There is no agreed 'gold standard' to indicate the effectiveness of a prevention program. The highest criteria are set by the Cochrane [1] and Campbell collaborations [2], the Blueprints for Violence Prevention [3] project, the US Government Department of Education's What Works Clearinghouse [4] and the Rand Corporation's Promising Practices [5] network. A combination of an experimental evaluation – or randomised controlled trial – replicated in several locations and with sustained effects beyond the period of the intervention is generally viewed as a necessary precondition for a program to be designated as 'proven'. Evidence from several trials re-analysed as part of a systematic review is a necessary precondition for ratification by the Cochrane and Campbell Collaborations.
Links:
[1] http://www.cochrane.org/index.htm
[2] http://www.campbellcollaboration.org/
[3] http://www.colorado.edu/cspv/blueprints/index.html
[4] http://www.whatworks.ed.gov/
[5] http://www.promisingpractices.net/programlist.asp