physicians to withdraw treatment of
critically ill patients being delayed by the
sunk-cost e
ffect?
Sunk-cost e
ffect in Dutch Intensive Care Units
By
Nena Dekema
Submitted for the degree of Master of Science University of Amsterdam
Medical Informatics August 2018
Nena Dekema, BSc.
Student number: 10641688 E-mail: h.dekema@amc.uva.nl
Mentor AMC
Dave Dongelmans, PhD.
Amsterdam UMC, locatie AMC - Intensive Care
E-mail: d.a.dongelmans@amc.uva.nl
Mentor KPMG Eva Tsjapanova, MSc.
KPMG, Health IT Audit & Assurance
E-mail: tsjapanova.eva@kpmg.nl
Tutor
Danielle Sent, PhD.
Amsterdam UMC, locatie AMC - Klinische Informatiekunde
E-mail: d.sent@amc.uva.nl
Location of Scientific Research Project KPMG - Health IT Audit & Assurance Laan van Langerhuize 1
1186 DS Amstelveen &
Amsterdam UMC, locatie AMC - Intensive Care Department Meibergdreef 9
1105 AZ Amsterdam
BackgroundIntensive Care Unit (ICU) physicians are responsible for multiple deci-sions while treating their patients. The decision to withdraw treatment is often de-layed, resulting in overtreatment, which attenuates the patient throughput and in-creases the total costs of the ICU. Cognitive biases affect the decision making process of physicians, which is a possible cause of overtreatment. The aim of this study is to provide a broad insight into the existence of a cognitive bias, called the sunk-cost ef-fect, in the decision making process of ICU physicians. Method The combination of a literature review and a case study provided insight into whether or not the sunk-cost effect is present in the decision making process of ICU physicians. In order to obtain a better understanding of the sunk-cost effect and the decision making process of physi-cians in general, multiple factors were identified in literature. With the use of real anonymized patient cases from three Dutch ICU’s, we observed whether the primary ICU physician was affected by the sunk-cost effect, based on the response of the par-ticipating ICU physicians. Additionally, we identified additional factors considered in the decision making process of ICU physicians. Results Using real anonymized pa-tient cases, we were able to show that the primary ICU physicians were affected by the sunk-cost effect in their decision making process, as the ICU physicians participating in this study withdrew the treatment sooner. An ICU physician’s age and experience are found to be related to the ability to avoid the sunk-cost effect. The results of this study indicate that both clinical and non-clinical factors are considered by ICU physicians in their decision making process. Clinical factors are considered during both decisions about continuation or about withdrawal of a treatment, whereas non-clinical factors are mostly considered when deciding about continuation of a treatment. Conclusion This study strongly suggests that ICU physicians taking care of critically ill patients are affected by the sunk-cost effect in their decision making process. Although ICU physi-cians who are not personally responsible for a patient’s treatment are less susceptible to the sunk-cost effect, primary ICU physicians are affected by this cognitive bias. The involvement of a more experienced ICU physician may decrease the occurrence of the sunk-cost effect. Factors that influence the decision making process of ICU physicians are of both clinical and non-clinical nature. More research is needed to confirm that the sunk-cost effect is present at the ICU and to investigate whether or not non-clinical factors influence the occurrence of the sunk-cost effect.
Keywords: Sunk-cost effect, Intensive Care Unit, Cognitive bias, Decision making
Achtergrond Intensive Care (IC) artsen zijn verantwoordelijk voor beslissingen ron-dom de behandeling van ernstig zieke pati¨enten. De beslissing om een behandeling te staken wordt vaak uitgesteld, wat resulteert in overbehandeling. Mogelijke oorzaken van overbehandeling zijn cognitieve biases, waardoor het beslissingsproces van artsen wordt be¨ınvloed. Het doel van deze studie is om inzicht te krijgen in het bestaan van een specifieke cognitieve bias, genaamd het verzonken kosten effect, in het besliss-ingsproces van IC artsen. Methode De combinatie van een literatuur review en een multicenter observatie casestudie heeft het mogelijk gemaakt om inzicht te krijgen in de eventuele aanwezigheid van het verzonken kosten effect in het beslissingsproces van IC artsen. Om het verzonken kosten effect en het algemene beslissingsproces van art-sen beter te begrijpen hebben we factoren ge¨ıdentificeerd in de literatuur. Met behulp van echte geanonimiseerde pati¨entcasussen van drie Nederlandse IC’s hebben we on-derzocht of verantwoordelijke IC artsen worden be¨ınvloed door het verzonken kosten effect, gebaseerd op het antwoord van de deelnemende IC artsen. Daarbij hebben we aanvullende factoren uit het beslissingsprocess van artsen ge¨ıdentificeerd. Resultaten Door het gebruik van echte geanonimiseerde pati¨entcasussen hebben wij aangetoond dat verantwoordelijke IC artsen worden be¨ınvloed door het verzonken kosten effect in hun beslissingsproces. De IC artsen die de casussen hebben ingevuld, staakten de be-handeling eerder dan de verantwoordelijke IC artsen. De leeftijd en ervaring van de IC artsen zijn gerelateerd aan de bekwaamheid van de IC artsen om het verzonken kosten effect te voorkomen. Klinische factoren worden overwogen in beslissingen zowel ron-dom het voortzetten als het staken van de behandeling, terwijl niet-klinische factoren het meest worden overwogen tijdens de beslissing om een behandeling voort te zetten. Conclusie Deze studie heeft sterke aanwijzingen voor het feit dat IC artsen worden be¨ınvloed door het verzonken kosten effect in hun beslissingsproces. Hoewel IC artsen die niet persoonlijk verantwoordelijk zijn voor de behandeling van een pati¨ent minder vatbaar zijn voor het effect, toont deze studie aan dat verantwoordelijke IC artsen door het verzonken kosten effect worden be¨ınvloed. De betrokkenheid van een ervaren IC arts zou het optreden van het verzonken kosten effect kunnen verminderen. Zowel klinische als non-klinische factoren be¨ınvloeden het beslissingsproces van IC artsen. Toekomstig onderzoek is nodig om te bevestigen of het verzonken kosten effect in-derdaad aanwezig is in het beslissingsproces van IC artsen en om te onderzoeken of niet-klinische factoren het voorkomen van het verzonken kosten effect be¨ınvloeden.
Sleutel-woorden: Verzonken kosten effect, Intensive Care, Cognitieve bias,
This thesis is the final assignment of the master program Medical Informatics of the University of Amsterdam (UvA). A combination of a literature study and a multicenter observational case study in three Dutch University hospitals made it possible to obtain a broad view of the existence of the sunk-cost effect in the decision making process of ICU physicians. This research was performed from November 2017 to August 2018, partially at the Intensive Care Unit (ICU) at the Amsterdam UMC and partially at the KPMG office in Amstelveen.
I would like to thank my supervisors Danielle Sent and Dave Dongelmans. I am very thankful for the opportunity to perform this research. Over the past three years, we collaborated a lot with each other for which I am very grateful . It all started with my bachelor thesis in 2016 and I am very pleased to finalize my master thesis under your supervision. Every time we discussed my research, you were able to provide me with new insights and opportunities to improve my thesis. Your feedback has been very valuable, both for the content of this thesis and my skills in performing a research. I would also like to thank my supervisor at KPMG, Eva Tsjapanova. In spite of your busy schedule, you were able to support me along the way and provide me with valu-able feedback. Due to my curiosity about the business life, I am very thankful for the fact that KPMG provided me with a work place to write my thesis. Of course, I would also like to thank my colleagues at KPMG for all the conviviality and support.
Furthermore, a special thanks my friends - especially Rianne - who provided me with useful tips and feedback in the last couple of months of writing my thesis. Your help really improved my writing skills, and I am very grateful for that.
