R Notebook

1. Defining malnutrition

1.1. Prerequisites

library(tidyverse)
library(zscorer)
library(readr)
library(dplyr)
library (lubridate)
library (ggplot2)
library (rmarkdown)
library(openxlsx)
library(writexl)
library(rstatix)
library(eeptools)
library(ggpubr)
library(readxl)

PSFI_df_malnutrition <- read_xlsx("Ben_cut_4_27.xlsx")

Sys.setenv(LANGUAGE = "en")

1.2. Anthroprometric analysis

# CREATE NEW CATEGORY THAT CALCULATES AGE IN YEARS BASED ON AGE IN DAYS
PSFI_df_malnutrition <- PSFI_df_malnutrition %>%
  mutate(age_years = age_days_exact / 365.25)

summary(PSFI_df_malnutrition$age_years)

    Min.  1st Qu.   Median     Mean  3rd Qu.     Max. 
 0.08768  0.79176  1.65535  3.07854  4.00176 14.92786 

# CREATES A SUMMARY OF CASES (ht=height, wt=weight, age_years= age in years, muac)
PSFI_df_malnutrition %>%
  filter(case_control == 1) %>%
  select(ht, wt, age_years, muac) %>%
  summary()

       ht               wt          age_years             muac      
 Min.   : 20.00   Min.   : 3.20   Min.   : 0.08768   Min.   : 9.00  
 1st Qu.: 67.00   1st Qu.: 7.00   1st Qu.: 0.73436   1st Qu.:14.00  
 Median : 78.00   Median : 9.50   Median : 1.40141   Median :15.00  
 Mean   : 85.23   Mean   :12.17   Mean   : 2.83539   Mean   :15.38  
 3rd Qu.: 98.00   3rd Qu.:14.00   3rd Qu.: 3.49267   3rd Qu.:17.00  
 Max.   :192.00   Max.   :72.00   Max.   :14.92786   Max.   :27.00  
 NAs    :1                                           NAs    :2      

# CREATES A HISTOGRAM AND BOXPLOT OF HEIGHT (CASES)
layout(mat = matrix(c(1,2),2,1, byrow=TRUE),  height = c(1,8))
par(mar=c(0, 3.1, 1.1, 2.1))
boxplot(PSFI_df_malnutrition$ht[PSFI_df_malnutrition$case_control == 1] , horizontal=TRUE , ylim=c(0,200), xaxt="n" , col=rgb(0.8,0.8,0,0.5) , frame=F)
par(mar=c(4, 3.1, 1.1, 2.1))

hist(PSFI_df_malnutrition$ht[PSFI_df_malnutrition$case_control == 1]
     , breaks=40 , col=rgb(1,0.8,0.8,1) , border=F , main="" , xlab="Height (cm)", xlim=c(0,200))

# CREATES A HISTOGRAM AND BOXPLOT OF WEIGHT (CASES)
layout(mat = matrix(c(1,2),2,1, byrow=TRUE),  height = c(1,8))
 par(mar=c(0, 3.1, 1.1, 2.1))
boxplot(PSFI_df_malnutrition$wt[PSFI_df_malnutrition$case_control == 1] , horizontal=TRUE , ylim=c(0,75), xaxt="n" , col=rgb(0.8,0.8,0,0.5) , frame=F)
par(mar=c(4, 3.1, 1.1, 2.1))

hist(PSFI_df_malnutrition$wt[PSFI_df_malnutrition$case_control == 1]
     , breaks=40 , col=rgb(1,0.8,0.8,1) , border=F , main="" , xlab="Weight (kg)", xlim=c(0,75))

# CREATES A HISTOGRAM AND BOXPLOT OF MUAC (CASES)
## BLUE LINE OUTLINES MODERATE MALNUTRITION (12.5cm)
## RED LINE OUTLINES SEVERE MALNUTRITION (11.5cm)
layout(mat = matrix(c(1,2),2,1, byrow=TRUE),  height = c(1,8))
par(mar=c(0, 3.1, 1.1, 2.1))
boxplot(PSFI_df_malnutrition$muac[PSFI_df_malnutrition$case_control == 1] , horizontal=TRUE , ylim=c(5,30), xaxt="n" , col=rgb(0.8,0.8,0,0.5) , frame=F)
par(mar=c(4, 3.1, 1.1, 2.1))

