Abstract
Objective
This study investigates decision-making processes among healthcare managers at Kawsar Hospital in Sanandaj, Iran, using a mixed-methods approach that integrates qualitative insights with data-driven analysis.
Methods
Semistructured interviews with 15 managers were analyzed through the Critical Incident Technique and the NatCen Framework, revealing distinct patterns in unstructured, multilevel decisions.
Results
Organizational knowledge played a central role in the initial stages of decision making, while external research became more influential in later phases. A dataset derived from these interviews was examined using Random Forest and K-means clustering. The machine learning results highlighted time pressure as the most influential predictor of decision quality, whereas managerial experience contributed only marginally. Clustering analysis further revealed heterogeneous managerial profiles, reflecting differences in experience, decision speed, and contextual stressors. Additionally, Response Surface Methodology (DOE/RSM) underscored the importance of information sources in optimizing decisions under uncertainty. All participants expressed satisfaction with their decisions, emphasizing the adaptive and naturalistic nature of their decision-making styles.
Conclusion
These findings, while exploratory given the limited sample size, provide a practical framework for enhancing healthcare management by integrating qualitative evidence with machine learning and experimental design. Future studies with larger datasets and expanded coding (e.g. a decision types table) are recommended to strengthen the generalizability and applications of this research.
Keywords
Highlights
Organizational knowledge is crucial in the initial stages of decision making, while external researches have a greater impact in the later stages.
Managerial experience and time pressure are key predictors of decision making.
The Response Surface Methodology (RSM) used in the experiments highlighted the significance of sources in improving decisions made under uncertainty.
Participants expressed satisfaction with their decisions and acknowledged the adaptive and natural nature of their decision-making styles. These styles integrate qualitative evidence with machine learning and experimental design.
Introduction
Effective decision making lies at the heart of successful healthcare management, as hospitals operate in dynamic environments characterized by uncertainty, limited resources, and high-stakes outcomes. Managers must continuously evaluate options, allocate resources, and respond to both routine and emergent challenges, all while maintaining quality of care and operational efficiency. In this context, decision making is not merely a technical process but a strategic function that directly influences patient outcomes, staff performance, and institutional sustainability. As healthcare systems increasingly adopt data-driven and evidence-based approaches, the ability of managers to integrate information from diverse sources—ranging from clinical data to organizational knowledge—becomes essential. Thus, understanding and improving decision-making processes is a critical step toward strengthening hospital performance and advancing public health goals.1,2
Kawsar Hospital, a prominent public healthcare institution in Sanandaj, Kurdistan Province, operates within a complex environment marked by evolving patient needs, limited resources, and policy uncertainty. 3 Healthcare managers at this facility routinely face unstructured, high-pressure decisions that require navigating across strategic, tactical, and operational levels. Despite the critical nature of their roles, there is limited empirical research on how such managers seek and use information when making nonclinical decisions. Prior studies have primarily focused on clinical decision making or policymaking, leaving a gap in understanding the behavioral and contextual factors influencing administrative healthcare decisions. 4 Furthermore, the literature has seldom explored the integration of qualitative insights with computational methods such as machine learning (ML) or experimental design to support managerial decision making. This study addresses these gaps by combining interview-based analysis with data-driven techniques to provide a more holistic understanding of the decision-making landscape at Kawsar Hospital. 5
Despite the recognized importance of informed decision making in healthcare management, existing literature reveals critical gaps in understanding how healthcare managers—particularly at the hospital level—access, interpret, and utilize information. Studies show that managerial decisions often neglect research-based evidence, relying instead on intuition, past experiences, or organizational routines.6,7 While several frameworks exist for clinical decisions and public health policies, the domain of administrative, nonclinical decision making—especially under uncertainty—remains underexplored. Moreover, few empirical investigations have examined when and why managers seek external versus internal sources of information, or how time pressure and organizational hierarchy influence their cognitive strategies. These knowledge gaps hinder the development of systematic support tools and training programs tailored to the realities of managerial work in hospitals. This study addresses these deficiencies by integrating qualitative exploration of decision episodes with ML and experimental design to identify data-driven patterns and actionable insights. 8
