A C TA Obstetricia et Gynecologica
AOGS M A I N R E SE A RC H A R TI C LE
Perineal support and risk of obstetric anal sphincter injuries: a Delphi survey KHALED M. K. ISMAIL1, ELENA PASCHETTA2, DIMITRIOS PAPOUTSIS3 & ROBERT M. FREEMAN
4
1
Birmingham Centre for Women’s and Children’s Health, School of Clinical & Experimental Medicine, College of Medical & Dental Sciences, University of Birmingham, Birmingham, UK, 2Postgraduate School of Psychiatry, Neuroscience Department, University of Turin, Turin, Italy, 3Department of Obstetrics and Gynaecology, Royal Shrewsbury Hospital, Shropshire, and 4Plymouth Hospitals NHS Trust and Plymouth University Peninsula Schools of Medicine and Dentistry, Plymouth, UK
Key words Obstetric anal sphincter injuries, Delphi survey, hands-on, hands-poised, birth Correspondence Khaled M. K. Ismail, Obstetrics and Gynaecology, 3rd Floor, Birmingham Women’s NHS Foundation Trust, Edgbaston B15 2TG, UK. E-mail: [email protected] Conflict of interest The authors have stated explicitly that there are no conflicts of interest in connection with this article. Please cite this article as: Ismail KMK, Paschetta E, Papoutsis D, Freeman RM. Perineal support and risk of obstetric anal sphincter injuries: a Delphi survey. Acta Obstet Gynecol Scand 2015; 94: 165–174. Received: 13 July 2014 Accepted: 20 November 2014 DOI: 10.1111/aogs.12547
Abstract Objective. To explore the views of a multidisciplinary group of experts and achieve consensus on the importance of perineal support in preventing obstetric anal sphincter injuries (OASIS). Design. A three-generational Delphi survey. Setting. A UK-wide survey of experts. Population. A panel of 20 members consisting of obstetricians, midwives and urogynecologists recommended by UK professional bodies. Methods. A 58-item web-based questionnaire was sent to all participants who were asked to anonymously rate the importance of each item on a six-point Likert scale. They were asked to rate their level of agreement on statements related to hands-on/hands-poised techniques, the association of hands-poised/hands-off approach with OASIS, the need to implement perineal support and the need to improve the evidence to support it. Systematic feedback of responses from previous rounds was provided to participants. Main outcome measures. To achieve consensus on key areas related to perineal support. Results. The response rate was 100% in all three iterations. There was consensus that current UK practice regarding perineal protection was not based on robust evidence. The respondents agreed that hands-poised/hands-off and OASIS are causally related and that hands-poised was misinterpreted by clinicians as hands-off. Although 90% of experts agreed that some form of randomized trial was required and that all would be prepared to take part, there was also consensus (75%) that in the meantime, hands-on should be the recommended technique. Conclusions. Our results highlight the current lack of evidence to support policies of perineal support at time of birth and the need to address this controversial issue. NICE, National Institute for Health and Care Excellence; OASIS, obstetric anal sphincter injuries; RCT, randomized controlled trial.
Abbreviations:
Introduction Obstetric anal sphincter injuries (OASIS) are a serious complication of vaginal birth with the incidence in the UK rising from 1.8 to 5.9% in primiparous women over a 12-year period (1). Based on an average OASIS rate of 5.9% it is estimated that 48 000 of these women sustain significant anal sphincter trauma in the UK per year.
Key Message Perineal support: hands-on or hands-off at birth? The current clinical practice is not evidence-based. Consensus among experts was that a randomized controlled trial is needed (90%) and that in the meantime, hands-on should be the recommended practice (75%).
ª 2014 Nordic Federation of Societies of Obstetrics and Gynecology, Acta Obstetricia et Gynecologica Scandinavica 94 (2015) 165–174
165
Perineal support and OASIS: a Delphi survey
OASIS represents a clearly identified patient safety problem with a negative impact on quality of life and with long-term adverse health-economic effects. Obstetric anal sphincter injuries is quoted to be the most important risk factor for female anal incontinence affecting young women (2,3). Despite optimal primary repair, 30–50% of affected women suffer from a varying degree of anal incontinence and experience a significant negative impact on their quality of life (4). Several studies have reported that the risk of anal incontinence is much higher with subsequent deliveries and that symptoms get worse by time (5–7). An estimated 40% of these women will opt for a future elective cesarean section to avoid worsening of symptoms or a repeat OASIS from a vaginal birth (8,9). An American study in 1996 reported at that time that the average cost of anal incontinence management following OASIS was $17 166 per patient (10). Hospital episode statistics from a recent UK study and information from the National Institute for Health and Care Excellence (NICE) fecal incontinence costing guide for secondary OASIS repair have helped estimate a cost of £1625 per patient, an annual cost of £78 million (1,11). Moreover, allegations of negligence for perineal damage caused during labor in the UK resulted in 441 claims in 2000–2010, thus representing the fourth highest number of claims in obstetrics. The total value of those claims is estimated to be £31.2 million, with 85% of them related to misdiagnosis of perineal trauma (12). The current UK-based practice in second stage of labor involves both the “hands-on” and “hands-poised” techniques, with approximately 50% of midwives using either method (13–15). The traditional teaching has been with “hands-on” but in recent years “hands-poised” or “hands-off” has become popular, although neither has been well described (16,17). The reason for the change from the traditional “hands-on” to “hands-poised/handsoff” is unclear and without any evidence base (18). It has been suggested that this change and the reluctance to use episiotomy might have contributed to the rising incidence of OASIS (1). In contrast, implementation of “hands-on” perineal support in Norway has resulted in a 50% reduction in OASIS rates (4,19,20). As there are differing views on this subject and a poor evidence base, we conducted a Delphi consensus survey among midwives, obstetricians and urogynecologists who were recommended by their national professional bodies in the UK. Our primary aim was to explore the level of consensus among a multidisciplinary group with regard the controversial area of “handson” vs. “hands-poised/hands-off” techniques at the time of childbirth.
166
K.M.K. Ismail et al.
