Read in Data

Upload our data
Transform Data so that all checked variables = 1 and unchecked = 0
Record ID 312 being omitted because sex = NA

Omarallergy<- as.data.frame(read.csv("AdultFoodAllergyCorr-ShellfishTrueAllergy_DATA_LABELS_2024-05-30_0822.csv"))


Omarallergy$Race..ethnicity..choice.otherRace. <- Omarallergy$Race..ethnicity..choice.other.
Omarallergy$Race..ethnicity..choice.other. <- NULL


##changing the variable names for ease
names(Omarallergy)<- gsub("Race..ethnicity..choice.", "", names(Omarallergy))

names(Omarallergy)<- gsub("shellfish.that.the.patient.reacted.to..or.suspected.reacted.to...choice.", "", names(Omarallergy))

Omarallergy[][Omarallergy[] == "Checked"] <- 1

Omarallergy[][Omarallergy[] == "Unchecked"] <- 0

First step, make race categories, categorize everyone as either white, black, or other

col_names <- colnames(Omarallergy)
duplicate_cols <- col_names[duplicated(col_names)]

Omarallergy <- mutate(Omarallergy, pdemrace = ifelse(White. == 1, "1", ifelse(Black.or.African.American. == 1, "2", ifelse(American.Indian.or.Alaska.Native. == 1, "3", ifelse(Asian. == 1, "3", ifelse(Native.Hawaiian.or.Other.Pacific.Islander. == 1, "3", ifelse(not.documented. == 1, "3", 
                                                                        ifelse(otherRace. == 1, "3", 
                                                                               ifelse(not.documented..1 == 1, 3, "")))))))))

                                                                        
# table(Omarallergy$pdemrace) 

filtered_data <- Omarallergy %>%
  filter(pdemrace != 3)


z <- table(filtered_data$pdemrace)
rownames(z) <- c( "White","Black")


kbl(z) %>%
  kable_paper(full_width = F) %>% kable_styling(bootstrap_options = "striped", font_size = 15)

Var1	Freq
White	93
Black	102

Categories of allergy

No Difference in Age by Race

Black and white individuals are not different for Age (W = 5154.5, p = .296)

filtered_data$race <- dplyr::recode(filtered_data$pdemrace, "1" = "White", "2" = "Black")

View(filtered_data)
filtered_data$pdemrace <- as.factor(filtered_data$pdemrace)
# anova(lm(age.in.years~ pdemrace, data=filtered_data))


wilcox.test(age.in.years ~ pdemrace, data = filtered_data, 
        exact = FALSE, alternative = "two.sided")

## 
##  Wilcoxon rank sum test with continuity correction
## 
## data:  age.in.years by pdemrace
## W = 5154.5, p-value = 0.2963
## alternative hypothesis: true location shift is not equal to 0

filtered_data %>%
  group_by(race) %>%
  get_summary_stats(age.in.years, type = "common")

ggboxplot(filtered_data, x = "race", y = "age.in.years", 
          color ="race", palette = c("#00AFBB", "#E7B800"),
          ylab = "age (years)", xlab = "Race") + theme(legend.position = "none") + ggtitle("Figure 1. No Difference in Age in \nYears by Race") +   theme(title  = element_text(face="bold", size = 15))

No Difference in Age of Onset by Race

Black and white individuals do not have different age of onsets for shellfish allergy (W = 1041.5, p = .296)

# anova(lm(shellfish.age.of.allergy.onset ~ pdemrace, data=filtered_data))

wilcox.test(shellfish.age.of.allergy.onset ~ pdemrace, data = filtered_data, 
        exact = FALSE, alternative = "two.sided")

## 
##  Wilcoxon rank sum test with continuity correction
## 
## data:  shellfish.age.of.allergy.onset by pdemrace
## W = 1041.5, p-value = 0.2967
## alternative hypothesis: true location shift is not equal to 0

filtered_data %>%
  group_by(race) %>%
  get_summary_stats(shellfish.age.of.allergy.onset, type = "common")

ggboxplot(filtered_data, x = "race", y = "shellfish.age.of.allergy.onset", 
          color ="race", palette = c("#00AFBB", "#E7B800"),
          ylab = "age of onset of allergy (years)", xlab = "Race") + theme(legend.position = "none") + ggtitle("Figure 2. No Racial differences \nin SF Allergy Onset Age") + theme(title  = element_text(face="bold", size = 15))

compare black and white patients on roach and mites, and symptoms, for eahc of the seven symtpoms, compare on sex, age, and age of onset of the allergy

Difference in Sex by Race

proportion of females differs by sex (i.e., more black females) at greater than chance levels (χ² =7.5132, p = .006 )

filtered_data$Sex <- dplyr::recode(filtered_data$Sex, female = "0", male = "1", na.action = "na.omit")


filtered_data$Sex <- as.factor(filtered_data$Sex)

# table(filtered_data$Sex)

cont_tablesex <- table(filtered_data$pdemrace, filtered_data$Sex)
chi_squaredsex <- chisq.test(cont_tablesex)




rownames(cont_tablesex) <- c( "White","Black")
colnames(cont_tablesex) <- c( "Female","Male")

chi_squaredsex

## 
##  Pearson's Chi-squared test with Yates' continuity correction
## 
## data:  cont_tablesex
## X-squared = 7.5132, df = 1, p-value = 0.006125

kbl(cont_tablesex) %>%
  kable_paper(full_width = F) %>% kable_styling(bootstrap_options = "striped", font_size = 15)

	Female	Male
White	53	40
Black	78	24

No Difference in Roach by Race

(χ² = 4.9651e-05, p = .99; fisher’s exact test, p = .88)

filtered_data$rhinitis.specific.sensitization..skin.test.or.serum.IgE...choice.cockroach. <- as.factor(filtered_data$rhinitis.specific.sensitization..skin.test.or.serum.IgE...choice.cockroach.)