hist(PSFI_df_malnutrition$muac[PSFI_df_malnutrition$case_control == 1]
     , breaks=20 , col=rgb(1,0.8,0.8,1) , border=F , main="" , xlab="Mid-upper arm circumference (cm)", xlim=c(5,30))
abline(v = 11.5, col = "red", lwd = 2, lty = 2)   # severe

abline(v = 12.5, col = "blue", lwd = 2, lty = 2)  # moderate

# HISTOGRAM AND BOXPLOT OF AGE (CASES)
layout(mat = matrix(c(1,2),2,1, byrow=TRUE),  height = c(1,8))
 par(mar=c(0, 3.1, 1.1, 2.1))
boxplot(PSFI_df_malnutrition$age_years[PSFI_df_malnutrition$case_control == 1] , horizontal=TRUE , ylim=c(0,15), xaxt="n" , col=rgb(0.8,0.8,0,0.5) , frame=F)
par(mar=c(4, 3.1, 1.1, 2.1))

hist(PSFI_df_malnutrition$age_years[PSFI_df_malnutrition$case_control == 1], breaks=40 , col=rgb(1,0.8,0.8,1) , border=F , main="" , xlab="Age (years)", xlim=c(0,15))

1.3. Z-scorer

# CREATES A NEW CATEGORY DEFINING SEX AS 1/2, INSTEAD OF 0/1 (NECESSARY FOR ZSCORER PACKAGE)
PSFI_df_malnutrition <- PSFI_df_malnutrition %>%
  mutate(sex_who = if_else(sex == 1, 1, 2))

summary(PSFI_df_malnutrition$sex_who)

   Min. 1st Qu.  Median    Mean 3rd Qu.    Max. 
  1.000   1.000   1.000   1.427   2.000   2.000 

# CREATES A NEW CATEGORY CALCULATING AGE IN MONTHS BASED ON AGE IN DAYS
PSFI_df_malnutrition <- PSFI_df_malnutrition %>%
  mutate(age_months = age_days_exact / 30.4375)

summary(PSFI_df_malnutrition$age_months)

   Min. 1st Qu.  Median    Mean 3rd Qu.    Max. 
  1.052   9.501  19.864  36.942  48.021 179.134 

# CREATES A NEW CATEGORY (AGE<6M = 0, AGE 6M - 5J = 1, AGE > 5J = 2)
PSFI_df_malnutrition <- PSFI_df_malnutrition %>%
  mutate(
    age_group = case_when(
      age_months < 6 ~ 0L,
      age_months >= 6 & age_years < 5 ~ 1L,
      age_years >= 5 ~ 2L,
      TRUE ~ NA_integer_
    )
  )

# ZSCORER PACKAGE CALCULATES ZSCORE OF WEIGHT FOR LENGTH (wflz), WEIGHT FOR AGE (wfaz), HEIGHT FOR AGE (hfaz), WEIGHT FOR HEIGHT (wfhz), BMI FOR AGE (baz)
PSFI_df_malnutrition <- PSFI_df_malnutrition %>%
  mutate(
    wflz = addWGSR(
      data = .,
      sex = "sex_who",
      firstPart = "wt",
      secondPart = "ht",
      index = "wfl"
    )$wflz,
    
    wfaz = addWGSR(
      data = .,
      sex = "sex_who",
      firstPart = "wt",
      secondPart = "age_days_exact",
      index = "wfa"
    )$wfaz,
    
    hfaz = addWGSR(
      data = .,
      sex = "sex_who",
      firstPart = "ht",
      secondPart = "age_days_exact",
      index = "hfa"
    )$hfaz,

    wfhz = addWGSR(
      data = .,
      sex = "sex_who",
      firstPart = "wt",
      secondPart = "ht",
      index = "wfh"
    )$wfhz,
    
    baz = addWGSR(
      data = .,
      sex = "sex_who",
      firstPart = "wt",
      secondPart = "ht",
      thirdPart = "age_days_exact",
      index = "bfa"
    )$bfaz
  )

============================================================================================================================================
============================================================================================================================================
============================================================================================================================================
============================================================================================================================================
============================================================================================================================================