Last but not least, a special thanks to my parents, brother, friends, and fellow students. Your continuous support really helped me to finalize my master.
Thank you all
Nena Dekema
Abstract iii Samenvatting iv Preface v Abbreviations ix 1 General introduction 1 1.1 Objectives . . . 3 1.2 Outline of thesis . . . 5 Bibliography . . . 6 2 Preliminaries 9 2.1 Sunk-cost effect . . . 9
2.2 Intensive Care Unit . . . 10
2.3 Scoring systems . . . 11
Bibliography . . . 12
3 Factors influencing the decision making process of physicians - a literature review 15 Abstract . . . 15
3.1 Introduction . . . 16
3.2 Method . . . 17
3.2.1 Literature search 1a/1b: sunk-cost effect . . . 17
3.2.2 Literature search 2: general decision making process of physicians 18 3.3 Results . . . 19
3.3.1 Literature search 1a/1b: sunk-cost effect . . . 19
3.3.2 Literature search 2: general decision making process of physicians 23 3.4 Discussion . . . 29
Bibliography . . . 32
4 Sunk-cost effect in Dutch Intensive Care Units - A multicenter observational study 39 Abstract . . . 39
4.1 Introduction . . . 40
4.2 Method . . . 41 vii
4.3 Results . . . 45
4.4 Discussion . . . 53
Bibliography . . . 56
5 Identification of factors considered by Intensive Care Unit physicians during treatment decisions 59 Abstract . . . 59
5.1 Introduction . . . 60
5.2 Method . . . 61
5.2.1 Factors: sunk-cost effect . . . 61
5.2.2 Factors: general decision making process . . . 62
5.2.3 Factors identified at to specific decision moment of ICU physician 63 5.3 Results . . . 64
5.3.1 Factors: sunk-cost effect . . . 64
5.3.2 Factors: general decision making process . . . 65
5.3.3 Factors identified at to specific decision moment of ICU physician 66 5.4 Discussion . . . 68
Bibliography . . . 72
6 Discussion and future work 75 Bibliography . . . 78
Bibliography 79
Appendix 1: Search strategy - sunk-cost effect 79 Appendix 2: Search strategy - general decision making process 81 Appendix 3: Study characteristics - sunk-cost effect articles 83 Appendix 4: Study characteristics - general decision making articles 87 Appendix 5: Basal information of patients cases 93
APACHE Acute Physiology and Chronic Health Evaluation EHR Electronic Health Record
ICU Intensive Care Unit QOL Quality of Life
SAPS Simplified Acute Physiology SCORE SOFA Sequential Organ Failure Assessment
General introduction
The Intensive Care Unit (ICU) is one of the most complex and expensive departments of a hospital. Worldwide, the total ICU expenses are responsible for 8 to 30% of the total hospital expenses [1]. The large dispersion can be attributed to the number of hospital beds available in each country and other differences in the healthcare system. The mean hospitalization costs of the ICU are 2.5 times higher than those of the regu-lar wards, due to the availability of expensive resources and highly qualified personnel [2-4]. Because of the constant throughput of patients, the availability of the beds at the ICU has become a complex issue. It is relatively easy to determine if someone is in need of intensive care treatment and should thus be admitted to the ICU. In contrast, it is much harder to reach agreement on the moment of withdrawal of intensive care treat-ment. Nevertheless, these decisions about withdrawing life sustaining treatment are common at the ICU, since 15 to 25% of the patients will die while being admitted to the ICU [5]. The decision of withdrawing treatment is challenging [6], for example in cases where the patient is young and has children or in cases where the patient’s family does not agree with the physician’s decision about whether or not further treatment should be provided [7]. As a consequence, physicians tend to overrate the chance on survival of a patient, resulting in overtreatment [8]. Overtreatment is defined as providing un-necessary care, a cause of preventable harm, and waste of resources in healthcare [9]. In 2010, the Institute of Medicine (IOC) brought this problem under attention, indicat-ing that overtreatment is “the largest contributor to waste in United States (US) health care, accounting for approximately $210 billion of the estimated $750 billion in excess spending each year” [3]. In addition to unnecessary costs, overtreatment contributes to more ICU beds being occupied, possibly resulting in missed opportunities to provide care to patients who are in need of intensive care treatment.
Apart from decisions about withdrawing treatment, ICU physicians are responsible for many other decisions concerning the treatment of critically ill patients. Observa-tion and examinaObserva-tion of a patient is the first step in the decision making process of physicians. After this, the physician can retrieve data about the patient from the Elec-tronic Health Record (EHR). This data can be in the form of past issues concerning the patient described in the EHR or in the form of information on vital signs. While be-ing admitted, multiple devices continuously monitor, measure, and influence the vital functioning of organ systems of patients. These data are stored in the EHR. After ex-amining the data, the physician deliberates the patient’s situation with colleague ICU physicians and if necessary, physicians from other disciplines, to be able to make the right treatment decision. The patient and the patient’s family are explicitly involved in this decision making process, whereas the physician takes the preferences of all the involved parties into consideration. By combining all these sources of information and by using evidence-based medicine, physicians are able to propose a treatment plan. Evidence-based medicine is defined as “the integration of best research evidence with clinical expertise and patient values” [10].
When taking into account that the decision making process is based on medical evi-dence and influence of other involved parties, physicians are faced with multiple chal-lenges. One of these challenges is to keep up with the rapidly growing amount of new insights on diseases and treatments from medical literature. Physicians of all specialisms are challenged by integrating medical evidence into their daily decision making [11-17]. Another challenge concerns cognitive biases, to which physicians are exposed in their decision making process. Cognitive biases are defined as errors in the process of thinking, whereby individuals perceive information based on their own preferences and experiences [18]. Examples of cognitive biases researched in the medi-cal domain are: confirmation bias [19-21] - bias of preferring information confirmed by predefined beliefs, loss aversion bias [22, 23] - bias of preferring to avoid losses in or-der to obtain gains, omission bias [24-26] - bias of preferring losses caused by omission over identical or lesser losses caused by action, and availability bias [19, 27] - bias of preferring recent information and events , which were personally observed and there-fore more memorable. In this study we focus on a cognitive bias that has not often been researched in the medical domain: ‘the sunk-cost effect’. The sunk-cost effect is defined as “the tendency to continue an endeavor once an investment in money, effort, or time has been made” [28]. Current decisions are influenced by previous investments, whereas decisions should be rationalized by only expected future outcomes influenc-ing the decision to continue a certain action [29]. In contrast to previous studies in healthcare [30, 31], we hypothesize that the sunk-cost effect is present in the decision making of ICU physicians. We suspect that, although a substantial amount of evidence
indicates that the patient will not survive, some physicians are inclined to continue the treatment due to already invested effort and resources in the care process of the patient.
In this study, we combine a literature review with a case study to provide a broad insight in whether or not the sunk-cost effect is present in the decision making pro-cess of ICU physicians. In order to obtain a better understanding of the sunk-cost effect and the decision making process of physicians in general, factors contributing to the sunk-cost effect and factors associated with the general decision making pro-cess of physicians were identified in literature. Within the case study, real anonymized patient cases from three Dutch Intensive Care Departments were used to observe the occurrence of the sunk-cost effect. Patients with complex life-threatening conditions are admitted to University Intensive Care Departments. Approximately 10% of the pa-tients spend more than seven days at the ICU, due to the complexity of their condition [32]. This group of prolonged stay patients accounts for 40 to 50% of the total ICU costs [32]. We hypothesize that the sunk-cost effect is present at these ICU’s due to the characteristics of patient population, the long hospitalization period with complex life threatening conditions. Creating awareness concerning the sunk-cost effect will be the first step in preventing overtreatment due to this bias at the ICU, which would result in avoiding unnecessary healthcare cost. However, in particular the creation of awareness about the sunk-cost effect will contribute to a more rational decision making through which unnecessary care is eliminated and resources become available for other patients who are in need of intensive care treatment.
