Factors Affecting Survival in Heart Failure Patients
Somaya AlGabry
05/08/2021
Introduction:
This is a survival analysis for death in HF patients. the analysis concluded that age, male gender, length of stay, prior missed outpatient appointments and many comorbidities have evidence that they affect death rate in heart failure.
Load Packages:
library(survival)
library(ggplot2)## Warning: package 'ggplot2' was built under R version 4.0.5library(survminer)## Warning: package 'survminer' was built under R version 4.0.5## Loading required package: ggpubr## Warning: package 'ggpubr' was built under R version 4.0.5library(gmodels)## Warning: package 'gmodels' was built under R version 4.0.5Load Data:
filename <- "HF.csv"
fileURL <- "https://d3c33hcgiwev3.cloudfront.net/6AiBbg-BEem6Gg6vVM6M8A_e872b4600f8111e9b2f4133a1edfbb40_simulated-HF-mort-data-for-GMPH-_1K_-final-_2_.csv?Expires=1628294400&Signature=W6JL~~LOJRnDWbOy80-uNeDtawKxqe~CwY9OUiOepmB2ALINxQN3ZlrsOzFegePzP6DXpQqnp3LavLWlHSnBIx6vNPfkhv6dWo8MElYTKll-dne15MYRKS0Xs8bw~Dx7xWxAj0WEDX4XGKIeHjudu4hgVQqawwtsq53A8x-~D3o_&Key-Pair-Id=APKAJLTNE6QMUY6HBC5A"
if (!file.exists(filename)){
download.file(fileURL, filename)
}
HF <- read.csv("HF.csv")Define Variables:
gender <- as.factor(HF[,"gender"]) # categorical variable
fu_time <- HF[,"fu_time"] # continuous variable (numeric)
death <- HF[,"death"] # binary variable (numeric)
age <- HF[,"age"]
ethnicgroup <- factor(HF[,"ethnicgroup"])
prior_dnas <- HF[,"prior_dnas"]
prior_dnas_c <- factor(HF[,"prior_dnas"])
copd <- factor(HF[,"copd"])
quintile <- factor(HF[, "quintile"])
los <- HF[,"los"]
ihd <- factor(HF[,"ihd"])
stroke <- factor(HF[,"stroke"])
arrhythmias <- factor(HF[,"arrhythmias"])
vd <- factor(HF[,"valvular_disease"])
pvd <- factor(HF[,"pvd"])
pneumonia <- factor(HF[,"pneumonia"])
renal <- factor(HF[,"renal_disease"])
htn <- factor(HF[,"hypertension"])
cancer <- factor(HF[,"cancer"])
met_cancer <- factor(HF[,"metastatic_cancer"])
dementia <- factor(HF[,"dementia"])
senile <- factor(HF[,"senile"])
cog_imp <- as.factor(ifelse(HF$dementia == 1 | HF$senile == 1, 1, 0))
mental <- factor(HF[,"mental_health"])
prior_attend <- HF[,"prior_appts_attended"]
cabg <- factor(HF[,"cabg"])
defib <- factor(HF[,"defib"])describe variables:
dim(HF)## [1] 1000 31table(gender, useNA = "always")## gender
## 1 2 <NA>
## 548 452 0summary(fu_time)## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 0.0 117.8 475.0 466.0 743.2 1107.0table(death, useNA = "always")## death
## 0 1 <NA>
## 508 492 0summary(age)## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 29.00 73.00 80.00 78.73 87.00 102.00hist(age)
table(ethnicgroup, useNA = "always")## ethnicgroup
## 1 2 3 9 <NA>
## 889 17 34 17 43summary(prior_dnas)## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 0.000 0.000 0.000 0.501 1.000 10.000table(prior_dnas_c)## prior_dnas_c
## 0 1 2 3 4 5 6 7 8 10
## 732 156 50 34 17 3 3 2 1 2CrossTable(copd, death)##
##
## Cell Contents
## |-------------------------|
## | N |
## | Chi-square contribution |
## | N / Row Total |
## | N / Col Total |
## | N / Table Total |
## |-------------------------|
##
##
## Total Observations in Table: 1000
##
##
## | death
## copd | 0 | 1 | Row Total |
## -------------|-----------|-----------|-----------|
## 0 | 393 | 365 | 758 |
## | 0.164 | 0.169 | |
## | 0.518 | 0.482 | 0.758 |
## | 0.774 | 0.742 | |
## | 0.393 | 0.365 | |
## -------------|-----------|-----------|-----------|
## 1 | 115 | 127 | 242 |
## | 0.512 | 0.529 | |
## | 0.475 | 0.525 | 0.242 |
## | 0.226 | 0.258 | |
## | 0.115 | 0.127 | |
## -------------|-----------|-----------|-----------|
## Column Total | 508 | 492 | 1000 |
## | 0.508 | 0.492 | |
## -------------|-----------|-----------|-----------|
##
## table(quintile, useNA = "always")## quintile
## 0 1 2 3 4 5 <NA>
## 4 138 205 211 220 216 6# there is an error here ( the 0 value). we can't use it as the reference value. so I am going to change it.
