Table 2.
Methodological quality ratings for included studies (n=39)
| Author and year | Selection bias | Design | Confounders | Blinding | Data collection methods |
Withdrawals and dropouts |
EPHPP global rating |
Total score |
| RCTs (n=7) | Of 18 | |||||||
| Blank and Smithkline30 2002 | 1 | 1 | 3 | 3 | 1 | 3 | 3 | 12 |
| Boden-Albala et al,34 2015 | 2 | 1 | 1 | 2 | 1 | 1 | 1 | 8 |
| Dracup et al,15 2009 | 1 | 1 | 1 | 1 | 3 | 1 | 2 | 8 |
| Luepker et al,31 2000 | 2 | 1 | 1 | 1 | 1 | 3 | 2 | 9 |
| Meischke et al,32 1997 | 3 | 1 | 3 | 1 | 3 | NA | 3 | 11/15 |
| Mooney et al,10 2014 | 2 | 1 | 1 | 1 | 1 | 1 | 1 | 7 |
| Müller-Nordhorn et al,33 2009 | 1 | 1 | 1 | 1 | 3 | 1 | 2 | 8 |
| Controlled before and after studies (n=3) | Of 12 | |||||||
| Morgenstern et al,35 2002 | 1 | 2 | 3 | 2 | 1 | NA | NA | 6 |
| Rowley et al,36 1982 | 3 | 2 | NA | 2 | 3 | NA | NA | 10 |
| Xin-gang et al,37 2013 | 2 | 2 | 3 | 2 | 3 | NA | NA | 9 |
| Uncontrolled before and after studies /ITS (n=26) | Of 12 | |||||||
| Addo et al,66 2012 | 2 | 2 | NA | 2 | 3 | NA | NA | 9 |
| Alberts et al,59 1992 | 2 | 2 | NA | 2 | 3 | NA | NA | 9 |
| Barsan et al,60 1994 | 2 | 2 | NA | 2 | 3 | NA | NA | 9 |
| Bett et al,45 2004 | 2 | 2 | NA | 2 | 3 | NA | NA | 9 |
| Bett et al,46 1993 | 2 | 2 | NA | 3 | 3 | NA | NA | 10 |
| Black and Brown,62 1973 | 2 | 2 | NA | 2 | 3 | 3 | NA | 9 |
| Breuer et al,54 1999 | 1 | 2 | NA | 2 | 3 | NA | NA | 8 |
| Camerlingo et al,64 2014 | 2 | 2 | NA | 2 | 3 | NA | NA | 9 |
| Diercks et al,56 2010 | 2 | 2 | NA | 3 | 3 | NA | NA | 10 |
| Gaspoz et al,48 1996 | 2 | 2 | NA | 3 | 3 | NA | NA | 10 |
| Herlitz et al,55 1992 | 2 | 2 | NA | 2 | 3 | NA | NA | 9 |
| Ho et al,1 1989 | 3 | 2 | NA | 2 | 3 | 3 | NA | 10 |
| Hodgson et al,38 2007 | 1 | 2 | NA | 2 | 3 | NA | NA | 8 |
| Luiz et al,49 2001 | 1 | 2 | NA | 2 | 3 | NA | NA | 8 |
| Maeso-Madronero et al,52 2000 | 1 | 2 | NA | 3 | 3 | NA | NA | 9 |
| Mellon et al,40 2014 | 2 | 2 | NA | 2 | 3 | NA | NA | 9 |
| Mitic and Perkins,57 1984 | 2 | 2 | NA | 2 | 3 | NA | NA | 9 |
| Moses et al,67 1991 | 2 | 2 | NA | 2 | 3 | NA | NA | 9 |
| Naegeli et al,63 2011 | 2 | 2 | NA | 2 | 3 | NA | NA | 9 |
| Nishijima et al,65 2016 | 2 | 2 | NA | 2 | 3 | NA | NA | 9 |
| Rau et al,53 2008 | 1 | 2 | NA | 2 | 3 | NA | NA | 8 |
| Rustige et al,41 1992 | 2 | 2 | NA | 2 | 3 | NA | NA | 9 |
| Thomassen et al,50 1999 | 3 | 2 | NA | 2 | 3 | NA | NA | 10 |
| Waters et al,47 1983 | 1 | 2 | NA | 2 | 3 | NA | NA | 8 |
| Wolters et al,51 2015 | 2 | 2 | NA | 2 | 3 | NA | NA | 9 |
| Wright et al,58 2001 | 1 | 2 | NA | 2 | 3 | NA | NA | 8 |
| Post-test only (n=2) | Of 12 | |||||||
| Bray et al,44 2015 | 2 | 3 | NA | 2 | 3 | NA | NA | 10 |
| Månsson et al,43 1999 | 2 | 3 | NA | 3 | 3 | NA | NA | 11 |
| Case–control study (n=1) | Of 12 | |||||||
| Tummala and Farshid,42 2015 | 2 | 2 | NA | 2 | 3 | NA | NA | 9 |
As the global rating, following EPHPP guidelines, could not be calculated for studies with non-randomised designs, we also calculated a sum total of the ratings across all dimensions that were applicable to the study design. This total score ranges from 6 to 18 if all six dimensions are applicable and from 4 to 12 if only four dimensions are applicable. Lower scores suggest better overall methodological quality and higher scores suggest poorer quality.
EPHPP, Effective Public Health Practice Project; ITS, interrupted time series; NA, not applicable; RCTs, randomised controlled trials.