library(tidyverse)
Registered S3 method overwritten by 'dplyr':
  method           from
  print.rowwise_df     
── Attaching packages ─── tidyverse 1.2.1 ──
✔ ggplot2 3.2.1     ✔ purrr   0.3.2
✔ tibble  2.1.3     ✔ dplyr   0.8.3
✔ tidyr   1.0.0     ✔ stringr 1.4.0
✔ readr   1.3.1     ✔ forcats 0.4.0
── Conflicts ────── tidyverse_conflicts() ──
✖ dplyr::filter() masks stats::filter()
✖ dplyr::lag()    masks stats::lag()
library(tableone) #create 'Table 1' to describe baseline characteristics
library(Matching) #multivariate and propensity score matching with balance optimization
Le chargement a nécessité le package : MASS

Attachement du package : ‘MASS’

The following object is masked from ‘package:dplyr’:

    select

## 
##  Matching (Version 4.9-6, Build Date: 2019-04-07)
##  See http://sekhon.berkeley.edu/matching for additional documentation.
##  Please cite software as:
##   Jasjeet S. Sekhon. 2011. ``Multivariate and Propensity Score Matching
##   Software with Automated Balance Optimization: The Matching package for R.''
##   Journal of Statistical Software, 42(7): 1-52. 
##

Dataset : right heart catheterization

Une description complète est disponible à cette adresse http://biostat.mc.vanderbilt.edu/wiki/pub/Main/DataSets/rhc.html. Il s’agit de données sur des patients (2184 traités et 3551 contrôles) admis aux urgences dans 5 hôpitaux, la variable de traitement est swang1 (right heat catheterization vs. non) et l’outcome est death (yes or no). Nous allons considérer les variables de confusion suivantes : - cat1 : Primary disease category - age - sex`` -meanbp1`: Mean blood pressure

Matching (greedy)

Questions :

rhc = read_csv("./Data/rhc_data.csv")
  • Créer la “Table 1”. Attention le package tableone ne supporte pas les variables discrètes à plusieurs facteurs, il faut donc d’abord construire la table disjonctive.

Correction

rhc_small_disjunctive = as_tibble(model.matrix(~ . , data=rhc_small)[,-1])
glimpse(rhc_small_disjunctive)
Observations: 5,735
Variables: 13
$ treatmentRHC            <dbl> 0, 1, 1, 0, 1…
$ deathYes                <dbl> 0, 1, 0, 1, 1…
$ cat1CHF                 <dbl> 0, 0, 0, 0, 0…
$ cat1Cirrhosis           <dbl> 0, 0, 0, 0, 0…
$ `cat1Colon Cancer`      <dbl> 0, 0, 0, 0, 0…
$ cat1Coma                <dbl> 0, 0, 0, 0, 0…
$ cat1COPD                <dbl> 1, 0, 0, 0, 0…
$ `cat1Lung Cancer`       <dbl> 0, 0, 0, 0, 0…
$ `cat1MOSF w/Malignancy` <dbl> 0, 0, 1, 0, 0…
$ `cat1MOSF w/Sepsis`     <dbl> 0, 1, 0, 0, 1…
$ age                     <dbl> 70.25098, 78.…
$ sexMale                 <dbl> 1, 0, 0, 0, 1…
$ meanbp1                 <dbl> 41, 63, 57, 5…
names(rhc_small_disjunctive) = c("treatment","Death","CHF","Cirr",
                                 "colcan","Coma","COPD",
                                 "lungcan","MOSF","sepsis",
                                 "age","male","meanbp1")

Correction

xvars = c("CHF","Cirr","colcan","Coma","COPD",
         "lungcan","MOSF","sepsis",
         "age","male","meanbp1") # variables de confusion
table1 = CreateTableOne(vars=xvars,strata="treatment", data=rhc_small_disjunctive, test=FALSE)
print(table1,smd=TRUE)
                     Stratified by treatment
                      0             1            
  n                    3551          2184        
  CHF (mean (SD))      0.07 (0.25)   0.10 (0.29) 
  Cirr (mean (SD))     0.05 (0.22)   0.02 (0.15) 
  colcan (mean (SD))   0.00 (0.04)   0.00 (0.02) 
  Coma (mean (SD))     0.10 (0.29)   0.04 (0.20) 
  COPD (mean (SD))     0.11 (0.32)   0.03 (0.16) 
  lungcan (mean (SD))  0.01 (0.10)   0.00 (0.05) 
  MOSF (mean (SD))     0.07 (0.25)   0.07 (0.26) 
  sepsis (mean (SD))   0.15 (0.36)   0.32 (0.47) 
  age (mean (SD))     61.76 (17.29) 60.75 (15.63)
  male (mean (SD))     0.54 (0.50)   0.59 (0.49) 
  meanbp1 (mean (SD)) 84.87 (38.87) 68.20 (34.24)
                     Stratified by treatment
                      SMD   
  n                         
  CHF (mean (SD))      0.095
  Cirr (mean (SD))     0.145
  colcan (mean (SD))   0.038
  Coma (mean (SD))     0.207
  COPD (mean (SD))     0.342
  lungcan (mean (SD))  0.095
  MOSF (mean (SD))     0.018
  sepsis (mean (SD))   0.415
  age (mean (SD))      0.061
  male (mean (SD))     0.093
  meanbp1 (mean (SD))  0.455
  • Faire un matching “greedy” basé sur la distance de Mahalanobis.

Correction

greedymatch = Match(Tr=rhc_small_disjunctive$treatment,
                    M=1,
                    X=rhc_small_disjunctive[xvars],replace=FALSE)

rhc_small_disjunctive_matched = rhc_small_disjunctive %>%
                                  slice(c(greedymatch$index.control,
                                          greedymatch$index.treated)) %>%
                                  mutate( match_id = rep(c(1:2184),2))
                                

matchedtab1 = CreateTableOne(vars=xvars,strata="treatment", data=rhc_small_disjunctive_matched, test=FALSE)
print(matchedtab1, smd = TRUE)
                     Stratified by treatment
                      0             1            
  n                    2184          2184        
  CHF (mean (SD))      0.10 (0.29)   0.10 (0.29) 
  Cirr (mean (SD))     0.02 (0.15)   0.02 (0.15) 
  colcan (mean (SD))   0.00 (0.02)   0.00 (0.02) 
  Coma (mean (SD))     0.04 (0.20)   0.04 (0.20) 
  COPD (mean (SD))     0.03 (0.16)   0.03 (0.16) 
  lungcan (mean (SD))  0.00 (0.05)   0.00 (0.05) 
  MOSF (mean (SD))     0.07 (0.26)   0.07 (0.26) 
  sepsis (mean (SD))   0.24 (0.43)   0.32 (0.47) 
  age (mean (SD))     61.25 (16.38) 60.75 (15.63)
  male (mean (SD))     0.56 (0.50)   0.59 (0.49) 
  meanbp1 (mean (SD)) 72.93 (34.14) 68.20 (34.24)
                     Stratified by treatment
                      SMD   
  n                         
  CHF (mean (SD))     <0.001
  Cirr (mean (SD))    <0.001
  colcan (mean (SD))  <0.001
  Coma (mean (SD))    <0.001
  COPD (mean (SD))    <0.001
  lungcan (mean (SD)) <0.001
  MOSF (mean (SD))    <0.001
  sepsis (mean (SD))   0.177
  age (mean (SD))      0.031
  male (mean (SD))     0.042
  meanbp1 (mean (SD))  0.138
  • Sur les données “matchées” étudiez l’outcome

