Case 1

A 23-year-old primiparous female is admitted in active labour at 39 weeks with 4 cm cervical dilatation after an uncomplicated pregnancy. The patient makes normal progress in cervical dilatation, but the second stage is prolonged and is augmented by an oxytocin infusion. Delivery is ultimately by vacuum extraction (VE) of a 4600-g boy and the placenta delivered, intact, within 5 minutes. Five units of oxytocin are given IM at delivery and an IV oxytocin infusion (40 units in 500 ml at 125 ml/h) is commenced. Twenty minutes after delivery the midwife notes that the patient has passed several large clots per vaginam; on abdominal examination the uterus is soft and at the level of the umbilicus. Clinical observations show a pulse of 90 and a blood pressure of 120/80.

SA Level 1: Perception

On entering the delivery room, perception of the scenario starts with scanning the room for visual cues. How does the patient look? What is connected to the patient? Who else is here? As time passes, further information can be gathered by focusing attention elsewhere. The patient’s physiological condition can be assessed by looking at the patient observations on an electronic monitor or on the patient observation chart, and also by formally physically examining the patient. The patient’s clinical history is formed by talking to the patient and other clinicians responsible for the patient’s care – for example, the midwife or accoucheur – and by looking at the clinical notes.

SA Level 2: Comprehension

Level 2 situational awareness encompasses the clinician’s comprehension and understanding of the perceived data gathered. This level goes beyond just being ‘aware’ of various elements of a situation: these disparate and sometimes patchy data are integrated with medical knowledge from long-term memory; mental models of maternal physiology and how this changes during childbirth; pharmacology knowledge; obstetric guidelines and training. Perhaps most importantly, these data are compared with previous experience of similar situations to find a mental ‘schema’ that fits the emerging situation (Schulz et al., 2013).

In Case 1, the clinical history (prolonged second stage, augmentation with oxytocin, vacuum extraction of a large baby) fits with the model of risk factors for postpartum haemorrhage. This mental model also fits with the patient examination of a soft, atonic uterus and passing large clots per vaginum.

SA Level 3: Projection

In Level 3 the clinician projects to the future expected progress of the patient’s condition. This is crucial to making early, proactive decisions about future investigations and treatments and the allocation of resources to achieving these therapeutic goals.

In this case, the clinician recognises that obstetric haemorrhage is a common (13% of maternities) and life-threatening emergency (10% of maternal deaths; Knight et al., 2014). She also recognises that a young woman will compensate well for ongoing blood loss and is not therefore falsely reassured by the reportedly normal clinical observations. This understanding allows her to plan further immediate treatment (additional uterotonic drugs) and to prepare herself and the clinical team for the probability of a developing major obstetric haemorrhage.

Case 1 Continued

Despite administration of ergometrine and carboprost the patient continues to bleed heavily per vaginum. Pelvic examination by the attending obstetric speciality trainee reveals bleeding from cervical os with no tears in cervix or vagina and a ‘boggy’ uterus; clots are expressed per vaginum. Clinical observations, taken by the support worker, now show a BP 90/60, pulse 130.

The developing team situational awareness of a major obstetric haemorrhage is built from the observations of the midwife, the clinical support worker and the attending clinician. Not every member of the team has to understand every piece of information, but all need a shared understanding of the ‘bigger picture’, to avoid becoming the chain’s proverbial weakest link (Endsley, 1995).

The team caring for the patient could share this understanding by declaring a ‘major obstetric haemorrhage’; this communicates the shared diagnosis within the team, so that they are now all acting towards a stated common goal, and it also acts to share this information with the wider organisation. It also allows a shared projection about the patient’s further course.

Distributed Situational Awareness

A further refinement of situational awareness encompasses not only people (individuals and teams) but also the equipment they use and the environment they work in. The complex, evolving interplay between people, environment and processes, where situational awareness is an emergent property of the whole system, has been termed ‘distributed situational awareness’ (Fioratou et al., 2010; Schulz et al., 2013). In this chapter we argue that when considering interventions to improve team performance, or incorporating new guidance or technology into our practice, it can help to explore how these impacts will play out at a system level.