df1 = filtered_data %>% 
  count(race, rhinitis.specific.sensitization..skin.test.or.serum.IgE...choice.cockroach.) %>% 
  group_by(race) %>% 
  mutate(prop=prop.table(n))


cont_table <- table(filtered_data$pdemrace, filtered_data$rhinitis.specific.sensitization..skin.test.or.serum.IgE...choice.cockroach.)
chi_squared <- chisq.test(cont_table)
fisher_result <- fisher.test(cont_table)



rownames(cont_table) <- c( "White","Black")
colnames(cont_table) <- c( "Not","Sensitized")


kbl(cont_table) %>%
  kable_paper(full_width = F) %>% kable_styling(bootstrap_options = "striped", font_size = 15)

	Not	Sensitized
White	63	30
Black	68	34

chi_squared

## 
##  Pearson's Chi-squared test with Yates' continuity correction
## 
## data:  cont_table
## X-squared = 4.9651e-05, df = 1, p-value = 0.9944

fisher_result

## 
##  Fisher's Exact Test for Count Data
## 
## data:  cont_table
## p-value = 0.88
## alternative hypothesis: true odds ratio is not equal to 1
## 95 percent confidence interval:
##  0.552748 1.999211
## sample estimates:
## odds ratio 
##   1.049763

ggplot(df1, aes(race, fill = rhinitis.specific.sensitization..skin.test.or.serum.IgE...choice.cockroach.))+
geom_bar(aes(y = prop*100),
    position = "dodge", stat = "identity") + scale_fill_discrete(name = "Roach Sensitization", labels = c("No", "Yes"),type=c("cornflowerblue","red")) +theme_classic() + ylab("Proportion (%)") + xlab("Race") + ggtitle("Figure 3. No Racial differences in \n Roach Sensitization") + theme(title  = element_text(face="bold", size = 15))

No Difference in Mites by Race

(χ² = 1.3362, p = 0.2477; fisher’s exact test, p = .199)

filtered_data$rhinitis.specific.sensitization..skin.test.or.serum.IgE...choice.dust.mite. <- as.factor(filtered_data$rhinitis.specific.sensitization..skin.test.or.serum.IgE...choice.dust.mite.)



df1 = filtered_data %>% 
  count(race, rhinitis.specific.sensitization..skin.test.or.serum.IgE...choice.dust.mite.) %>% 
  group_by(race) %>% 
  mutate(prop=prop.table(n))



cont_table <- table(filtered_data$pdemrace, filtered_data$rhinitis.specific.sensitization..skin.test.or.serum.IgE...choice.dust.mite.)
chi_squared <- chisq.test(cont_table)
fisher_result <- fisher.test(cont_table)



rownames(cont_table) <- c( "White","Black")
colnames(cont_table) <- c( "Not","Sensitized")


kbl(cont_table) %>%
  kable_paper(full_width = F) %>% kable_styling(bootstrap_options = "striped", font_size = 15)

	Not	Sensitized
White	46	47
Black	41	61

chi_squared

## 
##  Pearson's Chi-squared test with Yates' continuity correction
## 
## data:  cont_table
## X-squared = 1.3362, df = 1, p-value = 0.2477

fisher_result

## 
##  Fisher's Exact Test for Count Data
## 
## data:  cont_table
## p-value = 0.1987
## alternative hypothesis: true odds ratio is not equal to 1
## 95 percent confidence interval:
##  0.7941506 2.6713774
## sample estimates:
## odds ratio 
##   1.453293

ggplot(df1, aes(race, fill = rhinitis.specific.sensitization..skin.test.or.serum.IgE...choice.dust.mite.))+
geom_bar(aes(y = prop*100),
    position = "dodge", stat = "identity") + scale_fill_discrete(name = "Dust Mite Sensitization", labels = c("No", "Yes"),type=c("cornflowerblue","red")) +theme_classic() + ylab("Proportion (%)") + xlab("Race") + ggtitle("Figure 4. No Racial differences in \n Dust Mite  Sensitization") + theme(title  = element_text(face="bold", size = 15))

Difference in Symptoms

Hives/angioedema - marginal difference by race

non-significant difference in hives; black individuals >> white individuals

filtered_data$shellfish.symptoms.with.any.prior.reaction..choice.Skin..hives.or.angioedema.. <- as.factor(filtered_data$shellfish.symptoms.with.any.prior.reaction..choice.Skin..hives.or.angioedema..)

cont_table <- table(filtered_data$pdemrace, filtered_data$shellfish.symptoms.with.any.prior.reaction..choice.Skin..hives.or.angioedema..)
chi_squared <- chisq.test(cont_table)
fisher_result <- fisher.test(cont_table)



rownames(cont_table) <- c( "White","Black")
colnames(cont_table) <- c( "No","Yes")


kbl(cont_table) %>%
  kable_paper(full_width = F) %>% kable_styling(bootstrap_options = "striped", font_size = 15)

	No	Yes
White	31	62
Black	21	81

chi_squared

## 
##  Pearson's Chi-squared test with Yates' continuity correction
## 
## data:  cont_table
## X-squared = 3.4153, df = 1, p-value = 0.06459

fisher_result

## 
##  Fisher's Exact Test for Count Data
## 
## data:  cont_table
## p-value = 0.05228
## alternative hypothesis: true odds ratio is not equal to 1
## 95 percent confidence interval:
##  0.9649947 3.8888905
## sample estimates:
## odds ratio 
##   1.922028

df1 = filtered_data %>% 
  count(race, shellfish.symptoms.with.any.prior.reaction..choice.Skin..hives.or.angioedema..) %>% 
  group_by(race) %>% 
  mutate(prop=prop.table(n))

ggplot(df1, aes(race, fill = shellfish.symptoms.with.any.prior.reaction..choice.Skin..hives.or.angioedema..))+
geom_bar(aes(y = prop*100),
    position = "dodge", stat = "identity") + scale_fill_discrete(name = "SkinSymptom", labels = c("No", "Yes"),type=c("cornflowerblue","red")) +theme_classic() + ylab("Proportion (%)") + xlab("Race") + ggtitle("Figure 6. Marginal Diff in  Skin Symptoms") + theme(title  = element_text(face="bold", size = 15))

