Risk Models for Benchmarking Severe Perineal Tears during Vaginal Childbirth: a Cross-sectional Study of Public Hospitals in South Australia, 2002-08

Background: The incidence of third- and fourth-degree perineal tears during vaginal childbirth is being increasingly used as an indicator of the safety and quality of maternity health care services. In order to make fair comparisons across hospitals it may be necessary to estimate a probability of severe perineal tears for every woman, taking into account her risk profile. Methods: Logistic regression analysis was used to estimate the probabilities of third- and fourth-degree tears (n = 1582; 2.4%) in 65 598 vaginal births in publicly funded hospitals in South Australia, 200208. Results: Maternal age =25 years, primiparity, instrument assistance, Asian or African ethnicity, shoulder dystocia and increasing birthweight were all identified as factors that are associated with an increased risk of perineal tears. In parous women, episiotomy, with or without instrument assistance, was associated with more tears; but among nulliparous women, episiotomy was associated with significantly fewer tears when forceps assistance was required, and showed little or no association with tearing in vacuum-assisted or unassisted (spontaneous) births. Conclusion: The probabilities of severe perineal tears in first-time mothers giving birth to a term singleton with cephalic presentation, may range from under 1% to over 40%, according to a minimalist model containing only predictors unrelated to clinical management. If instrument assistance and episiotomy are also incorporated into the modelling, the estimated probability of tearing may exceed 50% in high risk individuals. Such variation highlights the need for risk adjustment when comparing hospitals with respect to their incidence of third or fourth degree perineal tears.