##chargement des packages----
library(questionr)
library(tidyverse)
## ── Attaching core tidyverse packages ──────────────────────── tidyverse 2.0.0 ──
## ✔ dplyr 1.1.1 ✔ readr 2.1.4
## ✔ forcats 1.0.0 ✔ stringr 1.5.0
## ✔ ggplot2 3.4.2 ✔ tibble 3.2.1
## ✔ lubridate 1.9.2 ✔ tidyr 1.3.0
## ✔ purrr 1.0.1
## ── Conflicts ────────────────────────────────────────── tidyverse_conflicts() ──
## ✖ dplyr::filter() masks stats::filter()
## ✖ dplyr::lag() masks stats::lag()
## ℹ Use the conflicted package (<http://conflicted.r-lib.org/>) to force all conflicts to become errors
library(tableone)
library(labelled)
library(gtsummary)
library(GGally)
## Registered S3 method overwritten by 'GGally':
## method from
## +.gg ggplot2
library(readxl)
library(effects)
## Le chargement a nécessité le package : carData
## lattice theme set by effectsTheme()
## See ?effectsTheme for details.
library(survival)
library(survminer)
## Le chargement a nécessité le package : ggpubr
##
## Attachement du package : 'survminer'
##
## L'objet suivant est masqué depuis 'package:survival':
##
## myeloma
library(ggplot2)
library(dplyr)
library(knitr)
library(cowplot)
##
## Attachement du package : 'cowplot'
##
## L'objet suivant est masqué depuis 'package:ggpubr':
##
## get_legend
##
## L'objet suivant est masqué depuis 'package:lubridate':
##
## stamp
##chargement des données
data_guillaume_dermato_ddi <- read_excel("Y:/fp/FAC/2023-2024/Memoires DES/Guillaume Flandin/20240705/data_guillaume_dermato_ddi.xlsx")
## New names:
## • `Détail` -> `Détail...39`
## • `Détail` -> `Détail...46`
## • `Détail` -> `Détail...70`
## • `Détail` -> `Détail...72`
dermato_g<-filter(data_guillaume_dermato_ddi, c(eligible_etude=="1" ))
data_guillaume_cbnpc_ddi <- read_excel("Y:/fp/FAC/2023-2024/Memoires DES/Guillaume Flandin/20240705/data_guillaume_cbnpc_ddi.xlsx")
## New names:
## • `Détail` -> `Détail...43`
## • `Détail` -> `Détail...50`
## • `Détail` -> `Détail...73`
## • `Détail` -> `Détail...75`
cbnpc<-filter(data_guillaume_cbnpc_ddi, c(eligible_etude=="1" ))
dermato<-subset(dermato_g, select =c(IPP, pftox, evt_tox, charlson, comprehension, patient_seul, actif_pro, ddi, ddi_surdosage, ddi_sousdosage, ipp_mdt, rque_pharma, rque_pharma_ei, polymedique, age,pftox_mediane, origine))
pneumo<-subset(cbnpc, select =c(IPP, pftox, evt_tox, charlson, comprehension, patient_seul, actif_pro, ddi, ddi_surdosage, ddi_sousdosage, ipp_mdt, rque_pharma, rque_pharma_ei, polymedique, age,pftox_mediane, origine))
##regroupement en un seul tableau
datpoolees<-bind_rows(dermato, pneumo)
datpoolees
## # A tibble: 172 × 17
## IPP pftox evt_tox charlson comprehension patient_seul actif_pro ddi
## <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl>
## 1 198802639 10.1 1 4 0 1 0 0
## 2 199205455 0.933 0 3 0 0 0 0
## 3 199605403 3.9 0 3 0 0 0 1
## 4 201202193 0.5 0 1 0 0 0 0
## 5 201205512 0.467 1 3 0 0 0 0
## 6 201403781 4.37 1 1 0 1 1 0
## 7 201405565 4.77 0 1 0 0 0 0
## 8 201406799 4.67 0 0 1 0 0 1
## 9 201408775 36.2 0 0 0 0 1 1
## 10 201409786 5.83 1 6 0 0 0 0
## # ℹ 162 more rows
## # ℹ 9 more variables: ddi_surdosage <dbl>, ddi_sousdosage <dbl>, ipp_mdt <dbl>,
## # rque_pharma <dbl>, rque_pharma_ei <dbl>, polymedique <dbl>, age <dbl>,
## # pftox_mediane <dbl>, origine <chr>
##recodage de données le cas échéant----
datpoolees$charlson_2<-ifelse(datpoolees$charlson>=2, 1, 0)
datpoolees$charlson_3<-ifelse(datpoolees$charlson>=3, 1, 0)
