## combine two vectors
epi.conf(omp,ctype = "inc.rate",method = "exact",N=1000,design = 1,
         conf.level = 0.95)*100
nskd<- 7  ## no of skin disease calf
nmatr<- 107.02 ## calf at 6 months at risk
skd<- as.matrix(cbind(nskd,nmatr)) ## combine two vectors
epi.conf(skd,ctype = "inc.rate",method = "exact",N=1000,design = 1,
         conf.level = 0.95)*100

nmis<- 10  ## no of skin disease calf
nmatr<- 107 ## calf at 6 months at risk
mis<- as.matrix(cbind(nmis,nmatr)) ## combine two vectors
epi.conf(mis,ctype = "inc.rate",method = "exact",N=1000,design = 1,
         conf.level = 0.95)*100
###estimate of cummulative incidence and 95%CI of calf morbidity
source("CumIncidence.R") ### find a function from the working directory
fit_morb<-CumIncidence (mydata$morbidity.interval, mydata$Morbidity, cencode = 0, xlab="Days",
                   t=c(0,30,60,90,120,150,180), level = 0.95)

##to estimate the commulative incidence of calf mortality and its 95%CI
fit_mort<-CumIncidence (mydata$mortality.interval, mydata$Mortality, cencode = 0, xlab="Days",
                        t=c(0,30,60,90,120,150,180), level = 0.95)

cuminc(mydata$morbidity.interval, mydata$Morbidity, cencode = 0,level=0.95)

### calculation of calf Mortality hazard at different time interval 
nmis<- 6## no of cases at day 30
nmatr<- 237 ## calf at  risk
mis<- as.matrix(cbind(nmis,nmatr)) ## combine two vectors
epi.conf(mis,ctype = "inc.rate",method = "exact",N=1000,design = 1,
         conf.level = 0.95)
nmis<- 5## no of cases at day 60
nmatr<- 211 ## calf at  risk at d60
mis<- as.matrix(cbind(nmis,nmatr)) ## combine two vectors
epi.conf(mis,ctype = "inc.rate",method = "exact",N=1000,design = 1,
         conf.level = 0.95)
nmis<- 1## no of cases at day 90
nmatr<- 148 ## calf at  risk at d90
mis<- as.matrix(cbind(nmis,nmatr)) ## combine two vectors
epi.conf(mis,ctype = "inc.rate",method = "exact",N=1000,design = 1,
         conf.level = 0.95)

nmis<- 1## no of cases at day 120
nmatr<- 94 ## calf at  risk at d120
mis<- as.matrix(cbind(nmis,nmatr)) ## combine two vectors
epi.conf(mis,ctype = "inc.rate",method = "exact",N=1000,design = 1,
         conf.level = 0.95)

nmis<- 2## no of cases at day 150
nmatr<- 52 ## calf at  risk at d150
mis<- as.matrix(cbind(nmis,nmatr)) ## combine two vectors
epi.conf(mis,ctype = "inc.rate",method = "exact",N=1000,design = 1,
         conf.level = 0.95)
nmis<- 0## no of cases at day 180
nmatr<- 2 ## calf at  risk at d180
mis<- as.matrix(cbind(nmis,nmatr)) ## combine two vectors
epi.conf(mis,ctype = "inc.rate",method = "exact",N=1000,design = 1,
         conf.level = 0.95)
### calculation of calf morbidity hazard at different time interval 
nmis<- 17## no of cases at day 30
nmatr<- 237 ## calf at  risk
mis<- as.matrix(cbind(nmis,nmatr)) ## combine two vectors
epi.conf(mis,ctype = "inc.rate",method = "exact",N=1000,design = 1,
         conf.level = 0.95)
nmis<- 15## no of cases at day 60
nmatr<- 198 ## calf at  risk at d60
mis<- as.matrix(cbind(nmis,nmatr)) ## combine two vectors
epi.conf(mis,ctype = "inc.rate",method = "exact",N=1000,design = 1,
         conf.level = 0.95)
nmis<- 5## no of cases at day 90
nmatr<- 129 ## calf at  risk at d90
mis<- as.matrix(cbind(nmis,nmatr)) ## combine two vectors
epi.conf(mis,ctype = "inc.rate",method = "exact",N=1000,design = 1,
         conf.level = 0.95)

nmis<- 13## no of cases at day 120
nmatr<- 82 ## calf at  risk at d120
mis<- as.matrix(cbind(nmis,nmatr)) ## combine two vectors
epi.conf(mis,ctype = "inc.rate",method = "exact",N=1000,design = 1,
         conf.level = 0.95)

