results aggregated

Author: ehsanx

1RHC.Rmd

@@ -168,7 +168,7 @@ We also conduct propensity score pair matching analysis, as follows.
 
 **Note**: In this workshop, we will not cover Propensity Score Matching (PSM) in this workshop. If you want to learn more about this, feel free to check out this other workshop: [Understanding Propensity Score Matching](https://ehsanx.github.io/psw/).
 
-```{r ps1, cache=TRUE, echo = TRUE}
+```{r ps16854, cache=TRUE, echo = TRUE}
 set.seed(123)
 require(MatchIt)
 ps.formula <- as.formula(paste("A~", 
@@ -216,11 +216,23 @@ tab1y <- CreateTableOne(vars = c("Y"),
                data = matched.data, strata = "A", 
                test = TRUE)
 print(tab1y, showAllLevels = FALSE, 
-      test = TRUE) 
+      test = TRUE)
 ```
 
-We also find the same conclusion based on propensity score pair matched data.
+- Hence, we also find the same conclusion based on propensity score pair matched data.
+- We can also estimate the effect of `RHC` on `length of stay` using propensity score-matched sample:
+
+```{r ps12ryy, cache=TRUE, echo = TRUE}
+fit.matched <- glm(Y~A,
+            family=gaussian, 
+            data = matched.data)
+publish(fit.matched)
+```
+
+```{r, cache=TRUE, echo = TRUE}
+saveRDS(fit.matched, file = "data/match.RDS")
+```
 
 ### TMLE in RHC data
 
-There are other papers that have used RHC data [@keele2021comparing;@keele2018pre]. Particularly, @keele2021comparing used TMLE method in estimating the impact of RHC on length of stay, and found point estimate $2.01 (95\% CI: 0.6-3.41)$. In today's workshop, we will learn about TMLE method.
+There are other papers that have used RHC data [@keele2021comparing;@keele2018pre]. Particularly, @keele2021comparing used TMLE (with super learner) method in estimating the impact of RHC on length of stay, and found point estimate $2.01 (95\% CI: 0.6-3.41)$. In today's workshop, we will learn about TMLE method.

2gcomp.Rmd

@@ -1,6 +1,6 @@
 # G-computation
 
-```{r setup01, include=FALSE}
+```{r setup01g, include=FALSE}
 require(boot)
 require(knitr)
 require(glmnet)
@@ -277,7 +277,8 @@ ObsData$Pred.TE <- ObsData$Pred.Y1 - ObsData$Pred.Y0
 Mean value of predicted treatment effect 
 
 ```{r reg2acnx1, cache=cachex, echo = TRUE}
-mean(ObsData$Pred.TE)
+TE <- mean(ObsData$Pred.TE)
+TE
 ```
 
 SD of treatment effect
@@ -341,8 +342,14 @@ Below are two versions of confidence interval.
 - Another is based on percentiles
 
 ```{r bootx3, cache=cachex, echo = TRUE}
-boot.ci(gcomp.res, type="norm") 
-boot.ci(gcomp.res, type="perc")
+CI1 <- boot.ci(gcomp.res, type="norm") 
+CI1
+CI2 <- boot.ci(gcomp.res, type="perc")
+CI2
 ```
 
+```{r, cache=TRUE, echo = TRUE}
+saveRDS(TE, file = "data/gcomp.RDS")
+saveRDS(CI2, file = "data/gcompci.RDS")
+```
 

2gcomp2.Rmd

@@ -1,6 +1,6 @@
 # G-computation using ML
 
-```{r setup01, include=FALSE}
+```{r setup01gm, include=FALSE}
 require(knitr)
 require(glmnet)
 require(kableExtra)
@@ -199,6 +199,8 @@ ObsData$Pred.TE <- ObsData$Pred.Y1 - ObsData$Pred.Y0
 Mean value of predicted treatment effect 
 
 ```{r reg2acnx1b, cache=cachex, echo = TRUE}
+TE1 <- mean(ObsData$Pred.TE)
+TE1
 summary(ObsData$Pred.TE)
 ```
 
@@ -257,6 +259,8 @@ ObsData$Pred.TE <- ObsData$Pred.Y1 - ObsData$Pred.Y0
 Mean value of predicted treatment effect 
 