GI, Emesis/Diarrhea - no diff by race

filtered_data$shellfish.symptoms.with.any.prior.reaction..choice.GI..nausea..vomiting..or.diarrhea.. <- as.factor(filtered_data$shellfish.symptoms.with.any.prior.reaction..choice.GI..nausea..vomiting..or.diarrhea..)


cont_table <- table(filtered_data$pdemrace, filtered_data$shellfish.symptoms.with.any.prior.reaction..choice.GI..nausea..vomiting..or.diarrhea..)
chi_squared <- chisq.test(cont_table)
fisher_result <- fisher.test(cont_table)



rownames(cont_table) <- c( "White","Black")
colnames(cont_table) <- c( "No","Yes")


kbl(cont_table) %>%
  kable_paper(full_width = F) %>% kable_styling(bootstrap_options = "striped", font_size = 15)

	No	Yes
White	71	22
Black	83	19

chi_squared

## 
##  Pearson's Chi-squared test with Yates' continuity correction
## 
## data:  cont_table
## X-squared = 0.46889, df = 1, p-value = 0.4935

fisher_result

## 
##  Fisher's Exact Test for Count Data
## 
## data:  cont_table
## p-value = 0.482
## alternative hypothesis: true odds ratio is not equal to 1
## 95 percent confidence interval:
##  0.347849 1.561518
## sample estimates:
## odds ratio 
##  0.7399337

df1 = filtered_data %>% 
  count(race, shellfish.symptoms.with.any.prior.reaction..choice.GI..nausea..vomiting..or.diarrhea..) %>% 
  group_by(race) %>% 
  mutate(prop=prop.table(n))

ggplot(df1, aes(race, fill = shellfish.symptoms.with.any.prior.reaction..choice.GI..nausea..vomiting..or.diarrhea..))+
geom_bar(aes(y = prop*100),
    position = "dodge", stat = "identity") + scale_fill_discrete(name = "GI", labels = c("No", "Yes"),type=c("cornflowerblue","red")) +theme_classic() + ylab("Proportion (%)") + xlab("Race") + ggtitle("Figure 7. GI Symptoms") + theme(title  = element_text(face="bold", size = 15))

Respiratory - No Racial differences by race

filtered_data$shellfish.symptoms.with.any.prior.reaction..choice.Respiratory..wheezing..stridor..voice.change..or.shortness.of.breath.. <- as.factor(filtered_data$shellfish.symptoms.with.any.prior.reaction..choice.Respiratory..wheezing..stridor..voice.change..or.shortness.of.breath..)

filtered_data$respiratory <- as.factor(filtered_data$shellfish.symptoms.with.any.prior.reaction..choice.Respiratory..wheezing..stridor..voice.change..or.shortness.of.breath..)


cont_table <- table(filtered_data$pdemrace, filtered_data$respiratory)
chi_squared <- chisq.test(cont_table)
fisher_result <- fisher.test(cont_table)



rownames(cont_table) <- c( "White","Black")
colnames(cont_table) <- c( "No","Yes")


kbl(cont_table) %>%
  kable_paper(full_width = F) %>% kable_styling(bootstrap_options = "striped", font_size = 15)

	No	Yes
White	46	47
Black	56	46

chi_squared

## 
##  Pearson's Chi-squared test with Yates' continuity correction
## 
## data:  cont_table
## X-squared = 0.37954, df = 1, p-value = 0.5378

fisher_result

## 
##  Fisher's Exact Test for Count Data
## 
## data:  cont_table
## p-value = 0.4753
## alternative hypothesis: true odds ratio is not equal to 1
## 95 percent confidence interval:
##  0.4402826 1.4675397
## sample estimates:
## odds ratio 
##  0.8048606

df1 = filtered_data %>% 
  count(race, respiratory) %>% 
  group_by(race) %>% 
  mutate(prop=prop.table(n))

ggplot(df1, aes(race, fill = respiratory))+
geom_bar(aes(y = prop*100),
    position = "dodge", stat = "identity") + scale_fill_discrete(name = "respiratory", labels = c("No", "Yes"),type=c("cornflowerblue","red")) +theme_classic() + ylab("Proportion (%)") + xlab("Race") + ggtitle("Figure 8. Respiratory Symptoms") + theme(title  = element_text(face="bold", size = 15))

Medical Comorbidities

Allergic Rhinitis - no difference in Race

cont_table_medcomorbidAR <- table(filtered_data$pdemrace, filtered_data$medical.comorbidities.at.the.time.of.food.allergy.diagnosis..choice.allergic.rhinitis)



chisq.test(cont_table_medcomorbidAR)

## 
##  Pearson's Chi-squared test with Yates' continuity correction
## 
## data:  cont_table_medcomorbidAR
## X-squared = 0.75331, df = 1, p-value = 0.3854

fisher.test(cont_table_medcomorbidAR)

## 
##  Fisher's Exact Test for Count Data
## 
## data:  cont_table_medcomorbidAR
## p-value = 0.322
## alternative hypothesis: true odds ratio is not equal to 1
## 95 percent confidence interval:
##  0.6970152 2.8681325
## sample estimates:
## odds ratio 
##   1.408603

rownames(cont_table) <- c( "White","Black")
colnames(cont_table) <- c( "No","Yes")


kbl(cont_table) %>%
  kable_paper(full_width = F) %>% kable_styling(bootstrap_options = "striped", font_size = 15)

