ABSTRACT

Objective: To test the reproducibility of British experts’ (eFM, K2MS, Gibb and Arulkumaran) [1-3] illustrations of fetal heart rate (FHR) decelerations by trained British Obstetricians and midwives. To analyze reasons for any discrepancies by examining factors relating to the participants, British experts’ descriptions and NICE guidelines [4]. Design: Prospective observational study. Setting: National Health Service (NHS) Hospitals. Participants: 38 Obstetric Consultants, 49 registrars and 45 midwives. Methods: Printed questionnaire. Statistical Analysis: Fisher’s Exact test. Results: This largest study of its kind showed almost unbelievably high disconnect between CTG interpretation by experts and participants. 98% - 100% midwives, 80% - 100% Registrars and 74% - 100% Consultants categorized FHR decelerations differently from the five experts’ illustrations/interpretations (p < 0.0001). Remarkably, the three experts’ illustrations of early (supposedly most benign) decelerations were classed as atypical variable by 56% Consultants, 78% Registrars and 99% midwives and the CTGs as pathological by 85% of the participants. Conclusions: The high degree of disagreement with the experts’ illustrations (p < 0.0001) did not appear to be due to participant factors. The immediate reasons seemed to be the conflicting illustrations and heterogeneity of experts’ descriptions. But most importantly, these appeared to stem from non-standardized ambiguous definitions of FHR decelerations and many intrinsic systemic flaws in the current NICE guidelines [4]. The NICE concept of “true uniform” (identical) early and late decelerations seems biologically implausible (a myth) and no examples can be found. Another myth seems to be that early and late decelerations should be gradual. Only very shallow decelerations will look “gradual” on the British CTG. These systemic flaws lead to dysfunctional CTG interpretation increasing intervention as well as impairing diagnosis of fetal hypoxemia. This is because the vast majority of FHR decelerations fall in a single heterogeneous “variable” group with many further classed as “atypical” (pathological) based on disproven and discredited criteria [5-7]. There is increasing evidence in USA that a system with variable decelerations as the majority is clinically unhelpful because of loss of information [5-9]. In the interest of patient care and safety, open debate is necessary regarding a better way forward. Classification of FHR decelerations based primarily and solely on time relationship to contractions appears more scientific and clinically useful.

1. INTRODUCTION

Electronic fetal monitoring (EFM) is one of the commonest medical procedures in the developed world. Reliable and clinically useful interpretation of intrapartum EFM is of paramount importance to prevent fetal hypoxemia/acidemia as well as to avoid unnecessary intervention. Secondly, the compensation claims for birth asphyxia account for up to 50% of the NHS litigation bill in UK [10]. The NHS in England paid out £3.1 billion for claims linked to maternity care in the past decade —the most for any specialty—mainly for cerebral palsy and errors in interpretation of cardiotocographs (CTG) [11]. Fetal heart rate (FHR) decelerations are the most common deviant feature on CTG associated with fetal hypoxemia often in the absence of FHR baseline abnormalities. Thus, FHR decelerations are the main determinants of classification into the 3 tier systems of CTG interpretation [5,12,13]. Hence, the clinical reliability and usefulness of 3 tier systems will to a large extent depend on standardized, unambiguous and scientifically robust definitions of FHR decelerations [12,14-16]. The current British definitions of decelerations were first introduced in the Royal College of Obstetricians and Gynaecologists (RCOG) guidelines of 2001 (personal correspondence with National Institute for Health and Clinical Excellence-NICE) [4,17]. In its brief guidelines on EFM, NICE has not repeated the definitions of FHR decalerations but the RCOG (2001) guidelines have been taken off-line. Because of the lack of easy access to these definitions, British Obstetricians and midwives interpret FHR decelerations based on the descriptions by a few expert groups and the rigorous mandatory training they undergo at least yearly. There seems to have been a major change in interpretation of FHR decelerations since NICE (2007) guidelines which state that “true uniform” early decalerations (head compression—a universal phenomenon in labor) are very rare and most decelerations in labor are variable (cord compression) [4], which seems counterintuitive. Also, this concept seems alien to established British practice over many decades before 2007. An opinion survey of 40 British Consultant Obstetricians revealed that “early” decelerations were the majority (more than 50%) before 2007 but this was expected to change because of fait accompli effect of NICE guidelines [13]. Another multicentre survey of British midwives showed similar finding (unpublished data). Many senior British Obstetricians and midwives have observed that CTG interpretation may have become “dysfunctional” in the last few years with many benign decelerations (previously termed “early”) being increasingly termed “variable” and then “atypical”, making CTGs pathological. To investigate this concern, a study of interpretation of FHR decelerations was undertaken. It is important to note that all classifications of FHR decelerations are not based on level 1, 2 or 3 evidence but on expert opinion only. Hence, these should stand the test of clinical observation, rationality and logical reasoning to be considered valid. This paper encourages a debate which is expected to enhance quality of care and improve patient safety in this very important area.

