A Six-Month Observational Study of Nursing Workload in 14 Latvian Intensive Care Units Using the Nursing Activities Score

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

Dear Authors, congratulations for the topic, is very interesting. To improve the overall quality of your manuscript, I want to suggest to you some strategies to address some issues:

1) The main issue concerns focus and proportionality. The manuscript is extremely rich in descriptive and inferential analyses, but the central research question risks being diluted. Many sections repeat similar messages (high workload, interunit variability, staffing mismatch) using different statistical lenses without clearly adding conceptual depth. A clearer hierarchy of results, with explicit justification of why each additional model or test meaningfully advances understanding beyond the descriptive findings, would greatly strengthen the narrative and reduce cognitive overload for the reader.

2) While NAS is rigorously applied and well described, the manuscript tends to equate NAS-derived workload with “care complexity” and staffing adequacy without sufficiently unpacking the conceptual assumptions behind this translation. A short reflective discussion acknowledging what NAS captures well, what it may miss (e.g., cognitive, relational, coordination work), and how this affects policy conclusions would make the argument more balanced and credible, especially for an international audience.

3) The Discussion could be strengthened by more explicitly linking the observed NAS-based workload to downstream clinical and organizational consequences. While the present study convincingly demonstrates persistent high workload and staffing mismatch, prior evidence suggests that nursing care complexity is not only a workload issue but also a marker of patient instability and risk. In pediatric acute care, higher nursing care complexity—measured through nursing diagnoses and nursing actions—has been shown to be a strong independent predictor of intra-hospital transfers and ICU admission, even after accounting for medical severity indicators such as DRG weight (PMID:  39602875). This consideration would help position NAS not merely as a descriptive staffing tool, but as part of a broader framework in which nursing-sensitive measures capture clinically meaningful risk and can inform proactive staffing and safety strategies. Please include this perspective, also considering other studies that have explored the relationship between nursing workload, care complexity, and patient outcomes across different clinical settings, to further strengthen the conceptual grounding and generalizability of the discussion.

4) Another issue involves implications and generalizability. The conclusions advocate strongly for NAS-based staffing reform, yet the discussion remains largely confined to the Latvian context and descriptive system performance. Readers would benefit from a more explicit bridge to practice and policy: what concrete staffing decisions could realistically change tomorrow, and which findings are context-specific versus transferable to other health systems?

Author Response

Dear Reviewer 1,

We sincerely thank the Reviewer for the thoughtful, constructive, and conceptually oriented feedback. We highly appreciate the recognition of the relevance of the topic and the detailed suggestions aimed at improving the focus, conceptual clarity, and practical relevance of the manuscript. In response to these comments, the manuscript has been substantially revised, and all modifications are clearly marked in the revised version of the text (tracked/strikethrough changes) to facilitate transparent review.

Following the Reviewer’s recommendation regarding focus and proportionality, we have carefully restructured the Results section to establish a clearer analytical hierarchy and reduce redundancy. While retaining the richness of the dataset, we streamlined overlapping statistical analyses that conveyed similar messages (e.g., repeated significance testing across multiple methods) and prioritised effect sizes, variance-explaining models, and structurally informative analyses. Detailed post-hoc comparisons, extended robustness checks, and secondary diagnostic outputs were either removed or relocated to the Supplementary Materials. Short interpretive linking statements were added to clarify how each analytical step meaningfully advances understanding beyond descriptive findings. This restructuring reduces cognitive overload and strengthens the narrative coherence around the central research question.

In line with the Reviewer’s second comment, we revised the Discussion to more explicitly address the conceptual assumptions underlying the use of NAS. We now clearly state that NAS primarily captures time-based and observable nursing activities, reflecting operational workload intensity rather than the full multidimensional construct of care complexity. The revised text explicitly acknowledges important dimensions that are not fully captured by NAS, including cognitive workload, coordination and organisational work, relational care, ethical decision-making, and emotional labour. We further discuss how these conceptual boundaries should be considered when interpreting staffing adequacy and policy implications, thereby presenting a more balanced and theoretically grounded argument for international readers.

