Clinical decision-making is a fundamental part of nursing care that impacts patient outcomes (Johansen and O'Brien, 2015). It is estimated that nurses make a clinical decision every 30 seconds in acute care (Yang et al, 2014) and five decisions per consultation in the community (Thompson and Yang, 2009).
Decision-making is a complex, non-linear process involving the collection and analysis of data (Tiffen et al, 2014), use of clinical judgement to weigh up treatment options (Thompson and Stapley, 2011) and use of systematic enquiry to problem-solve clinical conundrums (Johansen and O'Brien, 2015). Standing (2008) advocates reflecting on clinical decision-making to further enhance learning within practice (Johansen and O'Brien, 2015). Decisions can sometimes be based on personal judgements (Hammond et al (1975), so reflection can be used to recognise bias (Moyo et al, 2019). Therefore, it is essential to accurately assess the clinical presentation of a patient, identify their concerns and goals and address these by developing an appropriate plan of care.
Kolb's experiential learning theory (1984) asserts that, in order to learn from practice, there must be a connection between cognitive processes and actions (Table 1). Gibbs (1988) furthered this work with his six-stage reflective cycle:
- Description: describes the experience
- Feelings: explores thoughts and feelings during the experience and their impact
- Evaluation: evaluates the positive and negative aspects of the experience
- Analysis: making sense of the experience
- Conclusion: what can be learned from the experience
- Action plan: identifying what could be done differently next time and what changes are needed to achieve this
Table 1. Description of decision-making models and theories
| Model or theory | Description |
|---|---|
| Experiential learning theory (Kolb, 1984) | Learning through repeated experience. Kolb believed that new knowledge was developed through conceptualising experiences and gaining new insights, which could then be applied to future situations |
| Shared decision-making model | Shared decision-making is a collaborative process that involves the patient and the professional sharing personal values and preferences, along with the evidence base, to reach an informed decision about treatment |
| Think aloud method (Ericsson and Simon, 1993) | Described as the process of verbalising one's thoughts while completing a task. Employing the think aloud method can illuminate the thought processes involved in complex decision-making |
| Hypothetico-deductive reasoning (Elstein et al, 1978) | Hypothetico-deductive reasoning is a methodical process of decision-making. There are 4 steps:Cue acquisition, which involves gathering informationGenerating a hypothesisCue interpretation through tests and observationHypothesis evaluation, in which the initial hypothesis can be modified depending on the evidence |
| Dual processing theory | First conceived in the 1980s and further developed by Daniel Kahneman (2011), dual processing theory proposes that there are two systems involved in decision-making:System 1, which is intuitive, fast and based on previous experiences, but more open to logical fallacies and biasesSystem 2, which is analytical, slower and resource-intensive, but more systematic and logicalIn practice, both systems influence clinicians' decision-making |
Although Gibbs' work has been criticised for being limited and simplistic (Finlay, 2008), Middleton (2017) believes that it can be broadly applied in nursing practice. As nursing experiences are cyclical in nature, this can enable deeper understanding of clinical experiences, which can be applied to similar patients in the future (Standing, 2017).
Case study
John (a pseudonym used to protect his identity) was a 54 year-old man who was referred to the mental health crisis resolution team by his GP after attempting suicide. There had been a 1-month history of low mood, with the context of a relationship breakdown and the loss of his accommodation and business, resulting in debt. The aim of the assessment was to determine a diagnosis and an appropriate treatment plan.
John presented with emotional distress and mild low mood. He had a history of mild low mood, which had been treated by his GP with two different selective serotonin reuptake inhibitor (SSRI) antidepressants, with reported minimal benefit. A risk assessment established a history of mental health issues, recent significant losses likely to cause ongoing short-term distress and demographic factors that increased his risk, such as age, sex, social isolation and relationship status (Overholser et al, 2012). John reported attempting to end his life, but did not see this through for fear that it would not prove fatal. He indicated ongoing suicidal thoughts but no current planning or intent, and agreed to keep himself safe while receiving care from the team. A diagnosis of adjustment disorder was made, as his symptoms could not be assigned to any other disorder. He was heavily preoccupied with his losses and was experiencing difficulty adapting to his situation (Zelviene and Kazlauskas, 2018; World Health Organization (WHO), 2019) (Box 1).
