Bekassy 1996.

Methods	Retrospective cohort study Comparison group: A) External ‐ matching for age, parity and time of delivery B) Internal (self‐matching) Information source ‐ Anaesthetic records of University Hospital of Lund, Sweden, and National Medical Birth Registry at the National Board of Health and Welfare, Stockholm
Participants	A) Treated group ‐ 250 women who had undergone LC at University Hospital of Lund, Sweden, and had a subsequent delivery, between January 1980 and June 1988 Untreated group ‐ 250 women selected from the National Medical Birth Registry B) Of the 250 women of the treated group, 148 were parous. For these women, self‐matching was also possible.
Interventions	LC ('laser miniconisation')
Outcomes	PTB (< 37 weeks); PTB (< 37 weeks) (nulliparous); PTB (< 37 weeks) (parous); PTB (< 37 weeks) (single cone); PTB (< 37 weeks) (repeat cones); CS; instrumental deliveries (ventouse/forceps); prolonged labour (> 12 hours); cervical stenosis; LBW (< 2500 g); perinatal mortality; stillbirth
Notes	From January 1980 to June 1988, 1485 women between age 16 to 58 were treated by carbon dioxide laser miniconisation because of CIN at University Hospital of Lund. These women were identified retrospectively via certain operation code numbers in the Anaesthetic hospital records. Each woman had also a specific 10‐tailed patient identification number (PIN), which is also used by the National Medical Birth Registry to register births in Sweden. The information of these 1485 women was transferred to a magnetic tape which was then run against data held at the National Medical Birth registry and 250 women having a delivery after treatment (3 had twin pregnancies) were identified. Of these women, 245 delivered at the Department of Obstetrics and Gynaecology, University Hospital of Lund, and the other 45 at 21 different hospitals around Sweden. 20 women had LC twice before pregnancy.
Risk of bias
Bias	Authors' judgement	Support for judgement
Incomplete outcome data (attrition bias) All outcomes	Low risk	The information was obtained from hospital records and national registries
Selective reporting (reporting bias)	Low risk	No reporting bias is obvious
Other bias	Low risk	No other obvious source of bias
Relevant assignment described?	Low risk	Yes, treatment performed on clinical grounds
Representative intervention group?	Low risk	All eligible women having LC at the University Hospital of Lund between January 1980 to June 1988
Representative comparison group?	Unclear risk	A) The untreated group was not drawn from the same source as the treated group B) Internal comparison group (self‐matching)
Comparability of treatment groups?	Low risk	A) Matching for age, parity and time of delivery B) Internal comparison group (self‐matching)