Limitations

Identified Studies

1 High risk of potential bias due to lack of rigorous design based on Downs and Black quality rating checklist. Tul et al. (2011), Bosch et al. (2009), Badsha et al. (2009), Fishman et al. (2009), Yaģli & Űlger (2012)
3 Lack of standardized measurement Yaģli & Űlger (2012), Űlger & Yaģli (2010), Bukowski et al. (2006-2007), Babsha et al. (2009), Dash & Telles (2001)
4 Failure to complete a power analysis to determine appropriate sample size/small sample size Park et al. (2011), Park & McCaffrey (2012), Űlger & Yaģli (2010), Bukowski et al. (2006-2007), Groessl et al. (2008), Galantino et al. (2004), Bosch et al. (2009), Badsha et al. (2009), Dash & Telles (2001), Curtis et al. (2011)
5 Lack blinding (subject and main outcomes of the intervention.) Twenty five of the studies reviewed had no blinding to subjects or interven­tion. Only six met this criterion: Tilbrook et al. (2011), Williams et al. (2005), Williams et al. (2009), Saper et al. (2009), Tekur et al. (2008), Garfinkel et al. (1998)
8 Lack of sufficient follow-up to determine whether the effect was sustained (dose effect) All of the studies demonstrated a lack of followup EXCEPT Garfinkel et al. (1998), Tilbrook et al. (2011) ,Williams et al. (2005) (2009), Tellus (2009), Saper et al. (2009), Tekur et al. (2008), Sherman et al. (2005), Carson et al. (2010), da Silva et al. (2007)
9 Potential bias of administration of treatment and assessment of outcome Twenty-five of the studies had potential for bias due to lack of radomization. Only six met this criterion: Tilbrook et al. (2011), Williams et al. (2005), Williams et al. (2009), Saper et al. (2009), Tekur et al. (2008), Garfinkel et al. (1998)
Table 3: Summary of Main Methodological Limitations of the Reviewed Studies.