Definition of "Clinical Study"

[gauger]

Dear group,

I was just reading a hand-out about "clinical studies", where a the
notion "clinical study" was defined as: ".. a prospective comparison
of a (or several) intervention(s) against a control in humans".

My question is: do clinical studies necessarily involve control
groups? If so, why do we have the notion "randomised clinical trial"?

Further question: Are there any differences between "trial" and
"study".

Thanks in advance.

Uli

[Adrian Sayers]

You need to separate out the control and randomisation.

Control groups allows you to compare your intervention to. however
control does not nessecarily imply inert control, many controls are
active.

Randomisation is there to stop confounding, or minimise it best it can.
randomisation is supposed to ensure equal distribution of confounder
in both groups therefore all results should be unbiased estimates of
the intervention, or so the books say!

You still get many trials that arent randomised, and some with out
controls either.

bw
Adrian

2008/10/13 gauger <ulrich...@gmail.com>:

[Bland, M.]

Surely case studies of an individual patient, or case series, are
clinical studies too. This person has not thought this through.

Martin

--
***************************************************
J. Martin Bland
Prof. of Health Statistics
Dept. of Health Sciences
Seebohm Rowntree Building Area 2
University of York
Heslington
York YO10 5DD

Email: mb...@york.ac.uk
Phone: 01904 321334 Fax: 01904 321382
Web site: http://martinbland.co.uk/
***************************************************

[John Whittington]

At 05:12 13/10/2008 -0700, gauger wrote:
>I was just reading a hand-out about "clinical studies", where a the
>notion "clinical study" was defined as: ".. a prospective comparison
>of a (or several) intervention(s) against a control in humans".
>
>My question is: do clinical studies necessarily involve control
>groups? If so, why do we have the notion "randomised clinical trial"?
>
>Further question: Are there any differences between "trial" and
>"study".

These are obviously only semantic questions.

I think the only safe assumption about the words 'trial' and 'study' are
that they are, at least in the real
world, interchangeable. Investigations in animals are usually called
'studies', but rarely 'trials'. In terms of investigations in humans, it's
probably true that 'study' is more often used when subjects are healthy
volunteers, and 'trial' when they are patients with therapeutic needs, but
that is far from universal.

In terms of regulation and legislation, the phrase 'Clinical Trial' is
generally used for any investigation in humans, at least in Europe.

It is tempting to say that a 'clinical trial' (as opposed to a 'survey' or
observational study) requires that the investigation involves an
intervention (e.g. administration of a treatment), but regulations and
legislation talk in terms of 'non-interventional clinical trials', thereby
killing that semantic idea! However, I think that it is probably true to
say that most people think in terms of a clinical trial involving an
intervention.

A 'clinical trial' does not necessarily include a control group, even when
it involves an intervention. Merely observing the effect of an
intervention (i.e. an 'uncontrolled trial') constitutes a clinical trial,
even if not a very good one. Even when there is a control group,
allocation to treatments/interventions is not necessarily at random
(although it generally ought to be!). The term 'Randomised Clinical Trial'
is therefore very much needed - since a clinical trial can also be
uncontrolled or controlled but not randomised.

Kind Regards,

John

----------------------------------------------------------------
Dr John Whittington, Voice: +44 (0) 1296 730225
Mediscience Services Fax: +44 (0) 1296 738893
Twyford Manor, Twyford, E-mail: Joh...@mediscience.co.uk
Buckingham MK18 4EL, UK
----------------------------------------------------------------

[Bland, M.]

John writes "I think the only safe assumption about the words 'trial'

and 'study' are that they are, at least in the real world,

interchangeable." I disagree and I think a lot of epidemiologists would
too. How about "case-control studies", "cohort studies"? Trials are one
kind of study.

Martin

--

[John Whittington]

At 14:48 13/10/2008 +0100, Bland, M. wrote:
>John writes "I think the only safe assumption about the words 'trial' and
>'study' are that they are, at least in the real world,
>interchangeable." I disagree and I think a lot of epidemiologists would
>too. How about "case-control studies", "cohort studies"? Trials are one
>kind of study.

That makes complete sense, and is obviously in line with the everyday use
of the word 'trial' - - to 'try' something (i.e. have an
intervention'). If I were able to force the situation, I would certainly
like to see the distinction that Martin B is suggesting - i.e.
trial=intervention and study=non-intervention (just observation).

However, I fear that I still have to stick to my statement that 'the only
safe assumption', in the real world, is to assume that the words may be
used interchangeably. If one Googles "case control trials" (with the
quotes), one gets over 6,000 hits - many of which obviously relate to
published papers written by epidemiologists.

Furthermore, as I said before, the issue is confused by regulations and
legislation. For example, Directive 2001/20/EC of the European Parliament
(Generally known as "The Clinical Trials Directive", and now enacted, as
required, in UK legislation) defines a "non-interventional trial" in a
manner that encompasses case-control studies and cohort studies.

Having said all that, sort-of in my defence, I have to admit that when I
wrote my words, I was actually thinking 'the other way around' .... in
other words, it seems to be increasingly common for what you and I would
happily accept as a 'clinical trial' to be referred to as a 'clinical study'.

[Abhaya Indrayan]

Yes, calling all human studies as trials is taking it too far. The way I understand, all trials are studies anyway. Intervention studies are essentially experiments and can be conducted in laboratory, clinic or field. The last two are generally termed trials (clinical or field trials). Observational studies (no human intervention) can be retrospective (generally, case-control), prospective or cross-sectional. Intervention as well as observational studies of the type just cited are analytical form of studies (primary objective is to study antecedent-outcome type of relationship). The other form is descriptive, which is generally understood as survey (no prior assumption of antecedent or outcome in this case). Descriptive studies are different from cross-sectional analytical studies where antecedents and outcomes are studied together. These are generally accepted terms but exceptions can always be cited.

 

~Abhaya Indrayan

[gauger]

Thanks to all of you!

U.