1019_Homework Exercise 2
Pei-Yun,Li
2020-10-26
1 Description
Patients were randomized to receive one of two daily doses (low or high) of a new treatment for Alzheimer’s disease or a placebo in a study.
Each patient returned to the clinic every 2 months for 1 year for assessment of disease progression based on cognitive measurements on the Alzheimer’s Disease Assessment Scale (ADAS).
This test evaluates memory, language, and praxis function, and is based on the sum of scores from an 11-item scale, with a potential range of 0 to 70, higher scores indicative of greater disease severity.
The primary goal is to determine if the rate of disease progression is slowed with active treatment compared with a placebo.
Is there a difference in response profiles over time among the three groups?
Column 1: Treatment ID (L = Low dose, H = High dose, P = Placebo)
Column 2: Patient ID
Column 3: ADAS score at month 2
Column 4: ADAS score at month 4
Column 5: ADAS score at month 6
Column 6: ADAS score at month 8
Column 7: ADAS score at month 10
Column 8: ADAS score at month 12
2 Input Data
dta2 <- read.table("C:/Users/HANK/Desktop/HOMEWORK/adas.txt", header=T, na.strings='.')
dta2 <- mutate(dta2,
Treatment = factor(Treatment),
PID = factor(PID),
Baseline = adas02)
colnames(dta2)<-c("Treatment", "PID", "2", "4", "6", "8", "10" , "12", "Baseline")
str(dta2)## 'data.frame': 80 obs. of 9 variables:
## $ Treatment: Factor w/ 3 levels "H","L","P": 2 2 2 2 2 2 2 2 2 2 ...
## $ PID : Factor w/ 80 levels "1","2","3","4",..: 1 5 8 12 13 15 19 21 24 28 ...
## $ 2 : int 22 34 40 24 29 31 22 43 18 25 ...
## $ 4 : int 30 35 41 NA 26 36 27 49 28 24 ...
## $ 6 : int NA 46 41 21 29 41 28 42 29 27 ...
## $ 8 : int 33 37 46 28 26 46 24 48 NA 18 ...
## $ 10 : int 28 31 52 30 NA 52 27 48 25 21 ...
## $ 12 : int 30 35 48 27 36 57 28 46 28 22 ...
## $ Baseline : int 22 34 40 24 29 31 22 43 18 25 ...## Treatment PID 2 4 6 8 10 12 Baseline
## 1 L 1 22 30 NA 33 28 30 22
## 2 L 5 34 35 46 37 31 35 34
## 3 L 8 40 41 41 46 52 48 40
## 4 L 12 24 NA 21 28 30 27 24
## 5 L 13 29 26 29 26 NA 36 29
## 6 L 15 31 36 41 46 52 57 313 Wide to long format
dta2L <- gather(dta2, Month, ADAS, "2":"12")
dta2L <- mutate(dta2L,
Baseline = as.numeric(Baseline),
ADAS = as.numeric(ADAS),
Time_f = factor(as.numeric(Month)-2),
Month = factor (Month))
str(dta2L)## 'data.frame': 480 obs. of 6 variables:
## $ Treatment: Factor w/ 3 levels "H","L","P": 2 2 2 2 2 2 2 2 2 2 ...
## $ PID : Factor w/ 80 levels "1","2","3","4",..: 1 5 8 12 13 15 19 21 24 28 ...
## $ Baseline : num 22 34 40 24 29 31 22 43 18 25 ...
## $ Month : Factor w/ 6 levels "10","12","2",..: 3 3 3 3 3 3 3 3 3 3 ...
## $ ADAS : num 22 34 40 24 29 31 22 43 18 25 ...
## $ Time_f : Factor w/ 6 levels "0","2","4","6",..: 1 1 1 1 1 1 1 1 1 1 ...## Treatment PID Baseline Month ADAS Time_f
## 1 L 1 22 2 22 0
## 2 L 5 34 2 34 0
## 3 L 8 40 2 40 0
## 4 L 12 24 2 24 0
## 5 L 13 29 2 29 0
## 6 L 15 31 2 31 04 plot
p <- ggplot(dta2L, aes(Time_f, ADAS,
group = Treatment,
linetype = Treatment,
shape = Treatment)) +
stat_summary(fun = mean, geom = "line") +
stat_summary(fun = mean, geom = "point") +
stat_summary(fun.data = mean_se, geom = "errorbar", width = 0.2) +
scale_shape_manual(values=c(1, 2, 7)) +
scale_y_continuous(breaks=seq(20, 60, 10))+
labs(x = "Month", y = "ADAS", linetype = "Treatment", shape = "Treatment") +
theme_minimal() +
theme(legend.position=c(.15,.85))
suppressWarnings(suppressMessages(print(p)))
dta2w <- read.table("C:/Users/HANK/Desktop/HOMEWORK/adas.txt", h=T, na.strings='.')
