|
|
 |
| |
| The Medical Safety
Research Center has long been a place of statistical studies
for those actively involved in drug development, in the fields
of biology, pharmacy and medicine and related fields. "Questions
and Answers" sessions are supposed to be a dialog between
the experimental researcher and the statistician, but am I alone
in wondering sometimes whether the two are on the same wavelength? |
| |
| The problems faced
by the questioner were the biological indistinctness of an experimental
hypothesis, multiple uses of official approval for formal significant
differences, and the lack of knowledge/technical capabilities
to apply statistical models to covariance analysis. For the
respondent, the problems arose from lack of recognition of significant
biological differences and drug reaction relationship, an inability
to tackle problems etc. |
| |
| This homepage consists
of data and material distributed at the seminar, which was held
in the style of a private school on the morning of the regular
meeting of the Medical Safety Research Center. Since a large
number of people have made requests for this data, we have posted
it on our web page for the convenience of all concerned. |
| |
| The seminar was centered
around case studies of statistical analysis using modern statistical
models like superior Graphical User Interface (GUI) with JMP,
a statistical software concerned wholly with in-vitro and in-vivo
tests. This seminar also had its practical side, as participants
could personally experience, what is otherwise, mere statistics.
To facilitate learning, we have made the data used at the seminar
available for download.
Please feel free to use it.
|
| |
| The seminar was held
to explain the dosage reaction relation, which was the theme
of the bioassay method. Classics of this field such as, Finney
(1971), Finney (1978), " Probit Analysis ", and 'Statistical
methods in Biological Assay", have already been published,
but at this seminar, we introduced all these masterpieces together.
The statistics software JMP supports "reverse presumption"
which is the main theme of the bioassay method, and has proved
to be an excellent standard statistics software in this field.
Modern experimental data uses the basic data offered by a lot
of people, after omitting the name of the medication and changing
the dosage. |
| |
| Applied cases of various
statistical models have been prepared such as linear regression
model, linear mixed model, nonlinear regression model, response
aspect model, logistic regression model, Weibull regression
model for survival and the D optimum plan for experimental plan
etc. |
| |
| Links from cases and statistical model can be accessed easily. Please feel free to mail new themes to us.
You may omit the name of the medication, but please mention the purpose of the experiment, actual data and results. |
| |
| The end |
| 28th February 2001 |
| Takahashi Yukios |
|
|
|