# ASSIGNS Z-SCORE TO MUAC BASED ON THE 11.5 & 12.5CM LIMITS, THIS IS A PROXY SINCE LATER A Z-SCORE OF 0 WILL EQUAL NO MALNUTRITION, -2.5 WILL EQUAL MODERATE MALNUTRITION AND -4 WILL EQUAL SEVERE MALNUTRITION
PSFI_df_malnutrition <- PSFI_df_malnutrition %>%
  mutate(
    muacz = case_when(
      muac >= 12.5 ~ 0,
      muac >= 11.5 & muac < 12.5 ~ -2.5,
      muac < 11.5 ~ -4,
      TRUE ~ NA_real_
    )
  )

# CREATE MALNUTRITION ZSCORE BASED ON WFHL, BAZ, MUACZ (to be defined later)
PSFI_df_malnutrition <- PSFI_df_malnutrition %>%
  mutate(
    zscore_unified = case_when(
      age_group == 1 & ht >= 45 & ht < 65 ~ wflz,
      age_group == 1 & ht >= 65 & ht < 120 ~ wfhz,
      age_group == 1 & (ht < 45 | ht >= 120 | is.na(ht)) ~ muacz,
      age_group == 2 ~ baz,
      TRUE ~ NA_real_
    )
  )

# CREATE MALNUTRITION CATEGORY BASED ON PREVIOUS Z SCORE
PSFI_df_malnutrition <- PSFI_df_malnutrition %>%
  mutate(
    malnutrition = case_when(
  is.na(zscore_unified) ~ NA_integer_,
  zscore_unified < -3 ~ 2L,
  zscore_unified >= -3 & zscore_unified < -2 ~ 1L,
  TRUE ~ 0L
    )
  )

# ADD A MALNUTRITION SOURCE SO WE KNOW WHICH ANTHROPOMETRIC MEASURE IS BEING USED TO DEFINE MALNUTRITION
PSFI_df_malnutrition <- PSFI_df_malnutrition %>%
  mutate(
    malnutrition_source = case_when(
      age_group == 1 & ht >= 45 & ht < 65  ~ "WFL",
      age_group == 1 & ht >= 65 & ht < 120 ~ "WFH",
      age_group == 1 & (ht < 45 | ht >= 120 | is.na(ht)) ~ "MUAC",
      age_group == 2 ~ "BFA",
      TRUE ~ NA_character_
    )
  )

1.4 Z score check

subset_wfaz <- PSFI_df_malnutrition %>%
  filter(case_control == 1) %>%
  pull(wfaz)

# SUMMARY OF WEIGHT FOR AGE Z-SCORE (CASES)
summary(subset_wfaz)

   Min. 1st Qu.  Median    Mean 3rd Qu.    Max.     NAs 
-7.5400 -2.3050 -0.9700 -1.0971  0.0475 10.2500      49 

# NUMBER OF CASES IN WEIGHT FOR AGE Z SCORE
length(subset_wfaz)

[1] 755

subset_wfaz1 <- PSFI_df_malnutrition %>%
  filter(case_control == 1) %>%
  identify_outliers(wfaz) %>%
  pull(wfaz)

# NUMBER OF OUTLIERS IN WEIGHT FOR AGE Z SCORE
length(subset_wfaz1)

[1] 10

subset_hfaz <- PSFI_df_malnutrition %>%
  filter(case_control == 1) %>%
  pull(hfaz)

# SUMMARY OF HEIGHT FOR AGE Z SCORE (CASES)
summary(subset_hfaz)

    Min.  1st Qu.   Median     Mean  3rd Qu.     Max.      NAs 
-18.2600  -2.8700  -1.1800  -0.9338   0.6350  44.5600        1 

# NUMBER OF CASES IN HEIGHT FOR AGE Z SCORE
length(subset_hfaz)

[1] 755

subset_hfaz1 <- PSFI_df_malnutrition %>%
  filter(case_control == 1) %>%
  identify_outliers(hfaz) %>%
  pull(hfaz)

# NUMBER OF OUTLIERS FOR HEIGHT FOR AGE Z SCORE
length(subset_hfaz1)

[1] 29

subset_wfhz <- PSFI_df_malnutrition %>%
  filter(case_control == 1) %>%
  pull(wfhz)

# SUMMARY OF WEIGHT FOR HEIGHT Z SCORE (CASES)
summary(subset_wfhz)

    Min.  1st Qu.   Median     Mean  3rd Qu.     Max.      NAs 
-10.5900  -2.2300  -0.5700  -0.4132   1.1875  42.7700       93 

