See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/216827443

Clinical decision making

Book · June 2009

CITATIONS READS

9 7,249

3 authors, including:

Margaret Sque Max Chipulu

University of Wolverhampton University of Southampton
113 PUBLICATIONS   1,564 CITATIONS    92 PUBLICATIONS   1,185 CITATIONS   

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

PAPERLESS NHS View project

Special Issue @ PMJ: Career Paths and Career Systems for Project Managers View project

All content following this page was uploaded by Max Chipulu on 27 May 2014.

The user has requested enhancement of the downloaded file.

 CHAPTER 12

Clinical decision making

Magi Sque, Max Chipulu and Debra McGonigle

CONTENTS
➔ Introduction ➔ The nature of decision analysis
➔ Defining clinical decision making ⎯ Stages in decision analysis
⎯ Rules of thumb, heuristics and ⎯ Defining the problem
bias ⎯ Rational models
⎯ Intuition ⎯ Certainty, uncertainty and
⎯ Assessment probability
⎯ Judgements ⎯ The decision tree
⎯ Social judgement theory ⎯ Clinical decision support
⎯ Cognitive continuum theory systems
➔ Summary
➔ Key points for practice

INTRODUCTION
Decision making is at the heart of clinical encounters or operational decisions (deci-
sions that generate action). It therefore follows that the better our decisions, the more
successful and effective our clinical practice will be. Sound decision making and
operational decisions depend upon fully informed assessments and astute, analytic
judgements. Clinicians should be concerned about their decision making as a means
of improving accuracy and agreement among health professionals, and of ensuring
professional accountability. There is also the need to be aware of and recognise the
types of heuristics and biases1–4 that can influence judgement choices and decision
making.2,4,5
Clinical decisions span at least three areas of practice as clinicians make decisions
about diagnosis, monitoring and interventions. Diagnoses are determined for the
purpose they are required, for example identification of a state or condition, its tra-
jectory, the pattern of distress and severity, its treatment or prognosis. Monitoring is

235
236 ● A FOUNDATION FOR NEONATAL CARE

largely concerned with detecting changes in the practice situation and tends to ask
the questions: Is the patient getting better or worse? Is an intervention working? Is
there cause for revision of particular therapy? Are further diagnostic investigations
appropriate? Decisions about interventions underpin care, treatment and referral
options.
The aim of this chapter is to encourage reflection on the underpinning theories
of clinical decision making so that you will be able to acknowledge the processes that
influence the ways in which decisions are made. The chapter starts with a discussion
about the important elements that underpin clinical decisions, highlighting the sig-
nificance of the assessment process, judgements, information processing, heuristics
and intuition. It moves to a discussion covering humanistic intuitive discourses,
Brunswik’s Lens Model,6,7 social judgement theory6,8,9 and the cognitive continuum.6,7
The final part of the chapter addresses decision analysis, rational theories of decision
making, risk, uncertainty, and probability and decision trees.
Reflection on decisions made in clinical practice is important if we are to learn
from the experience and look to improving the ways decisions are made. There is
a need for everyone to be able to provide a clear rationale for the decisions made
in terms of professional accountability and also to ensure that, whenever possible,
decisions are based on the best available evidence. To demonstrate application of
some of the more theoretical elements of clinical decision making, a decision from
clinical practice has been used.
It is thought that decisions made in clinical practice are made ‘on-the-spot’ often
without much time, and the case used here (see Box 12.1) is a good example of this.
However, retrospective reflection on the decision clearly demonstrates how quickly
the many elements (including personal experience, knowledge, assessment, cues)
underpinning the decision are subconsciously assimilated and synthesised before a
reasoned decision is made.

DEFINING CLINICAL DECISION MAKING

Clinical decision making is a cognitive process concerned with problem recognition
through the identification of cues or of relevant clinical features, data gathering,
assimilation, analysis, evaluation and choice, to produce an operational decision. Data
gathering, assimilation and analysis are part of the assessment process. Evaluation
and choice are part of the judgement process. Clinical judgement is the outcome of
the assessment. Clinical judgement uses the skills of critical analysis, through which
the clinician makes an evaluation of the status and quality of the presenting phenom-
enon or condition and forms an opinion or conclusion (inference) about what is (or
will be) needed for an operational decision; it is an evaluation and choice between
alternative courses of action.
Judgements can be:
◗ descriptive – attributes observed directly or from other sources
(‘the jaundiced baby was as orange as orange juice’)
◗ evaluative – expresses a qualitative difference
(‘the baby’s temperature is normal today after the pyrexia recorded yesterday’)
◗ causal – attributes that explain a problem
(‘the baby’s weight gain can be attributed to increased feeding’)
 CLINICAL DECISION MAKING ● 237

