app.R

library(shiny)
library(shinyWidgets)
library(shinydashboard)
library(shinydashboardPlus)
library(leaflet)
library(tidyverse)
library(bsicons)
library(bslib)
 # add the icons then can publish

# remotes::install_github("deepanshu88/summaryBox")



load("d_zipcodes.rda")

theme <-bs_theme(
    version = 5,
    bg = "#ffd9b3",
    fg = "#000000",
    primary = "#000099",
    secondary = "#FF374B",
    base_font = "Arial"
)


# 

ui <- shiny::fluidPage(
  titlePanel("sdiWATCH-DM Score"),
  theme = theme,

fluidRow(
  # first column
  column(
    width = 4,
    sliderInput(
      inputId = "options_age", label = "Age(years)",
      min = 30, max = 90, step = 1, value = 50
    ),
    # bmi
    sliderInput(
      inputId = "options_bmi", label = "Body mass index (kg/m2)",
      min = 20, max = 55, value = 30, step = 1
    ),
    # hba1c
    sliderInput("options_hba1c", "HbA1c(%)", min = 5, max = 15, step = 0.1, value = 7),
    # zip code map
    value_box(
      title = "zip code Map",
      value =  leafletOutput("zmap")
    )
  ),
  # second column


  column(
    width = 4,

    # creatinine
    sliderInput("options_creat", "Creatinine(mg/dl)", min = 0.5, max = 5, step = 0.1, value = 1),

    # sbp
    sliderInput("options_sbp", "Systolic blood pressure(mmHg)", min = 70, max = 200, step = 1, value = 120),

    # dbp
    sliderInput("options_dbp", "Diastolic blood pressure(mmHg)", min = 50, max = 120, step = 1, value = 80),

    # zip code
    pickerInput(
      inputId = "options_zip",
      label = "5-digit Zip code",
      choices = d$ZIP,
      selected = d$ZIP[1],
      multiple = F,
      options = list(
        `live-search` = TRUE
      )
    ),
    
    br(),
    
    value_box(title = "Total WATCH-DM Points",value = textOutput("sum"), showcase = bs_icon("bar-chart")),
    value_box(title = "5-year HFH risk: Original WATCH-DM Score", value = textOutput("orig"), showcase = bs_icon("graph-up")),
    
    box(width = 12,
        h6("The WATCH-DM Score is a model to predict the 5-year heart failure hospitalization risk (HFH)
                   in patients with type 2 diabetes mellitus (T2DM). The model contains the following variables; age, body mass index, 
                   blood pressure (systolic & diastolic), 
                   HbA1c level, HDL-C level, serum creatinine, H/O myocardial infarction, and H/O coronary artery bypass grafting. Details regarding the original 
                   WATCH-DM score can be found at: https://pubmed.ncbi.nlm.nih.gov/35656988/."))
    )
  ,
  # third column
  column(
    width = 4,
    # hdlc
    sliderInput("options_hdlc", "HDL-C(mg/dl)", min = 20, max = 80, step = 1, value = 40),
    # mi
    prettyRadioButtons("options_mi", "Prior MI", choices = c("No", "Yes")),

    # cabg
    prettyRadioButtons("options_cabg", "Prior CABG", choices = c("No", "Yes")),
    
    # get risk
    actionButton("score", "Get Risk",icon = icon("calculator")),
    
    br(),
    br(),
    br(),
    
    value_box(title = "The SDI Quintile according to the zip code",value = textOutput("q2"), showcase = bs_icon("pie-chart")),
    value_box(title = "5-year recalibrated HFH risk: sdiWATCH-DM Score", value = textOutput("recab2"),
              showcase = bs_icon("lightning-fill")),
    
    box(width = 12,
        h6("We have recalibrated the score
                   using data from more than 1,000,000 US Veterans receiving outpatient care in the VA healthcare system. We have used their residential zip codes 
                   to determine their social deprivation index (SDI) and present the recalibrated 5-year risk for HFH. The score should not be interpreted as 
                   medical advice. It is to be used by healthcare professionals as an aid for shared clinical decision-making.")
        
