Data sources
Birth data for the time period July 1st 2000 until June 30th 2008 of all singleton births was provided by NSW Department of Health as recorded in the NSW Midwives Data Collection (MDC). This legislated, population based surveillance system contains maternal and infant data on all births of ≥400 grams birth weight or ≥20 weeks gestation.
The recording of perineal status was altered on the MDC in 2006. Prior to 2006, perineal status was recorded as intact/graze, 1st degree tear, 2nd degree tear, 3rd degree, 4th degree tear, episiotomy and combined episiotomy and tear. Post 2006 combined episiotomy and tear was removed. The two versions of the data were merged for the purpose herein. The data item ‘Episiotomy Yes/No’ was also utilised. The accuracy of the recording of perineal status has previously been shown to have a kappa of 0.84 and 0.82 in two separate and individual studies [
22,
23]. The positive predictive value (PPV) of 1st, 2nd, 3rd and 4th degree tears have been reported as 76.6, 96.6, 72.8 and 100.0 respectively. This PPV provides an overview of the validity of the recording of perineal status in various sources including electronic and paper based medical records. Only women recorded as having a vaginal birth were included in this study.
Data on all hospital admissions was provided by the census data collection, the Admitted Patient Data Collection (APDC). The clinical data component of the APDC utilises the International Classification of Diseases – Australian modification (ICD-10-AM). Probabilistic linkage of the two datasets was undertaken by the Centre for Health Record Linkage. The validity and accuracy of this process has been examined and these datasets have low rates of missing data when compared to medical records and high levels of agreement [
22,
23].
Seven thousand APDC codes related to all patient admissions over the period from 2000–2008 in NSW were sorted individually by the first author, and coded under sixteen sub headings. The purpose of sorting these codes was to determine the reason for, and frequency of, admission for women within a 12 month time period following birth for procedures related to perineal/pelvic floor trauma. The identified subheadings were: Nervous System; Skin; Skeleton; Renal/Ureter/Bladder; Fertility/Pregnancy; Miscellaneous (this included such codes as oncology related therapies); Eyes; Ears; Respiratory; Cardiac; Gastrointestinal; Pelvic/Sphincter/Urethral; Lymphatics; Breast; Psychiatric and Male Specific Codes. Any procedural codes relating to diagnosis and repair of the initial trauma were not included in the final coding categories. A list of potential Medicare Benefit Schedule (MBS) procedural codes was identified by the first author. This initial sorting of codes was completed and independently reviewed by two of the co-authors for accuracy. This list was then reviewed independently by three specialists in fields related to perineal trauma and outcomes including: 1. A midwife running a postnatal specialist perineal trauma clinic, 2. An obstetrician, and 3. A colorectal surgeon. Through this consensus process a refined list of codes that were specifically associated with therapies and treatment for morbidities potentially occurring as a result of severe perineal trauma from subheadings Renal/Ureter/Bladder, Fertility/Pregnancy, and Pelvic/Sphincter/Urethral was agreed upon. These codes were then grouped into the following four categories related by procedure and physiology: 1. Vaginal repair, 2. Fistula repair, 3. Faecal and urinary incontinence repair, and 4. Rectal/anal repair. There were 34 subgroups/diagnostic codes for vaginal repair, eight for fistula, eight for faecal and urinary incontinence and 11 for rectal/anal repair.
The first pregnancy recorded in the MDC with vaginal birth documented as the mode of birth was considered the index pregnancy for this dataset regardless of parity. For this reason, sub-analyses were undertaken according to parity.
Ethical approval was obtained from the NSW Population and Health Services Research Ethics Committee, Protocol No.2010/12/291.
Data analysis
Descriptive analyses of short and long term morbidity associated with all types of perineal trauma was produced utilising SPSS v.19 (IBM). Frequency distributions were used to classify the population and descriptive statistics the morbidity outcomes. Relative risk was calculated between factors and events, Odds Ratios (OR) are reported for rare outcomes. Due to the number of associations examined, the level of statistical significance was set at <0.001.