datpoolees$charlson_4<-ifelse(datpoolees$charlson>=4, 1, 0)
datpoolees$charlson_5<-ifelse(datpoolees$charlson>=5, 1, 0)
datpoolees$tox3mois<-ifelse(datpoolees$pftox<3 & datpoolees$evt_tox==1, 1, 0)
datpoolees$old75<-ifelse(datpoolees$age>=75, 1, 0)
datpoolees$origine.cat<-ifelse(datpoolees$origine =="dermato", 1, 0)
##modèles de cox ANALYSE UNIVARIEE ----
####age en valeur continue
modsurv<-coxph(Surv(pftox_mediane, evt_tox)~age, data=datpoolees)
modsurv%>%tbl_regression(exponentiate = TRUE)
| Characteristic |
HR |
95% CI |
p-value |
| age |
1.01 |
1.00, 1.03 |
0.14 |
####patient >=75 ans
modsurv<-coxph(Surv(pftox_mediane, evt_tox)~old75, data=datpoolees)
modsurv%>%tbl_regression(exponentiate = TRUE)
| Characteristic |
HR |
95% CI |
p-value |
| old75 |
1.24 |
0.79, 1.95 |
0.4 |
####charlson en valeur continue
modsurv<-coxph(Surv(pftox_mediane, evt_tox)~charlson, data=datpoolees)
modsurv%>%tbl_regression(exponentiate = TRUE)
| Characteristic |
HR |
95% CI |
p-value |
| charlson |
1.14 |
1.02, 1.27 |
0.019 |
####charlson >=2---
modsurv<-coxph(Surv(pftox_mediane, evt_tox)~charlson_2, data=datpoolees)
modsurv%>%tbl_regression(exponentiate = TRUE)
| Characteristic |
HR |
95% CI |
p-value |
| charlson_2 |
1.38 |
0.88, 2.18 |
0.2 |
####charlson >=3---
modsurv<-coxph(Surv(pftox_mediane, evt_tox)~charlson_3, data=datpoolees)
modsurv%>%tbl_regression(exponentiate = TRUE)
| Characteristic |
HR |
95% CI |
p-value |
| charlson_3 |
1.45 |
0.95, 2.20 |
0.081 |
####charlson >=4---
modsurv<-coxph(Surv(pftox_mediane, evt_tox)~charlson_4, data=datpoolees)
modsurv%>%tbl_regression(exponentiate = TRUE)
| Characteristic |
HR |
95% CI |
p-value |
| charlson_4 |
1.68 |
1.11, 2.55 |
0.015 |
####charlson >=5---
modsurv<-coxph(Surv(pftox_mediane, evt_tox)~charlson_5, data=datpoolees)
modsurv%>%tbl_regression(exponentiate = TRUE)
| Characteristic |
HR |
95% CI |
p-value |
| charlson_5 |
1.89 |
1.17, 3.07 |
0.010 |
####problème de compréhension
modsurv<-coxph(Surv(pftox_mediane, evt_tox)~comprehension, data=datpoolees)
modsurv%>%tbl_regression(exponentiate = TRUE)
| Characteristic |
HR |
95% CI |
p-value |
| comprehension |
2.10 |
1.11, 3.98 |
0.023 |
####patient vivant seul
modsurv<-coxph(Surv(pftox_mediane, evt_tox)~patient_seul, data=datpoolees)
modsurv%>%tbl_regression(exponentiate = TRUE)
| Characteristic |
HR |
95% CI |
p-value |
| patient_seul |
1.38 |
0.79, 2.40 |
0.3 |
####activité professionnelle
modsurv<-coxph(Surv(pftox_mediane, evt_tox)~actif_pro, data=datpoolees)
modsurv%>%tbl_regression(exponentiate = TRUE)
| Characteristic |
HR |
95% CI |
p-value |
| actif_pro |
0.64 |
0.40, 1.04 |
0.072 |
####interaction médicamenteuse
modsurv<-coxph(Surv(pftox_mediane, evt_tox)~ddi, data=datpoolees)
modsurv%>%tbl_regression(exponentiate = TRUE)
| Characteristic |
HR |
95% CI |
p-value |
| ddi |
0.89 |
0.59, 1.35 |
0.6 |
####interaction médicamenteuse à risque de surdosage
modsurv<-coxph(Surv(pftox_mediane, evt_tox)~ddi_surdosage, data=datpoolees)
modsurv%>%tbl_regression(exponentiate = TRUE)
| Characteristic |
HR |
95% CI |
p-value |
| ddi_surdosage |
0.79 |
0.51, 1.24 |
0.3 |
####prise d'ipp
modsurv<-coxph(Surv(pftox_mediane, evt_tox)~ipp_mdt, data=datpoolees)
modsurv%>%tbl_regression(exponentiate = TRUE)
| Characteristic |
HR |
95% CI |
p-value |
| ipp_mdt |
1.26 |
0.78, 2.04 |
0.3 |
####conclusion necessité suivi