To address the third comment, the Discussion has been strengthened by more explicitly linking the observed NAS-based workload patterns to downstream clinical and organisational consequences. We expanded the conceptual framing to position NAS not only as a descriptive staffing instrument, but also as part of a broader framework of nursing-sensitive indicators that reflect patient instability, safety risk, and system strain. Drawing on prior evidence from different clinical settings, including studies demonstrating that higher nursing care complexity predicts adverse patient trajectories independently of medical severity, we situate NAS within a wider perspective on proactive staffing, risk anticipation, and patient safety. This addition strengthens the generalisability and clinical relevance of the findings.

Finally, in response to the comment on implications and generalisability, we revised the Discussion and Proposals sections to more clearly distinguish between context-specific findings related to the Latvian healthcare system and transferable principles applicable to other health systems. The revised manuscript now more explicitly outlines which conclusions are shaped by national workforce constraints and organisational models, and which findings support broader implementation of workload-based, dynamic staffing approaches. We also clarified the practical implications by indicating which staffing decisions could be realistically informed by NAS data in the short term, while maintaining appropriate caution regarding system-level reforms.

Overall, we believe that these revisions have significantly improved the clarity, conceptual robustness, and international relevance of the manuscript. We are grateful to the Reviewer for the insightful comments, which have helped us refine both the analytical focus and the theoretical grounding of the study.

Reviewer 2 Report

Comments and Suggestions for Authors

Thank you for the opportunity to review this article.

This study assessed nursing workload in 14 Latvian intensive care units using the Nursing Activities Score (NAS) over a six-month period. With over 28,000 shift-based observations and rigorous statistical modeling, the study reports consistently high workloads and significant interunit and interlevel variation. The dataset is extensive and well-analyzed, and the findings are highly relevant to national workforce planning.

Below are suggestions to strengthen the manuscript.

First, while the analysis is detailed, the methods section is overly lengthy and could benefit from more concise structuring. Key procedures—particularly those involving data processing, staff training, and inclusion/exclusion criteria—should be summarized more efficiently to improve readability.

Second, although the study is observational and descriptive, parts of the discussion and conclusions imply causal relationships (e.g., high NAS leading to burnout or turnover). These should be rephrased to emphasize associations rather than causal effects.

Third, while the use of NAS is appropriate and supported by literature, the study would benefit from stronger integration with clinical indicators such as patient acuity (e.g., APACHE or SOFA scores) to better contextualize workload levels and variability.

Fourth, some tables are dense and could be more reader-friendly. Consider highlighting key figures (e.g., highest and lowest NAS values, shortage rates) or reorganizing into summarized formats for clarity.

Fifth, the figures are generally informative but resolution should be improved for publication. Where possible, use vector graphics (e.g., SVG) to ensure quality.

Sixth, while the discussion provides a strong national perspective, more cross-reference to international studies using NAS would help contextualize the findings, especially when discussing workload thresholds and policy implications.

Seventh, although the conclusions are well stated, a more focused limitations section would improve transparency—particularly regarding the absence of clinical outcomes, variability in unit staffing policies, and possible documentation bias.

Eighth, the use of the word “prospective” to describe the study design may be misleading. While data were collected over time, the study is essentially cross-sectional and observational; “prospective” is typically reserved for longitudinal designs and should be revised accordingly.

Overall, this is a valuable contribution to ICU workforce research. I hope the authors find these suggestions helpful.

Author Response

Dear Reviewer 2,

We sincerely thank the Reviewer for the careful evaluation of our manuscript and for recognising the scope, rigor, and national relevance of the dataset. We highly appreciate the constructive suggestions, which have helped us further improve the clarity, transparency, and international positioning of the study. In response, the manuscript has been revised throughout, and all changes are clearly indicated in the revised version of the manuscript (tracked/marked text).