Material and methods The Delphi technique is a formal consensus method used in medical and health service research to obtain and integrate the views and opinions of experts in a particular field (21,22). The Delphi approach involves asking a group of experts to participate in a number of consecutive rounds of questionnaires which they rate anonymously and independently using a numerical scale. Systematic feedback from the previous iterations is provided to the participants during the consecutive stages of the process, giving them the opportunity to reconsider their original opinions and to modify or persist with them, or move towards consensus (21). The Delphi method is a multistage process which usually ranges between two and four consecutive rounds depending on the rates of responses (23–25). The level of consensus can be considered achieved if a given proportion of participants are in agreement. This proportion varies between different studies, with most of them suggesting levels between 60 and 80% (26–29). The Delphi survey technique with its cardinal features of several iterations, feedback of results, anonymity of responses and the opportunity for participants to reconsider their views, is considered an effective method for transforming individuals’ opinions into group consensus, especially when greatly diverse responses are anticipated (30). The features of the Delphi method minimize the bias arising from factors such as powerful personalities, seniority and experience, which can potentially affect the group’s opinion (31). Therefore, this method allows the participants to express their views honestly without intimidation, inhibition or peer group pressure while achieving consensus among the group (21). Some researchers in the literature report that the major limitation of this methodology is the lack of scientific validation and sufficient measurement of the reliability of the findings, thus challenging the scientific quality of the Delphi study technique (32). Nevertheless, the role of the Delphi methodology comes into play when hard evidence is lacking and the conclusions rely on the views of a panel of experts. In this case it is argued that the Delphi method should not be subject to the same validation criteria as used for other research methodologies (33,34). The panel of experts for this Delphi survey was selected for their interest, level of knowledge and clinical experience in the field of OASIS in the UK. Selection of experts was done following recommendations from senior representatives of the Royal College of Obstetricians and Gynaecologists, Royal College of Midwives and British Society of Urogynaecology. Twenty experts were involved
ª 2014 Nordic Federation of Societies of Obstetrics and Gynecology, Acta Obstetricia et Gynecologica Scandinavica 94 (2015) 165–174
K.M.K. Ismail et al.
in the Delphi survey, 10 of whom were midwives and 10 obstetricians and urogynecologists. There are reports that a Delphi expert panel requires at least 10–15 experts to obtain reliable results (35). However, a panel of 20 members is usually seen in consensusbased research. Therefore, to account for non-response we approached 22 experts, 20 of whom agreed to participate (25,36,37). All the participants were provided with an invitation e-mail letter explaining the aims, procedures and requirements of the Delphi procedure. The Delphi procedure was web-based and was completed between May and November 2013. Although webbased and conventional Delphi processes have not been formally compared, web-based Delphi surveys have been found to be feasible, cost-effective, time saving, and better accepted by users than paper-based methods (38). A webbased questionnaire was developed and a three-round electronic-mail survey was conducted. Members of the panel of experts were asked to provide any questions or statements relating to hands-on, hands-poised or handsoff techniques and OASIS that they felt were relevant. All questions and statements sent by the panel members were included in the survey. The questionnaire included 58 items grouped in six progressive sections. The first section included items related to current evidence in relation to perineal support and risk of OASIS. In sections 2 and 3 we gathered the experts’ opinions on hands-on and hands-off techniques at the time of the delivery of baby’s head. In section 4 we explored the panellists’ views on perineal support at the time of delivery of baby’s shoulders. Section 5 included key items related to other intrapartum interventions possibly associated with increased risk of OASIS. Finally, in section 6 we explored the experts’ personal preferences for hands-on/hands-poised/ hands-off techniques based on their clinical experience and the need for future recommendations and research. The questionnaire was completed online by the participants (using the Survey Monkey facility). For each iteration, an accompanying e-mail including instructions and the web-based questionnaire was sent to all participants. To ensure security and confidentiality, each participant received a personal link with the e-mail invitation allowing individual access to the questionnaire on the online survey which was specifically designed for the present study. Two electronic reminders were sent for each iteration. Participants were asked to complete the questionnaire by rating the importance of each item on a sixpoint Likert scale: 1 – strongly disagree; 2 – disagree; 3 – slightly disagree; 4 – slightly agree; 5 – agree; 6 – agree strongly. The results from the questionnaires were automatically entered into an anonymous database and analyzed by obtaining group figures from an independent researcher
Perineal support and OASIS: a Delphi survey
with no expertise in the OASIS field (EP). A summary of the results showing the distribution of the group’s responses (mean SD; median value scores) and patterns of agreement/disagreement of the previous iteration were fed back at each stage to be evaluated by the panel members. In the subsequent iterations, participants were requested to answer the same questions taking into account the mean scores calculated for each item from the group answers to the previous iteration. Participants were asked to add free comments related to clarity of items and occasionally some of the items were reworded slightly in line with the group suggestions to clarify the meaning. Agreement and disagreement consensus was set a priori and this was considered to be achieved for an item if at least 60% of all experts scored ≥4 or ≤3, respectively. This is in keeping with other studies reporting rates between 60 and 80% (26–29). In the first iteration, the questionnaire was e-mailed to the expert group with a request for members to rate their level of agreement with each item on a scale from 1 to 6. Within the questionnaire an additional text box was provided for respondents to enter any further comments and suggestions. If panel members felt that important topics were missing from the provided list, they were encouraged to suggest additional potential items. In the second and third iteration the mean (SD) and median scores of each item from the previous iteration were reported back to the group. Changes or new items suggested by experts in the first and second round were added to the Delphi questionnaire list, and this expert group was requested to complete the questionnaire again in light of the results of the previous rounds.
Results Of the initial 22 experts invited, 20 finally participated in the Delphi survey. During the process the response rate was consistently 100%, with all 20 experts participating in all three rounds. Modifications were made in the nine questions based on the comments fed back by the participants during the subsequent rounds. The final results with levels of agreement or disagreement for each section after the third iteration are presented in Tables 1–3. The gradual change in the levels of consensus, as the participants were fed back the results from each iteration, is presented in Figures 1 and 2. The only major changes in the levels of consensus from round one to round three were identified in questions 36 and 37, which involve the hands-off technique (Figure 2). In section 1 of the questionnaire (Table 1) experts agreed (85%) that the change from the traditional method of hands-on perineal support to the hands-off/
ª 2014 Nordic Federation of Societies of Obstetrics and Gynecology, Acta Obstetricia et Gynecologica Scandinavica 94 (2015) 165–174
167
Perineal support and OASIS: a Delphi survey
K.M.K. Ismail et al.