summary(los)## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 0.00 3.00 7.00 10.77 13.00 89.00CrossTable(ihd, death)##
##
## Cell Contents
## |-------------------------|
## | N |
## | Chi-square contribution |
## | N / Row Total |
## | N / Col Total |
## | N / Table Total |
## |-------------------------|
##
##
## Total Observations in Table: 1000
##
##
## | death
## ihd | 0 | 1 | Row Total |
## -------------|-----------|-----------|-----------|
## 0 | 266 | 239 | 505 |
## | 0.349 | 0.360 | |
## | 0.527 | 0.473 | 0.505 |
## | 0.524 | 0.486 | |
## | 0.266 | 0.239 | |
## -------------|-----------|-----------|-----------|
## 1 | 242 | 253 | 495 |
## | 0.356 | 0.367 | |
## | 0.489 | 0.511 | 0.495 |
## | 0.476 | 0.514 | |
## | 0.242 | 0.253 | |
## -------------|-----------|-----------|-----------|
## Column Total | 508 | 492 | 1000 |
## | 0.508 | 0.492 | |
## -------------|-----------|-----------|-----------|
##
## CrossTable(stroke, death)##
##
## Cell Contents
## |-------------------------|
## | N |
## | Chi-square contribution |
## | N / Row Total |
## | N / Col Total |
## | N / Table Total |
## |-------------------------|
##
##
## Total Observations in Table: 1000
##
##
## | death
## stroke | 0 | 1 | Row Total |
## -------------|-----------|-----------|-----------|
## 0 | 500 | 480 | 980 |
## | 0.009 | 0.010 | |
## | 0.510 | 0.490 | 0.980 |
## | 0.984 | 0.976 | |
## | 0.500 | 0.480 | |
## -------------|-----------|-----------|-----------|
## 1 | 8 | 12 | 20 |
## | 0.459 | 0.474 | |
## | 0.400 | 0.600 | 0.020 |
## | 0.016 | 0.024 | |
## | 0.008 | 0.012 | |
## -------------|-----------|-----------|-----------|
## Column Total | 508 | 492 | 1000 |
## | 0.508 | 0.492 | |
## -------------|-----------|-----------|-----------|
##
## CrossTable(arrhythmias, death)##
##
## Cell Contents
## |-------------------------|
## | N |
## | Chi-square contribution |
## | N / Row Total |
## | N / Col Total |
## | N / Table Total |
## |-------------------------|
##
##
## Total Observations in Table: 1000
##
##
## | death
## arrhythmias | 0 | 1 | Row Total |
## -------------|-----------|-----------|-----------|
## 0 | 249 | 261 | 510 |
## | 0.392 | 0.405 | |
## | 0.488 | 0.512 | 0.510 |
## | 0.490 | 0.530 | |
## | 0.249 | 0.261 | |
## -------------|-----------|-----------|-----------|
## 1 | 259 | 231 | 490 |
## | 0.408 | 0.421 | |
## | 0.529 | 0.471 | 0.490 |
## | 0.510 | 0.470 | |
## | 0.259 | 0.231 | |
## -------------|-----------|-----------|-----------|
## Column Total | 508 | 492 | 1000 |
## | 0.508 | 0.492 | |
## -------------|-----------|-----------|-----------|
##
## CrossTable(vd, death)##
##
## Cell Contents
## |-------------------------|
## | N |
## | Chi-square contribution |
## | N / Row Total |
## | N / Col Total |
## | N / Table Total |
## |-------------------------|
##
##
## Total Observations in Table: 1000
##
##
## | death
## vd | 0 | 1 | Row Total |
## -------------|-----------|-----------|-----------|
## 0 | 380 | 376 | 756 |
## | 0.043 | 0.044 | |
## | 0.503 | 0.497 | 0.756 |
## | 0.748 | 0.764 | |
## | 0.380 | 0.376 | |
## -------------|-----------|-----------|-----------|
## 1 | 128 | 116 | 244 |
## | 0.132 | 0.136 | |
## | 0.525 | 0.475 | 0.244 |
## | 0.252 | 0.236 | |
## | 0.128 | 0.116 | |
## -------------|-----------|-----------|-----------|
## Column Total | 508 | 492 | 1000 |
## | 0.508 | 0.492 | |
## -------------|-----------|-----------|-----------|
##
## CrossTable(pvd, death)##
##
## Cell Contents
## |-------------------------|
## | N |
## | Chi-square contribution |
## | N / Row Total |
## | N / Col Total |
## | N / Table Total |
## |-------------------------|
##
##
## Total Observations in Table: 1000
##
##
## | death
## pvd | 0 | 1 | Row Total |
## -------------|-----------|-----------|-----------|
## 0 | 454 | 446 | 900 |
## | 0.022 | 0.023 | |
## | 0.504 | 0.496 | 0.900 |
## | 0.894 | 0.907 | |
## | 0.454 | 0.446 | |
## -------------|-----------|-----------|-----------|
## 1 | 54 | 46 | 100 |
## | 0.202 | 0.208 | |
## | 0.540 | 0.460 | 0.100 |
## | 0.106 | 0.093 | |
## | 0.054 | 0.046 | |
## -------------|-----------|-----------|-----------|
## Column Total | 508 | 492 | 1000 |
## | 0.508 | 0.492 | |
## -------------|-----------|-----------|-----------|
##
## CrossTable(pneumonia, death)##
##
## Cell Contents
## |-------------------------|
## | N |
## | Chi-square contribution |
## | N / Row Total |
## | N / Col Total |
## | N / Table Total |
## |-------------------------|
##
##
## Total Observations in Table: 1000
##
##
## | death
## pneumonia | 0 | 1 | Row Total |
## -------------|-----------|-----------|-----------|
## 0 | 457 | 426 | 883 |
## | 0.159 | 0.164 | |
## | 0.518 | 0.482 | 0.883 |
## | 0.900 | 0.866 | |
## | 0.457 | 0.426 | |
## -------------|-----------|-----------|-----------|
## 1 | 51 | 66 | 117 |
## | 1.197 | 1.236 | |
## | 0.436 | 0.564 | 0.117 |
## | 0.100 | 0.134 | |
## | 0.051 | 0.066 | |
## -------------|-----------|-----------|-----------|
## Column Total | 508 | 492 | 1000 |
## | 0.508 | 0.492 | |
## -------------|-----------|-----------|-----------|
##
## CrossTable(renal, death)##
##
## Cell Contents
## |-------------------------|
## | N |
## | Chi-square contribution |
## | N / Row Total |
## | N / Col Total |
## | N / Table Total |
## |-------------------------|
##
##
## Total Observations in Table: 1000
##
##
## | death
## renal | 0 | 1 | Row Total |
## -------------|-----------|-----------|-----------|
## 0 | 394 | 357 | 751 |
## | 0.409 | 0.422 | |
## | 0.525 | 0.475 | 0.751 |
## | 0.776 | 0.726 | |
## | 0.394 | 0.357 | |
## -------------|-----------|-----------|-----------|
## 1 | 114 | 135 | 249 |
## | 1.234 | 1.274 | |
## | 0.458 | 0.542 | 0.249 |
## | 0.224 | 0.274 | |
## | 0.114 | 0.135 | |
## -------------|-----------|-----------|-----------|
## Column Total | 508 | 492 | 1000 |
## | 0.508 | 0.492 | |
## -------------|-----------|-----------|-----------|
##
## CrossTable(htn, death)##
##
## Cell Contents
## |-------------------------|
## | N |
## | Chi-square contribution |
## | N / Row Total |
## | N / Col Total |
## | N / Table Total |
## |-------------------------|
##
##
## Total Observations in Table: 1000
##
##
## | death
## htn | 0 | 1 | Row Total |
## -------------|-----------|-----------|-----------|
## 0 | 187 | 192 | 379 |
## | 0.159 | 0.164 | |
## | 0.493 | 0.507 | 0.379 |
## | 0.368 | 0.390 | |
## | 0.187 | 0.192 | |
## -------------|-----------|-----------|-----------|
## 1 | 321 | 300 | 621 |
## | 0.097 | 0.100 | |
## | 0.517 | 0.483 | 0.621 |
## | 0.632 | 0.610 | |
## | 0.321 | 0.300 | |
## -------------|-----------|-----------|-----------|
## Column Total | 508 | 492 | 1000 |
## | 0.508 | 0.492 | |
## -------------|-----------|-----------|-----------|
##
## CrossTable(cancer, death)##
##
## Cell Contents
## |-------------------------|
## | N |
## | Chi-square contribution |
## | N / Row Total |
## | N / Col Total |
## | N / Table Total |
## |-------------------------|
##
##
## Total Observations in Table: 1000
##
##
## | death
## cancer | 0 | 1 | Row Total |
## -------------|-----------|-----------|-----------|
## 0 | 488 | 461 | 949 |
## | 0.072 | 0.075 | |
## | 0.514 | 0.486 | 0.949 |
## | 0.961 | 0.937 | |
## | 0.488 | 0.461 | |
## -------------|-----------|-----------|-----------|
## 1 | 20 | 31 | 51 |
## | 1.347 | 1.391 | |
## | 0.392 | 0.608 | 0.051 |
## | 0.039 | 0.063 | |
## | 0.020 | 0.031 | |
## -------------|-----------|-----------|-----------|
## Column Total | 508 | 492 | 1000 |
## | 0.508 | 0.492 | |
## -------------|-----------|-----------|-----------|
##
## CrossTable(met_cancer, death)##
##
## Cell Contents
## |-------------------------|
## | N |
## | Chi-square contribution |
## | N / Row Total |
## | N / Col Total |
## | N / Table Total |
## |-------------------------|
##
##
## Total Observations in Table: 1000
##
##
## | death
## met_cancer | 0 | 1 | Row Total |
## -------------|-----------|-----------|-----------|
## 0 | 507 | 484 | 991 |
## | 0.025 | 0.026 | |
## | 0.512 | 0.488 | 0.991 |
## | 0.998 | 0.984 | |
## | 0.507 | 0.484 | |
## -------------|-----------|-----------|-----------|
## 1 | 1 | 8 | 9 |
## | 2.791 | 2.881 | |
## | 0.111 | 0.889 | 0.009 |
## | 0.002 | 0.016 | |
## | 0.001 | 0.008 | |
## -------------|-----------|-----------|-----------|
## Column Total | 508 | 492 | 1000 |
## | 0.508 | 0.492 | |
## -------------|-----------|-----------|-----------|
##
## CrossTable(dementia, death)##
##
## Cell Contents
## |-------------------------|
## | N |
## | Chi-square contribution |
## | N / Row Total |
## | N / Col Total |
## | N / Table Total |
## |-------------------------|
##
##
## Total Observations in Table: 1000
##
##
## | death
## dementia | 0 | 1 | Row Total |
## -------------|-----------|-----------|-----------|
## 0 | 497 | 458 | 955 |
## | 0.290 | 0.299 | |
## | 0.520 | 0.480 | 0.955 |
## | 0.978 | 0.931 | |
## | 0.497 | 0.458 | |
## -------------|-----------|-----------|-----------|
## 1 | 11 | 34 | 45 |
## | 6.153 | 6.353 | |
## | 0.244 | 0.756 | 0.045 |
## | 0.022 | 0.069 | |
## | 0.011 | 0.034 | |
## -------------|-----------|-----------|-----------|
## Column Total | 508 | 492 | 1000 |
## | 0.508 | 0.492 | |
## -------------|-----------|-----------|-----------|
##
## CrossTable(senile, death)##
##
## Cell Contents
## |-------------------------|
## | N |
## | Chi-square contribution |
## | N / Row Total |
## | N / Col Total |
## | N / Table Total |
## |-------------------------|
##
##
## Total Observations in Table: 1000
##
##
## | death
## senile | 0 | 1 | Row Total |
## -------------|-----------|-----------|-----------|
## 0 | 497 | 459 | 956 |
## | 0.265 | 0.274 | |
## | 0.520 | 0.480 | 0.956 |
## | 0.978 | 0.933 | |
## | 0.497 | 0.459 | |
## -------------|-----------|-----------|-----------|
## 1 | 11 | 33 | 44 |
## | 5.765 | 5.953 | |
## | 0.250 | 0.750 | 0.044 |
## | 0.022 | 0.067 | |
## | 0.011 | 0.033 | |
## -------------|-----------|-----------|-----------|
## Column Total | 508 | 492 | 1000 |
## | 0.508 | 0.492 | |
## -------------|-----------|-----------|-----------|
##
## CrossTable(cog_imp, death)##
##
## Cell Contents
## |-------------------------|
## | N |
## | Chi-square contribution |
## | N / Row Total |
## | N / Col Total |
## | N / Table Total |
## |-------------------------|
##
##
## Total Observations in Table: 1000
##
##
## | death
## cog_imp | 0 | 1 | Row Total |
## -------------|-----------|-----------|-----------|
## 0 | 488 | 433 | 921 |
## | 0.866 | 0.894 | |
## | 0.530 | 0.470 | 0.921 |
## | 0.961 | 0.880 | |
## | 0.488 | 0.433 | |
## -------------|-----------|-----------|-----------|
## 1 | 20 | 59 | 79 |
## | 10.099 | 10.428 | |
## | 0.253 | 0.747 | 0.079 |
## | 0.039 | 0.120 | |
## | 0.020 | 0.059 | |
## -------------|-----------|-----------|-----------|
## Column Total | 508 | 492 | 1000 |
## | 0.508 | 0.492 | |
## -------------|-----------|-----------|-----------|
##
## CrossTable(mental, death)##
##
## Cell Contents
## |-------------------------|
## | N |
## | Chi-square contribution |
## | N / Row Total |
## | N / Col Total |
## | N / Table Total |
## |-------------------------|
##
##
## Total Observations in Table: 1000
##
##
## | death
## mental | 0 | 1 | Row Total |
## -------------|-----------|-----------|-----------|
## 0 | 454 | 457 | 911 |
## | 0.167 | 0.172 | |
## | 0.498 | 0.502 | 0.911 |
## | 0.894 | 0.929 | |
## | 0.454 | 0.457 | |
## -------------|-----------|-----------|-----------|
## 1 | 54 | 35 | 89 |
## | 1.708 | 1.764 | |
## | 0.607 | 0.393 | 0.089 |
## | 0.106 | 0.071 | |
## | 0.054 | 0.035 | |
## -------------|-----------|-----------|-----------|
## Column Total | 508 | 492 | 1000 |
## | 0.508 | 0.492 | |
## -------------|-----------|-----------|-----------|
##
## summary(prior_attend)## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 0.000 1.000 3.000 5.538 8.000 62.000hist(prior_attend)
CrossTable(cabg, death)##
##
## Cell Contents
## |-------------------------|
## | N |
## | Chi-square contribution |
## | N / Row Total |
## | N / Col Total |
## | N / Table Total |
## |-------------------------|
##
##
## Total Observations in Table: 1000
##
##
## | death
## cabg | 0 | 1 | Row Total |
## -------------|-----------|-----------|-----------|
## 0 | 495 | 491 | 986 |
## | 0.069 | 0.071 | |
## | 0.502 | 0.498 | 0.986 |
## | 0.974 | 0.998 | |
## | 0.495 | 0.491 | |
## -------------|-----------|-----------|-----------|
## 1 | 13 | 1 | 14 |
## | 4.875 | 5.033 | |
## | 0.929 | 0.071 | 0.014 |
## | 0.026 | 0.002 | |
## | 0.013 | 0.001 | |
## -------------|-----------|-----------|-----------|
## Column Total | 508 | 492 | 1000 |
## | 0.508 | 0.492 | |
## -------------|-----------|-----------|-----------|
##
## CrossTable(defib, death)##
##
## Cell Contents
## |-------------------------|
## | N |
## | Chi-square contribution |
## | N / Row Total |
## | N / Col Total |
## | N / Table Total |
## |-------------------------|
##
##
## Total Observations in Table: 1000
##
##
## | death
## defib | 0 | 1 | Row Total |
## -------------|-----------|-----------|-----------|
## 0 | 504 | 490 | 994 |
## | 0.002 | 0.002 | |
## | 0.507 | 0.493 | 0.994 |
## | 0.992 | 0.996 | |
## | 0.504 | 0.490 | |
## -------------|-----------|-----------|-----------|
## 1 | 4 | 2 | 6 |
## | 0.297 | 0.307 | |
## | 0.667 | 0.333 | 0.006 |
## | 0.008 | 0.004 | |
## | 0.004 | 0.002 | |
## -------------|-----------|-----------|-----------|
## Column Total | 508 | 492 | 1000 |
## | 0.508 | 0.492 | |
## -------------|-----------|-----------|-----------|
##
## Kaplan-Meier Plot:
km_fit <- survfit(Surv(fu_time, death) ~ 1)
plot(km_fit)
summary(km_fit, times = c(1:7,30,60,90*(1:10)))## Call: survfit(formula = Surv(fu_time, death) ~ 1)
##
## time n.risk n.event survival std.err lower 95% CI upper 95% CI
## 1 992 12 0.988 0.00346 0.981 0.995
## 2 973 7 0.981 0.00435 0.972 0.989
## 3 963 5 0.976 0.00489 0.966 0.985
## 4 954 6 0.970 0.00546 0.959 0.980
## 5 945 5 0.964 0.00590 0.953 0.976
## 6 938 1 0.963 0.00598 0.952 0.975
## 7 933 1 0.962 0.00606 0.951 0.974
## 30 865 39 0.921 0.00865 0.905 0.939
## 60 809 28 0.891 0.01010 0.871 0.911
## 90 770 24 0.864 0.01117 0.843 0.887
## 180 698 43 0.815 0.01282 0.790 0.841
## 270 653 24 0.787 0.01363 0.760 0.814
## 360 619 21 0.761 0.01428 0.733 0.789
## 450 525 44 0.705 0.01554 0.675 0.736
## 540 429 47 0.639 0.01681 0.607 0.673
## 630 362 32 0.589 0.01765 0.556 0.625
## 720 266 43 0.514 0.01876 0.479 0.552
## 810 190 31 0.448 0.01979 0.411 0.488
## 900 126 26 0.378 0.02098 0.339 0.421# split the curve by gender:
km_gender_fit <- survfit(Surv(fu_time, death) ~ gender)
