02a_Display_Hospitalisations.Rmd

---
title: "Display Emergency Hospitalisations"
output: html_document
---

```{r setup, include=FALSE}
library(ggplot2)
library(knitr)
knitr::opts_chunk$set(echo = FALSE, warning=FALSE, error=FALSE, message=FALSE)

descriptors = c ("ageYear", "n_risk_gps", "Sex", "age_gp", "covid_vs", "flu_vs", "urban_rural_classification", "simd2020_sc_quintile", "icu_admit_age", "death_28", "hosp_los_gp", "num_prev_admission_gp", "ethnic_gp", "health_board")
descriptors_labels = c ("Age", "Number of Risk Groups", "Sex", "Age Groups", "Covid Vaccination Status",
                        "Flu Vaccination Status", "Urban/Rural Classification", "SIMD", "ICU Admittance", "Death", 
                        "Length of Stay", "Number of previous admissions", "Ethnic Group", "Health Board")

var_labels = setNames(as.list(descriptors_labels), descriptors)
descriptors_death = c("ageYear", "n_risk_gps", "Sex", "age_gp", "covid_vs", "flu_vs", "urban_rural_classification", "simd2020_sc_quintile", "num_prev_admission_gp", "ethnic_gp", "health_board")
descriptors_death_labels = c ("Age", "Number of Risk Groups", "Sex", "Age Groups", "Covid Vaccination Status",
                        "Flu Vaccination Status", "Urban/Rural Classification", "SIMD",  
                        "Number of previous admissions", "Ethnic Group", "Health Board")

var_labels_death = setNames(as.list(descriptors_death_labels), descriptors_death)

```

## Number of Hospitalisations - respiratory emergency admission in any location
```{r}
resp_cohort = hosp_cohort %>% filter(acute_resp_admission == 1 | acute_resp_secondary == 1)# %>%
  #arrange(EAVE_LINKNO, ADMISSION_DATE) %>% filter(!duplicated(EAVE_LINKNO))
num_resp_hosp = nrow(resp_cohort)
print(num_resp_hosp)
```

## Number of Hospitalisations - respiratory emergency admission as main cause
```{r}
resp_cohort_strict = hosp_cohort %>% filter(acute_resp_admission == 1) %>%
  arrange(EAVE_LINKNO, ADMISSION_DATE) %>% filter(!duplicated(EAVE_LINKNO))

print(nrow(resp_cohort_strict))
```



## Hospitalisation - broad definition
```{r}
#resp_cohort = hosp_cohort %>% filter(acute_resp_admission == 1 | acute_resp_secondary == 1)

z_df = resp_cohort %>% mutate(value = TRUE) %>%
  mutate(MAIN_CONDITION = factor(MAIN_CONDITION)) %>%
  mutate(simd2020_sc_quintile = factor(simd2020_sc_quintile)) %>%
  mutate(icu_admit = as.factor(icu_admit)) %>%
  mutate(health_board = as.factor(health_board)) %>%
  mutate(ethnic_gp = as.factor(ethnic_gp)) %>%
  mutate(death_28 = as.factor(death_28))
z_df = z_df %>% set_variable_labels(.labels=var_labels, .strict=FALSE)

kable(summary_factorlist_wt(z_df, "value", descriptors))
```

## Overall cohort description - respiratory emergency admission as main cause (strict)
```{r}
z_df = resp_cohort_strict


z_df = z_df %>% mutate(value = TRUE) %>%
  mutate(MAIN_CONDITION = factor(MAIN_CONDITION)) %>%
  mutate(simd2020_sc_quintile = factor(simd2020_sc_quintile)) %>%
  mutate(icu_admit = as.factor(icu_admit))

z_df$death_28 = factor(z_df$death_28)

z_df = z_df %>% set_variable_labels(.labels=var_labels_death, .strict=FALSE)
kable(summary_factorlist_wt(z_df, "value",descriptors))
```

## ARI hosps by time period
```{r}
z_cut_off_date = as.Date("2022-11-15")
z_end = as.Date("2023-01-15")
z_df = resp_cohort %>% mutate(value = if_else(ADMISSION_DATE > z_cut_off_date, 1, 0)) %>%
  filter(ADMISSION_DATE < z_end) %>%
  mutate(MAIN_CONDITION = factor(MAIN_CONDITION)) %>%
  mutate(simd2020_sc_quintile = factor(simd2020_sc_quintile)) %>%
  mutate(icu_admit = as.factor(icu_admit)) %>%
  mutate(health_board = as.factor(health_board)) %>%
  mutate(ethnic_gp = as.factor(ethnic_gp)) %>%
  mutate(death_28 = as.factor(death_28))
z_df = z_df %>% set_variable_labels(.labels=var_labels, .strict=FALSE)

kable(summary_factorlist_wt(z_df, "value", descriptors))
```