Correction

y_trt = rhc_small_disjunctive_matched %>% filter(treatment==1) %>% 
                                      dplyr::select(Death) %>% pull()
y_con = rhc_small_disjunctive_matched %>% filter(treatment==0) %>%
                                      dplyr::select(Death)%>% pull()
table(y_trt,y_con)
     y_con
y_trt   0   1
    0 303 395
    1 508 978
mcnemar.test(matrix(table(y_trt,y_con),2,2))

    McNemar's Chi-squared test with continuity
    correction

data:  matrix(table(y_trt, y_con), 2, 2)
McNemar's chi-squared = 13.891, df = 1, p-value
= 0.0001937

Propensity score matching

Correction

psmodel = glm(treatment ~ . - Death ,
              family=binomial(),
              data=rhc_small_disjunctive)

summary(psmodel)

Call:
glm(formula = treatment ~ . - Death, family = binomial(), data = rhc_small_disjunctive)

Deviance Residuals: 
    Min       1Q   Median       3Q      Max  
-1.7013  -1.0097  -0.6336   1.1814   2.4791  

Coefficients:
              Estimate Std. Error z value Pr(>|z|)
(Intercept)  0.3934383  0.1388907   2.833 0.004615
CHF          0.3872748  0.1049917   3.689 0.000225
Cirr        -0.7697450  0.1694342  -4.543 5.55e-06
colcan      -1.2636489  1.0901722  -1.159 0.246404
Coma        -0.5918506  0.1263961  -4.683 2.83e-06
COPD        -1.2931956  0.1487784  -8.692  < 2e-16
lungcan     -1.3801526  0.4852428  -2.844 0.004452
MOSF         0.0114630  0.1126526   0.102 0.918951
sepsis       0.7501648  0.0726732  10.322  < 2e-16
age         -0.0031374  0.0017289  -1.815 0.069567
male         0.1697903  0.0583574   2.909 0.003620
meanbp1     -0.0109824  0.0008217 -13.366  < 2e-16
               
(Intercept) ** 
CHF         ***
Cirr        ***
colcan         
Coma        ***
COPD        ***
lungcan     ** 
MOSF           
sepsis      ***
age         .  
male        ** 
meanbp1     ***
---
Signif. codes:  
0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1

(Dispersion parameter for binomial family taken to be 1)

    Null deviance: 7621.4  on 5734  degrees of freedom
Residual deviance: 6983.6  on 5723  degrees of freedom
AIC: 7007.6

Number of Fisher Scoring iterations: 4
  • Représenter les distributions des scores de propension chez les contrôles et les traités.

Correction

ggplot(rhc_small_disjunctive %>% mutate(propensity = pscore) ,
       aes(as.factor(treatment),pscore,fill = treatment) )+ geom_violin()

  • Calculer le logit des scores de propension, puis leur écart-type et le “caliper”