This study is grounded in several foundational theories of decision making, notably Herbert Simon's distinction between structured and unstructured decisions. Structured decisions follow predefined procedures and are often routine in nature, while unstructured decisions occur in novel, complex situations lacking clear protocols—characteristics typical of many managerial challenges in healthcare settings.9,10 The research also draws on Mintzberg's managerial role framework, which categorizes managerial behavior into interpersonal, informational, and decisional roles. These roles influence how managers gather and process information under time constraints and hierarchical pressures. In addition, the concept of naturalistic decision making (NDM)—where decisions are shaped by experience, context, and urgency rather than formal analysis—provides a practical lens through which hospital decision making can be interpreted. 11 Finally, this study recognizes the multilevel nature of decisions—operational, tactical, and strategic—each of which carries different information needs and consequences. These theoretical foundations inform the study's analytical approach, ensuring a holistic understanding of how information, context, and experience interact in real-world hospital decision making.12–21
In recent years, ML has become an essential tool in healthcare research, providing advanced capabilities for pattern recognition and decision support. In this study, we apply Random Forest classification to analyze how variables such as managerial experience, time pressure, and information source are associated with decision quality/relate to decision quality. 12 The model reveals the relative importance of contextual and cognitive factors, with time pressure emerging as the most influential predictor of decision quality, while managerial experience showed only a marginal contribution. Additionally, K-means clustering was applied to identify distinct decision-making styles among managers, offering a data-driven segmentation of cognitive and contextual behaviors. These ML methods complement qualitative insights, enabling the identification of hidden patterns that traditional analyses may overlook. This hybrid approach, while exploratory given the limited dataset, provides a useful framework for understanding and enhancing managerial decision making in hospital environments. 22
To further explore the interaction of decision-making variables under uncertainty, the study employs Design of Experiments (DOE) using the Taguchi method within Design Expert 11 software. DOE enables a systematic investigation of multiple variables simultaneously, allowing the identification of optimal conditions that maximize decision quality. By analyzing 15 experimental runs derived from interviews, the Central Composite Design (CCD) orthogonal array helps assess the influence of decision level, decision type, time pressure, information source, and managerial experience. The analysis reveals that external information sources are most beneficial at later decision stages, while early reliance on internal organizational knowledge correlates with more effective decisions. DOE findings support the development of adaptive decision protocols tailored to the complexity and urgency of real-world hospital scenarios.23–30
This study addresses the critical need for improved managerial decision-making frameworks in healthcare by combining qualitative and quantitative methodologies. Through 15 semistructured interviews with hospital managers and the application of Random Forest, K-means clustering, and RSM-based DOE, we examine how different contextual and experiential factors shape decision quality under uncertainty. By integrating organizational behavior insights with data-driven techniques, the research not only identifies key determinants of effective decisions but also provides a replicable framework for improving hospital management. The findings aim to guide future efforts in enhancing healthcare decision making through tailored information strategies and optimized decision protocols, with broader applicability to other institutions through extended datasets such as the forthcoming decision types table. 31
Methodology
Inclusion and exclusion criteria
Staff with purely clinical roles (physicians and nurses) whose primary responsibilities did not include administrative/policy decisions. Managers on long-term leave or who had left the organization within the previous 6 months. Individuals who declined to be audio-recorded or to provide informed consent. Seventeen eligible managers were initially identified by the hospital's senior leadership. Two individuals were unavailable during the data-collection period, resulting in a final sample of 15 participants (88.2% response rate).
Qualitative phase
This study began with a qualitative inquiry involving semistructured interviews with 15 healthcare managers at Kawsar Hospital, selected via purposive sampling to ensure representation across operational, tactical, and strategic levels. At the first, written informed consent was obtained from all the participants. The Critical Incident Technique (CIT) was employed to focus the interviews on real-life, high-stakes decision-making experiences. Interview transcripts were analyzed using the NatCen Framework to systematically code decision phases, types, information sources, and contextual factors such as time pressure and managerial experience.
Quantitative phase
Data construction
Based on the qualitative findings, a structured dataset was constructed to quantitatively analyze decision-making patterns. Interview transcripts were first coded using the NatCen Framework and CIT, which allowed extraction of key variables such as decision type, level, time pressure, information source, and managerial experience. These coded variables were then transformed into numerical scales to enable statistical and ML analysis. Since the original data was qualitative and narrative-based, the dataset simulates decision scenarios reflecting the observed ranges and distributions identified in interviews. Each case represents a real decision episode reported by managers, converted into standardized quantitative form. This approach follows established mixed-methods practice, ensuring that narrative insights are systematically translated into analyzable data while preserving contextual validity.