In line with the Reviewer’s first comment, the Methods section has been streamlined and structurally refined. While preserving methodological transparency, we reduced excessive procedural detail and reorganised the section to improve readability. Key aspects related to data processing, staff training, and inclusion/exclusion criteria are now summarised more concisely, with redundant or overly granular descriptions removed or condensed.

Regarding the second comment, we carefully revised the Discussion and Conclusions to avoid any implication of causal inference. All statements suggesting that high NAS values lead to burnout, turnover, or adverse outcomes were rephrased to explicitly reflect associations rather than causality, consistent with the observational nature of the study. This distinction is now clearly stated throughout the interpretative sections.

In response to the third point, we strengthened the discussion of clinical context and acuity. While patient severity indices such as APACHE or SOFA were not available in the dataset, this limitation is now explicitly acknowledged and discussed in the Limitations section. We also clarified that NAS partially reflects patient severity through care intensity but does not replace clinical acuity scores, and we highlight this as an important direction for future research.

Addressing the fourth comment, several dense tables were revised for improved readability. Key values (e.g., highest and lowest NAS levels, extreme shortage rates, and interunit contrasts) are now more clearly highlighted, while extended numerical detail has been either summarised or relocated to Supplementary Materials to reduce visual and cognitive burden.

In response to the fifth suggestion, figure quality has been improved. All figures have been revised to ensure adequate resolution for publication, and where possible, formats suitable for high-quality rendering (e.g., vector-based graphics) have been prepared to meet journal standards.

Regarding international contextualisation (sixth comment), the Discussion section was expanded and refined to include stronger cross-referencing to international NAS-based studies. These additions help situate the Latvian findings within broader evidence on workload thresholds, staffing adequacy, and policy implications across different healthcare systems.

In line with the seventh comment, the Limitations section was reorganised and sharpened to more clearly address key constraints, including the absence of clinical outcome measures, interunit variability in staffing policies, and potential documentation bias inherent in real-world workload recording. This enhances transparency and supports cautious interpretation of the findings.

Finally, we addressed the eighth comment by revising the study design terminology. The term “prospective” has been removed, and the study is now consistently described as observational and cross-sectional with repeated measurements over time, which more accurately reflects the design and avoids methodological ambiguity.

We are grateful to the Reviewer for these insightful recommendations, which have substantially strengthened the manuscript. We believe that the revisions have improved clarity, methodological precision, and international relevance, while maintaining the scientific integrity of the original work.

Reviewer 3 Report

Comments and Suggestions for Authors

The manuscript investigates the actual nursing workload in Latvian ICUs as measured by the NAS, and how adequately do current staffing levels reflect this workload. This question is clearly stated, coherently followed. The study focuses on workload magnitude, inter-unit and inter-level variability, and the relationship between workload and staffing shortages.

The topic is highly relevant. Nursing workload and staffing adequacy in ICUs remain a globally issue, especially in the post-COVID. The Latvian healthcare setting is well justified as under-researched.

The study is not conceptually novel, as NAS has been widely applied internationally. The findings are largely confirmatory, reinforcing existing knowledge. The novelty lies in the large dataset NAS assessment in Latvia. The Introduction could be strengthened by explicitly stating what this Latvian dataset adds to global NAS research, which assumptions in prior literature are challenged or extended (unexpectedly high workload in level 2 ICUs…)

Several models explain limited variance (staff shortage R2 = 0.115) while statistically significant, these results should be interpreted cautiously.

The research design is strong. Some concern about selection bias: ICU participation was voluntary. The manuscript does not specify total number of ICUs nationwide, proportion represented by the 14 participating units… This introduces potential bias.

NAS observations are shift-based and likely repeated within units and patients.

While unit-level random effects are included, patient-level clustering is not discussed.

Clarification is needed regarding patient identifiers and repeated measures.

Overreliance on significance, as N very large, tiny effects become statistically significant. Many results with negligible effect sizes are discussed at length. Consider shift emphasis from p-values to effect sizes and clinical relevance.

