-
Notifications
You must be signed in to change notification settings - Fork 1
/
funnel-test.R
187 lines (136 loc) · 5.25 KB
/
funnel-test.R
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
# test if I have to pass everything over? YES I DO
test=function(x,y){sum(x)/y}
y=2
test(x=c(1,2,3),y)
####################################################
# March 2016. Test new funnel program
####################################################
# Lung data
x<-read.csv(file="~//Dropbox//DB-funnels-HSMR//funnels//lung.csv",header=T,sep=",")
N<- x$Cases
R<- N-x$deaths
xlabel<-"Number of operations per hospital"
scale<-0.8
xrange<-c(0,max(N))
xrange=c(0,600)
# ylabel<-"Mortality rate (%)"
#yrange<-c(0,max( R/N ))
ylabel<-"Survival rate (%)"
yrange<-c(0.94, 1)
#yrange=c(0,1)
names= as.character(x$X)
title<-"LCCOP data"
tails=c(0.001, 0.025)
source(file="~//Dropbox//DB-Rfunctions//funnel4.R")
funnel4(obs.prop=R/N, denom=N, pred.prop=P/N, names=names,
plot="slice", rank="precision", riskadj=F, RASRplot=F,
plot.target=F, title=title,xrange=xrange,
yrange=yrange, tails=tails,xlab=xlabel,ylab=ylabel,ypercent=T,
bandcols=c("white","lightcyan","cyan")
) #
####################################################
# Feb 2016. Test new funnel program
####################################################
# New York data using new function
x<-read.csv(file="~//Dropbox//DB-funnels-HSMR//funnels//CABG-hospitals-03.csv",header=T,sep=",")
N<- x$Cases
R<- N-x$Deaths
P = N -x$EMR*N/100
xlabel<-"Number of operations per hospital"
xlabel<-"Number of operations per hospital (adjusted)"
xrange<-c(0,max(N))
# ylabel<-"Mortality rate (%)"
#yrange<-c(0,max( R/N ))
yrange<-c(min(R/N )-0.01, 1)
#yrange=c(0,1)
names= as.character(x$Hospital)
tails=c(0.001,0.025)
source(file="~//Dropbox//DB-Rfunctions//funnel4.R")
# test using slices
# 1. not risk-adjusted
source(file="~//Dropbox//DB-Rfunctions//funnel4.R")
title<-"NY Cardiac Surgery - not risk-adjusted"
ylabel<-"Survival rate (%)"
funnel4(obs.prop=R/N, denom=N, pred.prop=P/N, names=names,
plot="funnel", rank="precision", riskadj=F, RASRplot=F,
plot.target=F, title=title,xrange=xrange,
yrange=yrange, tails=tails,xlab=xlabel,ylab=ylabel,ypercent=T,
bandcols=c("white","cyan","cyan3")
) #
# 2. risk-adjusted, using shifted slices
source(file="~//Dropbox//DB-Rfunctions//funnel4.R")
title<-"NY Cardiac Surgery - risk-adjusted"
ylabel<-"Survival rate (%)"
funnel4(obs.prop=R/N, denom=N, pred.prop=P/N, names=names,
plot="slice", rank="precision", riskadj=T, RASRplot=F,
mean.target=F,plot.target=F, title=title,xrange=xrange,
yrange=yrange, tails=tails,xlab=xlabel,ylab=ylabel,ypercent=T,
bandcols=c("white","cyan","cyan3")
) #
# 3. risk-adjusted, using RASR
source(file="~//Dropbox//DB-Rfunctions//funnel4.R")
title<-"NY Cardiac Surgery - risk-adjusted"
ylabel<-"Survival rate (%)"
funnel4(obs.prop=R/N, denom=N, pred.prop=P/N, names=names,
plot="funnel", rank="precision", riskadj=T, RASRplot=T,
mean.target=F,plot.target=F, title=title,xrange=xrange,
yrange=yrange, tails=tails,xlab=xlabel,ylab=ylabel,ypercent=T,
bandcols=c("white","cyan","cyan3")
) #
############################################################
# New York data using older function
x<-read.csv(file="~//Dropbox//DB-funnels-HSMR//funnels//CABG-hospitals-03.csv",header=T,sep=",")
N<- x$Cases
R<- N-x$Deaths
P = N -x$EMR*N/100
xlabel<-"Number of operations per hospital"
scale<-0.8
xrange<-c(0,max(N))
# ylabel<-"Mortality rate (%)"
#yrange<-c(0,max( R/N ))
ylabel<-"Survival rate (%)"
yrange<-c(min(R/N )-0.01, 1)
#yrange=c(0,1)
names= as.character(x$Hospital)
title<-"NY Cardiac Surgery - risk-adjusted"
source(file="~//Dropbox//DB-Rfunctions//funnel-3.R")
funnel.3(datatype="prop", obs.prop=R/N,denom=N, pred.prop=P/N, names=names,
frame="pos", slice=T, rank="outcome", riskadj=T, RASRplot=F,
normapprox=F, logtrans=F,
plot.target=F,
random.effects=0,tau=NA,
title=title,scale=scale,xrange=xrange,
yrange=yrange, xlab=xlabel,ylab=ylabel,symbolstretch=scale,
slicecols=c("pink","orange","white") ) #
par(mar=c(5,5,5,5))
par(mgp=c(3,1,1))
plot(c(1,0),c(0,1),axes=F,xlab="XXX")
# can I have non-integer binomial? NO! Ah, not so good
pbinom(6, 10.5,0.5)
####################################################################
# Robin's data
source(file="~//Dropbox//DB-Rfunctions//funnel-3.R")
# reading in data
x<-read.csv(file="~//Dropbox//DB-funnels-HSMR//funnels//Robin-EX1.txt",header=T,sep=",")
summary(x)
attach(x)
R=Dead
N=Total
min(N)
ylabel<-"Mortality rate (%)"
xlabel<-"Number of operations per surgeon"
yrange<-c(0,max(R/N ))
title<-""
xrange<-c(0,max(N))
scale<-1
# png("~//Dropbox//DB-articles//funnels//NYhospitals.png",width=4000,height=3000, res=600)
funnel.3(datatype="prop",
a1=R,b1=N,ratedenom=NA,normapprox=0,logtrans=0,meantarget=1,target=NA,
overdisp = 0, winsor = 10, phisig = 1,
random.effects=0,tau=NA,title=title,scale=scale,xrange=xrange,
yrange=yrange,pvalue=c(0.001,0.025),
Npoints=200,xlab=xlabel,ylab=ylabel,plottext=0,pointsymbol=16,symbolstretch=scale,
limittype=c(5,6) ,legend=1,ylogscale=0,ypercent=1) #
#dev.off()
# issues with this - most 0, a few extreme, need summaries in list
# eg p-values, q-values etc.