## Numbers
```{r}
print(table(z_df$value))
```

## Number of deaths - respiratory emergency admission & strict hospitalisation definition
```{r}
num_resp_deaths = nrow(resp_cohort %>% filter(death_28 == TRUE))
print(num_resp_deaths)
```

## Breakdown of admission diagnoses
```{r}
z_urti_count = resp_cohort %>% filter(urti_admit == 1 | urti_admit_secondary == 1) %>% 
  mutate(urti_admit = if_else(urti_admit == 1 | urti_admit_secondary == 1, 1, 0)) %>%
  count(urti_admit) %>% mutate(urti_admit = n) %>% select(-n) %>% pivot_longer(urti_admit)
z_lrti_count = resp_cohort %>% filter(lrti_admit == 1 | lrti_admit_secondary == 1) %>%
  mutate(lrti_admit = if_else(lrti_admit == 1 | lrti_admit_secondary == 1, 1, 0)) %>%
  count(lrti_admit) %>% mutate(lrti_admit = n) %>% select(-n) %>% pivot_longer(lrti_admit)
z_covid_admit_count = resp_cohort %>% filter(covid_admit == 1 | covid_admit_secondary == 1)  %>%
  mutate(covid_admit = if_else(covid_admit == 1 | covid_admit_secondary == 1, 1, 0)) %>%
  count(covid_admit) %>% mutate(covid_admit = n) %>% select(-n) %>% pivot_longer(covid_admit)
z_flu_admit_count = resp_cohort %>% filter(flu_admit == 1 | flu_admit_secondary == 1) %>%
  mutate(flu_admit = if_else(flu_admit == 1 | flu_admit_secondary == 1, 1, 0)) %>%
  count(flu_admit) %>% mutate(flu_admit = n) %>% select(-n) %>% pivot_longer(flu_admit)
z_rsv_admit_count = resp_cohort %>% filter(rsv_admit == 1 | rsv_admit_secondary == 1) %>%
  mutate(rsv_admit = if_else(rsv_admit == 1 | rsv_admit_secondary == 1, 1, 0)) %>%
  count(rsv_admit) %>% mutate(rsv_admit = n) %>% select(-n) %>% pivot_longer(rsv_admit)
z_bronch_admit_count = resp_cohort %>% filter(bronch_admit == 1 | bronch_admit_secondary == 1) %>%
    mutate(bronch_admit = if_else(bronch_admit == 1 | bronch_admit_secondary == 1, 1, 0)) %>%
  count(bronch_admit) %>% mutate(bronch_admit = n) %>% select(-n) %>% pivot_longer(bronch_admit)

z_counts = z_urti_count %>% rbind(z_lrti_count) %>% rbind(z_covid_admit_count) %>%
  rbind(z_flu_admit_count) %>% rbind(z_rsv_admit_count) %>% rbind(z_bronch_admit_count) %>%
  mutate(s = sum(value)) %>% mutate(perc = value*100/s) %>% select(-s)

# Calculate the overlaps in the admittance reasons
z_overlap = resp_cohort %>% select(urti_admit, urti_admit_secondary, lrti_admit, lrti_admit_secondary, 
                       covid_admit, covid_admit_secondary, flu_admit, flu_admit_secondary,
                       rsv_admit, rsv_admit_secondary, bronch_admit, bronch_admit_secondary) %>%
  mutate(bronch_admit = if_else(bronch_admit == 1 | bronch_admit_secondary == 1, 1, 0)) %>%
  mutate(rsv_admit = if_else(rsv_admit == 1 | rsv_admit_secondary == 1, 1, 0)) %>%
  mutate(flu_admit = if_else(flu_admit == 1 | flu_admit_secondary == 1, 1, 0)) %>%
  mutate(urti_admit = if_else(urti_admit == 1 | urti_admit_secondary == 1, 1, 0)) %>%
  mutate(lrti_admit = if_else(lrti_admit == 1 | lrti_admit_secondary == 1, 1, 0)) %>%
  mutate(num_overlap = rowSums(.)) %>% select(num_overlap) %>% count(num_overlap)

kable(z_counts)
```

### Overlap of diagnoses
```{r}
print(z_overlap)
```

## Distribution of admission diagnoses - respiratory emergency admission
```{r}
z_df = resp_cohort %>% count(MAIN_CONDITION) %>% filter(n > 10) %>% arrange(desc(n))

kable(redact_low_counts(z_df))
```