1.1
Objectives
The main purpose of this study is to investigate the sunk-cost effect in decision making of ICU physicians. Therefore we have divided our resource in three parts. Firstly, a literature search was conducted to identify factors associated with the sunk-cost effect and factors associated with the decision making process of physicians in general. Sec-ondly, we observed whether or not ICU physicians are affected by the sunk-cost effect in their decision making process by using real anonymized patient cases from three Dutch University hospitals. We hypothesize that physicians do suffer from the sunk-cost ef-fect as they let previously invested efforts influence their decision about continuation or withdrawal of a treatment. Thirdly, the presence of the factors associated with the sunk-cost effect and the general decision making process identified within the litera-ture search were investigated by the ICU physicians in practice.
The combination of a literature review and a case study in three Dutch University Hos-pitals provided a broad insight in the existence of the sunk-cost effect in the decision making process of ICU physicians. The research questions related to the study objec-tives are represented in figure 1.1.
1.2
Outline of thesis
This thesis consists of seven chapters, opening with a general introduction and closing with a general discussion and conclusion. After the introduction, we continue with pre-liminary information in Chapter 2, in order to specify the main concepts used within this study. Chapter 3 provides the results of the literature search on factors associated with the sunk-cost effect and the decision making process of physicians in general. Chapter 4 is dedicated to the sunk-cost effect, with the use of real anonymized patient cases to observe whether ICU physicians are affected by the sunk-cost effect in their de-cision making process. Whether or not the factors, identified in Chapter 3, are present in the decision making process of ICU physicians is presented in Chapter 5. We end this thesis by providing an general discussion and recommendation for future work in Chapter 6.
Bibliography
[1] Martin, C. M., Hill, A. D., Burns, K., & Chen, L. M. (2005). Characteristics and outcomes for critically ill patients with prolonged intensive care unit stays. Critical care medicine, 33(9), 1922-1927.
[2] Chalfin, D. B., Cohen, I. L., & Lambrinos, J. (1995). The economics and cost-effectiveness of critical care medicine. Intensive care medicine, 21(11), 952-961.
[3] Halpern, N. A., & Pastores, S. M. (2010). Critical care medicine in the United States 2000-2005: an analysis of bed numbers, occupancy rates, payer mix, and costs.Critical care medicine, 38(1), 65-71.
[4] M.L. Barrett, M.W. Smith, A. Elixhauser, et al.Utilization of Intensive Care Services, Accessed 20th June 2018, Retrieved from: http://www.hcup-us.ahrq.gov/reports/ statbriefs/sb185-Hospital-Intensive-Care-Units-2011.pdf (2011),
[5] Connolly, C., Miskolci, O., Phelan, D., & Buggy, D. J. (2016). End-of-life in the ICU: moving from ‘withdrawal of care’to a palliative care, patient-centred approach.British journal of anaesthesia, 117(2), 143-145.
[6] Munday, D., Petrova, M., & Dale, J. (2009). Exploring preferences for place of death with terminally ill patients: qualitative study of experiences of general practitioners and community nurses in England.Bmj, 339, b2391.
[7] Wilkinson, D., & Savulescu, J. (2011). Knowing when to stop: futility in the inten-sive care unit.Current Opinion in Anaesthesiology, 24(2), 160-165.
[8] Luce, J. M. (2010). End-of-life decision making in the intensive care unit.American journal of respiratory and critical care medicine, 182(1), 6-11.
[9] Ho, A. (2008). Relational autonomy or undue pressure? Family’s role in medical decisionmaking.Scandinavian journal of caring sciences, 22(1), 128-135.
[10] Sackett, D. L. (1997). Evidence-based Medicine How to practice and teach EBM. WB
Saunders Company.
[11] Casscells, W., Schoenberger, A., & Graboys, T. B. (1978). Interpretation by physi-cians of clinical laboratory results.New England Journal of Medicine, 299(18), 999-1001.
[12] Lenfant, C. (2003). Clinical research to clinical practicelost in translation?, New England Journal of Medicine, 349(9), 868-874.
[13] McAlister, F. A., Straus, S. E., Guyatt, G. H., Haynes, R. B., & Evidence-Based Medicine Working Group. (2000). Users’ guides to the medical literature: XX. Inte-grating research evidence with the care of the individual patient.Jama, 283(21),
2829-2836.
[14] Likosky, D. S. (2006). Integrating evidence-based perfusion into practices: the International Consortium for Evidence-Based Perfusion. The journal of extra-corporeal technology, 38(4), 297.
[15] Cook, D. J., & Giacomini, M. K. (2002). The integration of evidence based medicine and health services research in the ICU.Evaluating Critical Care (pp. 185-197). Springer,
Berlin, Heidelberg.
[16] Greenhalgh, T., Howick, J., & Maskrey, N. (2014). Evidence based medicine: a movement in crisis?.Bmj, 348, g3725.
[17] Clark, A. M., & Findlay, I. N. (2005). Improving evidence based cardiac care and policy implementation over the patient journey: the potential of coronary heart disease registers.
[18] Buss, D. M. (Ed.). (2005). The handbook of evolutionary psychology. John Wiley & Sons
[19] Klein JG. Five pitfalls in decisions about diagnosis and prescribing. BMJ: British Medical Journal. 2005;330(7494):781-783.
[20] Mendel, R., Traut-Mattausch, E., Jonas, E., Leucht, S., Kane, J. M., Maino, K., ... & Hamann, J. (2011). Confirmation bias: Why psychiatrists stick to wrong preliminary diagnoses.Psychological Medicine, 41(12), 2651-2659.
[21] Pines, J. M. (2006). Profiles in patient safety: confirmation bias in emergency medicine.Academic Emergency Medicine, 13(1), 90-94.
[22] McNeil, B. J., Pauker, S. G., Sox Jr, H. C., & Tversky, A. (1982). 23 On the Elicitation of Preferences for Alternative Therapies. Preference, Belief, and Similarity, 583.
[23] Rizzo, J. A., & Zeckhauser, R. J. (2003). Reference incomes, loss aversion, and physician behavior.Review of Economics and Statistics, 85 (4), 909-922.
[24] DiBonaventura, M. D., & Chapman, G. B. (2008). Do decision biases predict bad decisions? Omission bias, naturalness bias, and influenza vaccination.Medical Decision Making, 28(4), 532-539.
[25] Aberegg, S. K., Haponik, E. F., & Terry, P. B. (2005). Omission bias and decision making in pulmonary and critical care medicine.Chest, 128(3), 1497-1505.
[26] Chapman, G. B., & Elstein, A. S. (2000). Cognitive processes and biases in medical decision making. Decision making in health care: Theory, psychology, and applications,
183-210.
[27] Dale, W., Hemmerich, J., Ghini, E. A., & Schwarze, M. L. (2006). Can induced anx-iety from a negative earlier experience influence vascular surgeons’ statistical decision-making? A randomized field experiment with an abdominal aortic aneurysm analog.
Journal of the American College of Surgeons, 203(5), 642-652.
[28] Arkes, H. R., & Blumer, C. (1985). The psychology of sunk cost. Organizational behavior and human decision processes, 35(1), 124-140.
[29] Karlsson, N., Juliusson, ., & Grling, T. (2005). A conceptualisation of task di-mensions affecting escalation of commitment. European Journal of Cognitive Psychology,
17(6), 835-858.
[30] Braverman JA, Blumenthal-Barby JS. Assessment of the sunk-cost effect in clinical decision-making.Soc Sci Med. 2012 Jul;75(1):186-92.
[31] Bornstein BH, Emler AC, Chapman GB. Rationality in medical treatment deci-sions: is there a sunk-cost effect? Soc Sci Med. 1999 Jul;49(2):215-22.
