[GET-dev] GET-Evidence: adding "age of onset" to clinical importance

[Madeleine Ball]

At this point I'm not sure who to include in an email regarding potential
GET-Evidence improvements. I'm emailing this to Joe & Tom who have been
core people in designing the system (medical & software respectively). I'm
also cc'ing Wendy, who made the suggestion prompting this email, and I'm
sending this to the get-editors and get-dev mailing lists.

In her comments on the draft of a paper we're hoping to submit soon, Wendy
noted "I would add age of onset to the clinical importance" on a
GET-Evidence diagram. I've been thinking about adding this too, hearing it
come from another source convinces me we should think about doing it now
(rather than after the paper's published!).

So... it requires some sort of definition. I'm going to throw one out
there, anyone interested is invited to respond if they have some comments &
ideas.

Age of onset: score according to the median age of onset for the disease as
it manifests for individuals affected by this variant.
-- 0 stars: > 70 years of age
-- 1 star: 50-70 years of age
-- 2 stars: 30-50 years of age
-- 3 stars: 15-30 years of age
-- 4 stars: 5-15 years of age
-- 5 stars: Congenital or < 5 years of age

This is just a suggested set of ranges, comments are welcome. I have no
idea what the distributions there are for age of onsets, or if median
(rather than average) is a readily available number. (I just like medians
better.) Also, I guess the fact it scores higher for less age could be
confusing, but I think it's consistent with how other categories score more
highly when "worse/more severe/strong effect".

Someday, in the mythic future, we would like to split of the annotation
efforts into disease level pages (a variant implicated in the disease would
default to the values of the disease page for the clinical importance
section, but could be modified if that variant had a notably different
ranking in some category e.g. unusually early onset, more resistant to
standard treatment, etc). But for now, I think having this scoring on the
variant page is better than nothing?

Thanks!

Madeleine
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.arvados.org/pipermail/arvados/attachments/20111208/5d7367b8/attachment-0001.html>