	No	Yes
White	46	47
Black	56	46

Asthma - no difference in Race

cont_table_medcomorbid <- table(filtered_data$pdemrace, filtered_data$medical.comorbidities.at.the.time.of.food.allergy.diagnosis..choice.asthma)



chisq.test(cont_table_medcomorbid)

## 
##  Pearson's Chi-squared test with Yates' continuity correction
## 
## data:  cont_table_medcomorbid
## X-squared = 2.3469, df = 1, p-value = 0.1255

fisher.test(cont_table_medcomorbid)

## 
##  Fisher's Exact Test for Count Data
## 
## data:  cont_table_medcomorbid
## p-value = 0.1076
## alternative hypothesis: true odds ratio is not equal to 1
## 95 percent confidence interval:
##  0.8830289 3.0678121
## sample estimates:
## odds ratio 
##   1.638614

rownames(cont_table_medcomorbid) <- c( "White","Black")
colnames(cont_table_medcomorbid) <- c( "No","Yes")


kbl(cont_table_medcomorbid) %>%
  kable_paper(full_width = F) %>% kable_styling(bootstrap_options = "striped", font_size = 15)

	No	Yes
White	62	31
Black	56	46

Eczema - no difference in Race

cont_table_medcomorbidEC<- table(filtered_data$pdemrace, filtered_data$medical.comorbidities.at.the.time.of.food.allergy.diagnosis..choice.eczema)
chisq.test(cont_table_medcomorbidEC)

## 
##  Pearson's Chi-squared test with Yates' continuity correction
## 
## data:  cont_table_medcomorbidEC
## X-squared = 2.9762e-31, df = 1, p-value = 1

fisher.test(cont_table_medcomorbidEC)

## 
##  Fisher's Exact Test for Count Data
## 
## data:  cont_table_medcomorbidEC
## p-value = 1
## alternative hypothesis: true odds ratio is not equal to 1
## 95 percent confidence interval:
##  0.3896333 2.5146574
## sample estimates:
## odds ratio 
##  0.9860244

rownames(cont_table_medcomorbidEC) <- c( "White","Black")
colnames(cont_table_medcomorbidEC) <- c( "No","Yes")


kbl(cont_table_medcomorbidEC) %>%
  kable_paper(full_width = F) %>% kable_styling(bootstrap_options = "striped", font_size = 15)

	No	Yes
White	81	12
Black	89	13

Contact Derm - no difference in Race

cont_table_medcomorbidCD<- table(filtered_data$pdemrace, filtered_data$medical.comorbidities.at.the.time.of.food.allergy.diagnosis..choice.contact.dermatitis)
fisher.test(cont_table_medcomorbidCD)

## 
##  Fisher's Exact Test for Count Data
## 
## data:  cont_table_medcomorbidCD
## p-value = 0.4274
## alternative hypothesis: true odds ratio is not equal to 1
## 95 percent confidence interval:
##  0.03950996 3.20252303
## sample estimates:
## odds ratio 
##  0.4467991

rownames(cont_table_medcomorbidCD) <- c( "White","Black")
colnames(cont_table_medcomorbidCD) <- c( "No","Yes")


kbl(cont_table_medcomorbidCD) %>%
  kable_paper(full_width = F) %>% kable_styling(bootstrap_options = "striped", font_size = 15)

	No	Yes
White	89	4
Black	100	2

Oral Allergy Syndrome - no difference in Race

cont_table_medcomorbidOAS<- table(filtered_data$pdemrace, filtered_data$medical.comorbidities.at.the.time.of.food.allergy.diagnosis..choice.oral.allergy.syndrome)
fisher.test(cont_table_medcomorbidOAS)

## 
##  Fisher's Exact Test for Count Data
## 
## data:  cont_table_medcomorbidOAS
## p-value = 0.121
## alternative hypothesis: true odds ratio is not equal to 1
## 95 percent confidence interval:
##    0.6727025 266.9661642
## sample estimates:
## odds ratio 
##   5.708437

rownames(cont_table_medcomorbidOAS) <- c( "White","Black")
colnames(cont_table_medcomorbidOAS) <- c( "No","Yes")


kbl(cont_table_medcomorbidOAS) %>%
  kable_paper(full_width = F) %>% kable_styling(bootstrap_options = "striped", font_size = 15)

	No	Yes
White	92	1
Black	96	6

Anxiety - no difference in Race

cont_table_medcomorbidANXS<- table(filtered_data$pdemrace, filtered_data$medical.comorbidities.at.the.time.of.food.allergy.diagnosis..choice.anxiety)
chisq.test(cont_table_medcomorbidANXS)

## 
##  Pearson's Chi-squared test with Yates' continuity correction
## 
## data:  cont_table_medcomorbidANXS
## X-squared = 0.80343, df = 1, p-value = 0.3701

fisher.test(cont_table_medcomorbidANXS)

## 
##  Fisher's Exact Test for Count Data
## 
## data:  cont_table_medcomorbidANXS
## p-value = 0.2837
## alternative hypothesis: true odds ratio is not equal to 1
## 95 percent confidence interval:
##  0.2304729 1.5833503
## sample estimates:
## odds ratio 
##  0.6149051

rownames(cont_table_medcomorbidANXS) <- c( "White","Black")
colnames(cont_table_medcomorbidANXS) <- c( "No","Yes")


kbl(cont_table_medcomorbidANXS) %>%
  kable_paper(full_width = F) %>% kable_styling(bootstrap_options = "striped", font_size = 15)

	No	Yes
White	79	14
Black	92	10

Other Psych - White individuals >> Black Individuals

filtered_data$otherpsych <- as.factor(filtered_data$medical.comorbidities.at.the.time.of.food.allergy.diagnosis..choice.other.psychological.diagnosis..depression..bipolar..schizophrenia..etc.)




