Create a Reliable Change Index (RCI) object — CreateRCIObject • SciDataReportR

Learn regression-based Reliable Change Index (RCI) models relative to a user-defined reference visit and calculate projected RCI values.

Usage

CreateRCIObject(
  data,
  variables,
  DataFormat = c("wide", "long"),
  id_var,
  Method = "regression",
  BaselineSpecifier = NULL,
  FollowupSpecifier = NULL,
  SpecifierPosition = c("suffix", "prefix"),
  VisitColumn = NULL,
  VisitOrder = NULL,
  BaselineVisit = NULL,
  Confidence = 0.95,
  Relabel = TRUE,
  Data = lifecycle::deprecated(),
  Variables = lifecycle::deprecated(),
  ID = lifecycle::deprecated()
)

Arguments

data

A data frame.

variables

Character vector of canonical variable names.

DataFormat

Either "wide" or "long".

id_var

ID column.

Method

Currently only "regression" is supported.

BaselineSpecifier

Baseline visit identifier for wide data.

FollowupSpecifier

Follow-up visit identifier for wide data.

SpecifierPosition

Either "suffix" or "prefix".

VisitColumn

Visit column for long data.

VisitOrder

Optional ordering of visits.

BaselineVisit

Reference visit used for RCI calculations.

Confidence

Confidence interval threshold.

Relabel

Logical; use variable labels when available.

Interpretation guide

egression-based RCI values are interpreted similarly to z-scores.

RCI cutoff	Approximate confidence interval
+/-0.50	~38%
+/-1.00	~68%
+/-1.645	~90%
+/-1.96	~95%
+/-2.58	~99%

Traditional Jacobson-Truax RCI thresholds typically use +/-1.96, corresponding to approximately 95% confidence.

Data

Deprecated (since 19.15.0). Use data instead.

Variables

Deprecated (since 19.15.0). Use variables instead.

ID

Deprecated (since 19.15.0). Use id_var instead.

Value

A SciDataReportR_RCI object.

Details

Supports both wide and long longitudinal data structures.

Long format is recommended for datasets with more than two visits.

Examples

set.seed(20260803)
rci_data <- data.frame(
  id = rep(1:30, each = 2),
  visit = rep(c("Baseline", "Followup"), 30),
  Score = round(rnorm(60, mean = 50, sd = 10), 1)
)

# Use a +/-1 cutoff here so all three change classifications are visible.
# The default Confidence = 0.95 retains the conventional +/-1.96 cutoff.
rci <- CreateRCIObject(
  data = rci_data,
  variables = "Score",
  DataFormat = "long",
  id_var = "id",
  VisitColumn = "visit",
  BaselineVisit = "Baseline",
  Confidence = 0.68
)

# Individual trajectories across visits
rci$Plots$Spaghetti$Score


# Participant-level reliable-change values
rci$Plots$Waterfall$Score


# Baseline values relative to reliable change
rci$Plots$Quadrant$Score