**5. Discussion**

Thanks to face-to-face interviews, based on the Kahneman et al. (2011) checklist adjusted for orthopaedics, we derived some qualitative issues highlighting potential weaknesses of the cognitive path which steers orthopaedists in decisions. In particular, we preliminary understood that the follow-up choice (C.Q. **A**) can be considered as the only stage of the orthopaedic surgical process where physicians could potentially take decisions depending solely on the patient's real needs and conditions, without interferences from guidelines.

Nevertheless, the analysis of all the interviews discloses a high impact of consolidated practice also on clinical decision-making about follow-up. This might be due to two reasons: firstly, orthopaedics is a traditional specialisation, based on "standardisation of process" for this kind of healthcare pathway (Healy et al. 1998; Scranton 1999). The other reason is due to a "cultural issue" related to the physician's training, which makes practitioners feel more confident in doing what they have always done (as reported by SP during his interview); such behaviour can bring decision-makers to minimise risks of their practice and exacerbate its benefits (Kahneman et al. 2011) conducting to the **A** ff**ect Heuristic**.

Moreover, all respondents agree that a variable that certainly influences the decision-making is the coherence with decisions undertaken by other similar departments. According to Grove et al. (2015), orthopaedic surgeon decision-making is strongly influenced by the key-opinion leaders' guidelines and suggestions. This often brings the surgeon to take a decision in his/her operating context as an emulation of someone else's choice or endpoints; this kind of bias, however, could jeopardise the correct clinical decision.

This result is related to the scientific nature of medical science (which is based on scientific literature) as emerging:


In particular, evidence-based medicine (EBM) involves the use of the current best practices to make decisions regarding the patient's care (Sackett et al. 1996); in this way, decisions are steered together with the kind of data that a physician collects (Twells 2015). This approach, strongly related to the **Anchoring** bias, might speed up the doctors' decision-making with certain data and information (based on scholars' endpoints) without looking for (potentially useful) others.

Otherwise, the impact of other departments' evidence or key-opinion leader insights is a very meaningful aspect of understanding the decision process of surgeons; this concerns the **Halo E** ff**ect**. Accordingly, as reported by Cook et al. (2009), discussion with colleagues is more influential in clinical practice than empirical support; in this regard, SD is convinced that decisions are surely influenced also by the other departments' experience; and, as reported by SP, "for me, it is easier changing the surgical practice, if an opinion leader in the sector suggests me to change it". Moreover, according to all respondents' opinions, surgeons are more willing to change their practice on the basis of the experience of practitioners deemed most competent in the field.

Regarding the influence of past results (**Saliency**), according to Stewart and Stewart and Chambless (2007), all interviewed clinicians recognise that the past experience is more important in clinical treatment decisions than empirical research knowledge.

Nonetheless, there are some biases that are absent for some interviewees' opinions; to analyse more in detail these errors, it is important to notice that the respondents have some characteristics in common. In particular:


Therefore, the errors of **Groupthink** and **Availability** (recognised by SD and SH) are probably most concerning to the context of the public ownership of the hospital. Indeed, as reported by SP, the simultaneous absence of Groupthink and Availability in his practice "is related to the characteristic of my working place: I have my private patients whose information is owned mostly by me. In the private hospital, we are several physicians with our own patients; I just made decisions for mine". This result is in line with Smith et al. (2007) and Eisenberg (1979) that the identified external factors, features of clinicians, clinician's interaction (Robinson et al. 2017) with his/her profession, and the health care system as the factors that can modify the decision process. The context in which decisions are made can be very significant, but it has not been rigorously explored in prior studies (Durning et al. 2011).

Concerning the **Overconfidence** (the errors that SD and SP have in common), it may be caused by the lengthy experience of the surgeon. On the contrary, the lack of experience can be the explanation for the **Confirmation** bias as an error came out only from SH's interview. He is, indeed, the one with less than five years of experience.

Summarising, according to Kahneman's checklist, the analysis of the interviews showed the possibility that 8 biases out of 12 affect the decision-making process of the orthopaedic surgeons regarding the follow-up of patients undergoing hip and knee arthroplasty. Below, Table 3 reports the errors that emerged, their explanation, and medical scientific literature regarding them. The last column of Table 3 also reports which of the interviewees recognised the specific errors.


**Table 3.** literature on medical field our study's finding confirmed by literature.

Source: Authors' elaboration inspired by Cristofaro (2017a); \* All descriptions are taken from Cristofaro (2017a); \*\* Authors refer to Saliency as "Representativeness".

### **6. Conclusions and Implications**

This paper addresses the theme of the clinical reasoning process in healthcare. Precisely, the work aims at understanding if and how the decision-making process of orthopaedic surgeons can be affected by cognitive biases. Particularly, within the decision process sphere, the choice regarding the patient's follow-up after knee and hip arthroplasty has been analysed. To achieve the goal of this study, Kahneman's checklist was employed in order to recognise which kind of errors can mostly lead surgeons' decision-making process; accordingly, we conducted three semi-structured interviews with key-decisional orthopaedic surgeons working in different organisations. The results show several biases that can affect the clinical decision process regarding follow-up after knee and hip arthroplasty.

In particular, some are in common for all the interviews (Affect heuristic, Anchoring, Halo effect, Saliency); the others (Groupthink, Availability, Overconfidence, Confirmation) are related to the following two personal variables of surgeon: (i) working experience; (ii) working context.

