epidemiological statistics and Gun Ownership[5] but the focus is
epidemiological, even when the word is absent from the paper title. Firearm retailer’s willingness to participate in an illegal gun purchase[6]
is, from a data analytic
perspective (VPRP uses SAS), a study of the statistical likelihood of transmission by a given vector (the illegal sale) for a toxin (the fi rearm). There are also studies of epidemiological
Dr John Snow’s 1854 ‘ghost map’ of cholera in London (background) is recreated (foreground) in the open source epidemiological analysis software, Epi Info.
Publications from the Violence
Prevention Research Program (VPRP), run by a small but active and dedicated team within the Health Centre at University of California at Davis, include such titles as Gun Suicide by Young People in California: Descriptive Epidemiology
A risky business
The Population Biology and Disease Control research group at the Royal Veterinary College (RVC) is concerned with factors infl uencing animal health, and epidemiology lies at the core of its activities. It has a particular interest in development of new measures to deal with dangerous and infectious diseases, many of which impact both directly and indirectly on human health and welfare, providing advice to government and other organisations at national, European and international levels. As Professor Dirk Pfeiffer, head of the group, comments: ‘We deal with issues that have signifi cant health risks to animals and humans, as well as economic impact, and it is critical therefore that... we provide the various bodies that depend on our knowledge and experience with the best possible solution available to them.’
Risk, including the risk of transmission and the risk resulting from such transmission, is a central concern. At one end of the scale are risks such as incidence of disease associated with a specifi c feed, which are relatively simple to minimise; at the other lies tuberculosis or foot and mouth, which has more complex risk issues. Determination of ‘at risk’ groups and populations, maximum disease loci, development trends, and other common factors such as circumstances, temporality, exposure or lack of, is dependent on data analysis. Identifi cation of epidemiological risks through professional expertise is one half of the battle. The other is determination of responses, which involves
14 SCIENTIFIC COMPUTING WORLD
quantifi cation of those identifi ed risks. That, in turn, means Monte Carlo analysis of potential scenarios and their occurrence likelihood. As David Robson noted[7]
in connection with avian fl u in SCW
four years ago, the RVC makes use of a specifi c risk analysis package (Palisade’s Excel symbiote, ‘RISK’) for the purpose. From the results, advice is formulated on risk reduction strategies for any given disease. Inputs include likelihood of an infected inward vector in a given context and set of circumstances. To take a well known issue: how likely it is that a dairy herd might encounter a badger that carries Mycobacterium bovis (the bovine TB bacterium)? How common is M bovis in the badger population? How frequent are interactions which might enable infection, and on what other factors does that frequency depend? What is the probability, given such a contact of infective transfer? What factors in the containing biological and ecological systems infl uence this?
Risk factor calculations result in an outcome statistic that shapes decisions. Livestock slaughter on and around an infection locus in foot and mouth outbreaks, for example, is a drastic solution to be invoked only if risks associated with infection virulence and likelihood of a particular disease and circumstances combine to outweigh the costs of the policy. If the RVC’s purpose-written RISK model indicates a very low transmission probability, it may well be that other strategies, such as containment or barrier zones, are preferable.
interaction between violent behaviours and ill health of more traditional kinds. One such investigation is currently under way into the transfer of violence back to civil society from military service in zones of confl ict, including correlation between traumatic stress (aka ‘combat fatigue’) and subsequent civil violence. Another, suggested by use of role-playing
games as safe spaces within which to explore moral and practical choices in a hospital for the criminally insane, directly links outputs from such games to a worksheet for multifactorial analysis in a suite of
R routines. In this second case, precise game state and scenario data are recorded in parallel with player choice narrative, physiological signs and EEG streams. Correlations between situational triggers and responses are then mapped onto known past patterns, with metapatterns being sought across multiple respondents in an attempt to predict likely danger points. ‘Patterns of harm, actual and potential,’
as one doctor described them to me, are the basis for all medical planning. They cross every area of life and they are all, ultimately, interpretations of submodules in a global web of interrelations, in the same way as the weather. As with weather, growing computational reach and power increase understanding of the overarching mechanisms involved. Unlike weather, however, epidemiology offers considerable scope for local investigation of management of risks, plans, responses and outcomes at numerous levels. At both ends of the range, and across it, data analysis is key.
References and Sources For a full list of references and sources cited in this article, please visit
www.scientifi c-
computing.com/features/referencesjun11.php
A broader base for a
wider view King’s Health Partners is an Academic Health Sciences Centre – one of fi ve in Britain, comprising King’s College London and three NHS foundation trusts. As such, it combines research with education and frontline health care. Within this framework comes KCL’s Integrated Cancer Centre (ICC), which is in turn a partner in the Centre for Global OncoPolicy and moving towards ever-wider co-ordination of cancer epidemiology.
One of the centre’s latest developments is ORIS, the Oncology Research Information System designed at the ICC to act as a unifi ed platform for multi-disciplinary research information, unifying research data and analyses across the partner organisations. ORIS aims, in the words of ICC’s R&D lead Professor Peter Parker, to replace ‘protocols identifi ed through a process of trial and error’ with a targeted approach, based on strategic data analysis management. From the ICC design, ORIS was developed by IDBS and announced in March of this year. Apart from the anticipated immediate benefi ts, it is expected to act as a gravity well attracting other research foci into its orbit, thus further enhancing oncological integration and thereby encouraging a new level of epidemiological overview.
www.scientific-computing.com
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