2. STUDY OF INTERPRETATION OF FHR DECELERATIONS

Methods

A prospective observational multicentre study was conducted to compare interpretation by Consultant Obstetricians, Registrars and midwives of five definitive CTG traces given by the three British expert groups [1-3] to explain and illustrate their descriptions of FHR decelerations in training materials, in line with NICE (2007) guidelines [4]. The monogram “Fetal monitoring in practice” by Gibb and Arulkumaran (2008) is a widely accepted British reference book [3]. K2 Fetal Monitoring Training System, Plymouth (K2MS) pioneered the web based CTG training breaking new grounds and setting very high standards [2]. Their model has been emulated by many others around the world. It quite rightly remains popular with many Obstetricians and midwives. However, K2MS had to change its illustrations and descriptions of FHR decelerations to comply with NICE (2007) guidelines. The RCOG launched eFM (2011), its official award winning on-line CTG training module incorporating NICE guidelines [1,4]. The 5 experts’ CTG illustrations included in the study are shown in Figures 1-4 except for illustration 4 namely “early FHR decelerations by eFM [1] (Chapter-CTG changes caused by hypoxia, part 1, page 13/26)”. Permission to reproduce this illustration 4 was not obtainable, hence readers are referred to the original source. Although these illustrations have been a part of training manuals for a few years, it is important to note that web-based (on-line) materials could change in future.

Illustrations of early and late decelerations were particularly chosen for this study because of the controversy regarding their occurrence after NICE guidelines of 2007 [4]. The participants in the study were recruited on a non-selective voluntary basis by personal contact (e.g. at scientific meetings) using an anonymous printed questionnaire after obtaining approval of the Research and Development (R&D) committee of the author’s institute. Although the selection could not have been completely random, this was not necessary or relevant in common with other reported studies [18-20]. Participants were not informed of the source of illustrations. Ethical approval was not required by the R&D committee because no known patient data was involved. Total of 123 participants (38 Consultants, 50 Obstetric Specialist Registrars and 45 midwives) from NHS hospitals in different regions of UK who agreed to participate on anonymous

basis were asked to interpret five CTG illustrations (Figures 1-4) based strictly on the current NICE guidelines [4]. The images of CTG illustrations were resized (enlarged) proportionately where required to match the actual size used in clinical practice. In case of short (10 minutes) CTGs, the participants were asked to assume repetition of the same pattern for 30 minutes or more required for 3 tier classification. The participants were asked to categorize FHR decelerations into “early/late/

typical variable/atypical variable/no definite decelera tions”; and to classify a 30 minute CTG pattern into normal/suspicious/pathological tier based on NICE guidelines [4]. Space was provided for additional voluntary comments. The Obstetric Registrars and midwives had undergone regular CTG training by NICE approved trainers at intervals of about every 6 months on many occasions. All participants were certified to be competent in EFM. In fact the majority of CTGs are interpreted by registrars and midwives in actual practice. The results and any

discrepancies were analyzed to explore the underlying reasons including participant factors (interobserver variation, training issues etc) and any intrinsic systemic issues with the categorization of FHR decelerations. Cochrane recommended Fisher’s Exact test was used to check the statistical significance of any differences.

3. RESULTS

Detailed results are shown in Tables 1 and 2. In this paper the experts’ illustration of different types FHR decelerations will be used synonymously as “expert interpretation”. One Registrar had to be excluded from study because of incomplete answers leaving a total of 122 participants. Results revealed unbelievably high disagreement between participants and expert groups. It was observed that 98% - 100% midwives, 80% - 100% Registrars and 74% - 100% Consultants categorized FHR decelerations differently from the expert interpretation of 5 illustrations (p < 0.0001). On the other hand midwives were very consistent in their interpretation of type of FHR decelerations with 98% - 100% agreeing with each other for illustrations 1 - 4 and 84% for illustration 5. Registrars also showed high degree of consistency with 65% - 86% mutual agreement for illustrations 1-4 and 53% for illustration 5. Consultants showed more variation with 61% - 66% agreement with each other for illustrations 2 - 4 and 40% - 42% for illustrations 1 and 5; which was still much higher than their agreement with the experts (0% - 26%).