modsurv<-coxph(Surv(pftox_mediane, evt_tox)~rque_pharma_ei, data=datpoolees)
modsurv%>%tbl_regression(exponentiate = TRUE)
| Characteristic |
HR |
95% CI |
p-value |
| rque_pharma_ei |
0.73 |
0.47, 1.13 |
0.2 |
####patient polymédiqué
modsurv<-coxph(Surv(pftox_mediane, evt_tox)~polymedique, data=datpoolees)
modsurv%>%tbl_regression(exponentiate = TRUE)
| Characteristic |
HR |
95% CI |
p-value |
| polymedique |
1.32 |
0.87, 1.99 |
0.2 |
####pathologie (1 = dermato)
modsurv<-coxph(Surv(pftox_mediane, evt_tox)~origine.cat, data=datpoolees)
modsurv%>%tbl_regression(exponentiate = TRUE)
| Characteristic |
HR |
95% CI |
p-value |
| origine.cat |
2.61 |
1.68, 4.04 |
<0.001 |
##Analyse multivariée avec charlson >=2
modsurv<-coxph(Surv(pftox, evt_tox)~charlson_2+age+comprehension+actif_pro+polymedique+rque_pharma_ei+origine.cat, data=datpoolees)
modsurv%>%tbl_regression(exponentiate = TRUE)
| Characteristic |
HR |
95% CI |
p-value |
| charlson_2 |
1.42 |
0.67, 3.04 |
0.4 |
| age |
1.00 |
0.98, 1.03 |
0.7 |
| comprehension |
1.76 |
0.91, 3.40 |
0.092 |
| actif_pro |
0.95 |
0.51, 1.76 |
0.9 |
| polymedique |
1.34 |
0.83, 2.17 |
0.2 |
| rque_pharma_ei |
0.73 |
0.43, 1.22 |
0.2 |
| origine.cat |
3.69 |
2.28, 5.96 |
<0.001 |
step(modsurv)
## Start: AIC=812.07
## Surv(pftox, evt_tox) ~ charlson_2 + age + comprehension + actif_pro +
## polymedique + rque_pharma_ei + origine.cat
##
## Df AIC
## - actif_pro 1 810.10
## - age 1 810.23
## - charlson_2 1 810.90
## - polymedique 1 811.51
## - rque_pharma_ei 1 811.54
## <none> 812.07
## - comprehension 1 812.59
## - origine.cat 1 840.21
##
## Step: AIC=810.1
## Surv(pftox, evt_tox) ~ charlson_2 + age + comprehension + polymedique +
## rque_pharma_ei + origine.cat
##
## Df AIC
## - age 1 808.31
## - charlson_2 1 809.07
## - rque_pharma_ei 1 809.58
## - polymedique 1 809.60
## <none> 810.10
## - comprehension 1 810.60
## - origine.cat 1 839.07
##
## Step: AIC=808.31
## Surv(pftox, evt_tox) ~ charlson_2 + comprehension + polymedique +
## rque_pharma_ei + origine.cat
##
## Df AIC
## - rque_pharma_ei 1 807.76
## - polymedique 1 807.94
## <none> 808.31
## - comprehension 1 808.65
## - charlson_2 1 810.95
## - origine.cat 1 838.25
##
## Step: AIC=807.76
## Surv(pftox, evt_tox) ~ charlson_2 + comprehension + polymedique +
## origine.cat
##
## Df AIC
## - polymedique 1 806.38
## <none> 807.76
## - comprehension 1 808.07
## - charlson_2 1 810.37
## - origine.cat 1 846.47
##
## Step: AIC=806.38
## Surv(pftox, evt_tox) ~ charlson_2 + comprehension + origine.cat
##
## Df AIC
## <none> 806.38
## - comprehension 1 806.61
## - charlson_2 1 809.68
## - origine.cat 1 845.70
## Call:
## coxph(formula = Surv(pftox, evt_tox) ~ charlson_2 + comprehension +
## origine.cat, data = datpoolees)
##
## coef exp(coef) se(coef) z p
## charlson_2 0.5302 1.6993 0.2374 2.233 0.0255
## comprehension 0.5227 1.6865 0.3289 1.589 0.1121
## origine.cat 1.4206 4.1398 0.2300 6.175 6.6e-10
##
## Likelihood ratio test=45.89 on 3 df, p=5.99e-10
## n= 172, number of events= 93
##modèle final avec Charlson >=2 (covariables retenues sur le critère Aikake)
modsurv<-coxph(Surv(pftox, evt_tox)~charlson_2+comprehension+origine.cat, data=datpoolees)
modsurv%>%tbl_regression(exponentiate = TRUE)
| Characteristic |
HR |
95% CI |
p-value |
| charlson_2 |
1.70 |
1.07, 2.71 |
0.026 |
| comprehension |
1.69 |
0.89, 3.21 |
0.11 |
| origine.cat |
4.14 |
2.64, 6.50 |
<0.001 |
ggforest(modsurv)
