Salmonella Saintpaul: One Hump Or Two?

I was wondering the other day whether this outbreak was winding down or heating up. So I decided to poke around in the limited amount of data that CDC has provided in its outbreak updates since the beginning of June.

I am now firmly convinced that we are in the middle of a two-phase outbreak. The first phase, which probably peaked sometime in April/May, was likely due to tomatoes. The second phase, which is in progress right now, might very well be linked to jalapeño or serrano peppers in addition to tomatoes. Here's why.

1. The initial epidemiological investigations carried out by the state of New Mexico pointed very strongly to tomatoes sold in retail food markets as the source of the outbreak. Furthermore, two of the food market chains were selling only Mexican tomatoes at the time of the outbreak.

2. Epidemiological evidence from three recent large clusters of confirmed cases have pointed to jalapeño peppers. Among the current list of suspected foods, only jalapeño peppers were implicated in two of the clusters. No tomatoes were involved.

3. I have graphed all of the data reported by CDC. While CDC hasn't released information by date of onset, I was able to break down their data based on the number of new illnesses added to the total each day. I also analyzed separately the illnesses that were reported to have begun on or after June 1st.

This is what I found.

Series 1: Total number of accumulated cases as of each CDC update
Series 2: Number of new cases added to the total at each update
Series 3: Total number of accumulated cases with onset of June 1st or later
Series 4: Number of new cases with onset of June 1st or later at each update

Almost all of the new cases reported by CDC since the end of June had onset dates of June 1st or later. On June 30th, CDC reported that 173 people became ill between June 1st and June 20th – an average of just over 8 new illness per day. The number of new cases (with June 1st or later onset) jumped to 29 in the July 3rd report, and has stayed in double-digit territory ever since.

I am convinced that these data, if graphed correctly by date of onset, would show a clear two-phase outbreak pattern. But what would explain this behavior, and how could the outbreak strain jump from tomatoes to jalapeño peppers?

I'll suggest some possible answers to these questions in the next couple of days.