[32] Carden, G. P., Graham, J. W., McLennan, S., & Celi, L. A. (2008). Long-term out-come of long stay ICU and HDU patients in a New Zealand hospital. Critical care and shock, 11(1), 26.
Preliminaries
This chapter provides preliminary information on the sunk-cost effect. Accordingly, a brief explanation of the Intensive Care Unit (ICU) is provided, followed by the diverse patient scoring systems used at this department. Elucidation of these concepts con-tributes to understanding the importance of investigating the sunk-cost effect at the ICU.
2.1
Sunk-cost e
ffect
Cognitive biases are defined as errors in the process of thinking, whereby people per-ceive information based on their own preferences and experiences [1]. The sunk-cost effect is one of these cognitive biases and is manifested in “a greater tendency to con-tinue an endeavor once an investment in money, effort, or time has been made” [2]. Synonyms of this effect used within literature are: ‘the sunk-cost fallacy’, ‘sunk-cost bias’ and the ‘Concorde fallacy’. An example of the sunk-cost effect from Thaler (1980) is presented below [3]:
“A family pays $40 for tickets to a basketball game to be played 60 miles from their home. On the day of the game there is a snowstorm. They decide to go anyway, but note in passing that had the tickets been given to them, they would have stayed home.”
The familys decision to visit the basketball game is influenced by their previous in-vestment, even though they know that their decision is not pragmatic because of the snowstorm. When making a decision, one should only consider incremental costs and benefits, instead of historical costs, which cannot be repaired.
Originally, the sunk-cost effect is an economic phenomenon, that is also frequently studied by behavioral scientists. In the sociological field, the terms ‘escalation of com-mitment and ‘irrational escalation, are used to study the same behavior. ‘Escalation of commitment refers to the tendency of decision makers to persist with failing courses of action [4]. “Individuals frequently choose to continue a course of action or invest ad-ditional resources despite a guaranteed negative outcome and extensive sunk-cost”[5]. Some might imply that sunk-cost causes escalation of commitment, whereas others are persuaded that the sunk-cost effect and escalation of commitment are identical. Within this study, we assume that the sunk-cost effect and the escalation of commitment are similar, aligned with other studies [6-9]. In the subsequent sections, we will only refer to the sunk-cost effect. The sunk-cost effect is often related to the fact that people at-tempt to avoid wasting resources, whereby they pursue a course of action even though it is not pragmatic. Most frequently mentioned theories explaining the sunk-cost ef-fect are the prospect theory, cognitive dissonance and the self-justification theory. The prospect theory suggests that decisions are made based on the prospects of that de-cision, regardless of the final outcome [3, 10-11]. Cognitive dissonance implies that someone experiences conflicting behaviors, beliefs and attitudes when making a de-cision [5, 11-13]. The self-justification theory explains that people tend to pursue a course of action to affirm that their original decision was right, because they refuse to admit that they were wrong or might have made a mistake [5, 11, 14].
2.2
Intensive Care Unit
The ICU belongs to the most complex and expensive departments of a hospital, due to the fact that the right amount of qualified personnel and the availability of expen-sive resources and equipment are essential to provide for high qualitative care [15]. The complexity of the ICU is caused by both a wide diversity of patients with life-threatening conditions, and a high throughput of these patients. Patients who are no longer in need of lifesaving therapy are discharged to another department to make sure that beds at the ICU are available for patients who are in need of critical care. As a re-sult of this complexity, special trained nurses and physicians need to be available 24/7. Patients are supported by several devices that continuously monitor, measure and in-fluence the vital functioning of organ systems of patients. These devices generate large amounts of data, stored in the Electronic Health Record (EHR) of the patient. Addi-tionally, physicians are able to manually enter information, to ensure the integrity of the medical record. In addition to monitoring the patients progress, data from the EHR are also used to predict outcomes concerning the patient, e.g. the severity of illness.
2.3
Scoring systems
Multiple scoring systems are used within the ICU to determine the severity of ill-ness of a patient. Most of the scoring systems use data from the first ICU day, ex-amples of these scoring systems are: APACHE (Acute Physiology and Chronic Health Evaluation), SAPS (Simplified Acute Physiology Score) and MPM (Mortality Prediction Model). Other scoring systems include data from the complete duration of hospitaliza-tion at the ICU, whereby insight in the progress of the severity of the illness is obtained. Examples of scoring systems that calculate the score throughout ICU stay are: SOFA (Sequential Organ Failure Assessment), MODS (Multiple Organs Dysfunction Score), ODIN (Organ Dysfunction and Infection System) and LOD (Logistic Organ Dysfunc-tion) [16]. Some of the scoring systems are subjective, whereby the expert (i.e. physi-cian) determines the value for each variable. Other scoring systems are more objective, whereas logistic regression modeling techniques combined with clinical judgements determine the value of a variable.
We included both the APACHE and the SOFA score in this study. Since the beginning of the APACHE II score in 1985, there have been some extensions of the score, resulting in an APACHE III and APACHE IV score. The APACHE II score is calculated during the first 24 hours of a patients ICU stay, and is based on age, previous health status and 12 physiologic measurements. The result of this calculation is an integer score between 0 and 71. A high APACHE II score indicates a severe illness with a high risk of death [17]. APACHE III is the extended version of the APACHE II score, where variables were added and the collection of the data is different. The APACHE III score varies between 0 and 299 [18]. Again, some variables were added to the APACHE III score, resulting in the APACHE IV score [19]. It is up to the hospital to decide which version they use. In this study we mainly use the APACHE II score and occasionally the APACHE IV score, in case the hospital did not record the APACHE II score.
The SOFA score quantifies a patients organ dysfunction and is calculated every day throughout an ICU stay. The score is based on the degree of organ dysfunction of the following six organ systems: cardiovascular, respiratory, neurological, renal, hepatic, and coagulation [20]. Each organ system receives a score varying from 0 to 4, where 0 implies non-failure and 4 implies failure. As a result, the mortality prediction can be received by a combination of failure of the organ systems [15]. The total SOFA score is an integer score between 0 and 24.
Bibliography
[1] Buss, D. M. (Ed.). (2005). The handbook of evolutionary psychology. John Wiley &
Sons.
[2] Arkes, H. R., & Blumer, C. (1985). The psychology of sunk cost. Organizational behavior and human decision processes, 35(1), 124-140.
[3] Thaler, R. (1980). Toward a positive theory of consumer choice.Journal of Economic Behavior & Organization, 1(1), 39-60.
[4] Brockner, J. (1992). The escalation of commitment to a failing course of action: Toward theoretical progress.Academy of management Review, 17(1), 39-61.
[5] Braxton, C. C., Robinson, C. N., & Awad, S. S. (2017). Escalation of Commitment in the Surgical ICU.Critical care medicine, 45(4), e433-e436.
[6] Carnevale, J. J., Inbar, Y., & Lerner, J. S. (2011). Individual differences in need for cognition and decision-making competence among leaders. Personality and Individual Differences, 51(3), 274-278.
[7] Bornstein, B. H., & Chapman, G. B. (1995). Learning lessons from sunk costs. Jour-nal of Experimental Psychology: Applied, 1(4), 251.
[8] Braverman, J. A., & Blumenthal-Barby, J. S. (2012). Assessment of the sunk-cost effect in clinical decision-making. Social Science & Medicine, 75(1), 186-192.
[9] Bornstein BH, Emler AC, Chapman GB. Rationality in medical treatment decisions: is there a sunk-cost effect? Soc Sci Med. 1999 Jul;49(2):215-22.
[10] Zeelenberg, M., & Van Dijk, E. (1997). A reverse sunk cost effect in risky deci-sion making: Sometimes we have too much invested to gamble. Journal of Economic Psychology, 18(6), 677-691.
[11] Chulkov, D. V. (2017). Escalation of commitment and information security: theo-ries and implications. Information & Computer Security, 25(5), 580-592.
