Hi Jerrine!
I had the same thoughts as you regarding how the validity and reliability could be improved in the study that you selected! As you know, drawing conclusions from these large, national samples using ICD codes is challenging, given their limitations in accuracy/representation of the true diagnosis, and I see your second paper attempted to validate the algorithm, however it was used again in a large, national sample. I completely agree with you that these results should be studied in a cohort that is characterized in more detail, in order to overcome these limitations in large datasets. Thank you for sharing and allowing me to comment!
I had the same thoughts as you regarding how the validity and reliability could be improved in the study that you selected! As you know, drawing conclusions from these large, national samples using ICD codes is challenging, given their limitations in accuracy/representation of the true diagnosis, and I see your second paper attempted to validate the algorithm, however it was used again in a large, national sample. I completely agree with you that these results should be studied in a cohort that is characterized in more detail, in order to overcome these limitations in large datasets. Thank you for sharing and allowing me to comment!