nmis<- 3## no of cases at day 150
nmatr<- 40 ## calf at  risk at d150
mis<- as.matrix(cbind(nmis,nmatr)) ## combine two vectors
epi.conf(mis,ctype = "inc.rate",method = "exact",N=1000,design = 1,
         conf.level = 0.95)
nmis<- 0## no of cases at day 180
nmatr<- 16 ## calf at  risk at d180
mis<- as.matrix(cbind(nmis,nmatr)) ## combine two vectors
epi.conf(mis,ctype = "inc.rate",method = "exact",N=1000,design = 1,
         conf.level = 0.95)
##mydata$Interval<- as.numeric(mydata$Interval)
##mydata210<- subset(mydata, mydata$Interval_conception<=210)## interval below 210 days
### median survival time estimated using the following function
mydata$Morbidity<- as.numeric(levels(mydata$Morbidity))
morb<- survfit(Surv(mydata$morbidity.interval,mydata$Morbidity)~1)
summary(morb)
### ploting the survival curve for morbidity using ggsu
###fitting survival model to plot the morbidity curve
my.KMest<- survfit(Surv(mydata$morbidity.interval,mydata$Morbidity)~1)

### ploting the survival curve for morbidity using ggsurvplot

ggsurv <- ggsurvplot(my.KMest, data = mydata,  title = "survival curve of calf morbidity",
                     palette = "jco",       # jco palette
                     xlim = c(0,180),    # present narrower X axis, but not affect
                     ylim = c(0.4,1),
                     # survival estimates.
                     xlab = "Time (days)",   # customize X axis label.
                     break.time.by = 30,     # break X axis in time intervals by 500.
                     ggtheme = theme_survminer(),
                     ylab = "% surviving,s(t)", 
                     legend= "none",
                     risk.table = TRUE)                           # Add risk table

ggsurv

## fitting survival model to mortality
my.kmmor<- survfit(Surv(mydata$mortality.interval,mydata$Mortality)~1)

ggsurv_morta <- ggsurvplot(my.kmmor, data = mydata,  title = "survival curve of calf mortality",
                           palette = "jco",       # jco palette
                           xlim = c(0,180),    # present narrower X axis, but not affect
                           # survival estimates.
                           ylim =  c(0.75,1),
                           xlab = "Time (days)",   # customize X axis label.
                           break.time.by = 30,     # break X axis in time intervals by 500.
                           ggtheme = theme_survminer(),
                           ylab = "% surviving,s(t)", 
                           tables.height = 0.25,
                           legend = "none",
                           risk.table = TRUE)                           # Add risk table
ggsurv_morta   
    
 
##### log rank test of predictor variables on calf mortality
mor<- survfit(Surv(mydata$mortality.interval, mydata$Mortality)~1)
summary(mor)


#### confidence interval of proportions
prop.test(3,15,correct=FALSE)## 95% CI male
prop.test(41,53,correct=FALSE)## 95% CI female
prop.test(13,53,correct=FALSE)## 95% CI young
prop.test(38,53,correct=FALSE)## 95% CI adult

### univariate cox proportional regression for calf morbidity
attach(mydata)
my.surv<- Surv(mydata$morbidity.interval, mydata$Morbidity)
sink(file = "morb_univariate.txt")
se<- coxph(my.surv ~ mydata$Sex, mydata)
se
bree<- coxph(my.surv~mydata$Breed, mydata)
bree
par<- coxph(my.surv~mydata$parity, mydata)
par
del_sta<- coxph(my.surv~mydata$Delivery.status, mydata)
del_sta
post_del<- coxph(my.surv~mydata$Postdelivery.complication, mydata)
post_del
inc<- coxph(my.surv~mydata$income.source, mydata)
inc
exp<- coxph(my.surv~mydata$Dairying.experience, mydata)
exp
man<- coxph(my.surv~mydata$Management, mydata)
man
Her<- coxph(my.surv~mydata$Herd.size, mydata)
Her
breeding<- coxph(my.surv~mydata$Breeding.method, mydata)
breeding
floor<- coxph(my.surv~mydata$Floor.type, mydata)
floor
calfpen< coxph(my.surv~mydata$Calving.pen, mydata)
calfpen
vac<- coxph(my.surv~mydata$Dam.vaccination, mydata)
vac
umb<- coxph(my.surv~mydata$umbilical.disinfection, mydata)
umb
housing<-coxph(my.surv~mydata$Calf.housing, mydata)
housing
fre<- coxph(my.surv~mydata$pen.cleaning.frequency, mydata)
fre
bedding<- coxph(my.surv~mydata$pen.bedding, mydata)
bedding
colos<- coxph(my.surv~mydata$first.colostrum,mydata)
colos
method<- coxph(my.surv~mydata$colo.feeding.method, mydata)
method
amount<- coxph(my.surv~mydata$Amount.of.milk.fed, mydata)
amount
water<- coxph(my.surv~mydata$Water.provision, mydata)
water