BOX 12.1 THE CLINICAL DECISION

Edward was born in a district general hospital at 36 weeks’ gestation, weighing 2.6
kilograms. He was transferred to the regional neonatal surgical centre with a dia-
gnosis of oesophageal atresia and a tracheo-oesophageal fistula. Surgery to correct
these anomalies took place the following afternoon and Edward returned from theatre
ventilated via an endotracheal tube (ETT), and sedated with 20 µg/kg/h of morphine.
Operation notes indicated that the anastomosis was under some tension. Edward
was nursed on the neonatal unit in the intensive care room where four nurses were
caring for seven other babies.
The Senior Sister in charge on the night of Edward’s operation was made aware
of difficulties relating to his airway management, when asked to prepare drugs for
his reintubation. The unit was full, with 23 patients. Medical cover consisted of an
on-site Senior House Officer who was attending a delivery one floor below, a Neonatal
Registrar who had just finished a period of supervised practice because of concerns
relating to her competency, and an on-call Consultant who was a 15-minute drive
away. Edward’s allocated nurse, although experienced in caring for neonates, had
limited surgical experience.
Edward was bradycardic (heart rate of 60 beats per minute) and oxygen saturation
was 75% with no chest movement despite the Registrar giving ambubag ventilation
via his endotracheal tube. The Sister checked with the nurse caring for Edward that
she had checked his ETT for secretions and according to the nurse the tube was
patent. Rather than prepare reintubation drugs and remove the ETT, the Sister decided
to instil 0.5 mL of 0.9% sodium chloride into the ETT prior to performing ETT suction
again to clear any possible obstruction. She had major concerns relating to reintuba-
tion in terms of its effect on the anastomosis site and the ability of the Registrar to
perform the procedure.

◗ inference – not based on facts gathered from the patient

(‘if there is an increase in preterm deliveries then admissions to the neonatal
unit may be higher’)
◗ predictive – what might happen to the patient
(‘if the term ventilated baby remains poorly sedated, he may develop a
pneumothorax’).

An operational decision is what you do or do not do as a consequence of your judge-

ment. Operational decisions are made in the course of managing and delivering care.
They are the reasons for assessment and could be described as the assessment goal.
An operational decision is about acting on a judgement choice. Operational deci-
sions are concerned with action.

Rules of thumb, heuristics and bias

When making judgements and, consequently, decisions, we need to access the vast
amount of knowledge and information in our memories. To achieve this, heuristics
or ‘rules of thumb’ are frequently used as short cuts to this information.1–4 However,
238 ● A FOUNDATION FOR NEONATAL CARE

the inappropriate use of such selective information can be problematic. Heuristics

have been intensively and extensively studied in psychology, sociology, economics
and computer science.1,10 The work in psychology of Kahneman and Tversky11 is
notable in this area. Heuristics may lead to a quick, feasible solution but may also
cause system bias so that the way in which information is processed is imperfect. The
most common of these heuristics, availability, representativeness and anchoring, and
adjustment, are described by Cioffi.12
The availability heuristic tends to use information closest to hand. Evidence dem-
onstrates that the way individuals judge events is based on the evidence available to
them. For this reason, events that can be easily visualised or recalled may be judged
to be more likely than they are. Individuals also appear to judge events based on
how ‘typical’ of the class of the event they are judging the event in question appears,
regardless of other evidence outside their own experience of ‘typical’.
The representative heuristic involves reasoning influenced by prior experiences
of similar events which may lead to over-reliance on certain evidence while other
facts are ignored. As a consequence of using selective heuristics, people, including
clinicians, sometimes make substantial decision errors.13
Anchoring and adjustment heuristics draw on anchors, usually in the form of
cognitive reference points;9,12 for example, a poorly perfused neonate like Edward
could be expected to have a bluish complexion, indicative of cyanosis.

Intuition
In clinical decision making some health professionals have repeatedly stated reliance
on intuitive judgement.14 However, there appears to be endless debate about what
intuition is and whether it deserves recognition and legitimacy in praxis. Harbison15
describes a logic underlying intuitive judgement but it has been explained by others
as an inability to articulate the decision-making process. Hams14 explains that a
deeply grounded knowledge base, developed through critical thought, helps experi-
enced practitioners to practise intuitively. Therefore each clinical experience becomes
a lesson, which informs the next one. Buckingham and Adams16 discuss how cues
can stimulate recall and describe the term ‘pattern recognition’ as the linking of cues
with diagnostic categories. Intuition is seen by some as the skill of pattern recogni-
tion involving the rapid, unconscious processing of cues; giving rise to an inability
to articulate the decision-making process. Gobet and Chassy17 describe the work
done by Benner18 in the 1980s as being ‘too simple’ to account for the complexities of
expert intuition. They explain how intuitive decision making is linked to individual
perception and analytic and conscious problem solving at the expert level aided by
chunks of related information19 and complex data structures or templates20 that are
associated with long-term memory information.
McKinnon,21 on the other hand, argues that stored emotional interpretation of
clinical phenomena and its interface with reasoning, via neural pathways, creates
intuitive responses by the expert practitioner. He stresses the importance of the
link between cognition and emotion to explain intuitive thought and response.
McKinnon21 claims that: ‘Emotion stands at the helm of memory guidance, con-
solidation, storage and retrieval. Cognition is emotion gated’ (p. 42). Emotion is
therefore suggested to be the precursor to thought and response; that is, we feel before
 CLINICAL DECISION MAKING ● 239

we think.22 It would appear, therefore, that intuition is much more than a ‘gut feeling’.
It appears to have a deeply grounded knowledge base but because of its often very
rapid and unconscious nature, this becomes difficult to articulate. Later we will see
how intuition is used in practice.

