This reads in the NSQIP data for all shunt cases and all peds cases from 2019.

This finds all the unique cases of shunt placement and merges with record data. Then, we create our final working dataframes that only contains the factors of interest.

df0 <- merge(df_raw,df_shunt[,2:13], by="CaseID")

df <- df0[,c(1,3:6,14,16,22,94,126:130,196:199,348:358)]
df_first <- df[which(df$PED_SHNT_1STPERM == "Yes"),]
df_revise <- df[which(df$PED_SHNT_1STPERM == "No"),]

#rem null
#df_rem_NULL <- df[!df$ETHNICITY_HISPANIC == "NULL", ]
#df_rem_NULL <- df[!df$DELIVERY_MODE == "NULL", ]
df_rem_NULL <- df[!df$PREM_BIRTH == "Unknown", ]

below: closer examination of premature birth influence on shunt revision

df_premature0 <- df_rem_NULL[,c(8,19)]

premature <- as.data.frame(table(df_premature0),stringsAsFactors=FALSE)
prem_no <- premature[1:9,]
prem_yes <- premature[10:18,]

df_premature <- prem_no
df_premature$first_shunt <- prem_yes$Freq
df_premature$multi_shunt <- prem_no$Freq
df_premature <- df_premature[,-c(2,3)]
df_premature$proportion_revised <- (df_premature$multi_shunt/(df_premature$multi_shunt+df_premature$first_shunt))

df_premature[8,1] <- "24 and fewer weeks gestation"

ggplot of premature birth versus shunt revisions

ggplot(data=df_premature, aes(x=PREM_BIRTH, y=proportion_revised,fill=proportion_revised))+geom_bar(stat="identity")+
  scale_x_discrete(labels=c("<24 weeks","24 weeks","25-26 weeks","27-28 weeks","29-30 weeks","31-32 weeks","33-34 weeks","35-26","Full Term")) + 
  gghisto +theme(axis.text.x = element_text(size=10, angle=15)) + scale_y_continuous(labels = scales::percent)+scale_fill_gradient(low="mistyrose",high="lightcoral") +
  theme(legend.position="none") + xlab("Weeks Gestation") + ylab("Proportion reporting \nshunt revision") +ggtitle("Proportions of Premature Birth \nCategories Reporting Shunt Revisions ")

prem_table <- table(df_rem_NULL$PREM_BIRTH, df_rem_NULL$PED_SHNT_1STPERM)
chisq.test(df_rem_NULL$PREM_BIRTH, df_rem_NULL$PED_SHNT_1STPERM, correct=FALSE)
## 
##  Pearson's Chi-squared test
## 
## data:  df_rem_NULL$PREM_BIRTH and df_rem_NULL$PED_SHNT_1STPERM
## X-squared = 34.828, df = 8, p-value = 2.874e-05
library(chisq.posthoc.test)
chisq.posthoc.test(prem_table)
##                                 Dimension     Value         No        Yes
## 1            24 completed weeks gestation Residuals  2.2982125 -2.2982125
## 2            24 completed weeks gestation  p values  0.3878950  0.3878950
## 3         25-26 completed weeks gestation Residuals  3.1190716 -3.1190716
## 4         25-26 completed weeks gestation  p values  0.0326560  0.0326560
## 5         27-28 completed weeks gestation Residuals  1.2547522 -1.2547522
## 6         27-28 completed weeks gestation  p values  1.0000000  1.0000000
## 7         29-30 completed weeks gestation Residuals -1.4604842  1.4604842
## 8         29-30 completed weeks gestation  p values  1.0000000  1.0000000
## 9         31-32 completed weeks gestation Residuals -0.8819048  0.8819048
## 10        31-32 completed weeks gestation  p values  1.0000000  1.0000000
## 11        33-34 completed weeks gestation Residuals -2.7798978  2.7798978
## 12        33-34 completed weeks gestation  p values  0.0978770  0.0978770
## 13        35-36 completed weeks gestation Residuals -1.8116193  1.8116193
## 14        35-36 completed weeks gestation  p values  1.0000000  1.0000000
## 15 Less than 24 completed weeks gestation Residuals  2.4482438 -2.4482438
## 16 Less than 24 completed weeks gestation  p values  0.2583980  0.2583980
## 17                                     No Residuals -1.0753303  1.0753303
## 18                                     No  p values  1.0000000  1.0000000

preliminary findings: mode of birth predicts needing a second shunt. days from admissions to operation are related to shunt revision. premature birth signficantly related to first or second shunt.

