Table 3.
Assessment of the quality of studies included in current meta-analysis against the Newcastle-Ottawa criteria [16]
| Author | Year | Adequacy of case definition | Representativeness of cases | Selection of controls | Definition of controls | Comparability of cases and controls | Ascertainment of exposure | Same ascertainment for cases and controls | Non-response rate | Author comment |
|---|---|---|---|---|---|---|---|---|---|---|
| Daling, et al. [9] | 2009 | Yes, with independent validation (1) | Consecutive or obviously representative series of cases (2) | Community controls (3) | No history of disease (4) | Cases and controls comparable (study controls for age and other factors) (5) | Interview not blinded to case/control status (6) | Yes (7) | Rate different (Response rate: cases 67.5 %/controls 43.3 %) (8) | Low response rate among controls (risk of selection bias). Largest study; strongest contributor to summary estimates |
| Trabert et al. [7] | 2011 | Yes, with independent validation (9) | Consecutive or obviously representative series of cases (10) | Community controlsa (11) | No history of disease (12) | Cases and controls comparable (study controls for age and other factors) (13) | Self- completed questionnaire (14) | Yes (15) | Rate different (Response rate: cases 38.2 %/controls 73.3 %) (16) | Low response rate among cases. Controls recruited as friends of cases (risk of selection bias) |
| Lacson et al. [8] | 2012 | Yes, with independent validation (17) | Consecutive or obviously representative series of cases (18) | Community controls (19) | No history of disease (20) | Cases and controls comparable (study controls for age and other factors) (21) | Interview not blinded to case/control status (22) | Yes (23) | Same rate for both groups (Response rate: cases 81.0 %/controls 78.7 %) (24) | Minimised to those aged 18–35 (limits representativeness) |
Explanation of categorisations is presented in Additional file 2 alongside its corresponding number
aControls derived from friends of cases