plot(km_gender_fit)
survdiff(Surv(fu_time, death) ~ gender, rho=0) ##logrank test## Call:
## survdiff(formula = Surv(fu_time, death) ~ gender, rho = 0)
##
## N Observed Expected (O-E)^2/E (O-E)^2/V
## gender=1 548 268 271 0.0365 0.082
## gender=2 452 224 221 0.0448 0.082
##
## Chisq= 0.1 on 1 degrees of freedom, p= 0.8# split the curve by age categories:
age_categorised <- ifelse(age < 65, "under 65",
ifelse(age >= 65, "65 or above", NA))
age_categorised <- factor(age_categorised)
table(age_categorised, useNA = "always")## age_categorised
## 65 or above under 65 <NA>
## 885 115 0km_age_fit <- survfit(Surv(fu_time, death) ~ age_categorised)
plot(km_age_fit)
survdiff(Surv(fu_time, death) ~ age_categorised, rho=0)## Call:
## survdiff(formula = Surv(fu_time, death) ~ age_categorised, rho = 0)
##
## N Observed Expected (O-E)^2/E (O-E)^2/V
## age_categorised=65 or above 885 474 425 5.65 41.7
## age_categorised=under 65 115 18 67 35.85 41.7
##
## Chisq= 41.7 on 1 degrees of freedom, p= 1e-10Simple Cox Regression:
cox <- coxph(Surv(fu_time, death) ~ ethnicgroup)
summary(cox)## Call:
## coxph(formula = Surv(fu_time, death) ~ ethnicgroup)
##
## n= 957, number of events= 471
## (43 observations deleted due to missingness)
##
## coef exp(coef) se(coef) z Pr(>|z|)
## ethnicgroup2 -0.06428 0.93774 0.32000 -0.201 0.84078
## ethnicgroup3 -1.19586 0.30244 0.41108 -2.909 0.00362 **
## ethnicgroup9 0.07394 1.07674 0.35706 0.207 0.83596
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## exp(coef) exp(-coef) lower .95 upper .95
## ethnicgroup2 0.9377 1.0664 0.5008 1.7558
## ethnicgroup3 0.3024 3.3064 0.1351 0.6769
## ethnicgroup9 1.0767 0.9287 0.5348 2.1679
##
## Concordance= 0.516 (se = 0.006 )
## Likelihood ratio test= 12.99 on 3 df, p=0.005
## Wald test = 8.55 on 3 df, p=0.04
## Score (logrank) test = 9.61 on 3 df, p=0.02# add missing ethnic groups to new "unknown" category:
levels(ethnicgroup) <- c(levels(ethnicgroup),"8") # add level 8 to the factor
ethnicgroup[is.na(ethnicgroup)] <- "8" # Change NA to "None"
cox <- coxph(Surv(fu_time, death) ~ ethnicgroup)
summary(cox)## Call:
## coxph(formula = Surv(fu_time, death) ~ ethnicgroup)
##
## n= 1000, number of events= 492
##
## coef exp(coef) se(coef) z Pr(>|z|)
## ethnicgroup2 -0.06573 0.93638 0.31999 -0.205 0.83725
## ethnicgroup3 -1.19368 0.30310 0.41107 -2.904 0.00369 **
## ethnicgroup9 0.08160 1.08502 0.35706 0.229 0.81923
## ethnicgroup8 -0.02353 0.97675 0.22363 -0.105 0.91621
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## exp(coef) exp(-coef) lower .95 upper .95
## ethnicgroup2 0.9364 1.0679 0.5001 1.7532
## ethnicgroup3 0.3031 3.2992 0.1354 0.6784
## ethnicgroup9 1.0850 0.9216 0.5389 2.1846
## ethnicgroup8 0.9767 1.0238 0.6301 1.5140
##
## Concordance= 0.518 (se = 0.008 )
## Likelihood ratio test= 12.95 on 4 df, p=0.01
## Wald test = 8.53 on 4 df, p=0.07
## Score (logrank) test = 9.58 on 4 df, p=0.05Multiple Cox Regression Model:
reference article for affecting factor: https://www.ncbi.nlm.nih.gov/books/NBK513479/
mcox <- coxph(Surv(fu_time, death) ~ age + gender + ethnicgroup + los + ihd + stroke + arrhythmias + vd + pvd + copd + pneumonia + renal + htn + cancer + met_cancer + cog_imp + mental + prior_attend + prior_dnas + cabg + defib)
summary(mcox)## Call:
## coxph(formula = Surv(fu_time, death) ~ age + gender + ethnicgroup +
## los + ihd + stroke + arrhythmias + vd + pvd + copd + pneumonia +
## renal + htn + cancer + met_cancer + cog_imp + mental + prior_attend +
## prior_dnas + cabg + defib)
##
## n= 1000, number of events= 492
##
## coef exp(coef) se(coef) z Pr(>|z|)
## age 0.057228 1.058897 0.005805 9.859 < 2e-16 ***
## gender2 -0.245671 0.782180 0.098309 -2.499 0.012456 *
## ethnicgroup2 -0.210588 0.810108 0.357098 -0.590 0.555379
## ethnicgroup3 -0.798948 0.449802 0.416732 -1.917 0.055216 .
## ethnicgroup9 0.432757 1.541501 0.365419 1.184 0.236304
## ethnicgroup8 -0.131358 0.876904 0.232841 -0.564 0.572650
## los 0.011399 1.011465 0.003286 3.469 0.000522 ***
## ihd1 0.200584 1.222117 0.096332 2.082 0.037323 *
## stroke1 0.029573 1.030015 0.303815 0.097 0.922457
## arrhythmias1 -0.143632 0.866206 0.095310 -1.507 0.131809
## vd1 0.202418 1.224359 0.108949 1.858 0.063182 .
## pvd1 0.039183 1.039960 0.163059 0.240 0.810101
## copd1 0.084882 1.088588 0.108785 0.780 0.435230
## pneumonia1 0.294409 1.342332 0.141618 2.079 0.037627 *
## renal1 0.149282 1.161000 0.110007 1.357 0.174774
## htn1 -0.039846 0.960937 0.096734 -0.412 0.680404
## cancer1 0.313144 1.367719 0.209448 1.495 0.134890
## met_cancer1 2.191424 8.947943 0.400448 5.472 4.44e-08 ***
## cog_imp1 0.290274 1.336793 0.150246 1.932 0.053361 .
## mental1 -0.061308 0.940534 0.182404 -0.336 0.736788
## prior_attend -0.009295 0.990748 0.008091 -1.149 0.250644
## prior_dnas 0.126733 1.135114 0.043682 2.901 0.003716 **
## cabg1 -1.805277 0.164429 1.004864 -1.797 0.072409 .
## defib1 -0.601418 0.548034 0.723481 -0.831 0.405813
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## exp(coef) exp(-coef) lower .95 upper .95
## age 1.0589 0.9444 1.04692 1.0710
## gender2 0.7822 1.2785 0.64510 0.9484
## ethnicgroup2 0.8101 1.2344 0.40233 1.6312
## ethnicgroup3 0.4498 2.2232 0.19875 1.0180
## ethnicgroup9 1.5415 0.6487 0.75318 3.1549
## ethnicgroup8 0.8769 1.1404 0.55559 1.3840
## los 1.0115 0.9887 1.00497 1.0180
## ihd1 1.2221 0.8183 1.01185 1.4761
## stroke1 1.0300 0.9709 0.56785 1.8683
## arrhythmias1 0.8662 1.1545 0.71861 1.0441
## vd1 1.2244 0.8168 0.98894 1.5158
## pvd1 1.0400 0.9616 0.75548 1.4316
## copd1 1.0886 0.9186 0.87956 1.3473
## pneumonia1 1.3423 0.7450 1.01699 1.7718
## renal1 1.1610 0.8613 0.93582 1.4404
## htn1 0.9609 1.0407 0.79498 1.1615
## cancer1 1.3677 0.7311 0.90722 2.0620
## met_cancer1 8.9479 0.1118 4.08189 19.6148
## cog_imp1 1.3368 0.7481 0.99581 1.7945
## mental1 0.9405 1.0632 0.65783 1.3447
## prior_attend 0.9907 1.0093 0.97516 1.0066
## prior_dnas 1.1351 0.8810 1.04198 1.2366
## cabg1 0.1644 6.0817 0.02294 1.1785
## defib1 0.5480 1.8247 0.13273 2.2627
##
## Concordance= 0.713 (se = 0.012 )
## Likelihood ratio test= 237.2 on 24 df, p=<2e-16
## Wald test = 223.8 on 24 df, p=<2e-16
## Score (logrank) test = 244.5 on 24 df, p=<2e-16# backward deletion:
mcox2 <- coxph(Surv(fu_time, death) ~ age + gender + los + ihd + vd + pneumonia + met_cancer + cog_imp + prior_dnas )
summary(mcox2)## Call:
