RE: LLRC Email - FW: Four nuclear wars

["A Karam" <paksbi@rit.edu>]

Cancer is a stochastic process, which means it should occur randomly in

time and in space.  Any random phenomenon will have areas with higher

and lower numbers of events.  For example, if you throw a handful of

sand on a table, there will be places where you have clusters of sand

particles, even if the table is perfectly smooth and uniform.



Cancer is not a rare disease, so the overall incidence is high enough

that virtually every neighborhood and social group (family, circle of

friends, etc.) will experience it.  And, by pure chance, there will be

some areas in which a number of cancers manifest themselves in a

relatively limited region or social grouping.  Our tendency as humans is

to try to see patterns, so we want to find a pattern (a reason) for

these groupings.  We have trouble accepting that sometimes we get these

groupings due to chance, so we look for environmental factors.



As one example, my mother has 5 brothers and sisters, two of whom have

died of cancer.  This is twice the national average, so I suppose we

could look for a reason.  But Uncle Mart had a different type of cancer

than Uncle Mike, they died at different ages (45 versus 71), and they

had very different histories as adults.  To me, this is not a cancer

cluster - it's just bad luck.  Sad, but still bad luck.



I would suggest that, if we look hard enough, we can always find some

commonality among almost any random group of people.  This is especially

so if the group is associated by geography, genetics, or social factors.

So then the tendency is to say "This group of people have a higher

incidence of cancer and they all eat the same brand of pickles, so the

pickles must cause cancer."  By finding this "association", we can feel

happy that we "know" what caused the cancer, and we can also feel happy

that, by avoiding that particular "cause", we are making ourselves a

little safer.  It's like throwing all your darts at the wall and drawing

the target to give yourselves the highest number of points.



So I don't trust cancer clusters except when they're identified by

someone trained in epidemiology who has no stake in the analysis.



Andy



-----Original Message-----

From: owner-radsafe@list.vanderbilt.edu

[mailto:owner-radsafe@list.vanderbilt.edu] On Behalf Of John Jacobus



Cancer and other medical clusters are not that

unusual.



http://cis.nci.nih.gov/fact/3_58.htm

and

http://www.cdc.gov/nceh/clusters/



If you throw a number of markers onto a chess board,

you probably will find a cluster in one or more of the

squares.