dta2w <- dta2w %>%
mutate(Baseline= adas02)
dta2w %>%
dplyr::group_by(Treatment) %>%
dplyr::select(starts_with("adas")) %>%
furniture::table1(digits=2, total=FALSE, test=F, output="html")## Adding missing grouping variables: `Treatment`## Using dplyr::group_by() groups: Treatment| H | L | P | |
|---|---|---|---|
| n = 18 | n = 17 | n = 21 | |
| adas02 | |||
| 30.22 (8.56) | 32.82 (7.51) | 29.62 (6.78) | |
| adas04 | |||
| 32.44 (7.33) | 35.12 (8.64) | 33.43 (6.42) | |
| adas06 | |||
| 33.56 (8.56) | 37.65 (9.09) | 34.86 (6.43) | |
| adas08 | |||
| 33.33 (9.06) | 36.41 (11.09) | 36.38 (6.34) | |
| adas10 | |||
| 33.72 (7.44) | 38.35 (11.19) | 37.57 (7.03) | |
| adas12 | |||
| 34.28 (7.10) | 39.18 (10.77) | 39.05 (8.80) |
5 Covariance and correlation
| adas02 | adas04 | adas06 | adas08 | adas10 | adas12 | |
|---|---|---|---|---|---|---|
| adas02 | 1.0000000 | 0.8954000 | 0.7544220 | 0.7740050 | 0.7006653 | 0.7496129 |
| adas04 | 0.8954000 | 1.0000000 | 0.8144181 | 0.7936916 | 0.7006978 | 0.7400989 |
| adas06 | 0.7544220 | 0.8144181 | 1.0000000 | 0.8628966 | 0.6716075 | 0.7615946 |
| adas08 | 0.7740050 | 0.7936916 | 0.8628966 | 1.0000000 | 0.8472388 | 0.8337049 |
| adas10 | 0.7006653 | 0.7006978 | 0.6716075 | 0.8472388 | 1.0000000 | 0.9153379 |
| adas12 | 0.7496129 | 0.7400989 | 0.7615946 | 0.8337049 | 0.9153379 | 1.0000000 |
scatterplotMatrix(~ adas02 + adas04 + adas06 + adas08 + adas10 + adas12 | Treatment,
data=dta2w,
col="gray",
smooth=FALSE,
by.group = TRUE,
regLine=FALSE,
ellipse=list(levels=c(.975),
fill=TRUE,
fill.alpha=.1),
diagonal = T,
pch = c(1, 19, 21))## Warning in scatterplotMatrix.default(X[, -ncol], groups = X[, ncol], ...): number of groups exceeds number of available colors
## colors are recycled
## Warning: Missing values for following ID(s):
## 1, 7, 11, 12, 13, 24, 26, 27, 30, 31, 35, 41, 42, 45, 47, 48, 51, 52, 54, 61, 68, 70, 72, 74
## Removing those cases from the analysis.## Contrasts set to contr.sum for the following variables: Treatment| Effect | df | MSE | F | ges | p.value |
|---|---|---|---|---|---|
| Treatment | 2, 53 | 328.01 | 1.10 | .032 | .342 |
| Time_f | 3.27, 173.16 | 25.22 | 18.57 *** | .066 | <.001 |
| Treatment:Time_f | 6.53, 173.16 | 25.22 | 1.35 | .010 | .234 |
## Treatment = H:
## Time_f emmean SE df lower.CL upper.CL
## X0 30.2 1.93 82.6 26.4 34.1
## X2 32.4 1.93 82.6 28.6 36.3
## X4 33.5 1.93 82.6 29.7 37.4
## X6 33.3 1.93 82.6 29.5 37.2
## X8 33.7 1.93 82.6 29.9 37.6
## X10 34.3 1.93 82.6 30.4 38.1
##
## Treatment = L:
## Time_f emmean SE df lower.CL upper.CL
## X0 32.8 1.96 83.8 28.9 36.7
## X2 35.1 1.96 83.8 31.2 39.0
## X4 37.6 1.96 83.8 33.7 41.5
## X6 36.4 1.96 83.8 32.5 40.3
## X8 38.3 1.96 83.8 34.4 42.2
## X10 39.2 1.96 83.8 35.3 43.1
##
## Treatment = P:
## Time_f emmean SE df lower.CL upper.CL
## X0 29.6 1.87 79.5 25.9 33.3
## X2 33.4 1.87 79.5 29.7 37.1
## X4 34.8 1.87 79.5 31.1 38.6
## X6 36.4 1.87 79.5 32.7 40.1
## X8 37.6 1.87 79.5 33.8 41.3
## X10 39.0 1.87 79.5 35.3 42.7
##
## Warning: EMMs are biased unless design is perfectly balanced
## Confidence level used: 0.95xyplot(ADAS ~ Baseline | as.factor(Time_f), groups=Treatment, data=dta2L,
type=c("p","r","g"), layout=c(6,1),
xlab="Baseline ADAS",
ylab="ADAS",
auto.key = TRUE)
dta2L$trt_time <- dta2L$Time_f
dta2L$trt_time[dta2L$Treatment=="P"] <- "0"
dta2L$Time_f <- relevel(dta2L$Time_f, ref="0")
dta2L$trt_time <- relevel(dta2L$trt_time, ref="0")
str(dta2L)## 'data.frame': 480 obs. of 7 variables:
## $ Treatment: Factor w/ 3 levels "H","L","P": 2 2 2 2 2 2 2 2 2 2 ...
## $ PID : Factor w/ 80 levels "1","2","3","4",..: 1 5 8 12 13 15 19 21 24 28 ...
## $ Baseline : num 22 34 40 24 29 31 22 43 18 25 ...
## $ Month : Factor w/ 6 levels "10","12","2",..: 3 3 3 3 3 3 3 3 3 3 ...
## $ ADAS : num 22 34 40 24 29 31 22 43 18 25 ...
## $ Time_f : Factor w/ 6 levels "0","2","4","6",..: 1 1 1 1 1 1 1 1 1 1 ...
## $ trt_time : Factor w/ 6 levels "0","2","4","6",..: 1 1 1 1 1 1 1 1 1 1 ...