## coxph(formula = Surv(fu_time, death) ~ age + gender + los + ihd +
## vd + pneumonia + met_cancer + cog_imp + prior_dnas)
##
## n= 1000, number of events= 492
##
## coef exp(coef) se(coef) z Pr(>|z|)
## age 0.060630 1.062505 0.005684 10.667 < 2e-16 ***
## gender2 -0.266403 0.766130 0.095243 -2.797 0.005157 **
## los 0.011838 1.011908 0.003218 3.679 0.000234 ***
## ihd1 0.188195 1.207069 0.093562 2.011 0.044278 *
## vd1 0.201040 1.222674 0.107514 1.870 0.061497 .
## pneumonia1 0.322339 1.380352 0.138607 2.326 0.020042 *
## met_cancer1 2.457014 11.669914 0.364843 6.734 1.65e-11 ***
## cog_imp1 0.292821 1.340203 0.146576 1.998 0.045744 *
## prior_dnas 0.106943 1.112871 0.037477 2.854 0.004323 **
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## exp(coef) exp(-coef) lower .95 upper .95
## age 1.0625 0.94117 1.0507 1.0744
## gender2 0.7661 1.30526 0.6357 0.9234
## los 1.0119 0.98823 1.0055 1.0183
## ihd1 1.2071 0.82845 1.0048 1.4500
## vd1 1.2227 0.81788 0.9904 1.5095
## pneumonia1 1.3804 0.72445 1.0520 1.8112
## met_cancer1 11.6699 0.08569 5.7084 23.8574
## cog_imp1 1.3402 0.74616 1.0056 1.7862
## prior_dnas 1.1129 0.89858 1.0341 1.1977
##
## Concordance= 0.704 (se = 0.012 )
## Likelihood ratio test= 215.3 on 9 df, p=<2e-16
## Wald test = 208.9 on 9 df, p=<2e-16
## Score (logrank) test = 223.9 on 9 df, p=<2e-16# check the proportionality assumption:
zph <- cox.zph(mcox2, transform="km", global=TRUE)
zph## chisq df p
## age 0.547 1 0.4596
## gender 1.482 1 0.2234
## los 0.216 1 0.6423
## ihd 7.699 1 0.0055
## vd 0.539 1 0.4629
## pneumonia 1.234 1 0.2666
## met_cancer 0.488 1 0.4848
## cog_imp 0.350 1 0.5539
## prior_dnas 0.171 1 0.6790
## GLOBAL 12.332 9 0.1953ggcoxzph(zph)
# p-value of ihd is too low. I am going to remove it.
mcox2 <- coxph(Surv(fu_time, death) ~ age + gender + los + vd + pneumonia + met_cancer + cog_imp + prior_dnas)
summary(mcox2)## Call:
## coxph(formula = Surv(fu_time, death) ~ age + gender + los + vd +
## pneumonia + met_cancer + cog_imp + prior_dnas)
##
## n= 1000, number of events= 492
##
## coef exp(coef) se(coef) z Pr(>|z|)
## age 0.060063 1.061904 0.005632 10.665 < 2e-16 ***
## gender2 -0.290035 0.748237 0.094303 -3.076 0.002101 **
## los 0.011124 1.011187 0.003209 3.466 0.000528 ***
## vd1 0.198727 1.219849 0.107519 1.848 0.064558 .
## pneumonia1 0.349217 1.417956 0.138080 2.529 0.011436 *
## met_cancer1 2.501949 12.206264 0.364470 6.865 6.67e-12 ***
## cog_imp1 0.289850 1.336226 0.146679 1.976 0.048146 *
## prior_dnas 0.115907 1.122892 0.037286 3.109 0.001880 **
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## exp(coef) exp(-coef) lower .95 upper .95
## age 1.0619 0.94170 1.0502 1.0737
## gender2 0.7482 1.33647 0.6220 0.9001
## los 1.0112 0.98894 1.0048 1.0176
## vd1 1.2198 0.81977 0.9881 1.5060
## pneumonia1 1.4180 0.70524 1.0818 1.8586
## met_cancer1 12.2063 0.08193 5.9751 24.9356
## cog_imp1 1.3362 0.74838 1.0024 1.7813
## prior_dnas 1.1229 0.89056 1.0438 1.2080
##
## Concordance= 0.699 (se = 0.012 )
## Likelihood ratio test= 211.3 on 8 df, p=<2e-16
## Wald test = 206 on 8 df, p=<2e-16
## Score (logrank) test = 221.9 on 8 df, p=<2e-16Results:
In this survival analysis, factors affecting survival in Heart failure patients are: age (hazard: 1.06 per year, CI: 1.05 - 1.07, p-value: < 2e-16), gender-male as reference (hazard: 0.75 per year, CI: 0.62 - 0.9, p-value: 0.002), length of stay (los) (hazard: 1.01, CI: 1 - 1.02, p-value: 0.000), prior missed appointments (hazard: 1.12, CI: 1.04 - 1.2, p-value: 0.0019) and comorbidites: pneumonia (hazard: 0.35, CI: 1.08 - 1.86, p-value: 0.011), metastatic cancer (hazard: 12.21, CI: 5.98 - 24.94, p-value: 6.67e-12) and cognitive impairment (dementia and senility) (hazard: 1.34, CI: 1 - 1.78, p-value: 0.048).