-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathInstructions.html
63 lines (60 loc) · 3.41 KB
/
Instructions.html
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
<div id="using-this-software" class="section level3">
<h3>Using This Software:</h3>
<ol style="list-style-type: decimal">
<li><strong>Choose an outcome type and the smallest meaningful
effect:</strong> For a continous outcome, analyses are based on a
difference in means. For a binary outcome, analyses are based on a risk
difference (i.e. difference in proportions). For an ordinal outcome,
analyses are based on the <a href="https://covariateadjustment.github.io/estimands.html#mann-whitney-m-w-estimand">Mann-Whitney
estimand</a> or the distribution of outcomes under treatment and
control.</li>
<li><strong>Specify outcome parameters, separated by commas:</strong>
For continuous outcomes, enter the standard deviation in each treatment
arm: these values should all be positive. For binary outcomes, enter the
event probability under the treatment arm. If using the distribution
function for an ordinal outcome, enter the probability of each outcome
category in each treatment arm.</li>
<li><strong>Choose the study design:</strong> This includes the number
of interim analyses, their timing, and the decision rules applied at
each analysis. For more information about the designs implemented, see
the documentation on the <a href="https://cran.r-project.org/web/packages/rpact/index.html"><code>rpact</code>
package</a> <span class="citation">(Wassmer and Pahlke 2023)</span>. See
the <code>Background</code> tab for more information and references on
group sequential designs.</li>
</ol>
<p>Once these parameters are specified, plots will show the sample sizes
at which interim analyses might occur under different assumptions about
the outcome distribution. <strong>Note:</strong> these sample sizes are
approximate values based on asymptotic approximations. Actual levels of
information vary: simulation can be used to provide interval estimates
for sample size requirements. These simulations can be provided by a
statistical collaborator.</p>
</div>
<div id="monitoring-information-based-trials" class="section level3">
<h3>Monitoring Information-Based Trials</h3>
<p>The <a href="https://github.com/kelvlanc/GSDCovAdj">GSDCovAdj</a>
package can be used to monitor information levels, conduct analyses,
perform testing and inference <span class="citation">(Van Lancker, Betz,
and Rosenblum 2022)</span>.</p>
</div>
<div id="testing-and-validation" class="section level3">
<h3>Testing and Validation</h3>
<p>Software testing and documentation can be found in <a href="https://github.com/jbetz-jhu/PICARD">Github</a>. Report any bugs
by opening a <a href="https://github.com/jbetz-jhu/PICARD/issues">Github
Issue</a>.</p>
</div>
<div id="references" class="section level3 unnumbered">
<h3 class="unnumbered">References</h3>
<div id="refs" class="references csl-bib-body hanging-indent">
<div id="ref-VanLancker2022" class="csl-entry">
Van Lancker, Kelly, Joshua Betz, and Michael Rosenblum. 2022.
<span>“Combining Covariate Adjustment with Group Sequential, Information
Adaptive Designs to Improve Randomized Trial Efficiency.”</span>
<em>arXiv Preprint arXiv:1409.0473</em>. <a href="https://doi.org/10.48550/ARXIV.2201.12921">https://doi.org/10.48550/ARXIV.2201.12921</a>.
</div>
<div id="ref-rpact_package" class="csl-entry">
Wassmer, Gernot, and Friedrich Pahlke. 2023. <em>Rpact: Confirmatory
Adaptive Clinical Trial Design and Analysis</em>. <a href="https://CRAN.R-project.org/web/packages/rpact">https://CRAN.R-project.org/web/packages/rpact</a>.
</div>
</div>
</div>