The Abstract mentions time-series forecasting (L32), however, no formal forecasting models are presented. Either provide details or remove this claim.

Results are internally consistent and well structured.

Tables and figures are comprehensive. Table 1 is excessively detailed and difficult to interpret. Consider several indicators could be moved to Supplementary Materials.

Some figures (NAS scatterplots) add limited value. Consider consolidation or removal to improve clarity.

The finding about level 2 ICUs show the highest workload is interesting but underexplained.

Interpretation of regression results occasionally exceeds explanatory power. Consider provide stronger clinical and organizational explanations for unexpected patterns.

Conclusions align with stated objectives. Policy implications are clearly articulated

Some recommendations exceed what the data alone can support. Feasibility, cost, and implementation barriers of routine NAS use are not addressed. Add a short subsection on implementation considerations and system constraints.

References are generally appropriate and relevant. Limited inclusion of very recent (2023–2024) systematic reviews on ICU staffing.

L46: ICU are the healthcare environment…  => ICUs are…

Abbreviations (ICU, NAS..) should be defined consistently at first mention

Decimal notation is inconsistent (comma versus period L623-624; 690-692…)

Author Response

Dear Reviewer 3,

We sincerely thank the Reviewer for the thorough and balanced assessment of our manuscript. We highly appreciate the recognition of the study’s relevance, dataset scale, and internal coherence, as well as the constructive and detailed suggestions aimed at strengthening the conceptual framing, methodological transparency, and interpretability of the findings. In response, the manuscript has been revised throughout, and all changes are clearly marked in the revised version.

To address the comments regarding conceptual novelty and positioning, we have strengthened the Introduction to more explicitly articulate what the Latvian NAS dataset contributes to the international literature. In particular, we now emphasise how this large-scale, multicentre dataset extends existing NAS evidence by revealing unexpected workload patterns (notably the consistently higher workload in level 2 ICUs), challenging assumptions that the highest workload is necessarily concentrated in level 3 units, and providing national-level empirical evidence from a previously under-researched healthcare system.

In line with the Reviewer’s remarks on interpretation of statistical models and explained variance, we revised the Results and Discussion to adopt a more cautious and clinically grounded interpretation of regression findings with modest R² values (e.g., staffing shortage models). The text now explicitly acknowledges the limited explanatory power of single-variable models and avoids overinterpretation, focusing instead on effect sizes, directionality, and practical relevance.

Several revisions were made to address methodological transparency and potential sources of bias. The Methods section now clarifies the voluntary nature of ICU participation and explicitly acknowledges the possibility of selection bias. Where possible, we now contextualise the 14 participating ICUs relative to the national ICU landscape and clearly state this limitation. We also expanded the description of data structure to clarify that NAS observations are shift-based and include repeated measurements within units. While unit-level random effects were included, we now explicitly discuss the absence of patient-level clustering and identifiers as a limitation, acknowledging that patient-level longitudinal modelling was not feasible within the current data structure.

Responding to concerns about overreliance on statistical significance in a very large sample, we systematically revised the Results and Discussion to reduce emphasis on p-values for negligible effects. The narrative now prioritises effect sizes, clinical relevance, and consistency of patterns, and several statistically significant but clinically trivial findings have been shortened, consolidated, or moved to Supplementary Materials.

In response to comments on time-series analysis, we clarified the scope and role of temporal analyses. The Abstract and Results were revised to ensure consistency between claims and presented analyses. Forecasting statements were either explicitly supported by the ARIMA modelling section or rephrased to avoid overstatement where predictive inference was not central to the study aims.

Regarding tables and figures, Table 1 and other highly detailed tables were reorganised to improve readability, with selected indicators relocated to Supplementary Materials. Several figures with limited added interpretative value were consolidated or simplified, while key visualisations highlighting interunit variability and workload–staffing mismatch were retained.

The Discussion section was further refined to better explain unexpected findings, particularly the high workload observed in level 2 ICUs, with stronger clinical and organisational interpretation. We also ensured that conclusions remain aligned with the observational nature of the study and do not exceed the explanatory power of the data.

