| Methods |
'Case‐control study'. Cases were women with incident histologically confirmed ovarian cancer admitted to the major teaching and general hospitals in 4 geographical regions in Italy (women with borderline tumours were excluded) from 1992 to 1999. Controls were women from the same geographical area who were admitted to the same network of hospitals as cases for a wide range of acute non‐neoplastic conditions (women with hormonal or gynaecological diseases or bilateral oophorectomy were excluded) |
| Participants |
N = 1031 cases, median age 56, range 18 to 79 years. N = 2411 controls, median age 57, range 17 to 79 years |
| Interventions |
Fertility drugs, dosage, and number of cycles not reported. 15/1031 (1.5%) cases were exposed to fertility drugs, and 26/2411 (1.1%) controls were exposed to fertility drugs |
| Outcomes |
Epithelial ovarian cancer by histological diagnosis (see Table 2) |
| Notes |
Length of exposure not reported |
| Risk of bias |
| Bias |
Authors' judgement |
Support for judgement |
| Selection bias |
Unclear risk |
Cases and controls were recruited from the same geographical area |
| Confounding |
Unclear risk |
Factors adjusted for included age, education, menopausal status, age at menopause, parity, spontaneous miscarriages, termination of pregnancy, oral contraceptive use, family history for ovarian cancer, and history of infertility |
| Performance bias |
High risk |
Structured interviewer‐administered questionnaire and checked with medical records. Unclear if blinding of assessors to case‐control status was used |
| Detection bias |
High risk |
How cases were ascertained has not been specified, and it is unclear if blinding of assessors to exposure status was used |
| Attrition bias |
Unclear risk |
Cases and controls assessed for exposure and outcome at the same time as admitted to hospital |
| Selective reporting (reporting bias) |
Unclear risk |
Unclear if all fertility drugs used were investigated |