Box 1.Diagnostic criteria for adjustment disorder
- A maladaptive reaction to an identifiable psychosocial stressor or multiple stressors
- Usually emerges within a month of the stressor and resolves within 6 months of the stressor and consequences ending
- The symptoms are not better accounted for by another mental disorder
- The reaction to the stressor is characterised by preoccupation with the stressor or its consequences, including excessive worry, recurrent and distressing thoughts about the stressor, or constant rumination about its implications
- Results in significant impairment in personal, family, social, educational, occupational or other important areas of functioning. If functioning is maintained, it is only through significant additional effort
Adapted from World Health Organization (2019)
Feelings
John presented as challenging and dismissive. He believed that prescribers chose the cheapest treatment, rather than the most suitable, and was difficult to engage in meaningful discussion around his care. Despite these difficulties, it remained important to collaborate on a treatment plan (Standing, 2017). Adopting a person-centred approach contributed to an understanding of John's pessimistic views towards healthcare services. Such person-centred care can counter interpersonal difficulties, acknowledge the clinician's negative bias towards patient behaviours and the influence of the clinician's own personal values (Israelashvili, 2012). While John was ambivalent about the value of treatment, he did agree to begin medication.
Evaluation
There is a lack of evidence-based approaches to the management of adjustment disorder (Kazlauskas et al, 2017). Despite high prevalence (Casey et al, 2001; 2015), there is little research into the condition (Carta et al, 2009). There are no published clinical guidelines for adjustment disorder, although treatment is advised if symptoms are persistent (Kazlauskas et al, 2017) and distressing (Semple and Smyth, 2013).
Section 3.1 of the Nursing and Midwifery Council's (NMC) (2018)Code of Conduct states that nurses must promote wellbeing and prevent illness; section 4.1 requires nurses to weigh up acting in the patient's best interests and respecting their autonomy. John was deemed to have capacity to decide about his treatment. However, considering his ambivalence towards medication, it could be argued that there was a moral duty, from a deontological perspective (Seedhouse, 2009), to encourage John's participation with the proposed treatment plan of medication and a referral for therapy. Talking therapy is an important part of treatment for adjustment disorder to support the reduction of the impact of the stressor and the development of healthy coping mechanisms (Domhardt and Baumeister, 2018).
The evidence base for use of medication for adjustment disorder patients is insubstantial, dated and conflicting. Although limited by their retrospective design, Hameed et al (2005) found patients with adjustment disorder were twice as likely to respond to antidepressants compared to those with depression. Olfson et al (1998) established that patients with adjustment disorder responded favourably to antidepressant monotherapy, but no single agent was found to be superior; however, this evidence is outdated, considering advancements in treatments.
Conversely, de Leo (1989) and Casey (2009) found no clear evidence that antidepressants are effective in adjustment disorder. Semple and Smyth (2013) advise using medication in patients with adjustment disorder if symptoms endure and cause suffering. John fulfilled these criteria, as his symptoms had been present for 1 month and had caused much distress, along with suicidal ideation.
Beauchamp and Childress (2001) proposed four principles to provide ethical decision-making in healthcare: beneficence; nonmaleficence; autonomy; and justice. Although these principles have been criticised as being outdated (Huxtable, 2013), superficial and not aiding reflective practice (Seedhouse, 2009), they are a useful guide for busy clinicians in practice (Huxtable, 2013). Applying the principles offers the following insights:
- Beneficence is ‘doing good’. From a utilitarian perspective, prescribing John medication would do the most good in seeking to promote his wellbeing and management of suicidal thoughts
- Nonmaleficence is ‘doing no harm’. Although prescribing medication may cause harmful side effects (Joint Formulary Committee, 2022), the intention would be to provide positive outcomes (Beauchamp and Childress, 2017)
- In regards to respect for autonomy, John had agreed to commence medication, despite his ambivalence. Had he asserted a firm stance against medication, with his established capacity, there would have been an obligation to respect his autonomous decision.
The prescription of medication could be justified for the aforementioned reasons, but it was also important to be able to use the experience to potentially benefit patients with similar presentation in the future.
With John's agreement to begin medication, the next step was to engage him in the decision-making process to select the most appropriate treatment. The shared decision-making model proposes that patients should be involved in the entire decision-making process and that clinicians should share the evidence base and encourage the patient to consider the evidence along with their own wishes and values, to achieve a mutually agreed plan (The Health Foundation, 2016; Constanzo et al, 2019). Although this process can be hampered by a lack of willingness from both parties and resource issues (Constanzo et al, 2019), shared decision-making positively impacts on patients' concordance with treatment, choices and outcomes (The Health Foundation, 2016). The ability to implement shared decision-making is influenced by the clinician's organisation (Constanzo et al, 2019). Scholl et al (2018) found that organisational priorities, effective leadership and positive feedback greatly influenced successful implementation of shared decision-making in practice.
Employing a think aloud approach can enable clinicians to communicate their thought processes around selecting an appropriate medication. The think aloud method (Ericsson and Simon, 1993) encourages exploration and connection to previous learning (Levett-Jones et al, 2019). However, it has been criticised for only showing an interpretation of and not the true thought processes (Guss, 2018) and being open to observational bias (Forsberg et al, 2014). Using this method with John demonstrated that the medication choice was not based on what was cheapest, as he believed, but instead on the best option for his clinical presentation.