Correction

#logit = function(p) {log(p)-log(1-p)}
#logit_pscore = logit(pscore) # or 
predict(psmodel)
           1            2            3            4 
-1.400648912  0.206435553 -0.365702788 -0.446938036 
           5            6            7            8 
 0.386478803 -2.432791038 -0.333583743 -0.979428947 
           9           10           11           12 
-0.233769759 -0.560200391 -0.269666530 -0.200324292 
          13           14           15           16 
 0.496933291  0.165445731 -0.429765350 -1.010832061 
          17           18           19           20 
-0.351163270 -0.544522136  0.601348372 -0.252588293 
          21           22           23           24 
 0.638501114 -1.121127331  0.302395050 -1.386075423 
          25           26           27           28 
 0.063756725 -1.312987197 -1.485354631 -0.752852707 
          29           30           31           32 
-0.735231558 -0.263661837  0.553191164 -0.125712536 
          33           34           35           36 
-0.012037181 -2.169902474 -0.220063003  0.500564612 
          37           38           39           40 
-0.304270328 -0.037328206 -0.321020718 -0.392746715 
          41           42           43           44 
-0.009096550 -0.015754710 -0.417882807  0.030075445 
          45           46           47           48 
-0.073346976 -0.692253800 -0.311221247 -2.245760828 
          49           50           51           52 
-1.622956974 -1.094100164 -0.355751720 -1.022260594 
          53           54           55           56 
-1.615104551 -0.197929059  0.557622560 -0.366691537 
          57           58           59           60 
-0.332140615  0.816305732 -0.443913304 -0.590350667 
          61           62           63           64 
-0.350888542 -0.460618569  0.558692445 -0.235604151 
          65           66           67           68 
-0.423986112  0.439442291 -0.535514682 -2.468019128 
          69           70           71           72 
-1.144485692 -1.088701436  0.026776416 -0.115804169 
          73           74           75           76 
 0.788996260 -1.997475116 -1.062562512 -0.618636970 
          77           78           79           80 
 0.128631257 -0.607510122 -1.134748165  0.746155514 
          81           82           83           84 
-2.517407412 -2.097756628 -1.012519067  0.375518164 
          85           86           87           88 
-0.300756712 -0.313193023 -2.418509668 -1.351554781 
          89           90           91           92 
-0.453460554 -0.325369408 -1.832222170 -1.747207758 
          93           94           95           96 
-0.618048385 -0.429072240 -1.174785938 -1.037662008 
          97           98           99          100 
-0.474790115 -0.626512025 -0.568803160  0.986737159 
         101          102          103          104 
 0.745256732  0.131305928 -0.756925651 -1.180240587 
         105          106          107          108 
-0.994541841 -0.130795311 -0.252869087 -1.766164149 
         109          110          111          112 
-0.536342183 -0.497079576 -0.287012544 -0.451697532 
         113          114          115          116 
-1.497674133 -0.494227711 -1.681941425 -0.571193773 
         117          118          119          120 
-0.955575973  0.544051942 -1.137706566  0.584800140 
         121          122          123          124 
 0.498304271 -1.806786940  0.584086403 -0.247229842 
         125          126          127          128 
-1.293951261 -0.432721217 -0.222171538 -1.682684800 
         129          130          131          132 
-1.564921999 -0.183602042 -1.713368905 -0.353757820 
         133          134          135          136 
 0.319654292  0.185604913 -0.980498813 -0.264301790 
         137          138          139          140 
-0.359121109 -0.288403122 -1.710996526 -1.775848040 
         141          142          143          144 
 0.016331188 -0.353419106 -0.045089192 -0.619049809 
         145          146          147          148 
-0.278697487  0.401488066 -0.488519211  0.093230875 
         149          150          151          152 
-0.170433038 -0.771024543 -0.389570941 -1.280378942 
         153          154          155          156 
-0.788144850 -2.319550646 -0.456023600 -0.405585826 
         157          158          159          160 
-1.053467291 -0.219887413 -1.357760638  0.491395637 
         161          162          163          164 
-1.671719521 -0.391381374 -1.177919832 -0.888477164 
         165          166          167          168 
 0.304153904 -1.076748891 -2.217587167 -1.704265496 
         169          170          171          172 
-0.262462724 -1.023550165  0.212024338  0.404996979 
         173          174          175          176 
-0.398454528 -0.478881049  0.076401682 -0.346787447 
         177          178          179          180 
-0.059754045  0.551377629 -0.290487757  0.364033839 
         181          182          183          184 
-2.213005059 -0.323365979 -0.265815454 -0.815768058 
         185          186          187          188 
-0.586075910 -0.334340484  0.476789530 -1.081428403 
         189          190          191          192 
 0.482873223 -0.296900525 -0.691086858  0.270809801 
         193          194          195          196 
-0.979627349 -1.037431656 -0.228043824 -2.857726713 
         197          198          199          200 
-1.138839086 -1.883293225 -0.120164395  0.052449598 
         201          202          203          204 
 0.098517957 -0.036892086 -0.261303475 -0.271138147 
         205          206          207          208 
-0.902426184  0.270524342 -1.360308285 -0.152474706 
         209          210          211          212 
-1.553444166 -1.190108659 -1.175168574 -1.079304590 
         213          214          215          216 
-0.216571135 -1.276087418 -0.104262478  0.258611760 
         217          218          219          220 
-0.364915156 -1.478013051  0.517260118 -1.118516686 
         221          222          223          224 
-1.191807644  0.072371933 -0.986471988 -0.343638322 
         225          226          227          228 
-0.099439678 -0.138747942 -1.137294154  0.627066852 
         229          230          231          232 
-1.941603032 -1.182006493 -0.300526750 -1.365386165 
         233          234          235          236 
-1.071234903  0.401859540  0.265793132  0.605042526 
         237          238          239          240 
-0.852048653 -1.181525143 -0.310305264 -1.085691267 
         241          242          243          244 
-1.225979982 -2.003014139 -1.584985118 -0.233760901 
         245          246          247          248 
-0.000728897 -2.462214727 -1.010277558  0.491433446 
         249          250          251          252 
-2.054487680 -0.059421024 -2.078585749 -0.278114300 
         253          254          255          256 
-2.590342024 -0.304127542  0.487237292  0.012659905 
         257          258          259          260 
-0.121574618 -0.373316252 -0.964558582 -0.256057015 
         261          262          263          264 
-0.408450997 -0.265512900 -0.285439961  0.480148282 
         265          266          267          268 
 0.183941553 -0.186787476 -1.287407967 -0.926410446 
         269          270          271          272 
-0.953947268 -0.653102749 -0.070195345 -1.694159646 
         273          274          275          276 
-0.024215676  0.592499278 -1.266393741 -1.683637414 
         277          278          279          280 
-0.159678165 -0.408693183  0.414914122 -0.033499097 
         281          282          283          284 
 0.476562016 -0.322195218 -0.314826914 -1.136697763 
         285          286          287          288 
-0.516132931 -1.880364913 -1.044841058 -0.550021220 
         289          290          291          292 
-0.720334401  0.554146256 -0.217269286 -2.634042824 
         293          294          295          296 
-1.159392053 -0.649050555 -1.180778119 -1.327864719 
         297          298          299          300 
-1.424934900  0.050115325  0.728895153 -1.058673472 
         301          302          303          304 
-1.590547780 -1.650348030  0.022192671 -0.019123321 
         305          306          307          308 
-1.001061209 -0.432025950  0.363426919 -0.359263272 
         309          310          311          312 
 0.322498009 -2.494864063 -1.750472821 -1.400221810 
         313          314          315          316 
-0.773537387 -2.209288926  0.520341058 -0.679989930 
         317          318          319          320 
 0.097397226 -0.491154372  0.317785665  0.492478087 
         321          322          323          324 
-1.277277623 -0.166398233 -0.378726574  0.125643029 
         325          326          327          328 
-1.078124708 -0.454020437 -0.260525985 -0.250323383 
         329          330          331          332 
-1.911022014 -0.490456497 -2.404198378  0.614130076 