Table 1. Final results from section 1 of the questionnaire with items related to current evidence in relation to perineal support and risk of obstetric anal sphincter injuries (OASIS).
Section 1 (questions 1–25): current evidence The evidence on which UK current practice is based with regard to hands on or off/poised is robust The change from the traditional method of support (hands on) to hands off/poised was evidence-based The interpretation of the HOOP trial has changed practice Hands poised tends to be interpreted as hands off among clinicians The aim of hands off/poised is to encourage/potentiate normality The introduction of the hands off/poised approach is causally related to an increased risk of OASIS The increase incidence of OASIS is mainly due to improved assessment (PR examination) rather than hands off/poised The increase in incidence of OASIS is mainly due to increased awareness and improved recognition by the delivering clinician (acoucheur) The majority of Midwives and Obstetricians are confident with perineal management during the second stage of labor Midwives and Obstetricians in my unit have clear guidelines on whether to use hands on or hands off/poised Evidence based antenatal interventions to reduce perineal trauma are recommended to all women planning vaginal birth An RCT comparing hands on vs. hands off/poised is needed An RCT is unethical following the results of the Norwegian studies (Laine et al.) The data supporting hands on from Norway are not relevant to the UK If an RCT is funded, I would agree to participate I am in equipoise with regards to hands on vs. hands off/poised When conducting RCTs on hands on vs. hands off/poised we need to control for the speed and timing of delivery by communicating with the woman Good communication (and therefore collaboration between the woman and her attendant) is as important as what we do with the hands Good communication (and therefore collaboration between the woman and her attendant) is more important than what we do with the hands Hands on would be easy to re-introduce Training of both techniques should be facilitated and standardized by a video demonstration The technique for perineal support at instrumental delivery should be the same as normal delivery The best technique of perineal support to minimize perineal trauma has been established It is necessary to establish the best technique of perineal support to minimize perineal trauma before planning any clinical trials The practitioner should encourage the woman to deliver between contractions – minimize active pushing to prevent OASIS
Mean SD
Median
Disagreement scores ≤3 (%)
Agreement scores ≥4 (%)
2.2 0.89
2
90
10
2.35 1.13
2
85
15
0.67 0.74 0.71 1.18
5 5 5 4
5 0 5 20
95 100 95 80
3.7 0.97
4
35
65
4.3 0.86
4.5
15
85
2.6 1.18
2
80
20
2.2 1.00
2
90
10
2.15 0.87
2
95
5
4.85 0.98 2.65 1.34
5 2
10 80
90 20
2.5 5 2 5
80 0 80 0
20 100 20 100
4.85 5.15 4.75 4.15
2.75 5.15 2.7 5.2
1.11 0.67 1.17 0.52
5.05 0.75
5
0
100
3.3 1.12
3
70
30
3.75 1.25 4.8 1.15
4 5
40 10
60 90
4.75 1.20
5
15
85
2.6 1.14
2
75
25
4.65 1.03
5
15
85
3.5 1.23
4
45
55
Scores 1–6: 1 = disagree strongly; 2 = disagree; 3 = slightly disagree; 4 = slightly agree; 5 = agree; 6 = agree strongly. RCT: randomized controlled trial.
poised technique in the UK was not evidence-based. The reason given for this shift in practice was to encourage normality in labor and birth. All members strongly agreed (100%) that hands-poised tends to be interpreted as hands-off, with 80% of them agreeing that the hands-off/ poised approach is causally related to an increased risk of
168
OASIS. The majority of panellists (90%) recommended a randomized controlled trial (RCT) to compare hands-on vs. hands-off/poised in relation to OASIS; all of them (100%) agreed to participate in such a trial. However, there was consensus that it is necessary to establish the best technique of perineal support to minimize perineal
ª 2014 Nordic Federation of Societies of Obstetrics and Gynecology, Acta Obstetricia et Gynecologica Scandinavica 94 (2015) 165–174
K.M.K. Ismail et al.
Perineal support and OASIS: a Delphi survey
Table 2. Final results from sections 2–4 of the questionnaire with items related to hands-on and hands-off techniques at the time of the delivery of baby’s head and at the time of delivery of baby’s shoulders. Mean SD Section 2 (questions 26–33): hands-on technique The non-dominant, anterior, usually left hand is placed on the fetal head in order to slow down the passage through the perineal structures The non-dominant, anterior, usually left hand is placed on the fetal head in order to keep the flexion of the fetal head during the time of expulsion The non-dominant, anterior, usually left hand is placed on the fetal head in order to help the fetal head to be gradually extended during the time of expulsion About the dominant, posterior, usually right hand: the thumb and the index finger of the dominant hand are placed alongside the fourchette and vaginal opening About the dominant, posterior, usually right hand: the fingertips of the dominant hand are pressed against the perineum and a region of parietal eminences of the fetal head and pulled towards each other and towards the fourchette About the dominant, posterior, usually right hand: the palm of the dominant hand is placed against the central part of the perineum and the posterior fourchette About the dominant, posterior, usually right hand: the fingers of the dominant hand execute the Ritgen maneuver during expulsion to extend the fetal head About the dominant, posterior, usually right hand: the remaining part (apart from the thumb and the index finger) of the dominant hand executes a pressure on the fetal head to facilitate the extension Section 3 (questions 35–37): hands-off technique The hands are only applied to the head if it is delivering rapidly The hands are only applied to the head and perineum if the head is delivering rapidly Perineum should be assessed to decide whether to perform episiotomy or allow tearing Section 4 (questions 39–41): delivering the shoulders More attention should be given to supporting delivery of shoulders in preventing OASIS Adequate lateral flexion should be used to facilitate the delivery of the shoulders During the delivery of the posterior shoulder the acoucheur’s hand should be placed against the perineum to support
Median
Disagreement scores ≤3 (%)
Agreement scores ≥4 (%)
4.75 0.85
5
5
95
4.1 1.44
5
30
70
4.3 1.17
5
25
75
4.8 0.76
5
5
95
3.85 1.38
4
40
60
3.75 1.37
4
40
60
3.45 1.39
3.5
50
50
3.45 1.50
3
55
45
2.65 1.30
2
70
30
2.6 1.27
2
75
25
4.65 1.13
5
10
90
4.95 1.09
5
5
95
4.8 0.76
5
5
95
4.15 1.26
4
20
80
Scores 1–6: 1 = disagree strongly; 2 = disagree; 3 = slightly disagree; 4 = slightly agree; 5 = agree; 6 = agree strongly.