The dataset includes 15 decision cases, covering operational, tactical, and strategic levels. To enhance transparency, the construction process involved:
Simulated experimental dataset based on qualitative interviews (n = 15).
Conceptual framework for method integration
The integration of Random Forest, K-means clustering, and DOE is guided by a complementary conceptual rationale. Random Forest was applied to identify and rank the relative importance of decision-making variables, thereby operationalizing the structured analysis of managerial choices. K-means clustering was used to uncover distinct profiles of decision-makers, reflecting the heterogeneity and naturalistic styles emphasized in decision-making theory. DOE with RSM provided a systematic approach to explore nonlinear interactions and illustrate potential decision conditions under uncertainty. Together, these methods form a coherent framework that links decision-making theory with empirical analysis: Random Forest highlights determinants, K-means reveals behavioral clusters, and DOE suggests exploratory patterns of decision conditions. This hybrid approach aligns with Simon's distinction between structured and unstructured decisions, Mintzberg's managerial roles, and the principles of naturalistic decision making.
Machine learning analysis—Random Forest
A Random Forest Classifier was implemented in Python 3.13 to predict decision quality using the input features. This ensemble learning method was selected because it is robust to small datasets, reduces overfitting through bootstrap aggregation, and provides reliable estimates of variable importance. The model was configured with 100 decision trees, using the Gini index as the splitting criterion. To ensure methodological transparency and reliability, a five-fold cross-validation procedure was applied. Model performance was evaluated using R2 and root mean square error (RMSE), which confirmed acceptable predictive accuracy given the limited sample size. Results highlighted time pressure and managerial experience as the most influential variables associated with decision outcomes. Time pressure consistently reduced decision quality, while higher levels of experience partially buffered this effect. Other variables such as decision type and information source showed lower importance scores, suggesting that contextual stressors outweigh traditional factors in shaping decision quality.
Clustering analysis—K-means
Unsupervised K-means clustering was applied to categorize decision-making behaviors. The optimal number of clusters (k = 3) was determined using the Elbow Method and confirmed by the Silhouette Score, ensuring that the chosen segmentation reflects meaningful group differences.
The algorithm grouped the 15 decision cases into three clusters, each representing distinct managerial profiles:
This clustering highlights the heterogeneity of managerial decision-making styles, showing that healthcare managers are not homogeneous but fall into distinct behavioral categories. Profiling these clusters provides a basis for targeted managerial training and support strategies.
Design of experiments—response surface methodology
To explore the relationships between input factors and decision outcomes, RSM was applied using Design-Expert version 11. RSM uses a series of designed experiments to model and analyze the effects and interactions of multiple variables, fitting a regression model to illustrate potential response surfaces. This approach is suitable for complex, multifactor decision environments where nonlinear relationships may exist. The RSM analysis was conducted using the experimental dataset to examine how decision type, decision level, time pressure, information source, and managerial experience jointly influence decision quality and speed. The results are presented as exploratory patterns, offering quantitative insights into possible interactions under uncertainty rather than definitive optimization.
Results
Analysis of variance: Identifying key drivers of decision outcomes
To determine the influence of input factors on decision-making outcomes, an analysis of variance (ANOVA) was conducted for each of the three response variables: decision quality (R1) Table 2, decision speed (R2) Table 3, and decision impact (R3) Table 4. The ANOVA tables indicate how each variable contributes to variance in outcomes, based on F-values and corresponding p-values. For decision quality (R1), time pressure emerged as the only statistically significant predictor (F = 20.09, p = 0.0015), suggesting that higher time pressure negatively affects the quality of decisions. Other factors such as decision type, decision level, information source, and managerial experience did not show significant effects (p > 0.05). In the case of decision speed (R2), two factors significantly influenced outcomes. Decision level was the most impactful (F = 23.74, p = 0.0009), indicating that strategic-level decisions tend to take more time than operational or tactical ones. Time pressure was also significant (F = 5.19, p = 0.0488), confirming that urgency accelerates decision making. Managerial experience was marginally significant (p = 0.0712), hinting at a potential moderating role. For decision impact (R3)—which follows the same ANOVA pattern as R1—time pressure again showed a significant effect (F = 20.09, p = 0.0015), reinforcing its central role across multiple decision dimensions. Other variables showed no significant contribution. These findings collectively highlight that time pressure consistently affects both the quality and impact of decisions, often negatively, while decision level plays a central role in determining how fast decisions are made. Contrary to expectations, managerial experience and information source were not statistically significant, though they may interact with other variables in more complex ways not captured by main effects alone.