cont_table_medcomorbidPSY<- table(filtered_data$pdemrace, filtered_data$medical.comorbidities.at.the.time.of.food.allergy.diagnosis..choice.other.psychological.diagnosis..depression..bipolar..schizophrenia..etc.)
chisq.test(cont_table_medcomorbidPSY)

## 
##  Pearson's Chi-squared test with Yates' continuity correction
## 
## data:  cont_table_medcomorbidPSY
## X-squared = 7.2221, df = 1, p-value = 0.007201

fisher.test(cont_table_medcomorbidPSY)

## 
##  Fisher's Exact Test for Count Data
## 
## data:  cont_table_medcomorbidPSY
## p-value = 0.004599
## alternative hypothesis: true odds ratio is not equal to 1
## 95 percent confidence interval:
##  0.1060983 0.7386840
## sample estimates:
## odds ratio 
##  0.2936171

rownames(cont_table_medcomorbidPSY) <- c( "White","Black")
colnames(cont_table_medcomorbidPSY) <- c( "No","Yes")


kbl(cont_table_medcomorbidPSY) %>%
  kable_paper(full_width = F) %>% kable_styling(bootstrap_options = "striped", font_size = 15)

	No	Yes
White	72	21
Black	94	8

df1 = filtered_data %>% 
  count(race, otherpsych) %>% 
  group_by(race) %>% 
  mutate(prop=prop.table(n))

ggplot(df1, aes(race, fill = otherpsych))+
geom_bar(aes(y = prop*100),
    position = "dodge", stat = "identity") + scale_fill_discrete(name = "other psych", labels = c("No", "Yes"),type=c("cornflowerblue","red")) +theme_classic() + ylab("Proportion (%)") + xlab("Race") + ggtitle("Figure 9. Other Psych DX") + theme(title  = element_text(face="bold", size = 15))

Cardiac - No Racial differences

cont_table_medcomorbidhypertension<- table(filtered_data$pdemrace, filtered_data$medical.comorbidities.at.the.time.of.food.allergy.diagnosis..choice.cardiac..hypertension..coronary.artery.dz..etc.)
chisq.test(cont_table_medcomorbidhypertension)

## 
##  Pearson's Chi-squared test with Yates' continuity correction
## 
## data:  cont_table_medcomorbidhypertension
## X-squared = 1.9437, df = 1, p-value = 0.1633

fisher.test(cont_table_medcomorbidhypertension)

## 
##  Fisher's Exact Test for Count Data
## 
## data:  cont_table_medcomorbidhypertension
## p-value = 0.145
## alternative hypothesis: true odds ratio is not equal to 1
## 95 percent confidence interval:
##  0.8338467 3.3820097
## sample estimates:
## odds ratio 
##   1.664177

rownames(cont_table_medcomorbidhypertension) <- c( "White","Black")
colnames(cont_table_medcomorbidhypertension) <- c( "No","Yes")


kbl(cont_table_medcomorbidhypertension) %>%
  kable_paper(full_width = F) %>% kable_styling(bootstrap_options = "striped", font_size = 15)

	No	Yes
White	73	20
Black	70	32

GI - No Racial differences

cont_table_GI<- table(filtered_data$pdemrace, filtered_data$medical.comorbidities.at.the.time.of.food.allergy.diagnosis..choice.gastrointestinal)
chisq.test(cont_table_GI)

## 
##  Pearson's Chi-squared test with Yates' continuity correction
## 
## data:  cont_table_GI
## X-squared = 0.028037, df = 1, p-value = 0.867

fisher.test(cont_table_GI)

## 
##  Fisher's Exact Test for Count Data
## 
## data:  cont_table_GI
## p-value = 0.7616
## alternative hypothesis: true odds ratio is not equal to 1
## 95 percent confidence interval:
##  0.5790866 2.1087649
## sample estimates:
## odds ratio 
##   1.102891

rownames(cont_table_GI) <- c( "White","Black")
colnames(cont_table_GI) <- c( "No","Yes")


kbl(cont_table_GI) %>%
  kable_paper(full_width = F) %>% kable_styling(bootstrap_options = "striped", font_size = 15)

	No	Yes
White	64	29
Black	68	34

Alpha Gal - No Racial differences

cont_table_alphagal<- table(filtered_data$pdemrace, filtered_data$medical.comorbidities.at.the.time.of.food.allergy.diagnosis..choice.alpha.gal.allergy)
fisher.test(cont_table_alphagal)

## 
##  Fisher's Exact Test for Count Data
## 
## data:  cont_table_alphagal
## p-value = 0.04999
## alternative hypothesis: true odds ratio is not equal to 1
## 95 percent confidence interval:
##  0.000000 1.361982
## sample estimates:
## odds ratio 
##          0

rownames(cont_table_alphagal) <- c( "White","Black")
colnames(cont_table_alphagal) <- c( "No","Yes")


kbl(cont_table_alphagal) %>%
  kable_paper(full_width = F) %>% kable_styling(bootstrap_options = "striped", font_size = 15)

	No	Yes
White	89	4
Black	102	0

Other Food Allergy - No Racial differences

cont_table_otherfoodallergy<- table(filtered_data$pdemrace, filtered_data$medical.comorbidities.at.the.time.of.food.allergy.diagnosis..choice.other.food.allergy)
chisq.test(cont_table_otherfoodallergy)

## 
##  Pearson's Chi-squared test with Yates' continuity correction
## 
## data:  cont_table_otherfoodallergy
## X-squared = 0.19538, df = 1, p-value = 0.6585

fisher.test(cont_table_otherfoodallergy)

