EBM and Clinical Support Librarians@UCHC

A blog for medical students, faculty and librarians about their use of evidence based medicine, clinical literature, Web 2.0, sources and search strategies

Tag Archives: GIDEON (Global Infectious Disease Epidemiology Online Network)

Teaching & Learning in Medicine, Research Methodology, Biostatistics: Show Me the Evidence (Part 4): Causality, Airplanes and GIDEON

As accident re-enactments go, this one is pretty Riveting

Links courtesy of NJ.com and Exosphere3D - All rights reserved – Copyright 2011

.

The focus and calm of U.S. Airways Captain Chesley Sullenberger can be appreciated by listening to the audio portion of this re-enactment, as he made critical analyses – over a period of only a few minutes – on how (and where) to land a disabled aircraft sinking earthward over a densely populated area. His decisions saved 100% of the lives on board that day.

Hang out with a bunch of epidemiologists long enough and eventually the conversation while turn to Causality. They will tell you that there are some big differences in semantics between linking causation, etiology and “proof” that X exposure caused Z disease or health condition. When I looked up the terms “causation” and “causality” (on Google) there were major sidetracks, such as WBA (Why-Because-Analysis) as in, Why did the airplane crash? Why did the reactor experience a meltdown?

Why? From the viewpoint of a physician, an engineer or an epidemiologist… because of X-Y-Z.  

X-Y-Z could be bird strikes, human error, engine failure, weather conditions, crazed people bearing guns, lack of fuel or a thousand other accidents waiting to happen. Often the causes can be identified. Sometimes one can only attribute unanticipated events to gauzy, fuzzy concepts such as “it was a one in a million chance” or “this was God’s will”,  “causes unknown” or just plain Karma. The harder (non-fuzzy) data can be applied towards improvements in systems design, development of new vaccines, engineering safety… all targeted towards avoidance of future accidents (or disease outbreaks).

___________________________________

Epidemiologists save lives.  Sometimes this association accumulates slowly… as in proving associations between Exposure X and development of Disease Y decades later.  In other cases, evidence mounts up as an emergency, such as the identification of a novel virus identified as SARS in 2002.  But linking health effects of exposures over a human life-span is so much more elusive than showing evidence that birds got sucked into a jet engine at 2,500 feet (as above).

Turning to the PubMed database, a screenshot below from Medical Subject Headings List (MeSH) reveals how the term “causation” is mapped in the online thesaurus of medical indexing terms:

Image Source: NLM (http://www.ncbi.nlm.nih.gov/mesh) – All rights reserved – copyright 2011

One way to search a large database such as PubMed is to simply type in some words – for an example, Liver Cancer AND Epidemiology.  This pulls up over 18,000 retrievals…  too many (!), but by then selecting and applying standard Limit Fields such as Language, Journal Subset, Age Group, Gender and others, the retrievals can be filtered down to a more-manageable number.

A more precise way to search a large database like PubMed is to use the Medical Subject Headings list. In the example below, the term Liver Cancer was typed into the MeSH search page, which maps automatically to the preferred MeSH term — Liver Neoplasms. While this search still retrieves thousands of citation, they can be limited by selecting and applying any MeSH Subheading (or clinical qualifiers) that are appropriate to the search. These subheadings include clinical concepts such as Virology, Immunology, Genetics, Epidemiology, Transmission and 80 others.  Following is a screenshot of that type of search:

Image Source: NLM (http://www.ncbi.nlm.nih.gov/mesh) – All rights reserved – copyright 2011

Remember that a librarian’s idea of “causality” could be defined, in part, by the number or types of clinical subqualifiers selected (immunology, virology, epidemiology and those types of “background” concepts) to be combined with the formal MeSH term.

There are many ways to search. It helps a novice medical searcher sometimes to tell them just that: There is no one right way to search. Sounds enigmatic and it is.

Here is an example which I found recently in the medical literature, a 2010 article which discusses application of Bradford Hill criteria (listed here):

  • strength of association
  • consistency
  • specificity
  • temporality
  • biological gradient (dose-response)
  • biological plausibility
  • biological coherence
  • experimental evidence and
  • analogy

 Image: http://www.ncbi.nlm.nih.gov/pubmed/20644061 – All rights reserved – Copyright 2011

_________________________________

Those searching for practical answers about causality, transmissible agents, disease etiology, global prevalence or current treatments might want to search GIDEON (Global Infectious Disease Epidemiology Online Network), an interesting “niche” information source targeting the research requirements of epidemiologists, clinical & translational researchers, MPH students, toxicologists or anyone interested in tracking or diagnosing infectious diseases on a country- or world-wide scale (subscription required). Updated weekly, the database is produced by Gideon Informatics and hosted on the EBSCO platform.

A search for causality or epidemiology done in the resources indexed by GIDEON is quite unlike a search done in PubMed about the etiology of Liver Cancer. The producers collect, review and index factual data collected from around the globe; their data encompasses a wide and diverse group of human cultures, agricultural, societal, economic or environmental practices.

Below is a screenshot of one example of the type of data that can be searched on GIDEON: bacterium (causative agent), typhoid (identifiable infectious agent) and United States (location, recorded incidence over time):

.Finally, a screenshot of What’s New at GIDEON (May 11 2011):

Image Source: GIDEON – All rights reserved – copyright 2011