# NUMBER OF CASES IN WEIGHT FOR HEIGHT Z SCORE
length(subset_wfhz)

[1] 755

subset_wfhz1 <- PSFI_df_malnutrition %>%
  filter(case_control == 1) %>%
  identify_outliers(wfhz) %>%
  pull(wfhz)

# NUMBER OF OUTLIERS IN WEIGHT FOR HEIGHT Z SCORE
length(subset_wfhz1)

[1] 20

subset_baz <- PSFI_df_malnutrition %>%
  filter(case_control == 1) %>%
  pull(baz)

# SUMMARY OF BMI FOR AGE Z SCORE (CASES)
summary(subset_baz)

   Min. 1st Qu.  Median    Mean 3rd Qu.    Max.     NAs 
-12.560  -2.250  -0.695  -0.583   1.060  50.500       1 

# NUMBER OF CASES IN BMI FOR AGE Z SCORE
length(subset_baz)

[1] 755

subset_baz1 <- PSFI_df_malnutrition %>%
  filter(case_control == 1) %>%
  identify_outliers(baz) %>%
  pull(baz)

# NUMBER OF OUTLIERS IN BMI FOR AGE Z SCORE
length(subset_baz1)

[1] 29

subset_muac <- PSFI_df_malnutrition %>%
  filter(age_group == 1 & case_control == 1) %>%
  pull(muac)

# SUMMARY OF MUAC (CASES, 6-59 MONTHS)
summary(subset_muac)

   Min. 1st Qu.  Median    Mean 3rd Qu.    Max.     NAs 
   9.00   14.00   15.00   15.02   16.00   23.00       2 

# NUMBER OF CASES IN MUAC
length(subset_muac)

[1] 535

subset_muac1 <- PSFI_df_malnutrition %>%
  filter(age_group == 1 & case_control == 1) %>%
  identify_outliers(muac) %>%
  pull(muac)

# NUMBER OF OUTLIERS IN MUAC
length(subset_muac1)

[1] 18

# CREATES BOXPLOT AND HISTOGRAM FOR WEIGHT FOR HEIGHT Z SCORE (CASES)
layout(mat = matrix(c(1,2),2,1, byrow=TRUE),  height = c(1,8))
 par(mar=c(0, 3.1, 1.1, 2.1))
boxplot(PSFI_df_malnutrition$wfhz[PSFI_df_malnutrition$case_control == 1] , horizontal=TRUE , ylim=c(-11,43), xaxt="n" , col=rgb(0.8,0.8,0,0.5) , frame=F)
par(mar=c(4, 3.1, 1.1, 2.1))

hist(PSFI_df_malnutrition$wfhz[PSFI_df_malnutrition$case_control == 1]
     , breaks=50 , col=rgb(1,0.8,0.8,1) , border=F , main="" , xlab="Weight-for-height (z-score)", xlim=c(-11,43))

# CREATES BOXPLOT AND HISTOGRAM OF WEIGHT FOR AGE (CASES)
layout(mat = matrix(c(1,2),2,1, byrow=TRUE),  height = c(1,8))
 par(mar=c(0, 3.1, 1.1, 2.1))
boxplot(PSFI_df_malnutrition$wfaz[PSFI_df_malnutrition$case_control == 1] , horizontal=TRUE , ylim=c(-8,11), xaxt="n" , col=rgb(0.8,0.8,0,0.5) , frame=F)
par(mar=c(4, 3.1, 1.1, 2.1))

hist(PSFI_df_malnutrition$wfaz[PSFI_df_malnutrition$case_control == 1]
     , breaks=50 , col=rgb(1,0.8,0.8,1) , border=F , main="" , xlab="Weight-for-age (z-score)", xlim=c(-8,11))

# CREATES BOXPLOT AND HISTOGRAM OF BMI FOR AGE (CASES)
layout(mat = matrix(c(1,2),2,1, byrow=TRUE),  height = c(1,8))
 par(mar=c(0, 3.1, 1.1, 2.1))
boxplot(PSFI_df_malnutrition$baz[PSFI_df_malnutrition$case_control == 1] , horizontal=TRUE , ylim=c(-13,51), xaxt="n" , col=rgb(0.8,0.8,0,0.5) , frame=F)
par(mar=c(4, 3.1, 1.1, 2.1))

hist(PSFI_df_malnutrition$baz[PSFI_df_malnutrition$case_control == 1]
     , breaks=50 , col=rgb(1,0.8,0.8,1) , border=F , main="" , xlab="BMI-for-age (z-score)", xlim=c(-13,51))

