Wrapper function to calculate MCID using anchor-based, distribution-based, or both approaches.
Usage
calculate_mcid(
data,
score_var,
anchor_var = NULL,
baseline_var = NULL,
method = c("both", "anchor", "distribution"),
reliability = NULL,
anchor_minimal = "Minimally Improved",
conf_level = 0.95
)Arguments
- data
Data frame with PRO data
- score_var
Character. Name of the score/change variable
- anchor_var
Character. Name of anchor variable (required for anchor method)
- baseline_var
Character. Name of baseline score variable (for distribution method)
- method
Character. "anchor", "distribution", or "both" (default: "both")
- reliability
Numeric. Test-retest reliability for SEM calculation
- anchor_minimal
Character vector. Anchor values for minimal improvement
- conf_level
Numeric. Confidence level (default: 0.95)
Value
List containing MCID results. Structure depends on method:
If method = "anchor": anchor-based results with mcid, ci, n, method
If method = "distribution": distribution-based results (half_sd, one_sem, etc.)
If method = "both": both anchor and distribution results
Examples
if (FALSE) { # \dontrun{
# Anchor-based MCID
anchor_result <- calculate_mcid(
data = pro_data,
score_var = "AVAL",
anchor_var = "PGIC",
responder_value = "Much improved",
method = "anchor"
)
# Distribution-based MCID
dist_result <- calculate_mcid(
data = pro_data,
score_var = "AVAL",
method = "distribution",
reliability = 0.85
)
} # }