ª 2014 Nordic Federation of Societies of Obstetrics and Gynecology, Acta Obstetricia et Gynecologica Scandinavica 94 (2015) 165–174
169
Perineal support and OASIS: a Delphi survey
K.M.K. Ismail et al.
Table 3. Final results from sections 5 and 6 of the questionnaire with items related to other intrapartum interventions possibly associated with increased risk of OASIS and the need for future recommendations and research. Mean SD Section 5 (questions 43–48): other intrapartum interventions The following is important as a determinant 4.5 of risk of OASIS: time allowed for head descent The following is important as a determinant 4.6 of risk of OASIS: length of pushing The following is important as a determinant 4.4 of risk of OASIS: position in labor The following is important as a determinant 3.6 of risk of OASIS: antenatal perineal massage The following is important as a determinant 2.55 of risk of OASIS: intrapartum perineal massage Different hands-on techniques are required 4.15 for women giving birth in different positions Section 6 (questions 49–58): future needs We need to conduct a descriptive summary 5.2 of details (even filmed demonstrations) of all possible techniques/methods used by different clinicians to protect the perineum 4.5 Current midwifery and obstetric education and training should include both techniques (hands on and hands off/poised) 2.2 I would prefer that the practitioner delivering (i.e. midwife or obstetrician) uses hands off/poised technique if I’m having a vaginal delivery or if my partner or daughter or other close relative is having a vaginal delivery 4.95 I would prefer that the practitioner delivering (i.e. midwife or obstetrician) uses the hands-on technique if I’m having a vaginal delivery or if my partner or daughter or other close relative is having a vaginal delivery The decision about hands on or hands off/poised 3.2 should be left to the woman’s preference 3.05 The decision about hands on or hands off/poised should be left to the practitioner’s (midwife. obstetrician) preference I believe that women are fully informed about 2.15 hands-on and hands-off/poised techniques I believe that women are fully informed about 2.3 risk of OASIS and its consequences I think that all women should be informed about 4.8 risk of OASIS antenatally A return to hands on should be recommended 4.4 practice
Median
Disagreement scores ≤3 (%)
Agreement scores ≥4 (%)
0.76
5
10
90
0.88
5
10
90
0.94
4
10
90
1.23
4
40
60
0.82
2
85
15
0.98
4
20
80
0.52
5
0
100
1.35
5
20
80
1.32
2
85
15
1.31
5
15
85
1.23
3.5
50
50
1.27
2.5
70
30
1.69
1
80
20
1.49
2
80
20
1.23
5
15
85
1.35
4
25
75
Scores 1–6: 1 = disagree strongly; 2 = disagree; 3 = slightly disagree; 4 = slightly agree; 5 = agree; 6 = agree strongly.
trauma before planning any clinical trials (85% of respondents). Indeed, 80% of participants believe that the majority of midwives and obstetricians are not confident with perineal management during the second stage of labor. There was strong agreement with six of eight items in section 2 of the questionnaire regarding the hands-on technique (Table 2). Concerning the hands-off technique,
170
there was a high level of disagreement on crucial aspects of the hands-poised/hands-off procedure. Nevertheless, 90% of respondents believe that the perineum should be assessed to decide whether to perform an episiotomy or allow tearing (Table 2; section 3). In delivering the shoulders, there was strong agreement with all items of this section (Table 2; section 4). Regarding the other intrapartum interventions there was strong agreement with
ª 2014 Nordic Federation of Societies of Obstetrics and Gynecology, Acta Obstetricia et Gynecologica Scandinavica 94 (2015) 165–174
K.M.K. Ismail et al.
Perineal support and OASIS: a Delphi survey
Figure 1. The gradual change in the levels of consensus after feedback of the results from each iteration to the participants. The results presented are from sections 1 and 2 and relate to questions 1–33. Each line represents the mean values of responses to each item in the questionnaire following each iteration.
Figure 2. The gradual change in the levels of consensus after feedback of the results from each iteration to the participants. The results presented are from sections 3–5 and relate to questions 35–58. Each line represents the mean values of responses to each item in the questionnaire following each iteration. Questions 38 and 42 are open questions and therefore no mean values are available.
five of six items in this section (Table 3; section 5). The time allowed for head to descend, the length of pushing and the position during labor have been found as the most important determinants of risk for OASIS (90%). The majority of panellists also believe that different hands-on techniques are required for women giving birth in different positions. There was unanimous agreement that there is a need to conduct a descriptive summary of details (even filmed demonstrations) of all possible techniques used by different clinicians to protect the perineum (Table 3; section 6). Of all experts, 85% would prefer the accoucheur to use the hands-on technique during a vaginal delivery of their partner or daughter and 75% of them recommend a return to hands-on in clinical practice. Finally, only 20% of panel members believed that women are fully informed about hands-on and handsoff/poised techniques, and 85% of them recognized the importance of informing women about the risk of OASIS antenatally.
