Difference between revisions of "Ask Fido"
(An expert level differential diagnosis clinical decision support system for the veterinary community) |
|||
| Line 32: | Line 32: | ||
== Future Development == | == Future Development == | ||
| − | Plans are in place after the beta testing trial has taken place to improve several aspects of the search including: natural language processing, diagnostic information, order sets, treatment suggestions and links to relevant papers on the topic. | + | Plans are in place after the beta testing trial has taken place to improve several aspects of the search including: natural language processing, diagnostic information, order sets, treatment suggestions and links to relevant papers on the topic. A public pet owner port of this system has been built [http://www.theonlinevet.co.uk The Online Vet] |
== Support and Help == | == Support and Help == | ||
Revision as of 10:10, 28 January 2016
Overview
Ask Fido aims to provide a solution as a single portal diagnostic search engine for veterinary students and practitioners. Drawing on existing knowledge bases and web resources, this application allows users to conduct a meta-search across multiple platforms using a unique clinical diagnostic interface. This evidence based approach will allow the user to access recent papers, treatments, clinical signs, descriptors, suggested diagnostic tests and even images and videos. The diagnostic interface has already ben trialled with students and proven to be intuitive and well received by students and vets.
The development of this app benefits from collaboration with significant partners in the veterinary informatics, knowledge base and technical fields. A key advantage of this approach will be that clinical coding systems used in electronic patients records will be dynamically linked to relevant clinical fact sheets and references. In addition, mobile access to this application will provide the opportunity for access to information at the point of clinical care.
Whilst the application has been initially designed for use by UK veterinary students in their clinical studies, it will have a wider relevance to new graduates and clinicians in practice. Given the common approach to international veterinary practice, a wider relevance for global practice is anticipated. The commercial potential of this market will underpin the intention to provide a sustainable business plan at the end of the pilot project.
Current Development
The website is still in beta testing with the aim to improve the diagnostic accuracy of clinical results. It relies on data accumulation from user searches and user submitted feedback to refine the analytics and improve accuracy. The feedback goes straight back into development and improves the search accuracy.
How to use it
The system relies on three parts to conduct a clinical search.
1. Species - The database contains information on 8 different clinical species: Avian, Bovine, Canine, Equine, Feline, Caprine, Porcine and Ovine. This section has to be completed.
2. Signalment - This is an optional section. To further refine results, the system allows for the optional addition of a patients clinical information first. This allows information such as sex, age, onset of clinical signs and neutered status to provide tailored results depending on the patient.
3. Clinical Signs - This is where users can input clinical signs. It includes VeNom clinical coding to make search terms more relevant. There is in total 1000 different clinical signs to search with and use. the search bar utilises an autocomplete function so will try to match what the user has started typing.
Once this information has been submitted, the terms are then matched with all the relevant diseases in the database and displayed in a diagnosis card format. The 'card' term is used as users can delete and reorder the list of cards to suit their need. There is a refine button that suggests potential clinical signs to add to the search based on the resulting diagnoses, and a copy to clipboard for fast transfer into notes. With regards to the Canine and Feline searches, the prevalence data has been drawn from Vetcompass and amalgamated into the search results to order in terms of prevalence. Users can click on the cards for access to information from large trusted veterinary databases such as Wikivet and Vetstream.
The Database
The database behind Ask Fido contains approximately 4000 diseases across 8 different species and includes around 1,000 clinical signs.
Future Development
Plans are in place after the beta testing trial has taken place to improve several aspects of the search including: natural language processing, diagnostic information, order sets, treatment suggestions and links to relevant papers on the topic. A public pet owner port of this system has been built The Online Vet
Support and Help
The team headed by Charles Shirazi and in collaboration with Wikivet, Vetstream, the RVC and other Veterinary Informatic companies are always looking for help and support building this tool. Whether it be wanting to get involved in the development and working with us, or financial aid to help the development continue - please get in touch with Charles at cshirazi2@rvc.ac.uk.