Assessment
Assessment is the first stage of the decision-making process and it is used to ‘build a
picture’ of the patient’s situation and it is fundamental and crucial to accurate clini-
cal judgements and reliable operational decisions. Key data-gathering strategies of
the assessment process are: listening, asking, observing, doing, filtering and finally
synthesising and interpreting. More specifically, data gathering will include exploring
the history of the clinical event, which may involve asking questions, both general and
focused, of individuals who are able to give an accurate account of events. This may
be the patient or, in the case of the neonate, the parent or relevant health professional.
A thorough examination of the patient through observation, using look, feel and
move strategies, and both visual and non-visual cues (see Box 12.2) and investigations
to help confirm or refute suspicions will complete the data-gathering process.
Awareness of the need for assessment in Edward’s case was brought about by the
request from clinical staff for the Sister to prepare medication for reintubation. One
of the medications, a muscle relaxant, would require the use of bag-and-mask ventila-
tion and replacing Edward’s endotracheal tube would require extension of his neck.

BOX 12.2 THE VISUAL AND NON-VISUAL CUES USED

DURING THE ASSESSMENT PROCESS OF EDWARD

Visual cues

● Mechanical ventilation via an endotracheal tube

● Heart rate 60 BPM
● No chest movement
● Cyanotic
● Sedated on a continuous morphine infusion
● Surgical wound
● Night time – reduced personnel and services

Non-visual cues

● Post-operative oesophageal atresia and tracheo-oesophageal repair

● Recent anaesthetic
● Anastomotic tension
● Bag-and-mask ventilation to be avoided if at all possible
● Senior House Officer attending a delivery elsewhere
● Inexperienced Registrar – recent period of supervised practice
● On-call Consultant off site – approximately 15-minute drive away
● Relatively inexperienced nurse caring for Edward
240 ● A FOUNDATION FOR NEONATAL CARE

Both of these actions were perceived by the Sister as having a detrimental effect on
the anastomosis site. The first step in the Sister’s decision making and a fundamental
part of the assessment process was that of cue acquisition. In a very brief period of
time, both visual and non-visual cues (Box 12.2) led the Sister to an assessment of
Edward’s situation and the need for a course of action or an operational decision.
Building a picture of the patient situation is the cognitive processing element of
assessment. It involves matching past experience with the present case, which could
be flawed.9 The literature suggests that there are significant inadequacies in the
assessment process caused by inadequate data search and early closure of the consul-
tation.2 Assessment is not only about data gathering. Knowledge is fundamental to
the ability to accurately interpret and synthesise data. Table 12.1 identifies six broad
types of knowledge that impact on the assessment process and consequent decision
making.23–25 The table also demonstrates how these types of knowledge were appli-
cable to the Sister’s decision making in Edward’s case.

TABLE 12.1 Types of knowledge applied in Edward’s case

Type of Definition Sources and relevance Application to the
knowledge decision made
Empirical23 Scientific, technical or Literature, formal Knowledge used in
factual knowledge that education, research or relation to neonates,
underpins clinical practice clinical audit. oesophageal atresia,
and may take the form of Relevance to health surgical treatment,
science, learning theories, assessment, physical complications, and
organisation of facts, examination, diagnosis, ventilation obtained from
describing and explaining prognosis, treatment. specialist courses, study
phenomena. days, journals.
Ethical23 Moral experience gained Experience, custom and Principles of rights
through life, knowing passing on of values and responsibilities,
what is right and wrong. and beliefs, through beneficence and non-
observation and teaching. maleficence applied; to
Moral framework that respect Edward’s right
underpins clinical action. to life, to act responsibly
to do positive good for
Edward and to do him no
harm.
Aesthetic23 Aesthetic knowledge Gained through Neonatal surgical
appears to have no experience and the experience over many
rational or logical individual characteristics years leading to ‘intuitive’
explanation. It is of the person such as reasoning.
individualised through their values and beliefs.
experience and is Decisions are made
regarded as ‘intuitive’. rapidly without apparent
rationale or logical
reasoning. Commonly
part of the repertoire of
the expert practitioner.
(continued)
 CLINICAL DECISION MAKING ● 241

Type of Definition Sources and relevance Application to the

knowledge decision made
Personal23 Self-knowledge, knowing Experience, learning from Aware of own personal
oneself, individual self and others. limitations (strengths and
perceptions, feelings, Awareness of one’s own weaknesses).
prejudices or biases. perceptions, feelings, Acknowledges bias
prejudices or biases. towards Registrar
because of recent
concerns about capability.
Socio- Social, economic, political Awareness, experience. Clinical governance
political24 considerations relevant How the individual patientissues: risk and quality.
for a particular society. fits within the bigger Economic: issue of
It is about the bigger picture. increased financial and
context than an individual personal cost related
patient. to any post-operative
complications.
Woman’s25 Feminist perspective Women’s life experiences. Not knowingly aware
over science as male Women’s socialisation of its application in this
dominated. Argues that into the role of health decision.
science and technology professional.
undermine feminist
knowledge, which
involves women’s
life instincts and
the humanness of
relationships.

