malaria_do_analysis: Code for calculating Malaria disease cases attributable to extreme rainfall and extreme temperature Run Full Malaria-Climate Analysis Pipeline

Description

The Malaria_do_analysis() function executes the complete workflow for analyzing the association between malaria cases and climate variables. It integrates health, climate, and spatial data; fits spatio-temporal models using INLA; and generates a suite of diagnostic and inferential outputs, including plots and attributable risk estimates.

Usage

malaria_do_analysis(
  health_data_path,
  climate_data_path,
  map_path,
  region_col,
  district_col,
  date_col = NULL,
  year_col,
  month_col,
  case_col,
  tot_pop_col,
  tmin_col,
  tmean_col,
  tmax_col,
  rainfall_col,
  r_humidity_col,
  runoff_col,
  geometry_col,
  spi_col = NULL,
  ndvi_col = NULL,
  max_lag = 2,
  nk = 2,
  basis_matrices_choices,
  inla_param,
  param_term,
  level,
  param_threshold = 1,
  filter_year = NULL,
  family = "nbinomial",
  group_by_year = FALSE,
  cumulative = FALSE,
  config = FALSE,
  save_csv = FALSE,
  save_model = FALSE,
  save_fig = FALSE,
  output_dir = NULL
)

Value

A named list containing:

inla_result - Fitted INLA model object and summaries.
plot_malaria, plot_tmax, plot_rainfall - Exploratory time-series plots.
reff_plot_monthly - Monthly random effects plot.
reff_plot_yearly - Yearly spatial random effects plot.
contour_plot - Exposure-response contour plot.
rr_map_plot - Spatial relative risk map.
rr_plot, rr_df - Relative risk plot and associated data.
attr_frac_num - Attributable risk summary table.
plot_AR_num, plot_AR_frac, plot_AR_per_100k - Plots of attributable number, fraction, and rate.

Arguments

health_data_path: Character. Path to the processed health data file.
climate_data_path: Character. Path to the processed climate data file.
map_path: Character. Path to the spatial data file (e.g., shapefile).
region_col: Character. Column name for the region variable.
district_col: Character. Column name for the district variable.
date_col: Character (optional). Column name for the date variable. Defaults to NULL.
year_col: Character. Column name for the year variable.
month_col: Character. Column name for the month variable.
case_col: Character. Column name for malaria case counts.
tot_pop_col: Character. Column name for total population.
tmin_col: Character. Column name for minimum temperature.
tmean_col: Character. Column name for mean temperature.
tmax_col: Character. Column name for maximum temperature.
rainfall_col: Character. Column name for cumulative monthly rainfall.
r_humidity_col: Character. Column name for relative humidity.
runoff_col: Character. Column name for monthly runoff data.
geometry_col: Character. Column name of the geometry column in the shapefile (usually "geometry").
spi_col: Character (optional). Column name for the Standardized Precipitation Index (SPI). Defaults to NULL.
ndvi_col: Character (optional). Column name for the Normalized Difference Vegetation Index (NDVI). Defaults to NULL.
max_lag: Numeric. Maximum temporal lag to include in the distributed lag model (e.g., 2-4). Defaults to 4.
nk: Numeric. Number of internal knots for the natural spline of each predictor, controlling its flexibility: nk = 0 produces a linear effect with one basis column, nk = 1 generates a simple spline with two columns, nk = 2 yields a more flexible spline with three columns, and higher values of nk further increase flexibility but may also raise collinearity among spline terms. Defaults to 2.
basis_matrices_choices: Character vector. Specifies which climate variables to include in the basis matrix (e.g., c("tmax", "rainfall", "r_humidity")).
inla_param: Character vector. Specifies exposure variables included in the INLA model (e.g., c("tmin", "rainfall", "r_humidity")).
param_term: Character or vector. Exposure variable(s) of primary interest for relative risk and attribution (e.g., "tmax", "rainfall").
level: Character. Spatial disaggregation level; must be one of "country", "region", or "district".
param_threshold: Numeric. Threshold above which exposure is considered "attributable." Defaults to 1.
filter_year: Integer or vector (optional). Year(s) to filter the data by. Defaults to NULL.
family: Character. Probability distribution for the outcome variable. Options include "poisson" (default) and "nbinomial" for a negative binomial model.
group_by_year: Logical. Whether to group attributable metrics by year. Defaults to FALSE.
cumulative: Boolean. If TRUE, plot and save cumulative risk of all year for the specific exposure at region and district level. Defaults to FALSE.
config: Logical. Whether to enable additional INLA model configurations. Defaults to TRUE.
save_csv: Logical. If TRUE, saves intermediate datasets to CSV. Defaults to TRUE.
save_model: Logical. If TRUE, saves fitted INLA model results. Defaults to TRUE.
save_fig: Logical. If TRUE, saves generated plots. Defaults to TRUE.
output_dir: Character. Directory where output files (plots, datasets, maps) are saved. Defaults to NULL.