         333          334          335          336 
-0.122992605  0.442839949 -0.393006281 -1.656136317 
         337          338          339          340 
 0.459633100 -0.896054906 -1.232910225  0.535692178 
         341          342          343          344 
-0.335975055  0.104100282 -0.954618957 -1.337192129 
         345          346          347          348 
-2.068706029 -0.063504935 -1.956498321  0.188713624 
         349          350          351          352 
-0.226721467 -0.083111364 -0.290798354 -0.099170605 
         353          354          355          356 
-0.333488765 -0.337840373 -1.641659130  0.594301858 
         357          358          359          360 
 0.409378346 -0.252197849 -1.004652139 -0.431126985 
         361          362          363          364 
-0.695881056 -1.415279865 -1.629752679 -0.841487094 
         365          366          367          368 
-0.338011393 -2.171834587  0.498564690  0.580993023 
         369          370          371          372 
-1.800085489  0.540520942  0.167856387 -1.441431649 
         373          374          375          376 
-0.347112894 -0.210175118 -0.259978591 -0.455598514 
         377          378          379          380 
-0.783834074 -0.349510783  0.528388524 -0.649929439 
         381          382          383          384 
-0.830305662 -1.183211704 -0.636771215  0.287128878 
         385          386          387          388 
 0.569604267 -0.452800022 -1.174589955  0.261675294 
         389          390          391          392 
 0.259566566 -1.097953084  0.486578199 -0.323334521 
         393          394          395          396 
 0.048896333 -0.449237634 -1.105273024 -1.051155043 
         397          398          399          400 
-0.110588274  0.204782132 -1.592446401 -1.710297059 
         401          402          403          404 
-0.739150986  0.485488148  0.462031410 -0.361009359 
         405          406          407          408 
-0.180988730  0.750731390 -1.011168238 -0.438768385 
         409          410          411          412 
-1.971475784 -0.359024010 -0.173575890 -0.741455862 
         413          414          415          416 
 0.346375029  0.351118823 -2.620429665 -0.221035836 
         417          418          419          420 
-0.246782789  0.624536592 -1.068817504 -1.227151831 
         421          422          423          424 
 0.324796255  0.244851060  0.019677025 -0.474868527 
         425          426          427          428 
-0.323329902  0.586979158 -1.263237930 -0.399849817 
         429          430          431          432 
 0.546513193 -2.240538890 -1.583653669 -1.572023724 
         433          434          435          436 
 0.201735393 -0.479663577 -0.152979851  0.642653517 
         437          438          439          440 
-2.337556015 -0.209940470 -1.262715621 -0.220753387 
         441          442          443          444 
-0.191092657 -1.759064581 -2.107047973  0.347667867 
         445          446          447          448 
-0.348411820 -0.373091894 -0.104457040 -0.236486549 
         449          450          451          452 
-0.460143121 -0.991741618 -1.959870929 -0.352332698 
         453          454          455          456 
-0.036521693  0.315164482  0.083150424  0.313737190 
         457          458          459          460 
-1.011199868 -0.716040044 -0.211581149 -0.129579557 
         461          462          463          464 
-0.246683085 -0.813215515 -0.275204050 -0.830955796 
         465          466          467          468 
-0.549081553  0.614953376 -1.018498883 -0.508377761 
         469          470          471          472 
 0.475382466 -0.773430498  0.577679896 -0.414934237 
         473          474          475          476 
-0.361021633 -0.958666891 -0.543088994 -1.917004690 
         477          478          479          480 
-0.226907998 -1.055864925 -0.308398318  0.342362399 
         481          482          483          484 
-0.339376072 -1.825286932 -0.291816718 -0.723783072 
         485          486          487          488 
-1.297980970 -1.088696845  0.491067840 -0.894870333 
         489          490          491          492 
-0.405256185 -0.289010244  0.483914850 -0.790163577 
         493          494          495          496 
 0.679414083 -0.564342529  0.665926499 -0.188529465 
         497          498          499          500 
 0.603017349 -1.104648205 -0.261249909 -0.247882268 
         501          502          503          504 
-0.388688810 -1.539397777 -1.684515886 -0.927361155 
         505          506          507          508 
 0.435766282 -2.180974264 -1.612018942 -0.120663092 
         509          510          511          512 
-1.347767247 -0.314908641 -1.598614030  0.419077563 
         513          514          515          516 
-0.138511681 -0.474923561  0.256697858 -0.451693854 
         517          518          519          520 
 0.252633581 -0.060643396 -0.855645690 -1.125434867 
         521          522          523          524 
-0.916272746 -0.387602336 -0.868593370 -0.434439392 
         525          526          527          528 
-0.801797154 -0.398827761 -1.049896367 -1.197397675 
         529          530          531          532 
-1.821295770 -2.017578230 -0.224861963  0.041051459 
         533          534          535          536 
 0.637158389 -0.400046753 -0.462789216 -0.941143624 
         537          538          539          540 
-1.744957638 -0.917514399 -0.597801908 -0.260858377 
         541          542          543          544 
 0.239517667 -0.283365521 -1.223570404  0.454239643 
         545          546          547          548 
-0.816418718 -0.362398063 -0.992140429 -2.473457677 
         549          550          551          552 
-1.681936714  0.025400950  0.476703322 -0.454601895 
         553          554          555          556 
-0.764749222 -0.428137224 -0.387492612  0.380621300 
         557          558          559          560 
-0.220116490 -0.618583690  0.448474649 -0.357425512 
         561          562          563          564 
-1.215443148 -0.121041267  0.288103113 -0.487813285 
         565          566          567          568 
-1.787114375 -0.306462818  0.476491587 -0.954711326 
         569          570          571          572 
-0.239839876  0.005805957 -0.451201710  0.235921779 
         573          574          575          576 
-1.071031699  0.376710512 -0.165399009  0.380296377 
         577          578          579          580 
 0.507629933  0.508765895 -1.113857786  0.489301715 
         581          582          583          584 
-2.358894928 -0.511128909 -1.278664486  0.418977068 
         585          586          587          588 
 0.207059816 -0.288649256 -0.810357274 -1.867449769 
         589          590          591          592 
-0.202760457 -1.643737718 -1.619022655 -0.233563282 
         593          594          595          596 
-1.952203288  0.300074774  0.328696951  0.303360019 
         597          598          599          600 
-1.820652430 -0.060030175 -0.370795182  0.118767490 
         601          602          603          604 
 0.592491603 -1.433012550 -0.279663947  0.412504751 
         605          606          607          608 
-1.139909954  0.181155956  0.050435317 -0.221169212 
         609          610          611          612 
-0.685577857  0.495257940 -2.455588859 -2.623584356 
         613          614          615          616 
-1.716317609 -0.091353007  0.401881851  0.053745368 
         617          618          619          620 
 0.428408121 -0.106208835 -1.649629218 -0.474524912 
         621          622          623          624 
-1.800951512 -0.472744091  0.616969244 -0.340609356 
         625          626          627          628 
 0.080350093 -0.340070297 -0.727540047  0.127467201 
         629          630          631          632 
-0.215293652  0.508372168  0.701645457 -0.573995463 
         633          634          635          636 
-2.083354524 -1.554913701 -2.103311262 -1.209872812 
         637          638          639          640 
-1.651166786 -1.496641266  0.153780443 -0.886266808 
         641          642          643          644 
-1.572053151  0.442469798  0.538469067 -0.168775759 
         645          646          647          648 
-0.119122747 -1.509247455  1.131679479 -0.232733485 
         649          650          651          652 
-0.113304427 -0.241347080  0.603630766 -0.683201149 
         653          654          655          656 
 0.097246661  0.405001697 -1.739548112 -1.930278753 
         657          658          659          660 
-0.967993234 -0.346581017 -0.247697093 -0.643059983 
         661          662          663          664 
-0.444685210 -0.425800004 -0.248973457 -0.073412646 
         665          666          667          668 
-1.695239018 -2.456442244  0.372926719  0.311336928 
         669          670          671          672 