Of particular importance to the Sister’s decision were empirical, ethical, personal

and aesthetic knowledge. The confidence that comes with knowledge, experience
and knowing one’s personal limitations led the Sister directly and quickly to the
decision. Equally important was personal knowledge of the Registrar in relation to
clinical capability, the absence of the Senior House Officer and the time factor in
summoning the on-call Consultant. However, it appears the most important type
of knowledge used was aesthetic, in the form of ‘intuition’. The need for immediate
action to re-establish ventilation and consequently improve Edward’s heart rate and
colour meant the decision-making process advanced rapidly over a period of 10–15
seconds. The large numbers of simultaneous cues were used intuitively to make the
decision not to remove the ETT but to ensure patency by instilling 0.5 mL of 0.9%
sodium chloride and applying suction.
Underpinning all clinical decision making, as in the Sister’s operational decision,
is ethical knowledge. Making ethical clinical decisions is facilitated through cogni-
tion and objectivity, which is more than merely following rules or acting intuitively.
Applying ethical knowledge is about positively deciding the correct action to take in a
given situation that will be in the best interest of the patient. This requires the ability
to facilitate dialogue based on options and optimal solutions, the ability to anticipate
circumstances and recognise the subtleties of moral issues. Clearly underpinning
242 ● A FOUNDATION FOR NEONATAL CARE

the Sister’s decision was Edward’s right to life and her professional responsibility to
exert beneficence; to take action that would potentially have the safest positive out-
come for Edward, i.e. ensuring patency of the endotracheal tube, whilst upholding
the principle of non-maleficence ‘to do no harm’ through not causing a potential
rupture of the anastomosis and further post-operative complications through the
act of reintubation.
Two other important domains that impact on the assessment and decision-making
process are domain-specific knowledge and prior knowledge of the patient and their
circumstances.26 Clearly a neonatal practitioner will have specific domain know-
ledge that will assist them in making appropriate decisions about their tiny clients.
Clinical staff will also develop a familiarity with the particular baby and their indi-
vidual responses so they are better able to make judgements about clinical signs and
the illness trajectory. Specific knowledge and expectations about a neonate’s family
situation that might affect decisions about support issues such as discharge planning,
are also very important.
To access this knowledge from memory, we must apply heuristics, pattern recogni-
tion or rules. A combination of cues may bring hypotheses to mind and they can be
easily retrieved from memory. Therefore, to make sense of cue gathering to inform
the reasoning process, hypotheses generation and consequent judgements, cues need
to be sorted and ‘chunked’ or coalesced.19 Box 12.3 shows the sorting and coalescing
of Edward’s assessment cues.

Judgements
It could be argued that, in making the decision not to remove Edward’s ETT, the Sister
used both anchoring and representative heuristics. The cognitive reference points of
low heart rate, cyanosis and no chest movement immediately led the Sister to the
judgement that the endotracheal tube was blocked, whilst there was recognition of a
similarity between Edward’s situation and previous experience of caring for babies
with oesophageal atresia. Whilst heuristics are perceived as an aid to reducing uncer-
tainty, they can also be the cause of poor decision making and bias.1–4,9 Biases such as
ignoring base rates (prevalence), hindsight or stereotyping may occur for numerous
reasons including over-confidence, time pressures or one’s own values and beliefs.2
On reflection, the Sister’s bias in relation to the ability of the Registrar to reintubate
Edward carried some weight in the decision.
The information processing theory is perhaps the most influential theory which
describes how decisions are made. Bohinc and Gradisar27 state that its basis is in
studies on human problem solving, but in healthcare it is defined as a process of
hypothetical-deductive reasoning. Thompson and Dowding9 cite Elstein et al.’s28
hypothetico-deductive model, listing the four-stage process as: cue acquisition,
hypothesis generation, cue interpretation and hypothesis evaluation. Thompson
and Dowding9 discuss the generation of between four and six initial and tentative
hypotheses as a norm. They argue, however, that these may be flawed, as human rea-
soning is ‘bounded’ by the capacity of the human memory. There is also the danger
of being too focused and missing potential outcomes. Table 12.2 shows the tentative
hypotheses generated by the Sister from the coalescence of cues.
 CLINICAL DECISION MAKING ● 243

BOX 12.3 CHUNKING AND COALESCING OF EDWARD’S ASSESSMENT CUES

Clinical history

● Repair of oesophageal atresia and tracheo-oesophageal fistula 12 hours

previously
● Anastomotic tension
● Bag-and-mask ventilation to be avoided if at all possible

Current status

● Preterm baby 36/40

● Recent anaesthetic
● Continuous morphine infusion
● Mechanically ventilated

Clinical observations

● Heart rate 60 BPM

● No chest movement
● Cyanotic

Staffing context

● Inexperienced Registrar present

● Senior House Officer elsewhere
● On-call Consultant off site
● Inexperienced attending nurse
● Night-time – less urgent care support

TABLE 12.2 The hypotheses generated from coalescence of cues

Coalescence of cues Tentative hypotheses
● Repair of oesophageal atresia and tracheo-oesophageal Risk of anastomotic breakdown
fistula 12 hours previously
● Anastomotic tension
● Bag-and-mask ventilation to be avoided if at all possible
● Preterm baby 36/40 Edward is unlikely to breathe if
● Recent anaesthetic endotracheal tube removed
● Continuous morphine infusion
● Mechanically ventilated
● Heart rate 60 BPM Blocked ETT
● No chest movement
● Cyanotic
(continued)
 244 ● A FOUNDATION FOR NEONATAL CARE

Coalescence of cues Tentative hypotheses

● Inexperienced Registrar Risk of failure to reintubate
● Senior House Officer elsewhere
● On-call Consultant off site
● Inexperienced attending nurse
● Night-time – less urgent care support

These tentative hypotheses form first-order judgements, which become second-

order cues. Interpretation of the second-order cues and whether to accept the opinion
and the probability that the opinion is correct leads to second-order judgements and
the choice of an operational decision. The Sister’s decision-making process using the
hypothetico-deductive process is illustrated in Figure 12.1.

