Episiotomy healing assessment: Redness, Oedema, Ecchymosis, Discharge, Approximation (REEDA) scale reliability

OBJECTIVE: to analyse the Redness, Oedema, Ecchymosis, Discharge, Approximation (REEDA) scale reliability when evaluating perineal healing after a normal delivery with a right mediolateral episiotomy. METHOD: observational study based on data from a clinical trial conducted with 54 randomly selected women, who had their perineal healing assessed at four time points, from 6 hours to 10 days after delivery, by nurses trained in the use of this scale. The kappa coefficient was used in the reliability analysis of the REEDA scale. RESULTS: the results indicate good agreement in the evaluation of the discharge item (0.75< Kappa ≥0.88), marginal and good agreement in the first three assessments of oedema (0.16< Kappa ≥0.46), marginal agreement in the evaluation of ecchymosis (0.25< Kappa ≥0.42) and good agreement regarding redness (0.46< Kappa ≥0.66). For the item coaptation, the agreement decreased from excellent in the first assessment to good in the last assessment. In the fourth evaluation, the assessment of all items displayed excellent or good agreement among the evaluators. CONCLUSION: the difference in the scores among the evaluators when applying the scale indicates that this tool must be improved to allow an accurate assessment of the episiotomy healing process.


Introduction
Episiotomy, a common procedure in obstetric care, is associated with the need for suture and healing complications in the postpartum period, such as blood loss, oedema, haematoma, infection wound dehiscence and perineal pain (1) .
Inflammatory signs, such as oedema, ecchymosis, redness and pain, occur from the first hours after delivery and may remain beyond the hospitalization period. A randomized controlled trial which compared two different perineal repair techniques identified that oedema, redness and ecchymosis occurred in 26.2%, 6.6% e 3.3% of women who had episiotomy or second degree laceration at the first 24 hours after childbirth, respectively. On the fourth day after delivery, the distribution of these signs was 11.5% de oedema, 4.9% redness and 8.2% ecchymosis (2) .
In an online survey completed by 2,400 women who gave birth in American hospitals from July 2011 through June 2012, 41% of those who had a vaginal birth reported a painful perineum for two months postpartum. Seven per cent of these women reported the same problem after 6 months postpartum. Perineal pain was strongly related to whether the woman had an episiotomy (18%) or did not (9%) (p < 0.01) (3) . Birth position, fundal pressure, guided pushing, birth weight, perineal management manoeuvres during labour and suture material and technique might also influence postpartum perineal pain, as these parameters influence the rates and severity of spontaneous perineal trauma and episiotomies (4)(5) .
Beyond the perineal pain, perineal trauma complications in the postpartum period may include wound infection and dehiscence. There is limited data on the prevalence of perineal wound dehiscence related to episiotomy or perineal tears, but rates ranging from 0,1% to 5,5% have been reported (6) .
Despite the effects of perineal healing complications on maternal recovery, the prevalence of these morbidities is poorly known, mainly as a consequence of the difficulty of healthcare professionals to identify them in clinical practice. The fact that breastfeeding issues and newborn care are considered as more important than maternal wellbeing and also the lack of a defined tool to assess the perineal condition impairs the detection of these problems. Assessment tools have been proposed for assessing perineal healing in the postpartum period, such as the PAT (Perineal Assessment Tool) and REEDA (Redness, Oedema, Ecchymosis, Discharge, Approximation) scales (7) .
These scales use similar categories and descriptors to assess the same items. However, the main difference between them is that the PAT operational settings are less objective than in the REEDA scale, and therefore, the former has low reliability (7) . A systematic clinical evaluation of the postpartum perineal condition, with the use of these scales, is not part of the standard care provided to postpartum women.
The REEDA scale is a tool for assessing perineal healing that was primarily developed by Davidson (8) and later reviewed by Carey (9) . It includes five items related to the healing process: hyperaemia, oedema, ecchymosis, discharge and coaptation of the wound edges (Redness, Oedema, Ecchymosis, Discharge, Approximation -REEDA) (8)(9) . It can be used to assess all types of postpartum perineal trauma.
This scale has been used in recent studies that have investigated interventions aiming to assess perineal suture techniques (10) , perineal pain in the suture (11) , postpartum perineal care (12)(13) and the effect of laser irradiation on perineal pain (14) . However, this instrument lacks validation to be incorporated in the clinical practice.
The validation of a scale involves steps that include analysing its reliability, which refers to error (in the statistical sense) inherent in the scores (15) . The reliability includes the degree of agreement between observers in simultaneous and independent assessments in relation to the scores of an instrument (16) .
Health professionals use scales, questionnaires and tests to identify signs and symptoms and to assess the results of interventions. Repeated measures of a given condition, often undertaken by different professionals, should agree well enough in order to allow comparisons and to identify real change in an individual condition when it occurs (15) . The aim of this study is to analyse the reliability of the scale REEDA as a tool for the clinical assessment of perineal healing after episiotomy.