Discussion There was consensus that the change from the traditional method of hands-on perineal support to the current hands-off/poised technique in the UK was not evidencebased, in agreement with previous reports (18). The reason given for this shift in practice was that its primary objective was to encourage and potentiate normality in labor and birth. Nevertheless, it was a common belief that
the introduction of the hands-off/poised approach was implicated in the increased risk for OASIS, in keeping with the suggestion from the recent hospital episode statistics report (1). There was a high level of agreement on how to apply hands-on perineal support and what the hands-poised technique involves. However, all experts agreed that the hands-poised technique is interpreted by practitioners as hands-off. There was consensus on the importance of other intrapartum interventions such as supporting the perineum while delivering the shoulders and using different hands-on techniques for women giving birth in different positions. Another issue highlighted was the belief that women are not informed antenatally about the hands-on and hands-off/poised techniques and the risk of OASIS and its consequences. There was agreement that they should therefore be provided with such information. Finally, many experts believe that the recent data from Norway showing a protective effect for perineal support are relevant to the UK (4,19,20). The consensus was that this hands-on technique should be the recommended practice. All experts agreed that an RCT study would provide the evidence for the most effective method of perineal support. If such a study was to be funded, all experts would agree to participate. On review of the literature there is evidence that the OASIS incidence is steadily increasing (1). In Norway, the OASIS rate has increased from
AOGS M A I N R E SE A RC H A R TI C LE
Perineal support and risk of obstetric anal sphincter injuries: a Delphi survey KHALED M. K. ISMAIL1, ELENA PASCHETTA2, DIMITRIOS PAPOUTSIS3 & ROBERT M. FREEMAN
4
1
Birmingham Centre for Women’s and Children’s Health, School of Clinical & Experimental Medicine, College of Medical & Dental Sciences, University of Birmingham, Birmingham, UK, 2Postgraduate School of Psychiatry, Neuroscience Department, University of Turin, Turin, Italy, 3Department of Obstetrics and Gynaecology, Royal Shrewsbury Hospital, Shropshire, and 4Plymouth Hospitals NHS Trust and Plymouth University Peninsula Schools of Medicine and Dentistry, Plymouth, UK
Key words Obstetric anal sphincter injuries, Delphi survey, hands-on, hands-poised, birth Correspondence Khaled M. K. Ismail, Obstetrics and Gynaecology, 3rd Floor, Birmingham Women’s NHS Foundation Trust, Edgbaston B15 2TG, UK. E-mail: [email protected] Conflict of interest The authors have stated explicitly that there are no conflicts of interest in connection with this article. Please cite this article as: Ismail KMK, Paschetta E, Papoutsis D, Freeman RM. Perineal support and risk of obstetric anal sphincter injuries: a Delphi survey. Acta Obstet Gynecol Scand 2015; 94: 165–174. Received: 13 July 2014 Accepted: 20 November 2014 DOI: 10.1111/aogs.12547
Abstract Objective. To explore the views of a multidisciplinary group of experts and achieve consensus on the importance of perineal support in preventing obstetric anal sphincter injuries (OASIS). Design. A three-generational Delphi survey. Setting. A UK-wide survey of experts. Population. A panel of 20 members consisting of obstetricians, midwives and urogynecologists recommended by UK professional bodies. Methods. A 58-item web-based questionnaire was sent to all participants who were asked to anonymously rate the importance of each item on a six-point Likert scale. They were asked to rate their level of agreement on statements related to hands-on/hands-poised techniques, the association of hands-poised/hands-off approach with OASIS, the need to implement perineal support and the need to improve the evidence to support it. Systematic feedback of responses from previous rounds was provided to participants. Main outcome measures. To achieve consensus on key areas related to perineal support. Results. The response rate was 100% in all three iterations. There was consensus that current UK practice regarding perineal protection was not based on robust evidence. The respondents agreed that hands-poised/hands-off and OASIS are causally related and that hands-poised was misinterpreted by clinicians as hands-off. Although 90% of experts agreed that some form of randomized trial was required and that all would be prepared to take part, there was also consensus (75%) that in the meantime, hands-on should be the recommended technique. Conclusions. Our results highlight the current lack of evidence to support policies of perineal support at time of birth and the need to address this controversial issue. NICE, National Institute for Health and Care Excellence; OASIS, obstetric anal sphincter injuries; RCT, randomized controlled trial.
Abbreviations:
Introduction Obstetric anal sphincter injuries (OASIS) are a serious complication of vaginal birth with the incidence in the UK rising from 1.8 to 5.9% in primiparous women over a 12-year period (1). Based on an average OASIS rate of 5.9% it is estimated that 48 000 of these women sustain significant anal sphincter trauma in the UK per year.
Key Message Perineal support: hands-on or hands-off at birth? The current clinical practice is not evidence-based. Consensus among experts was that a randomized controlled trial is needed (90%) and that in the meantime, hands-on should be the recommended practice (75%).
ª 2014 Nordic Federation of Societies of Obstetrics and Gynecology, Acta Obstetricia et Gynecologica Scandinavica 94 (2015) 165–174
165
Perineal support and OASIS: a Delphi survey
OASIS represents a clearly identified patient safety problem with a negative impact on quality of life and with long-term adverse health-economic effects. Obstetric anal sphincter injuries is quoted to be the most important risk factor for female anal incontinence affecting young women (2,3). Despite optimal primary repair, 30–50% of affected women suffer from a varying degree of anal incontinence and experience a significant negative impact on their quality of life (4). Several studies have reported that the risk of anal incontinence is much higher with subsequent deliveries and that symptoms get worse by time (5–7). An estimated 40% of these women will opt for a future elective cesarean section to avoid worsening of symptoms or a repeat OASIS from a vaginal birth (8,9). An American study in 1996 reported at that time that the average cost of anal incontinence management following OASIS was $17 166 per patient (10). Hospital episode statistics from a recent UK study and information from the National Institute for Health and Care Excellence (NICE) fecal incontinence costing guide for secondary OASIS repair have helped estimate a cost of £1625 per patient, an annual cost of £78 million (1,11). Moreover, allegations of negligence for perineal damage caused during labor in the UK resulted in 441 claims in 2000–2010, thus representing the fourth highest number of claims in obstetrics. The total value of those claims is estimated to be £31.2 million, with 85% of them related to misdiagnosis of perineal trauma (12). The current UK-based practice in second stage of labor involves both the “hands-on” and “hands-poised” techniques, with approximately 50% of midwives using either method (13–15). The traditional teaching has been with “hands-on” but in recent years “hands-poised” or “hands-off” has become popular, although neither has been well described (16,17). The reason for the change from the traditional “hands-on” to “hands-poised/handsoff” is unclear and without any evidence base (18). It has been suggested that this change and the reluctance to use episiotomy might have contributed to the rising incidence of OASIS (1). In contrast, implementation of “hands-on” perineal support in Norway has resulted in a 50% reduction in OASIS rates (4,19,20). As there are differing views on this subject and a poor evidence base, we conducted a Delphi consensus survey among midwives, obstetricians and urogynecologists who were recommended by their national professional bodies in the UK. Our primary aim was to explore the level of consensus among a multidisciplinary group with regard the controversial area of “handson” vs. “hands-poised/hands-off” techniques at the time of childbirth.
166
K.M.K. Ismail et al.
