Calculate Study Weights Based on Risk of Bias

Computes inverse-variance weights adjusted for risk of bias assessments. High risk studies receive reduced weights (default 0), and some concerns receive partial weights (default 0.5). Low risk studies receive full weights.

Usage

calculate_rob_weights(
  meta_result,
  rob_results,
  weight_high = 0,
  weight_concerns = 0.5,
  weight_moderate = 0.75,
  weight_serious = 0.25,
  weight_critical = 0
)

Arguments

meta_result

A MetaResult object from meta_analysis().

rob_results

List of RoB2Result or ROBINSIResult objects.

weight_high

Numeric. Weight multiplier for high-risk studies. Default: 0 (exclude completely).

weight_concerns

Numeric. Weight multiplier for "some concerns" (RoB 2) or "Moderate" (ROBINS-I) studies. Default: 0.5.

weight_moderate

Numeric. Weight multiplier for "Moderate" risk (ROBINS-I). Default: 0.75.

weight_serious

Numeric. Weight multiplier for "Serious" risk (ROBINS-I). Default: 0.25.

weight_critical

Numeric. Weight multiplier for "Critical" risk (ROBINS-I). Default: 0.

Value

A list with components:

weights

Numeric vector of adjusted weights

study_ids

Character vector of study identifiers

rob_judgments

Character vector of RoB judgments

multipliers

Numeric vector of weight multipliers applied

method

Character string describing the weighting method

Examples

if (FALSE) { # \dontrun{
# Create meta-analysis result
meta_res <- meta_analysis(
  yi = log(c(0.75, 0.82, 0.68, 0.91, 0.77)),
  sei = c(0.12, 0.15, 0.18, 0.14, 0.11),
  study_labels = paste("Study", 1:5),
  effect_measure = "hr"
)

# Create RoB 2 assessments
rob_results <- list(
  assess_rob2("Study 1", "Low", "Low", "Low", "Some concerns", "Low"),
  assess_rob2("Study 2", "Low", "Low", "Low", "Low", "Low"),
  assess_rob2("Study 3", "High", "Low", "Low", "Low", "Low"),
  assess_rob2("Study 4", "Low", "Low", "Low", "Low", "Low"),
  assess_rob2("Study 5", "Low", "Low", "Low", "Some concerns", "Low")
)

# Calculate RoB-adjusted weights
rob_weights <- calculate_rob_weights(
  meta_res,
  rob_results,
  weight_high = 0,
  weight_concerns = 0.5
)
rob_weights$weights
rob_weights$rob_judgments
} # }