Fluctuations in influenza-like illness epidemics and suicide mortality: A time-series regression of 13-year mortality data in South Korea

Aims We explored the association between influenza epidemic and suicide mortality rates in a large population using a time-series regression of 13-year mortality data in South Korea. Methods Weekly suicide mortalities and influenza-like illness (ILI) were analyzed using time series regression. Regression coefficient for suicide mortality based on percentage change of ILI was calculated using a quasi-Poisson regression. Non-linear distributed lag models with quadratic function up to 24 weeks were constructed. Results The association between ILI and suicide mortality increased significantly up to 8 weeks post-influenza diagnosis. A significant positive association between ILI and suicide mortality was observed from 2009, when a novel influenza A(H1N1)pdm09 virus provoked a worldwide pandemic. No meaningful association between these factors was observed before 2009. Conclusion There was a significant positive relationship between ILI and suicide mortality after 2009, when a novel influenza A(H1N1)pdm09 virus provoked a worldwide pandemic.

Review on manuscript "Fluctuations in in uenza epidemics and suicide mortality: A time-series regression of 13-year mortality data" Summary of the manuscript is manuscript investigates an association between suicide mortality and the rate of in uenzalike illness (ILI) in Korea.
e proposed study is exciting and timely as we experience a virus outbreak.Still, I am afraid that the presentation of statistical methods, data description, the measure of the unit of variables, and analysis is confusing.
I strongly recommend using consistent words for suicide mortality and ILI rate.roughout the manuscript, I read suicide a empts, in uenza symptoms, in uenza infections were used instead of suicide mortality, and ILI rate that could distract any reader.Another confusing word was novel A(H1N1)pdm09.Do we have a di erence between novel A(H1N1)pdm09, classic novel A(H1N1)pdm09, the new novel A(H1N1)pdm09, A(H1N1)pdm09?I think all are the same, but I would choose one name for the entire text.

Abstract:
Objective: e title is entirely di erent from the objective of the study.1) As I understood, this study is only in Korea, so I would recommend re ecting this in the title.2) title says in uenza epidemics and suicide mortality, but the objective is about in uenza infections and neuropsychiatric symptoms.
Aims: I would recommend being speci c about the population under study -Korea.
Methods: I couldn't understand the sentence that the regression coe cient for suicide based on the percentage change of in uenza.

Lines 2-9:
is study focused on the population of Korea, so why wouldn't the authors mention any previous study in Korea in the introduction?3. Line 10: "In population studies …" I would suggest being speci c (name it) about the population.
4. Line 11: Do authors consider weekly/monthly suicide mortality?I would recommend saying the month instead of the spring season because seasons are di erent across countries.
5. Line 11: I would either use overlap or follow, not both.
7. Line 14: I am not sure about using the "etiology of suicide." Could we use the cause of suicide mortality?Moreover, if the prevention strategies for Korea, then authors could revise the sentence to re ect this in the sentence.
10. Line 23: I would say subject information is blinded.
11. Line 24: What does "current study" refer to?Would it be a study in this paper?Correct me if I am wrong: I assume that this paper uses a retrospective study data.
Line 37: It is confusing that the introduction reads on neuropsychiatric symptoms, but this line picks only suicide.I would rewrite clearly whether this study considers only suicide mortality or the number of suicides per week or whatever appropriate throughout the text.
13. Line 40-41: Correct me if I am wrong.I assume that the authors didn't de ne the codes so that I would rewrite this sentence.For example, we chose cases with X60-X84 and Y870 related to intentional self-harm and sequelae of intentional self-harm, respectively.
14. Line 41-42: is sentence is hard to understand.If the authors would revise the previous sentence inline 40-41, then this sentence is redundant so that this sentence can be removed from the text.
15. Line 42-43: Would it be the same data or same database?
16. Statistical analysis: is section can be rewri en with su cient detail.Some comments are the following: 17. Line 47-48: is sentence could be simpli ed.For example, we used the weekly percent of ILI.
18. Lines 48-49: I would suggest to write down the regression model and de ne the terms in the model.• I would write the caption for each gure below the gure.
• Figure 3: What is the y-axis?Is it a rate ratio of suicide?What is the grey line?Why is the label " % increment of in uenza-like illness?" here?
• Figure 1: Why Figure 3 appears before presenting Figure 1?For the caption of gure 1, I would recommend to write the model and de ne the terms in the model.What is P ?What is the time series adjusted?
• Figure 2: Does each boxplot include weekly suicide in each month?If this is the case, I will write this in the caption.
19. Line 51: e model description and ing method can be rewri en with model speci cation, as suggested in the previous comment.38.Line 108: "However, a signi cant association was observed between depression and Inuenza A infection."-I would recommend rewriting this sentence with an appropriate transition word than "However."39.Line 121: I would specify what are the signi cant results.Would it be a signi cant association or signi cant periodic association?What variables involved in these associations?40.Line 134 -135: I would write this sentence in the introduction.41.Line 136: It is confusing to me that A(H1N1) has been considered as a new virus subtype in lines 134-135, but it is named as classic (previous) seasonal in uenza A(H1N1).Why do both sentences contradicting?42." is" -It is confusing to me what does this refer to?43.I thought the authors selected weekly ILI cases from K-CDC for the complete analysis and mentioned in the results that a er 2009 and until eight weeks, there was a signi cant association between suicide rate and percent ILI.If the HIN1 pandemic was in 2009, why did the authors couldn't consider this data for their fact-checking?44.Line 152: I would choose one word from ILI or in uenza infections or in uenza symptoms.As I understood, ILI is the most appropriate among these three words.