Politics bites science

For a clear outline as to why you shouldn't believe that 45,000 people die each year because they don't have insurance, read the blog entry by John Milloy, publisher of "Junk Science" on last week's report in the American Journal of Public Health.

Forget that the authors are blatantly biased members of the "Physicians for a National Health Plan."

All you need to know is:
# Among the many problems with the study, here are four of the most glaring; all of which will likely be missed by the media:

* The researchers assumed that study subjects lacking health insurance at the time of the interviews did not subsequently gain or regain insurance coverage. In fact, a study subject could have received health coverage the very next day after the interview and this would not have been considered by the researchers.
* The researchers essentially assume that lack of health insurance at the time of interview is the causal factor in the deaths that occurred. No data was gathered to back up this assumption.
* None of the data collected during the interviews, including insurance status, was validated by the researchers.
* The study result is statistically weak. Combined with the peculiar date-of-death cutoff (the year 2000 as opposed to any other year), it raises questions as to whether the study result was produced by “data dredging” – essentially cherry-picking data that provides the desired result.