In addition, hypothetico-deductive reasoning was used as part of the information processing theory (Banning, 2007), which proposes that clinicians undergo reasoned, logical steps to come to a decision. First, information was collated from several sources (Standing, 2017). John was not known to the local secondary mental health service, having recently moved to the area to live with his parents following the breakdown of his relationship. He did not want his parents to be contacted regarding his mental health issues, something clinicians must respect under the Data Protection Act 2018 (Royal College of Nursing, 2019).
There were no other sources of information available, as his GP notes had not been transferred. John reported feeling very depressed but was deemed, through observation and assessment, to have mild symptoms of depression (WHO, 2019). Despite reported suicidal thoughts, there was no immediate plan, and he had approached services for support. Nevertheless, in comparison to those with depression, patients with adjustment disorder are considered more vulnerable, impulsive and prone to early suicidal behaviours in the course of the illness (Carta et al, 2009; Casey et al, 2015). John had previously been prescribed sertraline and fluoxetine by his GP, with no noticeable benefits, despite adequate trial periods.
At this point, the think aloud method was employed to verbalise the thought process. An initial hypothesis was shared that a low dose of mirtazapine could be considered the next logical step following the trial of two different SSRIs (Taylor et al, 2015). Although this advice centres on diagnosed depression, it was thought that mirtazapine's sedative properties (Joint Formulary Committee, 2022) would also help with John's reported poor sleep, and its low toxicity levels have been found to be beneficial in cases of overdose (Semple and Smyth, 2013). Through exploring his physical health, it was found that John had type 2 diabetes; although he did not appear malnourished or dehydrated, he reported a poor diet.
John was informed that is not recommended to prescribe mirtazapine for patients with diabetes (Joint Formulary Committee, 2022), although Song et al (2014) found that, despite mirtazapine causing weight gain, which could affect diabetic patients, there were no other statistically significant results from blood tests. Therefore, Song et al (2014) concluded that mirtazapine could be used in stabilised patients. Information about his physical health had not been included in his GP referral, so it was not possible to corroborate any information; therefore, a decision was made to adopt a cautious approach.
Research shows a correlation between metabolic dysfunction, diabetes and depression (Lustman and Clouse, 2005; Taylor et al, 2015), meaning John could be vulnerable to developing more severe depressive symptoms in future (Kazlauskas et al, 2017). Given this new information, the initial hypothesis was evaluated, and consideration was given to trialling a third SSRI, as recommended by Taylor et al (2015). John was initially dismissive of this, seeing little value in trialling a third SSRI. In their definitive meta-analysis of antidepressants, Cipriani et al (2018) found that fluoxetine, sertraline and escitalopram were best tolerated; however, of the three, only escitalopram was deemed highly effective.
Although limited by validity issues and lack of information in some studies, Cipriani et al's (2018) meta-analysis is still considered the most complete review of prescribing information on antidepressants. Cipriani et al (2018) found that patients who had not adequately responded to one SSRI could benefit from another, despite similarities in their mode of action.
Sharing this evidence base with John enabled discussion around the study's findings, which ensured his understanding and his ability to use the information to inform any decisions, as advocated in the shared decision-making model. After discussion, it was concluded that John would commence on 10mg of escitalopram, as evidence suggests this treatment is well tolerated, effective and safe for patients with diabetes.
Analysis
Kahneman (2011), a pioneer of decision-making theory, proposes that humans have two different forms of thinking: one that is fast and intuitive, and another that is slower and more analytical. This is known as the dual-processing theory. Researchers suggest the analytical and intuitive thinking patterns work simultaneously to complement each other (Schuwirth et al, 2019). Intuitive thought processes are quick and unconscious (Croskerry, 2013) and built through experience (Cooper, 2016; Melin-Johansson et al, 2017), as knowledge and pattern recognition develops via repeated exposure. Intuitive thinking also corresponds with tacit knowledge (Cooper, 2016) which is gained through personal experience. However, these processes are more prone to error (Croskerry, 2013). When facing doubt, it is expected that clinicians should seek information to reduce uncertainty; however, Thompson and Yang (2009) found that, in practice, clinicians often rely on intuitions to guide their decision-making process, which can lead to bias.
Analytical thinking is defined as a slower cognitive process (Levett-Jones et al, 2019) that is logical, reasoned and methodical (Croskerry, 2009; Bjork and Hamilton, 2011). Analytical thinking is less prone to error (Croskerry, 2009); however, it uses greater cognitive resources and is, therefore, impractical to employ for every decision faced in practice (Croskerry, 2013).
