notebook.community

Edit and run 





In [1]:



    


library(gdata)
library(data.table)



    


















    






gdata: read.xls support for 'XLS' (Excel 97-2004) files ENABLED.

gdata: read.xls support for 'XLSX' (Excel 2007+) files ENABLED.

Attaching package: ‘gdata’

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

    nobs

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

    object.size

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

    startsWith


Attaching package: ‘data.table’

The following objects are masked from ‘package:gdata’:

    first, last


















In [2]:



    


data <- read.table("/media/igna/Elements/HotelDieu/Cochrane/MappingRCTs_vs_Burden/database_RCTs_regions_27diseases.txt")
nrow(data)



    


















    






117180
















In [3]:



    


table(is.na(data$Sample))
sum(data$Sample,na.rm=TRUE)
max(data$Sample,na.rm=TRUE)



    


















    








 FALSE   TRUE 
109211   7969 










    






77618020









    






20121212
















In [4]:



    


#Shrinking sample sizes
table(data$Sample < 10)
table(data$Sample > 200000)
sum(data$Sample[data$Sample>=10 & data$Sample<=200000],na.rm=TRUE)
table(is.na(data$Sample) | data$Sample<10 | data$Sample > 200000)



    


















    








 FALSE   TRUE 
107582   1629 










    








 FALSE   TRUE 
109187     24 










    






43952638









    








 FALSE   TRUE 
107558   9622 

















In [5]:



    


data <- data[!is.na(data$Sample) & data$Sample<=200000 & data$Sample>=10,]



    











In [6]:



    


#Distribution sample sizes
summary(data$Sample)



    


















    








    Min.  1st Qu.   Median     Mean  3rd Qu.     Max. 
    10.0     48.0    100.0    408.6    220.0 200000.0 

















In [7]:



    


spl_qt <- c(0,10,20,40,60,100,200,400,1000,2000,10000,20000,100000,200000)

data$Sple_cl <- cut(data$Sample,spl_qt,right=F)
data$Sple_cl <- as.character(data$Sple_cl)
data$Sple_cl <- as.factor(data$Sple_cl)
levels(data$Sple_cl)



    


















    







    
	
  1. '[10,20)'
    2. 
	
  2. '[100,200)'
    3. 
	
  3. '[1e+03,2e+03)'
    4. 
	
  4. '[1e+04,2e+04)'
    5. 
	
  5. '[1e+05,2e+05)'
    6. 
	
  6. '[20,40)'
    7. 
	
  7. '[200,400)'
    8. 
	
  8. '[2e+03,1e+04)'
    9. 
	
  9. '[2e+04,1e+05)'
    10. 
	
  10. '[40,60)'
    11. 
	
  11. '[400,1e+03)'
    12. 
	
  12. '[60,100)'
    13. 



















In [8]:



    


levels(data$Sple_cl) <- c(levels(data$Sple_cl),'[0,10)')



    











In [9]:



    


table(data$Sple_cl)



    


















    








      [10,20)     [100,200) [1e+03,2e+03) [1e+04,2e+04) [1e+05,2e+05) 
         4598         22308          2861           270            31 
      [20,40)     [200,400) [2e+03,1e+04) [2e+04,1e+05)       [40,60) 
        14304         15898          1968           212         14124 
  [400,1e+03)      [60,100)        [0,10) 
        10348         20633             0 

















In [10]:



    


data$Sple_cl <- reorder(data$Sple_cl,new.order=c('[0,10)',
                                                 '[10,20)',
                                                 '[20,40)',
                                                 '[40,60)',
                                                 '[60,100)',
                                                 '[100,200)',
                                                 '[200,400)',
                                                 '[400,1e+03)',
                                                 '[1e+03,2e+03)',
                                                 '[2e+03,1e+04)',
                                                 '[1e+04,2e+04)',
                                                 '[2e+04,1e+05)',
                                                 '[1e+05,2e+05)'
                                                 ))
levels(data$Sple_cl) <- c('[0,10)',
                         '[10,20)',
                         '[20,40)',
                         '[40,60)',
                         '[60,100)',
                         '[100,200)',
                         '[200,400)',
                         '[400,1e3)',
                         '[1e3,2e3)',
                         '[2e3,1e4)',
                         '[1e4,2e4)',
                         '[2e4,1e5)',
                         '[1e5,2e5)'
                        )