To address concerns about policy recommendations and feasibility, we added a short reflective component discussing implementation considerations and system constraints related to routine NAS use, including organisational readiness, data infrastructure, and the need for phased integration rather than immediate system-wide reform.

Finally, several editorial and formal issues were corrected, including grammatical errors (e.g., “ICUs are”), consistent definition of abbreviations at first mention, harmonisation of decimal notation, and minor language polishing. The reference list was also updated to include more recent international literature where appropriate.

We sincerely thank the Reviewer for these insightful and constructive comments. We believe that the revisions have substantially improved the manuscript’s clarity, balance, methodological transparency, and international relevance, while preserving the strength of its empirical contribution.

Round 2

Reviewer 1 Report

Comments and Suggestions for Authors

While the authors have expanded the Discussion and incorporated some of the terminology suggested in the previous round of review, the core conceptual concern raised by the previous reviewer in Comment 3 remains insufficiently addressed. The revised text largely reproduces the language of the earlier review in a more polished form, but it does not demonstrate a corresponding level of conceptual reasoning or interpretative decision-making.
The Discussion continues to frame NAS primarily as a descriptive staffing tool. Although patient instability, system strain, and nursing-sensitive indicators are mentioned, these concepts are not analytically unpacked, nor are they used to reinterpret the findings. As a result, several key questions remain unanswered. Do the authors conceptualise NAS-based workload as a proxy for clinically meaningful patient risk, as a correlate of organisational strain, or as an indicator that is conceptually distinct from but complementary to nursing care complexity grounded in nursing clinical reasoning? Which of these positions do the authors explicitly adopt, and what are the theoretical implications of this choice for interpreting high NAS values beyond staffing adequacy?
Importantly, in the previous round of review the authors were directed to empirical evidence demonstrating that nursing care complexity, operationalised through nursing diagnoses and nursing actions, predicts adverse patient trajectories independently of medical severity indicators. In the current revision, this evidence is acknowledged only in passing and is not substantively engaged. How do the authors reconcile their NAS-based findings with evidence showing that nursing care complexity captures clinically meaningful risk beyond workload intensity alone?
Without explicit answers to these questions, the revision appears rhetorically aligned with the previous review but conceptually unchanged. A major revision requires the authors to take a clear conceptual position and to demonstrate how engagement with existing nursing-sensitive evidence, including the previously indicated study, meaningfully reshapes the interpretation and implications of their findings. For these reasons, the concerns raised in the previous round should be considered unresolved.

Author Response

Dear Reviewer 1,

Thank you for this important conceptual comment. We acknowledge that in the previous revision the Discussion did not sufficiently clarify the interpretative framework used for the Nursing Activities Score (NAS).  In response, we have revised the Discussion to explicitly define the conceptual role of NAS in this study. NAS is now clearly positioned as an indicator of realised care demand and organisational workload intensity, rather than as a proxy for nursing care complexity or direct patient risk. We explicitly distinguish NAS-based workload from nursing care complexity as defined through nursing diagnoses and clinical reasoning. We have also strengthened engagement with nursing-sensitive literature by clarifying that evidence demonstrating the independent predictive role of nursing care complexity is not contradicted by our findings. Instead, we conceptualise NAS-based workload and nursing care complexity as distinct but complementary constructs, where high NAS values indicate conditions of system strain under which nursing complexity may have amplified downstream effects. These revisions clarify our conceptual position and reshape the interpretation of high NAS values beyond staffing adequacy alone. We believe this addresses the reviewer’s concern regarding conceptual reasoning and interpretative decision-making.

Reviewer 2 Report

Comments and Suggestions for Authors

Thank you for addressing my comments.

Author Response

Thank you very much for your answer and comments.

Reviewer 3 Report

Comments and Suggestions for Authors

The authors responded adequately to my comments. I have no further comments. 

Author Response

Thank you very much for your answer and comments.