[12] Festinger, L. (1962). A theory of cognitive dissonance (Vol. 2). Stanford university
press.
[13] Chung, S. H., & Cheng, K. C. (2018). How does cognitive dissonance influence the sunk cost effect?. Psychology research and behavior management, 11, 37.
[14] Mannion, R., & Thompson, C. (2014). Systematic biases in group decision-making: implications for patient safety. International Journal for Quality in Health Care, 26(6),
606-612.
[15] Seidel, J., Whiting, P. C., & Edbrooke, D. L. (2006). The costs of intensive care.
Continuing Education in Anaesthesia Critical Care & Pain, 6(4), 160-163.
[16] Rapsang, A. G., & Shyam, D. C. (2014). Scoring systems in the intensive care unit: A compendium. Indian Journal of Critical Care Medicine: Peer-Reviewed, Official
Publication of Indian Society of Critical Care Medicine, 18(4), 220-228.
[17] Knaus, W. A., Draper, E. A., Wagner, D. P., & Zimmerman, J. E. (1985). APACHE II: a severity of disease classification system.Critical care medicine, 13(10), 818-829.].
[18] Knaus, W. A., Wagner, D. P., Draper, E. A., Zimmerman, J. E., Bergner, M., Bastos, P. G., ... & Harrell, F. E. (1991). The APACHE III prognostic system: risk prediction of hospital mortality for critically III hospitalized adults.Chest, 100(6), 1619-1636.].
[19] Zimmerman, J. E., Kramer, A. A., McNair, D. S., Malila, F. M., & Shaffer, V. L. (2006). Intensive care unit length of stay: Benchmarking based on Acute Physiology and Chronic Health Evaluation (APACHE) IV.Critical care medicine, 34(10), 2517-2529.
[20] Vincent, J. L., Moreno, R., Takala, J., Willatts, S., De Mendona, A., Bruining, H., ... & Thijs, L. G. (1996). The SOFA (Sepsis-related Organ Failure Assessment) score to describe organ dysfunction/failure
Factors influencing the decision
making process of physicians - a
literature review
Abstract
BackgroundPhysicians are challenged in their clinical decision making. Cognitive biases affect the decision making process of physicians, which is a possible cause of overtreatment. In this study we focus on one cognitive bias in particular: the sunk-cost effect. In order to understand where the sunk-cost effect arises from in clinical decision making, we aim to identify factors associated with the sunk-cost effect and factors associated with the general decision making process of physicians. Method We performed a literature search in Scholar to identify factors associated with the sunk-cost effect, in both the general non-medical and medical domain. In order to obtain an overview of factors associated with the general decision making process, a second literature search has been performed in Pubmed, Embase, PsychINFO, and Medline. All identified factors were presented based on their frequency, to get insight in the most im-portant factors. The general decision making factors were divided in patient, physician and environment related factors. Results The search for factors associated with the sunk-cost effect yielded 71 articles, of which we included 17 in our study. The search for factors associated with the general decision making process of physicians yielded 453 articles, of which we included 29 in our study. Conclusion Although factors related to the sunk-cost effect are limited, it is suggested that individuals who are personally responsible for a decision, tend to be more influ-enced by the sunk-cost effect than people who are not personally responsible. An individuals’ experience, emotions and their desire not to waste are also found to be risk factors for the occur-rence of the sunk-cost effect. A variety of factors associated with the general decision making process of physicians were found, which can be divided in internal and external factors. Pa-tients’ preferences, age, influences of family and friends, and physicians’ experience were the most identified factors associated with the general decision making process of physicians.
3.1
Introduction
Clinical decision making is one of the main ongoing activities in healthcare [1]. Clinical decision making is defined as “the process of making an informed judgement about the treatment necessary for patients” [2]. The process of clinical decision making involves multiple steps in which the patient is the point of focus [3]. A combination of physical examination, clinical experience, electronic information resources, colleague opinions, and a patient’s preferences are considered by physicians when making a decision [4]. In addition, the preferences of surrogate decision makers, such as a patient’s family or friends, are often considered when patients are not able to communicate their own preferences [5]. Research indicates that the decision making process of physicians is influenced by multiple factors, both clinical and non-clinical [6, 7].
In an ideal world, the only factors contributing to a physicians’ decision about a pa-tient’s treatment are based upon medical evidence, including medical guidelines, med-ical education, experience achieved over the years, and complementary training. Un-fortunately, this ideal world is not the real world we are currently living in. On the daily basis, physicians are faced with multiple challenges influencing their decision making process, varying from pressure of patients and/or relatives [8-10] and fear of malpractice [10], to cognitive biases [11-13]. These challenges cause overtreatment [10], diagnostic errors [14], and missed opportunities to provide the best available care to a patient.
In this study, we focus on a specific cognitive bias called the sunk-cost effect, as pre-viously described in Chapter 1. The sunk-cost effect may be present in the decision making process, which leads to irrational decisions. In healthcare, the sunk-cost effect can contribute to overtreatment, whereby unnecessary care is provided to patients. In order to be able to avoid the sunk-cost effect in healthcare, we need to understand where the sunk-cost effect arises from in the decision making process of physicians. Therefore, this study focusses on both the sunk-cost effect and the decision making process of physicians in general. Firstly, we will obtain an overview of factors associ-ated with the sunk-cost effect, by performing a literature search in general non-medical and medical domain. Secondly, we will perform an additional literature search to ob-tain an overview of factors associated with the decision making process of physicians in general. The combination of both factors associated with the sunk-cost effect and fac-tors associated with the general decision making process of physicians will contribute in understanding where the sunk-cost effect arises from in medical domain. Research suggests that the sunk-cost effect can be avoided by creating awareness of the effect [15, 16]. Therefore, our aim is to make physicians aware of factors influencing the sunk-cost effect, whereupon the occurrence of the sunk-cost effect might decrease.
3.2
Method
We conducted a literature study to identify factors influencing the potential occur-rence of the sunk-cost effect during a decision making process, both in the general non-medical as well as in the medical domain. Due to limited literature on factors associated with the sunk-cost effect, a second search was performed to identify gen-eral factors associated with the decision making process of physicians. Identification of these factors contributes to understanding the occurrence of the sunk-cost effect in medical decision making.
3.2.1 Literature search 1a/1b: sunk-cost effect
Two literature searches were performed to identify factors associated with the sunk-cost effect in both general non-medical domain (search 1a) as well as in the medical domain (search 1b). The Scopus database was used to search for articles containing fac-tors about the sunk-cost effect in relation to the decision making process. This database is comprised of multiple clinical databases. Because our subject has not often been re-searched, this facilitates the finding of useful articles. Various studies focusing on the sunk-cost effect use rats and pigeons. We specified the exclusion of these studies in the search strategy, because articles investigating the sunk-cost effect with rats and pi-geons are not relevant for our study. A combination of the following keywords was used: (“Sunk-cost effect” OR “Escalation of Commitment”) AND “Clinical Decision
Mak-ing” AND “Factor*” NOT (“Rat” OR “Pigeon”), to search for articles containing factors
in the medical domain. General non-medical articles were searched using a combina-tion of resembling keywords. Only the term“Clinical Decision Making”, was replaced
by“Decision Making”. The complete search strategy for both general non-medical and
medical domain is presented in Appendix 1.
