setwd("~/My Stuff/Masters/MPA/PAD 6833/Data")tab1<-read.delim("PUDF_base1_1q2020_tab.txt")library(dplyr)##
## Attaching package: 'dplyr'## The following objects are masked from 'package:stats':
##
## filter, lag## The following objects are masked from 'package:base':
##
## intersect, setdiff, setequal, unionlibrary(ggplot2)
library(tidyverse)## ── Attaching core tidyverse packages ──────────────────────── tidyverse 2.0.0 ──
## ✔ forcats 1.0.1 ✔ stringr 1.6.0
## ✔ lubridate 1.9.5 ✔ tibble 3.3.1
## ✔ purrr 1.2.2 ✔ tidyr 1.3.2
## ✔ readr 2.2.0## ── Conflicts ────────────────────────────────────────── tidyverse_conflicts() ──
## ✖ dplyr::filter() masks stats::filter()
## ✖ dplyr::lag() masks stats::lag()
## ℹ Use the conflicted package (<http://conflicted.r-lib.org/>) to force all conflicts to become errorsdiabetes_subset <- subset(
tab1,
grepl("^E11", PRINC_DIAG_CODE)
)illness_model <- lm(ILLNESS_SEVERITY ~ PAT_AGE, data=diabetes_subset)summary(illness_model)##
## Call:
## lm(formula = ILLNESS_SEVERITY ~ PAT_AGE, data = diabetes_subset)
##
## Residuals:
## Min 1Q Median 3Q Max
## -1.5588 -0.4582 -0.1909 0.5531 2.3023
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 1.000e+00 7.549e-01 1.325 0.1853
## PAT_AGE02 1.250e+00 8.440e-01 1.481 0.1386
## PAT_AGE03 1.333e+00 8.716e-01 1.530 0.1261
## PAT_AGE04 9.091e-01 7.662e-01 1.186 0.2355
## PAT_AGE05 6.977e-01 7.636e-01 0.914 0.3609
## PAT_AGE06 8.056e-01 7.653e-01 1.053 0.2925
## PAT_AGE07 1.017e+00 7.581e-01 1.342 0.1796
## PAT_AGE08 1.102e+00 7.570e-01 1.456 0.1454
## PAT_AGE09 1.109e+00 7.562e-01 1.466 0.1426
## PAT_AGE10 1.183e+00 7.556e-01 1.565 0.1175
## PAT_AGE11 1.237e+00 7.554e-01 1.637 0.1016
## PAT_AGE12 1.337e+00 7.552e-01 1.771 0.0766 .
## PAT_AGE13 1.316e+00 7.552e-01 1.743 0.0813 .
## PAT_AGE14 1.421e+00 7.551e-01 1.882 0.0599 .
## PAT_AGE15 1.447e+00 7.551e-01 1.916 0.0554 .
## PAT_AGE16 1.501e+00 7.551e-01 1.988 0.0468 *
## PAT_AGE17 1.530e+00 7.552e-01 2.025 0.0428 *
## PAT_AGE18 1.458e+00 7.553e-01 1.931 0.0535 .
## PAT_AGE19 1.493e+00 7.554e-01 1.976 0.0481 *
## PAT_AGE20 1.494e+00 7.558e-01 1.976 0.0482 *
## PAT_AGE21 1.506e+00 7.572e-01 1.989 0.0467 *
## PAT_AGE22 -1.972e-11 1.068e+00 0.000 1.0000
## PAT_AGE23 1.191e+00 7.560e-01 1.575 0.1152
## PAT_AGE24 1.408e+00 7.555e-01 1.864 0.0624 .
## PAT_AGE25 1.488e+00 7.580e-01 1.963 0.0497 *
## PAT_AGE26 1.559e+00 7.659e-01 2.035 0.0418 *
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 0.7549 on 13497 degrees of freedom
## Multiple R-squared: 0.02897, Adjusted R-squared: 0.02717
## F-statistic: 16.11 on 25 and 13497 DF, p-value: < 2.2e-16c_diabetes<- diabetes_subset |> drop_na(PAT_AGE)c_diabetes$PAT_AGE <- as.numeric(c_diabetes$PAT_AGE)## Warning: NAs introduced by coercionillness_model <- lm(ILLNESS_SEVERITY ~ PAT_AGE, data=c_diabetes)summary(illness_model)##
## Call:
## lm(formula = ILLNESS_SEVERITY ~ PAT_AGE, data = c_diabetes)
##
## Residuals:
## Min 1Q Median 3Q Max
## -1.6111 -0.4077 -0.3059 0.5923 1.8365
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 2.082158 0.026727 77.91 <2e-16 ***
## PAT_AGE 0.020344 0.001691 12.03 <2e-16 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 0.7612 on 13520 degrees of freedom
## (1 observation deleted due to missingness)
## Multiple R-squared: 0.01059, Adjusted R-squared: 0.01052
## F-statistic: 144.7 on 1 and 13520 DF, p-value: < 2.2e-16c_diabetes <- c_diabetes %>%
filter(!FIRST_PAYMENT_SRC %in% c("`", "", "ZZ", NA)) %>%
mutate(
INSURED = ifelse(FIRST_PAYMENT_SRC == "09", 0, 1)
)c_diabetes <- c_diabetes %>% filter (!SEX_CODE %in% c("U", "'", NA)) %>% mutate(SEX = ifelse(SEX_CODE == "M", 0, 1))c_diabetes <- c_diabetes %>% filter (!RACE %in% c("'","5", NA)) %>% mutate(WHITE= ifelse(SEX_CODE == "4", 1, 0))c_diabetes <- c_diabetes %>% filter (!ETHNICITY %in% c("'", NA)) %>% mutate(HISPANIC= ifelse(ETHNICITY == "1", 1, 0))illness_model <- lm(ILLNESS_SEVERITY ~ PAT_AGE + PAT_AGE + SEX_CODE + RACE + ETHNICITY + FIRST_PAYMENT_SRC, data=c_diabetes)summary(illness_model)##
## Call:
## lm(formula = ILLNESS_SEVERITY ~ PAT_AGE + PAT_AGE + SEX_CODE +
## RACE + ETHNICITY + FIRST_PAYMENT_SRC, data = c_diabetes)
##
## Residuals:
## Min 1Q Median 3Q Max
## -1.7207 -0.4829 -0.2168 0.5276 2.0315
##
## Coefficients: (1 not defined because of singularities)
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 1.209340 0.542910 2.228 0.02594 *