### univariate cox proportional regression for calf mortality
attach(mydata)

my.mor<- Surv(mydata$mortality.interval, mydata$Mortality)
sink(file = "mortality_univariate.txt")
se_mo<- coxph(my.mor ~ mydata$Sex, mydata)
se_mo
bree_mo<- coxph(my.mor~mydata$Breed, mydata)
bree_mo
par_mo<- coxph(my.mor~mydata$parity, mydata)
par_mo
del_sta_mo<- coxph(my.mor~mydata$Delivery.status, mydata)
del_sta_mo
post_del_mo<- coxph(my.mor~mydata$Postdelivery.complication, mydata)
post_del_mo
inc_mo<- coxph(my.mor~mydata$income.source, mydata)
inc_mo
exp_mo<- coxph(my.mor~mydata$Dairying.experience, mydata)
exp_mo
man_mo<- coxph(my.mor~mydata$Management, mydata)
man_mo
Her_mo<- coxph(my.mor~mydata$Herd.size, mydata)
Her_mo
breeding_mo<- coxph(my.mor~mydata$Breeding.method, mydata)
breeding_mo
floor_mo<- coxph(my.mor~mydata$Floor.type, mydata)
floor_mo
calfpen_mo< coxph(my.mor~mydata$Calving.pen, mydata)
calfpen_mo
vac_mo<- coxph(my.mor~mydata$Dam.vaccination, mydata)
vac_mo
umb_mo<- coxph(my.mor~mydata$umbilical.disinfection, mydata)
umb_mo
housing_mo<-coxph(my.mor~mydata$Calf.housing, mydata)
housing_mo
fre_mo<- coxph(my.mor~mydata$pen.cleaning.frequency, mydata)
fre_mo
bedding_mo<- coxph(my.mor~mydata$pen.bedding, mydata)
bedding_mo
colos_mo<- coxph(my.mor~mydata$first.colostrum,mydata)
colos_mo
method_mo<- coxph(my.mor~mydata$colo.feeding.method, mydata)
method_mo
amount_mo<- coxph(my.mor~mydata$Amount.of.milk.fed, mydata)
amount_mo
water_mo<- coxph(my.mor~mydata$Water.provision, mydata)
water_mo