-2.142305784 -1.490322835 -0.213896732  0.395584837 
         673          674          675          676 
-0.387394595 -1.662513258 -0.211720818 -0.712796097 
         677          678          679          680 
-2.360678258  0.422988375 -2.308880512  0.470340870 
         681          682          683          684 
-1.856597705 -0.291484190 -0.903194134 -1.702282518 
         685          686          687          688 
-1.398377170  0.487403406 -0.687888554 -0.299582564 
         689          690          691          692 
-0.136925692 -1.300898703 -0.497852139 -0.339838781 
         693          694          695          696 
-1.981171728 -1.044329221 -0.356832562 -0.918359902 
         697          698          699          700 
-1.671147445 -0.255059733 -0.350496523 -0.359557311 
         701          702          703          704 
-0.346441628 -1.073820243 -0.104844598 -1.019273767 
         705          706          707          708 
-0.726982507  0.246476240 -0.533081035 -0.135750998 
         709          710          711          712 
-2.509548536 -1.413113410 -0.027161573  0.670827178 
         713          714          715          716 
 0.362401110 -0.223703926 -0.108555921 -0.720833821 
         717          718          719          720 
-0.302971498 -0.300033236 -0.522821634  0.114472126 
         721          722          723          724 
-0.537862206  0.122129884 -0.806187817  0.100669511 
         725          726          727          728 
-1.111664151 -0.413904406 -0.990899406 -1.183317735 
         729          730          731          732 
-0.515128677 -1.603982146  0.500843933 -0.305144303 
         733          734          735          736 
-1.581077590 -0.379149755 -0.910727722 -1.768504918 
         737          738          739          740 
-1.128295933  0.390274266 -0.462096256 -1.067199273 
         741          742          743          744 
-1.795102173 -2.017093038 -0.961130245  0.258697899 
         745          746          747          748 
-0.119097662  0.700920765 -0.177543769 -0.197448993 
         749          750          751          752 
-0.161998173 -0.891822409 -0.134277907  0.612765173 
         753          754          755          756 
-0.368657300  0.373879156 -0.874028936 -0.538232513 
         757          758          759          760 
 0.402457437  0.553059095  0.123700982 -1.709435841 
         761          762          763          764 
-0.010046261 -0.230065188 -1.324954760 -0.324394156 
         765          766          767          768 
-0.111429063 -1.174893872 -0.777237793 -0.494947094 
         769          770          771          772 
-0.560503119 -0.682869996 -0.245926252 -1.070704989 
         773          774          775          776 
-0.289455577 -0.597384596 -0.213902884 -0.498099220 
         777          778          779          780 
-2.225565665 -0.103214462 -0.350320726  0.602739661 
         781          782          783          784 
-0.065838360 -1.025289660 -0.925535255  0.582062884 
         785          786          787          788 
-0.879859722  0.632212202 -1.332162737 -0.065157470 
         789          790          791          792 
-0.213620604 -1.303433931 -1.000200402  0.256968960 
         793          794          795          796 
-0.410285404 -1.629150572 -3.025573018 -1.273385755 
         797          798          799          800 
-0.331897636 -1.271312642 -1.662456664 -0.611717729 
         801          802          803          804 
-1.698354569 -0.341835791 -0.863854390 -1.158556432 
         805          806          807          808 
-0.298794195  0.184482856 -0.597936485 -0.463749407 
         809          810          811          812 
-0.458026288 -0.273748354  0.567213607 -2.201558069 
         813          814          815          816 
 0.610674096  0.220176778  0.186274399 -0.961298060 
         817          818          819          820 
-0.120595903 -0.303278200  0.480301958  0.261116815 
         821          822          823          824 
-2.426356005 -0.146676668 -0.287331553  0.619323942 
         825          826          827          828 
-0.334078452 -0.471677713 -1.045784659  0.115361854 
         829          830          831          832 
-0.402622537  0.693681352 -2.453543216 -1.581763511 
         833          834          835          836 
-2.760830218 -0.468234553 -0.852124718 -1.692269888 
         837          838          839          840 
-0.422530085 -0.355749421 -1.461364855 -0.845546335 
         841          842          843          844 
 0.167836112  0.260004708  0.174650864 -0.909787145 
         845          846          847          848 
-1.339276854 -1.747092317 -0.677506707 -0.992319148 
         849          850          851          852 
-1.584154156 -0.349254033  0.615038238 -0.540342182 
         853          854          855          856 
-1.556460516 -0.209596814 -0.235940936 -0.738706198 
         857          858          859          860 
-2.407550588 -0.150647182  0.410986558 -1.333218679 
         861          862          863          864 
-1.416995771 -0.153750913  0.456801557 -2.383560908 
         865          866          867          868 
-0.842229093 -0.354612557  0.263331882  0.653679701 
         869          870          871          872 
-0.461773034 -0.235381593 -0.305070507 -2.695182225 
         873          874          875          876 
 0.478093405 -1.774896968 -0.996464036 -1.079717259 
         877          878          879          880 
-0.278491349 -1.633150808 -0.226991323 -0.155626095 
         881          882          883          884 
 0.505156330  0.156207253  0.450813636 -0.078027758 
         885          886          887          888 
-0.238220861 -0.097251854 -1.102241562  0.430047203 
         889          890          891          892 
-0.314918690 -0.303073338  0.594968559 -2.168145272 
         893          894          895          896 
-0.384146576  0.444893548  0.287635596  0.504106889 
         897          898          899          900 
 0.167147447 -0.587985193 -0.218069458  0.196627589 
         901          902          903          904 
-0.365005149 -0.181900109 -0.308647850 -1.592736578 
         905          906          907          908 
 0.612716658 -2.170254032 -0.273560007 -0.133405610 
         909          910          911          912 
-0.560689845 -2.481537724  0.498551974 -1.640330164 
         913          914          915          916 
-2.340338897  0.380084615 -2.000137140 -0.169214070 
         917          918          919          920 
-0.755312043 -0.365068239 -1.910667062 -0.568718190 
         921          922          923          924 
-0.399988853 -0.367094875  0.593349541 -0.095016654 
         925          926          927          928 
-1.025834401 -1.455844663 -0.135661471  0.553416977 
         929          930          931          932 
-0.545161319 -0.372028513 -0.080108838 -0.384640004 
         933          934          935          936 
-0.274719322 -0.609604599 -1.148229537 -1.188666505 
         937          938          939          940 
-0.146232179 -1.503699282 -0.319876697 -2.472513156 
         941          942          943          944 
-1.069732231 -1.393236581 -1.810694341 -0.324233731 
         945          946          947          948 
-1.884857474 -1.107143974 -1.025623780 -0.088473738 
         949          950          951          952 
-0.195008064 -1.909829272  0.235445238  1.056407529 
         953          954          955          956 
-1.663328816 -2.540048592 -0.344974748  0.320071227 
         957          958          959          960 
-2.330732159  0.394663420 -0.241159714  0.571610739 
         961          962          963          964 
-0.403807798  0.633707245  0.452686874 -0.741528241 
         965          966          967          968 
-1.265917365  0.070891191 -2.207259034 -1.925541700 
         969          970          971          972 
 0.032155507 -0.994483663  0.771670310  0.156353563 
         973          974          975          976 
 0.009953937 -0.856336321 -0.164352215 -0.959666752 
         977          978          979          980 
-0.273887916 -0.120023285 -1.479593408 -0.591708360 
         981          982          983          984 
-0.577790930 -0.827642953 -0.236059232  0.258396639 
         985          986          987          988 
 0.317871962 -0.305911391 -0.389589939 -0.224244565 
         989          990          991          992 
-1.264793982 -0.296597011 -0.216898619  0.580244473 
         993          994          995          996 
-0.370569312  0.283045557  0.348455422  0.057444241 
         997          998          999         1000 
-0.870011453  0.248071300 -0.986103848  0.982169152 
 [ reached getOption("max.print") -- omitted 4735 entries ]
  • Appliquer le greedy matching sur le logitdu score de propension avec le “capiler” calculé à la question précédente

