Home / Topics / Participants / Pool of Subjects / Inclusion Criteria / Moderators as Predictors of Outcome

Moderators as Predictors of Outcome

Posted on July 22, 2009

Joseph R. Calabrese (bio) passes along advice from a mentor when he says, "Study what you see."


When studying treatment in refractory cohorts of patients, basically, what we're doing is we're taking moderators of illness, features of patients' history that predict non-response, and we're focusing on them.

So I believe the key thing to do, then, is to develop homogeneous cohorts of patients who present with these variants of illness and then to study them. This is very difficult to do because that's like taking the most difficult subgroup of patients and just studying them.

But this, I think, is the next step for our field: Teasing out moderators of illness and then designing studies that specifically target the treatment of these subgroups of patients.

So for me, I took a homogeneous cohort of patients with rapid cycling and randomized them to two different treatments; this is essentially a moderator trial. The problem, though, is that it took me ten years to do that.

So these are, I think, very difficult trials to do, and in the future, we need to have big teams of people dedicated to the conduct of these trials that specifically focus on cohorts of patients the pharmaceutical industry has avoided in the past.

I have always thought that the focus on moderators in illness, and in this instance, bipolar disorder, to be intuitive. So if you take a cohort of patients and you look at what makes them fail treatment, that's a moderator if it's something that they presented with. So long time ago, it became obvious to me that we should study homogeneous cohorts of patients who don't do well and take them and randomize to two different treatments. And that is what now is called a focus on moderators as predictors of outcome.

One of my mentors once told me, "Study what you see." And by that, he meant if you see a lot of something, then do a study on that something. And that's basically all I did. And I think to a great extent these moderator and mediator analyses, focus on these things.

But the challenge is developing a homogeneous cohort of patients that present with these things. This is where the networks, the Bipolar Trials Network, I think, can really make a contribution. Get a big bunch of people together, work on the same topic, and enroll enough patients to make a difference.

Viewing Preferences




Excerpted from interview with researcher at the 2009 Career Development Institute for Bipolar Disorder in Pittsburgh, PA.

Additional Reference:
Nierenberg, A. A., Sylvia, L. G., Leon, A., Reilly-Harrington, N., Ketter, T. A., Calabrese, J. R., Thase, M., Bowden, C. L., Friedman, E. S., Ostacher, M. J., Iosifescu, D., Novak, L., & The LiTMUS Study Group. (2009). Lithium treatment - moderate dose use study (LiTMUS): Rationale and design. Clinical Trials, 6(6), 637-48.


Please note that the feedback is viewed only by 4researchers staff and is not intended for communication with individual contributors.


Use the form below to submit feedback about this article. If you would like a response, please be sure to include your e-mail address.

More About "Inclusion Criteria"


Benefiting Real-World Patients


Design Studies that Industry Won't Do

Show All...


More From Joseph R. Calabrese (bio)


A Well-Rounded Commitment


Conflict of Interest

Show All...