#summary of ALL shunts   
tbl_summary(df[,c(2:21)],label = `PED_SHNT_1STPERM` ~ "First Shunt",by=`PED_SHNT_1STPERM`)%>%
  bold_labels()  %>%add_p()
## There was an error in 'add_p()/add_difference()' for variable 'RACE', p-value omitted:
## Error in stats::fisher.test(c("White", "Unknown/Not Reported", "White", : FEXACT error 7(location). LDSTP=18240 is too small for this problem,
##   (pastp=259.676, ipn_0:=ipoin[itp=489]=515, stp[ipn_0]=259.449).
## Increase workspace or consider using 'simulate.p.value=TRUE'
## There was an error in 'add_p()/add_difference()' for variable 'TRANST', p-value omitted:
## Error in stats::fisher.test(c("Admitted from home/clinic/doctor's office", : FEXACT error 7(location). LDSTP=18240 is too small for this problem,
##   (pastp=324.318, ipn_0:=ipoin[itp=379]=786, stp[ipn_0]=321.999).
## Increase workspace or consider using 'simulate.p.value=TRUE'
## There was an error in 'add_p()/add_difference()' for variable 'DISCHDEST', p-value omitted:
## Error in stats::fisher.test(c("NULL", "Home", "Home", "Home", "Home", : FEXACT error 7(location). LDSTP=18240 is too small for this problem,
##   (pastp=35.1155, ipn_0:=ipoin[itp=126]=57, stp[ipn_0]=29.4798).
## Increase workspace or consider using 'simulate.p.value=TRUE'
## There was an error in 'add_p()/add_difference()' for variable 'PED_SHNT_HC_ETIOL', p-value omitted:
## Error in stats::fisher.test(c("Intraventricular hemorrhage (IVH) of prematurity without other cause", : FEXACT error 7(location). LDSTP=18210 is too small for this problem,
##   (pastp=96.8752, ipn_0:=ipoin[itp=122]=10, stp[ipn_0]=40.3951).
## Increase workspace or consider using 'simulate.p.value=TRUE'
Characteristic No, N = 5591 Yes, N = 1,7391 p-value2
AGE_DAYS 942 (233, 3,600) 131 (50, 598) <0.001
SEX 0.8
Female 248 (44%) 780 (45%)
Male 311 (56%) 959 (55%)
RACE
American Indian or Alaska Native 6 (1.1%) 8 (0.5%)
Asian 11 (2.0%) 43 (2.5%)
Black or African American 99 (18%) 353 (20%)
Native Hawaiian or Other Pacific Islander 5 (0.9%) 5 (0.3%)
Unknown/Not Reported 95 (17%) 317 (18%)
White 343 (61%) 1,013 (58%)
ETHNICITY_HISPANIC 0.11
No 397 (71%) 1,263 (73%)
NULL 47 (8.4%) 178 (10%)
Yes 115 (21%) 298 (17%)
TRANST
Admitted from home/clinic/doctor's office 269 (48%) 775 (45%)
Admitted through ER, including outside ER with direct hospital admission 234 (42%) 460 (26%)
Chronic care/Rehab/Intermediate Care/Spinal Cord 1 (0.2%) 7 (0.4%)
Other 8 (1.4%) 28 (1.6%)
Transferred from outside hospital (NICU, PICU, Inpatient on General floor, Adult ICU) 47 (8.4%) 469 (27%)
HtoODay 1 (0, 11) 2 (0, 15) <0.001
PREM_BIRTH <0.001
24 completed weeks gestation 32 (5.7%) 63 (3.6%)
25-26 completed weeks gestation 62 (11%) 125 (7.2%)
27-28 completed weeks gestation 39 (7.0%) 99 (5.7%)
29-30 completed weeks gestation 13 (2.3%) 64 (3.7%)
31-32 completed weeks gestation 16 (2.9%) 65 (3.7%)
33-34 completed weeks gestation 16 (2.9%) 104 (6.0%)
35-36 completed weeks gestation 36 (6.4%) 158 (9.1%)
Less than 24 completed weeks gestation 30 (5.4%) 56 (3.2%)
No 286 (51%) 958 (55%)
Unknown 29 (5.2%) 47 (2.7%)
DELIVERY_MODE <0.001
NULL 490 (88%) 1,018 (59%)
Scheduled C-Section 5 (0.9%) 179 (10%)