## Warning in .get_data(model, data = data): The `data` argument is not provided.
## Data will be extracted from model fit.
###Modèles alternatifs à 2 covariables
modsurv<-coxph(Surv(pftox, evt_tox)~charlson_2+origine.cat, data=datpoolees)
modsurv%>%tbl_regression(exponentiate = TRUE)
| Characteristic |
HR |
95% CI |
p-value |
| charlson_2 |
1.76 |
1.11, 2.80 |
0.016 |
| origine.cat |
4.23 |
2.70, 6.64 |
<0.001 |
ggforest(modsurv)
## Warning in .get_data(model, data = data): The `data` argument is not provided.
## Data will be extracted from model fit.
##Analyse multivariée avec charlson >=3
modsurv<-coxph(Surv(pftox, evt_tox)~charlson_3+age+comprehension+actif_pro+polymedique+rque_pharma_ei+origine.cat, data=datpoolees)
modsurv%>%tbl_regression(exponentiate = TRUE)
| Characteristic |
HR |
95% CI |
p-value |
| charlson_3 |
1.60 |
0.83, 3.09 |
0.2 |
| age |
1.00 |
0.98, 1.02 |
0.9 |
| comprehension |
1.73 |
0.90, 3.31 |
0.10 |
| actif_pro |
0.96 |
0.52, 1.77 |
0.9 |
| polymedique |
1.30 |
0.80, 2.11 |
0.3 |
| rque_pharma_ei |
0.72 |
0.43, 1.22 |
0.2 |
| origine.cat |
3.80 |
2.35, 6.15 |
<0.001 |
step(modsurv)
## Start: AIC=810.89
## Surv(pftox, evt_tox) ~ charlson_3 + age + comprehension + actif_pro +
## polymedique + rque_pharma_ei + origine.cat
##
## Df AIC
## - actif_pro 1 808.91
## - age 1 808.93
## - polymedique 1 810.04
## - rque_pharma_ei 1 810.43
## <none> 810.89
## - charlson_3 1 810.90
## - comprehension 1 811.27
## - origine.cat 1 840.21
##
## Step: AIC=808.91
## Surv(pftox, evt_tox) ~ charlson_3 + age + comprehension + polymedique +
## rque_pharma_ei + origine.cat
##
## Df AIC
## - age 1 806.97
## - polymedique 1 808.09
## - rque_pharma_ei 1 808.47
## <none> 808.91
## - charlson_3 1 809.07
## - comprehension 1 809.30
## - origine.cat 1 838.94
##
## Step: AIC=806.97
## Surv(pftox, evt_tox) ~ charlson_3 + comprehension + polymedique +
## rque_pharma_ei + origine.cat
##
## Df AIC
## - polymedique 1 806.18
## - rque_pharma_ei 1 806.52
## <none> 806.97
## - comprehension 1 807.30
## - charlson_3 1 810.95
## - origine.cat 1 837.57
##
## Step: AIC=806.18
## Surv(pftox, evt_tox) ~ charlson_3 + comprehension + rque_pharma_ei +
## origine.cat
##
## Df AIC
## - rque_pharma_ei 1 804.83
## <none> 806.18
## - comprehension 1 806.47
## - charlson_3 1 811.43
## - origine.cat 1 841.96
##
## Step: AIC=804.83
## Surv(pftox, evt_tox) ~ charlson_3 + comprehension + origine.cat
##
## Df AIC
## <none> 804.83
## - comprehension 1 805.10
## - charlson_3 1 809.68
## - origine.cat 1 845.03
## Call:
## coxph(formula = Surv(pftox, evt_tox) ~ charlson_3 + comprehension +
## origine.cat, data = datpoolees)
##
## coef exp(coef) se(coef) z p
## charlson_3 0.5643 1.7583 0.2178 2.591 0.00957
## comprehension 0.5257 1.6917 0.3284 1.601 0.10937
## origine.cat 1.4419 4.2286 0.2316 6.225 4.83e-10
##
## Likelihood ratio test=47.43 on 3 df, p=2.814e-10
## n= 172, number of events= 93
###modèle final (covariables retenues sur le critère Aikake)
modsurv<-coxph(Surv(pftox, evt_tox)~charlson_3+origine.cat+comprehension, data=datpoolees)
modsurv%>%tbl_regression(exponentiate = TRUE)
| Characteristic |
HR |
95% CI |
p-value |
| charlson_3 |
1.76 |
1.15, 2.69 |
0.010 |
| origine.cat |
4.23 |
2.69, 6.66 |
<0.001 |
| comprehension |
1.69 |
0.89, 3.22 |
0.11 |
ggforest(modsurv)
## Warning in .get_data(model, data = data): The `data` argument is not provided.
## Data will be extracted from model fit.
###modèles alternatifs à 2 covariables
modsurv<-coxph(Surv(pftox, evt_tox)~charlson_3+origine.cat, data=datpoolees)
modsurv%>%tbl_regression(exponentiate = TRUE)
| Characteristic |
HR |
95% CI |
p-value |
| charlson_3 |
1.81 |
1.19, 2.77 |
0.006 |
| origine.cat |
4.30 |
2.73, 6.76 |
<0.001 |
ggforest(modsurv)