Correction

psmatch = Match(Tr=rhc_small_disjunctive$treatment,M=1,X=logit_pscore,
             replace=FALSE,caliper=caliper)

n_matched = length(psmatch$index.control)

rhc_small_disjunctive_matched_propensity =rhc_small_disjunctive %>%
                                  slice(c(psmatch$index.control,
                                          psmatch$index.treated)) %>%
                                  mutate( match_id = rep(c(1:n_matched),2))
  • Calculer la “Table 1”

Correction

matchedtab1<-CreateTableOne(vars=xvars, strata ="treatment", 
                            data=rhc_small_disjunctive_matched_propensity, test = FALSE)
print(matchedtab1, smd = TRUE)
                     Stratified by treatment
                      0             1            
  n                    1925          1925        
  CHF (mean (SD))      0.09 (0.29)   0.09 (0.29) 
  Cirr (mean (SD))     0.03 (0.16)   0.03 (0.16) 
  colcan (mean (SD))   0.00 (0.02)   0.00 (0.02) 
  Coma (mean (SD))     0.04 (0.21)   0.05 (0.22) 
  COPD (mean (SD))     0.03 (0.16)   0.03 (0.17) 
  lungcan (mean (SD))  0.00 (0.06)   0.00 (0.05) 
  MOSF (mean (SD))     0.08 (0.27)   0.08 (0.27) 
  sepsis (mean (SD))   0.25 (0.43)   0.25 (0.43) 
  age (mean (SD))     61.14 (17.78) 60.94 (15.53)
  male (mean (SD))     0.57 (0.50)   0.57 (0.49) 
  meanbp1 (mean (SD)) 71.37 (33.68) 71.06 (35.06)
                     Stratified by treatment
                      SMD   
  n                         
  CHF (mean (SD))      0.009
  Cirr (mean (SD))     0.003
  colcan (mean (SD))  <0.001
  Coma (mean (SD))     0.025
  COPD (mean (SD))     0.019
  lungcan (mean (SD))  0.010
  MOSF (mean (SD))    <0.001
  sepsis (mean (SD))   0.012
  age (mean (SD))      0.012
  male (mean (SD))     0.015
  meanbp1 (mean (SD))  0.009
  • Sur les données “matchées” étudiez l’outcome

Correction

y_trt = rhc_small_disjunctive_matched_propensity %>% filter(treatment==1) %>% 
                                      dplyr::select(Death) %>% pull()
y_con = rhc_small_disjunctive_matched_propensity %>% filter(treatment==0) %>%
                                      dplyr::select(Death)%>% pull()
table(y_trt,y_con)
     y_con
y_trt   0   1
    0 231 381
    1 468 845
mcnemar.test(matrix(table(y_trt,y_con),2,2))

    McNemar's Chi-squared test with continuity
    correction

data:  matrix(table(y_trt, y_con), 2, 2)
McNemar's chi-squared = 8.7114, df = 1, p-value
= 0.003162
  • Faire varier la valeur du “caliper” et observer les changements dans la “Table 1” et le test sur l’outcome

Correction

for (cal_const in c(0.0001,0.01,0.1,0.5,1,5)){
  
  caliper = cal_const *sd(logit_pscore)
  print(paste(c("Le caliper vaut ",caliper),collapse=""))
  psmatch = Match(Tr=rhc_small_disjunctive$treatment,M=1,X=logit_pscore,
               replace=FALSE,caliper=caliper)
  n_matched = length(psmatch$index.control)
  
  rhc_small_disjunctive_matched_propensity =rhc_small_disjunctive %>%
                                    slice(c(psmatch$index.control,
                                            psmatch$index.treated)) %>%
                                    mutate( match_id = rep(c(1:n_matched),2))
  print(paste(c("Il y a ",length(psmatch$index.treated)," données matchées"),collapse=""))
  matchedtab1<-CreateTableOne(vars=xvars, strata ="treatment", 
                              data=rhc_small_disjunctive_matched_propensity, test = FALSE)
  print(matchedtab1, smd = TRUE)
  y_trt = rhc_small_disjunctive_matched_propensity %>% filter(treatment==1) %>% 
                                        dplyr::select(Death) %>% pull()
  y_con = rhc_small_disjunctive_matched_propensity %>% filter(treatment==0) %>%
                                        dplyr::select(Death)%>% pull()
  print(table(y_trt,y_con))
  print(mcnemar.test(matrix(table(y_trt,y_con),2,2)))
  }
[1] "Le caliper vaut 7.7179970051288e-05"
[1] "Il y a 335 données matchées"
                     Stratified by treatment
                      0             1            
  n                     335           335        
  CHF (mean (SD))      0.05 (0.21)   0.05 (0.22) 
  Cirr (mean (SD))     0.04 (0.19)   0.04 (0.19) 
  colcan (mean (SD))   0.00 (0.00)   0.00 (0.00) 
  Coma (mean (SD))     0.05 (0.22)   0.04 (0.19) 
  COPD (mean (SD))     0.01 (0.12)   0.01 (0.12) 
  lungcan (mean (SD))  0.00 (0.00)   0.00 (0.05) 
  MOSF (mean (SD))     0.08 (0.28)   0.13 (0.33) 
  sepsis (mean (SD))   0.24 (0.43)   0.21 (0.41) 
  age (mean (SD))     62.28 (17.65) 61.94 (15.10)
  male (mean (SD))     0.52 (0.50)   0.56 (0.50) 
  meanbp1 (mean (SD)) 69.27 (31.02) 68.71 (31.26)
                     Stratified by treatment
                      SMD   
  n                         
  CHF (mean (SD))      0.014
  Cirr (mean (SD))    <0.001
  colcan (mean (SD))  <0.001
  Coma (mean (SD))     0.058
  COPD (mean (SD))    <0.001
  lungcan (mean (SD))  0.077
  MOSF (mean (SD))     0.146
  sepsis (mean (SD))   0.057
  age (mean (SD))      0.020
  male (mean (SD))     0.072
  meanbp1 (mean (SD))  0.018
     y_con
y_trt   0   1
    0  38  81
    1  90 126

    McNemar's Chi-squared test with continuity
    correction

data:  matrix(table(y_trt, y_con), 2, 2)
McNemar's chi-squared = 0.37427, df = 1,
p-value = 0.5407

[1] "Le caliper vaut 0.0077179970051288"
[1] "Il y a 1863 données matchées"
                     Stratified by treatment
                      0             1            
  n                    1863          1863        
  CHF (mean (SD))      0.09 (0.29)   0.08 (0.27) 
  Cirr (mean (SD))     0.03 (0.16)   0.03 (0.16) 
  colcan (mean (SD))   0.00 (0.02)   0.00 (0.00) 
  Coma (mean (SD))     0.05 (0.21)   0.05 (0.22) 
  COPD (mean (SD))     0.03 (0.16)   0.03 (0.17) 
  lungcan (mean (SD))  0.00 (0.06)   0.00 (0.05) 
  MOSF (mean (SD))     0.08 (0.27)   0.08 (0.28) 
  sepsis (mean (SD))   0.25 (0.43)   0.25 (0.44) 
  age (mean (SD))     60.83 (17.91) 60.92 (15.58)
  male (mean (SD))     0.57 (0.50)   0.57 (0.50) 
  meanbp1 (mean (SD)) 71.63 (33.51) 71.25 (34.94)
                     Stratified by treatment
                      SMD   
  n                         
  CHF (mean (SD))      0.033
  Cirr (mean (SD))     0.007
  colcan (mean (SD))   0.033
  Coma (mean (SD))     0.020
  COPD (mean (SD))     0.020
  lungcan (mean (SD))  0.019
  MOSF (mean (SD))     0.008
  sepsis (mean (SD))   0.011
  age (mean (SD))      0.006
  male (mean (SD))     0.003
  meanbp1 (mean (SD))  0.011
     y_con
y_trt   0   1
    0 230 376
    1 467 790

    McNemar's Chi-squared test with continuity
    correction

data:  matrix(table(y_trt, y_con), 2, 2)
McNemar's chi-squared = 9.6085, df = 1, p-value
= 0.001937