Concerning the biases which differ among interviewees, the following Figure 2 summarises the main contribution of the work.

**Figure 2.** Matrix connection between physician's working experience and context with the cognitive biases in orthopaedics. Source: Author's elaboration.

**.**

Reading the figure as a matrix, it highlights the connection between working experience and working context with specific biases of the decision-making process in the orthopaedic field, in coherence with the main literature listed in Table 3.

Figure 2 shows that the less experienced a surgeon is, the more likely is he/she to look for **Confirmation** (Smith et al. 2010); this might be because he or she is very tied to theory and tends to look for what he knows. Ierano et al. (2019) confirmed that less experienced surgeons/junior health professionals always look for confirmation through guidelines. On the contrary, **Overconfidence** incurs due to high experience. Besides the working context, high experience leads the surgeons to think mostly on positive scenarios (Kahneman et al. 2011) regarding their work/task as they are more confident in their skills.

**Groupthink** and **Availability**, however, are both mostly related to the public context characteristics; the first one is typical of hospital ward teamwork, that is out of the SP's working sphere. The second one, instead, may be related to the data present as support to the decision-making path, whose availability (per timing, quality, and quantity) are mandatory only in public hospitals, according to public performance measurement roles (Bouckaert and Halligan 2007; Pinnarelli et al. 2015).

Hence, based on the level of experience (low/high) of surgeons and the ownership of the healthcare institution (public/private) in which they work, the matrix points out the most probable biases in the orthopaedic field according to the specific features of decision-makers. This could be very useful for the managemen<sup>t</sup> of the healthcare institution in terms of prompt reaction to the expected cognitive errors.

Furthermore, for sure our results confirm that a qualitative cognitive tool, as the Kahneman et al. (2011) checklists, could potentially help physicians avoid these errors, but it needs to be integrated in daily practice, also as a more usable electronic version (Raymond et al. 2017; Otokiti 2019). For sure this manuscript presents some limitations; most of them are those related to the use of the qualitative method which implies the interpretive role of researchers and limited extension of data. The first limitation of this qualitative inquiry regards the lack of distinction between "heuristics" and "traps", which both fall under the bias umbrella term. Although heuristics could generally have also a positive impact on decisions (thus, in particular circumstances, they should not be reduced), however, this is not proved in the medical field according to Ryan et al. (2018). In addition, only three physicians are interviewed. Nevertheless, as explained in the methodology section, the differences in terms of working experience and context of our interviewees would reduce such a limitation by giving a good representation of the Italian orthopaedic environment. Moreover, given that this research is moving its first steps and it is at an original level of investigation in a still understudied field, according to literature Cristofaro (2017a) and Jette et al. (2003), three interviewees can be considered enough if they represent the apical position for the decision-making context of their organisational environment. For sure to overcome these limitations, a higher number of interviews should be made in future research. Further studies could focus on theoretical exploration (e.g., systematic literature reviews, bibliometric literature reviews, etc.) of clinical decision-making in the surgical field, which would arise the difference between specialisation sub-fields. Moreover, given the fact that the magnitude of heuristic effects on complex clinical decisions is still unexplored, this specific aspect would also deserve to be investigated in future streams of research. In addition, the study leaves some areas of investigation uncovered; particularly surprising is the lack of connection with the Kahneman's bias of Self Interest. No direct evidence of this arose from the interviews; only insights regarding the decisions about the frequency of follow-ups came out. According to all respondents, the schedule of follow-up is established in line with literature, practice, and in some rare cases according to the patient's peculiarity (this would confirm all other results). Nevertheless, as highlighted by the control question E, a potential reason that could modify the follow-up frequency and schedule would be the specific interest of physicians in collecting information (both clinical and epidemiologic) according to healthcare managemen<sup>t</sup> and scientific production based on big-data (Roski et al. 2014; Yan et al. 2017). Also, these aspects deserve to be further investigated.

In conclusion, this study has highlighted the linkage between the clinical decision-making process and managemen<sup>t</sup> tools to improve decisions, by fostering debate in these fields. For practitioners, this study shows the experience of quality decision-making process tool (To et al. 2018) employment that brings out some cognitive shortcuts that can lead the clinical decision process. From an academic point of view, this article represents a preliminary contribution to the influence of cognitive biases in limiting the rational thinking of decision-makers in the specificity of the orthopaedic field.

With this regard, the study can surely contribute to the debate by both scholars and practitioners about the application of tools in improving the quality of the clinical decision process. According to Antonacci et al. (2020), indeed, the improvement of decision-making is one of the main leverages for enhancing better healthcare outcomes, which, in turn, can be translated into better performance (Skaržauskiene 2010; Safi and Burrell 2007; Oyewobi et al. 2016) for the healthcare organisation.

**Author Contributions:** Conceptualization, I.S. and G.P.; methodology, G.P.; software, I.S.; validation, G.P., I.S. and A.C.; formal analysis, I.S. and G.P.; investigation, G.P. and I.S.; resources, G.P., I.S. and A.C.; data curation, I.S. and G.P.; writing—original draft preparation, I.S. and G.P.; writing—review and editing, G.P. and I.S.; visualization, A.C.; supervision, A.C. and G.P.; project administration, I.S. and G.P. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research received no external funding.

**Conflicts of Interest:** The authors declare no conflict of interest.