Only time pressure (C) significantly affected decision quality (p = 0.0015); other factors were not significant.
Decision level (B) and time pressure (C) had significant effects on decision speed (p < 0.05); experience showed a marginal effect.
Time pressure (C) was the only factor significantly influencing decision impact (p = 0.0015); others were not significant.
Interaction plot interpretation
The interaction plot for Decision Quality reveals a notable interplay between time pressure and managerial experience. Under high time pressure, decision quality decreases significantly when managerial experience is low. However, for more experienced managers, time pressure has a less detrimental effect, suggesting that experience buffers stress-induced quality loss. Additionally, the interaction between decision type and decision level shows that unstructured decisions made at the strategic level are more prone to lower quality, while structured decisions at the tactical level yield higher-quality outcomes (Figure 1(a)). For Decision Speed, the interaction plot demonstrates that decision level and information source jointly influence how quickly decisions are made. Tactical-level decisions supported by hybrid information sources (organizational + external) tend to be made more rapidly. Moreover, there is a clear interaction between time pressure and decision type: under high time pressure, structured decisions are executed faster than unstructured ones, implying that formalized procedures facilitate faster responses in urgent situations (Figure 1(b)). In the case of Decision Impact, the strongest interaction is observed between managerial experience and information source. Highly experienced managers using hybrid information sources tend to achieve the highest impact. Conversely, novice managers relying solely on internal knowledge experience limited decision impact. Another relevant interaction is between decision type and time pressure, indicating that unstructured decisions made under low time pressure result in more positive long-term effects, likely due to the allowance for deeper analysis and reflection (Figure 1(c)).
Interaction plots illustrating joint effects of input variables on (a) decision quality, (b) decision speed, and (c) decision impact. These plots highlight critical combinations such as time pressure × experience and decision level × information source, underscoring the contextual complexity of healthcare management decisions.
3D surface plot interpretation
The 3D surface in Figure 2(a) illustrates how managerial experience and time pressure interact to influence decision quality. The surface is relatively flat at low experience levels, indicating little variation in quality regardless of time pressure. However, as managerial experience increases, the surface slopes upward—particularly under moderate to low time pressure—highlighting that experienced managers perform better in environments with manageable urgency. The gradient indicates a strong positive effect of experience on decision quality when decision time is not severely constrained. In Figure 2(b), the plot represents the combined effects of decision type and decision level on decision speed. The tilted surface suggests that tactical-level, structured decisions are executed most rapidly. As the decision level shifts from operational to strategic and the type becomes more unstructured, the slope flattens and shifts downward, indicating slower decision making. This suggests that formalized and localized decisions accelerate action, while broader, strategic, and ambiguous decisions tend to require more deliberation. Figure 2(c) focuses on the interaction between information source and managerial experience in shaping decision impact. The surface displays a noticeable upward slope along the experience axis, especially when hybrid (internal + external) information sources are utilized. This implies that the highest decision impact is achieved when experienced managers integrate diverse data sources. Conversely, novice managers relying solely on internal sources generate flatter outcomes, with minimal impact.
Three-dimensional surface plots showing the interaction effects of key decision variables on (a) decision quality, (b) decision speed, and (c) decision impact. The surfaces demonstrate how combinations such as time pressure × experience and decision type × decision level shape decision performance metrics in healthcare management contexts.
Predicted versus actual plot interpretation for three response variables
To assess the predictive power and robustness of the fitted models for the three response variables, a set of predicted versus actual plots was generated using the RSM framework. These plots, presented in Figure 3, demonstrate the relationship between the model's predicted values and the actual experimental observations for each output. The scatter plot in Figure 3(a) reveals a strong linear relationship between the predicted and actual values. The data points cluster closely along the 45-degree reference line, indicating minimal deviation. This pattern confirms that the model accurately estimates the quality of decision outcomes under varying input conditions (e.g. decision level, time pressure, and experience). Figure 3(b) exhibits a moderate to strong correlation between predicted and actual values. Although some dispersion is evident around the line of equality, the general alignment of data points suggests acceptable predictive performance. Slight deviation at the extremes may indicate the influence of high-order interactions or nonlinear factors affecting decision-making speed. In Figure 3(c), the alignment of points is less concentrated, implying a weaker model fit for this response. Greater variance between actual and predicted values may stem from latent variables or noise in the experimental setup that were not fully captured in the model. Nonetheless, the directionality remains consistent, and the plot still provides valuable insights for qualitative interpretation.
