User:Elisabetta Lambertini/Notebook/Salmonella/2010/11/01

{| width="800"
 * style="background-color: #EEE"|[[Image:owwnotebook_icon.png|128px]] Project name
 * style="background-color: #F2F2F2" align="center"|  |Main project page
 * style="background-color: #F2F2F2" align="center"|  |Main project page


 * colspan="2"|
 * colspan="2"|

Risk from Salmonella in almonds
Debugging. Slight change of plans on how to look at outcomes. Danyluk 2006 used the geometric mean of #cases/yr. It makes sense since the distribution of both Prob(ill) and #cases/yr is highly skewed. Note that zeros are seen as errors when taking the Log and eliminated from the vector, so the distribution of LogCases/yr has less data points than #cases/yr. This doesn't change the result.

Outbreak scenario: we expect a result of ~ 1500-2000 cases (93 reported cases * (100/20), assuming a reporting rate of 20%. Test:
 * point estimate across the range of back-calculated concentrations: from 200 MPN/100 g, to 444 MPN/100 g;
 * "high" distribution, using detection limit as lower boundary (1 CFU/100 ml to have + enrichment, back calculated as ~ 56 MPN/100 g)
 * "low" distribution using 0.1 as lower boundary.

Note: MPN distribution fitted as cumulative.

Assumptions:
 * the whole amount shipped btw feb and april 2001 was contaminated at the same level as the 26 samples analyzed (also try with just the amount of lots C+D)


 * }