At this stage, we searched for relevant studies in seven electronic databases, including IranMedex, Scientific Information Database (SID), MagIran, Science Direct, PubMed, Google Scholar, and ProQuest. Our search was restricted to papers published in English and Persian languages after 1990. The search terms were defined based on the research question and the inclusion criteria. We used a combination of keywords and Boolean operators to retrieve relevant studies from the database. The keywords were “nurse” AND “uncertainty” AND “decision-making” AND “instrument” OR “tool” OR “scale” OR “questionnaire”. We also used truncation () and wildcard (?) symbols to expand the search results. For example, we searched for “nurs” to include “nurse”, “nurses”, “nursing”, etc. and “decisio?” to include “decision” and “decisions”. We also applied filters, such as publication date, language, and type of publication to narrow down the search results. Our search strategy yielded a total of 137 published papers.

In this step, we screened the retrieved studies for inclusion criteria using a four-step process. First, we excluded 50 duplicates. Second, we screened the abstracts of 87 articles and excluded 47 articles that did not meet our criteria (Table 1). At this stage, we reviewed only the abstracts of the articles without considering the full texts. We identified 47 articles that did not meet the appropriate criteria at this stage of abstract review. For example, some of them focused on the uncertainty of patients or diseases, the degree of uncertainty in nursing students, etc. Third, we screened the full texts of the remaining articles (n=40) and included 12 articles with accessible full texts (Fig. 1). Fourth, we assessed the quality of the included articles using two quality criteria instruments: one for empirical sources and one for theoretical sources.

Fig. (1). PRISMA flowchart for searching and selecting the final studies.

At this stage, we analyzed and synthesized the data from the included articles using five steps: data display, data comparison, data reduction, drawing the conclusion, and verification [9]. We used tables and charts to display the data and compare the similarities and differences among the studies. We reduced the data by grouping them into themes and categories based on our research question. We concluded by summarizing the main findings and implications of the studies. We verified our conclusion by checking its validity and reliability with other sources.

Despite the increasing number of studies on physicians’ clinical uncertainty, limited studies are available on nurses’ clinical uncertainty [3]. The IUS has been used in some Iranian and international studies on nurses [10, 18]. Zahrayi et al. performed a review study in Iran to determine the psychometric properties of the IUS. In this study, the intolerance of uncertainty was defined as an intrinsic feature caused by negative beliefs about uncertainty and its consequences, such as worry and the need for confidence and control. The authors emphasized that intolerance of uncertainty is an essential predictor of OCD, worry, anxiety, and eating disorders [25]. Nonetheless, no specific instrument is available for measuring uncertainty in nurses’ clinical decision-making.

A precise definition of uncertainty is necessary for designing an uncertainty measurement instrument. According to the literature, ambiguity [4, 30], lack of confidence and control [28], complexity [26, 29], and being on the horns of a dilemma [12] are the most critical characteristics of uncertainty. However, the unanswered question is whether these characteristics adequately measure nurses' uncertainty in clinical decision-making. A review study on nurses' uncertainty in decision-making reported that although nurses may face uncertainty in many situations, the probability of uncertainty increases when there is a need for interpretation or intervention in case of emergencies. The authors also emphasized that detecting and controlling an uncertain situation are essential skills for nurses to make effective clinical decisions [1]. Furthermore, this review concluded that an uncertain situation could result in cognitive, emotional, and behavioral reactions. However, nurses often use problem-solving skills to better understand the situation when facing uncertain situations. In addition to using their knowledge and experience, nurses also consult with experienced colleagues to deal with uncertain situations. Nevertheless, most studies on uncertainty have been conducted on physicians or patients, and there are few relevant studies on nurses. Therefore, further studies are needed to fill this gap [1].

In the process of patient care, clinical decision-making is the outcome of critical thinking, clinical reasoning, and clinical judgment [45]. Appropriate instruments are needed to assess and detect nurses' uncertainty, especially in the stage of clinical judgment. However, no specific instrument is available in this regard, and the existing instruments are general and do not measure all dimensions of uncertainty in nursing. Therefore, they are inappropriate for assessing uncertainty in clinical nursing practice [3].