Case 2

The on-call anaesthetist has given a general anaesthetic to facilitate an urgent caesarean section in the operating theatre. Shortly after delivery the patient becomes tachycardic and hypotensive; the airway pressures needed to ventilate the patient increase. The obstetrician observes that the uterus is soft and it is noted that there is over 800 ml of blood in the theatre suction system. The obstetrician recommends an oxytocin infusion to treat a presumptive atonic postpartum haemorrhage and while connecting the oxytocin infusion, the anaesthetist notices redness around the intravenous cannula site on the forearm. After pulling back the surgical drapes, a widespread urticarial rash is noted over the patient’s chest.

SA Level 1 Errors

The most common type of situational awareness error is at Level 1, with a failure to perceive relevant information (Endsley, 2006). This type of error can be caused by barriers to detectability; for example, a physical obstruction to sight or auditory masking from background noise. In this case the surgical drape is a visual barrier to detecting the developing urticarial rash. Another example is the misperception and misreading of the name of a medication (e.g. metronidazole for metformin), when the presentation and packaging of the drugs are similar.

A key factor in situational awareness is allocation of attention. In errors of this kind, the necessary information is all present, but the attention of the clinician is inappropriately focused on only part of it: for example, the accoucheur attends to the examination of the cervical dilatation and fetal head position, but neglects to notice the gradually deteriorating ‘early warning’ scores on the patient observation chart.

SA Level 2 Errors

Errors can also be made in the interpretation and understanding of the available cues and how they pertain to the clinical goals. A practitioner may lack the mental model to properly integrate all the available data or fail to recognise prototypical situations; they may also apply the wrong mental model to a given scenario.

In Case 2, the available information could fit with the typical picture of a haemorrhage, anaphylactic shock, an amniotic fluid or pulmonary embolus. A less-experienced clinician, through lack of experience or training, may not have all of these models, especially that of amniotic fluid embolus, for immediate recall. In the case above, the clinicians select the wrong mental model (haemorrhage) and there is a danger of then interpreting evolving signs as fitting with this existing diagnosis (a confirmation bias) rather than considering alternative diagnoses, such as anaphylactic shock.

SA Level 3 Errors

In errors at this level, there is a failure to predict future states completely or accurately, even though the existing situation is correctly understood. In Case 2, this could present as a comprehension that the patient has anaphylactic shock, but there is a lack of appreciation that this is an imminently life-threatening emergency that needs immediate, specific treatment (steroids, antihistamines and subcutaneous or low-dose intravenous adrenaline). Projection of different future states, depending on the action taken by the team, is crucial in clinical decision-making.

Decision-Making

Once SA is built, the clinician can begin to make decisions about diagnosis and appropriate treatments, or the course of action to be taken. This section examines first, analytical and second, more intuitive, ‘real-life’ cognitive models of decision-making; last, we examine the role of the patient in decision-making and methods of integrating these approaches.

Case 3

In clinical practice, a variety of cognitive and emotional factors and subjective judgements come into play in making decisions. In the following case, based on real events, these factors have deliberately been elaborated.

It has been a busy day on the labour ward and you are tired. At the lunchtime meeting there was an interesting case presentation on the expectant management of placental abruption; this afternoon, the labour ward is busy with six women in active labour. You are asked to see a 28-year-old Caucasian woman, J, because her midwife is concerned.

J is at 28 weeks gestation in her third pregnancy, having had a caesarean section with her first delivery and a successful VBAC with her second. The CTG shows a persistent bradycardia and the contractions, which were frequent, have now become constant abdominal pain. The midwife notes that the fetal head, which had been well engaged, has now moved up and is palpable abdominally.

Analytical Decision-Making

Evidence-based medicine advocates often promote a Bayesian probabilistic model to making diagnostic decisions. In principle, a constellation of clinical findings should generate a list of differential diagnoses which can then be ordered in terms of the probability of their occurrence. As treatment progresses, the results of further diagnostic tests or therapeutic interventions can be used to modify these probabilities and narrow the list of possible diagnoses, until the correct diagnosis or decision is made.

A Bayesian approach to Case 3 would be to compare the baseline risk of placental abruption (∼1:200, Crowther et al., 2012) with that of a ruptured uterus (∼1:500, Fitzpatrick et al., 2012) and then look at features, such as the persistent bradycardia and change of fetal station, which tend towards one diagnostic path rather than the other.

This approach suffers from assuming complete probabilistic knowledge of all potential choices and supposing that the pay-offs of each decision are known. In real-life decision-making information may be incomplete, the pay-offs can be unclear, patients and physicians may vary their ordering of preferred outcomes and trade-offs and compromises may be necessary.