*********************************************

From cfrey@PSGHS.EDU Sun Aug 23 17:17:00 1998
Date: Fri, 21 Aug 1998 14:41:24 -0400
From: "Carolin M. Frey"

Subject: Stat ethics wording

For Section II.D add the following element:
Avoid unnecessary use of research subjects by making informed
recommendations for study size. These may be based on prospective power
analysis or the planned precision of the study endpoint with consideration
given to the feasibility of obtaining subjects and the value of the data
elements to be collected.

For Section II.D #4 replace with:
Prior to accepting data for analysis or manuscripts for review, ensure that
appropriate subject approvals were obtained.

[The reason for replacing the existing wording for the item is that
statisticians are rarely interact directly with research subjects. More
typical is that statisticians are asked to work with subject data and
therefore have the opportunity to verify whether these were obtained in an
ethical manner.]

--------------
That aside, all of section II.D could benefit from an overhaul. The
guidelines are vague. What does it mean to "know and adhere to appropriate
guidelines for human subjects protection?" Exactly how does one "protect the
privacy and confidentiality of research subjects?" Item 6 concerning
protection of vulnerable populations is really subsumed under 1. It is
unclear what this section is intended to achieve. For example, should
statisticians be knowledgeable about federal regulations concerning human
and animal research? Or should they be familiar with the underlying
principles (e.g. The Belmont Report as it relates to human subjects).

Without the knowledge base to understand what is being alluded to in Section
II.D, these ethics elements will likely be ignored. What is really being
communicated here is that statisticians should (a strong word!) educate
themselves about the ethical use of research subjects. If so, a couple of
references to key documents might be valuable. The reference(s) might follow
a statement to the effect that "statisticians, in making recommendations for
study design and procedures, should give consideration to how these affect
the autonomy, risk/benefit ratio and justice of prospective human subjects."
Equivalent specificity in wording should be sought for animal studies and I
admit that I am glossing over these, but I think you get the point.

Hope this helps,

Carolin (Lin) Frey
Penn State Geisinger Health System
717-271-5189