Letter to the Editor. Reporting continuous variables and statistical significance in a study of 30-day hospital readmissions after mechanical thrombectomy for acute ischemic stroke

View More View Less
  • 1 Mudanjiang Medical University, Affiliated Hongqi Hospital, Heilongjiang, China
  • 2 Heilongjiang Key Laboratory of Ischemic Stroke Prevention and Treatment, Heilongjiang, China
Free access

TO THE EDITOR: We read with great interest the article by Mouchtouris et al.1 (Mouchtouris N, Al Saiegh F, Valcarcel B, et al. Predictors of 30-day hospital readmission after mechanical thrombectomy for acute ischemic stroke. J Neurosurg. Published online May 1, 2020. doi:10.3171/2020.2.JNS193249). In this study, it was concluded that hypertension, length of hospital stay, and hemorrhagic conversion were important prognostic factors for readmission within 30 days in acute ischemic stroke patients treated with mechanical thrombectomy. However, the following issues should be addressed carefully.

First, it can be seen from the descriptions in their Tables 1 and 2 that continuous variables are represented by means and standard deviations (SDs) and that the differences between groups are compared by an unpaired t-test. However, we doubt whether those continuous variables described in these tables conform to a normal distribution and whether it is appropriate to describe those continuous variables that do not conform to a normal distribution as means and SDs. Obviously, the SD of some variables is almost equal to the mean, sometimes even larger than the mean. For example, the mean hospital length of stay in the readmitted group was 15.6 days, while the SD was 29.3 days, indicating that the data may not conform to a normal distribution. In this case, it is not appropriate to use an unpaired t-test to compare the differences between groups. We suggest using the Mann-Whitney U-test for statistical analysis of this study.

Second, as described in their Table 1, the hospital length of stay in the readmitted group was significantly higher than that in the non-readmitted group (15.6 ± 29.3 vs 9.1 ± 7.5 days). However, the p value described in their article suggests no statistical difference between the two groups. Based on the data provided in Table 1, the p value recalculated by Stata 11 software is < 0.001, which is notably different from the authors’ results. Is this a statistical mistake? We look forward to receiving a reasonable explanation from the authors. In addition, since the follow-up time has been reported in this study, why not use Cox regression analysis2,3 to explore prognostic factors?

Disclosures

The authors report no conflict of interest.

References

  • 1

    Mouchtouris N, Al Saiegh F, Valcarcel B, Predictors of 30-day hospital readmission after mechanical thrombectomy for acute ischemic stroke. J Neurosurg. Published online May 1, 2020. doi:10.3171/2020.2.JNS193249

    • Search Google Scholar
    • Export Citation
  • 2

    Arts EE, Fransen J, Den Broeder AA, Low disease activity (DAS28≤3.2) reduces the risk of first cardiovascular event in rheumatoid arthritis: a time-dependent Cox regression analysis in a large cohort study. Ann Rheum Dis. 2017;76(10):16931699.

    • Search Google Scholar
    • Export Citation
  • 3

    ElHafeez SA, Torino C, D’Arrigo G, An overview on standard statistical methods for assessing exposure-outcome link in survival analysis (Part II): the Kaplan-Meier analysis and the Cox regression method. Aging Clin Exp Res. 2012;24(3):203206.

    • Search Google Scholar
    • Export Citation
View More View Less
  • Thomas Jefferson University and Jefferson Hospital for Neuroscience, Philadelphia, PA
Keywords:

Response

We would like to thank Li et al. for their interest in our study and their critical review of our work, as well as the editor for enabling this scientific discourse. After thorough review of our data, we confirm that the continuous variables of age and National Institutes of Health Stroke Scale (NIHSS) scores are normally distributed, which was verified using histogram plots and the normal distribution curve. For that reason, the mean and SD are the appropriate descriptive statistics for these variables. We appreciate the authors’ point regarding the hospital and ICU lengths of stay (LOS), both of which are not normally distributed, and the nonparametric Mann-Whitney U-test was performed to analyze them. The median value and interquartile range are reported for these variables in Table 1 in our article. Similarly, the median and interquartile range are the appropriate descriptive statistics for the time to readmission and readmission LOS (Table 2 in our article).

Furthermore, the hospital LOS was compared between patients who were readmitted and those who were not by using the Mann-Whitney U-test, and a statistically significant difference was seen with a p value of 0.045 (see Table 1). We appreciate the authors’ suggestion of using the Mann-Whitney U-test; however, the suggested p value is not correct based on our data. However, this was updated and did not impact the rest of the statistical analysis.

Lastly, the Cox regression analysis was not performed because our study was not sufficiently powered to do so, and this was outside of our original scope. We aimed to identify significant risk factors of the binary outcome of readmission within 30 days from discharge. However, we certainly encourage larger-scale studies in the future that allow for such statistical analysis.

We thank Li et al. once again for their critical review, and we look forward to further research on this topic.

If the inline PDF is not rendering correctly, you can download the PDF file here.

Contributor Notes

Correspondence Changhao Yin: aoicg46@163.com.

INCLUDE WHEN CITING Published online August 21, 2020; DOI: 10.3171/2020.5.JNS201857.

Disclosures The authors report no conflict of interest.

  • 1

    Mouchtouris N, Al Saiegh F, Valcarcel B, Predictors of 30-day hospital readmission after mechanical thrombectomy for acute ischemic stroke. J Neurosurg. Published online May 1, 2020. doi:10.3171/2020.2.JNS193249

    • Search Google Scholar
    • Export Citation
  • 2

    Arts EE, Fransen J, Den Broeder AA, Low disease activity (DAS28≤3.2) reduces the risk of first cardiovascular event in rheumatoid arthritis: a time-dependent Cox regression analysis in a large cohort study. Ann Rheum Dis. 2017;76(10):16931699.

    • Search Google Scholar
    • Export Citation
  • 3

    ElHafeez SA, Torino C, D’Arrigo G, An overview on standard statistical methods for assessing exposure-outcome link in survival analysis (Part II): the Kaplan-Meier analysis and the Cox regression method. Aging Clin Exp Res. 2012;24(3):203206.

    • Search Google Scholar
    • Export Citation

Metrics