We are currently working on the forum. For the short-term, all forum content will be in read-only format. We apologize for the interruption and look forward to collaborating with you shortly. All the best in your research!
How to design response labels for different response types
Comments
-- If you have a chance to use Openclinica, please post your opinions on the
forum.
-- There is no argument about values that standards bring.
-- There seems to have a difference in opinions with regards to the
following two approaches:
a. CDISC from the ground up: GUI implements CDISC to exporting data
in CDISC
b. Excel to allow for gathering data from researchers --> export data
in CDISC. My guess is that with this approach validate of data will be
different than if a software starts with CDISC from the ground up. (Note
that this is a lot of risk for a software company -- the software company
invests completely on CDISC).
I don't' believe OC facilitates importing data in CDISC format, at
this moment.
-- With regards to Validation -- I think there is something here that I
don't fully understand. I do know that OC does validate its excel sheet.
But with regards to the issue that you were mentioned about 2 different
users of CRF (the author researcher who created the CRF, and another
researcher who use the same crf), I have not envisioned it would be a
problem with OC.
Scenario 1: both researchers using ONE central database -- OC will check all
its Excel (for general syntax as I mentioned in couple e-mail before); when
OC loads the excel, it definitely check the central db. There are many
sub-cases that ran through my head, but I don't see a problem yet.
Scenario 2: both researchers use separate databases. They both can change
the CRF to fit with THEIR needs. Upload and use both CRFs (might or might
not be same any more) in their own databases. A problem will arise when
both decides to upload their data (and crf) into same db. (I believe this
is my original counter-argument with Michael Sweeney with regards to 'there
is no system wide' response label, values, types). At the same time, I
don't see how a software that adopted CDISC from the ground up, would escape
this problem without a proper design. (this is a classic problem of
distributed environment; it is orthogonal to whether one accepts standards
or not.)
As always, it is my humble opinions and always want to learn more from
experts.
My C.
-----Original Message-----
From: dr. Jozef Aerts [mailto:[email protected]]
Sent: Saturday, December 02, 2006 3:29 PM
To: [email protected]; 'Sidhu, Bobby [CORP]'
Cc: [email protected]
Subject: Re: [Users] How to design response labels for different response
types
Thanks,
It's getting to become an interesting discussion indeed !
I am located in Europe, that's probably why you haven't seen me at
conferences in the US yet. I try to catch 1-2 conferences in the US each
year, but that is not always possible. Most of my customers are EDC vendors
(both in the US and Europe) that need help with CDISC implementation.
One of my recent customers specializes on rescue studies, and without CDISC
standards this would be very difficult or nearly impossible.
I think OpenClinica is a great development, is the first open source
clinical system after Phosco that may become successful (others failed).
I agree that if researchers feel more comfortable with Excel for CRFs, than
they can use it,but the problem of portability between EDC system remains.
OpenClinica does already have export in CDISC format which I think is very
(but also the easiest to implement), but there is still a way to go (Lab
data for example, what format does OC use for that ?)
I think researchers indeed shouldn't have themselves ask what kind of widget
should be used (i.e. the visual stuff) but that is also not what I meant. A
study design tool (at least if based on CDISC standards), only requires that
the user defines the question, the datatype (integer, float, date ...),
whether there is a list of possible answers (codelist) and whether an answer
to the question is mandatory or not. The EDC system should then
automatically create the eCRFs and the visual layout (look-and-feel) based
on some predefined rules. So this should not have to cost a single dollar
and the researcher can concentrate on the science.
I do not know what Excel costs in the US nowadays, but it is not for free,
so also there there is a cost, although it may be hidden. For an open source
system I was a bit surprised that it encourages to use Excel, and not for
example OpenOffice, which is open source and freely available.
And I do agree, some study design tools are overpriced !
About the validation, this might be less an issue when the designer of the
CRF is at the same time the user of the EDC system. But what in the case the
designer of the CRF is at another location, has another function, and thus
never sees the CRFs after they have been ported to the EDC system. What
guarantee is there that the CRF is the same in the EDC system than in the
worksheet ? I know that study design tools need to be validated, and the
forms that come out of it, so I presume that also the Excel systems need to
be validated. I know from practice however that usually this isn't done, as
it is difficult to do...
I think also the FDA issue is important. Sponsors need to submit the study
data AND metadata to the FDA. The FDA is encouraging sponsors to do this in
CDISC format. They are more and more also asking for the annotated eCRFs
(i.e. unfilled, but annotated with submission - SDTM - information, i.e.
which clinical domain and which submission SDTM variable), and the ideal
format for this is CDISC ODM. I do not mean a visual representation of the
eCRF, but the questions, codelists used, data types, application logic
etc.., also the science and logic behind the CRF.
To finalize, it is not an utopia that an EDC system is being set up fully
automatically from a study desing in CDISC ODM format. I know of at least 5
EDC systems that have this feature. The study design is either made using a
study design tool or an XML editor (which can be cheaper than Excel). The
whole DB and EDC system is then set up in usually less than a minute. It
might be less an issue in academic research, but doing so, the cost of EDC
design is decreased enormously (sponsors sometimes pay pretty large amounts
of money for that).
I downloaded OC a few weeks ago, but haven't found the time yet to start
working with it. I hope to have so during the next week. I will also try to
set up CRFs using the worksheet mechanism (although I will try to use
OpenOffice). So I can see how that works. I am especially interested how
question dependency logic (e.g. automatically skip the question about
pregnancy when the investigator enters that the subject is male) and range
checks can be and are implemented.
I am very positive about OC so far, but I find their statement "OpenClinica
uses an extensive data model based on the CDISC standard for clinical
research" pretty exagerated. But maybe I should first have a look at the
source code to find out what the data model exactly is.
Have a very nice weekend,
Jozef Aerts
XML4Pharma
P.S. Just a remark about the naming "OC". In large pharma and at CROs, it is
used for "Oracle Clinical" so indeed confusing isn't it ?
===========================================================================