# CREATE BOXPLOT AND HISTOGRAM OF HEIGHT FOR AGE 
layout(mat = matrix(c(1,2),2,1, byrow=TRUE),  height = c(1,8))
 par(mar=c(0, 3.1, 1.1, 2.1))
boxplot(PSFI_df_malnutrition$hfaz[PSFI_df_malnutrition$case_control == 1] , horizontal=TRUE , ylim=c(-19,45), xaxt="n" , col=rgb(0.8,0.8,0,0.5) , frame=F)
par(mar=c(4, 3.1, 1.1, 2.1))

hist(PSFI_df_malnutrition$hfaz[PSFI_df_malnutrition$case_control == 1]
     , breaks=50 , col=rgb(1,0.8,0.8,1) , border=F , main="" , xlab="Height-for-age (z-score)", xlim=c(-19,45))

# CREATES HISTOGRAM AND BOXPLOT OF MUAC (CASES, 6-59 MONTHS)
layout(mat = matrix(c(1,2),2,1, byrow=TRUE),  height = c(1,8))
par(mar=c(0, 3.1, 1.1, 2.1))
boxplot(PSFI_df_malnutrition$muac[PSFI_df_malnutrition$case_control == 1] , horizontal=TRUE , ylim=c(5,30), xaxt="n" , col=rgb(0.8,0.8,0,0.5) , frame=F)
par(mar=c(4, 3.1, 1.1, 2.1))

hist(PSFI_df_malnutrition$muac[PSFI_df_malnutrition$case_control == 1]
     , breaks=20 , col=rgb(1,0.8,0.8,1) , border=F , main="" , xlab="Mid-upper arm circumference (cm)", xlim=c(5,30))
abline(v = 11.5, col = "red", lwd = 2, lty = 2)   # severe

abline(v = 12.5, col = "blue", lwd = 2, lty = 2)  # moderate

# MAKES A TABLE OF OUTLIERS FOR WEIGHT FOR HEIGHT
wfhz_outliers <- PSFI_df_malnutrition %>%
  filter(case_control == 1) %>%
  identify_outliers(wfhz) %>%
  select(record_id, wt, ht, wfhz, age_months, is.outlier, is.extreme)

wfhz_outliers

# MAKES A TABLE OF OUTLIERS FOR WEIGHT FOR AGE
wfaz_outliers <- PSFI_df_malnutrition %>%
  filter(case_control == 1) %>%
  identify_outliers(wfaz) %>%
  select(record_id,ht, wt, age_months, wfaz, is.outlier, is.extreme)

wfaz_outliers

# MAKES A TABLE OF OUTLIERS FOR HEIGHT FOR AGE
hfaz_outliers <- PSFI_df_malnutrition %>%
  filter(case_control == 1) %>%
  identify_outliers(hfaz) %>%
  select(record_id, ht, wt, age_months, hfaz, is.outlier, is.extreme)

hfaz_outliers

# MAKES A TABLE OF OUTLIERS FOR BMI FOR AGE
baz_outliers <- PSFI_df_malnutrition %>%
  filter(case_control == 1) %>%
  identify_outliers(baz) %>%
  select(record_id, wt, ht, baz, age_months, is.outlier, is.extreme)

baz_outliers

# MAKES A TABLE OF OUTLIERS FOR MUAC

muac_outliers <- PSFI_df_malnutrition %>%
  filter(age_group == 1 & case_control == 1) %>%
  identify_outliers(muac) %>%
  select(record_id, ht, wt, age_months, muac, is.outlier, is.extreme)

muac_outliers

# CREATES AN EXCEL FILE OF ALL OUTLIER TABLES
write.xlsx(
  list(
    WFHZ = wfhz_outliers,
    WFAZ = wfaz_outliers,
    HAZ = hfaz_outliers,
    MUAC = muac_outliers, 
    BAZ = baz_outliers
  ),
  file = "anthropometric_outliers.xlsx"
)

