| Methods |
'Retrospective cohort'. Women who registered with at least 1 of 10 participating clinics in Australia before 1994: 30% before 1986, 70% from 1986 to 1996. Linked to cancer registry |
| Participants |
Women who received at least 1 IVF treatment. N = 29,700, median age 31 (range 18 to 50) in exposed, median age 30 (range 18 to 53) in unexposed |
| Interventions |
Fertility treatment used: 1182 (6.9%) with clomiphene citrate, 6543 (38.2%) with clomiphene citrate + HMG, 1464 (8.5%) with HMG, 11,153 (65%) with HMG + GnRH agonist, 1771 (8.6%) with other treatments NR. Dosage NR. 6346 (37.0%) with 1 cycle, 3712 (21.6%) with 2 cycles, 5157 (30.1%) between 3 and 5 cycles, 1933 (11.3%) with more than 6 cycles. 134,240 person‐years follow‐up in exposed, 96,794 person‐years in unexposed. Median follow‐up in exposed 7 (range < 1 to 21) years; in unexposed 10 (< 1 to 22) years |
| Outcomes |
Invasive ovarian cancer by histological diagnosis from the Victoria Cancer Registry (see Table 3) |
| Notes |
80% of the cohort sample was followed up until 1996 |
| Risk of bias |
| Bias |
Authors' judgement |
Support for judgement |
| Selection bias |
Low risk |
No person had ovarian cancer at the start of the study and had at least 1 ovary |
| Confounding |
High risk |
No adjusted analysis reported, and groups were not matched or balanced for confounding factors at baseline |
| Performance bias |
High risk |
Medical records; no blinding of assessors to case‐control status |
| Detection bias |
High risk |
Cancer registry; no blinding of assessors to exposure status |
| Attrition bias |
Low risk |
81% exposed and 72% unexposed were followed up |
| Selective reporting (reporting bias) |
Low risk |
All drugs used were reported |