Material and methods The Delphi technique is a formal consensus method used in medical and health service research to obtain and integrate the views and opinions of experts in a particular field (21,22). The Delphi approach involves asking a group of experts to participate in a number of consecutive rounds of questionnaires which they rate anonymously and independently using a numerical scale. Systematic feedback from the previous iterations is provided to the participants during the consecutive stages of the process, giving them the opportunity to reconsider their original opinions and to modify or persist with them, or move towards consensus (21). The Delphi method is a multistage process which usually ranges between two and four consecutive rounds depending on the rates of responses (23–25). The level of consensus can be considered achieved if a given proportion of participants are in agreement. This proportion varies between different studies, with most of them suggesting levels between 60 and 80% (26–29). The Delphi survey technique with its cardinal features of several iterations, feedback of results, anonymity of responses and the opportunity for participants to reconsider their views, is considered an effective method for transforming individuals’ opinions into group consensus, especially when greatly diverse responses are anticipated (30). The features of the Delphi method minimize the bias arising from factors such as powerful personalities, seniority and experience, which can potentially affect the group’s opinion (31). Therefore, this method allows the participants to express their views honestly without intimidation, inhibition or peer group pressure while achieving consensus among the group (21). Some researchers in the literature report that the major limitation of this methodology is the lack of scientific validation and sufficient measurement of the reliability of the findings, thus challenging the scientific quality of the Delphi study technique (32). Nevertheless, the role of the Delphi methodology comes into play when hard evidence is lacking and the conclusions rely on the views of a panel of experts. In this case it is argued that the Delphi method should not be subject to the same validation criteria as used for other research methodologies (33,34). The panel of experts for this Delphi survey was selected for their interest, level of knowledge and clinical experience in the field of OASIS in the UK. Selection of experts was done following recommendations from senior representatives of the Royal College of Obstetricians and Gynaecologists, Royal College of Midwives and British Society of Urogynaecology. Twenty experts were involved
ª 2014 Nordic Federation of Societies of Obstetrics and Gynecology, Acta Obstetricia et Gynecologica Scandinavica 94 (2015) 165–174
K.M.K. Ismail et al.
in the Delphi survey, 10 of whom were midwives and 10 obstetricians and urogynecologists. There are reports that a Delphi expert panel requires at least 10–15 experts to obtain reliable results (35). However, a panel of 20 members is usually seen in consensusbased research. Therefore, to account for non-response we approached 22 experts, 20 of whom agreed to participate (25,36,37). All the participants were provided with an invitation e-mail letter explaining the aims, procedures and requirements of the Delphi procedure. The Delphi procedure was web-based and was completed between May and November 2013. Although webbased and conventional Delphi processes have not been formally compared, web-based Delphi surveys have been found to be feasible, cost-effective, time saving, and better accepted by users than paper-based methods (38). A webbased questionnaire was developed and a three-round electronic-mail survey was conducted. Members of the panel of experts were asked to provide any questions or statements relating to hands-on, hands-poised or handsoff techniques and OASIS that they felt were relevant. All questions and statements sent by the panel members were included in the survey. The questionnaire included 58 items grouped in six progressive sections. The first section included items related to current evidence in relation to perineal support and risk of OASIS. In sections 2 and 3 we gathered the experts’ opinions on hands-on and hands-off techniques at the time of the delivery of baby’s head. In section 4 we explored the panellists’ views on perineal support at the time of delivery of baby’s shoulders. Section 5 included key items related to other intrapartum interventions possibly associated with increased risk of OASIS. Finally, in section 6 we explored the experts’ personal preferences for hands-on/hands-poised/ hands-off techniques based on their clinical experience and the need for future recommendations and research. The questionnaire was completed online by the participants (using the Survey Monkey facility). For each iteration, an accompanying e-mail including instructions and the web-based questionnaire was sent to all participants. To ensure security and confidentiality, each participant received a personal link with the e-mail invitation allowing individual access to the questionnaire on the online survey which was specifically designed for the present study. Two electronic reminders were sent for each iteration. Participants were asked to complete the questionnaire by rating the importance of each item on a sixpoint Likert scale: 1 – strongly disagree; 2 – disagree; 3 – slightly disagree; 4 – slightly agree; 5 – agree; 6 – agree strongly. The results from the questionnaires were automatically entered into an anonymous database and analyzed by obtaining group figures from an independent researcher
Perineal support and OASIS: a Delphi survey
with no expertise in the OASIS field (EP). A summary of the results showing the distribution of the group’s responses (mean SD; median value scores) and patterns of agreement/disagreement of the previous iteration were fed back at each stage to be evaluated by the panel members. In the subsequent iterations, participants were requested to answer the same questions taking into account the mean scores calculated for each item from the group answers to the previous iteration. Participants were asked to add free comments related to clarity of items and occasionally some of the items were reworded slightly in line with the group suggestions to clarify the meaning. Agreement and disagreement consensus was set a priori and this was considered to be achieved for an item if at least 60% of all experts scored ≥4 or ≤3, respectively. This is in keeping with other studies reporting rates between 60 and 80% (26–29). In the first iteration, the questionnaire was e-mailed to the expert group with a request for members to rate their level of agreement with each item on a scale from 1 to 6. Within the questionnaire an additional text box was provided for respondents to enter any further comments and suggestions. If panel members felt that important topics were missing from the provided list, they were encouraged to suggest additional potential items. In the second and third iteration the mean (SD) and median scores of each item from the previous iteration were reported back to the group. Changes or new items suggested by experts in the first and second round were added to the Delphi questionnaire list, and this expert group was requested to complete the questionnaire again in light of the results of the previous rounds.
Results Of the initial 22 experts invited, 20 finally participated in the Delphi survey. During the process the response rate was consistently 100%, with all 20 experts participating in all three rounds. Modifications were made in the nine questions based on the comments fed back by the participants during the subsequent rounds. The final results with levels of agreement or disagreement for each section after the third iteration are presented in Tables 1–3. The gradual change in the levels of consensus, as the participants were fed back the results from each iteration, is presented in Figures 1 and 2. The only major changes in the levels of consensus from round one to round three were identified in questions 36 and 37, which involve the hands-off technique (Figure 2). In section 1 of the questionnaire (Table 1) experts agreed (85%) that the change from the traditional method of hands-on perineal support to the hands-off/
ª 2014 Nordic Federation of Societies of Obstetrics and Gynecology, Acta Obstetricia et Gynecologica Scandinavica 94 (2015) 165–174
167
Perineal support and OASIS: a Delphi survey
K.M.K. Ismail et al.