## 
##  Fisher's Exact Test for Count Data
## 
## data:  cont_table_otherfoodallergy
## p-value = 0.6231
## alternative hypothesis: true odds ratio is not equal to 1
## 95 percent confidence interval:
##  0.6094929 2.4638628
## sample estimates:
## odds ratio 
##   1.219878

rownames(cont_table_otherfoodallergy) <- c( "White","Black")
colnames(cont_table_otherfoodallergy) <- c( "No","Yes")

kbl(cont_table_otherfoodallergy) %>%
  kable_paper(full_width = F) %>% kable_styling(bootstrap_options = "striped", font_size = 15)

	No	Yes
White	71	22
Black	74	28

Drug Allergy - racial differences

filtered_data$drugallergy <- as.factor(filtered_data$medical.comorbidities.at.the.time.of.food.allergy.diagnosis..choice.drug.allergy)


cont_table_otherdrugallergy<- table(filtered_data$pdemrace, filtered_data$medical.comorbidities.at.the.time.of.food.allergy.diagnosis..choice.drug.allergy)
chisq.test(cont_table_otherdrugallergy)

## 
##  Pearson's Chi-squared test with Yates' continuity correction
## 
## data:  cont_table_otherdrugallergy
## X-squared = 4.0853, df = 1, p-value = 0.04326

fisher.test(cont_table_otherdrugallergy)

## 
##  Fisher's Exact Test for Count Data
## 
## data:  cont_table_otherdrugallergy
## p-value = 0.03195
## alternative hypothesis: true odds ratio is not equal to 1
## 95 percent confidence interval:
##  0.2891080 0.9819947
## sample estimates:
## odds ratio 
##    0.53491

rownames(cont_table_otherdrugallergy) <- c( "White","Black")
colnames(cont_table_otherdrugallergy) <- c( "No","Yes")


kbl(cont_table_otherdrugallergy) %>%
  kable_paper(full_width = F) %>% kable_styling(bootstrap_options = "striped", font_size = 15)

	No	Yes
White	44	49
Black	64	38

df1 = filtered_data %>% 
  count(race, drugallergy) %>% 
  group_by(race) %>% 
  mutate(prop=prop.table(n))

ggplot(df1, aes(race, fill = drugallergy))+
geom_bar(aes(y = prop*100),
    position = "dodge", stat = "identity") + scale_fill_discrete(name = "drug allergy", labels = c("No", "Yes"),type=c("cornflowerblue","red")) +theme_classic() + ylab("Proportion (%)") + xlab("Race") + ggtitle("Figure 10. Drug Allergy Greater in White Individuals") + theme(title  = element_text(face="bold", size = 15))

Insect Allergy - racial differences

filtered_data$insectallergy <- as.factor(filtered_data$medical.comorbidities.at.the.time.of.food.allergy.diagnosis..choice.stinging.insect.allergy..hymenoptera.)


cont_table_insectallergy<- table(filtered_data$pdemrace, filtered_data$medical.comorbidities.at.the.time.of.food.allergy.diagnosis..choice.stinging.insect.allergy..hymenoptera.)
chisq.test(cont_table_insectallergy)

## 
##  Pearson's Chi-squared test with Yates' continuity correction
## 
## data:  cont_table_insectallergy
## X-squared = 4.0889, df = 1, p-value = 0.04317

fisher.test(cont_table_insectallergy)

## 
##  Fisher's Exact Test for Count Data
## 
## data:  cont_table_insectallergy
## p-value = 0.02745
## alternative hypothesis: true odds ratio is not equal to 1
## 95 percent confidence interval:
##  0.01927547 0.94409617
## sample estimates:
## odds ratio 
##  0.1880703

rownames(cont_table_insectallergy) <- c( "White","Black")
colnames(cont_table_insectallergy) <- c( "No","Yes")


kbl(cont_table_insectallergy) %>%
  kable_paper(full_width = F) %>% kable_styling(bootstrap_options = "striped", font_size = 15)

	No	Yes
White	84	9
Black	100	2

df1 = filtered_data %>% 
  count(race, insectallergy) %>% 
  group_by(race) %>% 
  mutate(prop=prop.table(n))

ggplot(df1, aes(race, fill = insectallergy))+
geom_bar(aes(y = prop*100),
    position = "dodge", stat = "identity") + scale_fill_discrete(name = "insect allergy", labels = c("No", "Yes"),type=c("cornflowerblue","red")) +theme_classic() + ylab("Proportion (%)") + xlab("Race") + ggtitle("Figure 11. Insect Allergy") + theme(title  = element_text(face="bold", size = 15))

Medications at Time of DX

PPI - No Racial differences

PPI <- table(filtered_data$pdemrace, filtered_data$medications.prescribed.at.time.of.food.allergy.diagnosis..choice.proton.pump.inhibitor.)
chisq.test(PPI)

## 
##  Pearson's Chi-squared test with Yates' continuity correction
## 
## data:  PPI
## X-squared = 0.45953, df = 1, p-value = 0.4978

fisher.test(PPI)

## 
##  Fisher's Exact Test for Count Data
## 
## data:  PPI
## p-value = 0.4655
## alternative hypothesis: true odds ratio is not equal to 1
## 95 percent confidence interval:
##  0.3323381 1.5928631
## sample estimates:
## odds ratio 
##  0.7311916

Beta Blockers - No Racial differences

BB <- table(filtered_data$pdemrace, filtered_data$medications.prescribed.at.time.of.food.allergy.diagnosis..choice.beta.blocker.)
fisher.test(BB)

## 
##  Fisher's Exact Test for Count Data
## 
## data:  BB
## p-value = 0.6199
## alternative hypothesis: true odds ratio is not equal to 1
## 95 percent confidence interval:
##  0.435641 4.324629
## sample estimates:
## odds ratio 
##   1.333438