The existing instruments measure situations that cause uncertainty, people’s feelings in dealing with uncertainty, and adaptive responses in an uncertain situation [15]. Cranley investigated the strategies used by ICU nurses in dealing with uncertainty. The study reported that coping with uncertainty includes three stages, namely, identifying uncertainty and its emotional and physiological responses (e.g., stress), management of the situation (e.g., receiving help from colleagues), and outcomes (e.g., learning and gaining experience from new situations). Cranley concluded that nurses use cognitive-behavioral, emotional, and behavioral strategies to adapt to an uncertain situation [4]. Given the acceptable reliability of the specific ICU nurses’ uncertainty measurement tool [15], this tool can be adapted and used to measure the nurses' responses to uncertainty. However, it is essential to design culture-specific instruments for measuring nurses' uncertainty in clinical decision-making. Furthermore, if researchers intend to consider the weight of clinical judgment in the degree of uncertainty of nurses when making clinical decisions, specific tools should be designed for this purpose. In this respect, qualitative methods, such as content analysis for finding the meanings of concepts and processes and reducing the challenges of facing contradictory theories, would be suitable [46]. Qualitative methods allow researchers to interpret the originality and truth of data objectively and scientifically. They also allow to reveal the hidden themes and patterns [47]. Investigating nurses’ lived experiences of uncertainty in clinical decision-making helps researchers discover the characteristics of uncertainty in clinical decision-making and design an instrument based on the discovered characteristics.

Given that a person constantly understands and processes evidence at any given time, the state of uncertainty is dynamic [28]. 'Dynamic' refers to the interaction between antecedents over time [3]. Uncertainty is also a discomforting situation [28, 41] because it disrupts a person's daily life [28].

The experience of uncertainty is pervasive in human experience and is influenced by a sense of self-confidence and control; it can be very specific (event-focused) or more general, as well as non-normative and detrimental to human development [28]. The three attributes of uncertainty include probability, temporality, and perception, which may be explicit or implicit in uncertainty. All the attributes of uncertainty overlap, and they all affect each other [31]. Feelings of confidence and a sense of control are the main factors that determine the nature of the experience of uncertainty [28].

In the health system, uncertainty in decision-making can include diagnostic, prognostic, treatment, interventional [16], technical, personal, and conceptual aspects [32, 39]. Regarding the uncertainty in intervention decisions, we can mention the uncertainty of nurses whether to request emergency care for their patients [37], and nurses' uncertainty is rooted in their uncertainty about their interpretation and choice of interventions [1]. In this regard, it has been observed that nurses may face uncertainty in decision-making during all clinical decision-making stages (review, use of intuition and intervention, etc.). However, it seems that most cases occur during the interpretation of findings [1]. Nurses make better decisions with structured information (e.g., diagnostic procedures) than with unstructured details [48]. Prognostic uncertainty can be due to the unpredictability of future events or the limited accuracy of probability estimation [40]. The source of technical uncertainty can be related to the information of the doctor or nurse and can also be associated with the rapid growth of health system knowledge. Individual uncertainty arises from the relationship between the patient and the physician or nurse. Conceptual uncertainty means not being aware of the patient's competitive needs for shared resources [32, 39].

According to the classification for nurses' uncertainty, micro-uncertainty refers to a person's degree of confidence in their decisions. Due to the different decisions of different people, macro-uncertainty is observed in the decisions made by individuals [33].

Nurses care for patients with different personal characteristics who respond differently to nursing interventions, and this can cause uncertainty in them. On this account, nurses face a significant challenge when using judgment and decision-making skills; whatever decision they make, they must use the available evidence to think about the uncertain future. Despite their uncertainty, they must decide what to do [16]. A correct understanding of nurses' uncertainty can lead to strategies that support nurses in their clinical decisions [1]. Therefore, it seems necessary to conduct specific studies on the concept of uncertainty in decision-making in nursing.

The Intolerance of Uncertainty Scale (IUS), developed by Freeston et al., is a self-report instrument designed to understand why people worry. In our review, all the participants in the included studies were undergraduate psychology students with an average age of 22 years [27].

Validity: The methodological quality of content validity and hypothesis testing was “very good”. The IUS showed acceptable convergent and divergent validity when used concurrently with the Worry Domain Questionnaire (WDQ) (r = 0.57) [49], Beck’s Anxiety Inventory (BAI) (r = 0.57) [50], Beck’s Depression Inventory (BDI) (r = 0.52) [51], Questionnaire on Generalized Anxiety Disorder-Modified (QGAD-M) [52], and Penn State Worry Questionnaire (PSWQ) (r = 0.63) [53]. Both construct validity and criterion validity were identified as “very good”. Moreover, the measure of sampling adequacy for the intercorrelation matrix was 0.80, indicating that factor analysis was appropriate.

Reliability: The IUS also showed acceptable reliability with an internal consistency of 0.91. Based on the IUS and the Worry Scale responses, people can be categorized into three groups: those who meet somatic criteria, those who meet both somatic and cognitive criteria, and those who meet neither cognitive nor somatic criteria [27].