## PAT_AGE 0.020217 0.003125 6.470 1.03e-10 ***
## SEX_CODEF 0.142735 0.044128 3.235 0.00122 **
## SEX_CODEM 0.215530 0.043896 4.910 9.26e-07 ***
## RACE1 0.683593 0.557017 1.227 0.21976
## RACE2 0.547819 0.537143 1.020 0.30781
## RACE3 0.535417 0.532455 1.006 0.31465
## RACE4 0.461062 0.532308 0.866 0.38643
## ETHNICITY` 0.005805 0.251216 0.023 0.98156
## ETHNICITY1 0.084164 0.018869 4.460 8.28e-06 ***
## ETHNICITY2 NA NA NA NA
## FIRST_PAYMENT_SRC12 0.144410 0.079496 1.817 0.06932 .
## FIRST_PAYMENT_SRC13 0.040307 0.110596 0.364 0.71553
## FIRST_PAYMENT_SRC14 -0.221935 0.537359 -0.413 0.67961
## FIRST_PAYMENT_SRC15 0.138838 0.092327 1.504 0.13268
## FIRST_PAYMENT_SRC16 0.251665 0.078351 3.212 0.00132 **
## FIRST_PAYMENT_SRCBL 0.054311 0.082061 0.662 0.50809
## FIRST_PAYMENT_SRCCH 0.073328 0.122604 0.598 0.54980
## FIRST_PAYMENT_SRCCI 0.151344 0.085043 1.780 0.07517 .
## FIRST_PAYMENT_SRCHM 0.173289 0.082801 2.093 0.03639 *
## FIRST_PAYMENT_SRCMA 0.326040 0.077160 4.226 2.41e-05 ***
## FIRST_PAYMENT_SRCMB -0.192488 0.146860 -1.311 0.18999
## FIRST_PAYMENT_SRCMC 0.206706 0.079438 2.602 0.00928 **
## FIRST_PAYMENT_SRCOF 0.031861 0.132501 0.240 0.80998
## FIRST_PAYMENT_SRCVA 0.041431 0.137732 0.301 0.76357
## FIRST_PAYMENT_SRCWC 0.458060 0.239149 1.915 0.05547 .
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 0.7522 on 9612 degrees of freedom
## Multiple R-squared: 0.03121, Adjusted R-squared: 0.02879
## F-statistic: 12.9 on 24 and 9612 DF, p-value: < 2.2e-16illness_model <- lm(ILLNESS_SEVERITY ~ PAT_AGE + SEX + HISPANIC + WHITE + INSURED, data=c_diabetes)summary(illness_model)##
## Call:
## lm(formula = ILLNESS_SEVERITY ~ PAT_AGE + SEX + HISPANIC + WHITE +
## INSURED, data = c_diabetes)
##
## Residuals:
## Min 1Q Median 3Q Max
## -1.6333 -0.4642 -0.3172 0.5568 1.8479
##
## Coefficients: (2 not defined because of singularities)
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 2.128289 0.035160 60.532 < 2e-16 ***
## PAT_AGE 0.020994 0.002191 9.581 < 2e-16 ***
## SEX -0.081209 0.015741 -5.159 2.53e-07 ***
## HISPANIC 0.061396 0.017954 3.420 0.00063 ***
## WHITE NA NA NA NA
## INSURED NA NA NA NA
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 0.759 on 9633 degrees of freedom
## Multiple R-squared: 0.01141, Adjusted R-squared: 0.0111
## F-statistic: 37.06 on 3 and 9633 DF, p-value: < 2.2e-16#This model has an r-squared of 0.01141 suggesting it only covers about 1% of the data set. This it is overall significant with the p-value: < 2.2e-16. This is a really big data set with a lot of different factors that can relate to illness severity so that makes sense that I am not covering much with my model, but that seems kind of crazy.
#The model created does suggest that patient age, sex, and hispanic status are significant. I am not sure why NA appears for some of my variables. Patient age is a misnomer because it is suggesting that each section is one year so that as the year increases by 1, the illness severity increases by 0.02. Looking at the dictionary though, the numbers correlate to an age range so 11 actually covers 40-44 and 12 covers 45-49 (looking at it now I also need to cut out 22-26 in the future because those are specifically HIV and drug/alcohol use patients). Would this work for the binomials like sex? If the patient is male then they it is predicted their illness severity would go down 0.08 and if they are hispanic then their severity increases by 0.06?
table(c_diabetes$FIRST_PAYMENT_SRC)##
## 11 12 13 14 15 16 BL CH CI HM MA MB MC OF VA WC
## 99 947 87 2 202 1786 557 61 378 503 3904 36 973 48 43 11table(c_diabetes$FIRST_PAYMENT_SRC)##
## 11 12 13 14 15 16 BL CH CI HM MA MB MC OF VA WC
## 99 947 87 2 202 1786 557 61 378 503 3904 36 973 48 43 11