Figures and data in Region of Attainable Redaction, an extension of Ellipse of Insignificance analysis for gauging impacts of data redaction in dichotomous outcome trials

Version of Record

Accepted for publication after peer review and revision.

Download
Cite
Share
CommentOpen annotations (there are currently 0 annotations on this page).

Version of Record published: February 16, 2024 (This version)
Accepted Manuscript published: January 29, 2024 (Go to version)
Accepted: January 23, 2024
Received: October 4, 2023
Preprint posted: September 25, 2023 (Go to version)

Figures
Tables
Additional files

2 figures, 2 tables and 2 additional files

Figures

Figure 1

Download asset Open asset

ROAR analysis implementation example.

(a) A simulated example of the Region of Attainable Redaction (ROAR) for $a = 70$ , $b = 30$ , $c = 50$ , $d = 50$ ( $R R_{E} > 1$ ) with all points bounded by the shaded region depict a degree of redaction which would not lead to the null being rejected. (b) Relevant vectors for ascertaining possible redaction thresholds in $R R_{E} > 1$ case. (c) ROAR analysis of the similar data but with ( $R R_{E} < 1$ ) ( $a = 50$ , $b = 50$ , $c = 70$ , $d = 30$ ). Note that $R R_{E} > 1$ case is a transform of $R R_{E} < 1$ situation. (d) Relevant vectors for ascertaining possible redaction thresholds in case $R R_{e} < 1$ .

Figure 2

Download asset Open asset

ROAR example for large data-set.

(a) A simulated example of the Region of Attainable Redaction (ROAR) for a meta-analysis of $N = 39, 197$ as described in the text. (b) Relevant vectors for ascertaining possible redaction thresholds.

Tables

Table 1

Reported groups and related variables.

Redaction for $R R_{e} > 1$
	Endpoint positive	Endpoint negative
Experimental group	$a$	$b + x$
Control group	$c + y$	$d$
Redaction for $R R_{e} < 1$
	Endpoint positive	Endpoint negative
Experimental group	$a + x$	$b$
Control group	$c$	$d + y$

Table 2

ROAR-derived metrics for published data (see ‘Results’ for details).

Statistics for simulated example case
ROAR statistic ( $α = 0.05$ )	Calculated ROAR value (simulated $R R_{E} > 1$ data)	Calculated ROAR value (simulated $R R_{E} < 1$ data)
Total subjects reported $N$	200	200
Relative risk (95% CI) $R R_{E}$	1.40 (1.11–1.77)	0.60 (0.46–0.76)
FOCK point	$(x_{e}, y_{e}) = (6.89, 4.79)$	$(x_{e}, y_{e}) = (4.79, 6.89)$
$r_{m i n}$	12 subjects	12 subjects
Experimental redaction tolerance $ρ_{E}$	9.51% ( $⌈ x_{c} ⌉$ = 10 subjects)	14.32% ( $⌈ x_{c} ⌉$ = 14 subjects)
Control redaction tolerance $ρ_{C}$	14.32% ( $⌈ y_{c} ⌉$ = 16 subjects)	9.51% ( $⌈ y_{c} ⌉$ = 10 subjects)
Total redaction tolerance $ρ_{A}$	5.66% (12 subjects)	5.66% (12 subjects)
Statistics for large meta-analysis
ROAR statistic ( $α = 0.05$ )	Calculated ROAR value
Total subjects reported $N$	39,197
Relative risk (95% CI)	0.85 (0.74–0.97)
FOCK point	$(x_{e}, y_{e}) = (14.17, 0.32)$
$r_{m i n}$	14 subjects
Experimental redaction tolerance $ρ_{E}$	0.07% (13 subjects)
Control redaction tolerance $ρ_{C}$	3.21% (649 subjects)
Total redaction tolerance $ρ_{A}$	0.04% (14 subjects)

FOCK, Fewest Observations/Censored Knowledge; ROAR, Region of Attainable Redaction.

Additional files

MDAR checklist: https://cdn.elifesciences.org/articles/93050/elife-93050-mdarchecklist1-v2.docx
Download elife-93050-mdarchecklist1-v2.docx
Supplementary file 1 Coefficients for $g (x, y)$ (Table S1) and related identity $h (x, y)$ (Table S2).: https://cdn.elifesciences.org/articles/93050/elife-93050-supp1-v2.pdf
Download elife-93050-supp1-v2.pdf

Download links

A two-part list of links to download the article, or parts of the article, in various formats.

Downloads (link to download the article as PDF)

Article PDF

Open citations (links to open the citations from this article in various online reference manager services)

Mendeley

Cite this article (links to download the citations from this article in formats compatible with various reference manager tools)

David Robert Grimes

(2024)

Region of Attainable Redaction, an extension of Ellipse of Insignificance analysis for gauging impacts of data redaction in dichotomous outcome trials

eLife 13:e93050.

https://doi.org/10.7554/eLife.93050

Share this article

Cite this article

ROAR analysis implementation example.

ROAR example for large data-set.

Reported groups and related variables.

ROAR-derived metrics for published data (see ‘Results’ for details).

MDAR checklist

Supplementary file 1

Downloads (link to download the article as PDF)

Open citations (links to open the citations from this article in various online reference manager services)

Cite this article (links to download the citations from this article in formats compatible with various reference manager tools)