## Warning in .get_data(model, data = data): The `data` argument is not provided.
## Data will be extracted from model fit.
#ANALYSE SUR CRITERE DE TOXICITE PRECOCE
##Analyses univariées
##analyses univariée avec tableau de résultat, ex tox selon aucmoy en valeur continue
datpoolees |>
tbl_uvregression(
y = tox3mois,
include = c(age, old75, charlson, charlson_2, charlson_3, charlson_4, charlson_5, ddi, ddi_surdosage, rque_pharma_ei, polymedique, patient_seul, actif_pro, comprehension, origine.cat),
method = glm,
method.args = list(family = binomial),
exponentiate = TRUE
) |>
bold_labels()
| Characteristic |
N |
OR |
95% CI |
p-value |
| charlson |
172 |
1.18 |
0.98, 1.42 |
0.075 |
| comprehension |
172 |
1.77 |
0.56, 5.20 |
0.3 |
| patient_seul |
172 |
1.30 |
0.46, 3.35 |
0.6 |
| actif_pro |
172 |
0.54 |
0.23, 1.20 |
0.15 |
| ddi |
172 |
0.91 |
0.47, 1.80 |
0.8 |
| ddi_surdosage |
172 |
0.74 |
0.36, 1.48 |
0.4 |
| rque_pharma_ei |
172 |
1.12 |
0.56, 2.19 |
0.8 |
| polymedique |
172 |
1.49 |
0.77, 2.94 |
0.2 |
| age |
172 |
1.03 |
1.00, 1.05 |
0.038 |
| charlson_2 |
172 |
2.06 |
0.94, 4.91 |
0.082 |
| charlson_3 |
172 |
1.85 |
0.94, 3.74 |
0.078 |
| charlson_4 |
172 |
1.67 |
0.83, 3.32 |
0.15 |
| charlson_5 |
172 |
1.89 |
0.82, 4.27 |
0.13 |
| old75 |
172 |
1.43 |
0.67, 2.96 |
0.3 |
| origine.cat |
172 |
1.95 |
1.00, 3.84 |
0.052 |
##courbe ROC sur score de charlson
library(pROC)
## Type 'citation("pROC")' for a citation.
##
## Attachement du package : 'pROC'
##
## Les objets suivants sont masqués depuis 'package:stats':
##
## cov, smooth, var
roc1<-roc(tox3mois~charlson, datpoolees)
## Setting levels: control = 0, case = 1
## Setting direction: controls < cases
coords(roc1, "best", ret=c("threshold", "specificity", "1-npv"))
## threshold specificity 1-npv
## threshold 2.5 0.495935 0.2179487
##courbe ROC sur age
library(pROC)
roc1<-roc(tox3mois~age, datpoolees)
## Setting levels: control = 0, case = 1
## Setting direction: controls < cases
coords(roc1, "best", ret=c("threshold", "specificity", "1-npv"))
## threshold specificity 1-npv
## threshold 58.5 0.4065041 0.137931
##modele multivarié
mod<-glm( tox3mois ~ charlson_2+age+origine.cat+polymedique+actif_pro, data=datpoolees, family="binomial")
summary(mod)
##
## Call:
## glm(formula = tox3mois ~ charlson_2 + age + origine.cat + polymedique +
## actif_pro, family = "binomial", data = datpoolees)
##
## Deviance Residuals:
## Min 1Q Median 3Q Max
## -1.2086 -0.7954 -0.7065 1.1951 2.0163
##
## Coefficients:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) -2.81208 1.13504 -2.478 0.0132 *
## charlson_2 0.60935 0.63873 0.954 0.3401
## age 0.01387 0.01946 0.713 0.4760
## origine.cat 0.81485 0.37180 2.192 0.0284 *
## polymedique 0.24002 0.35672 0.673 0.5010
## actif_pro 0.01379 0.50320 0.027 0.9781
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for binomial family taken to be 1)
##
## Null deviance: 205.54 on 171 degrees of freedom
## Residual deviance: 194.95 on 166 degrees of freedom
## AIC: 206.95
##
## Number of Fisher Scoring iterations: 4
exp(coefficients(mod))
## (Intercept) charlson_2 age origine.cat polymedique actif_pro
## 0.06007987 1.83923731 1.01397041 2.25883286 1.27127366 1.01388580
exp(confint(mod, level=0.95))
