Re: how to test inferiority in SPSS

Posted by Rich Ulrich on Dec 31, 2011; 9:52pm
URL: http://spssx-discussion.165.s1.nabble.com/how-to-test-inferiority-in-SPSS-tp5109102p5112677.html

I'm pleased to see that the regulators have reached some
apparent consensus on how to test for equivalence -- The
approaches were fairly ad-hoc, back 10 years ago, when I
tried to figure out what to advise someone in a Usenet sci.stat.*
group.  Today, Wiki separately outlines what regulators want
in Australia, the U.S., and Europe; and the requirements look
pretty similar. 

Wellek's book is the top citation today in Google Scholar - but
using Google Scholar is trickier than it used to be, since Scholar
is no longer a listed choice.  Putting scholar.google.com  in the
address field worked for me.  Searching for < bioequivalence >
returned a different selection of citations, which might also
be useful.  But nothing before about 2003 will give *requirements*
since those were still in flux.

--
Rich Ulrich

> Date: Sat, 31 Dec 2011 08:52:59 -0500

> From: [hidden email]
> Subject: Re: how to test inferiority in SPSS
> To: [hidden email]
>
> SPSS does not contain a module specifically for testing non-inferiority.
> The statistics that you need to carry out these computations by hand can be
> obtained from the descriptive procedure (for interval data) or from
> frequencies (for categorical data). The logic of the procedure requires
> that you establish a range of differences (or ratios) in treatment outcomes
> that is clinically insignificant - that is a clinical issue rather than a
> statistical one. Having established a range of indifference, you will need
> to compute the confidence interval for the treatment effect in your data.
> To support a claim of non-inferiority, the confidence interval must fall
> within the specified range of clinical indifference.
>
> You may find the following book helpful. It provides a comprehensive
> treatment of the subject in understandable terms.
>
> Stefan Wellek (2010). Testing statistical hypotheses of equivalence, 2nd
> Edition. Chapman and Hall.
>
> Best,
>
> Stephen Brand
>
> www.StatisticsDoc.com
>
> [snip, previous]