    











In [11]:



    


table(data$Sple_cl)



    


















    








   [0,10)   [10,20)   [20,40)   [40,60)  [60,100) [100,200) [200,400) [400,1e3) 
        0      4598     14304     14124     20633     22308     15898     10348 
[1e3,2e3) [2e3,1e4) [1e4,2e4) [2e4,1e5) [1e5,2e5) 
     2861      1968       270       212        31 

















In [12]:



    


regs <- sort(unique(unlist(strsplit(as.character(data$Regions),"&"))))
LR <- lapply(regs,function(x){1:nrow(data)%in%grep(x,data$Regions)})
LR <- do.call('cbind',LR)
LR <- data.table(LR)
LR$TrialID <- data$TrialID

#Nb of patients per region per trial
#Nb countries per region per trial to distribute sample size equally across countries
nb_ctrs <- lapply(strsplit(as.character(data$Nb_ctr_per_reg),'&'),as.numeric)
RGs <-strsplit(as.character(data$Regions),'&')
pats <- data.frame(TrialID = rep(data$TrialID,sapply(nb_ctrs,length)),
                   Nb_ctrs = unlist(nb_ctrs),
                   Region = unlist(RGs),
                   Tot_sample = rep(data$Sample,sapply(nb_ctrs,length)))

pats$tot_ctrs <- rep(sapply(nb_ctrs,sum),sapply(nb_ctrs,length))
pats$sample_per_reg <- pats$Tot_sample*pats$Nb_ctrs/pats$tot_ctrs
pats <- data.table(pats)



    











In [13]:



    


pats$Sple_cl <- cut(pats$sample_per_reg,spl_qt,right=F)
pats$Sple_cl <- as.character(pats$Sple_cl)
pats$Sple_cl <- as.factor(pats$Sple_cl)

levels(pats$Sple_cl)



    


















    







    
	
  1. '[0,10)'
    2. 
	
  2. '[10,20)'
    3. 
	
  3. '[100,200)'
    4. 
	
  4. '[1e+03,2e+03)'
    5. 
	
  5. '[1e+04,2e+04)'
    6. 
	
  6. '[1e+05,2e+05)'
    7. 
	
  7. '[20,40)'
    8. 
	
  8. '[200,400)'
    9. 
	
  9. '[2e+03,1e+04)'
    10. 
	
  10. '[2e+04,1e+05)'
    11. 
	
  11. '[40,60)'
    12. 
	
  12. '[400,1e+03)'
    13. 
	
  13. '[60,100)'
    14. 



















In [14]:



    


pats$Sple_cl <- reorder(pats$Sple_cl,new.order=c('[0,10)',
                                                 '[10,20)',
                                                 '[20,40)',
                                                 '[40,60)',
                                                 '[60,100)',
                                                 '[100,200)',
                                                 '[200,400)',
                                                 '[400,1e+03)',
                                                 '[1e+03,2e+03)',
                                                 '[2e+03,1e+04)',
                                                 '[1e+04,2e+04)',
                                                 '[2e+04,1e+05)',
                                                 '[1e+05,2e+05)'
                                                 ))
levels(pats$Sple_cl) <- c('[0,10)',
                         '[10,20)',
                         '[20,40)',
                         '[40,60)',
                         '[60,100)',
                         '[100,200)',
                         '[200,400)',
                         '[400,1e3)',
                         '[1e3,2e3)',
                         '[2e3,1e4)',
                         '[1e4,2e4)',
                         '[2e4,1e5)',
                         '[1e5,2e5)'
                        )



    











In [15]:



    


regs <- names(sort(table(pats$Region),decreasing=TRUE))
regs



    


















    







    
	
  1. 'High-income'
    2. 
	
