Random effects model in Malaria Journal …

The Bayesian errors-in-variables + random effects Poisson regression model I built to estimate the relationship between prevalence and incidence for our team’s (i.e. MAP’s) vivax malaria mapping ambitions has appeared in Malaria Journal recently. That’s the second epidemiology paper I’m on as second author, so once I get my equivalent paper on the falciparum prevalence-incidence relationship published (fingers crossed) I can probably start calling myself an epidemiologist.  While it wasn’t an overly complicated model to build or fit (w/ JAGS) I’d claim that it’s still more sophisticated than the norm for regressions in astronomical studies: for starters our likelihood function acknowledges the discrete sampling distribution of the response (as we encourage in this COIN paper) and ditto for the predictor.  In this case we also use a hierarchical structure with a Dirichlet prior to borrow strength between the various phenotype relapse zones: that is, we allow each to have a unique intercept and slope if necessary, but where the observations are sparse we allow multiple zones to share a common intercept and slope if warranted by the data.  If only I’d caught the typo in our description of the likelihood function in the paper where the rate is erroneously exponentiated twice (instead of just once).

s12936-015-0706-3-8

Advertisements
This entry was posted in Uncategorized. Bookmark the permalink.

2 Responses to Random effects model in Malaria Journal …

  1. Congratulations! That’s a nice looking plot. Is there any culture in epidemiology of putting papers on the arxiv, or similar freely accessible places?

    • Some colleagues have asked me about arXiv but no one really seems to use it in epi. Having said that, Malaria Journal is largely open access and you can even post online comments on the papers … a dangerous feature of the journal, perhaps!

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s