Run a Batch Association Analysis

This function automates running a series of association models, iterating through a list of predictors and outcomes for a given dataset. It can run both cross-sectional (lm) and longitudinal (lmer) models.

run_association_analysis(
  data,
  predictors,
  outcomes,
  covariates,
  dataset_name,
  analysis_type = "Cross-Sectional",
  use_delta_outcome = TRUE,
  random_effects = "(1 | PTID)",
  interaction_variable = NULL
)

Arguments

data

The data frame containing all necessary variables.

predictors

A character vector of predictor variable names.

outcomes

A character vector of outcome variable names.

covariates

A character string of covariates to include in all models (e.g., "Age + Sex").

dataset_name

A character string to label the results (e.g., "ADNI").

analysis_type

A character string, either "Cross-Sectional" (default) or "Longitudinal".

use_delta_outcome

A logical. If TRUE and analysis_type is "Longitudinal", it will look for outcome variables with a _delta suffix.

random_effects

An optional character string for the lmer random effects (e.g., "(1 | PTID)"). Only used when analysis_type is "Longitudinal".

interaction_variable

An optional character string specifying the variable to interact with the predictor in longitudinal models (e.g., "yearsbl"). This is required for longitudinal analyses.

Value

A single tibble (data frame) containing the summarized results from all successful model fits.

Examples

if (FALSE) { # \dontrun{
# Example for a cross-sectional analysis
cs_results <- run_association_analysis(
  data = my_cross_sectional_data,
  predictors = c("PredictorA", "PredictorB"),
  outcomes = c("MMSE", "ADAS13"),
  covariates = "Age + Sex + Education",
  dataset_name = "MyStudy_CS"
)

# Example for a longitudinal analysis with an interaction
long_results <- run_association_analysis(
  data = my_longitudinal_data,
  predictors = c("Biomarker1", "Biomarker2"),
  outcomes = c("CognitionScore"),
  covariates = "Age_at_baseline + Sex + Education + yearsbl",
  dataset_name = "MyStudy_Long",
  analysis_type = "Longitudinal",
  use_delta_outcome = FALSE, # Use the raw score over time
  random_effects = "(1 | SubjectID)",
  interaction_variable = "yearsbl" # Specify the interaction term
)
} # }