Abstract
To the Editor,
I have read with interest the recent article by Setzer et al. However, I have concerns with the methodology.
In regard to the meteorological variables, the data were provided by 2 local offices responsible for the Rhein main area. On the basis of the patients' location at the time point of bleeding, the patients were referred to one of 2 offices. The measured values were compared between days of subarachnoid hemorrhage (SAH) and days without SAH.
I could not determine what had been done when the authors had 2 or more cases of SAH in 1 day from both centers and whether they had combined the data from 2 centers or not. There is no statement regarding how values were measured in nonbleeding days and whether one or both centers' data had been used.
As for bleeding days, one of the office's data for each day was used; they should have compared them with the data for nonbleeding days from the same center. In other words, there should be 2 sets of data from each of the meteorological centers, comparing the variables in each center separately between bleeding and nonbleeding days. Otherwise, the result does not seem to be reliable. Any combinations of the data from 2 centers like calculation of means or pooling of data do not give reliable results.
My other concern is about the population factor. If we compare variables in 2 different datasets from each center, the results should be adjusted for the population covered by each center. Otherwise, inconclusive comparison in the less populated center may dilute the results from the other center(s) and make them inconclusive too.
I have read with interest the recent article by Setzer et al. However, I have concerns with the methodology.
In regard to the meteorological variables, the data were provided by 2 local offices responsible for the Rhein main area. On the basis of the patients' location at the time point of bleeding, the patients were referred to one of 2 offices. The measured values were compared between days of subarachnoid hemorrhage (SAH) and days without SAH.
I could not determine what had been done when the authors had 2 or more cases of SAH in 1 day from both centers and whether they had combined the data from 2 centers or not. There is no statement regarding how values were measured in nonbleeding days and whether one or both centers' data had been used.
As for bleeding days, one of the office's data for each day was used; they should have compared them with the data for nonbleeding days from the same center. In other words, there should be 2 sets of data from each of the meteorological centers, comparing the variables in each center separately between bleeding and nonbleeding days. Otherwise, the result does not seem to be reliable. Any combinations of the data from 2 centers like calculation of means or pooling of data do not give reliable results.
My other concern is about the population factor. If we compare variables in 2 different datasets from each center, the results should be adjusted for the population covered by each center. Otherwise, inconclusive comparison in the less populated center may dilute the results from the other center(s) and make them inconclusive too.
Original language | English |
---|---|
Pages (from-to) | 401-401 |
Number of pages | 1 |
Journal | Surgical Neurology |
Volume | 71 |
Issue number | 3 |
DOIs | |
Publication status | Published - Mar 2009 |
Externally published | Yes |