Tell Me! How to Diminish Stress and Sick Leave in an Organization: A Cognitive Behavioral Quasi-Experimental Intervention

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

Dear Authors,

I appreciate the work and effort. The manuscript addresses a relevant topic, focusing on stress reduction and sick leave through a combined web‑based resource (WBR) and cognitive‑behavioral therapy (CBT) intervention. The theoretical framing is drawing on COR, JD‑R, and Demand‑Control models, and the study offers practical implications for workplace well‑being programs. The manuscript is generally well written and clearly structured.

Below are my comments that in my point of view may improve the work - Major Comments

Title & Abstract 

The title is clear and reflects the study’s purpose, but consider specifying that the design is quasi‑experimental.

The abstract should more explicitly acknowledge baseline differences between groups and the limitations of self‑reported sick‑leave measures.

Introduction

The theoretical background is comprehensive, but some concepts (e.g., JD‑R, COR) are repeated. Streamlining would improve readability.

The rationale for combining WBR and CBT is sound, but the introduction could better articulate why these two modalities were expected to interact synergistically.

Explicitly present the study’s contributions. Although the context, theoretical background, and hypotheses are well developed, the authors do not clearly articulate what this study adds to the existing literature. It is recommended that the Introduction ends with a concise paragraph outlining the main contributions and the specific gap the study addresses.

Methods

Design

The description of the quasi‑experimental design lacks detail. Please clarify the pseudo‑randomization procedure and justify the pre‑assignment of the WBR+CBT group.

Consider discussing potential selection bias more explicitly.

Participants

Provide more information on recruitment, inclusion/exclusion criteria, and attrition.

Baseline differences between groups should be highlighted earlier in the manuscript.

Measures

The use of a single self‑reported item for sick leave is a major limitation. Please justify this choice and discuss its implications.

Consider adding reliability indices for all scales used.

Interventions

The WBR intervention is insufficiently described. Please include duration, frequency, and structure; adherence metrics; examples of modules or activities

The CBT intervention requires more detail regarding protocol, techniques, therapist qualifications, and fidelity monitoring.

Statistical Analysis

MANCOVA is appropriate, but the manuscript should explain how baseline differences were controlled.

Consider adding ANCOVA or mixed‑effects models as robustness checks.

The post‑hoc power analysis should be interpreted cautiously or removed.

Results

The results are clearly presented, but the manuscript should explicitly acknowledge the large baseline differences and discuss their impact.

Consider adding effect sizes for all comparisons.

Discussion

The discussion overstates the causal impact of the intervention. Please adopt a more cautious tone given the quasi‑experimental design.

Alternative explanations (regression to the mean, differential motivation, intervention intensity) should be discussed.

The limitations section should be expanded, particularly regarding measurement issues and generalizability.

Conclusion

The conclusion is coherent but should reflect the methodological constraints more explicitly.

Tables & Figures

Consider adding a comparative table summarizing the components of WBR vs. WBR+CBT.

A flow diagram of participant progression (CONSORT‑style) would improve transparency.

Language & Style

Overall writing quality is good, but some sentences are long and could be simplified.

Minor redundancies in the theoretical section could be removed.

 

Author Response

Reviewer #1:

Title & Abstract

The title is clear and reflects the study’s purpose but consider specifying that the design is quasi‑experimental. The abstract should more explicitly acknowledge baseline differences between groups and the limitations of self reported ‑sick leave‑ measures.

We appreciate this comment and follow the recommendations in the title and abstract, as suggested by the reviewer

Introduction

The theoretical background is comprehensive, but some concepts (e.g., JD‑R, COR) are repeated. Streamlining would improve readability.

Thank you for this suggestion. We have revised and streamlined the theoretical background by rewriting several sentences throughout the Introduction to reduce repetition of key concepts (e.g., JD–R and COR) and improve clarity and readability in a more concise and coherent manner.

The rewriting text can be found specifically on page 3 lines 102-107, and on page 4, line 170 of the manuscript.

The rationale for combining WBR and CBT is sound, but the introduction could better articulate why these two modalities were expected to interact synergistically.

Thank you for this useful comment. We have revised the introduction to more clearly articulate the rationale for the synergistic interaction between WBR and CBT. Specifically, we have expanded the theoretical justification by explaining how WBR and CBT simultaneously may enhance overall resource gain and improve intervention effectiveness.