NICE guidelines [4] are unclear about the place of early decelerations in the 3 tier classification. However, NICE does specify that early decelerations are “benign and not significant” (i.e. not a non-reassuring feature); hence the CTG could be called “normal”. Alternative opinion could be that early decelerations should be classed into the “suspicious” tier. In either case there was significant disparity between experts and study participants. The over-whelming majority of participants (98% - 100% midwives, 65% - 86% Registrars and 40% - 66% Consultants) classed illustrations 1, 2 and 4 of supposeedly very benign early decelerations as “atypical variable” and hence into pathological tier. It seems reasonable to construe that expert teaching examples of early, typical variable and late decelerations would fall in normal, suspicious and pathological tiers respectively in the 3 tier classification of NICE [4] when same pattern was presumed to persist for 30 minutes or more. However, a different assumption would not materially change the finding of major disagreement seen in Table 2. The participants showed somewhat lower congruence with each other for the 3 tier classification of the illustrations (Midwives 62% - 98%, registrars 44% - 82% and Consultants 52% - 87%) but this was still significantly higher than their agreement with expert interpretation (0% - 52%).

4. DISCUSSION There have been a few studies reporting variable interobserver interpretation of intrapartum CTG traces but generally a moderate agreement [18-20]. A multinational European study showed fair agreement among 21 obstetricians for the classification of accelerations, baseline and decelerations but poor for baseline variability [18]. Knowledge of adverse neonatal outcome lead to a more severe retrospective classification of the intrapartum CTG by five Obstetricians, derived mainly from decalerations and variability [19]. An American study of interpretation of CTG by three maternofetal specialists reported moderate interobserver agreement for categories I and II but poor for category III tracings [20]. The current study is somewhat unique and novel in that it has subjected published definitive teaching illustrations by acknowledged British experts to interpretation by a large group of trained clinicians of three different cadres. One would expect a much higher agreement (lot more than 50%) by practicing clinicians for CTG illustrations meant to demonstrate the types of FHR decelerations for training purposes. This paper does not claim that the experts’ illustrations are necessarily wrong but reports that only 0% - 26% participants agreed with the expert interpretation which is most remarkable. It could be argued that if illustrations of “atypical variable” decelerations were chosen for the study, the agreement would have been quite high; but that would not be surprising given that the vast majority of decelerations are classed as variable (and then many atypical) by the clinicians anyway based

on the NICE guidelines [4]. Thus the validity of the current system of categorization of decelerations is best examined by focusing on early and late decelerations. It is clearly very vital to have a debate on the reasons for the remarkable finding of this study to improve the patient safety and quality of care. The paper presents some analytical explanation and remedies but other explanations cannot be ruled out and are open to debate and the readers can draw their own conclusions.

The results were analyzed to look for “participant factors”. The discrepancy between experts and participants did not appear to be “a (inadequate or inconsistent) training issue” because it applied to Consultants as well and there did not seem to be a regional variation. Moreover the Registrars, midwives and Consultants were intensively trained at least once a year on many occasions and certified to be competent. They were recruited from different regions of UK and had been following the same standard guidelines for 5 years or so. Moreover, there was very good agreement amongst the participants themselves. Participants varied a bit more for the 3 tier classification of the illustrations which may have been because of quite complex and confusing rules about significance of frequency, proportion and duration of subtypes of variable decelerations. Importantly the interpretation by the participants could be explained by logical reasoning. In the absence of “participant factors” it became necessary to critically look for any intrinsic reasons within the experts’ illustrations themselves. It also seems simplistic to assume that the reason for high disagreement may simply be an oversight or inadvertent lack of attention to detail by the expert groups. The following in depth analysis suggests that any inaccuracies in experts’ illustrations may be “enforced errors” stemming from more fundamental systemic flaws in the current categorization/definitions of FHR decelerations in the NICE guidelines [4]. These flaws seem to make it almost impossible to provide valid examples early and late decelerations and manifest as substantial heterogeneity of expert description particularly inconsistent emphasis on debatable attributes like “uniform”, “repetitive” and rapid/gradual/ bell shape etc.

4.1. Bias, Confounding, Other Reasons for Caution and Generalisability

Based on the additional comments, midwives and Obstetric Registrars appeared to rigorously follow teaching of NICE guidelines. However, Consultants often use more flexibility and discretion in their practice based on their previous experience (put old hat on) and personal views. However, all participants were asked to interpret illustrations based strictly on NICE guidelines only [4]. Very high (98% - 100%) disagreement by midwives may have been because of preponderance of hospitals where midwives were taught and instructed not to class any FHR decelerations as “early” or “late” because no decelerations are truly uniform, a strange notion indeed. The results of the 3 tier classification are more subjective. Since, NICE guidelines have been universally adopted in the UK, the results of this study would be very highly generalisable. Moreover, analytical conclusions of this study would be applicable to other national guidelines with a possibility of achieving meaningful uniformity of EFM guidelines in North America, Europe and Australasia, critical for future research and progress in this very important field.

