Skip to contents

Create Binary Environmental Suitability Indicator from Case Data

Source: R/get_cases_binary.R
get_cases_binary.Rd

This function creates a binary indicator of environmental suitability for cholera transmission based on case count data. It uses sophisticated temporal logic to identify not only outbreak periods but also the environmental conditions leading up to outbreaks, which may be more predictive of future transmission risk.

Usage

get_cases_binary(data, cutoff = 1)

Arguments

data

A data frame containing cholera surveillance data with columns:

  • iso_code: ISO country code

  • cases: Weekly cholera case counts

  • date_start: Start date of each week (for ordering)

cutoff

Numeric threshold for case counts. Weeks with cases >= cutoff are considered outbreak periods. Default is 1 (any cases).

Value

The input data frame with an additional cases_binary column containing the binary environmental suitability indicator (0 = unsuitable, 1 = suitable).

Details

The environmental suitability logic works as follows:

  1. Primary identification: Weeks with cases >= cutoff are marked as suitable (1)

  2. Lead-up period: The week immediately before any outbreak week is marked as suitable

  3. Extended lead-up: If two consecutive weeks have cases >= cutoff, then 2-3 weeks before are also marked as suitable

This approach recognizes that environmental conditions favorable for cholera transmission typically develop in the weeks preceding observable outbreaks. By marking these lead-up periods as "suitable," the model can learn to predict environmental suitability before cases are reported, which is more useful for early warning systems.

Missing Data Handling: Weeks with missing case data (NA) are preserved as NA in the binary indicator and are never retroactively marked as suitable during the lead-up period logic. This prevents false environmental suitability signals during periods with poor surveillance coverage.

The function processes each country separately to ensure that temporal relationships are maintained within each location's outbreak history.

Note

The temporal logic requires data to be ordered by date within each country. The function will sort the data appropriately, but for best performance, input data should already be ordered by iso_code and date_start.

See also

compile_suitability_data for the main data processing pipeline where this function is used.

Examples

if (FALSE) { # \dontrun{
# Load cholera case data
data <- read.csv("cholera_weekly_data.csv")

# Create binary suitability indicator with default cutoff
data_with_binary <- get_cases_binary(data)

# Use higher cutoff for more conservative outbreak definition
data_with_binary <- get_cases_binary(data, cutoff = 10)
} # }