The revised text can be found on pages 4-5, lines 179-185 of the manuscript.

Explicitly present the study’s contributions. Although the context, theoretical background, and hypotheses are well developed, the authors do not clearly articulate what this study adds to the existing literature. It is recommended that the Introduction ends with a concise paragraph outlining the main contributions and the specific gap the study addresses.

Thank you for this important suggestion. We have revised the end of the introduction to explicitly the study’s main contributions and the specific gap it addresses. Specifically, we have added a concise paragraph outlining the contributions of the study. These revisions clarify the novelty and relevance of the present research.

The revised text can be found on page 2, lines 68-73 of the manuscript.

 

 

Methods

Design

The description of the quasi-experimental design lacks detail. Please clarify the ‑pseudo randomization procedure and justify the ‑preassignment‑ of the WBR+CBT groupnsider discussing potential selection bias more explicitly.

We appreciate this suggestion. To clarify the quasi-experimental design, we have added more information about how pseudo-randomization was developed and justified in accordance with ethical standards and the subjects' needs and preferences. At the same time, we have discussed potential selection bias more explicitly and acknowledged its existence. 

The revised text can be found on page 9, lines 330-332, 333-334, 336-337, 343-349, and 352-357

Participants

Provide more information on recruitment, inclusion/exclusion criteria, and attrition. Baseline differences between groups should be highlighted earlier in the manuscript.

We are grateful for this comment, and we have added more information about recruitment, inclusion/exclusion criteria, and attrition. Additionally, we report baseline differences between groups in this section.

The revised text can be found on page 7, lines 256, and 266-277 of the manuscript.

Measures  

The use of a single self-reported item for sick leave is a major limitation. Please justify this choice and discuss its implications.

We thank for this comment. In fact, following Sackett and Larson (1990), the use of single items may be particularly appropriate for unidimensional constructs with a sufficiently narrow meaning, for instance, sickness absence. These authors argued that when a construct has a narrow scope, is unidimensional, and is unambiguous to respondents, a single-item measure is the most appropriate. This is one of the reasons why we use a one-item scale. Additionally, empirical evidence has shown that a single-item predictor has similar predictive validity to a scale (Bergkvist and Rossiter 2007, 2009). Lastly, we found a study that analyzed the prediction of sickness absence (measured by a single item) using two perceived health scales: a single-item scale and a multiple-item scale. It compares the predictive power of self-perceived health, measured by a single item, for sick leave. Although the SF-12 predicts sick leave with greater accuracy and stability, a single item is sufficient to identify workers at high risk of prolonged absences (Roelen et al., 2015). So, this single-item discriminates high vs low sickness leave. Furthermore, Song et al. (2023) have shown that time and resource constraints are particularly pronounced in momentary ecological assessments of repeated measures, where full multi-item scales for measuring specific constructs are impractical, as in our quasi-experiment. Therefore, it is sometimes more efficient to use a brief measure in applied research, work environments, and repeated-measures studies. Overall, if a construct has a narrow meaning, it is unidimensional; however, we acknowledge the advantages of multiple-item measures, as they may demonstrate high validity and reliability under certain circumstances.

 

REFERENCES

Bergkvist, L., & Rossiter, J. R. 2007. 'The predictive validity of multiple-item versus single-item measures of the same constructs.' Journal of Marketing Research, 44(2), 175–184.

Bergkvist, L., & Rossiter, J. R. 2009.' Tailor-made single-item measures of doubly concrete constructs.' International Journal of Advertising, 28(4), 607–621.

Heise, David R. “Separating Reliability and Stability in Test-Retest Correlation.” American Sociological Review 34, no. 1 (1969): 93–101. https://doi.org/10.2307/2092790.

Roelen, Corne A.M., Martijn W. Heymans, Jos W.R. Twisk, Mikko Laaksonen, Ståle Pallesen, Nils Magerøy, Bente E. Moen, and Bjørn Bjorvatn. 2015. 'Health measures in prediction models for high sickness absence: Single-item self-rated health versus multi-item SF-12.' European Journal of Public Health 25 (4): 668–672.

Sackett, P. R., & Larson, J. R. 1990. 'Research strategies and tactics in I/O psychology'. In M. D. Dunnette, P. L. Ackerman, L. M. Hough, & H. C. Triandis (Eds.), Handbook of industrial and organizational psychology Vol 1 (pp. 419–489). Palo Alto: Consulting Psychologists Press.