#### multivariate cox proportional hazard model for calf morbidity
final_morbidity<- coxph(my.surv~mydata$Breed + mydata$parity + mydata$Delivery.status + mydata$Postdelivery.complication + mydata$income.source + mydata$Dairying.experience + mydata$Herd.size + mydata$Floor.type + mydata$Calving.pen + mydata$Calf.housing + mydata$pen.cleaning.frequency + mydata$pen.bedding + mydata$first.colostrum + mydata$Amount.of.milk.fed + mydata$Water.provision, control = coxph.control(iter.max = 50), mydata)
summary(final_morbidity)
#### multivariate cox proportional hazard model for calf mortality
final_mortality<- coxph(my.mort~mydata$parity + mydata$Delivery.status + mydata$Postdelivery.complication + mydata$Dairying.experience + mydata$Herd.size + mydata$Floor.type + mydata$Calving.pen + mydata$Calf.housing + mydata$pen.cleaning.frequency + mydata$pen.bedding + mydata$first.colostrum + mydata$colo.feeding.method + mydata$Amount.of.milk.fed, control = coxph.control(iter.max = 50), mydata)
summary(final_mortality)
#### survival curve for dam parity
parity<- survfit(Surv(mydata$morbidity.interval, mydata$Morbidity) ~ mydata$parity, data = mydata)
ggsurv_parity <- ggsurvplot(parity, data = mydata,  title = "Survival curve of dam parity",
                            linetype = "strata",
                            pval = TRUE, # Add p-value
                            conf.int = TRUE,
                            palette = c("#E7B800", "#2E9FDF"), # custom color palette
                            xlim = c(0,180),    # present narrower X axis
                            xlab = "Time (days)",   # customize X axis label.
                            break.time.by = 30,     # break X axis in time intervals by 500.
                            ggtheme = theme_survminer(aspect.ratio = 1),
                            ylab = "% surviving,s(t)", 
                            font.x= c(14,"black"),
                            font.y= c(14,"black"),
                            surv.plot.height= 2,
                            legend = "bottom", 
                            legend.title = "Parity",
                            legend.labs = c("Multiparous","Primiparous"))                           # Add risk table
ggsurv_parity   
#### survival curve for cleaning freq.
cleaning_freq<- survfit(Surv(mydata$morbidity.interval, mydata$Morbidity) ~ mydata$pen.cleaning.frequency, data = mydata)
ggsurv_freq <- ggsurvplot(cleaning_freq, data = mydata,  title = "Survival curve of pen cleaning freq.",
                            linetype = "strata",
                            pval = TRUE, # Add p-value
                            conf.int = TRUE,
                            palette = c("#E7B800", "#2E9FDF"), # custom color palette
                            xlim = c(0,180),    # present narrower X axis
                            xlab = "Time (days)",   # customize X axis label.
                            break.time.by = 30,     # break X axis in time intervals by 500.
                            ggtheme = theme_survminer(),
                            ylab = "% surviving,s(t)",
                            font.x= c(14,"black"),
                            font.y= c(14,"black"),
                            legend = "bottom", 
                            legend.title = "cleaning frequency",
                            legend.labs = c("Less frequently","Regularly"))
ggsurv_freq   
#### survival curve for first colostrum
colostr<- survfit(Surv(mydata$morbidity.interval, mydata$Morbidity) ~ mydata$first.colostrum, data = mydata)
ggsurv_colos <- ggsurvplot(colostr, data = mydata,  title = "Survival curve of first colostrum feeding",
                          linetype = "strata",
                          pval = TRUE, # Add p-value
                          conf.int = TRUE,
                          palette = c("#E7B800", "#2E9FDF"), # custom color palette
                          xlim = c(0,180),    # present narrower X axis
                          xlab = "Time (days)",   # customize X axis label.
                          break.time.by = 30,     # break X axis in time intervals by 500.
                          ggtheme = theme_survminer(),
                          ylab = "% surviving,s(t)", 
                          font.x= c(14,"black"),
                          font.y= c(14,"black"),
                          legend = "bottom", 
                          legend.title = "First Colostrum",
                          legend.labs = c("within 6 hours","After 6 hours"))
ggsurv_colos 
### survival curve of morbidity with delivery status
delivery<- survfit(Surv(mydata$morbidity.interval, mydata$Morbidity) ~ mydata$Delivery.status, data = mydata)
ggsurv_delivery <- ggsurvplot(delivery, data = mydata,  title = "Survival curve of delivery status",
                            linetype = "strata",
                            pval = TRUE, # Add p-value
                            conf.int = TRUE,
                            palette = c("#E7B800", "#2E9FDF"), # custom color palette
                            xlim = c(0,180),    # present narrower X axis
                            xlab = "Time (days)",   # customize X axis label.
                            break.time.by = 30,     # break X axis in time intervals by 500.
                            ggtheme = theme_survminer(aspect.ratio = 1),
                            ylab = "% surviving,s(t)", 
                            font.x= c(14,"black"),
                            font.y= c(14,"black"),
                            surv.plot.height= 2,
                            legend = "bottom", 
                            legend.title = "delivery status",
                            legend.labs = c("Normal","Dystocia"))                        
ggsurv_delivery   
##survival curves for mortality


#### survival curves for multivariate significant variables for mortality

#### survival curve for dam parity
parity_mor<- survfit(Surv(mydata$mortality.interval, mydata$Mortality) ~ mydata$parity, data = mydata)
ggsurv_parity_mor <- ggsurvplot(parity_mor, data = mydata,  title = "Survival curve of dam parity",
                            linetype = "strata",
                            pval = TRUE, # Add p-value
                            conf.int = TRUE,
                            palette = c("#E7B800", "#2E9FDF"), # custom color palette
                            xlim = c(0,180),    # present narrower X axis
                            xlab = "Time (days)",   # customize X axis label.
                            break.time.by = 30,     # break X axis in time intervals by 500.
                            ggtheme = theme_survminer(aspect.ratio = 1),
                            ylab = "% surviving,s(t)", 
                            font.x= c(14,"black"),
                            font.y= c(14,"black"),
                            surv.plot.height= 2,
                            legend = "bottom", 
                            legend.title = "Parity",
                            legend.labs = c("Multiparous","Primiparous"))                           # Add risk table

