Announcement

Collapse
No announcement yet.

CDC Changes Reporting Methodology

Collapse
X
  • Filter
  • Time
  • Show
Clear All
new posts

  • #61
    Re: CDC Changes Reporting Methodology

    Originally posted by Missouriwatcher View Post
    So multiply numbers by 4. Either way, comparisons can be made of partial data without multiplying. While P&I are definitely up for weeks 38-44, total deaths (due to any cause) are not. They are BELOW averages of last five years. So people are now being listed as dying of H1N1 when they probably would have died from their heart diseases, respiratory diseases. cancers, etc during this time-frame. ILI's may be high based on fear and panic. The numbers are not there to support that H1N1 is killing more people than would have otherwise died.

    RoRo, could you graph total deaths of past few years? Missing city data or not, the numbers don't change much, yet will show that 2009 ranks 4th as far as total deaths compared with the last five years.
    But thinking this through logically, if what you are claiming is true and not an artifact due to partial and late reporting within cities (not normally mentioned in the reports - only complete absences usually are) and inconsistent reporting methodology over time then it would perhaps seem that we should be spraying swine flu from helicopters to even further reduce the US national death rate and that sounds absurd.

    Even the CDC now says there have been around 4000 excess deaths. Do you disagree with that and are you suggesting that less children are dying during the swine flu epidemic than normally do?

    Comment


    • #62
      Re: CDC Changes Reporting Methodology

      There have been excess deaths due to H1N1, but not excess total deaths (due to any cause, including H1N1). Instead of someone dying of emphysema, heart disease, etc, those are now being placed into H1N1 category. People had been questioning where CDC got it's numbers for estimate of 36,000 dying annually of seasonal flu. They are now trying to paint a more realistic version of those numbers of people dying of H1N1. Yet, they have done a poor job explaining that those numbers have not translated to increases in total deaths (any cause including H1N1).

      In the transcript from yesterday, Anne Schuchat attempts at an explanation by saying, "We haven't so far seen anything that is of concern or extra concern but we're reviewing reports that we get every day and we will be updating you on information when new information that's critical comes to light" and "...many people can have a bacterial pneumonia following a flu illness and may or may not be recognized as flu." This new "system gives us additional detail about underlying diseases that children have, about issues like the bacterial pneumonia problem." "The death reporting in 2004 focused in on children with laboratory confirmed influenza infection.It would be better to compare our 129 number with the previous years of around 40 to 88 or so deaths each year than to look in on that 540 number." They have also made corrections to cases because "tests aren't perfectly accurate so a certain number of patients who really do have h1n1 flu have a negative test result." She goes on to state, "Things haven't really changed from last week to this week". http://cdc.gov/media/transcripts/2009/t091112.htm

      IMO, the CDC needs to better explain data, because misinterpretation can instill fear and panic. I wouldn't be surprised if ILI's jump drastically this week because of the way news reported.

      Comment


      • #63
        Re: CDC Changes Reporting Methodology

        Originally posted by Missouriwatcher View Post
        There have been excess deaths due to H1N1, but not excess total deaths (due to any cause, including H1N1).

        ...I wouldn't be surprised if ILI's jump drastically this week because of the way news reported
        I would argue that we do not know that. We have a set of numbers from a report of voluntarily submitted cities which could change in total population number covered for any number of reasons. You are taking the numbers at absolute face value to accurately scale up to the US nationally. Statistically I do not believe the evidence is there to justify this because we don't have enough evidence of the processes behind the data gathering. Your interpretation could be correct but I suspect there are others factors in the data which we are not aware of. When we are looking at small changes in relative values then a slight change in the data gathering process can appear to show artifacts that just aren't actually there.

        Do we have accurate deaths from all causes numbers for the USA other than the MMWR sample? If so how do these compare over the years?

        Oh and I'll bet that the US ILI rate doesn't increase after this CDC estimate. 95% of the population won't even have noticed it.