Judgement Process

1st Order Cues 1st Order Judgement 2nd Order Judgement

Decision
2nd Order Cues

• Post-operative oesophageal atresia Risk of anastomatic

repair 12 hours previously breakdown
• Anastomotic tension
• Bag-and-mask ventilation to be avoided
if at all possible

• Preterm baby 36/40 wks Risk will not breathe if Instill normal saline
• Recent anaesthetic ETT removed and suction
• Morphine infusion
• Ventilated

• Heart rate 60 BPM Blocked ETT Reintubate

• No chest movement
• Cyanotic

• Inexperienced Registrar Risk of failure to reintubate

• Senior House Officer elsewhere
• On-call Consultant off site
• Inexperienced attending nurse
• Night-time – less urgent care support

Instill normal saline

and suction

FIGURE 12.1 Demonstrates the decision-making process using the hypothetico-

deductive process

However, it is possible for clinicians to come to different decisions even though

they are presented with the same set of cues. One way of explaining this is through
social judgement theory.6

Social judgement theory

Using social judgement theory we can answer the question: Did the clinician get
it right? This theory helps us to evaluate and learn from our decisions to improve
 CLINICAL DECISION MAKING ● 245

our decision making. Social judgement theory considers the relationship between
judgement and the selection of information. Judgements are analysed, after they have
been made, for their accuracy and the relative importance given to each cue. Both
the patient and the clinical context need to be considered equally.
The theoretical foundation for this judgement analysis is the Brunswik’s Lens
Model.6,8 In 19436 Brunswik, an Austrian psychologist, proposed the Len’s Model
that functions like a convex lens. The model applied to the clinical situation assumes
the actual state of the patient offers a clinician tangible surface cues, i.e. signs and
symptoms of their intangible states. The outcome of the assessment, judgement and
decision, in this example, not to remove the ETT but to instil 0.5 mL 0.9% sodium
chloride and apply suction, depended on the qualities of the judge, in this case the
Sister; the qualities of the patient, Edward; and the quality of the interaction. Because
tangible cues do not indicate the depth of conditions with certainty (inferred states
such as severity or a blocked endotracheal tube), they are fallible.
Thompson and Dowding9 explain that the outcome of a judgement is a func-
tion of how the cues are used and interpreted. During shared decision making, the
Brunswik’s Lens Model may be useful in that it enables clinicians to explain their
rationale in terms of making a judgement by assigning different ‘weights’ to the
presenting cues. In this context it is possible that two clinicians faced with the same
patient, the same history and the same cues could make quite different judgements
depending on their assessment of the patient, which might have been influenced by
a number of factors including their level of knowledge, previous experience, personal
biases and values.

Cognitive continuum theory

Developed by Hammond,6 the cognitive continuum theory is a unifying theory
of human judgement and decision making that acknowledges the analytical and
humanistic-intuitive approaches. It focuses on the tension and distinction between
intuitive and analytical thinking, recognising the diversity of individual cognitive
strategies. The theory suggests that reasoning is neither purely intuitive nor purely
analytical but is rather a continuum between the two poles, with judgements located
at points somewhere in between,6,9 allowing judges to alternate back and forth
between analysis and intuition. The theory offers a set of interrelated predictions
about task properties and cognition, suggesting that the cognitive approach of a judge
relies upon three principles, namely:
◗ the complexity of the task to be undertaken
◗ the clarity of the task, and
◗ the way in which the task is presented.

Tasks that tend to induce intuitive judgements, as in the Sister’s case, tend to be
complex; lack clarity; are unstructured; present many simultaneous cues; have a
variety of solutions; and require a rapid response. On the other hand, tasks to which
analytical judgements best respond are clear, highly structured, have few cues that
present sequentially, have definitive or few solutions and allow for time to plan a
considered response.6
 246 ● A FOUNDATION FOR NEONATAL CARE

Well structured
1

High
Scientific
experiment

2
Controlled
trial

Possibility of manipulation
3
Quasi
Task structured

experiment

4
System aided
judgement

5
Peer aided
judgement
Ill structured

6
Intuitive

Low
judgement

Intuition Cognitive mode Analysis

No time to plan Time to plan

FIGURE 12.2 The cognitive continuum

The cognitive continuum theory provides a framework for clinicians to aim for
accuracy in their decision making. Not adhering to its basic premise is likely to lead
to judgement inaccuracies.29 Using cognitive continuum theory we can answer the
question: How did the healthcare professional get it (i.e. the judgement) right? The
cognition mode to use should depend on the task structure, number of cues and time
available to complete the task.9 In the neonatal environment, mode 6 represents the
position at which the articulation of evidence and Sister’s cognition occurred.15,29
Because of the large number of cues, very limited time and poorly structured task,
the Sister’s decision sits at mode 6 on the continuum (see Figure 12.2).
Kahneman et al.10 cautioned that over-reliance on intuition to the exclusion of
analysis can severely limit the use of knowledge and evidence. This can have det-
rimental effects for clinicians not producing an informed rationale to justify their
decisions. The cognitive continuum highlights this need to underpin practice deci-
sions with a sound evidence base. So we turn now to the options that allow us to
calculate and provide a rational basis for clinical decisions.