The searches were limited to articles written in English and no restrictions for the year of publication were defined. Additionally, the reference list of all included articles was screened in order to identify additional articles. The articles were screened on their title and abstract. Afterwards, full-text screening was performed on articles that were potentially relevant, based on pre-defined inclusion and exclusion criteria. There was only one inclusion criterion: 1) factors influencing the sunk-cost effect in decision mak-ing. The exclusion criteria were: 1) sunk-cost causing escalation of commitment, 2) factors reducing the sunk-cost effect and/or de-escalation of commitment, and 3) fac-tors improving the decision making process. An overview of the facfac-tors influencing the sunk-cost effect was made. The included articles were verified by two indepen-dent reviewers (A.V., D.D) to increase screening reliability and to ensure there were
no duplicates or there was no ambiguity among the factors. Consensus was reached by discussing the results in detail. Study characteristics (first author, year of publica-tion, title, journal, country, study design, (period), domain, (subject) and population size) were documented to provide a complete overview of the articles used within this study.
3.2.2 Literature search 2: general decision making process of physicians
The search criteria for identifying factors associated with the general decision mak-ing process of physicians included a combination of the followmak-ing keywords: (“Clini-cal Decision Making” AND “Physician” AND “Treatment” AND “Factor*”) NOT “Shared Decision Making” NOT “Primary Health Care” NOT “Psychiatry” NOT “General Practi-tioner”. The complete version of the search strategy can be found in Appendix 2. Four
electronic databases were utilized for this search: Pubmed, Embase, PsychINFO, and Medline. Similar to literature search 1a and 1b, the searches were limited to articles written in English, and no restrictions for the year of publication were defined. Ad-ditionally, the reference list of all included articles was screened in order to identify additional articles. The articles were screened based on their title and abstract. We proceeded to read the full version of the potentially relevant articles, based on pre-defined inclusion and exclusion criteria. The following inclusion criteria were applied: 1) the article contains factors influencing the general decision making process of physi-cians, and 2) the factors are restricted to the decision making process of physicians working in secondary care. The following exclusion criteria were applied: 1) factors that are present in the decision making process of a particular disease, 2) factors from the psychiatry department, due to differences in communication between the physi-cian and psychiatric patients in relation to other patients, 3) factors influencing the general decision making process of patients, 4) factors influencing the general decision making process of nurses and/or other healthcare workers other than physicians, and 5) factors facilitating shared decision making between physician and patient. For the remaining articles we identified the factors and categorized them into three groups: 1) patient related factors, 2) physician related factors, and 3) environment related factors, as proposed by Hajjaj et al. (2010) [3].
To reiterate the process of search 1a and 1b, the included articles were verified by two independent reviewers (A.V., D.D) to increase screening reliability and to ensure there were no duplicates or there was no ambiguity among the factors. Consensus was reached by discussing the results in detail. Study characteristics (first author, year of publication, title, journal, country, study design, (period), domain, (subject) and
population size) were documented to provide a complete overview of the articles used within this study.
3.3
Results
3.3.1 Literature search 1a/1b: sunk-cost effect
Two searches were performed to identify factors influencing the potential occurrence of the sunk-cost effect, in both general non-medical (study 1a) and medical domain (study 1b). In the general non-medical domain, we identified 63 studies of which 12 studies met our criteria and were subsequently included (Figure 3.1). Of the 8 studies identified within the medical domain, 4 studies met our criteria and were thus included in the study (Figure 3.2). Appendix 3 summarizes the study characteristics of the in-cluded articles.
Figure 3.1:Results of search strategy 1a - sunk-cost effect (general non-medical domain).
Figure 3.2:Results of search strategy 1b - sunk-cost effect (medical domain).
In total, 20 unique factors were identified in both general non-medical and medical domain. Table 3.1 provides an overview of the factors associated with the sunk-cost effect, found in both general non-medical and medical domain. Of these factors, 7 (35.0%) were exclusively present in the general non-medical domain and 5 (25.0%) were exclusively present in the medical domain. Figure 3.3 shows the frequency of each factor found in both general non-medical and medical domain. Responsibility and emotional values were identified the most in both general non-medical and medical domain, in 5 and in 5 articles, respectively.
Figure 3.3: Frequency of each factor associated with the sunk-cost effect, found in both general non-medical and medical domain.
Table 3.1: Factors associated with the sunk-cost effect, found in both general non-medical and non-medical domain. A short description of each factor is presented, followed
General non-medical domain
The most frequently identified factors in general non-medical domain were emotions, desire do not to waste, responsibility, disappointment because of negative feedback, cultural differences and age. Emotions were identified in 4 articles [21-23, 29]. Dif-ferent emotions, such as anger and fear, are associated with the sunk-cost effect. The desire not to waste was mentioned in relation to the sunk-cost effect and identified in 2 articles [25, 27]. In 2 articles, responsibility was identified [21, 33]. Being person-ally responsible for a decision influences the occurrence of the sunk-cost effect. People tend to be less influenced by the sunk-cost effect when someone else is responsible for the decision. Disappointment because of negative feedback was identified in 2 articles [30-31]. Some people are disappointed after receiving negative feedback. Literature indicates that these people are more likely to be affected by the sunk-cost effect than people who do not experience disappointment. Cultural differences were identified in 2 articles, suggesting that culture is associated with the sunk-cost effect [27, 33]. Chi-nese people are more likely to be affected by the sunk-cost effect than ChiChi-nese people [27]. Age was identified in 2 articles [23-24]. People with a higher age are less likely to be influenced by the sunk-cost effect.
Medical domain
The most frequently mentioned factors identified in medical domain were responsi-bility, number of years of practice, desire do not to waste, consistency and cognitive dissonance. Responsibility was identified in 3 articles [18-20]. People who make per-sonal decisions tend to be more affected by the sunk-cost effect than when they make decisions on behalf of others. Number of years of practice, i.e. the experience of the de-cision maker, was identified in 3 articles [17, 19-20]. People with more experience are less likely to be affected by the sunk-cost effect, since they are familiar with situations and know how to act. The desire not to waste was mentioned in relation to the sunk-cost effect and identified in 2 articles [17, 19]. People consider previous investments in something as a waste, when they decide to deviate of the original plan. Therefore, peo-ple are tended to continue as planned, which results in the sunk-cost effect. Cognitive dissonance was also identified in 2 articles [18, 20]. Cognitive dissonance explains the phenomenon of people experiencing conflicting behaviors, beliefs and attitudes when making a decision. When someone has to choose between two conflicting options, they are likely to be affected by the sunk-cost effect.
3.3.2 Literature search 2: general decision making process of physicians
Factors associated with the general decision making process of physicians were searched for in a second literature search. Of the 453 initial studies identified, 29 studies were in compliance with our criteria and included in this study (Figure 3.4). Appendix 4 sum-marizes the study characteristics of the included articles. An overview of the identified factors categorized into three groups: 1) patient related factors, 2) physician related factors, and 3) environment related factors is presented in Table 3.2/3.3. Figure 3.5, 3.6, and 3.7 show the frequency in which each factor is found. Of all the factors iden-tified, the experience of the physicians is mentioned the most (51.7%). Three factors are outstanding when considering factors related to the patient: 1) influences of family and friends, 2) the patients’ age, and 3) the patients’ preferences, mentioned in 41.4%, 37.9%, and 37.9% of the articles, respectively. Only a few factors are related to the clinical environment, with the characteristics of the hospital (31.0%) mentioned the most.
Figure 3.4:Results of search strategy 2 - factors associated with the general decision making process of physicians.
Table 3.2: P atien t, ph ysician, and en vironmen t rela ted factors associa ted with the g ener al decision making process of ph ysicians. A short description of each factor is presen ted, foll ow ed by the ref erences of the articles in which the factor w as found.
Table 3.3: P atien t, ph ysician, and en vironmen t rela ted factors associa ted with the g ener al decision making process of ph ysicians. A short descrip-tion of each factor is presen ted, foll ow ed by the ref erences of the articles in which the factor w as found.