## Attente de la réalisation du profilage...
## 2.5 % 97.5 %
## (Intercept) 0.005889147 0.518380
## charlson_2 0.531724119 6.606512
## age 0.976592293 1.054458
## origine.cat 1.098861194 4.745839
## polymedique 0.631082534 2.570193
## actif_pro 0.368158623 2.696288
mod%>%tbl_regression(intercept = TRUE, exponentiate = TRUE)
| Characteristic |
OR |
95% CI |
p-value |
| (Intercept) |
0.06 |
0.01, 0.52 |
0.013 |
| charlson_2 |
1.84 |
0.53, 6.61 |
0.3 |
| age |
1.01 |
0.98, 1.05 |
0.5 |
| origine.cat |
2.26 |
1.10, 4.75 |
0.028 |
| polymedique |
1.27 |
0.63, 2.57 |
0.5 |
| actif_pro |
1.01 |
0.37, 2.70 |
>0.9 |
step(mod)
## Start: AIC=206.95
## tox3mois ~ charlson_2 + age + origine.cat + polymedique + actif_pro
##
## Df Deviance AIC
## - actif_pro 1 194.95 204.95
## - polymedique 1 195.40 205.40
## - age 1 195.47 205.47
## - charlson_2 1 195.87 205.87
## <none> 194.95 206.95
## - origine.cat 1 199.87 209.87
##
## Step: AIC=204.95
## tox3mois ~ charlson_2 + age + origine.cat + polymedique
##
## Df Deviance AIC
## - polymedique 1 195.40 203.40
## - age 1 195.50 203.50
## - charlson_2 1 195.89 203.89
## <none> 194.95 204.95
## - origine.cat 1 200.03 208.03
##
## Step: AIC=203.4
## tox3mois ~ charlson_2 + age + origine.cat
##
## Df Deviance AIC
## - age 1 196.12 202.12
## - charlson_2 1 196.32 202.32
## <none> 195.40 203.40
## - origine.cat 1 200.81 206.81
##
## Step: AIC=202.12
## tox3mois ~ charlson_2 + origine.cat
##
## Df Deviance AIC
## <none> 196.12 202.12
## - charlson_2 1 201.70 205.70
## - origine.cat 1 202.27 206.27
##
## Call: glm(formula = tox3mois ~ charlson_2 + origine.cat, family = "binomial",
## data = datpoolees)
##
## Coefficients:
## (Intercept) charlson_2 origine.cat
## -2.0906 0.9836 0.8822
##
## Degrees of Freedom: 171 Total (i.e. Null); 169 Residual
## Null Deviance: 205.5
## Residual Deviance: 196.1 AIC: 202.1
##modele final sur critere aikake
mod<-glm( tox3mois ~ charlson_2+origine.cat, data=datpoolees, family="binomial")
summary(mod)
##
## Call:
## glm(formula = tox3mois ~ charlson_2 + origine.cat, family = "binomial",
## data = datpoolees)
##
## Deviance Residuals:
## Min 1Q Median 3Q Max
## -1.0836 -0.7558 -0.7230 1.2742 2.1010
##
## Coefficients:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) -2.0906 0.4650 -4.495 6.94e-06 ***
## charlson_2 0.9836 0.4378 2.247 0.0246 *
## origine.cat 0.8822 0.3604 2.448 0.0144 *
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for binomial family taken to be 1)
##
## Null deviance: 205.54 on 171 degrees of freedom
## Residual deviance: 196.12 on 169 degrees of freedom
## AIC: 202.12
##
## Number of Fisher Scoring iterations: 4
exp(coefficients(mod))
## (Intercept) charlson_2 origine.cat
## 0.1236161 2.6741034 2.4162412
exp(confint(mod, level=0.95))
## Attente de la réalisation du profilage...
## 2.5 % 97.5 %
## (Intercept) 0.04683633 0.292799
## charlson_2 1.17572226 6.631286
## origine.cat 1.20180150 4.961262
mod%>%tbl_regression(intercept = TRUE, exponentiate = TRUE)
| Characteristic |
OR |
95% CI |
p-value |
| (Intercept) |
0.12 |
0.05, 0.29 |
<0.001 |
| charlson_2 |
2.67 |
1.18, 6.63 |
0.025 |
| origine.cat |
2.42 |
1.20, 4.96 |
0.014 |
##Analyse multivariée avec Charlson à 3
mod<-glm( tox3mois ~ charlson_3+age+origine.cat+polymedique+actif_pro, data=datpoolees, family="binomial")
summary(mod)
##
## Call:
## glm(formula = tox3mois ~ charlson_3 + age + origine.cat + polymedique +
## actif_pro, family = "binomial", data = datpoolees)
##
## Deviance Residuals:
## Min 1Q Median 3Q Max
## -1.2333 -0.8157 -0.6859 1.1914 1.9509
##
## Coefficients:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) -2.84186 1.18548 -2.397 0.0165 *
## charlson_3 0.27109 0.52865 0.513 0.6081
## age 0.01981 0.01908 1.038 0.2991
## origine.cat 0.74328 0.35931 2.069 0.0386 *
## polymedique 0.21467 0.35696 0.601 0.5476
## actif_pro -0.02888 0.50068 -0.058 0.9540
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for binomial family taken to be 1)
##
## Null deviance: 205.54 on 171 degrees of freedom
## Residual deviance: 195.61 on 166 degrees of freedom
## AIC: 207.61
##
## Number of Fisher Scoring iterations: 4
exp(coefficients(mod))
## (Intercept) charlson_3 age origine.cat polymedique actif_pro
## 0.05831732 1.31139668 1.02000878 2.10281828 1.23945518 0.97153060
exp(confint(mod, level=0.95))