THE NATURE OF DECISION ANALYSIS

As a discipline, ‘Decision Analysis’ is relatively new, dating back to the 1960s. In
1968 Howard Raiffa30 published the first significant text on the subject. Usually,
decision analysis involves the application of a mathematical or statistical model and
is thus generally perceived as being objective and quantitative. However, this is a
 CLINICAL DECISION MAKING ● 247

misconception, since decision analysis should use all types of information, some of
which may be subjective. The nature of decision analysis can be summarised as:
◗ being systematic and structured
◗ integrating intuitive and subjective judgements
◗ involving qualitative and quantitative components
◗ requiring rational analysis.

Stages in decision analysis

Decision analysis uses a structured framework which enables decision making to be
shared with colleagues and patients. It can provide the opportunity for numerical,
quantitative evaluation of the risks and benefits associated with some decisions.

Defining the problem

At the heart of the decision is the problem to be solved. It appears an overstatement
but to solve the problem effectively it is necessary to solve the right problem. To
understand fully what the problem is, the scope of the problem must be known.
This allows the identification of who must take part in the decision-making process.
In Edward’s case: Should the Sister sit in conclave with the medical staff? Should
Edward’s allocated nurse be involved? The Sister decided to define the problem by
herself. This is how she appeared to perceive it: ‘To remove or not to remove the
endotracheal tube?’ Would the problem definition have been different if others had
been involved in defining it? Probably not, for this is an obvious problem. But bigger,
less obvious strategic problems are almost entirely shaped by the worldviews of those
who take part in the decision.

TABLE 12.3 Elements of the rational decision-making process applied to Edward’s

case
Element of Definition Example in Edward’s case
the decision
Objective(s) The aims or goals to be achieved Ensure patency of blocked endotracheal tube
as a result of the solution found
Strategies The alternative approaches ● Remove the tube immediately and

that can be used to achieve the reintubate

objective(s) ● Instil 0.5 mL 0.9% sodium chloride, apply

suction and reintubate later

States of the Environmental factors that Success or failure could lead to:
world are beyond the control of the ● anastomotic breakdown

decision maker but can influence ● tube patent but Edward’s condition fails to
how successfully the objectives improve
are achieved ● failure to reintubate

Payoffs The value (e.g. monetary, utility E.g. a minimum score of zero if Edward’s
or other score according to the condition fails to improve, reintubation
decision maker(s)’ view) of each fails and help does not arrive immediately;
strategy under each state of the alternatively, a maximum score of 100 if
world Edward’s condition improves and there are no
other side effects.
248 ● A FOUNDATION FOR NEONATAL CARE

At the end of the problem-definition stage, the objective(s), strategies and context
of the problem should be clear. This will allow the decision maker(s) to choose the
most appropriate model to analyse the value of each strategy.31

Rational models
Rational decision models are normative but they can also be prescriptive. A very good
exposition of the normative-prescriptive distinction can be found in Koehler and
Harvey.32 The aim of rational models is to maximise the ‘payoff ’; i.e. to choose the
best strategy. To do this the decision maker must be perfectly rational and be able to
perfectly calculate the ‘payoff ’ of each strategy given. ‘Perfectly rational’ and ‘perfect
information’ are two very big phrases and they point to the difficulties in applying
rational models, which we will revisit below.

Certainty, uncertainty and probability

The environments in which decisions are made can be categorised as certain, risky
or uncertain. Of these environments, decisions made under certainty are the easiest.
The greater the lack of certainty or the greater the degree of uncertainty, the more
complex the decision becomes. The idea, therefore, is that when faced with an uncer-
tain environment, the decision maker must employ a technique that reassigns the
uncertainty so that the end result is something akin to certainty. For this to happen it
is necessary to gather sufficient information to allow the decision maker to quantify
the degree of uncertainty or probability.
Statisticians use the word ‘probability’ to express numerically the degree of likeli-
hood that a given event will happen in the future.33 Thus, if an event is impossible, it
has a probability of 0 or 0%. If an event is certain to happen, it has a probability of 1
or 100%. The degree of other uncertain events can then be measured using numbers
between 0 and 1 or between 0% and 100%.
There are three generally accepted classes of probability:
◗ logical or classical probability
◗ frequentist or observed probability
◗ subjective probability.

Logical or classical probability

If it is possible to work out all the possible outcomes of a process and to count them,
then we can logically work out with exactitude the probability of an event. This is
then an exact, objective quantity.

