Table 3.
Effect estimates by treatment group
| Outcome | Unadjusted Estimate (95% CI) | Unadjusted p value | Adjusted Estimate d (95% CI) | Adjusted p value |
|---|---|---|---|---|
| Wound Infection | 0.117* | 0.119* | ||
| Non-absorbable versus absorbable suture with fascia non-closure | 1.02 (0.53, 1.94) | 0.954 | 1.01 (0.53, 1.91) | 0.984 |
| Non-absorbable versus absorbable suture with fascia closure | 2.21 (1.07, 4.55) | 0.031 | 2.17 (1.05, 4.45) | 0.035 |
| Fascia non-closure versus closure with non-absorbable suture | 0.60 (0.28, 1.27) | 0.181 | 0.59 (0.28, 1.25) | 0.169 |
| Fascia non-closure versus closure with absorbable suture | 1.29 (0.71, 2.37) | 0.402 | 1.27 (0.70, 2.31) | 0.437 |
| Haematoma/Seromaa | ||||
| - Fishers Exact Test | 0.574 | |||
| Re-Operationa | ||||
| - Fishers Exact Test | 0.013 | |||
| Readmissiona | ||||
| - Fishers Exact Test | 0.246 | |||
| Emergency Presentation | 0.045* | 0.046* | ||
| - Absorbable suture and fascia non-closure | 0.58 (0.25, 1.36) | 0.211 | 0.58 (0.25, 1.35) | 0.205 |
| - Absorbable suture and fascia closure | 1.90 (0.87, 4.16) | 0.110 | 1.87 (0.85, 4.08) | 0.118 |
| - Fascia closure and non-absorbable suture | 0.65 (0.29, 1.47) | 0.304 | 0.64 (0.29, 1.45) | 0.288 |
| - Fascia closure and absorbable suture | 2.12 (0.94, 4.82) | 0.071 | 2.07 (0.92, 4.69) | 0.080 |
| Day 3 Pain Score | 0.223* | 0.221* | ||
| - Absorbable suture and fascia non-closure | 0.15 (−0.29, 0.59) | 0.505 | 0.15 (−0.29, 0.59) | 0.514 |
| - Absorbable suture and fascia closure | −0.24 (−0.69, 0.21) | 0.291 | −0.24 (− 0.69, 0.20) | 0.282 |
| - Fascia closure and non-absorbable suture | 0.22 (−0.22, 0.67) | 0.327 | 0.23 (−0.21, 0.68) | 0.309 |
| - Fascia closure and absorbable suture | −0.17 (− 0.61, 0.27) | 0.459 | − 0.16 (− 0.60, 0.28) | 0.477 |
* Denotes p value for test of interaction (is the effect of one treatment affected by the other treatment)
a Too few events to allow modelling. Fisher’s Exact test was performed to test for any association between group (suture type x fascia closure) and outcome
b These outcomes were assessed at two time points; planned analyses were log binomial regression model (for binary categorisation) and ordinal logistic regression (for original 3 categories), with GEE to account for repeated measures. Due to extremely small numbers in ‘dissatisfied’ category, neither of these modelling approaches was possible. Instead, Fisher’s Exact test has been performed separately at each time point
c Log Poisson regression with robust variance has been used for the adjusted model instead of log binomial regression due to convergence issues
d Adjusted analyses included stratification variables (emergency vs elective caesarean section, BMI category (≤25 vs >25), age, smoking status, diabetes and previous caesarean section