MH_Codebook_V1

Load Packages

tidyverse

if (!require(tidyverse)){
  install.packages("tidyverse", dependencies = TRUE)
  library(tidyverse)
}

codebookr

if (!require(codebookr)){
  install.packages("codebookr", dependencies = TRUE)
  library(codebookr)
}

summarytools

if (!require(summarytools)){
  install.packages("summarytools", dependencies = TRUE)
  library(summarytools)
}

Import Data

mh_deid <- read.csv("~/Desktop/R-Code/mh_deid.csv")

Codebook

mh_deid %>%
  select(age_yrs, PATIENT_GENDER_CD, PATIENT_RACE_DESC, PATIENT_ETHNIC_GROUP_DESC, PATIENT_LANGUAGE_DESC, PATIENT_RELIGION_DESC, PATIENT_MARITAL_STATUS_DESC, PATIENT_STATE_CD, EDU_YEARS, TOBACCO_DESC, depression, anxiety, ptsd, bipolar, body_image, ocd, stress, seasonalAD, panic, any_psych_dx,ST_ABBR, E_TOTPOP, RPL_THEMES, RPL_THEME1, RPL_THEME2, RPL_THEME3, RPL_THEME4) -> exampledf1
print(dfSummary(exampledf1), method = 'render') 

Data Frame Summary

exampledf1

Dimensions: 11306 x 27
Duplicates: 0

No	Variable	Stats / Values	Freqs (% of Valid)	Graph	Valid	Missing
1	age_yrs [numeric]	Mean (sd) : 49.8 (18.4) min ≤ med ≤ max: 2.8 ≤ 49.3 ≤ 100.4 IQR (CV) : 30.6 (0.4)	9012 distinct values		11306 (100.0%)	0 (0.0%)
2	PATIENT_GENDER_CD [character]	1. F 2. M	6024 ( 53.3% ) 5282 ( 46.7% )		11306 (100.0%)	0 (0.0%)
3	PATIENT_RACE_DESC [character]	1. WHITE OR CAUCASIAN 2. BLACK OR AFRICAN AMERICAN 3. OTHER 4. ASIAN 5. UNKNOWN 6. CHOOSE NOT TO DISCLOSE 7. AMERICAN INDIAN AND ALASK 8. ASIAN INDIAN 9. OTHER ASIAN 10. NATIVE HAWAIIAN AND OTHER [ 5 others ]	9823 ( 86.9% ) 726 ( 6.4% ) 261 ( 2.3% ) 255 ( 2.3% ) 113 ( 1.0% ) 57 ( 0.5% ) 39 ( 0.3% ) 11 ( 0.1% ) 7 ( 0.1% ) 5 ( 0.0% ) 9 ( 0.1% )		11306 (100.0%)	0 (0.0%)
4	PATIENT_ETHNIC_GROUP_DESC [character]	1. CHOOSE NOT TO DISCLOSE 2. HISPANIC 3. NON-HISPANIC 4. UNKNOWN	66 ( 0.6% ) 203 ( 1.8% ) 10719 ( 94.8% ) 318 ( 2.8% )		11306 (100.0%)	0 (0.0%)
5	PATIENT_LANGUAGE_DESC [character]	1. ENGLISH 2. ARABIC 3. JAPANESE 4. SPANISH 5. CHINESE, MANDARIN 6. UNKNOWN 7. KOREAN 8. ALBANIAN 9. SIGN LANGUAGE 10. RUSSIAN [ 12 others ]	11202 ( 99.1% ) 35 ( 0.3% ) 22 ( 0.2% ) 13 ( 0.1% ) 5 ( 0.0% ) 5 ( 0.0% ) 4 ( 0.0% ) 3 ( 0.0% ) 3 ( 0.0% ) 2 ( 0.0% ) 12 ( 0.1% )		11306 (100.0%)	0 (0.0%)
6	PATIENT_RELIGION_DESC [character]	1. NONE 2. CATHOLIC 3. CHRISTIAN 4. UNKNOWN 5. LUTHERAN 6. PROTESTANT 7. BAPTIST 8. METHODIST 9. JEWISH 10. PATIENT REFUSED [ 40 others ]	4438 ( 39.3% ) 2222 ( 19.7% ) 1156 ( 10.2% ) 487 ( 4.3% ) 460 ( 4.1% ) 437 ( 3.9% ) 408 ( 3.6% ) 262 ( 2.3% ) 214 ( 1.9% ) 183 ( 1.6% ) 1039 ( 9.2% )		11306 (100.0%)	0 (0.0%)
