Skip to main content

Table 4 Risk of bias assessment of the observational studies a

From: Screening for cervical cancer: a systematic review and meta-analysis

Cohort study Representativeness of exposed cohort Selection of unexposed cohort Ascertainment of exposure Demonstration that outcome not present at study start Comparability of cohorts on age Comparability of cohorts on other factors Assessment of outcome Adequate length of follow-up Adequacy of follow-up cohorts Overall scoreb
Herbert et al. [21] Truly representative Same community Secure record Yes Yes No Record linkage Yes (3 years) All subjects followed 8
Rebolj et al. [22]
Case control study Adequate case definition Representative cases Selection of controls Definition of controls Comparability of controls on age Comparability of controls on other factors Ascertainment of exposure Method of ascertainment Non-response rate Overall score b
Andrae et al. [24] Independently validated Consecutive cases Community controls No history of disease Yes No Secure record Same for both groups Same for both groups 8
Aristizabal et al.[15] Independently validated Representative cases Community and hospital controls Not stated Yes Neighborhood Secure record and non-blinded interview Same for both groups Not stated 7
Berrino et al.[16] Independently validated Consecutive cases Hospital controls No history of disease Yes No Secure record Same for both groups Same for both groups 7
Clarke and Anderson [17] Independently validated Representative cases Community controls Not stated Yes Neighborhood and type of dwelling Secure record and non-blinded interview Same for both groups Non-respondents described 7
Decker et al.[25] Independently validated Consecutive cases Community controls No history of disease Area of residence Secure record Same for both groups Same for both groups 9
Hernández-Avila et al.[26] Independently validated Representative cases Community controls Not stated Yes Age of sexual debut, # normal births, # sex partners, SES Non-blinded interview Same for both groups Rate different/no designation 6
Herrero et al.[18] Independently validated Not stated Hospital controls No history of disease No No Non-blinded interview Same for both groups Same for both groups 4
Hoffman et al.[27] Independently validated Not stated Hospital controls No history of disease Yes Race, area of residence, hospital Interview Same for both groups Same for both groups 6
Jiménez-Pérez and Thomas [28] Independently validated Consecutive cases Hospital controls No history of disease Yes Area of residence Non-blinded interview Same for both groups Same for both groups 7
Kasinpila et al.[38] Independently validated Consecutive cases Hospital controls No history of disease Yes Significant risk factors Non-blinded interview Same for both groups Same for both groups 7
La Vecchia et al.[19] Independently validated Representative cases Hospital controls No history of disease Yes No Interview Same for both groups Same for both groups 6
Makino et al.[29] Independently validated Potential for selection bias Community controls No history of disease Yes Area of residence Self-report Same for both groups Same for both groups 7
Miller et al.[31] Independently validated Potential for selection bias Hospital controls No history of disease Yes Length of membership in health program, race/ethnicity Secure record Same for both groups Same for both groups 7
Nieminen et al.[23] Independently validated Consecutive cases Community controls Not stated Yes Socio-demographics, parity, smoking Self-report Same for both groups Rate different/no designation 6
Sasieni et al.[3235] Independently validated Consecutive cases Community controls Not stated Yes Area of residence Secure record Same for both groups Same for both groups 8
Talbott et al.[30] Independently validated Consecutive cases Community controls Not stated Yes Sex, race, street or neighborhood Non-blinded interview Same for both groups Same for both groups 7
Yang et al.[36] Record linkage Not stated Hospital controls No history of disease Yes No Secure record Same for both groups Same for both groups 5
Zappa et al.[37] Independently validated Consecutive cases Community controls Not stated Yes No Secure record Same for both groups Same for both groups 7
  1. aRisk of bias was assessed using the Newcastle-Ottawa Scale [39]; b A higher overall score (maximum = 9) corresponds to a lower risk of bias.
  2. The study met this assessment criterion. ✗ This study could not be assessed with this scale. There was no unexposed cohort; both cohorts received screening. The two groups differed in terms of their age at exposure.