[1] "Le caliper vaut 0.077179970051288"
[1] "Il y a 1915 données matchées"
                     Stratified by treatment
                      0             1            
  n                    1915          1915        
  CHF (mean (SD))      0.10 (0.30)   0.09 (0.28) 
  Cirr (mean (SD))     0.03 (0.17)   0.03 (0.16) 
  colcan (mean (SD))   0.00 (0.03)   0.00 (0.02) 
  Coma (mean (SD))     0.04 (0.21)   0.05 (0.22) 
  COPD (mean (SD))     0.03 (0.17)   0.03 (0.17) 
  lungcan (mean (SD))  0.00 (0.05)   0.00 (0.05) 
  MOSF (mean (SD))     0.08 (0.27)   0.08 (0.27) 
  sepsis (mean (SD))   0.25 (0.44)   0.25 (0.43) 
  age (mean (SD))     60.96 (17.72) 60.86 (15.61)
  male (mean (SD))     0.57 (0.50)   0.57 (0.50) 
  meanbp1 (mean (SD)) 71.82 (33.96) 71.17 (35.12)
                     Stratified by treatment
                      SMD   
  n                         
  CHF (mean (SD))      0.042
  Cirr (mean (SD))     0.019
  colcan (mean (SD))   0.019
  Coma (mean (SD))     0.022
  COPD (mean (SD))     0.009
  lungcan (mean (SD)) <0.001
  MOSF (mean (SD))     0.006
  sepsis (mean (SD))   0.004
  age (mean (SD))      0.006
  male (mean (SD))     0.009
  meanbp1 (mean (SD))  0.019
     y_con
y_trt   0   1
    0 239 373
    1 460 843

    McNemar's Chi-squared test with continuity
    correction

data:  matrix(table(y_trt, y_con), 2, 2)
McNemar's chi-squared = 8.8788, df = 1, p-value
= 0.002885

[1] "Le caliper vaut 0.38589985025644"
[1] "Il y a 1957 données matchées"
                     Stratified by treatment
                      0             1            
  n                    1957          1957        
  CHF (mean (SD))      0.09 (0.29)   0.10 (0.30) 
  Cirr (mean (SD))     0.03 (0.16)   0.03 (0.16) 
  colcan (mean (SD))   0.00 (0.02)   0.00 (0.02) 
  Coma (mean (SD))     0.04 (0.20)   0.05 (0.21) 
  COPD (mean (SD))     0.03 (0.16)   0.03 (0.17) 
  lungcan (mean (SD))  0.00 (0.06)   0.00 (0.05) 
  MOSF (mean (SD))     0.08 (0.28)   0.08 (0.27) 
  sepsis (mean (SD))   0.25 (0.43)   0.25 (0.43) 
  age (mean (SD))     61.24 (17.71) 60.91 (15.49)
  male (mean (SD))     0.57 (0.50)   0.58 (0.49) 
  meanbp1 (mean (SD)) 71.51 (34.12) 70.72 (34.91)
                     Stratified by treatment
                      SMD   
  n                         
  CHF (mean (SD))      0.021
  Cirr (mean (SD))     0.013
  colcan (mean (SD))  <0.001
  Coma (mean (SD))     0.029
  COPD (mean (SD))     0.022
  lungcan (mean (SD))  0.010
  MOSF (mean (SD))     0.015
  sepsis (mean (SD))   0.006
  age (mean (SD))      0.020
  male (mean (SD))     0.014
  meanbp1 (mean (SD))  0.023
     y_con
y_trt   0   1
    0 251 374
    1 461 871

    McNemar's Chi-squared test with continuity
    correction

data:  matrix(table(y_trt, y_con), 2, 2)
McNemar's chi-squared = 8.8575, df = 1, p-value
= 0.002919

[1] "Le caliper vaut 0.77179970051288"
[1] "Il y a 2041 données matchées"
                     Stratified by treatment
                      0             1            
  n                    2041          2041        
  CHF (mean (SD))      0.09 (0.29)   0.10 (0.30) 
  Cirr (mean (SD))     0.02 (0.15)   0.02 (0.15) 
  colcan (mean (SD))   0.00 (0.02)   0.00 (0.02) 
  Coma (mean (SD))     0.05 (0.21)   0.05 (0.21) 
  COPD (mean (SD))     0.03 (0.16)   0.03 (0.17) 
  lungcan (mean (SD))  0.00 (0.06)   0.00 (0.05) 
  MOSF (mean (SD))     0.09 (0.28)   0.08 (0.27) 
  sepsis (mean (SD))   0.24 (0.43)   0.28 (0.45) 
  age (mean (SD))     60.70 (17.98) 60.92 (15.53)
  male (mean (SD))     0.58 (0.49)   0.58 (0.49) 
  meanbp1 (mean (SD)) 70.48 (33.19) 69.94 (34.56)
                     Stratified by treatment
                      SMD   
  n                         
  CHF (mean (SD))      0.021
  Cirr (mean (SD))    <0.001
  colcan (mean (SD))  <0.001
  Coma (mean (SD))     0.005
  COPD (mean (SD))     0.006
  lungcan (mean (SD))  0.018
  MOSF (mean (SD))     0.034
  sepsis (mean (SD))   0.088
  age (mean (SD))      0.013
  male (mean (SD))     0.003
  meanbp1 (mean (SD))  0.016
     y_con
y_trt   0   1
    0 243 410
    1 500 888

    McNemar's Chi-squared test with continuity
    correction

data:  matrix(table(y_trt, y_con), 2, 2)
McNemar's chi-squared = 8.7044, df = 1, p-value
= 0.003174

[1] "Le caliper vaut 3.8589985025644"
[1] "Il y a 2184 données matchées"
                     Stratified by treatment
                      0             1            
  n                    2184          2184        
  CHF (mean (SD))      0.09 (0.28)   0.10 (0.29) 
  Cirr (mean (SD))     0.02 (0.15)   0.02 (0.15) 
  colcan (mean (SD))   0.00 (0.02)   0.00 (0.02) 
  Coma (mean (SD))     0.05 (0.21)   0.04 (0.20) 
  COPD (mean (SD))     0.03 (0.17)   0.03 (0.16) 
  lungcan (mean (SD))  0.00 (0.04)   0.00 (0.05) 
  MOSF (mean (SD))     0.09 (0.28)   0.07 (0.26) 
  sepsis (mean (SD))   0.23 (0.42)   0.32 (0.47) 
  age (mean (SD))     61.00 (17.80) 60.75 (15.63)
  male (mean (SD))     0.56 (0.50)   0.59 (0.49) 
  meanbp1 (mean (SD)) 70.09 (32.48) 68.20 (34.24)
                     Stratified by treatment
                      SMD   
  n                         
  CHF (mean (SD))      0.027
  Cirr (mean (SD))     0.006
  colcan (mean (SD))  <0.001
  Coma (mean (SD))     0.011
  COPD (mean (SD))     0.014
  lungcan (mean (SD))  0.010
  MOSF (mean (SD))     0.054
  sepsis (mean (SD))   0.197
  age (mean (SD))      0.015
  male (mean (SD))     0.044
  meanbp1 (mean (SD))  0.057
     y_con
y_trt   0   1
    0 244 454
    1 545 941

    McNemar's Chi-squared test with continuity
    correction

data:  matrix(table(y_trt, y_con), 2, 2)
McNemar's chi-squared = 8.1081, df = 1, p-value
= 0.004407
---
title: "Exemples pour la causalité : appariement et score de propension (chapitre 2)"
output: html_notebook
---

```{r}
library(tidyverse)
library(tableone) #create 'Table 1' to describe baseline characteristics
library(Matching) #multivariate and propensity score matching with balance optimization
```

