Wednesday, April 29, 2009

An Argument for GIS and Evaluation

Before you comment about the fact that I haven't written for a while, please understand that I'm doing this is a side gig to my day job, being a husband, and father. That being said, I had hoped to write on a weekly basis. It looks like the pace will be more along the lines of every two weeks.

Now that we've got that covered, we can move to the topic of today's blog. While I had originally planned to write my thoughts on evaluation of really complex systems (social determinants of health), a little thing came up - you may have heard of it - Swine Flu.

Now some people might be scratching their heads at this point wondering, what does Swine Flu have to do with evaluation? Well, our colleagues in the world of epidemiology are engaging in a very large evaluation of the origins and spread of the virus. It is interesting from a methodological standpoint in that they are doing both a retrospective and formative evaluation of the disease and interventions. Many tools are brought to bare - one of which that I find very useful, GIS mapping. Granted, most of the public websites are pretty rudimentary, but in aggregate, they contain useful information, that can be further "mashed up" to tell an interesting story of trends. In some cases, they report the geocoded location of a possible or confirmed case and the status of the case (e.g. patient mortality). On other maps, the counts by region are all that is offered. What all these maps provide though is a picture of the spread of the illness, possible indicators of the virulence and impact of interventions (the maps are reporting the outcomes of the intersection of both). Those applying interventions, who have additional data, not reported in the metadata attached to each point, have a better idea as to whether the interventions are working or not. This informs them as to what options are available for the future.

From a public standpoint, I had a conversation with someone here in St. Louis yesterday. The question on both of our lips was what should we be doing at this point. Should we start reducing the amount of time we have in the public? Should we send our children to childcare? What sorts of precautions should we take and to what extreme? The CDC provides general guidelines for reducing the risk of transmittion of the flu - http://www.cdc.gov/swineflu/ along with cautions about areas of high concentration of virulent cases (identified by higher mortality rates). Frankly, I also took a look at the GIS maps and realized that while we are at some level of risk due to air traffic and the timing of spring break for the high schools and colleges in the area, but we have no confirmed cases as yet and the level of spread would indicate that things are pretty quiet now. However, I am watching the trends to see if the infection rates change. I also realize that this is the start of this and that there will be the dormant summer before this really hits hard this fall. The case for prevention and treatment (and subsequent evaluation) will be much more complex (and as such, very interesting).

Here are links to the maps I've been following:
http://maps.google.com/maps/ms?ie=UTF8&hl=en&t=p&msa=0&msid=106484775090296685271.0004681a37b713f6b5950&ll=32.650649,-116.139221&spn=2.062781,3.99353&z=8
Prettier version: http://maps.google.com/maps/ms?hl=en&ie=UTF8&msa=0&msid=106484775090296685271.0004681a37b713f6b5950&t=h&ll=35.46067,-94.746094&spn=66.482866,112.5&z=3&source=embed
More "qualitative information": http://healthmap.org/swineflu
Timeline example - isn't accurate as datestamps for the some of the data is incorrect, but gives you a sense of what can be done to view spread over time: http://homepage.ntlworld.com/keir.clarke/flu.htm
My favorite static version: http://maps.google.co.uk/maps/ms?hl=en&ie=UTF8&msa=0&msid=109496610648025582911.0004686892fbefe515012&ll=26.115986,-9.140625&spn=144.408712,283.007813&t=p&z=2
An interesting map regarding travel to Mexico: http://maker.geocommons.com/maps/4884

The last map is interesting if only because it seems to reflect the case distribution of Swine Flu in the US. In other words, there might be a correlation between level of travel to Mexico and incidence rates of Swine Flu. I state might be, as I haven't crunched the numbers, nor know of a study as yet. If someone has, it would be great to see. However, it appears that what we see right now is the "first wave" of cases that were contracted in Mexico. Their impact on the regions they were identified in (in Missouri's case, right now 12 unconfirmed cases) will show up in a few days if the virus is maintaining its level of virulence and the preventative measures have not either been implemented and/or are ineffective. In any case, it will be the evaluation process and the GIS maps that will best inform the public health officials as to what is happening and allow them to better assess what is working or not.

Best regards,

Charles Gasper