        Comment


        • #64
          Re: CDC Changes Reporting Methodology

          Look at ACHA (college tracking of ILI). They serve a population of over 3 million and have had nearly 74,000 cases (ILI) up to Nov. 6 (they began reporting Aug22-28).
          http://www.acha.org/ILI_Cumulative.cfm

          Of note, some of the schools haven't reported since Sept. While no deaths have been reported, FluTrackers has reported one as of 13Nov from University of New Mexico (which may be one of the schools reporting to ACHA) http://www.flutrackers.com/forum/sho...d.php?t=132988

          Comment


          • #65
            Re: CDC Changes Reporting Methodology

            Originally posted by Missouriwatcher View Post
            RoRo, could you graph total deaths of past few years? Missing city data or not, the numbers don't change much, yet will show that 2009 ranks 4th as far as total deaths compared with the last five years.
            Sure. So I went back through the MMWRs 2005-2009 for weeks 38-44, corrected for population as described in post 53. Cumulative fatalities for weeks 38-44 in 2009 are higher than all except 2007.
            Name:  a9bfb64b0fca83bca13c80c422e8970d.jpg
Views: 1
Size:  46.2 KB
            Twitter: @RonanKelly13
            The views expressed are mine alone and do not represent the views of my employer or any other person or organization.

            Comment


            • #66
              Re: CDC Changes Reporting Methodology

              You did a wonderful job on the graph! I know that was a lot of work. Thank you.

              This wasn't quite how I figured it, but it still gives a good view. One thing I would like to correct is that the avg of the five years is 78,154. That would make it a difference of 322 (1288 for this 7-week period). Breaking it down:

              +184 deaths weekly for 2009
              -81 deaths weekly for 2008
              +770 deaths weekly for 2007
              -206 deaths weekly for 2006
              -603 deaths weekly for 2005

              Comment


              • #67
                Re: CDC Changes Reporting Methodology

                Originally posted by Missouriwatcher View Post
                You did a wonderful job on the graph! I know that was a lot of work. Thank you.

                This wasn't quite how I figured it, but it still gives a good view. One thing I would like to correct is that the avg of the five years is 78,154. That would make it a difference of 322 (1288 for this 7-week period). Breaking it down:

                +184 deaths weekly for 2009
                -81 deaths weekly for 2008
                +770 deaths weekly for 2007
                -206 deaths weekly for 2006
                -603 deaths weekly for 2005
                Thanks, MW. It's not perfect, I decided to include city populations only in the corrections and not the metropolitan areas. I purposely kept 2009 seperate while averaging 2005-8 just to see how 2009 compared to the other four. Lots of fall deaths in 2007. Any one know what that was all about? I can't recall if we had lots of heatwaves that year, or was anything else going on?
                Last edited by Ronan Kelly; November 13th, 2009, 07:43 PM. Reason: fall not summer
                Twitter: @RonanKelly13
                The views expressed are mine alone and do not represent the views of my employer or any other person or organization.

                Comment


                • #68
                  Re: CDC Changes Reporting Methodology

                  Not sure those extra deaths in 2007 were due to flu in this time-frame. The P&I was not particularly high.

                  "During the 2007--2008 season, influenza activity peaked in mid-February and was associated with greater overall mortality, and higher rates of hospitalizations among children aged 0-4 years compared with each of the previous three seasons."
                  http://www.cdc.gov/flu/weekly/weekly...-08summary.htm

                  Comment


                  • #69
                    Re: CDC Changes Reporting Methodology

                    % of deaths due to P&I updated for week 45 (7.5%). Are we post peak?

                    Name:  9237a80e1baec3671e634c2df85862ef.jpg
Views: 1
Size:  44.5 KB

                    Name:  928bea4a58036a976f91a799533d04e5.jpg
Views: 1
Size:  43.3 KB
                    Twitter: @RonanKelly13
                    The views expressed are mine alone and do not represent the views of my employer or any other person or organization.

                    Comment


                    • #70
                      Re: CDC Changes Reporting Methodology

                      I think we need to wait another week or two to determine if we past peak. It appears that there were similar decreases in Week 45 in 2007 and 2008.
                      "I know God will not give me anything I can't handle. I just wish that He didn't trust me so much." - Mother Teresa of Calcutta

                      Comment

                      Working...
                      X