Correlate food expenditure with disease costs — ics_correlate_food

Given health expenditure data by state (or district), compute correlations between every food group/item expenditure share and a specified disease category's hospitalization cost.

Usage

ics_correlate_food_disease(
  con,
  health_df,
  disease = "mean_exp",
  round = "HCES-2022-23",
  level = c("both", "group", "item"),
  min_states = 10L
)

Arguments

con: A DBI connection from ics_connect().
health_df: A data frame with columns state_name and at least one numeric disease expenditure column. Typically from indiahealthsurvey::ihs_state_summary().
disease: Column name in health_df containing the disease cost to correlate against. Default "mean_exp".
round: Consumption survey round. Default "HCES-2022-23".
level: One of "group" (13 food groups), "item" (individual items), or "both".
min_states: Minimum states with data for a correlation to be computed. Default 10.

Value

A tibble with columns: food, type ("group" or "item"), correlation, n_states, sorted by absolute correlation.

Details

This answers: "Which foods are most associated with spending on cardiovascular disease / diabetes / cancer at the state level?"

Examples

if (FALSE) { # \dontrun{
library(indiahealthsurvey)
hcon <- ihs_connect("~/data/nss_health")
cvd <- ihs_state_summary(hcon, "hospitalization", by_disease = TRUE) |>
  dplyr::filter(disease_category == "Cardiovascular")
ihs_disconnect(hcon)

con <- ics_connect("~/data/consumption_parquet")
ics_correlate_food_disease(con, cvd)
} # }