## Attente de la réalisation du profilage...
## 2.5 % 97.5 %
## (Intercept) 0.005058668 0.5449513
## charlson_3 0.464410561 3.7262911
## age 0.983705025 1.0606028
## origine.cat 1.046224062 4.3027642
## polymedique 0.614614786 2.5055763
## actif_pro 0.354177482 2.5678190
mod%>%tbl_regression(intercept = TRUE, exponentiate = TRUE)
| Characteristic |
OR |
95% CI |
p-value |
| (Intercept) |
0.06 |
0.01, 0.54 |
0.017 |
| charlson_3 |
1.31 |
0.46, 3.73 |
0.6 |
| age |
1.02 |
0.98, 1.06 |
0.3 |
| origine.cat |
2.10 |
1.05, 4.30 |
0.039 |
| polymedique |
1.24 |
0.61, 2.51 |
0.5 |
| actif_pro |
0.97 |
0.35, 2.57 |
>0.9 |
step(mod)
## Start: AIC=207.61
## tox3mois ~ charlson_3 + age + origine.cat + polymedique + actif_pro
##
## Df Deviance AIC
## - actif_pro 1 195.61 205.61
## - charlson_3 1 195.87 205.87
## - polymedique 1 195.97 205.97
## - age 1 196.73 206.73
## <none> 195.61 207.61
## - origine.cat 1 199.97 209.97
##
## Step: AIC=205.61
## tox3mois ~ charlson_3 + age + origine.cat + polymedique
##
## Df Deviance AIC
## - charlson_3 1 195.89 203.89
## - polymedique 1 195.98 203.98
## - age 1 196.88 204.88
## <none> 195.61 205.61
## - origine.cat 1 200.13 208.13
##
## Step: AIC=203.89
## tox3mois ~ age + origine.cat + polymedique
##
## Df Deviance AIC
## - polymedique 1 196.32 202.32
## <none> 195.89 203.89
## - origine.cat 1 200.21 206.21
## - age 1 200.58 206.58
##
## Step: AIC=202.32
## tox3mois ~ age + origine.cat
##
## Df Deviance AIC
## <none> 196.32 202.32
## - origine.cat 1 200.95 204.95
## - age 1 201.70 205.70
##
## Call: glm(formula = tox3mois ~ age + origine.cat, family = "binomial",
## data = datpoolees)
##
## Coefficients:
## (Intercept) age origine.cat
## -3.15560 0.02856 0.74478
##
## Degrees of Freedom: 171 Total (i.e. Null); 169 Residual
## Null Deviance: 205.5
## Residual Deviance: 196.3 AIC: 202.3
##modele final sur critere aikake
mod<-glm( tox3mois ~ age+origine.cat, data=datpoolees, family="binomial")
summary(mod)
##
## Call:
## glm(formula = tox3mois ~ age + origine.cat, family = "binomial",
## data = datpoolees)
##
## Deviance Residuals:
## Min 1Q Median 3Q Max
## -1.2461 -0.8266 -0.6818 1.2343 1.9696
##
## Coefficients:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) -3.15560 0.90146 -3.501 0.000464 ***
## age 0.02856 0.01275 2.240 0.025076 *
## origine.cat 0.74478 0.34926 2.132 0.032969 *
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for binomial family taken to be 1)
##
## Null deviance: 205.54 on 171 degrees of freedom
## Residual deviance: 196.32 on 169 degrees of freedom
## AIC: 202.32
##
## Number of Fisher Scoring iterations: 4
exp(coefficients(mod))
## (Intercept) age origine.cat
## 0.04261302 1.02897629 2.10598801
exp(confint(mod, level=0.95))
## Attente de la réalisation du profilage...
## 2.5 % 97.5 %
## (Intercept) 0.006654776 0.2324878
## age 1.004329568 1.0561046
## origine.cat 1.068047373 4.2207938
mod%>%tbl_regression(intercept = TRUE, exponentiate = TRUE)
| Characteristic |
OR |
95% CI |
p-value |
| (Intercept) |
0.04 |
0.01, 0.23 |
<0.001 |
| age |
1.03 |
1.00, 1.06 |
0.025 |
| origine.cat |
2.11 |
1.07, 4.22 |
0.033 |
##Analyse multivariée avec Charlson à 3 et age binaire
datpoolees$old58<-ifelse(datpoolees$age>58, 1, 0)
mod<-glm( tox3mois ~ charlson_3+old58+origine.cat+polymedique+actif_pro, data=datpoolees, family="binomial")
summary(mod)
##
## Call:
## glm(formula = tox3mois ~ charlson_3 + old58 + origine.cat + polymedique +
## actif_pro, family = "binomial", data = datpoolees)
##
## Deviance Residuals:
## Min 1Q Median 3Q Max
## -1.2239 -0.8138 -0.6609 1.1939 2.2757
##
## Coefficients:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) -2.7568 0.6338 -4.350 1.36e-05 ***
## charlson_3 -0.1476 0.4781 -0.309 0.7576
## old58 1.7240 0.6287 2.742 0.0061 **
## origine.cat 0.8962 0.3740 2.397 0.0165 *
## polymedique 0.2453 0.3650 0.672 0.5016
## actif_pro 0.4503 0.5415 0.832 0.4057
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for binomial family taken to be 1)
##
## Null deviance: 205.54 on 171 degrees of freedom
## Residual deviance: 188.49 on 166 degrees of freedom
## AIC: 200.49
##
## Number of Fisher Scoring iterations: 4
exp(coefficients(mod))
## (Intercept) charlson_3 old58 origine.cat polymedique actif_pro
## 0.0634974 0.8628127 5.6068212 2.4503550 1.2780407 1.5687351
exp(confint(mod, level=0.95))