## Number of extended hospital stays
```{r}
z_df = resp_cohort %>% filter(extended_los == 1)

print(nrow(z_df))
```

## Demographics of extended hospital stays
```{r}
z_df = resp_cohort %>% mutate(value = TRUE) %>%
  mutate(MAIN_CONDITION = factor(MAIN_CONDITION)) %>%
  mutate(simd2020_sc_quintile = factor(simd2020_sc_quintile)) %>%
  mutate(icu_admit = as.factor(icu_admit))

z_df$death_28 = factor(z_df$death_28)
z_df = z_df %>% set_variable_labels(.labels=var_labels, .strict=FALSE)

kable(summary_factorlist_wt(z_df, "extended_los",descriptors))
```

## Number of Hospitalisations - RSV
```{r}
num_rsv_hosp = nrow(resp_cohort %>% filter(rsv_admit == 1))
print(num_rsv_hosp)
```

## Overall cohort description - RSV
```{r}
z_df = resp_cohort %>% filter(rsv_admit == 1) %>%
  mutate(value = TRUE) %>%
  mutate(MAIN_CONDITION = factor(MAIN_CONDITION)) %>%
  mutate(simd2020_sc_quintile = factor(simd2020_sc_quintile)) %>%
  mutate(icu_admit = as.factor(icu_admit))

z_df$death_28 = factor(z_df$death_28)
z_df = z_df %>% set_variable_labels(.labels=var_labels, .strict=FALSE)

kable(summary_factorlist_wt(z_df, "value", descriptors))
```


## Number of deaths - emergency admission - RSV
```{r}
num_rsv_deaths = nrow(resp_cohort %>% filter(rsv_admit == 1) %>% filter(death_28 == TRUE))
print(num_rsv_deaths)
```

## Deaths - RSV
```{r}
z_df = z_df %>% set_variable_labels(.labels=var_labels_death, .strict=FALSE)

kable(summary_factorlist_wt(z_df, "death_28", descriptors_death))
```

## Distribution of number of admissions - RSV
```{r}
z_df = resp_cohort %>% filter(rsv_admit == 1) %>% count(EAVE_LINKNO) %>% rename(num_admissions = n) %>% count(num_admissions)
 
kable(redact_low_counts(z_df))
```

## Number of Hospitalisations - Flu/Pneumonia
```{r}
num_flu_hosp = nrow(resp_cohort %>% filter(flu_admit == 1))
print(num_flu_hosp)
```


## Overall cohort description - Flu/Pneumonia
```{r}
z_df = resp_cohort %>% filter(flu_admit == 1) %>%
  mutate(value = TRUE) %>%
  mutate(MAIN_CONDITION = factor(MAIN_CONDITION)) %>%
  mutate(simd2020_sc_quintile = factor(simd2020_sc_quintile)) %>%
  mutate(icu_admit = as.factor(icu_admit))

z_df$death_28 = factor(z_df$death_28)
z_df = z_df %>% set_variable_labels(.labels=var_labels, .strict=FALSE)

kable(summary_factorlist_wt(z_df, "value", descriptors))
```

## Number of deaths - Flu/Pneumonia
```{r}
num_flu_deaths = nrow(resp_cohort %>% filter(flu_admit == 1) %>% filter(death_28 == TRUE))
print(num_flu_deaths)
```

## Deaths - Flu/Pneumonia
```{r}
z_df = z_df %>% set_variable_labels(.labels=var_labels_death, .strict=FALSE)

kable(summary_factorlist_wt(z_df, "death_28", descriptors_death))
```

## Distribution of number of admissions - Flu/Pneumonia
```{r}
z_df = resp_cohort %>% filter(flu_admit == 1) %>% count(EAVE_LINKNO) %>% rename(num_admissions = n) %>% count(num_admissions)

kable(redact_low_counts(z_df))
```

## Number of Hospitalisations - COVID-19
```{r}
num_covid_hosp = nrow(resp_cohort %>% filter(covid_admit == 1))
print(num_covid_hosp)
```

## Overall cohort description - COVID-19
```{r}
z_df = resp_cohort %>% filter(covid_admit == 1) %>%
  mutate(value = TRUE) %>%
  mutate(MAIN_CONDITION = factor(MAIN_CONDITION)) %>%
  mutate(simd2020_sc_quintile = factor(simd2020_sc_quintile)) %>%
  mutate(icu_admit = as.factor(icu_admit))

z_df$death_28 = factor(z_df$death_28)
z_df = z_df %>% set_variable_labels(.labels=var_labels, .strict=FALSE)

kable(summary_factorlist_wt(z_df, "value", descriptors))
```

## Number of deaths - COVID-19
```{r}
num_covid_deaths = nrow(resp_cohort %>% filter(covid_admit == 1) %>% filter(death_28 == TRUE))
print(num_covid_deaths)
```

## Deaths - COVID-19
```{r}
z_df = z_df %>% set_variable_labels(.labels=var_labels_death, .strict=FALSE)

kable(summary_factorlist_wt(z_df, "death_28", descriptors_death))
```

