dec-manuscript.Rmd

---
output: pdf_document
---

```{r setup, include=FALSE}
# Load used packages
library(package = "tidyverse")

# Configure knitting
knitr::opts_chunk$set(echo = FALSE, message = FALSE, warning = FALSE)
```




```{r get-data}
# Set a snapshot file
vico_snapshot <- "data/vico_snapshot.RData"

# Define used variables
vico_variables <- c(
  # general case data
  "SubjectID",
  "SiteDepartamento", "epiYearAdmision", "SiteType", "SiteName", "HUSarea",
  "salaIngreso",
  "fechaHoraAdmision",
  "sexo", "edadAnios", "fechaDeNacimiento",
  
  # case definition
  "elegibleDiarrea", # compare with actual case definition
  "diarreaComenzoHaceDias", "diarreaMaximoAsientos1Dia",
  "diarreaOtroEpisodioSemanaAnterior",
  "pacienteInscritoViCo", "SASubjectID",
  
  # lab info
  "muestraHecesColecta", "pruebaCultivoHizo", "ETEC", "EPEC",
  
  # DEC virulence
  "sxt1", "stx2", "eaeA", "ST1a", "ST1b", "LT",
  
  # outcomes
  "egresoTipo", "egresoMuerteFecha", "moribundoViCo", "moribundoViCoFecha",
  # TODO missing ICU data
  "cuidadoIntensivoDias",
  "gradoDeshidratacion", "gradoDeshidratacionRegistrado",
  "centroRehidratacionTipo",
  "hxD_ConSangre", "hxD_ConMoco", "hxC_Vomitos",
  
  # other etiology data
  "rotavirus", "RTqPCR_NV1", "RTqPCR_NV2", "SAPOVIRUS", "ASTROVIRUS",
  
  # clinical
  "tomadoSuerosSalesUltimas72hora",
  "medUltimas72HorasAntiB",
  "ingresoPlanBPlanC",
  
  # ses
  "familiaIngresosMensuales", "casaCuantasPersonasViven", "casaCuantosDormitorios",
  "casaMaterialPiso", "casaEnergiaElectrica", "patienteGradoEscolarCompleto",
  "parentescoGradoEscolarCompleto",
  
  # water sources
  "fuentesAguaChorroDentroCasaRedPublica", "fuentesAguaChorroPublico",
  "fuentesAguaChorroPatioCompartidoOtraFuente", "fuentesAguaLavaderosPublicos",
  "fuentesAguaPozoPropio", "fuentesAguaPozoPublico",
  "fuentesAguaCompranAguaEmbotellada", "fuentesAguaDeCamionCisterna",
  "fuentesAguaLluvia", "fuentesAguaRioLago",
  
  # other sanitation
  "TxAgua", "TipoSanitario", "inodoroComparte", "inodoroLimpia", "inodoroLimpia_frec",
  "desechoHeces",
  
  # risks
  "contactoAnimalesCasa"
)

# Check if a snapshot is available
if(!file.exists(vico_snapshot)){
  # if the snapshot is not available, read from server
  # Use the DBI package
  
  #TODO add loading data from server
  # data_base <- DBI::dbConnect()
  
  # Depto Santa Rosa
  # epiYearAdmission %in% c(2008, 2009, 2014, 2015)
} else {
  # if the snapshot is available, load it
  load(file = vico_snapshot)
}
```



```{r select-variables-subsets}
enterics <- Diarr %>%
  as_tibble() %>%
  # only keep relevant subsets
  filter(
    # Screened in Santa Rosa
    SiteDepartamento == "SR",
    # During the period with DEC testing
    epiYearAdmision %in% c(2008, 2009, 2014, 2015)
  ) %>%
  select(one_of(vico_variables)) %>%
  # Classify
  mutate(
    # Define pathotype
    STEC_ONE = ifelse(sxt1 == 1 | stx2 == 1, 1, 0),
    EPEC_ONE = ifelse(eaeA == 1, 1, 0),