 ```{r reg2acnx1br, cache=cachex, echo = TRUE}
+TE2 <- mean(ObsData$Pred.TE)
+TE2
 summary(ObsData$Pred.TE)
 ```
 
@@ -461,6 +465,8 @@ ObsData$Pred.TE <- ObsData$Pred.Y1 - ObsData$Pred.Y0
 Mean value of predicted treatment effect 
 
 ```{r reg2acnx1bs, cache=cachex, echo = TRUE}
+TE3 <- mean(ObsData$Pred.TE)
+TE3
 summary(ObsData$Pred.TE)
 ```
 
@@ -479,4 +485,10 @@ fit.sl2 <- recombineSL(fit.sl, Y = Y, method = "method.CC_LS")
 fit.sl2$coef
 fit.sl4 <- recombineSL(fit.sl, Y = Y, method = "method.CC_nloglik")
 fit.sl4$coef
-```
+```
+
+```{r, cache=TRUE, echo = TRUE}
+saveRDS(TE1, file = "data/gcompxg.RDS")
+saveRDS(TE2, file = "data/gcompls.RDS")
+saveRDS(TE3, file = "data/gcompsl.RDS")
+```

3ipw.Rmd

@@ -2,7 +2,7 @@
 
 In this chapter, we are not only interested in outcome modelling, but also **exposure modelling**.
 
-```{r setup01, include=FALSE}
+```{r setup01i, include=FALSE}
 require(knitr)
 require(glmnet)
 require(kableExtra)
@@ -147,3 +147,6 @@ publish(out.fit)
 ```
 
 
+```{r, cache=TRUE, echo = TRUE}
+saveRDS(out.fit, file = "data/ipw.RDS")
+```

3ipw2.Rmd

@@ -2,7 +2,7 @@
 
 Similar to G-computation, we will try to use machine learning methods, particularly Superlearner in estimating IPW estimates
 
-```{r ipw2setup01, include=FALSE}
+```{r ipw2setup01ic, include=FALSE}
 require(knitr)
 require(glmnet)
 require(kableExtra)
@@ -162,3 +162,6 @@ out.fit <- glm(out.formula,
 publish(out.fit)
 ```
 
+```{r, cache=TRUE, echo = TRUE}
+saveRDS(out.fit, file = "data/ipwsl.RDS")
+```

4tmle.Rmd

@@ -1,6 +1,6 @@
 # TMLE
 
-```{r tmlesetup01, include=FALSE}
+```{r tmlesetup01t, include=FALSE}
 require(knitr)
 require(glmnet)
 require(kableExtra)
@@ -449,13 +449,15 @@ ci.estimate <- function(data = ObsData, H.AL.components = 1){
 ### $\hat\epsilon$ = $\hat\epsilon_0$ and $\hat\epsilon_1$ 
 
 ```{r tmleinf2b, cache=TRUE}
-ci.estimate(data = ObsData, H.AL.components = 2)
+CI2 <- ci.estimate(data = ObsData, H.AL.components = 2)
+CI2
 ```
 
 ### Only 1 $\hat\epsilon$
 
 ```{r tmleinf2c, cache=TRUE}
-ci.estimate(data = ObsData, H.AL.components = 1)
+CI1 <- ci.estimate(data = ObsData, H.AL.components = 1)
+CI1
 ```
 
 ```{r tmleinf, cache=TRUE, include = FALSE}
@@ -477,3 +479,8 @@ ci.estimate(data = ObsData, H.AL.components = 1)
 # ATE_TMLE1_CI <- c(ATE_TMLE1 - 1.96*sqrt(varHat.IC), ATE_TMLE1 + 1.96*sqrt(varHat.IC))
 # cat("ATE: ", ATE_TMLE1, "  (", ATE_TMLE1_CI[1], ", ", ATE_TMLE1_CI[2], ")", sep = "")
 ```
+
+```{r, cache=TRUE, echo = TRUE}
+saveRDS(ATE.TMLE1, file = "data/tmlepointh.RDS")
+saveRDS(CI2, file = "data/tmlecih.RDS")
+```