Patient related factors
The most frequently identified patient related factors were influences of family and friends, age, and patient preferences. Influences of family and friends were identi-fied in 12 articles [34-36, 38, 41, 46-47, 52-54, 59, 62]. Family and friends are able to provide the physician with information about the patient’s opinion and preferences concerning the treatment of the patient. They know the patient, whereupon they can provide the physician with information. The age of the patient was identified in 11 ar-ticles [34, 38, 40, 44, 46, 48, 54-55, 57, 59, 62]. Physicians tend to consider a relatively young age to be a reason to put more effort in a treatment. The patient’s preferences were identified in 11 articles [34-36, 38, 44-45, 47-48, 58-59, 62]. The patient is in-volved in the decision making process concerning his treatment. Physicians consider the patient’s preferences when making a decision.
Figure 3.5:Frequency of factors concerning the patient, which are associated with the general decision making process of physicians.
Physician related factors
Experience, influences of colleagues and the physician’s ethnicity were the most fre-quently identified physician related factors. Experience was identified in 15 articles [37-38, 42-44, 46-51, 56-57, 59, 62]. Physicians use their experience, when making a decision about a patient. The influences of colleagues is identified in 9 articles [34-35, 38, 48, 50, 52, 56, 62]. Multiple physicians are involved in the treatment process of a patient. When making a decision, the influences of colleagues are often considered by the physician. The ethnicity of the physician was identified in 6 articles [39, 42, 47-48, 52, 57]. Research indicates that different decisions are made by physicians with different ethnicities.
Figure 3.6: Frequency of factors concerning the physician, which are associated with the general decision making process of physicians.
Environment related factors
The most frequently identified factors related to environment were hospital character-istics, resource availability, and guidelines. The hospital characteristics were identified in 9 articles [34, 38, 41-42, 48-49, 52, 55, 62]. The hospital characteristics comprise all characteristics of the hospital or healthcare center, from the amount of physicians to the involved specialties. Resource availability was identified in 6 articles [34-35, 37, 55-56, 62]. Different resources are needed when providing care to patients. The avail-ability of resources is considered when making a decision. When certain resources are not available, the physician will decide to start an alternative treatment option. Guidelines were identified in 6 articles [43-44, 46, 50, 58-59]. Guidelines provide de-tailed descriptions about different treatment options for a disease. Physicians use these guidelines when making a decision.
Figure 3.7: Frequency of factors concerning the environment, which are associated with the general decision making process of physicians.
3.4
Discussion
A literature review identified factors associated with the sunk-cost effect - the tendency to continue an action due to previous investments which cannot be repaired. Factors of both general non-medical and medical domain were found and ranked based on their frequency. The feeling to be personally responsible for a decision, emotions such as anger and fear, the desire not to be wasteful, and (work) experience are associated with the sunk-cost effect. In an additional literature review, factors associated with the general decision making process of physicians were identified. These factors can either be patient, physician, or environment related. When making a decision, physicians consider patient related factors, such as influences of family and friends, a patient’s age and a patient’s preferences. Factors related to the physician are also involved in their decision making process, and include the physician’s experience and influences of colleagues.
Sunk-cost effect
Individuals who are personally responsible for a decision tend to be more influenced by the sunk-cost effect than individuals who are not personally responsible [18-21, 33]. The findings of a study where no relation between personal responsibility and the sunk-cost effect were identified can be disputed, since business students partici-pating in this study had been taught how to avoid the sunk-cost effect [63]. Another study supports this finding by reporting that the sunk-cost effect is not dependent of the person who makes the decision [64]. More research is needed to investigate the impact of personal responsibility in relation to the sunk-cost effect, in order to support our finding that personal responsibility is augmenting the sunk-cost effect. In the gen-eral non-medical domain, the sunk-cost effect is caused by individuals who experience emotions, such as anger and fear [20-23, 29]. Emotions affect an individual’s way of thinking, whereby they cannot think properly [65-66]. Fear contributes to the feeling of being uncertain, and anger causes individuals to take more risks. Future research should focus on investigating other emotions, to see whether other emotions are also risk factors for the sunk-cost effect. In the medical domain, we found that more experi-enced individuals are less likely to be influexperi-enced by the sunk-cost effect [17, 19-20, 67] As was found by two authors, we also found that team decision making may serve as a way to decrease the sunk-cost effect [30, 33]. Experienced individuals can share their knowledge and collaborate with less experienced individuals. As a result, individuals are no longer personally responsible for a decision. Future research should focus on the impact of team decision making in relation to the sunk-cost effect. It was found in multiple studies that individuals often have a desire to avoid wasting resources when making a decision, which enables the sunk-cost effect [17, 19, 25, 27]. We propose to
teach individuals about the existence of the sunk-cost effect and factors contributing to the effect, as it has been found by other authors that this will decrease the occurrence of the sunk-cost effect [15-16].
General decision making process
Factors related to the patient, the physician, and the environment are associated with the general decision making process of physicians. As we found multiple factors for each group, we decided arguably to only discuss the factors we considered to be impor-tant. We propose to view these factors from the internal and external factors perspec-tive. It is an omission that we have not done this from the start of the study. Internal factors can be changed by the decision maker, whereas external factors are beyond a person’s reach to change. The distinction makes that for purposes of influencing the possible bias it is logical to focus on internal factors, although knowing the external factors could help in preventing the bias. The distinction between internal and exter-nal factors is sometimes disputable. Therefore, we provide some examples from our study. Internal patient factors, such as a patient’s preferences, can change based on the current healthcare conditions. External patient factors, such as a patient’s age and the influence of family and friends, are factors a patient is not able to influence. It should be noted that in our study, the influence of family and friends is identified more fre-quently than a patient’s preferences. This does not imply that a physician considers the influences of family and friends more often than he considers a patient’s prefer-ences. At some departments, the influence of family and friends is common due to the fact that patients are too ill to communicate their own preferences, for example at the ICU. Factors related to the physician are most often external factors, e.g. colleagues’ opinions, which are often considered in a physician’s decision making process [34-35, 38, 48, 50, 52, 56, 62]. The physician’s experience is an example of an internal factor [37-38, 42-44, 46-51, 56-57, 59, 62].
The combination of factors associated with the sunk-cost effect and factors associated with the general decision making process of physicians enabled us to get more insight in a major part of factors that are considered when making a decision. This enables fur-ther investigation of these factors when investigating the sunk-cost effect in the ICU in the next study (Chapter 4 and 5). Another strength of our study concerns the inclusion of up-to-date articles. Recent published articles were found in the literature search and included in the study respectively. Although we were able to include up-to-date arti-cles, we should have used a more specific scope for the search strategy for identifying factors associated in the general decision making process of physicians. Though when investigating the sunk-cost effect in literature, the broad scope of sunk-cost helped us to identify the limited factors available in relation to this effect. A minority of factors
was found in multiple articles, which in our view increases their reliability. On the con-trary, the majority of factors was only found in one article, which in our view decreases the reliability of these factors. We have to acknowledge that probably not all factors were found during our search. For the factors associated with the sunk-cost effect, this can be explained by the limited amount of literature that is available about the sunk-cost effect. The broad search used to identify general decision making factors might have resulted in generic results, due to which we missed some important factors. Some of the identified factors had a similar meaning. We therefore combined these factors and provided them with a clear description. This process of merging factors with sim-ilar meaning was challenging. Some factors were not merged because of differences in context. The factor ‘disappointment because of negative feedback’ was separated from the factor ‘emotions’, since emotions were not related to receiving negative feedback. It might be the case that some factors were merged even though they should have been presented separately, or vice-versa.
Altogether, this research identified factors associated with the sunk-cost effect and tors associated with the general decision making process of physicians. Although fac-tors related to the sunk-cost effect are limited, it is suggested that individuals who are personally responsible for a decision, tend to be more influenced by the sunk-cost effect than people who are not personally responsible. An individual’s experience, emotions, and desire not to waste are also found to be factors for the occurrence of the sunk-cost effect. In addition, there is a variety of factors associated with the general decision making process of physicians found in literature that can be divided in internal and ex-ternal factors. Physicians’ decisions are affected by both patient and physician related factors, as well as environmental factors. In the following chapters, the application of the factors discussed in this chapter in relation to the sunk-cost effect is investigated.
