Ninety-Day Cost, Mortality and Hospital Disparities in Ischemic Stroke: Real-World Evidence from a Czech Administrative Database

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

Introduction

• The EUR 45 billion estimate quoted was not obtained by Wafa et al. 2020, but rather by “Wilkins E, Wilson L, Wickramasinghe K, Bhatnagar P, Leal J, Luengo-Fernandez R, Burns R, Rayner M, Townsend N. European Cardiovascular Disease Statistics 2017. Brussels: European Heart Network; 2017.” In addition, more up to date sources now exist on the costs of stroke across Europe: https://academic.oup.com/eurheartj/article/44/45/4752/7251239

 

Methods

• I was unsure as to how secondary care resource use was valued. A micro-costing study would mean that all unit costs were estimated by the authors by taking into account staff costs, capital equipment, consumables, cost of buildings etc… to then derive, say, a cost per hospital day. Is this what was actually done? Or did the authors value resource use by using DRG tariffs, insurance claims etc…?

 

• In the methods I saw no mention of any statistical analyses to adjust for confounding. Therefore, if one of the objectives is to analyse differences in costs between different hospital providers, some level of adjustment would be expected. This could include age, sex, history of previous comorbidity (e.g. Charlson Comorbidity Index, based on previous admissions), urban vs. rural setting etc…

 

Results

• The authors report that 18,219 (i.e. the majority) patients received treatments other than thrombectomy/thrombolysis. What would these treatments entail? Mainly secondary prevention through pharmaceutical treatment?

 

• I am not very convinced that presenting mortality rates by type of treatment received is helpful. As presented, one can draw the conclusion that thrombectomy leads to death. Of course, this is because patients undergoing thrombectomy have more severe strokes, which therefore results in higher death. One could argue that type of treatment could be a surrogate for stroke severity, but other factors such as age, prior disability etc… will also play a role in determining the type of treatment received.

 

Conclusion

• I think the conclusion should be rewritten as it can lead to misconceptions. From their conclusion, I would conclude that stroke patients should not be treated with thrombectomy or thrombolysis as this leads to higher costs and higher mortality. However, these findings are because patients undergoing these two treatments are more likely to have severe events. I believe that the conclusions should focus on the fact that administrative data provides useful evidence on the resource needs, costs and outcomes of patients with stroke, which can help plan service provision, and might help benchmark costs and outcomes across different providers.

Author Response

Reviewer 1

Reviewer 1 – Comment 1

Introduction

The EUR 45 billion estimate quoted was not obtained by Wafa et al. 2020, but rather by “Wilkins E, Wilson L, Wickramasinghe K, Bhatnagar P, Leal J, Luengo-Fernandez R, Burns R, Rayner M, Townsend N. European Cardiovascular Disease Statistics 2017. Brussels: European Heart Network; 2017.” In addition, more up to date sources now exist on the costs of stroke across Europe: https://academic.oup.com/eurheartj/article/44/45/4752/7251239

Response:
We thank the reviewer for this important comment. Following the reviewer’s recommendation, we removed the original citation (Wilkins et al., 2017) and replaced it with a more recent reference providing updated data on the economic burden of stroke in Europe (Luengo-Fernandez et al., 2023).

Reviewer 1 – Comment 2

Methods

I was unsure as to how secondary care resource use was valued. A micro-costing study would mean that all unit costs were estimated by the authors by taking into account staff costs, capital equipment, consumables, cost of buildings etc… to then derive, say, a cost per hospital day. Is this what was actually done? Or did the authors value resource use by using DRG tariffs, insurance claims etc…?

Response:
We appreciate this comment. We have clarified the costing methodology in the Methods section. Although the term “micro-costing” was originally used, the analysis is based on aggregated administrative claims data submitted by healthcare providers and reimbursed by health insurance funds. Therefore, costs reflect payer expenditures (including hospitalizations, outpatient care, pharmaceuticals, and medical materials), rather than a bottom-up estimation of unit costs. The terminology has been revised accordingly to avoid confusion.

Reviewer 1 – Comment 3

Methods

In the methods I saw no mention of any statistical analyses to adjust for confounding. Therefore, if one of the objectives is to analyse differences in costs between different hospital providers, some level of adjustment would be expected. This could include age, sex, history of previous comorbidity (e.g. Charlson Comorbidity Index, based on previous admissions), urban vs. rural setting etc…

Response:
We agree that adjustment for confounding is essential when comparing providers or treatment effects. However, the primary aim of this study was descriptive—to assess the distribution of costs and outcomes in a real-world administrative dataset. We have clarified this in the Methods and explicitly acknowledged the lack of risk adjustment (beyond age standardization) as a limitation in the Discussion.

Reviewer 1 – Comment 4

Results

The authors report that 18,219 (i.e. the majority) patients received treatments other than thrombectomy/thrombolysis. What would these treatments entail? Mainly secondary prevention through pharmaceutical treatment?

Response:
We thank the reviewer for this comment. The term “other treatment” has been clarified in the Methods section to refer to patients who did not receive reperfusion therapy (thrombolysis or thrombectomy), and who were managed with standard medical care, including supportive treatment and secondary prevention.

