Page 88 - AIH-2-3

Artificial Intelligence in Health Organizational culture’s impact on burnout

2.7. Burnout prediction work, job home-life interference, irritability, anxiety,

Linear and logistic regression models can be used to depersonalization, mood swings, and overall burnout.
predict and diagnose burnout based on survey results. 23,24 Further information on the hypotheses can be found in the
32
Indicators such as exhaustion, cognitive performance literature.
issues, and lack of enjoyment in one’s work are predictors of 2.9. Gaps in existing literature
burnout. Similarly to OC, low performance also predicts
23
burnout. Highly neurotic individuals are more likely to The existing literature on the relationship between OC and
experience burnout than those with high self-efficacy. burnout highlights several unexplored or underexplored
areas:
Additional ML techniques, such as k-means clustering, (i) Integration of ML in burnout studies
cluster analysis, and multitask learning, have been While burnout is a well-researched topic, few studies
employed to classify and predict burnout. 25-27 These studies employ ML methods to analyze the relationship
demonstrate the potential of ML techniques to predict between OC and burnout. Traditional statistical
behavioral issues like burnout using survey data. However, methods dominate the field, leaving a gap in the
these studies do not focus on employees in health systems. application of advanced techniques such as random

2.8. OC, burnout, and ML forests, neural networks, or ensemble methods for
deeper and more nuanced insights.
Survey results have been used to predict components of (ii) Non-patient-facing health system employees
burnout using aspects of OC, such as partial least squares Most research on burnout focuses on patient-facing
regression and ordinary least squares regression. 22,28 The roles in healthcare, such as doctors and nurses, due
social environment has been identified as influencing work to their high-stress environments. However, there is
engagement, which is the opposite of burnout. Engagement limited literature addressing burnout among non-
29
is comprised of vigor, dedication, and absorption. Job patient-facing employees in health systems, despite
30
quality, which involves work conditions valued by employees, their critical roles in organizational functioning.
also influences burnout and overall employee well-being. (iii) OC as a predictor
Creating a culture that fosters feedback, autonomy, work- Although many studies acknowledge that OC affects
life balance, a positive climate, and open communication employee well-being, few quantitatively evaluate
also influences engagement and, consequently, reduces which specific cultural factors (e.g., communication,
burnout. 31
leadership style, and work-life balance) most
While numerous studies have identified predictors significantly predict burnout. There is a gap in
and influencers of burnout, it remains unclear whether identifying and ranking these predictors using robust
these influencers can be used to predict burnout using models.
ML techniques. For example, although a decision tree (iv) Dynamic and contextual nature of burnout
model has shown that OC can predict burnout, more Existing research often treats burnout as a static
advanced and complex techniques, such as random outcome rather than a dynamic process. Limited
forest models, have not been employed to demonstrate exploration of how changes in OC over time influence
32
the same results. This research builds on previous work burnout leaves a gap in understanding the temporal
by investigating how ML, specifically a more complex and adaptive aspects of these relationships.
and advanced random forest model, can be applied to (v) Interdisciplinary approaches
predict burnout among non-patient-facing employees in Research on burnout and OC is often conducted
health systems. Based on existing literature regarding in isolation, either focusing on psychological
32
the prediction of burnout using OC and ML techniques, aspects or management theories. There is a need
the research question is: Can a random forest ML model for interdisciplinary approaches that integrate
predict burnout among non-patient-facing employees in psychological, sociological, and computational
health systems? perspectives to study these complex interactions.
Previous literature has established a relationship (vi) Generalizability across industries and cultures
between OC and burnout. The dependent variables and Much of the current literature is geographically or
independent variables described in prior research were culturally specific, with a strong focus on Western
utilized in this study’s random forest model. It was organizational practices. There is a need for studies
32
hypothesized that whether employees perceive their that explore the cross-cultural applicability of findings
organization’s culture as positive will predict several and consider how the impact of OC on burnout may
burnout symptoms, including emotional exhaustion after vary across industries and regions.

Volume 2 Issue 3 (2025) 82 doi: 10.36922/aih.5127