Table 1. Final results from section 1 of the questionnaire with items related to current evidence in relation to perineal support and risk of obstetric anal sphincter injuries (OASIS).
Section 1 (questions 1–25): current evidence The evidence on which UK current practice is based with regard to hands on or off/poised is robust The change from the traditional method of support (hands on) to hands off/poised was evidence-based The interpretation of the HOOP trial has changed practice Hands poised tends to be interpreted as hands off among clinicians The aim of hands off/poised is to encourage/potentiate normality The introduction of the hands off/poised approach is causally related to an increased risk of OASIS The increase incidence of OASIS is mainly due to improved assessment (PR examination) rather than hands off/poised The increase in incidence of OASIS is mainly due to increased awareness and improved recognition by the delivering clinician (acoucheur) The majority of Midwives and Obstetricians are confident with perineal management during the second stage of labor Midwives and Obstetricians in my unit have clear guidelines on whether to use hands on or hands off/poised Evidence based antenatal interventions to reduce perineal trauma are recommended to all women planning vaginal birth An RCT comparing hands on vs. hands off/poised is needed An RCT is unethical following the results of the Norwegian studies (Laine et al.) The data supporting hands on from Norway are not relevant to the UK If an RCT is funded, I would agree to participate I am in equipoise with regards to hands on vs. hands off/poised When conducting RCTs on hands on vs. hands off/poised we need to control for the speed and timing of delivery by communicating with the woman Good communication (and therefore collaboration between the woman and her attendant) is as important as what we do with the hands Good communication (and therefore collaboration between the woman and her attendant) is more important than what we do with the hands Hands on would be easy to re-introduce Training of both techniques should be facilitated and standardized by a video demonstration The technique for perineal support at instrumental delivery should be the same as normal delivery The best technique of perineal support to minimize perineal trauma has been established It is necessary to establish the best technique of perineal support to minimize perineal trauma before planning any clinical trials The practitioner should encourage the woman to deliver between contractions – minimize active pushing to prevent OASIS
Mean SD
Median
Disagreement scores ≤3 (%)
Agreement scores ≥4 (%)
2.2 0.89
2
90
10
2.35 1.13
2
85
15
0.67 0.74 0.71 1.18
5 5 5 4
5 0 5 20
95 100 95 80
3.7 0.97
4
35
65
4.3 0.86
4.5
15
85
2.6 1.18
2
80
20
2.2 1.00
2
90
10
2.15 0.87
2
95
5
4.85 0.98 2.65 1.34
5 2
10 80
90 20
2.5 5 2 5
80 0 80 0
20 100 20 100
4.85 5.15 4.75 4.15
2.75 5.15 2.7 5.2
1.11 0.67 1.17 0.52
5.05 0.75
5
0
100
3.3 1.12
3
70
30
3.75 1.25 4.8 1.15
4 5
40 10
60 90
4.75 1.20
5
15
85
2.6 1.14
2
75
25
4.65 1.03
5
15
85
3.5 1.23
4
45
55
Scores 1–6: 1 = disagree strongly; 2 = disagree; 3 = slightly disagree; 4 = slightly agree; 5 = agree; 6 = agree strongly. RCT: randomized controlled trial.
poised technique in the UK was not evidence-based. The reason given for this shift in practice was to encourage normality in labor and birth. All members strongly agreed (100%) that hands-poised tends to be interpreted as hands-off, with 80% of them agreeing that the hands-off/ poised approach is causally related to an increased risk of
168
OASIS. The majority of panellists (90%) recommended a randomized controlled trial (RCT) to compare hands-on vs. hands-off/poised in relation to OASIS; all of them (100%) agreed to participate in such a trial. However, there was consensus that it is necessary to establish the best technique of perineal support to minimize perineal
ª 2014 Nordic Federation of Societies of Obstetrics and Gynecology, Acta Obstetricia et Gynecologica Scandinavica 94 (2015) 165–174
K.M.K. Ismail et al.
Perineal support and OASIS: a Delphi survey
Table 2. Final results from sections 2–4 of the questionnaire with items related to hands-on and hands-off techniques at the time of the delivery of baby’s head and at the time of delivery of baby’s shoulders. Mean SD Section 2 (questions 26–33): hands-on technique The non-dominant, anterior, usually left hand is placed on the fetal head in order to slow down the passage through the perineal structures The non-dominant, anterior, usually left hand is placed on the fetal head in order to keep the flexion of the fetal head during the time of expulsion The non-dominant, anterior, usually left hand is placed on the fetal head in order to help the fetal head to be gradually extended during the time of expulsion About the dominant, posterior, usually right hand: the thumb and the index finger of the dominant hand are placed alongside the fourchette and vaginal opening About the dominant, posterior, usually right hand: the fingertips of the dominant hand are pressed against the perineum and a region of parietal eminences of the fetal head and pulled towards each other and towards the fourchette About the dominant, posterior, usually right hand: the palm of the dominant hand is placed against the central part of the perineum and the posterior fourchette About the dominant, posterior, usually right hand: the fingers of the dominant hand execute the Ritgen maneuver during expulsion to extend the fetal head About the dominant, posterior, usually right hand: the remaining part (apart from the thumb and the index finger) of the dominant hand executes a pressure on the fetal head to facilitate the extension Section 3 (questions 35–37): hands-off technique The hands are only applied to the head if it is delivering rapidly The hands are only applied to the head and perineum if the head is delivering rapidly Perineum should be assessed to decide whether to perform episiotomy or allow tearing Section 4 (questions 39–41): delivering the shoulders More attention should be given to supporting delivery of shoulders in preventing OASIS Adequate lateral flexion should be used to facilitate the delivery of the shoulders During the delivery of the posterior shoulder the acoucheur’s hand should be placed against the perineum to support
Median
Disagreement scores ≤3 (%)
Agreement scores ≥4 (%)
4.75 0.85
5
5
95
4.1 1.44
5
30
70
4.3 1.17
5
25
75
4.8 0.76
5
5
95
3.85 1.38
4
40
60
3.75 1.37
4
40
60
3.45 1.39
3.5
50
50
3.45 1.50
3
55
45
2.65 1.30
2
70
30
2.6 1.27
2
75
25
4.65 1.13
5
10
90
4.95 1.09
5
5
95
4.8 0.76
5
5
95
4.15 1.26
4
20
80
Scores 1–6: 1 = disagree strongly; 2 = disagree; 3 = slightly disagree; 4 = slightly agree; 5 = agree; 6 = agree strongly.