Unknown/Not Documented 3 (0.5%) 53 (3.0%)
Unscheduled C-Section 38 (6.8%) 279 (16%)
Vaginal Delivery 23 (4.1%) 210 (12%)
ANESURG 54 (42, 70) 52 (42, 66) 0.034
SURGANE 23 (16, 30) 23 (17, 32) 0.2
DPATRM 137 (110, 178) 126 (101, 155) <0.001
ANETIME 147 (118, 194) 138 (111, 168) <0.001
OPTIME 69 (47, 94) 56 (41, 75) <0.001
DOptoDis 2 (1, 4) 2 (1, 5) 0.3
DISCHDEST
Expired 3 (0.5%) 9 (0.5%)
Facility which was home 2 (0.4%) 7 (0.4%)
Home 478 (86%) 1,298 (75%)
NULL 57 (10%) 319 (18%)
Rehab 10 (1.8%) 20 (1.2%)
Separate acute care 2 (0.4%) 65 (3.7%)
Skilled care, not home 7 (1.3%) 19 (1.1%)
Unknown 0 (0%) 1 (<0.1%)
Unskilled facility, not home 0 (0%) 1 (<0.1%)
TOTHLOS 2 (1, 10) 2 (1, 9) 0.079
STILLINHOSP <0.001
NULL 502 (90%) 1,420 (82%)
Yes 57 (10%) 319 (18%)
PED_SHNT_HC_ETIOL
Congenital hydrocephalus 0 (0%) 525 (30%)
Intracranial cyst 0 (0%) 116 (6.7%)
Intraventricular hemorrhage (IVH) of prematurity without other cause 0 (0%) 373 (21%)
Neoplastic 0 (0%) 220 (13%)
NULL 559 (100%) 0 (0%)
Other - insufficient information to classify type 0 (0%) 155 (8.9%)
Post hemorrhagic from other vascular lesion 0 (0%) 58 (3.3%)
Post infectious 0 (0%) 77 (4.4%)
Post traumatic 0 (0%) 40 (2.3%)
Spina bifida with Chiari malformation / myelomeningocele 0 (0%) 118 (6.8%)
Syndromic, not otherwise specified 0 (0%) 38 (2.2%)
Vascular anomaly 0 (0%) 19 (1.1%)
PED_SHNT_PRIOR_HCPROC <0.001
No 0 (0%) 1,124 (65%)
NULL 559 (100%) 0 (0%)
Yes 0 (0%) 615 (35%)
1 Median (IQR); n (%)
2 Wilcoxon rank sum test; Pearson's Chi-squared test 
#summary removing specified null values
tbl_summary(df_rem_NULL[,c(2:21)],label = `PED_SHNT_1STPERM` ~ "First Shunt",by=`PED_SHNT_1STPERM`)%>%
  bold_labels()  %>%add_p()
## There was an error in 'add_p()/add_difference()' for variable 'RACE', p-value omitted:
## Error in stats::fisher.test(c("White", "White", "Unknown/Not Reported", : FEXACT error 7(location). LDSTP=18240 is too small for this problem,
##   (pastp=226.606, ipn_0:=ipoin[itp=244]=26, stp[ipn_0]=225.063).
## Increase workspace or consider using 'simulate.p.value=TRUE'
## There was an error in 'add_p()/add_difference()' for variable 'TRANST', p-value omitted:
## Error in stats::fisher.test(c("Admitted from home/clinic/doctor's office", : FEXACT error 7(location). LDSTP=18240 is too small for this problem,
##   (pastp=327.251, ipn_0:=ipoin[itp=353]=1011, stp[ipn_0]=329.048).
## Increase workspace or consider using 'simulate.p.value=TRUE'
## There was an error in 'add_p()/add_difference()' for variable 'DISCHDEST', p-value omitted:
## Error in stats::fisher.test(c("NULL", "Home", "Home", "Home", "NULL", : FEXACT error 7(location). LDSTP=18240 is too small for this problem,
##   (pastp=39.8074, ipn_0:=ipoin[itp=145]=2922, stp[ipn_0]=34.0131).
## Increase workspace or consider using 'simulate.p.value=TRUE'
## There was an error in 'add_p()/add_difference()' for variable 'PED_SHNT_HC_ETIOL', p-value omitted:
## Error in stats::fisher.test(c("Intraventricular hemorrhage (IVH) of prematurity without other cause", : FEXACT error 7(location). LDSTP=18240 is too small for this problem,
##   (pastp=94.3609, ipn_0:=ipoin[itp=96]=103, stp[ipn_0]=215.755).
## Increase workspace or consider using 'simulate.p.value=TRUE'
Characteristic No, N = 5301 Yes, N = 1,6921 p-value2