ACE/ARB - No Racial differences

AceArb <- table(filtered_data$pdemrace, filtered_data$medications.prescribed.at.time.of.food.allergy.diagnosis..choice.ACE.ARB.)
fisher.test(AceArb)

## 
##  Fisher's Exact Test for Count Data
## 
## data:  AceArb
## p-value = 0.2204
## alternative hypothesis: true odds ratio is not equal to 1
## 95 percent confidence interval:
##  0.7255089 4.5721512
## sample estimates:
## odds ratio 
##   1.773398

NSAID - No Racial differences

NSAID <- table(filtered_data$pdemrace, filtered_data$medications.documented.at.the.time.of.chart.review..PMHx.section...not.already.included.in.previous.medication.question...choice.NSAID..prescribed.or.documented.in.note..)
fisher.test(NSAID)

## 
##  Fisher's Exact Test for Count Data
## 
## data:  NSAID
## p-value = 0.08581
## alternative hypothesis: true odds ratio is not equal to 1
## 95 percent confidence interval:
##  0.8814582 4.9878702
## sample estimates:
## odds ratio 
##   2.042624

Sulfasalazine - noone taking it

Sulfasalazine <- table(filtered_data$pdemrace, filtered_data$medications.documented.at.the.time.of.chart.review..PMHx.section...not.already.included.in.previous.medication.question...choice.sulfasalazine..azulfidine..or.mesalamine..5.ASA..)

Immune modulator - No Racial differences

ImmuneMod <- table(filtered_data$pdemrace, filtered_data$medications.documented.at.the.time.of.chart.review..PMHx.section...not.already.included.in.previous.medication.question...choice.immune.modulator..immune.suppressant..biologic..cyclophsophamide..MMF..calcineurin.inhibitor..mTOR.inhibitor..methotrexate..azathioprene..)
fisher.test(ImmuneMod)

## 
##  Fisher's Exact Test for Count Data
## 
## data:  ImmuneMod
## p-value = 1
## alternative hypothesis: true odds ratio is not equal to 1
## 95 percent confidence interval:
##  0.2689808 4.7249200
## sample estimates:
## odds ratio 
##   1.099444

Oral steroid - No Racial differences

oralsteroid <- table(filtered_data$pdemrace, filtered_data$medications.prescribed.at.time.of.food.allergy.diagnosis..choice.daily.oral.steroid.)
fisher.test(oralsteroid)

## 
##  Fisher's Exact Test for Count Data
## 
## data:  oralsteroid
## p-value = 0.7232
## alternative hypothesis: true odds ratio is not equal to 1
## 95 percent confidence interval:
##   0.2906562 10.2230425
## sample estimates:
## odds ratio 
##   1.543004

Shellfish allergy severity

No difference in antihistamine only or undocumented any prior reaction

notxantihisonly <- table(filtered_data$pdemrace, filtered_data$shellfish.severity.of.any.prior.reaction..choice.no.treatment.or.antihistamine.only)
fisher.test(notxantihisonly)

## 
##  Fisher's Exact Test for Count Data
## 
## data:  notxantihisonly
## p-value = 0.8809
## alternative hypothesis: true odds ratio is not equal to 1
## 95 percent confidence interval:
##  0.5044299 1.8033134
## sample estimates:
## odds ratio 
##  0.9533317

notdocument <- table(filtered_data$pdemrace, filtered_data$shellfish.severity.of.any.prior.reaction..choice.not.documented)
fisher.test(notdocument)

## 
##  Fisher's Exact Test for Count Data
## 
## data:  notdocument
## p-value = 0.1987
## alternative hypothesis: true odds ratio is not equal to 1
## 95 percent confidence interval:
##  0.3738865 1.2546803
## sample estimates:
## odds ratio 
##  0.6864614

Severity, Requiring ED and/or EPI, Black individuals >> White Individuals

filtered_data$epied <- as.factor(filtered_data$shellfish.severity.of.any.prior.reaction..choice.presented.to.ED.and..or.epinephrine.use)



EDepinephrine <- table(filtered_data$pdemrace, filtered_data$shellfish.severity.of.any.prior.reaction..choice.presented.to.ED.and..or.epinephrine.use)
fisher.test(EDepinephrine)

## 
##  Fisher's Exact Test for Count Data
## 
## data:  EDepinephrine
## p-value = 0.03521
## alternative hypothesis: true odds ratio is not equal to 1
## 95 percent confidence interval:
##  1.023373 6.042139
## sample estimates:
## odds ratio 
##   2.405833

df1 = filtered_data %>% 
  count(race, epied) %>% 
  group_by(race) %>% 
  mutate(prop=prop.table(n))

ggplot(df1, aes(race, fill = epied))+
geom_bar(aes(y = prop*100),
    position = "dodge", stat = "identity") + scale_fill_discrete(name = "Epi-ED", labels = c("No", "Yes"),type=c("cornflowerblue","red")) +theme_classic() + ylab("Proportion (%)") + xlab("Race") + ggtitle("Figure 12. Reaction presenting to ED \n Epi Use") + theme(title  = element_text(face="bold", size = 15))

Prescribed EPI at Time of DX - No Racial differences by Race

EPIDX <- table(filtered_data$pdemrace, filtered_data$prescribed.epinephrine.at.time.of.food.allergy.diagnosis)
fisher.test(EPIDX)

## 
##  Fisher's Exact Test for Count Data
## 
## data:  EPIDX
## p-value = 0.3569
## alternative hypothesis: true odds ratio is not equal to 1
## 95 percent confidence interval:
##  0.4856944 9.5813046
## sample estimates:
## odds ratio 
##   1.987209

chisq.test(EPIDX)

## 
##  Pearson's Chi-squared test with Yates' continuity correction
## 
## data:  EPIDX
## X-squared = 0.60715, df = 1, p-value = 0.4359