Predicted versus actual plots for the three response variables derived from the Response Surface Methodology (RSM) model: (a) Decision Accuracy, (b) Decision Speed, and (c) Decision Impact. A strong correlation between predicted and actual values in plots (a) and (b) confirms model reliability, while plot (c) suggests areas for model refinement.
Interpretation of feature importance
The feature importance plot generated by the Random Forest algorithm highlights the relative contribution of each predictor variable to the model's ability to accurately predict Decision Quality (Figure 4). Among the examined features, Time Pressure and Decision Impact emerged as the two most influential factors, with importance scores of approximately 0.40 and 0.42, respectively. These results suggest that managers’ decision quality is highly sensitive to the time constraints under which decisions are made, as well as the perceived or actual impact of those decisions. In contrast, variables such as Experience (years) and Information Source contributed minimally to the model, indicating that, within the context of this dataset, these factors played a lesser role in determining decision quality. Interestingly, Decision Speed also demonstrated moderate importance, reinforcing the notion that not only the conditions but also the pace of decision making influence perceived quality. This analysis provides practical insights for hospital administrators and healthcare decision-makers: optimizing time management and emphasizing the consequences of decisions may enhance overall decision quality more effectively than focusing on traditional factors like experience or hierarchical level alone.
Random Forest variable importance for decision quality.
Interpretation of K-means clustering results
Figure 5 illustrates the output of K-means clustering applied to the decision-making dataset, projected into two dimensions using Principal Component Analysis (PCA) for visualization. The algorithm identified three distinct clusters, each represented by a different color: cluster 0 (green), cluster 1 (orange), and cluster 2 (blue). The spatial separation among the clusters suggests meaningful patterns in the underlying data. For instance, cluster 1 includes data points primarily concentrated in the negative space of Principal Component 1, indicating a group of decision-makers that may share characteristics such as lower time pressure or less hierarchical responsibility. Conversely, cluster 2 lies in the far-right region of the plot, potentially representing individuals under high decision speed or greater impact pressure. Cluster 0 appears to be more balanced, centered in the middle of the PCA space, suggesting moderate levels across key variables. This unsupervised analysis reinforces the idea that healthcare decision making is not monolithic, but rather consists of distinct profiles of managerial behavior, each with different combinations of pressures, experience, and decision outcomes. Such segmentation could inform targeted strategies for training, support, or policy development. Cluster profiling revealed three distinct managerial styles (Table 5). Cluster 1 represents routine managers operating under low time pressure and relying primarily on organizational knowledge, achieving high-quality but slower decisions. Cluster 2 includes crisis-driven managers facing high time pressure, often using external or hybrid sources, resulting in faster but lower-quality decisions. Cluster 3 consists of experienced adaptive managers who integrate hybrid information sources, achieving balanced decision quality and impact. These profiles provide statistical justification for the chosen k = 3 segmentation and highlight the heterogeneity of healthcare decision-making behaviors.
Interpretation of K-means clustering results.
Cluster profiles derived from K-means analysis.
Discussion
Interpretation of findings in context of literature
This study explored the factors influencing managerial decision-making quality in a healthcare context under uncertainty, using both DOE and ML approaches. The Random Forest analysis revealed that time pressure and decision impact were the two most influential variables affecting decision quality. While these findings align with the literature emphasizing the role of environmental stressors and high-stakes conditions in shaping managerial judgment, 31 they should be interpreted with caution given the limited sample size (n = 15) and the use of a simulated dataset. Managers under acute time constraints and high-impact contexts often rely on heuristic strategies, which may reduce deliberative quality. 32 In parallel, the K-means clustering analysis identified three distinct clusters of decision-makers, suggesting heterogeneity in managerial profiles. Each cluster exhibited unique patterns of experience, decision speed, and contextual pressure, echoing prior research on situational leadership and adaptive expertise in complex settings.33–35 Although exploratory in nature, these results indicate that decision quality is not only a function of individual competence but also strongly mediated by organizational and situational variables. This highlights the need for healthcare management practices that tailor training and support systems to different decision-making profiles, and that provide structural mechanisms to mitigate the negative effects of time pressure in high-impact contexts.
