trial_participant_flow

source	target	n	stage
Assessed	Randomized	900	Screening
Assessed	Excluded	300	Screening
Excluded	Ineligible	180	Screening
Excluded	Declined	90	Screening
Excluded	Other reasons	30	Screening
Randomized	Intervention	450	Allocation

source

target

stage

Assessed

Randomized

900

Screening

Assessed

Excluded

300

Screening

Excluded

Ineligible

180

Screening

Excluded

Declined

Screening

Excluded

Other reasons

Screening

Randomized

Intervention

450

Allocation

#	变量	类型	缺失	统计
1	sourceTrial-state node from which participants move.	字符型	—	不同值 9Excluded · Treated · Received control · Assessed · Randomized · Intervention · Control · Intervention follow-up
2	targetTrial-state node receiving the participants.	字符型	—	不同值 21Randomized · Excluded · Ineligible · Declined · Other reasons · Intervention · Control · Treated
3	nkeyNumber of participants in the flow.	整数	—	min 3q1 —中位 180.0均值 233.3q3 —max 900.0
4	stageScreening, Allocation, Follow-up or Analysis.	字符型	—	不同值 4Allocation · Follow-up · Screening · Analysis

变量

类型

缺失

统计

sourceTrial-state node from which participants move.

字符型

—

不同值 9Excluded · Treated · Received control · Assessed · Randomized · Intervention · Control · Intervention follow-up

targetTrial-state node receiving the participants.

字符型

—

不同值 21Randomized · Excluded · Ineligible · Declined · Other reasons · Intervention · Control · Treated

nkeyNumber of participants in the flow.

整数

—

min 3q1 —中位 180.0均值 233.3q3 —max 900.0

stageScreening, Allocation, Follow-up or Analysis.

字符型

—

不同值 4Allocation · Follow-up · Screening · Analysis

library(readr) trial_participant_flow <- read_csv( "https://assets.evanzhou.org/tessera/csv/trial_participant_flow.csv", col_types = cols( source = col_character(), target = col_character(), n = col_integer(), stage = col_character() ) )

Source

A deterministic synthetic cohort created for the Tessera Sankey recipe. Of 1,200 people assessed for eligibility, 900 are randomized equally to intervention and control; subsequent branches represent non-receipt, loss to follow-up, discontinuation and exclusion from complete-case analysis.

The values do not describe a real trial. Every internal node is balanced exactly, and intervention and control remain separate through the analysis stage.

Use cases

Participant-flow Sankey diagrams
Demonstrating attrition at several trial stages
Checking conservation at internal nodes
Contrasting Sankey quantity flow with a formal CONSORT diagram

# Generate the synthetic edge table used by the trial-participant Sankey recipe. script_dir <- local({ arg <- grep("^--file=", commandArgs(trailingOnly = FALSE), value = TRUE) path <- if (length(arg)) sub("^--file=", "", arg[[1L]]) else sys.frame(1)$ofile dirname(normalizePath(path, mustWork = TRUE)) }) out_csv <- file.path(script_dir, "..", "csv", "trial_participant_flow.csv") dir.create(dirname(out_csv), recursive = TRUE, showWarnings = FALSE) trial_participant_flow <- data.frame( source = c( "Assessed", "Assessed", "Excluded", "Excluded", "Excluded", "Randomized", "Randomized", "Intervention", "Intervention", "Control", "Control", "Treated", "Treated", "Treated", "Received control", "Received control", "Received control", "Intervention follow-up", "Intervention follow-up", "Control follow-up", "Control follow-up" ), target = c( "Randomized", "Excluded", "Ineligible", "Declined", "Other reasons", "Intervention", "Control", "Treated", "Not treated", "Received control", "No control", "Intervention follow-up", "Intervention lost", "Intervention stopped", "Control follow-up", "Control lost", "Control stopped", "Intervention analyzed", "Intervention excluded", "Control analyzed", "Control excluded" ), n = c( 900, 300, 180, 90, 30, 450, 450, 430, 20, 425, 25, 375, 35, 20, 370, 40, 15, 372, 3, 366, 4 ), stage = c( rep("Screening", 5), rep("Allocation", 6), rep("Follow-up", 6), rep("Analysis", 4) ), stringsAsFactors = FALSE ) internal_nodes <- intersect(trial_participant_flow$source, trial_participant_flow$target) for (node in internal_nodes) { stopifnot( sum(trial_participant_flow$n[trial_participant_flow$target == node]) == sum(trial_participant_flow$n[trial_participant_flow$source == node]) ) } write.csv(trial_participant_flow, out_csv, row.names = FALSE)