We have added a summary of these arguments in the limitation paragraph of the discussion. It can be found on page 18, lines 674-679.

Consider adding reliability indices for all scales used.

We have considered this comment, but we do not think that we could calculate reliability for the two variables that have a single-item each one. Following the classic work of Heise (1969), it is not possible to measure the stability (test-retest reliability) of the measurement at two points in time without obtaining an erroneous coefficient; they recommend at least three points in time. Additionally, if the goal of the intervention is to modify the variable mono item, the stability of correlations is compromised. In other words, we know of no other way to measure the reliability of single-item measurements. Therefore, we consider that it is not possible, strictly speaking, to add reliability to the two missing single-item variables.

Interventions

The WBR intervention is insufficiently described. Please include duration, frequency, and structure; adherence metrics; examples of modules or activities The CBT intervention requires more detail regarding protocol, techniques, therapist qualifications, and fidelity monitoring.

We thank you for your comment. We have added more information and a comparative table of differences between WBR and WBR+T interventions.

The new information can be found on pages 10-12, lines 396-398, 404-412, 422-431and 437-439 of the manuscript.

Statistical Analysis

MANCOVA is appropriate, but the manuscript should explain how baseline differences were controlled. Consider adding ANCOVA or mixed ‑effects models as robustness checks.

We have added two ANCOVA to check the robustness of the model.  

The new statistical analysis can be found on page 12, lines 444-456 of the manuscript.

The posthoc‑ power analysis should be interpreted cautiously or removed.

We have doubts about this suggestion. We think that post-hoc power analysis is valid. To follow your recommendation, we are going to add that this analysis should be interpreted with caution due to particularities of design.

This new paragraph can be found on page 14, lines 510-513 of the manuscript.

Results

The results are clearly presented, but the manuscript should explicitly acknowledge the large baseline differences and discuss their impact.

Thank you for your comment. We have added a paragraph acknowledging the baseline differences and discussed their impact on results (see page 12, lines 470-482). Anyway, it is important to note that we also conducted two ANCOVAs, with time 1 as the outcome and time 2 as the covariate, to confirm that the intervention had a significant impact. We have found a significant change from time 2 to time 1 in the groups that received an intervention. So, there is an improvement in workers' sick leave and stress due to the intervention.

Consider adding effect sizes for all comparisons.

Additionally, we added to eta square (η_p²) (Lakens, 2013) and d of Cohen (Cohen, 1988), such as an indicator of effect sizes (see tables 3 and 4 results).

These changes can be found on page 12, lines 465-468, and pages 12-13, lines 470-491. The corresponding results are also presented in Tables 3 and 4.

Discussion

The discussion overstates the causal impact of the intervention. Please adopt a more cautious tone given the quasi-experimental design‑.  Alternative explanations (regression to the mean, differential motivation, intervention intensity) should be discussed. The limitations section should be expanded, particularly regarding measurement issues and generalizability

We appreciate this comment. We have adopted a more cautious tone and explained the methodological issues in the limitations paragraph. We have explained the limitations and advantages of single-item measurements in the limitations paragraph and baseline differences in the quasi-experiment.

The revised text can be found on page 18, lines 674-679.

Conclusion

The conclusion is coherent but should reflect the methodological constraints more explicitly.

We have modified the conclusion, adding methodological constraints.

The revised text can be found on page 19, lines 706-710 of the manuscript.

Tables & Figures

Consider adding a comparative table summarizing the components of WBR vs. WBR+CBT.

We have added a comparative table summarizing the components of WBR vs. WBR + CBT intervention (Table 2, page 11).

A flow diagram of participant progression (CONSORTstyle‑) would improve transparency.

Thank you for this helpful suggestion. We have followed Consort-Style and, we have added a paragraph describing participant progression (page 7, lines 282-286) to improve transparency. In addition, a CONSORT-style flow diagram of participant progression has been incorporated as Figure 1, page 8.

Language & Style

Overall writing quality is good, but some sentences are long and could be simplified.

Thank you for this helpful comment. We have carefully reviewed the manuscript and revised several sentences to improve clarity, conciseness, and readability. These changes contribute to a clearer and more accessible presentation of the study.