ggsurv_parity_mor   
#### survival curve for cleaning frequency
cleaning_freq_mor<- survfit(Surv(mydata$mortality.interval, mydata$Mortality) ~ mydata$pen.cleaning.frequency, data = mydata)
ggsurv_freq_mor <- ggsurvplot(cleaning_freq_mor, data = mydata,  title = "Survival curve of pen cleaning freq.",
                          linetype = "strata",
                          pval = TRUE, # Add p-value
                          conf.int = TRUE,
                          palette = c("#E7B800", "#2E9FDF"), # custom color palette
                          xlim = c(0,180),    # present narrower X axis
                          xlab = "Time (days)",   # customize X axis label.
                          break.time.by = 30,     # break X axis in time intervals by 500.
                          ggtheme = theme_survminer(),
                          ylab = "% surviving,s(t)",
                          font.x= c(14,"black"),
                          font.y= c(14,"black"),
                          legend = "bottom", 
                          legend.title = "cleaning frequency",
                          legend.labs = c("Less frequently","Regularly"))

ggsurv_freq_mor   

### survival curve of mortality with delivery status
delivery_mor<- survfit(Surv(mydata$mortality.interval, mydata$Mortality) ~ mydata$Delivery.status, data = mydata)
ggsurv_delivery_mor <- ggsurvplot(delivery_mor, data = mydata,  title = "Survival curve of delivery status",
                              linetype = "strata",
                              pval = TRUE, # Add p-value
                              conf.int = TRUE,
                              palette = c("#E7B800", "#2E9FDF"), # custom color palette
                              xlim = c(0,180),    # present narrower X axis
                              xlab = "Time (days)",   # customize X axis label.
                              break.time.by = 30,     # break X axis in time intervals by 500.
                              ggtheme = theme_survminer(aspect.ratio = 1),
                              ylab = "% surviving,s(t)", 
                              font.x= c(14,"black"),
                              font.y= c(14,"black"),
                              surv.plot.height= 2,
                              legend = "bottom", 
                              legend.title = "delivery status",
                              legend.labs = c("Normal","Dystocia"))                           # Add risk table

ggsurv_delivery_mor   


##test the proportional hazard assumption for covariates in the final model of calf mortality
attach(mydata)
my.surv<- Surv(mydata$mortality.interval,mydata$Mortality)
my_coxph_mortality<- coxph(my.surv~ parity + pen.cleaning.frequency + Delivery.status, control = coxph.control(iter.max = 50), mydata)
test.ph<- cox.zph(my_coxph_mortality)
test.ph

ggcoxzph(test.ph)

#### test the proportional hazard assumptions for covariates in the final model of calf morbidity

my.morb<- Surv(mydata$morbidity.interval,mydata$Morbidity)
my_coxph_morb<- coxph(my.morb~ parity + pen.cleaning.frequency + first.colostrum, control = coxph.control(iter.max = 50), mydata)
test.ph1<- cox.zph(my_coxph_morb)
test.ph1
ggcoxzph(test.ph1)

##### box plot of different causes of disease on morbidity

box<- read.csv(file= "cause.csv", header = T)
attach(box)
box1<- subset(box,nchar(as.character(box$morbidity.cause))>=5)
box2<- subset(box,nchar(as.character(box$mortality.cause))>=5)

#### boxplot of median calf morbidity based on their syndrome
theme_set(theme_pubr())
ggplot(data=box1, mapping=aes(x=morbidity.cause, y=interval.morbidity, fill = morbidity.cause))+geom_boxplot() + 
    labs(
        title = "Median age of calf morbidity days by syndrome",
        x = "Causes of morbidity",
        y = "Days at risk") +
        theme(plot.title = element_text(size = 16, face = "bold", hjust = 0.5), 
        legend.title= element_text(colour="blue", size=10, face="bold"))
       
#### boxplot of median calf mortality based on their syndrome
theme_set(theme_pubr())
ggplot(data=box2, mapping=aes(x=mortality.cause, y=interval.mortality, fill = mortality.cause))+geom_boxplot() + 
    labs(
        title = "Median age of calf mortality days by syndrome",
        x = "Causes of mortality",
        scale_y_continuous(breaks=c(0,30,60,90,120,150,180)),
        y = "Days at risk") +
    theme(plot.title = element_text(size = 16, face = "bold", hjust = 0.5), 
          legend.title= element_text(colour="blue", size=10, face="bold"))


#### median days of morbidity for disease syndrome
survfit(Surv(mydata$morbidity.interval,mydata$Morbidity)~mydata$Cause.morbidity)

##### median days of mortality
survfit(Surv(mydata$mortality.interval,mydata$Mortality)~mydata$Cause.mortality)