Shellfish that patient is allergic to

No diff in clam, crab, lobster, scallops, or shrimp

clam <- table(filtered_data$pdemrace, filtered_data$clam.)
fisher.test(clam)

## 
##  Fisher's Exact Test for Count Data
## 
## data:  clam
## p-value = 0.1049
## alternative hypothesis: true odds ratio is not equal to 1
## 95 percent confidence interval:
##  0.003653862 1.611823791
## sample estimates:
## odds ratio 
##  0.1756135

crab <- table(filtered_data$pdemrace, filtered_data$crab.)
fisher.test(crab)

## 
##  Fisher's Exact Test for Count Data
## 
## data:  crab
## p-value = 1
## alternative hypothesis: true odds ratio is not equal to 1
## 95 percent confidence interval:
##  0.5155848 2.1391279
## sample estimates:
## odds ratio 
##   1.047585

chisq.test(crab)

## 
##  Pearson's Chi-squared test with Yates' continuity correction
## 
## data:  crab
## X-squared = 1.0765e-30, df = 1, p-value = 1

lobster <- table(filtered_data$pdemrace, filtered_data$lobster.)
fisher.test(lobster)

## 
##  Fisher's Exact Test for Count Data
## 
## data:  lobster
## p-value = 0.4359
## alternative hypothesis: true odds ratio is not equal to 1
## 95 percent confidence interval:
##  0.3031279 1.6503377
## sample estimates:
## odds ratio 
##  0.7124862

chisq.test(lobster)

## 
##  Pearson's Chi-squared test with Yates' continuity correction
## 
## data:  lobster
## X-squared = 0.45242, df = 1, p-value = 0.5012

scallops <- table(filtered_data$pdemrace, filtered_data$scallops.)
fisher.test(scallops)

## 
##  Fisher's Exact Test for Count Data
## 
## data:  scallops
## p-value = 0.3922
## alternative hypothesis: true odds ratio is not equal to 1
## 95 percent confidence interval:
##  0.1360041 1.9888490
## sample estimates:
## odds ratio 
##   0.549369

chisq.test(scallops)

## 
##  Pearson's Chi-squared test with Yates' continuity correction
## 
## data:  scallops
## X-squared = 0.55833, df = 1, p-value = 0.4549

shrimp <- table(filtered_data$pdemrace, filtered_data$shrimp.)
fisher.test(shrimp)

## 
##  Fisher's Exact Test for Count Data
## 
## data:  shrimp
## p-value = 0.3176
## alternative hypothesis: true odds ratio is not equal to 1
## 95 percent confidence interval:
##  0.7310606 2.4542388
## sample estimates:
## odds ratio 
##   1.337252

chisq.test(shrimp)

## 
##  Pearson's Chi-squared test with Yates' continuity correction
## 
## data:  shrimp
## X-squared = 0.75257, df = 1, p-value = 0.3857

Oyster - more white individuals allergic

oyster <- table(filtered_data$pdemrace, filtered_data$oyster.)
fisher.test(oyster)

## 
##  Fisher's Exact Test for Count Data
## 
## data:  oyster
## p-value = 0.02329
## alternative hypothesis: true odds ratio is not equal to 1
## 95 percent confidence interval:
##  0.0000000 0.9731435
## sample estimates:
## odds ratio 
##          0

rownames(oyster) <- c( "White","Black")
colnames(oyster) <- c( "No","Yes")


kbl(oyster) %>%
  kable_paper(full_width = F) %>% kable_styling(bootstrap_options = "striped", font_size = 15)

	No	Yes
White	88	5
Black	102	0

filtered_data$oysteral <- as.factor(filtered_data$oyster.)

df1 = filtered_data %>% 
  count(race, oysteral) %>% 
  group_by(race) %>% 
  mutate(prop=prop.table(n))

ggplot(df1, aes(race, fill = oysteral))+
geom_bar(aes(y = prop*100),
    position = "dodge", stat = "identity") + scale_fill_discrete(name = "oyster allergy", labels = c("No", "Yes"),type=c("cornflowerblue","red")) +theme_classic() + ylab("Proportion (%)") + xlab("Race") + ggtitle("Figure 13. Oyster Allergy") + theme(title  = element_text(face="bold", size = 15))

Omar AI project

Noura Elsayed

05/18/24

Read in Data

First step, make race categories, categorize everyone as either white, black, or other

Categories of allergy

No Difference in Age by Race

No Difference in Age of Onset by Race

Difference in Sex by Race

No Difference in Roach by Race

No Difference in Mites by Race

Difference in Symptoms

Hives/angioedema - marginal difference by race

GI, Emesis/Diarrhea - no diff by race

Respiratory - No Racial differences by race

Medical Comorbidities

Allergic Rhinitis - no difference in Race

Asthma - no difference in Race

Eczema - no difference in Race

Contact Derm - no difference in Race

Oral Allergy Syndrome - no difference in Race

Anxiety - no difference in Race

Other Psych - White individuals >> Black Individuals

Cardiac - No Racial differences

GI - No Racial differences

Alpha Gal - No Racial differences

Other Food Allergy - No Racial differences

Drug Allergy - racial differences

Insect Allergy - racial differences

Medications at Time of DX

PPI - No Racial differences

Beta Blockers - No Racial differences

ACE/ARB - No Racial differences

NSAID - No Racial differences

Sulfasalazine - noone taking it

Immune modulator - No Racial differences

Oral steroid - No Racial differences

Shellfish allergy severity

No difference in antihistamine only or undocumented any prior reaction

Severity, Requiring ED and/or EPI, Black individuals >> White Individuals

Prescribed EPI at Time of DX - No Racial differences by Race

Shellfish that patient is allergic to

No diff in clam, crab, lobster, scallops, or shrimp

Oyster - more white individuals allergic