BosenbergSED < Main < Vanderbilt Biostatistics Wiki

You are here: Vanderbilt Biostatistics Wiki>Main Web>StatCollabCourse>BosenbergSED (01 Oct 2012, MattShotwell) (raw view)EditAttach

---+ Bosenberg's (1995) Paper and SED Data

---++ Notes
   *Bosenberg, A. T. (1995) "Skin-epidural distance in children." Anaesthesia. 50: 895-897.*
   * SED Data (digitized from paper figure): [[http://biostat.mc.vanderbilt.edu/wiki/pub/Main/ReproducibleResearchTutorial/Bosenberg1995.csv][Bosenberg1995.csv]]
   * Scanned image from the Bosenberg (1995) paper: [[%ATTACHURLPATH%/Bosenberg1995.png][Bosenberg1995.png]]
   * Digitized SED figure from the Bosenberg paper (Fig. 2): [[%ATTACHURL%/Bosenberg1995.svg][Bosenberg1995.svg]]
   * Fitted linear regression line: =SED = 3.93 + 0.80 * WT=, where =SED= is in mm units, and =WT= is in kg units (R<sup>2</sup> = 0.74).
   * From the paper:
<blockquote>
<p><em>Statistical methods</em>: Regression analysis using the least squares procedure was used to determine the association between the skin-epidural distance and weight. For the difference between the two equations, i.e. the regression equation and the hypothesised 1 mm = 1 kg body weight, a Z-test was performed.</p>
<p>In children over 10 years (n = 19) there was also poor correlation. The data from these patients were not used for further analysis.</p>
<p>Plotting the best fit regression line and its 95% confidence limits shows that the y = x line, or skin-epidural distance (mm) is equal to weight (kg) falls within these limits for weight <40 kg (or 10 years of age) in this series (Fig. 2). Therefore the difference between any two points on these two lines is not statistically significant.</p>
<p>In addition a Z-test performed on any two points obtained from the two equations for a specific weight, i.e. skin-epidural distance (mm) = 0.80 weight (kg) 3.93, and skin-epidural distance (mm) = weight (kg) was found to be not statistically significant.</p>
*...*
<p>In summary, we have demonstrated that 1 mm/kg is a useful guideline for judging the skin-epidural distance in children between 6 months and 10 years of age.</p>
</blockquote>

---++ Discussion Points
   * Was the confidence band really a prediction band?
   * Was it simultaneous or pointwise?
   * Was Bosenberg's reasoning sound?
   * How can we verify Bosenberg's conclusions?

---++ Assignment
Modify the Bosenberg Sweave file from the previous assignment as follows, and regenerate the PDF report. Ensure that the document remains reproducible. 
   1 Add Bosenberg's simplified prediction rule (i.e., 1 mm/kg) to the scatterplot.
   1 Modify the figure caption to reflect the addition above.
   1 Use a statistical technique to assess whether Bosenberg's simplified rule was reasonable.
   1 Add a short discussion of your findings.

---++ Solution
Here is one solution to the assignment above: [[%ATTACHURL%/Bosenberg.Rnw][Bosenberg.Rnw]] [[%ATTACHURL%/Bosenberg.pdf][Bosenberg.pdf]]

Topic attachments
I	Attachment	Action	Size	Date	Who	Comment
Rnw	Bosenberg.Rnw	manage	4.2 K	01 Oct 2012 - 08:51	MattShotwell	Solution to the Bosenberg assignment
pdf	Bosenberg.pdf	manage	89.3 K	01 Oct 2012 - 08:52	MattShotwell	Solution to the Bosenberg assignment
png	Bosenberg1995.png	manage	38.6 K	17 Sep 2012 - 07:56	MattShotwell	Scanned image from the Bosenberg (1995) paper.
svg	Bosenberg1995.svg	manage	185.3 K	16 Sep 2012 - 10:02	MattShotwell	Digitized SED data from Bosenberg (1995) paper

Topic revision: r4 - 01 Oct 2012, MattShotwell

Main

Department Home Page

Biostatistics Graduate Program

Vanderbilt University Medical Center

Biostatistics Webs
- Archive
- Main
- Sandbox
- System

Copyright © 2013-2022 by the contributing authors. All material on this collaboration platform is the property of the contributing authors.
Ideas, requests, problems regarding Vanderbilt Biostatistics Wiki? Send feedback