# CREATE A TABLE OF ALL NA'S
na_table <- PSFI_df_malnutrition %>%
  filter(case_control == 1) %>%
  mutate(
    hfaz_missing = is.na(hfaz),
    wfaz_missing  = is.na(wfaz),
    wfhz_missing = is.na(wfhz), 
    muac_missing = is.na (muac), 
    baz_missing = is.na (baz)
  ) %>%
  filter(hfaz_missing | wfaz_missing | wfhz_missing | baz_missing | muac_missing) %>%
  select(
    record_id,
    ht,
    wt,
    age_months,
    hfaz,
    wfaz,
    wfhz,
    baz,
    muac, 
    hfaz_missing,
    wfaz_missing,
    wfhz_missing, 
    baz_missing,
    muac_missing
  )

na_table

anthro_vars <- c("hfaz", "wfaz", "wfhz", "baz", "muac")

cases_df <- PSFI_df_malnutrition %>%
  filter(case_control == 1)

anthro_summary <- lapply(anthro_vars, function(var) {

  x <- cases_df[[var]]

  outlier_info <- cases_df %>%
    select(all_of(var)) %>%
    identify_outliers(!!sym(var))

  tibble(
    measure = var,
    n = sum(!is.na(x)),
    n_missing = sum(is.na(x)),
    pct_missing = round(mean(is.na(x)) * 100, 2),
    n_outliers = sum(outlier_info$is.outlier, na.rm = TRUE),
    n_extreme_outliers = sum(outlier_info$is.extreme, na.rm = TRUE)
  )

}) %>%
  bind_rows()

anthro_summary

write.xlsx(
  list(
    Missing = na_table,
    Summary = anthro_summary
  ),
  file = "NA.xlsx")

1.5 Data check

PSFI_df_malnutrition %>%
  group_by(malnutrition, case_control) %>%
  summarize(
    malnutrition_missing = sum(is.na(malnutrition))
  )

 #103 mort_inhosp_missing  = controls

PSFI_df_malnutrition %>%
  filter (case_control == 1) %>%
  group_by(malnutrition_source) %>%
  summarize(
    count_malnut = n()
  )

# NA due to child with height of 20 cm (under WHO curve for wfh/wfl) and 38 days old (too young for MUAC)

PSFI_df_malnutrition %>%
  filter (case_control == 1) %>%
  group_by(malnutrition) %>%
  summarize(
    count_malnut = n()
  )

PSFI_df_malnutrition %>%
  drop_na(malnutrition, mort_inhosp) %>%
  ggplot(aes(x = factor(malnutrition))) +
  geom_bar() +
  labs(
    x = "Malnutrition",
    y = "N"
  )

NA

PSFI_df_malnutrition%>%
  filter (case_control == 1) %>%
  count(malnutrition, mort_inhosp) %>%
  group_by(malnutrition) %>%
  mutate(prop = n / sum(n))

PSFI_df_malnutrition %>%
  drop_na(mort_inhosp, malnutrition) %>% #only cases & minus the one previously mentioned NA
  ggplot +
  (aes(x = factor(malnutrition), fill = factor(mort_inhosp))) +
  geom_bar(stat = "count", position = "fill") +
  labs(
    x = "Malnutrition",
    y = "Proportion",
    fill = "Mortality"
  )

PSFI_df_malnutrition %>%
  count(malnutrition, case_control) %>%
  group_by(malnutrition) %>%
  mutate(prop = n / sum(n))

PSFI_df_malnutrition %>%
  drop_na(malnutrition) %>%
  ggplot +
  (aes(x = factor(malnutrition), fill = factor(case_control))) +
  geom_bar(stat = "count", position = "fill") +
  labs(
    x = "Malnutrition",
    y = "Proportion",
    fill = "Case/Control"
  )

PSFI_df_malnutrition %>%

ggplot +
  (aes(x = whz, y = wfhz)) +
  geom_point(alpha = 0.5) +
  geom_abline(slope = 1, intercept = 0, linetype = "dashed") +
    stat_cor(method = "pearson") +
  theme_minimal()

PSFI_df_malnutrition %>%

ggplot +
  (aes(x = waz, y = wfaz)) +
  geom_point(alpha = 0.5) +
  geom_abline(slope = 1, intercept = 0, linetype = "dashed") +
    stat_cor(method = "pearson") +
  theme_minimal()

write.xlsx(PSFI_df_malnutrition, file = "PSFI_final_malnutrition.xlsx")