Method
This is an observational study based on data obtained from a randomised, triple-blind, controlled trial on the effectiveness of Low-level Laser Therapy (LLLT) for the healing of episiotomies.
Women were recruited in the rooming-in unit of the University Hospital of University of São Paulo, Brazil (HU-USP). The sample size was calculated based on the outcomes of a randomised clinical trial (14) . A 2.0-point reduction in the pain score reported by women after the LLLT irradiation was the main outcome. With a significance level of 5% and a test power of 90%, a study sample size of at least 24 women in each group was obtained. In the current study, the final sample size was 54 women, who were randomly divided into two groups: the experimental group (n = 29), who received LLLT irradiation, and the control group (n = 25), who did not receive LLLT irradiation.
The current study used all of the women who participated in the original trial because the results of the study indicated that the groups were homogeneous regarding sociodemographic and clinical characteristics and postpartum perineal pain. The main outcome of the study (perineal healing) did not differ between the groups after LLLT irradiation (17) .
Women who met the following inclusion criteria were The Peri-Rule Ruler™ (18) was used to assess the scale items requiring measurement. It was packed in a layer of PVC film and reused after cleaning with soap and water, followed by disinfection with 70% alcohol.
The item hyperaemia in this study was assessed only regarding its area, independently of being unilateral or bilateral, as there is no such option in the REEDA scale.
The reliability analysis investigated the degree of agreement between the observers' evaluations. A greater agreement between the evaluations provided by the professionals was considered greater reliability.
For this analysis, we used the Kappa Coefficient, which ranges from 0 to 1. A kappa value ≥ 0.75 was considered an excellent agreement, and a result > 0.45 and < 0.75 indicated good agreement. A value ≤ 0.45 was considered marginal agreement (19) .  Figure 1 -Redness, oedema, ecchymosis, discharge and approximation of the edges of the lesion assessment scale (REEDA) (7) www.eerp.usp.br/rlae Alvarenga MB, Francisco AA, Oliveira SMJV, Silva FMB, Shimoda GT, Damiani LP. At the first assessment, the evaluators assigned the same total score on the scale REEDA to 44 (81.5%) women. The differences in the scores among the remaining 10 women ranged from 1 to 5 points. At the second assessment, the score was the same in 72.2% of the postpartum women, and the differences among the remaining 15 women ranged from 1 to 3 points. At the third evaluation, the total score was the same for 83.3% of the women, and among the remaining women, their differences ranged from 1 to 3 points. At the fourth assessment, the scores coincided in 83.6% of cases, and the differences ranged from 1 to 2 points among the remaining women (Table 1). At the first evaluation of the REEDA items, a few differences were observed among the means of three out of the five score items, however the means of coaptation and discharge items were similar. For the total scoring, the means were also similar ( Table 2).
At the second evaluation, the mean values of the ecchymosis, discharge and coaptation items were similar. For the oedema and hyperaemia items, the difference ranged from 0.11 to 0.28, respectively. The difference in the mean total score was 0.28 ( Table 2).
The third evaluation revealed that the mean of the score of each item analysed by the main researcher and by the judge were similar to each other, except for hyperaemia. This similarity also occurred with the mean total score (Table 3)

Discussion
Adopting protocols with well-defined criteria is essential for systematically assessing and treating injury. This study aimed to assess the inter-observer reliability of the REEDA scale as a tool for the quantitative assessment of perineal healing after episiotomy.
The excellent agreement obtained in the evaluation of the discharge item is related to the low frequency of this event in the women of this sample. Only two women experienced this event at the third or fourth assessment. When the elements of the sample are very similar regarding the studied event, it is more difficult for the instrument to reliably indicate different item degrees (16) .
The smallest REEDA score for the item coaptation material has been fully absorbed. In this healing stage, it is expected that the perineal tissue is undergoing a proliferation process (6) , however the perineal wound may be partially or totally dehisced, involving superficial tissues such skin or as the deeper layers, such as muscles.
The inability of professionals to differentiate normal and abnormal wound healing, associated with the millimetre dimensions of REEDA scale to assess the approximation of the wound edges might justify the lower value of the Kappa coefficient observed in this assessment.
In the hyperaemia item, difficulties when applying the REEDA scale arise from the fact that this item is bilaterally assessed. In clinical practice, hyperaemia might be observed in only one side of the incision.
Consequently, in this study, this item was assessed only regarding its area when a unilateral occurrence prevented a full evaluation.
The marginal agreement in the oedema and ecchymosis evaluation, obtained in this study, highlights the complexity of the application of the REEDA scale resulting from the precision with which they are assessed.
The ecchymosis can occur discretely. Moreover, it might be difficult to distinguish between the occurrence of hyperaemia and ecchymosis, even when the evaluators are trained (7) .
The difficulties in defining and measuring the perineal oedema are related to the fact that the REEDA scale classifies its extension from one to two centimetres from the incision. This measurement can be confused depending on the protrusions of tissue resulting from tight stitches of the suture. Moreover, oedema is assessed only regarding the width from the edge of the incision, not the length and depth of the tissue that presents induration (7) .
Other studies also highlight the difficulty of identifying and assessing perineal oedema and ecchymosis in clinical practice with the use of other measurement instruments.
In a study (20)  cases there was difficulty in the oedema classification, and there was difficulty in 4 cases of ecchymosis. The less experienced professionals displayed more uncertainty in the application of the scale (20) .
The data of our study indicate that the REEDA scale scores also had better agreement among the evaluators when used at the follow-up visit, when the items with less agreement (hyperaemia, oedema and ecchymosis) were no longer present. These local inflammatory signs are expected in an early phase of the healing process and decrease with the evolution of local reactions and absorption of the suture material. After nearly two weeks, the cell matrix formation and tissue remodelling is generally complete, even though this process can take several months (21) . These results indicate the need for further research to redefine the criteria for evaluating those items.
Limitations of this study included a small sample size, which was not calculated to detect a difference when comparing the evaluation of the judges. Notwithstanding, sample was enough to identify the items for which there was a low inter-rater agreement. The assessments were carried out by several professionals, which increase the variability of the data but it also allows to verify the use of the REEDA scale in a clinical setting.

Conclusions
Of the five items of the REEDA scale, the A reliable instrument for assessing perineal healing is valuable to nurse-midwives, midwives and other caregivers, as a concise evaluation tool may help facilitate measures to improve perineal care.