TY - JOUR

T1 - Response to the letters of Dr Amos and Dr Preti and colleagues

AU - van der Gaag, Mark

AU - Smit, Filip

AU - French, Paul

AU - Yung, Alison R.

AU - McGorry, Patrick

AU - Cuijpers, Pim

PY - 2014/3

Y1 - 2014/3

N2 - Reply by the current author to the comments made by Andrew James Amos et al. (see record 2014-09375-015) and Antonio Preti et al. (see record 2014-09375-016) on the original article (see record 2013-26281-001). The letters of Amos et al. and Preti et al. provide with the opportunity to rebut some of their criticisms and to further clarify the design and outcome of meta-analysis of randomized trials in the prevention of first psychosis. In his criticism of meta-analysis, Amos relies on the arguments of the pre meta-analysis era: this can be characterized as "vote counting". Vote counting concludes that if most individual studies do not find statistically significant results, then there is no effect. studies. The Button paper referred to by Amos states that publication bias and low quality of the primary studies may inflate true effect-sizes in meta-analysis. This is correct and therefore the authors evaluated and reported the impact of publication bias and study quality on their meta-analytic outcomes. If Dr Amos can agree with the scientific community that p-values of individual trials are less important in demonstrating risk reduction in rare disorders and that risk reduction is the key focus in preventing rare diseases, then he should also agree that meta-analysis is an appropriate method to pool the available evidence.

AB - Reply by the current author to the comments made by Andrew James Amos et al. (see record 2014-09375-015) and Antonio Preti et al. (see record 2014-09375-016) on the original article (see record 2013-26281-001). The letters of Amos et al. and Preti et al. provide with the opportunity to rebut some of their criticisms and to further clarify the design and outcome of meta-analysis of randomized trials in the prevention of first psychosis. In his criticism of meta-analysis, Amos relies on the arguments of the pre meta-analysis era: this can be characterized as "vote counting". Vote counting concludes that if most individual studies do not find statistically significant results, then there is no effect. studies. The Button paper referred to by Amos states that publication bias and low quality of the primary studies may inflate true effect-sizes in meta-analysis. This is correct and therefore the authors evaluated and reported the impact of publication bias and study quality on their meta-analytic outcomes. If Dr Amos can agree with the scientific community that p-values of individual trials are less important in demonstrating risk reduction in rare disorders and that risk reduction is the key focus in preventing rare diseases, then he should also agree that meta-analysis is an appropriate method to pool the available evidence.

KW - Health sciences

KW - Psychology

UR - http://www.scopus.com/inward/record.url?scp=84895821403&partnerID=8YFLogxK

U2 - 10.1016/j.schres.2014.01.004

DO - 10.1016/j.schres.2014.01.004

M3 - Journal articles

C2 - 24440496

VL - 153

SP - 237

EP - 239

JO - Schizophrenia Research

JF - Schizophrenia Research

SN - 0920-9964

IS - 1-3

ER -