# Dataset : right heart catheterization
Une description complète est disponible à cette adresse http://biostat.mc.vanderbilt.edu/wiki/pub/Main/DataSets/rhc.html. Il s'agit de données sur des patients (2184 traités et 3551 contrôles) admis aux urgences dans 5 hôpitaux, la variable de traitement est `swang1` (right heat catheterization vs. non) et l'outcome est `death` (yes or no). Nous allons considérer les variables de confusion suivantes :
- `cat1` : Primary disease category
- `age`
- `sex``
- `meanbp1`: Mean blood pressure

# Matching (greedy)
## Questions : 
- Charger les données http://biostat.mc.vanderbilt.edu/wiki/Main/DataSets, 
```{r}
rhc = read_csv("./Data/rhc_data.csv")
rhc_small = rhc %>% mutate(treatment = swang1 ) %>% 
                dplyr::select(treatment,death, cat1, age,sex,meanbp1,-swang1)
                
glimpse(rhc_small)
unique(rhc_small$cat1)
```
- Créer la "Table 1". Attention le package `tableone` ne supporte pas les variables discrètes à plusieurs facteurs, il faut donc d'abord construire la table disjonctive.
```{r}

```
#### Correction
```{r}
rhc_small_disjunctive = as_tibble(model.matrix(~ . , data=rhc_small)[,-1])
glimpse(rhc_small_disjunctive)
names(rhc_small_disjunctive) = c("treatment","Death","CHF","Cirr",
                                 "colcan","Coma","COPD",
                                 "lungcan","MOSF","sepsis",
                                 "age","male","meanbp1")
```

```{r}

```
#### Correction
```{r}
xvars = c("CHF","Cirr","colcan","Coma","COPD",
         "lungcan","MOSF","sepsis",
         "age","male","meanbp1") # variables de confusion
table1 = CreateTableOne(vars=xvars,strata="treatment", data=rhc_small_disjunctive, test=FALSE)
print(table1,smd=TRUE)
```

- Faire un matching "greedy" basé sur la distance de Mahalanobis.
```{r}

```
#### Correction
```{r}
greedymatch = Match(Tr=rhc_small_disjunctive$treatment,
                    M=1,
                    X=rhc_small_disjunctive[xvars],replace=FALSE)

rhc_small_disjunctive_matched = rhc_small_disjunctive %>%
                                  slice(c(greedymatch$index.control,
                                          greedymatch$index.treated)) %>%
                                  mutate( match_id = rep(c(1:2184),2))
                                

matchedtab1 = CreateTableOne(vars=xvars,strata="treatment", data=rhc_small_disjunctive_matched, test=FALSE)
print(matchedtab1, smd = TRUE)
```
- Sur les données "matchées" étudiez l'outcome
```{r}

```
#### Correction
```{r}
y_trt = rhc_small_disjunctive_matched %>% filter(treatment==1) %>% 
                                      dplyr::select(Death) %>% pull()
y_con = rhc_small_disjunctive_matched %>% filter(treatment==0) %>%
                                      dplyr::select(Death)%>% pull()
table(y_trt,y_con)
mcnemar.test(matrix(table(y_trt,y_con),2,2))
```
# Propensity score matching

- Estimer le score propension grâce à une régression logistique et calculer les scores de propension.
```{r}

```
#### Correction
```{r}
psmodel = glm(treatment ~ . - Death ,
              family=binomial(),
              data=rhc_small_disjunctive)

summary(psmodel)
````

```{r}
#create propensity scores
pscore = predict(psmodel,type="response")
```
- Représenter les distributions des scores de propension chez les contrôles et les traités.
```{r}

```
#### Correction
```{r}
ggplot(rhc_small_disjunctive %>% mutate(propensity = pscore) ,
       aes(as.factor(treatment),pscore,fill = treatment) )+ geom_violin()
```
- Calculer le logit des scores de propension, puis leur écart-type et le "caliper"
```{r}

```
#### Correction
```{r}
#logit = function(p) {log(p)-log(1-p)}
#logit_pscore = logit(pscore) # or 
logit_pscore = predict(psmodel)
caliper = 0.2 *sd(logit_pscore)
```

- Appliquer le greedy matching sur le `logit`du score de propension avec le "capiler" calculé à la question précédente
```{r}

```
#### Correction
```{r}
psmatch = Match(Tr=rhc_small_disjunctive$treatment,M=1,X=logit_pscore,
             replace=FALSE,caliper=caliper)

n_matched = length(psmatch$index.control)

rhc_small_disjunctive_matched_propensity =rhc_small_disjunctive %>%
                                  slice(c(psmatch$index.control,
                                          psmatch$index.treated)) %>%
                                  mutate( match_id = rep(c(1:n_matched),2))


```
- Calculer la "Table 1"
```{r}

```
#### Correction
```{r}
matchedtab1<-CreateTableOne(vars=xvars, strata ="treatment", 
                            data=rhc_small_disjunctive_matched_propensity, test = FALSE)
print(matchedtab1, smd = TRUE)
```

- Sur les données "matchées" étudiez l'outcome
```{r}

```
#### Correction
```{r}
y_trt = rhc_small_disjunctive_matched_propensity %>% filter(treatment==1) %>% 
                                      dplyr::select(Death) %>% pull()
y_con = rhc_small_disjunctive_matched_propensity %>% filter(treatment==0) %>%
                                      dplyr::select(Death)%>% pull()
table(y_trt,y_con)
mcnemar.test(matrix(table(y_trt,y_con),2,2))
```
- Faire varier la valeur du "caliper" et observer les changements dans la "Table 1" et le test sur l'outcome
```{r}

```
#### Correction
```{r}
for (cal_const in c(0.0001,0.01,0.1,0.5,1,5)){
  
  caliper = cal_const *sd(logit_pscore)
  print(paste(c("Le caliper vaut ",caliper),collapse=""))
  psmatch = Match(Tr=rhc_small_disjunctive$treatment,M=1,X=logit_pscore,
               replace=FALSE,caliper=caliper)
  n_matched = length(psmatch$index.control)
  
  rhc_small_disjunctive_matched_propensity =rhc_small_disjunctive %>%
                                    slice(c(psmatch$index.control,
                                            psmatch$index.treated)) %>%
                                    mutate( match_id = rep(c(1:n_matched),2))
  print(paste(c("Il y a ",length(psmatch$index.treated)," données matchées"),collapse=""))
  matchedtab1<-CreateTableOne(vars=xvars, strata ="treatment", 
                              data=rhc_small_disjunctive_matched_propensity, test = FALSE)
  print(matchedtab1, smd = TRUE)
  y_trt = rhc_small_disjunctive_matched_propensity %>% filter(treatment==1) %>% 
                                        dplyr::select(Death) %>% pull()
  y_con = rhc_small_disjunctive_matched_propensity %>% filter(treatment==0) %>%
                                        dplyr::select(Death)%>% pull()
  print(table(y_trt,y_con))
  print(mcnemar.test(matrix(table(y_trt,y_con),2,2)))
  }
```