4.2. British (RCOG, 2001; NICE, 2007) Definitions of FHR Decelerations [4,17]

The importance of standardized and unambiguous definitions has been well documented [14-16]. The British definitions were described in RCOG guidelines of 2001 [17] and were adopted by NICE in 2007 [4]; but are not accessible on-line at present.

Early decelerations: “Uniform, repetitive, periodic slowing of FHR with onset early in the contraction and return to baseline at the end of the contraction”.

Late decelerations: “Uniform, repetitive, periodic slowing of FHR with onset mid to end of the contraction and nadir more than 20 seconds after the peak of the contraction and ending after the contraction.

Variable decelerations: “Variable, intermittent periodic slowing of FHR with rapid onset and recovery. Time relationships with contraction cycle are variable and they may occur in isolation. Sometimes they resemble other types of deceleration patterns in timing and shape.”

It needs to be debated whether RCOG/NICE definitions (adopted from description by Edward Hon [21], personal correspondence with NICE) are unambiguous. Hon et al (1968) described early and late decelerations to have a “uniform pattern” of a specific time relationship to contraction cycle [21]. However, quoted without proper context, the term “uniform” in RCOG definition has been misinterpreted as identical i.e. “uniform in duration and depth and shape” (a biological implausibility) even by expert groups like eFM of RCOG and NICE [1,4]. The first sentence in the RCOG definition of “variable” decelerations specifies rapid onset (descent) hence this may be construed as the main characteristic. By corollary “early” and “late” decelerations may be required to have “gradual” descent but this has not been specified in their definitions. The second sentence in the definition of “variable” definition corresponds to traditional British definition [22-25] in that the time relationships to contraction cycle must be variable. The last sentence in the definition of variable decelerations appears an “abstraction” because not only it contradicts the first two sentences, but based on it any decelerations could be justifiably called “variable”!

4.3. Heterogeneity of British Expert and NICE Descriptions of FHR Decelerations

NICE (2007) categorically states that early decelerations are “truly uniform” and there is a mention of rapid descent for variable decelerations in RCOG definitions [4,17]. It is difficult to imagine any other meaning of “truly uniform” apart from “same in depth and duration i.e. identical”. RCOG experts (eFM, CTG changes caused by hypoxia, part 1, pages 13-14/26) have indeed confirmed this by specifying that early and late decelerations must be uniform in both length and depth, but without mentioning gradual shape [1]. Gibb and Arulkumaran (2008) described of “early” decelerations to be “bell” (gradual) shaped without mentioning “uniform” shape [3]. K2MS experts (Chapter-Decelerations, page 14/21) state “Early decelerations are uniform in shape, have rounded nadir (U shaped) and have symmetry with contractions”. But it is unclear what K2MS experts mean by uniform. Thus it can be surmised that the three British expert and NICE/RCOG descriptions of different FHR decelerations vary significantly. None appear standardized, comprehensive or unambiguous.

Table 1 shows that 77% Consultants, 87% SpRs and 100% midwives categorized illustration 1 of early decalerations by Gibb and Arulkumaran [3] as “variable” because of different sizes and deep “V” shape. 75% participants further classed majority decelerations as “atypical” because of lack of shoulders, possibly making the CTG “pathological”. These decelerations do not fulfil “bell or gradual” shaped description by Gibb and Arulkumaran themselves. However, they would be classed as early based on traditional British practice before 2007 [22-25] (Figure 5). Thus a definitive example of most benign FHR decelerations (early) was classed as “pathological” by 66% of Obstetricians and midwives (Table 2) based on their current mandatory CTG training. It appears that in order to give example of hypothetical and elusive ‘U’ shaped (gradual) early decelerations, K2MS experts may have inadvertently shown somewhat shallow decelerations with late time relationship to contractions in illustration 2. The vast majority of participants (89%) categorized illustration 3 decelerations as “atypical variable” and CTG as “pathological” (Tables 1-2). The reason quite correctly seems to be absence of shoulders (initial small accelerations). It is relatively common for shoulders to be absent. The positive predictability of “atypical” features has never been tested in a British study. RCOG adopted the definitions of atypical variable decelerations from American authors Krebs et al who reported that the “loss of initial acceleration” was the commonest atypical feature but was also found to be quite benign [26]. The only good quality evidence (level IIA) available subsequently has refuted and invalidated the significance of “atypical features”. An important study of 5388 deliveries by Cahill et al. (2012) reported no correlation between “atypical features” and incidence of neonatal acidemia [6]. Hamilton et al. [7] analyzed digital EFM tracings of 3695 patients specifically testing the 6 atypical features described by the RCOG. They found that only one feature namely “loss of variability during a