    ETEC_ONE = ifelse(ST1a == 1 | ST1b == 1 | LT == 1, 1, 0),
    STEC_Solo = ifelse(STEC_ONE == 1 & !(EPEC_ONE == 1 | ETEC_ONE == 1), 1, 0),
    EPEC_Solo = ifelse(EPEC_ONE == 1 & !(STEC_ONE == 1 | ETEC_ONE == 1), 1, 0),
    ETEC_Solo = ifelse(ETEC_ONE == 1 & !(STEC_ONE == 1 | EPEC_ONE == 1), 1, 0),
    other_dec = (STEC_ONE == 1 | EPEC_ONE == 1 | ETEC_ONE == 1) &
      (STEC_Solo != 1 & ETEC_Solo != 1 & EPEC_Solo != 1),
    # categories
    site_type = ifelse(
      test = SiteType == "H",
      yes = "hospital",
      no = "clinic"
    ),
    period = case_when(
      epiYearAdmision %in% c(2008, 2009) ~ "2008-2009",
      epiYearAdmision %in% c(2014, 2015) ~ "2014-2015"
    ),
    # Cuantify screening
    screened = 1,
    # TODO define eligibility
    eligible = elegibleDiarrea == 1 & !is.na(elegibleDiarrea),
    #   (
    #   (diarreaComenzoHaceDias <= 14 & !is.na(diarreaComenzoHaceDias)) &
    #   (diarreaMaximoAsientos1Dia >= 3 & !is.na(diarreaMaximoAsientos1Dia)) &
    #   (diarreaOtroEpisodioSemanaAnterior != 1 & !is.na(diarreaOtroEpisodioSemanaAnterior)) &
    #     # (salaIngreso != 6) &
    #     (actualAdmitido == 1)
    # ),
    enrolled = eligible & !is.na(SASubjectID),
    sampled = muestraHecesColecta == 1,
    cultured = pruebaCultivoHizo == 1,
    assumed_ecoli = !is.na(ETEC) | !is.na(EPEC),
    dec_tested = !is.na(sxt1) | !is.na(stx2) | !is.na(eaeA) | !is.na(ST1a) |
      !is.na(ST1b) | !is.na(LT),
    # 
    dec_positive = (!is.na(STEC_Solo) & STEC_Solo == 1) |
      (!is.na(EPEC_Solo) & EPEC_Solo == 1) |
      (!is.na(ETEC_Solo) & ETEC_Solo == 1) |
      (!is.na(other_dec) & other_dec == 1)
  )


# sites per period
enterics %>%
  group_by(period, site_type, SiteName) %>%
  summarize() %>%
  summarize(n = n()) %>%
  spread(period, n)


# Enrollment table
enterics %>%
  select(
    site_type, screened, eligible, enrolled, sampled, cultured,
    assumed_ecoli, dec_tested, dec_positive
  ) %>%
  group_by(site_type) %>%
  summarize_at(
    vars(-one_of(c("site_type"))),
    funs(sum(., na.rm = TRUE))
  ) %>%
  gather(
    step, value, -site_type, factor_key = TRUE
  ) %>%
  spread(site_type, value) %>%
  # add percentage
  mutate_at(
    vars(clinic, hospital),
    funs(percent = round(. / lag(.) * 100, 1))
  )


# dec tested by year
enterics %>%
  group_by(epiYearAdmision, site_type) %>%
  summarize(
    dec_tested = sum(dec_tested, na.rm = TRUE)
  ) %>%
  ungroup() %>%
  spread(site_type, dec_tested) %>%
  mutate_at(
    vars(clinic, hospital),
    funs(percent = round(. / sum(.) * 100, 1))
  )

# dec positive by year
enterics %>%
  group_by(site_type, epiYearAdmision) %>%
  summarize(
    n = sum(dec_tested),
    dec_positive = sum(dec_positive, na.rm = TRUE),
    percent = round(dec_positive / n * 100, 1)
  )

```



```{r flow-diagram}
DiagrammeR::DiagrammeR("
  graph LR
    A-->B
    A-->C
    C-->E
    B-->D
    C-->D
    D-->F
    E-->F
")
```



```{r classify-patients}

enterics 


```