Implications for healthcare management practices
The findings, while exploratory and based on a limited simulated dataset, nonetheless offer several practical insights for healthcare administrators. These implications should be considered as indicative rather than definitive, but they highlight actionable directions for improving managerial decision making under uncertainty:
Taken together, these implications emphasize that healthcare management practices must move beyond traditional reliance on experience and hierarchy, toward context-sensitive, data-driven, and adaptive support systems that strengthen decision quality under uncertainty.
Limitations and future research directions
While the study yields valuable insights, several limitations must be acknowledged:
Sample size: The sample consisted of 15 decision scenarios from a single hospital, which may limit the generalizability of findings. Self-reported outcomes: Decision quality and impact were partly based on subjective assessments, which could introduce bias. Cross-sectional design: The research design was cross-sectional, capturing decisions at one point in time. A longitudinal approach would offer a deeper understanding of evolving decision behaviors. Model constraints: The Random Forest model, while robust, does not capture causal relationships. Future work could integrate causal inference methods or Bayesian models to enrich the analysis.
Future studies may also explore hybrid models combining qualitative insights (e.g. from interviews) with algorithmic classification, and extend the model across multiple healthcare settings to validate and refine decision-making frameworks under uncertainty. The reliance on simulated quantitative data derived from qualitative interviews is both a strength and a limitation: it allowed systematic analysis of narrative insights, but future research should incorporate larger, real-world datasets to enhance external validity.
Conclusion
This study explored the decision-making behavior of healthcare managers at Kawsar Hospital under conditions of uncertainty by combining experimental design (RSM-based DOE) and advanced ML techniques, namely Random Forest and K-means clustering. The findings revealed that time pressure and decision impact are the most influential factors affecting decision quality, whereas traditionally assumed variables like managerial experience played a minor role in predictive significance. This challenges some prior assumptions in the literature and highlights the need for context-sensitive decision support systems. The Random Forest model proved highly effective in identifying and quantifying the importance of input variables. It showed that decisions made under higher time pressure and with greater potential impact require more structured support, as these factors significantly determine quality outcomes. Meanwhile, K-means clustering allowed the segmentation of decision-makers into three distinct behavioral clusters based on their decision attributes. This provides a novel, data-driven framework for profiling healthcare managers beyond job titles or years of experience. These insights offer valuable practical implications for healthcare organizations. For Kawsar Hospital, it is recommended to integrate digital tools that can assess real-time decision contexts, track time-sensitive choices, and support high-impact decisions. Such tools could include AI-enabled dashboards or automated risk alerts. Furthermore, by mapping decision-makers into behavioral clusters, hospital management can align roles and responsibilities with decision profiles, enhancing both efficiency and accountability. Looking ahead, expanding the dataset to include decisions from other hospitals and contexts will help validate and generalize these findings. Incorporating qualitative data such as cognitive style or leadership traits could further enrich model performance. Additionally, building a live decision-monitoring dashboard would allow real-time data collection and model retraining, transforming the decision-making process into a dynamic and learning-driven system. These steps will pave the way for a more intelligent, adaptive, and evidence-based approach to healthcare management.
Footnotes
Acknowledgements
We would like to express our gratitude to all the participants and staff at Kawsar Hospital in Sanandaj.
Ethical consideration
This research was ethically approved by the Ethics Committee of the Sanandaj Branch, Islamic Azad University, Sanandaj, Iran (IR.IAU.SDJ.REC.1402.039).
Consent to participate
Verbal consent to participate was obtained and informed from all of the participants.
Consent for publication
All authors have read and approved the final version of the manuscript and consent to its publication in this journal.
Author contributions
Farhad Sattar Mohammed: Writing—review & editing; Conceptualization; Data curation; Investigation; Methodology; Validation; Writing—original draft.
Kaveh Bahmanpour: Investigation; Methodology; Resources; Supervision; Validation; Visualization; Writing—review & editing.
Sina Valiee: Writing—review & editing.
Adel Fatemi: Writing—review & editing; Conceptualization; Data curation; Formal analysis.
Funding
The authors received no financial support for the research, authorship, and/or publication of this article.
Declaration of conflicting interests
The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Data availability statement
All data generated or analyzed during this study are included in this article.
Guarantor
Kaveh Bahmanpour.