AGE_DAYS 834 (214, 3,436) 126 (48, 496) <0.001
SEX >0.9
Female 236 (45%) 757 (45%)
Male 294 (55%) 935 (55%)
RACE
American Indian or Alaska Native 6 (1.1%) 8 (0.5%)
Asian 9 (1.7%) 43 (2.5%)
Black or African American 93 (18%) 347 (21%)
Native Hawaiian or Other Pacific Islander 5 (0.9%) 4 (0.2%)
Unknown/Not Reported 86 (16%) 299 (18%)
White 331 (62%) 991 (59%)
ETHNICITY_HISPANIC 0.10
No 378 (71%) 1,235 (73%)
NULL 42 (7.9%) 166 (9.8%)
Yes 110 (21%) 291 (17%)
TRANST
Admitted from home/clinic/doctor's office 256 (48%) 762 (45%)
Admitted through ER, including outside ER with direct hospital admission 219 (41%) 430 (25%)
Chronic care/Rehab/Intermediate Care/Spinal Cord 1 (0.2%) 7 (0.4%)
Other 8 (1.5%) 28 (1.7%)
Transferred from outside hospital (NICU, PICU, Inpatient on General floor, Adult ICU) 46 (8.7%) 465 (27%)
HtoODay 1 (0, 11) 2 (0, 16) <0.001
PREM_BIRTH <0.001
24 completed weeks gestation 32 (6.0%) 63 (3.7%)
25-26 completed weeks gestation 62 (12%) 125 (7.4%)
27-28 completed weeks gestation 39 (7.4%) 99 (5.9%)
29-30 completed weeks gestation 13 (2.5%) 64 (3.8%)
31-32 completed weeks gestation 16 (3.0%) 65 (3.8%)
33-34 completed weeks gestation 16 (3.0%) 104 (6.1%)
35-36 completed weeks gestation 36 (6.8%) 158 (9.3%)
Less than 24 completed weeks gestation 30 (5.7%) 56 (3.3%)
No 286 (54%) 958 (57%)
DELIVERY_MODE <0.001
NULL 461 (87%) 971 (57%)
Scheduled C-Section 5 (0.9%) 179 (11%)
Unknown/Not Documented 3 (0.6%) 53 (3.1%)
Unscheduled C-Section 38 (7.2%) 279 (16%)
Vaginal Delivery 23 (4.3%) 210 (12%)
ANESURG 54 (42, 70) 52 (42, 66) 0.076
SURGANE 22 (16, 30) 23 (17, 32) 0.11
DPATRM 136 (109, 176) 125 (101, 155) <0.001
ANETIME 147 (117, 192) 137 (111, 168) <0.001
OPTIME 67 (47, 92) 56 (41, 74) <0.001
DOptoDis 2 (1, 3) 2 (1, 5) 0.6
DISCHDEST
Expired 3 (0.6%) 9 (0.5%)
Facility which was home 1 (0.2%) 6 (0.4%)
Home 453 (85%) 1,258 (74%)
NULL 56 (11%) 314 (19%)
Rehab 8 (1.5%) 20 (1.2%)
Separate acute care 2 (0.4%) 64 (3.8%)
Skilled care, not home 7 (1.3%) 19 (1.1%)
Unknown 0 (0%) 1 (<0.1%)
Unskilled facility, not home 0 (0%) 1 (<0.1%)
TOTHLOS 2 (1, 9) 2 (1, 9) 0.2
STILLINHOSP <0.001
NULL 474 (89%) 1,378 (81%)
Yes 56 (11%) 314 (19%)
PED_SHNT_HC_ETIOL
Congenital hydrocephalus 0 (0%) 524 (31%)
Intracranial cyst 0 (0%) 109 (6.4%)
Intraventricular hemorrhage (IVH) of prematurity without other cause 0 (0%) 373 (22%)
Neoplastic 0 (0%) 193 (11%)
NULL 530 (100%) 0 (0%)
Other - insufficient information to classify type 0 (0%) 149 (8.8%)
Post hemorrhagic from other vascular lesion 0 (0%) 58 (3.4%)
Post infectious 0 (0%) 73 (4.3%)
Post traumatic 0 (0%) 39 (2.3%)
Spina bifida with Chiari malformation / myelomeningocele 0 (0%) 118 (7.0%)
Syndromic, not otherwise specified 0 (0%) 37 (2.2%)
Vascular anomaly 0 (0%) 19 (1.1%)
PED_SHNT_PRIOR_HCPROC <0.001
No 0 (0%) 1,097 (65%)
NULL 530 (100%) 0 (0%)
Yes 0 (0%) 595 (35%)
1 Median (IQR); n (%)
2 Wilcoxon rank sum test; Pearson's Chi-squared test 
#summary of first time shunts, characterized by surgical adjuncts 
#tbl_summary(df_first[,c(2:5,7:16)], by=PED_SHNT_SURGADJS)%>%
  #bold_labels()  %>%add_p()