Minor redundancies in the theoretical section could be removed

Thank you for this helpful comment. We have revised the theoretical section to eliminate minor redundancies and streamline the presentation of key concepts, resulting in a more concise and coherent narrative.

Reviewer 2 Report

Comments and Suggestions for Authors

The topic is timely and highly relevant within the field of occupational health psychology, particularly in light of the increasing attention to employee well-being and sustainable organizational practices. The study is theoretically well grounded, and it addresses an important gap in applied intervention research by comparing different intervention formats within a real organizational context. Nevertheless, several issues need to be addressed to strengthen the manuscript.

 

Introduction

The statement reporting that approximately 20% of the working population is affected by mental health problems each year, resulting in an estimated 12 billion lost workdays, requires clearer sourcing and contextualization. The reference should be explicitly indicated immediately after the claim, and it should be specified whether these figures refer to global, European, or other geographical contexts. (lines 37-38)

The theoretical section appears somewhat superficial in its treatment of the Demand–Control model, the JD–R framework, and Conservation of Resources theory. While these models are highly relevant, they are mainly mentioned rather than conceptually integrated.  (lines 71-85)

The statement regarding the role of “personal resources” in facilitating emotion regulation and stress management is rather generic. It would strengthen the manuscript to specify which personal resources are being referred to (e.g., self-efficacy, coping strategies, emotional regulation skills) and to support this claim with more targeted empirical evidence.  (lines 84-85)

The manuscript appears to overemphasize job control as a key  variable across different theoretical frameworks. While job control is central within the Demand–Control model, the JD–R framework conceptualizes stress as resulting from a broader configuration of job demands and multiple job resources, of which control represents only one possible component.  (lines 88-89)

The statement suggesting that research has primarily focused on providing resources to help workers manage stressors may inadvertently legitimize work stress as an unavoidable condition. It would be important to clarify that, while individual and personal resources are relevant, a substantial body of research emphasizes the need to address stressors at the organizational and work-design level, in line with primary prevention approaches.(lines 86-96).

 

Participants 

The description of the participants is rather limited. Providing additional information on key demographic and occupational characteristics (e.g., age, tenure, educational level, contract type) would improve the interpretability and generalizability of the findings.

 

Results

Baseline differences between groups—particularly the higher initial levels of stress and sick leave observed in the WBR+CBT condition—should be carefully considered when interpreting the observed effects. Although ANCOVA was appropriately used to control for covariates, baseline non-equivalence between groups warrants caution in interpreting the magnitude and specificity of the results.

Discussion

Although the discussion refers to JD–R and COR frameworks, the links between the empirical findings and the specific theoretical mechanisms proposed by these models could be articulated more clearly.

 

 

Author Response

Reviewer #2:

Comments and Suggestions for Authors

The topic is timely and highly relevant within the field of occupational health psychology, particularly in light of the increasing attention to employee well-being and sustainable organizational practices. The study is theoretically well grounded, and it addresses an important gap in applied intervention research by comparing different intervention formats within a real organizational context. Nevertheless, several issues need to be addressed to strengthen the manuscript.

Introduction

The statement reporting that approximately 20% of the working population is affected by mental health problems each year, resulting in an estimated 12 billion lost workdays, requires clearer sourcing and contextualization. The reference should be explicitly indicated immediately after the claim, and it should be specified whether these figures refer to global, European, or other geographical contexts. (lines 37-38)

Thank you for this valuable comment. We have revised the manuscript to clarify the source and contextualization. Specifically, we have now explicitly indicated the references immediately after the relevant claims and specified that these figures refer to global estimates. In addition, we have expanded the paragraph, and we have also added new references to provide further contextualization regarding the economic and productivity-related impact of mental health conditions worldwide, including their contribution to lost workdays, indirect societal costs, and reduced productivity. These revisions improve the accuracy, clarity, and contextualization of the information presented.

The revised text can be found on page 2, lines 39-53 of the manuscript.

The theoretical section appears somewhat superficial in its treatment of the Demand–Control model, the JD–R framework, and Conservation of Resources theory. While these models are highly relevant, they are mainly mentioned rather than conceptually integrated.  (lines 71-85)

Thank you for this valuable comment. We have revised the theoretical section to provide a more integrated and conceptually coherent presentation of the Demand–Control model, the JD–R framework, and conservation of resources theory. Specifically, we clarified how these models complement each other in explaining the role of job demands and both organizational and personal resources in shaping employee well-being. These revisions strengthen the theoretical foundation and improve the conceptual integration of the framework guiding the study.