Frequentist or observed probability

However, most events that clinicians encounter may not be easily countable or
logically decomposable. What, for example, is the probability that a baby born in
England may be preterm? In this situation, we may use survey or census data to
estimate the probability. According to the Audit Commission,34 635 748 babies were
born in England in 2006. Of these, 62 471 were admitted to neonatal units. Thus we
can estimate:
 CLINICAL DECISION MAKING ● 249

Probability that a baby will need neonatal care = Number of babies admitted to
neonatal units/Total number of babies born in England = 62 471/63 5748 = 0.1 or
10%

Again, this is an objective quantity. It is, however, not exact; it is only an estimate
because it is based on survey data.

Subjective probability
However, it may not be possible to collect survey data. For example, we know in
Edward’s case that the Sister decided not to remove the ETT because the risk of failure
to reintubate was ‘high’. But what exactly was the probability of risk of failure? Box
12.4 below shows the number of factors that could have contributed to risk of failure
to reintubate.

BOX 12.4 POTENTIAL INFLUENCING FACTORS ON

RISK OF FAILURE TO REINTUBATE EDWARD

● Inexperienced Registrar
● Senior House Officer elsewhere
● On-call Consultant off site
● Night-time – fewer staff to manage an emergency
● Success rate of reintubation in previous cases

Whilst it is unrealistic in this case (the Sister reports that the decision was made
‘in a matter of seconds’), perhaps in another context it would have been possible to
go back to previous records involving cases of reintubation of babies and calculate
the proportion of cases in which reintubation failed. But is this the best estimate of
the chance of failure for Edward? The answer is ‘No’, because it is almost impossible
that all the babies in previous cases were in exactly the same situation and reacted to
the treatment in exactly the same way as Edward did. So what to do?
In this situation, there were a number of factors all of which could have been
significantly influential. It was necessary for the Sister to use her judgement and
expertise to estimate as accurately as possible the probability of failure to reintubate
based on Edward’s unique circumstances and taking into account all the factors. She
judged the probability to be ‘substantial’. Given more time, it would have been worth-
while assigning a specific number to this ‘high’ judgement (e.g. 10% probability),
since precisely stated quantities, even when based on subjective judgement, allow for
much more effective rational analysis. Now, unlike the probabilities calculated above,
this quantity is neither objective nor exact. It is a subjective estimate.

The decision tree

For complex, multi-stage decision problems (unlike the case of Edward), it is often
helpful to represent the options, uncertainties and outcomes in the form of a deci-
sion tree.35 A decision tree is a mathematical, graphical illustration of all the potential
250 ● A FOUNDATION FOR NEONATAL CARE

possible decision options for a given choice. For each of the possible decision options,
the consequences of taking that decision are quantified and expressed in terms of
likelihoods and utilities. In specific clinical contexts where difficult shared decision
making is important, a decision tree could offer the potential for all those involved
to choose the option that best maximises benefit for the patient whilst minimising
risk.

Clinical decision support systems

Clinical decision support systems (CDSS) are computer systems that integrate cues
from medical and patient information sources with the aim of helping clinicians
make decisions. Within the neonatal context their use is limited, a Cochrane Review
reported there was insufficient evidence to enable evaluation of any potential ben-
efit.36 More recently, the sensitivity and speed of ‘Isabel’ (Isabel Healthcare Inc.,
USA), a new web-based clinical decision support system has been evaluated in both
paediatrics37 and in adult medicine.38 In both studies it was reported that CDSS may
have a future role as a diagnostic aid in the decision-making process.

SUMMARY
This chapter used a clinical example, Edward’s case, to illustrate some of the key
considerations in making judgements and decisions in the practice environment and
some of the many ways in which decision making can be approached. This retrospec-
tive review of acute decisions may help either to validate such actions for the future
or to influence changes in practice.
Many of the issues crucial to effective decision making have been discussed,
including the importance of a sound assessment process and the ability to keep it
as wide as possible and not to narrow the judgement focus too soon; the potential
impact that heuristics and biases can have on creating decision errors; and the use
of social judgement and the cognitive continuum as tools to both understand and
analyse decisions, as a means of improving decision making. There remains much
work to be done in the field, accruing evidence in the science of decision making
in healthcare generally and in neonatal practice particularly. It is hoped this chap-
ter might act as a springboard not only to improving your decision making but to
deliberations with regard to contributing to further research and evidence gathering
in this most important of skills.

KEY POINTS FOR PRACTICE

1 Sound assessment is based on the quality of information gathered.
2 The process of information gathering relies on cognitive searches, utilising reli-
able and accurate sources of information and knowledge, and the interpretation
of visual and non-visual cues.
3 Judgement is a product of the assessment process and personal cognitive ability.
4 Clinicians need to reflect on the decision-making process in order that they
acknowledge their personal biases, values and the role that heuristics may play.
5 Effective clinical decisions depend on sound judgements which take account of
the patient’s clinical needs, priorities, and values.
 CLINICAL DECISION MAKING ● 251