Bibliography
[1] Schwartz, A., & Bergus, G. (2008). Medical decision making: a physician’s guide.
Cambridge University Press.
[2] Hardy, D., & Smith, B. (2008). Decision making in clinical practice. British Journal of Anaesthetic & Recovery Nursing, 9(1), 19-21.
[3] Hajjaj, F., Salek, M., Basra, M., & Finlay, A. (2010). Non-clinical influences on clini-cal decision-making: a major challenge to evidence-based practice.Journal of the Royal Society of Medicine, 103(5), 178-187.
[4] Hay, M. C., Weisner, T. S., Subramanian, S., Duan, N., Niedzinski, E. J., & Kravitz, R. L. (2008). Harnessing experience: exploring the gap between evidencebased medicine and clinical practice.Journal of evaluation in clinical practice, 14(5), 707-713.
[5] Schenker, Y., Crowley-Matoka, M., Dohan, D., Tiver, G. A., Arnold, R. M., & White, D. B. (2012). I don’t want to be the one saying ‘we should just let him die’: intraper-sonal tensions experienced by surrogate decision makers in the ICU.Journal of General Internal Medicine, 27(12), 1657-1665.
[6] McKinlay, J. B., Potter, D. A., & Feldman, H. A. (1996). Non-medical influences on medical decision-making.Social science & medicine, 42(5), 769-776.
[7] Eisenberg, J. M. (1979). Sociologic influences on decision-making by clinicians.
Annals of Internal Medicine, 90(6), 957-964.
[8] Strumio, J., Chlabicz, S., Pytel-Krolczuk, B., Marcinowicz, L., Rogowska-Szadkowska, D., & Milewska, A. J. (2016). Combined assessment of clinical and patient factors on doctors’ decisions to prescribe antibiotics.BMC family practice, 17(1), 63.
[9] Ho, A. (2008). Relational autonomy or undue pressure? Family’s role in medical decisionmaking.Scandinavian journal of caring sciences, 22(1), 128-135.
[10] Lyu, H., Xu, T., Brotman, D., Mayer-Blackwell, B., Cooper, M., Daniel, M., ... & Makary, M. A. (2017). Overtreatment in the United States. PloS one, 12(9), e0181970.
[11] Saposnik, G., Redelmeier, D., Ruff, C. C., & Tobler, P. N. (2016). Cognitive biases associated with medical decisions: a systematic review. BMC medical informatics and decision making, 16(1), 138.
[12] Hussain, A., & Oestreicher, J. (2018). Clinical decision-making: heuristics and cognitive biases for the ophthalmologist.Survey of ophthalmology, 63(1), 119-124.
[13] Groopman, J. E., & Prichard, M. (2007). How doctors think (Vol. 82). Boston:
Houghton Mifflin.
[14] National Academies of Sciences, Engineering, and Medicine. (2016). Improving diagnosis in health care. National Academies Press.
[15] Larrick, R. P., Morgan, J. N., & Nisbett, R. E. (1990). Teaching the use of cost-benefit reasoning in everyday life.Psychological Science, 1(6), 362-370.
[16] Larrick, R. P., Morgan, J. N., & Nisbett, R. E. (1990). Teaching the use of cost-benefit reasoning in everyday life.Psychological Science, 1(6), 362-370.
[17] Braverman, J. A., & Blumenthal-Barby, J. S. (2012). Assessment of the sunk-cost effect in clinical decision-making. Social Science & Medicine, 75(1), 186-192.
[18] Mannion, R., & Thompson, C. (2014). Systematic biases in group decision-making: implications for patient safety. International Journal for Quality in Health Care, 26(6),
606-612.
[19] Bornstein, B. H., Emler, A. C., & Chapman, G. B. (1999). Rationality in medical treatment decisions: is there a sunk-cost effect?. Social science & medicine, 49(2), 215-222.
[20] Braxton, C. C., Robinson, C. N., & Awad, S. S. (2017). Escalation of Commitment in the Surgical ICU.Critical care medicine, 45(4), e433-e436.
[21] Wong, K. F. E., & Kwong, J. Y. (2007). The role of anticipated regret in escalation of commitment.Journal of Applied Psychology, 92(2), 545.
[22] Tsai, M. H., & Young, M. J. (2010). Anger, fear, and escalation of commitment.
Cognition and Emotion, 24(6), 962-973.
[23] Jarmolowicz, D. P., Bickel, W. K., Sofis, M. J., Hatz, L. E., & Mueller, E. T. (2016). Sunk costs, psychological symptomology, and help seeking.SpringerPlus, 5(1), 1699.
[24] Strough, J., Bruine de Bruin, W., Parker, A. M., Karns, T., Lemaster, P., Pichayayothin, N., ... & Stoiko, R. (2016). What were they thinking? Reducing sunk-cost bias in a life-span sample. Psychology and aging, 31(7), 724.
[25] Chulkov, D. V. (2017). Escalation of commitment and information security: theo-ries and implications. Information & Computer Security, 25(5), 580-592.
[26] Carnevale, J. J., Inbar, Y., & Lerner, J. S. (2011). Individual differences in need for cognition and decision-making competence among leaders. Personality and Individual Differences, 51(3), 274-278.
[27] Liang, B., Kale, S. H., & Cherian, J. (2014). Is the future static or dynamic? The role of culture on escalation of commitment in new product development. Industrial Marketing Management, 43(1), 155-163.
[28] Schott, J. P., Scherer, L. D., & Lambert, A. J. (2011). Casualties of war and sunk costs: Implications for attitude change and persuasion. Journal of Experimental Social Psychology, 47(6), 1134-1145.
[29] Coleman, M. D. (2010). Sunk cost, emotion, and commitment to education. Cur-rent Psychology, 29(4), 346-356.
[30] Rausch, A., Wall, F., & Mdritscher, G. (2012). Sunk costs and the need for justifica-tion: an experimental study on de-escalation.Problems and Perspectives in Management, 10(2), 93-106.
[31] Jones, P. E., & Roelofsma, P. H. (2000). The potential for social contextual and group biases in team decision-making: Biases, conditions and psychological mecha-nisms.Ergonomics, 43(8), 1129-1152.
[32] Roetzel, P. G. (2015, May). The Impact of Creativity and Information Load on Escalation of Commitment. InECIS.
[33] Drummond, H. (1994). Too little too late: A case study of escalation in decision making.Organization Studies, 15(4), 591-607.
[34] Aita, K., Takahashi, M., Miyata, H., Kai, I., & Finucane, T. E. (2007). Physicians’ attitudes about artificial feeding in older patients with severe cognitive impairment in Japan: a qualitative study.BMC geriatrics, 7(1), 22.
[35] Ates, F., Zor, M., Ylmaz, O., Tuncer, M., Ozturk, M., Gurbuz, C., ... & Kucuk, E. V. (2016). Management behaviors of the urology practitioners to the small lower calyceal stones: the results of a web-based survey.Urolithiasis, 44(3), 277-281.
[36] Bluhm, M. V. (2011). Factors Influencing Oncologists’ Use of Chemotherapy in Patients at the End of Life: A Qualitative Study.
[37] Branche, A. R., Walsh, E. E., Jadhav, N., Karmally, R., Baran, A., Peterson, D. R., & Falsey, A. R. (2016). Provider decisions to treat respiratory illnesses with antibiotics: insights from a randomized controlled trial.PloS one, 11(4), e0152986.
[38] Carmel, S., Werner, P., & Ziedenberg, H. (2007). Physicians’ and nurses’ prefer-ences in using life-sustaining treatments.Nursing ethics, 14(5), 665-674.