    )
  ),
  column(width = 6,
         h6("Developed by Salil Deo (svd14@case.edu), Ayush Patel and Sadeer Al-Kindi")
         )
  
)  
)







server <- function(input, output, session) {
  observeEvent(input$score, {
    # age

    age <-
      ifelse(input$options_age < 50, 0,
        ifelse(
          input$options_age >= 50 & input$options_age < 55, 1,
          ifelse(
            input$options_age >= 55 & input$options_age < 60, 2,
            ifelse(
              input$options_age >= 60 & input$options_age < 65, 3,
              ifelse(
                input$options_age >= 65 & input$options_age < 70, 4,
                ifelse(
                  input$options_age >= 70 & input$options_age < 75, 5, 6
                )
              )
            )
          )
        )
      )

    # bmi

    bmi <- ifelse(input$options_bmi < 30, 0,
      ifelse(
        input$options_bmi >= 30 & input$options_bmi < 35, 1,
        ifelse(
          input$options_bmi >= 35 & input$options_bmi < 40, 3, 4
        )
      )
    )


    # hba1c

    hba1c <- ifelse(input$options_hba1c < 7, 0,
      ifelse(
        input$options_hba1c >= 7 & input$options_hba1c < 9, 1,
        ifelse(
          input$options_hba1c >= 9 & input$options_hba1c < 9.9, 4,
          ifelse(
            input$options_hba1c >= 10 & input$options_hba1c < 11.9, 5, 6
          )
        )
      )
    )


    # hdlc

    hdlc <- ifelse(input$options_hdlc < 30, 5,
      ifelse(
        input$options_hdlc >= 30 & input$options_hdlc < 35, 1,
        ifelse(
          input$options_hdlc >= 35 & input$options_hdlc < 40, 3, 4
        )
      )
    )


    # sbp
    #

    sbp <- ifelse(input$options_sbp < 100, 0,
      ifelse(
        input$options_sbp >= 100 & input$options_sbp < 140, 2,
        ifelse(
          input$options_sbp >= 140 & input$options_sbp < 160, 4, 5
        )
      )
    )

    # dbp
    #

    dbp <- ifelse(input$options_dbp < 60, 4,
      ifelse(
        input$options_dbp >= 60 & input$options_sbp < 79, 2, 0
      )
    )


    # prior mi
    #

    prior_mi <- ifelse(
      input$options_mi == "No", 0, 3
    )

    prior_cabg <- ifelse(
      input$options_cabg == "No", 0, 3
    )

    n <- age + bmi + hdlc + hba1c + sbp + dbp + prior_mi + prior_cabg

    # output$sum <- renderText(paste("The WATCH-DM Score is", n))
    output$sum <- renderText(n)

    # get the risk groups
    #

    g <- ifelse(
      n < 11, "Very Low",
      ifelse(
        n %in% c(12, 13), "Low",
        ifelse(
          n %in% c(14, 15), "Moderate",
          ifelse(
            n %in% c(16, 17, 18), "High", "Very High"
          )
        )
      )
    )

    # output$risk <- renderText(paste("The WATCH-DM Score category is", g))
    output$risk <- renderText(g)

    # baseline survival
    #

    S0t <- 0.9934

    beta <- 0.118668

    orig <- 1 - S0t**exp(n * beta)

    orig2 <- orig * 100

    # output$orig <- renderText(paste("The 5-year WATCH-DM heart failure hospitalization predicted risk is", paste0(round(orig2,2),"%",sep =" ")))
    output$orig <- renderText(paste0(round(orig2, 2), "%", sep = " "))



    # get the quintile for the risk using the zip code data

    #        d <- read_csv("map_data.csv")

    t <- reactive({
      d %>% filter(ZIP %in% input$options_zip)
    })

    quintile <- t()$quintile

    # output$q2 <- renderText(paste("Your SDI is in the quintile",t()$quintile))
    output$q2 <- renderText(t()$quintile)


    log_eo <- ifelse(
      quintile == 1, 0.3160,
      ifelse(
        quintile == 2, 0.1508,
        ifelse(
          quintile == 3, 0.0268,
          ifelse(
            quintile == 4, -0.0545, -0.2260
          )
        )
      )
    )

    recab <- (1 - S0t**exp(n * beta - log_eo)) * 100

    # output$recab2 <-
    #    renderText(paste("The sdiWATCH-DM recalibrated heart failure hospitalization risk is",
    #                     paste0(round(recab,2),"%", sep = " ")))


    output$recab2 <-
      renderText(paste0(round(recab, 2), "%", sep = " "))


    # map

    output$zmap <- renderLeaflet({
      leaflet() %>%
        addTiles() %>%
        addMarkers(lng = t()$lng, lat = t()$lat, label = "Mapped Zip code")
    })
  })
}



shiny::shinyApp(ui, server)