The English version of the IUS, developed by Buhr et al., has four subscales. This scale reflects the idea that uncertainty is stressful and upsetting, uncertain situations are negative and should be avoided, being uncertain is unfair, and uncertainty results in failure to act [11]. The English version was tested and validated in a sample of students just admitted to various undergraduate programs.

Validity: This scale has four factors, and its scores have an acceptable correlation with the scores of PSWQ [53], Worry and Anxiety Questionnaire (WAQ) [54], BDI-II [55], and BAI [50]. The correlation coefficients showed that the English version of IUS is more correlated with worry criteria (r = 0.60) and less with anxiety and depression criteria, confirming its convergent and divergent validity.

Reliability: The English version of the IUS had acceptable internal consistency (Cronbach’s alpha = 0.94) and test-retest reliability (r = 0.74). This scale can differentiate between people who meet the criteria for generalized anxiety disorder (GAD), people who meet only the somatic criteria, and those who meet none of the GAD criteria. It has been shown that while people who meet the criteria for GAD have the highest scores in the English version of IUS, those who meet none of the criteria for GAD have the lowest scores [11].

Norton assessed the psychometric properties of IUS to ascertain if the structure of the instrument differs among four racial (cultural) non-clinical African-American, Caucasian, Spanish/Latin, and Southeast Asian groups. The participants were bachelor’s students of Houston University, and their mean age was 21.65 ±4.93 (MD) years [19].

Validity: The factor structure of each non-clinical group indicated a poor and unstable representation. The factor analysis has shown the five- or six-factor solutions. The IUS had a high correlation with the PSWQ and GADQ-IV [19].

Reliability: The overall internal consistency of the IUS was good (α=0.945). Moreover, the alpha coefficients of the African (α =0.951), Caucasian (α =0.942), Spanish (α =0.934), and Southeast Asian (α =0.954) participants were similar [19].

Carleton et al. developed a short version of IUS, known as IUS-12 [13]. They reduced the 27 items of the original IUS [27] to 12 items and tested the new scale on two academic groups. This version of IUS-12 has two subscales, including prospective anxiety and inhibitory anxiety.

Validity: The IUS-12 showed acceptable convergent validity when it was administered concurrently with BDI-II (r = 0.56) [55], BAI (r = 0.57) [50], PSWQ (r = 0.54) [53], and GADQ-IV (r = 0.61) [56].

Reliability: A significant association was observed between the scores obtained from IUS-27 and IUS-12 (r = 0.96), indicating the closeness of the results of these two scales; this confirms the validity of IUS-12. Furthermore, IUS-12 had a high internal consistency (α = 0.91), which was considered “acceptable”, and the highest inter-item correlations were low to moderate (between 0.27 and 0.65) [13].

Gosselin et al. (2008) also developed a self-report Intolerance of Uncertainty Inventory (IUI) to measure uncertainty intolerance. The initial instrument had 72 items, which were reduced to 45 items via factor analysis. The IUI has two parts. The first part includes three subscales: intolerance of uncertainty and uncertain situation, intolerance of unexpectedness, and difficulty waiting in an uncertain situation. The second part consists of six subscales: overestimation of the probability of an adverse event, control, reassurance-seeking, avoidance, worry, and doubt [35].

Validity: The content validity analysis was conducted by researchers and two experienced individuals. Fifteen university students also confirmed its comprehensibility. The convergent validity of the IUI was also confirmed through concurrent administration with the IUS [27], PSWQ [53], WAQ [54], and Padua Inventory [57].

Reliability: The instrument's stability and internal consistency were assessed and confirmed through test-retest (a five-week interval) and Cronbach's alpha coefficient. All scores of the IUI were correlated with the IUS (developed by Freeston) and had a moderate and high correlation with the scores resulting from the validated measures of GAD and obsessive-compulsive disorder (OCD). Besides, individuals who met the GAD criteria on the WAQ questionnaire scored higher on the IUI. Overall, study results supported the temporal stability of the IUI after a 5-week interval.

Vadivel et al. conducted a study on Iranian students to assess the psychometric properties of the Persian version of the IU-12; the age range was 18-24 in this study [21].

Validity: The values of the content validity index (CVI) and the content validity ratio (CVR) were over 0.7 and 0.78, respectively. In this study, the results of previous studies on the PSWQ and the Depression Anxiety Stress Scale (DASS-21) were assessed along with the Persian version to check the validity. Two subscales of the translated IUS-12 had direct relationships with worry (r=0.75), stress (r=0.61), anxiety (r=0.61), and depression (r=54) [21].

Reliability: The values of CR and Cronbach’s alpha were 0.86 and 0.89, respectively, and the internal consistency was good [21].