## Distribution of number of admissions - COVID-19
```{r}
z_df = resp_cohort %>% filter(covid_admit == 1) %>% count(EAVE_LINKNO) %>% rename(num_admissions = n) %>% count(num_admissions)

kable(redact_low_counts(z_df))
```

## Number of hospital admissions per pathogen
```{r}
data = data.frame(name = c("Total Respiratory",
                           "RSV", "Flu/Pneumonia", "COVID-19"),
                  values = c(num_resp_hosp, num_rsv_hosp, num_flu_hosp, num_covid_hosp)
                  )

g = ggplot(data, aes(x=name, y=values)) + geom_bar(stat="identity")

print(g)
```

## Number of deaths per pathogen
```{r}
data = data.frame(name = c("Total Respiratory",
                           "RSV", "Flu/Pneumonia", "COVID-19"),
                  values = c(num_resp_deaths, num_rsv_deaths, num_flu_deaths, num_covid_deaths)
                  )

g = ggplot(data, aes(x=name, y=values)) + geom_bar(stat="identity")

print(g)
```

# Hospital Length of stays
## All respiratory pathogens
```{r}
g = ggplot(resp_cohort, aes(hosp_los_gp)) + geom_bar()

print(g)
```

## Flu/Pneumonia admissions
```{r}
z_cohort = resp_cohort %>% filter(flu_admit == 1)
g = ggplot(z_cohort, aes(hosp_los_gp)) + geom_bar()

print(g)
```

## COVID admissions
```{r}
z_cohort = resp_cohort %>% filter(covid_admit == 1)
g = ggplot(z_cohort, aes(hosp_los_gp)) + geom_bar()

print(g)
```

## RSV admissions
```{r}
z_cohort = resp_cohort %>% filter(rsv_admit == 1)
g = ggplot(z_cohort, aes(hosp_los_gp)) + geom_bar()

print(g)
```
# Weekly counts

## All respiratory admissions
```{r}
z_count = resp_cohort %>% 
  select(EAVE_LINKNO, ADMISSION_DATE) %>% mutate(t = floor(difftime(ADMISSION_DATE, a_begin, units="days"))) %>%
  count(t) %>% mutate(t = t + a_begin) 
g = ggplot() + geom_line(z_count, mapping=aes(x=t, y=n)) + 
  scale_x_date(breaks="1 month", date_labels="%b %Y") + ylab("Number of admissions") +
  xlab("Month and Year") 

print(g)
```

## Flu/Pneumonia admissions
```{r}
z_count = resp_cohort %>% filter(flu_admit == 1) %>% 
  select(EAVE_LINKNO, ADMISSION_DATE) %>% mutate(t = floor(difftime(ADMISSION_DATE, a_begin, units="days"))) %>%
  count(t) %>% mutate(t = t + a_begin)
g = ggplot() + geom_line(z_count, mapping=aes(x=t, y=n)) + 
  scale_x_date(breaks="1 month", date_labels="%b %Y") + ylab("Number of admissions") +
  xlab("Month and Year") 
print(g)
```

## COVID-19 admissions
```{r}
z_count = resp_cohort %>% filter(covid_admit == 1) %>% 
  select(EAVE_LINKNO, ADMISSION_DATE) %>% mutate(t = floor(difftime(ADMISSION_DATE, a_begin, units="days"))) %>%
  count(t) %>% mutate(t = t + a_begin)

g = ggplot() + geom_line(z_count, mapping=aes(x=t, y=n)) + 
  scale_x_date(breaks="1 month", date_labels="%b %Y") + ylab("Number of admissions") +
  xlab("Month and Year") 
print(g)
```


## RSV admissions
```{r}
z_count = resp_cohort %>% filter(rsv_admit == 1) %>% 
  select(EAVE_LINKNO, ADMISSION_DATE) %>% mutate(t = floor(difftime(ADMISSION_DATE, a_begin, units="days"))) %>%
  count(t) %>% mutate(t = t + a_begin)

g = ggplot() + geom_line(z_count, mapping=aes(x=t, y=n)) + 
  scale_x_date(breaks="1 month", date_labels="%b %Y") + ylab("Number of admissions") +
  xlab("Month and Year") 
print(g)
```

# Long Hospital Stays
```{r}
z_df = resp_cohort %>%
    mutate(value = TRUE) %>%
  mutate(MAIN_CONDITION = factor(MAIN_CONDITION)) %>%
  mutate(simd2020_sc_quintile = factor(simd2020_sc_quintile)) %>%
  mutate(icu_admit = as.factor(icu_admit))

z_df$death_28 = factor(z_df$death_28)

kable(summary_factorlist_wt(z_df, "extended_los", descriptors))
```