## Attente de la réalisation du profilage...
## 2.5 % 97.5 %
## (Intercept) 0.01681847 0.2044697
## charlson_3 0.33881409 2.2377592
## old58 1.70517472 20.4132433
## origine.cat 1.18826952 5.1762253
## polymedique 0.62444819 2.6287801
## actif_pro 0.53731426 4.6003191
mod%>%tbl_regression(intercept = TRUE, exponentiate = TRUE)
| Characteristic |
OR |
95% CI |
p-value |
| (Intercept) |
0.06 |
0.02, 0.20 |
<0.001 |
| charlson_3 |
0.86 |
0.34, 2.24 |
0.8 |
| old58 |
5.61 |
1.71, 20.4 |
0.006 |
| origine.cat |
2.45 |
1.19, 5.18 |
0.017 |
| polymedique |
1.28 |
0.62, 2.63 |
0.5 |
| actif_pro |
1.57 |
0.54, 4.60 |
0.4 |
step(mod)
## Start: AIC=200.49
## tox3mois ~ charlson_3 + old58 + origine.cat + polymedique + actif_pro
##
## Df Deviance AIC
## - charlson_3 1 188.58 198.58
## - polymedique 1 188.94 198.94
## - actif_pro 1 189.18 199.18
## <none> 188.49 200.49
## - origine.cat 1 194.40 204.40
## - old58 1 196.73 206.73
##
## Step: AIC=198.58
## tox3mois ~ old58 + origine.cat + polymedique + actif_pro
##
## Df Deviance AIC
## - polymedique 1 188.98 196.98
## - actif_pro 1 189.32 197.32
## <none> 188.58 198.58
## - origine.cat 1 194.67 202.67
## - old58 1 199.15 207.15
##
## Step: AIC=196.98
## tox3mois ~ old58 + origine.cat + actif_pro
##
## Df Deviance AIC
## - actif_pro 1 189.69 195.69
## <none> 188.98 196.98
## - origine.cat 1 195.38 201.38
## - old58 1 199.89 205.89
##
## Step: AIC=195.69
## tox3mois ~ old58 + origine.cat
##
## Df Deviance AIC
## <none> 189.69 195.69
## - origine.cat 1 195.47 199.47
## - old58 1 201.70 205.70
##
## Call: glm(formula = tox3mois ~ old58 + origine.cat, family = "binomial",
## data = datpoolees)
##
## Coefficients:
## (Intercept) old58 origine.cat
## -2.3534 1.4012 0.8536
##
## Degrees of Freedom: 171 Total (i.e. Null); 169 Residual
## Null Deviance: 205.5
## Residual Deviance: 189.7 AIC: 195.7
##modele final sur critere aikake
mod<-glm( tox3mois ~ old58+origine.cat, data=datpoolees, family="binomial")
summary(mod)
##
## Call:
## glm(formula = tox3mois ~ old58 + origine.cat, family = "binomial",
## data = datpoolees)
##
## Deviance Residuals:
## Min 1Q Median 3Q Max
## -1.1359 -0.8079 -0.6348 1.2195 2.2110
##
## Coefficients:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) -2.3534 0.4564 -5.156 2.52e-07 ***
## old58 1.4012 0.4401 3.184 0.00145 **
## origine.cat 0.8536 0.3597 2.373 0.01764 *
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for binomial family taken to be 1)
##
## Null deviance: 205.54 on 171 degrees of freedom
## Residual deviance: 189.69 on 169 degrees of freedom
## AIC: 195.69
##
## Number of Fisher Scoring iterations: 4
exp(coefficients(mod))
## (Intercept) old58 origine.cat
## 0.09504835 4.06018406 2.34818049
exp(confint(mod, level=0.95))
## Attente de la réalisation du profilage...
## 2.5 % 97.5 %
## (Intercept) 0.03613699 0.2190187
## old58 1.79214277 10.2389908
## origine.cat 1.16953422 4.8154942
mod%>%tbl_regression(intercept = TRUE, exponentiate = TRUE)
| Characteristic |
OR |
95% CI |
p-value |
| (Intercept) |
0.10 |
0.04, 0.22 |
<0.001 |
| old58 |
4.06 |
1.79, 10.2 |
0.001 |
| origine.cat |
2.35 |
1.17, 4.82 |
0.018 |