I want to find source code for computing a "permuted-block
randomization"?
If that's not possible, a well-defined description of how to compute
one would be ok. Not a proof, just a well-described implementation.

I have found exhaustive proofs of this subject, but I just want a
simple description. I don't want to
know why it works, just how to do it.

This is not an exam or homework question! It may possibly sound like
one, but I'm
administering an NIH clinical trial.

---

Request for Question Clarification by efn-ga on 07 Dec 2002 15:01 PST

Does it matter what programming language the source code is written in?

---

Clarification of Question by linuxgeek-ga on 07 Dec 2002 16:12 PST

Preferably Java. C or C++ would be good.
If this can't be found I'll take any language as it will still be
valuable. I'll tip extra if it's either java, C, or C++.

Thanks,
Oscar

---

Request for Question Clarification by answerguru-ga on 07 Dec 2002 20:27 PST

Hi there,

After doing some preliminary research, I realized that there are
several variations on this concept, many of which appear to be
proprietary information held by the health industry. Do you have a
specific algorithm that you can provide or reference? If this is the
case, I could easily code it for you in Java, C, or C++.

answerguru-ga

---

Clarification of Question by linuxgeek-ga on 08 Dec 2002 16:23 PST

OK. I obviously need to do some more research myself. I should be able
to get a more specific algorithm for you tomorrow, Monday.

Thanks,
Oscar

---

Clarification of Question by linuxgeek-ga on 08 Dec 2002 17:26 PST

This is all the information I have on this right now. This is from the
protocol.

The randomization will be performed using a permuted block
randomization design.  Subjects will be stratified jointly on
"DISEASE" severity (mild, moderate, and severe), gender, and
race/ethnicity (non-Hispanic Whites and minorities).  As stated
previously, this randomization approach will accomplish the following:
1) tend to balance quantities of subjects between the Active and
Sub-therapeutic Groups throughout the entire study, 2) reduce bias due
to differences in subject characteristics.

I will hopefully get an exact algorithm tomorrow, and in any case I'll
post again here tomorrow.

Thanks,
Oscar

---

Clarification of Question by linuxgeek-ga on 09 Dec 2002 22:56 PST

I still don't have that much information. But it looks like I might
not need this question answered.

I'm not sure if I should go ahead and close it. I need to read the
google answer FAQ. If I close it, I may open it up again later. In any
case, thanks for your inquiry. This is the first time I've used google
answer, and I definitely will try it again.

Thanks,
Oscar