The revised text can be found on page 3, lines 89-93 and 100-113 of the manuscript.

The statement regarding the role of “personal resources” in facilitating emotion regulation and stress management is rather generic. It would strengthen the manuscript to specify which personal resources are being referred to (e.g., self-efficacy, coping strategies, emotional regulation skills) and to support this claim with more targeted empirical evidence.  (lines 84-85)

Thank you for this helpful suggestion. We have revised the manuscript to clarify which specific personal resources are being referred to, explicitly including self-efficacy, coping strategies, emotional regulation skills, and psychological resilience. These revisions improve the conceptual clarity and empirical grounding of the argument.

The revised text can be found on page 3, lines 97-100 of the manuscript.

The manuscript appears to overemphasize job control as a key variable across different theoretical frameworks. While job control is central within the Demand–Control model, the JD–R framework conceptualizes stress as resulting from a broader configuration of job demands and multiple job resources, of which control represents only one possible component.  (lines 88-89)

Thank you for this clarification. We agree that job control plays a central role within the Demand–Control model but represents only one of several possible job resources within the JD–R framework. We have revised the manuscript to avoid overemphasizing job control and to more accurately reflect the broader configuration of job demands and multiple job resources proposed by the JD–R model. The revised text clarifies that job control is one relevant resource among others. These changes improve the theoretical precision and alignment of the manuscript.

The revised text can be found on page 3, lines 115-119 of the manuscript.

The statement suggesting that research has primarily focused on providing resources to help workers manage stressors may inadvertently legitimize work stress as an unavoidable condition. It would be important to clarify that, while individual and personal resources are relevant, a substantial body of research emphasizes the need to address stressors at the organizational and work-design level, in line with primary prevention approaches. (lines 86-96).

Thank you for this thoughtful comment. We agree that work stress should not be framed as an unavoidable condition and that primary prevention approaches targeting organizational and work-design factors are essential.

We have revised the paragraph to clarify that, while individual resources remain relevant, a substantial body of research emphasizes the importance of addressing psychosocial risk factors at the organizational level. These revisions ensure a more balanced perspective and align the manuscript with primary prevention principles.

The revised text can be found on page 3, lines 121-128 of the manuscript.

Participants 

The description of the participants is rather limited. Providing additional information on key demographic and occupational characteristics (e.g., age, tenure, educational level, contract type) would improve the interpretability and generalizability of the findings.

We really appreciate this comment. We have added more information to the sample’s description.

The initial sample consisted of 154 participants (82 women, 72 men), all employed full-time under indefinite contracts within the food manufacturing industry. Occupational roles included 81 operational employees and 40 administrative staff, with 33 participants not reporting their role classification. Due to organizational data access limitations, information on age and educational level was unavailable. All demographics available are described in Table 1 and in the participant’s section.

The revised text can be found on page 7, lines 266-277 of the manuscript.

 

Results

Baseline differences between groups—particularly the higher initial levels of stress and sick leave observed in the WBR+CBT condition—should be carefully considered when interpreting the observed effects. Although ANCOVA was appropriately used to control for covariates, baseline non-equivalence between groups warrants caution in interpreting the magnitude and specificity of the results.

Thank you for this important comment. We conducted two ANCOVAs controlling for baseline levels (Time 1) of the outcome variables. The results showed that, after controlling for baseline levels, the intervention had a significant effect on stress at Time 2, whereas no significant effect was found for sick leave at Time 2, possibly due to limited statistical power. We have clarified these results in the first paragraph of the Results section. In addition, we have included an explanation discussing how baseline differences may have influenced the observed intervention effects (see the last paragraph of the Results section).

Specifically, the revised text can be found on pages 12 and 16, lines 470-482 and 566-579 of the manuscript.

Discussion

Although the discussion refers to the JD–R and COR frameworks, the links between the empirical findings and the specific theoretical mechanisms proposed by these models could be more clearly articulated.

Thank you for calling our attention to this issue. In the discussion section we have added a paragraph to explain how cognitive-behavioral therapy contributes to provide additional personal resources, and how these additional resources contributed to reduce stress and sick leave.

The revised text can be found on pages 16-17, lines 643-655 of the manuscript.