Kaplan Meier Survival Curves without an addin?

  • Thread starter Thread starter TC2004
  • Start date Start date
T

TC2004

Is it possible to draw Kaplan Meier survival curves in Excel without using an
add-in?

The main difference between these and a normal line plot are that the line
should fall straight down toward the x-axis after each data point rather than
connecting one data point with the next on a diagonal.

Thanks.
 
How about using negative error bars with value the same as the y-value of
each point?
best wishes & Happy New Year
 
Hi,

1. Make a line chart
2. Double-click the line and on the Patterns tab set line to None and on the
Options tab check Drop lines
 
No! The Kaplan-Meier Survival Analysis is more than a step curve. While the
result looks like a step chart, the calculation of the values requires a
specific survival analysis technique and handling of censored data.

See this link for the basics:
http://www.cra.org/Activities/craw/dmp/awards/2003/Mower/KaplanMeier.html#Basic
idea

Note that the author explains that a program (language not specified) was
used to calculate the values which were plotted in Excel.
The Kaplan -Meier curve for censored data requires a marker for sections
with censored data. This is not typical of a step chart.

Here's another example of a Kalplan-Meier curve:
http://www.bmj.com/content/vol317/issue7172/images/large/survprob.f1.jpeg
Again, while it looks like as step chart, it also has the censored data
markers.

Here's a link to a full Kaplan-Meier analysis including standard error
calculations and mean survival times with 95% confidence intervals.
http://www.childrens-mercy.org/stats/model/survival.asp

As Tamino http://tamino.wordpress.com/2008/12/18/lovehate-relationship/ at
Open Mind recently discussed the use of Excel and R and his concern that
" Any idiot can load data into Excel and fit ..." . Care in developing and
applying techniques is required.

Kelly O'Day
http://processtrends.com
 
The OP said "draw curves", so I incorrectly assumed that the problem was
with turning known values into the stepped appearance of the chart. I did
not mean to trivialize the foregoing analysis.

Re your cite from Tamino:
"Any idiot can load data into Excel and fit ..."
This is very true (and inflammatory), but as is stated a few lines later:
"I don’t know if Excel, or R, or something else was used to fit the trend
line in question, and it doesn’t really matter."

When a statement like the first is made, the sense of the second statement
is missed, and the blog entry and its comments becomes polarized; one is
side bashing Excel and Windows and PCs and Corporate IT departments
(predominantly based on anecdotes), and the other much less vocal side is
saying that anybody can misuse analytical tools, but with foresight and
common sense you can make any platform work.

I'm glad you've taken the latter POV: "Care in developing and applying
techniques is required."

- Jon
-------
Jon Peltier, Microsoft Excel MVP
Peltier Technical Services, Inc.
http://PeltierTech.com/WordPress/
_______
 
Back
Top