7	PATIENT_MARITAL_STATUS_DESC [character]	1. DIVORCED 2. LEGALLY SEPARATED 3. MARRIED 4. OTHER 5. SIGNIFICANT OTHER 6. SINGLE 7. UNKNOWN 8. WIDOWED	295 ( 2.6% ) 53 ( 0.5% ) 4824 ( 42.7% ) 10 ( 0.1% ) 121 ( 1.1% ) 3551 ( 31.4% ) 2226 ( 19.7% ) 226 ( 2.0% )		11306 (100.0%)	0 (0.0%)
8	PATIENT_STATE_CD [character]	1. MI 2. OH 3. IN 4. FL 5. CA 6. IL 7. NC 8. NY 9. IA 10. TX [ 23 others ]	10874 ( 96.2% ) 256 ( 2.3% ) 47 ( 0.4% ) 23 ( 0.2% ) 8 ( 0.1% ) 7 ( 0.1% ) 7 ( 0.1% ) 7 ( 0.1% ) 6 ( 0.1% ) 6 ( 0.1% ) 65 ( 0.6% )		11306 (100.0%)	0 (0.0%)
9	EDU_YEARS [integer]	Mean (sd) : 15.3 (3.7) min ≤ med ≤ max: 0 ≤ 16 ≤ 30 IQR (CV) : 5 (0.2)	28 distinct values		822 (7.3%)	10484 (92.7%)
10	TOBACCO_DESC [character]	1. NEVER 2. NOT ASKED 3. PASSIVE 4. QUIT 5. YES	6149 ( 56.2% ) 10 ( 0.1% ) 325 ( 3.0% ) 3443 ( 31.5% ) 1005 ( 9.2% )		10932 (96.7%)	374 (3.3%)
11	depression [integer]	1 distinct value	0 : 11306 ( 100.0% )		11306 (100.0%)	0 (0.0%)
12	anxiety [integer]	1 distinct value	0 : 11306 ( 100.0% )		11306 (100.0%)	0 (0.0%)
13	ptsd [integer]	1 distinct value	0 : 11306 ( 100.0% )		11306 (100.0%)	0 (0.0%)
14	bipolar [integer]	1 distinct value	0 : 11306 ( 100.0% )		11306 (100.0%)	0 (0.0%)
15	body_image [integer]	1 distinct value	0 : 11306 ( 100.0% )		11306 (100.0%)	0 (0.0%)
16	ocd [integer]	1 distinct value	0 : 11306 ( 100.0% )		11306 (100.0%)	0 (0.0%)
17	stress [integer]	1 distinct value	0 : 11306 ( 100.0% )		11306 (100.0%)	0 (0.0%)
18	seasonalAD [integer]	1 distinct value	0 : 11306 ( 100.0% )		11306 (100.0%)	0 (0.0%)
19	panic [integer]	1 distinct value	0 : 11306 ( 100.0% )		11306 (100.0%)	0 (0.0%)
20	any_psych_dx [integer]	1 distinct value	0 : 11306 ( 100.0% )		11306 (100.0%)	0 (0.0%)
21	ST_ABBR [character]	1. IL 2. IN 3. MI 4. OH	1 ( 0.0% ) 40 ( 0.4% ) 11013 ( 97.4% ) 252 ( 2.2% )		11306 (100.0%)	0 (0.0%)
22	E_TOTPOP [integer]	Mean (sd) : 4350.1 (1694.9) min ≤ med ≤ max: 182 ≤ 4113 ≤ 20695 IQR (CV) : 2162 (0.4)	1925 distinct values		11306 (100.0%)	0 (0.0%)
23	RPL_THEMES [numeric]	Mean (sd) : -0.2 (23) min ≤ med ≤ max: -999 ≤ 0.3 ≤ 1 IQR (CV) : 0.4 (-141.6)	2281 distinct values		11306 (100.0%)	0 (0.0%)
24	RPL_THEME1 [numeric]	Mean (sd) : -0.2 (23) min ≤ med ≤ max: -999 ≤ 0.3 ≤ 1 IQR (CV) : 0.4 (-124.6)	2244 distinct values		11306 (100.0%)	0 (0.0%)
25	RPL_THEME2 [numeric]	Mean (sd) : -0.1 (23) min ≤ med ≤ max: -999 ≤ 0.3 ≤ 1 IQR (CV) : 0.4 (-161.8)	2212 distinct values		11306 (100.0%)	0 (0.0%)
26	RPL_THEME3 [numeric]	Mean (sd) : 0.5 (0.3) min ≤ med ≤ max: 0 ≤ 0.5 ≤ 1 IQR (CV) : 0.5 (0.6)	1620 distinct values		11306 (100.0%)	0 (0.0%)
27	RPL_THEME4 [numeric]	Mean (sd) : -0.1 (23) min ≤ med ≤ max: -999 ≤ 0.4 ≤ 1 IQR (CV) : 0.5 (-240.6)	2206 distinct values		11306 (100.0%)	0 (0.0%)

Generated by summarytools 1.0.1 (R version 4.1.2)
2022-10-11