  2. 'North Africa and Middle East'
    3. 
	
  3. 'Southeast Asia, East Asia and Oceania'
    4. 
	
  4. 'Central Europe, Eastern Europe, and Central Asia'
    5. 
	
  5. 'South Asia'
    6. 
	
  6. 'Latin America and Caribbean'
    7. 
	
  7. 'Sub-Saharian Africa'
    8. 



















In [16]:



    


reg_labs <- c('High-income',
              'North Africa\nand Middle East',
              'Southeast and East Asia\nand Oceania',
              'Eastern Europe\nand Central Asia',
              'South Asia',
              'Latin America\nand Caribbean',
              'Sub-Saharian Africa')



    











In [17]:



    


options(repr.plot.width=10, repr.plot.height=5)
par(mfrow=c(2,4),mar=c(5,4.9,3,0),las=2)

barplot(100*table(data$Sple_cl)/nrow(data),ylim=c(0,35),main="Worldwide",cex.names=0.8)
mtext("Distribution of RCTs\nacross sample sizes (%)",side=2,line=2.5,las=3,cex=0.8)

#barplot(table(pats$Sple_cl)/nrow(pats),ylim=c(0,0.5))

for(r in 1:length(regs)){barplot(100*table(pats$Sple_cl[pats$Region==regs[r]])/sum(pats$Region==regs[r]),
                                                   ylim=c(0,35),main=reg_labs[r],cex.names=0.8)
                    if(r==4) mtext("Distribution of RCTs\nacross sample sizes (%)",side=2,line=2.5,las=3,cex=0.8)
                        }



    


















    
























In [18]:



    


pdf("../Figures/Sample_size_distribution_within_regions.pdf",width=10,height=5)
options(repr.plot.width=10, repr.plot.height=5)
par(mfrow=c(2,4),mar=c(5,4.9,3,0),las=2)

barplot(100*table(data$Sple_cl)/nrow(data),ylim=c(0,35),main="Worldwide",cex.names=0.8)
mtext("Distribution of RCTs\nacross sample sizes (%)",side=2,line=2.5,las=3,cex=0.8)

#barplot(table(pats$Sple_cl)/nrow(pats),ylim=c(0,0.5))

for(r in 1:length(regs)){barplot(100*table(pats$Sple_cl[pats$Region==regs[r]])/sum(pats$Region==regs[r]),
                                                   ylim=c(0,35),main=reg_labs[r],cex.names=0.8)
                    if(r==4) mtext("Distribution of RCTs\nacross sample sizes (%)",side=2,line=2.5,las=3,cex=0.8)
                        }
dev.off()



    


















    






PNG: 2
















In [19]:



    


head(pats)



    


















    









TrialIDNb_ctrsRegionTot_sampletot_ctrssample_per_regSple_cl


	
1EUCTR2007-006175-36-FI3                     High-income           105                   3                     105                   [100,200)             

	
2NCT005901741          High-income1175       1          1175       [1e3,2e3)  

	
3NCT005903561          High-income532        1          532        [400,1e3)  

	
4NCT005914871          High-income60         1          60         [60,100)   

	
5NCT010725261          High-income20         1          20         [20,40)    

	
6NCT005899661          High-income62         1          62         [60,100)   





















In [26]:



    


hist(log10(pats$sample_per_reg),breaks=log10(c(0.1,10,50,100,200,500,1000,5000,10000,20000,100000,200000)))



    


















    
























In [ ]:



    






    











In [ ]:



    






    











In [ ]:



    






    











In [ ]:

Content source: iatal/RCTvsBurden

Similar notebooks:

notebook.community | gallery | about