REFERENCES
1 Gilovich T, Griffin D, Kahneman D, editors. Heuristics and Biases: the psychology of intuitive
judgement. Cambridge: Cambridge University Press; 2002.
2 Graber ML, Berner ES. Diagnostic error: is overconfidence the problem? Am J Med. 2008;
121(5A): S2–46.
3 Elstein AS, Schwartz A. Evidence base of clinical diagnosis: clinical problem solving and
diagnostic decision making: selective review of the cognitive literature. BMJ. 2002; 324:
729–32.
4 Klein JG. Five pitfalls in decisions about diagnosis and prescribing. BMJ. 2005; 330:
781–3.
5 Goldstein WM, Hogarth RM, editors. Research on Judgment and Decision Making: currents,
connections, and controversies. Cambridge: Cambridge University Press; 1997.
6 Cooksey RW. Judgement Analysis: theory, methods and applications. London: Academic Press;
1996.
7 Standing M. Clinical judgement and decision-making in nursing – nine modes of practice
in a revised cognitive continuum. J Adv Nurs. 2008; 62(1): 124–34.
8 Hammond KR. How convergence of research paradigms can improve research on diagnostic
judgement. Med Decis Making. 1996; 16(3): 281–7.
9 Thompson C, Dowding D. Clinical Decision Making and Judgement in Nursing. London:
Churchill Livingstone; 2002.
10 Kahneman D, Slovic P, Tversky A. Judgement Under Uncertainty: heuristics and biases.
Cambridge: Cambridge University Press; 1982.
11 Kahneman D, Tversky A. Subjective probability: a judgment of representativeness. Cognitive
Psychology. 1972; 3: 430–54.
12 Cioffi J. A study of the use of past experiences in clinical decision making in emergency
situations. Int J Nurs Stud. 2001; 38(5): 591–9.
13 Payne JW, Bettman JR, Johnson EJ. The Adaptive Decision Maker. Cambridge: Cambridge
University Press; 1993.
14 Hams SP. A gut feeling? Intuition and critical care nursing. Intensive Crit Care Nurs. 2000;
16(5): 310–18.
15 Harbison J. Clinical decision making in nursing: theoretical perspectives and their relevance
to practice. J Adv Nurs. 2001; 35(1): 126–33.
16 Buckingham CD, Adams A. Classifying clinical decision making: interpreting nursing intui-
tion, heuristics and medical diagnosis. J Adv Nurs. 2000; 32(4): 990–8.
17 Gobet F, Chassy P. Towards an alternative to Benner’s theory of expert intuition in nursing:
a discussion paper. Int J Nurs Stud. 2008; 45(1): 129–39.
18 Benner P. From Novice to Expert. London: Addison Wesley; 1984.
19 Gobet F, Clarkson G. Chunks in expert memory: evidence for the magical number four . . .
or is it two? Memory. 2004; 12(6): 732–47.
20 Gobet F, Simon HA. Five seconds or sixty? Presentations in time in expert memory. Cognitive
Science. 2000; 24: 651–82.
21 McKinnon J. Feeling and knowing: neural scientific perspectives on intuitive practice. Nurs
Stand. 2005; 20(1): 41–6.
22 Panksepp J. Affective Neuroscience: the foundations of human and animal emotions. Oxford:
Oxford University Press; 1998.
23 Carper BA. Fundamental patterns of knowing in nursing. Adv Nurs Sci. 1978; 1(1): 13–23.
24 Heath H. Reflection and patterns of knowing in nursing. J Adv Nurs. 1998; 27: 1054–9.
 252 ● A FOUNDATION FOR NEONATAL CARE

25 Hagnell E. Nursing knowledge: women’s knowledge: A sociological perspective. J Adv Nurs.

1988; 14: 226–33.
26 Benner P, Tanner CA, Chesla CA. Expertise in Nursing Practice: caring, clinical judgement,
and ethics. New York: Springer; 1996.
27 Bohinc M, Gradisar M. Decision-making model for nursing. J Nurs Admin. 2003; 33(12):
627–9.
28 Elstein AS, Shulman LS, Sprafka SA. Medical Problem Solving: an analysis of clinical reasoning.
Cambridge: Harvard University Press; 1978.
29 Cader R, Campbell S, Watson D. Cognitive Continuum Theory in nursing decision-making.
J Adv Nurs. 2005; 49(4): 331–447.
30 Raiffa H. Decision Analysis: introductory lectures on choices under uncertainty. London:
Random House; 1968.
31 Goodwin P, Wright G. Decision Analysis for Management Judgement. 3rd ed. New York: John
Wiley; 2004.
32 Koehler DJ, Harvey N, editors. Blackwell Handbook of Judgment and Decision Making. New
York: Blackwell; 2004.
33 Rowntree D. Probability Without Tears: a primer for non-mathematicians. New York: Barnes
& Noble; 1994.
34 Report by the Comptroller and Auditor General. Caring for Vulnerable Babies: the reorganisa-
tion of neonatal services in England. HC 0101 Session 2007–2008. London: HMSO; 2007.
35 Jones MJ. Decision analysis using spreadsheets. Eur J Oper Res. 1986; 26; 385–400.
36 Tan K, Dear PRF, Newell SJ. Clinical decision support systems for neonatal care. Cochrane
Database Syst Rev. 2005; 2: CD004211.
37 Ramnarayan P, Tomlinson A, Rao A, et al. ISABEL: a web-based differential diagnosis aid
for paediatrics: results from an initial performance evaluation. Arch Dis Child. 2003; 88:
408–13.
38 Graber ML, Mathew A. Performance of a web-based clinical diagnosis support system for
internists. J Gen Intern Med. 2008; 23(Suppl. 1): 37–40.

View publication stats