Work by Bjork and Hamilton (2011), although limited by survey method and geographical area, demonstrates that nurses employ both intuitive and analytical approaches, incorporating reasoned judgement, intuition, tacit knowledge and analytical thinking into a quasirational process (Baake, 2007).
There were several areas of doubt within John's assessment. Lack of information from outside sources, absence of treatment guidelines and an inability to corroborate self-reported information can lead clinicians to exercise caution. This is influenced by diagnostic and treatment responsibilities (Thompson and Yang, 2009), lawful duty of care (Cox, 2010) and professional accountability (NMC, 2018).
While intuition could lead towards prescribing medication for John to treat his symptoms and impact his suicidal thoughts, there is a risk, when considering beneficence, of pathologising life events (Zelviene and Kazlauskas, 2018). Johansen and O'Brien (2015) assert that clinicians assign importance to available information through the decision-making process in a way that is personal and contextual, but caution that this must be subject to reflection.
Reflecting on the assessment, early anchoring towards prescribing medication was present, rather than an exploration of non-pharmacological treatment, especially in consideration of the fact that symptoms of adjustment disorder are time-limited, usually resolving within 6 months. Anchoring can lead to confirmation bias (Cooper, 2016), whereby the clinician selects information to confirm their original belief. While this can be influenced by several factors, risk is of particular significance for clinicians in mental health, as patients with adjustment disorder are considered vulnerable to completed suicide (Casey and Bailey, 2011; Casey et al, 2015).
When determining John's risk, his demographic information and social circumstances could place him at a higher risk of suicide. Clinical decision-making is influenced by the organisation in which it is conducted, which needs to support clinicians in these processes (Bjork and Hamilton, 2011; Guerrasio and Lessing, 2019).
Mental health trusts are guided by Care Quality Commission (2018) requirements on safety and implementing learning from adverse events into practice. This focus on safety, the influence of investigations into serious events and media reporting impact upon the willingness of clinicians to engage in positive risk-taking.
Anecdotal evidence suggests that, in some decision-making areas, clinicians ‘work backwards', a form of abductive reasoning (Hughes and Nimmo, 2016), and consider whether they would be able to justify their decision to investigators if the patient lost their life. Although this can help to guide decision-making, it can also promote defensive practice. Therefore, despite the wish to engage in shared decision-making, an unconscious process of guiding John towards accepting medication to satisfy the clinicians’ own needs may have been present.
Action plan
Action planning is perhaps the most important part of the reflective process, as it guides the clinician on planning how to apply what was learnt to future practice (Howatson-Jones, 2016; Koshy et al, 2017).
The NMC (2019) strongly advocates reflective practices and sharing knowledge. Continued professional development is part of the NMC's Code of Conduct, although this can be hindered by time and resource issues (Scala et al, 2016). A clinician's trust also has the responsibility to enable this process (Williams et al, 2015).
Maintaining a positive attitude towards reflection and regular supervision enhances its effects (Gonge and Buus, 2014). Reflecting on both positive and negative experiences within teams can lead to meaningful change and reduce isolated practices (Koshy et al, 2017; NMC, 2019). Domingo Galutira (2018) found that effective reflection leads to personal and professional development, better patient outcomes and enhanced quality of care. A whole-team approach and organisational support to would be needed to implement reflection effectively.
John's case study was shared with colleagues in group supervision and local prescribing forums. This approach shared lessons learned and supported new learning opportunities and professional development for colleagues. There was discussion around the issue of prescribing within a limited evidence base. The outcome of these discussions established a need for management strategies to employ in the future for similar presentations.
The agreed strategies included regular reviews of clinical guidelines and research findings and the continued need for individual and group supervision to share concerns within the team. Additionally, there were discussions with the management team about how to embed shared decision-making within the team to improve patient outcomes, and how to acquire the resources for effective implementation.
Discussion
This case has highlighted the importance of placing the patient at the heart of the decision-making process. Although nursing practice is unavoidably influenced by cognitive biases, these can be acknowledged through the reflective process (Croskerry, 2013). Through the use of a person-centred approach and the shared decision-making model, it was possible to address John's negative views of mental health services and promote concordance with the mutually agreed treatment plan.
In practice, nurses can rely on intuitive decision-making, especially in resource-limited scenarios. However, taking time to reflect upon decision-making processes, legal and ethical implications and current organisational culture can serve to enhance practice and perhaps allay any fears around risks, while building confidence in the ability to face future doubt.
Taking time to familiarise themselves with current research and the evidence base will enable clinicians to confidently and fluidly recall pertinent information to guide future decision-making (Ritter and Witte, 2019) and improve services.
Conclusions
Reflecting upon decisions made in practice can uncover a variety of influences that can help or hinder the decision-making process. Examining these influences can ensure that biases are recognised and addressed, and that ethical, professional and organisational needs are balanced within the decision-making process.