Reviewer 1 – Comment 5

Results

I am not very convinced that presenting mortality rates by type of treatment received is helpful. As presented, one can draw the conclusion that thrombectomy leads to death. Of course, this is because patients undergoing thrombectomy have more severe strokes, which therefore results in higher death. One could argue that type of treatment could be a surrogate for stroke severity, but other factors such as age, prior disability etc… will also play a role in determining the type of treatment received.

Response:
We fully agree with the reviewer. Treatment groups in this study reflect underlying stroke severity and clinical indication rather than comparable populations. We have revised the Results and Discussion to explicitly state that differences in mortality between treatment groups should not be interpreted as causal effects of treatment.

Reviewer 1 – Comment 6

Conclusion

I think the conclusion should be rewritten as it can lead to misconceptions. From their conclusion, I would conclude that stroke patients should not be treated with thrombectomy or thrombolysis as this leads to higher costs and higher mortality. However, these findings are because patients undergoing these two treatments are more likely to have severe events. I believe that the conclusions should focus on the fact that administrative data provides useful evidence on the resource needs, costs and outcomes of patients with stroke, which can help plan service provision, and might help benchmark costs and outcomes across different providers.

Response:
We agree with the reviewer and have substantially rewritten the Conclusion. The revised version emphasizes that observed differences reflect patient characteristics and care pathways rather than treatment effectiveness. The conclusions now focus on the value of administrative data for system-level analysis, resource planning, and benchmarking.

Author Response File: Author Response.pdf

Reviewer 2 Report

Comments and Suggestions for Authors

The manuscript presents a real and high-impact problem worldwide, namely 90-day mortality and the financial impact, specifically the costs of care for patients with ischemic stroke in the Czech Republic, using administrative data from health insurance databases rather than databases from healthcare facilities where patients were hospitalized. At the same time, the cohort of patients included in the study is very large, and the topic is relevant because the assessment of costs and the performance of stroke centers are very important for the organization of healthcare systems. However, as the manuscript is reviewed, several concerns arise, which I will detail point by point below:

1. In the introduction, the reason for the study should be explained, namely its relevance to the analysis of costs and 90-day mortality from onset in patients with ischemic stroke.
2. The study objective should be formulated more clearly and presented strictly.
3. 'Other treatment' is a vague term that should be explained. What exactly does it refer to? Conservative treatment? Other interventional procedures?
   4. Unfortunately, one of the major limitations of the study is that data on stroke severity, namely NIHSS score and type of occlusion (large vessel occlusion, etc.), as well as data on patient functionality before and after stroke – the Rankin scale (mRS at admission and at 90 days), comorbidities – were not included. All of these influence patient mortality and, at the same time, costs.
   5. Another observation regarding the manuscript is that there is no data on the complications that patients experienced during hospitalization (hemorrhagic transformation of stroke, malignant edema, etc.) – these generate higher costs and higher mortality.
   6. The comparison should be centered on describing the distribution of costs and mortality in the stroke care system and not as a comparative evaluation of interventions (thrombectomy is known to be an expensive intervention, but life-saving and absolutely necessary under certain conditions). In this sense, the conclusion should also be reformulated.

Author Response

Reviewer 2

Reviewer 2 – Comment 1

In the introduction, the reason for the study should be explained, namely its relevance to the analysis of costs and 90-day mortality from onset in patients with ischemic stroke.

Response:
We have expanded the Introduction to more clearly explain the rationale of the study, particularly its relevance for understanding both costs and 90-day mortality in a real-world national dataset.

 

Reviewer 2 – Comment 2

The study objective should be formulated more clearly and presented strictly.

Response:
The study objective has been reformulated for clarity and precision.

 

Reviewer 2 – Comment 3

'Other treatment' is a vague term that should be explained. What exactly does it refer to? Conservative treatment? Other interventional procedures?

Response:
This has been clarified as described above.

 

Reviewer 2 – Comment 4

Unfortunately, one of the major limitations of the study is that data on stroke severity, namely NIHSS score and type of occlusion (large vessel occlusion, etc.), as well as data on patient functionality before and after stroke – the Rankin scale (mRS at admission and at 90 days), comorbidities – were not included. All of these influence patient mortality and, at the same time, costs.

Response:
We fully acknowledge this limitation. Administrative datasets do not include detailed clinical variables such as stroke severity (NIHSS), functional outcomes (mRS), or occlusion type. This limitation has been expanded in the Discussion.

 

 

 

Reviewer 2 – Comment 5

Another observation regarding the manuscript is that there is no data on the complications that patients experienced during hospitalization (hemorrhagic transformation of stroke, malignant edema, etc.) – these generate higher costs and higher mortality.
Response:
We agree that complications are important drivers of both costs and mortality. However, such data were not available in the dataset. This has been explicitly added as a limitation.

 

Reviewer 2 – Comment 6

The comparison should be centered on describing the distribution of costs and mortality in the stroke care system and not as a comparative evaluation of interventions (thrombectomy is known to be an expensive intervention, but life-saving and absolutely necessary under certain conditions). In this sense, the conclusion should also be reformulated.

Response:
We fully agree. The manuscript has been revised to emphasize a descriptive, system-level perspective. We have removed or reformulated language that could imply causal comparisons between interventions. The Conclusion has been rewritten accordingly.

Author Response File: Author Response.pdf

Round 2

Reviewer 2 Report

Comments and Suggestions for Authors

The authors have addressed all my recommendations.

 

Comments on the Quality of English Language

None