epidemiological statistics
a very reasonable cost but, in principle, any statistical or mathematical software product can be applied to epidemiological work. As always, appropriateness for purpose and user is the main requirement. Because much of the work done in the field is driven by medical or life sciences practitioners, products with roots in those areas, such as GenStat, are popular and effective. Equally though, many medical practitioners doing active epidemiological research lack an extensive statistical training or experience. Inexperienced researchers benefit from products that offer friendly interfaces, robust black box methods, and fulsome help written for the nervous: products like Sigmaplot and Unistat are popular. The free and open source, Epi Info (see Mapping the territory) is very widely used. Diseases, of course, to misquote
Shakespeare only slightly, ‘come not single spies but in battalions’ and cholera tends often to arrive with other enteric fellow travellers, such as food poisoning by the likes of Escherichia coli and Shigella dysenteriae. From one point of view, this complicates matters. On the other hand, it means that combating one will often have the collateral benefit of blocking others – even, in many cases, when the others are ill defined or completely unidentified beyond a vague description, such as ‘diarrhoea’ or ‘tummy bug’. Either way, data analytic approaches are helpful in establishing linkages and even in suggesting black box treatments. They can also assist in the identification of prophylactic factors across similarly transmitted infections or those with similar mechanisms. Becker and others[3]
,
for example, recently used GenStat assay analysis to investigate the extent to which various dietary components interfere with toxin binding to a ganglioside receptor – a mechanism applicable not just to cholera or dysentery, nor only in humans. Becker et al specifically refer, in their
paper on inhibition of toxin binding, to ‘cholera... or colibacillosis in pigs...’ There are few areas of study closer than epidemiology to the truism that human beings are most likely to forget: that they are only one twig on the schematic tree of life. While many issues are, for the most part, confined to their ecological niche, jumping species or genus, or even domain barriers, is not a rare occurrence in the bigger picture of things. Edward Jenner’s exploitation of bovine to human crossover by the Orthopoxvirus vaccinia (as cowpox) is probably the best known fact in medical history, while hysteria over avian
www.scientific-computing.com
For each of the 10 departments in Haiti, the population is divided into ‘susceptible,’ ‘infectious,’ and ‘recovered.’ Infection spreads through contact with infectious persons within a given department and in other departments, as well as through contamination of water sources. Illustration from Tuite et al, 2011
influenza provides an equally prominent present example. It is possible, and useful, to
compartmentalise human epidemiology for routine purposes and treat inward or outward infections as exceptional transaction events. Where human populations rub up against others, however, especially where the others are genetically restricted monocultures, such as those that occur in farming, the exceptional becomes common enough to require full time study of its own. Veterinary epidemiology (as in A risky business) is vital, not only to agriculture and other managed non-human
Mapping the territory One of the most widely-used analytic and reporting tools in epidemiological practice is published by the Centres for Disease Prevention and Control (CDC), based in Atlanta, Georgia, and the de facto lead body for much international co-ordination in the field. Epi Info is a free and open-source Windows platform package, available in a dozen languages, aimed at health professionals rather than professional data analysts. It dispenses with high- end bells and whistles, which most practitioners are unlikely to require, in favour of rugged, low maintenance implementation of straightforward controls and methods accessible to users without a deep background in statistics. Central to its functionality, in the true epidemiological spirit of Dr Snow, is GIS-based data mapping. A comprehensive downloadable collection of ESRI format geospatial vector
populations, but also as a gatekeeper for breaches in the membrane between those populations and human health. Just as epidemiology isn’t limited to
human health concerns, nor need it be confined to organic disease. A considerable part of medical effort is concentrated, for instance, on dealing with deliberate harm done by one human being to another – either collectively, as in the study of war and peace[4]
in the last issue, or on a
smaller civilian scale. There is, therefore, a flourishing study area in the epidemiology of violence, particularly of the criminal or domestic varieties.
shapefiles is available for most administrative boundaries across the world. Data is stored in Microsoft Access MDB files, making it accessible to most other analytic alternatives, although Epi Info does offer more than enough power, sophistication and fetch to meet the probable needs of its intended user base.
A number of demonstrations are available
for interfacing Epi Info with Google’s Mesh4x. This promises tantalising opportunities for either symmetric or asymmetric cloud-based collaborative epidemiological projects between users participating at widely different scales and with disparate technologies. The level of hands-on computing commitment required to make this work is likely to make most natural Epi Info users blanch, but that will almost certainly change with time to provide off-the-peg connection kits.
JUNE/JULY 2011 13
Page 1 |
Page 2 |
Page 3 |
Page 4 |
Page 5 |
Page 6 |
Page 7 |
Page 8 |
Page 9 |
Page 10 |
Page 11 |
Page 12 |
Page 13 |
Page 14 |
Page 15 |
Page 16 |
Page 17 |
Page 18 |
Page 19 |
Page 20 |
Page 21 |
Page 22 |
Page 23 |
Page 24 |
Page 25 |
Page 26 |
Page 27 |
Page 28 |
Page 29 |
Page 30 |
Page 31 |
Page 32 |
Page 33 |
Page 34 |
Page 35 |
Page 36 |
Page 37 |
Page 38 |
Page 39 |
Page 40 |
Page 41 |
Page 42 |
Page 43 |
Page 44 |
Page 45 |
Page 46 |
Page 47 |
Page 48 |
Page 49 |
Page 50 |
Page 51 |
Page 52