Reader Comments
Post a new comment on this article
Post Your Discussion Comment
Please follow our guidelines for comments and review our competing interests policy. Comments that do not conform to our guidelines will be promptly removed and the user account disabled. The following must be avoided:
- Remarks that could be interpreted as allegations of misconduct
- Unsupported assertions or statements
- Inflammatory or insulting language
Thank You!
Thank you for taking the time to flag this posting; we review flagged postings on a regular basis.
closeAn important confounding factor
Posted by noritanida on 05 Jan 2012 at 01:06 GMT
In 1995, the Ministry of Health and Welfare (MHW) introduced the ICD-10 system into Japan. As part of its policy, MHW enforced doctors to improve a death certificate. Until 1995, it was a common practice for doctors to write down “heart failure” or “respiratory failure” as a cause of death. But, such a trivial death certificate was not allowed anymore. Instead, many doctors used “pneumonia” as a cause of death. Accordingly, a number of death certificates with “pneumonia” jumped up since 1995. I was wondering to what extents this fundamental change of the Japanese death certification system affected influenza-related excess mortality rates.
RE: An important confounding factor
vivek_charu replied to noritanida on 09 Jan 2012 at 18:16 GMT
Thank you for this comment. We examined the transition between ICD-9 and ICD-10-coded pneumonia and influenza (P&I) deaths in Japanese death certificates very carefully. In a method similar to that used by the US National Center for Health Statistics, we studied the ratio of ICD-10-coded P&I deaths to ICD-9-coded P&I deaths. We did not find any evidence to correct ICD-10 coded deaths by any factor.
Furthermore, if systematic changes in the coding of P&I deaths exist after 1995, we would expect that the increase in P&I-coded deaths would be present in the summer-months as well (months without influenza circulation). As shown in Fig. 1 of our paper, we do not see an abrupt change in P&I mortality rates before and after 1995 in Japan. Further, our estimation of baseline P&I mortality rates using cyclical regression models and spline functions allows us to correct for both short- and long-term trends in P&I mortality rates.
We therefore do not believe that changes in the coding of P&I deaths effect our calculation of influenza-related excess mortality rates in Japan.