ª 2014 Nordic Federation of Societies of Obstetrics and Gynecology, Acta Obstetricia et Gynecologica Scandinavica 94 (2015) 165–174
169
Perineal support and OASIS: a Delphi survey
K.M.K. Ismail et al.
Table 3. Final results from sections 5 and 6 of the questionnaire with items related to other intrapartum interventions possibly associated with increased risk of OASIS and the need for future recommendations and research. Mean SD Section 5 (questions 43–48): other intrapartum interventions The following is important as a determinant 4.5 of risk of OASIS: time allowed for head descent The following is important as a determinant 4.6 of risk of OASIS: length of pushing The following is important as a determinant 4.4 of risk of OASIS: position in labor The following is important as a determinant 3.6 of risk of OASIS: antenatal perineal massage The following is important as a determinant 2.55 of risk of OASIS: intrapartum perineal massage Different hands-on techniques are required 4.15 for women giving birth in different positions Section 6 (questions 49–58): future needs We need to conduct a descriptive summary 5.2 of details (even filmed demonstrations) of all possible techniques/methods used by different clinicians to protect the perineum 4.5 Current midwifery and obstetric education and training should include both techniques (hands on and hands off/poised) 2.2 I would prefer that the practitioner delivering (i.e. midwife or obstetrician) uses hands off/poised technique if I’m having a vaginal delivery or if my partner or daughter or other close relative is having a vaginal delivery 4.95 I would prefer that the practitioner delivering (i.e. midwife or obstetrician) uses the hands-on technique if I’m having a vaginal delivery or if my partner or daughter or other close relative is having a vaginal delivery The decision about hands on or hands off/poised 3.2 should be left to the woman’s preference 3.05 The decision about hands on or hands off/poised should be left to the practitioner’s (midwife. obstetrician) preference I believe that women are fully informed about 2.15 hands-on and hands-off/poised techniques I believe that women are fully informed about 2.3 risk of OASIS and its consequences I think that all women should be informed about 4.8 risk of OASIS antenatally A return to hands on should be recommended 4.4 practice
Median
Disagreement scores ≤3 (%)
Agreement scores ≥4 (%)
0.76
5
10
90
0.88
5
10
90
0.94
4
10
90
1.23
4
40
60
0.82
2
85
15
0.98
4
20
80
0.52
5
0
100
1.35
5
20
80
1.32
2
85
15
1.31
5
15
85
1.23
3.5
50
50
1.27
2.5
70
30
1.69
1
80
20
1.49
2
80
20
1.23
5
15
85
1.35
4
25
75
Scores 1–6: 1 = disagree strongly; 2 = disagree; 3 = slightly disagree; 4 = slightly agree; 5 = agree; 6 = agree strongly.
trauma before planning any clinical trials (85% of respondents). Indeed, 80% of participants believe that the majority of midwives and obstetricians are not confident with perineal management during the second stage of labor. There was strong agreement with six of eight items in section 2 of the questionnaire regarding the hands-on technique (Table 2). Concerning the hands-off technique,
170
there was a high level of disagreement on crucial aspects of the hands-poised/hands-off procedure. Nevertheless, 90% of respondents believe that the perineum should be assessed to decide whether to perform an episiotomy or allow tearing (Table 2; section 3). In delivering the shoulders, there was strong agreement with all items of this section (Table 2; section 4). Regarding the other intrapartum interventions there was strong agreement with
ª 2014 Nordic Federation of Societies of Obstetrics and Gynecology, Acta Obstetricia et Gynecologica Scandinavica 94 (2015) 165–174
K.M.K. Ismail et al.
Perineal support and OASIS: a Delphi survey
Figure 1. The gradual change in the levels of consensus after feedback of the results from each iteration to the participants. The results presented are from sections 1 and 2 and relate to questions 1–33. Each line represents the mean values of responses to each item in the questionnaire following each iteration.
Figure 2. The gradual change in the levels of consensus after feedback of the results from each iteration to the participants. The results presented are from sections 3–5 and relate to questions 35–58. Each line represents the mean values of responses to each item in the questionnaire following each iteration. Questions 38 and 42 are open questions and therefore no mean values are available.
five of six items in this section (Table 3; section 5). The time allowed for head to descend, the length of pushing and the position during labor have been found as the most important determinants of risk for OASIS (90%). The majority of panellists also believe that different hands-on techniques are required for women giving birth in different positions. There was unanimous agreement that there is a need to conduct a descriptive summary of details (even filmed demonstrations) of all possible techniques used by different clinicians to protect the perineum (Table 3; section 6). Of all experts, 85% would prefer the accoucheur to use the hands-on technique during a vaginal delivery of their partner or daughter and 75% of them recommend a return to hands-on in clinical practice. Finally, only 20% of panel members believed that women are fully informed about hands-on and handsoff/poised techniques, and 85% of them recognized the importance of informing women about the risk of OASIS antenatally.
Discussion There was consensus that the change from the traditional method of hands-on perineal support to the current hands-off/poised technique in the UK was not evidencebased, in agreement with previous reports (18). The reason given for this shift in practice was that its primary objective was to encourage and potentiate normality in labor and birth. Nevertheless, it was a common belief that
the introduction of the hands-off/poised approach was implicated in the increased risk for OASIS, in keeping with the suggestion from the recent hospital episode statistics report (1). There was a high level of agreement on how to apply hands-on perineal support and what the hands-poised technique involves. However, all experts agreed that the hands-poised technique is interpreted by practitioners as hands-off. There was consensus on the importance of other intrapartum interventions such as supporting the perineum while delivering the shoulders and using different hands-on techniques for women giving birth in different positions. Another issue highlighted was the belief that women are not informed antenatally about the hands-on and hands-off/poised techniques and the risk of OASIS and its consequences. There was agreement that they should therefore be provided with such information. Finally, many experts believe that the recent data from Norway showing a protective effect for perineal support are relevant to the UK (4,19,20). The consensus was that this hands-on technique should be the recommended practice. All experts agreed that an RCT study would provide the evidence for the most effective method of perineal support. If such a study was to be funded, all experts would agree to participate. On review of the literature there is evidence that the OASIS incidence is steadily increasing (1). In Norway, the OASIS rate has increased from
