Presented by Crawford Revie, Takele Beyene Tufa, Gennaro Imperatore, Alemayehu Hailemariam, Klara Saville, Samson Leta, Tariku Jibat Beyene and Amir Ibrahim at the HEARD Project Stakeholder Workshop−PPP Models for Veterinary Service Delivery ILRI, Addis Ababa, 20 June 2019
The use of a mobile app to support diagnosis across multiple animal species and improve disease surveillance in remote settings
1. The use of a mobile app to support diagnosis
across multiple animal species and improve
disease surveillance in remote settings
Crawford Revie, Takele Beyene Tufa, Gennaro
Imperatore, Alemayehu Hailemariam, Klara
Saville, Samson Leta, Tariku Jibat Beyene and
Amir Ibrahim
HEARD Project Stakeholder Workshop−PPP
Models for Veterinary Service Delivery,
ILRI, Addis Ababa, 20 June 2019
8. D3F and the AHSO (Animal Health Surveillance Ontology)
Animal Info
D3F
Signs Diseases
Diagnostic
Engine(s)
Likely diagnosis
and TX options
TX given, case
outcome, expert
judgement
Machine
Learning
Image
Capture
Image
Bank
Case
DB
AHSOAHSO
AgroVoc
GenEpiO**
Uberon
schema.org + many more
WGS-84
Vets
AHA
Farmers?
EDDiE
App2
App3
Unique
Anim. ID
+ Blockchain /
Distrib. ledger
Dose
Estimation
Sign
Detection
External Databases
Dashboards:
- vet experts
- survl. auth.
- policy
A
P
I
9. Thank you to the funding partners for this project
Hinweis der Redaktion
We have just come back from a couple of days touring around animal health clinics in Oromia where a smart phone application is being piloted. Many of you may already know Takele, he couldn’t be with us today as he is marking DVM exam papers.
The smartphone application is primarily for clinical decision support.
The animal health assistant or vet inputs information into the application to describe the animal, its clinical history and observable signs. The app provides prompts for diagnoses which are ranked according to likelihood. Recommendations for appropriate treatment are made along with advice on correct dosing. This application is not designed to replace animal health practitioners but support clinical decision making and encourage professional development.
Each case is saved on the phone and uploaded to a cloud when there is internet access.
Algorithms that generate the diagnosis prompts were developed using a cohort of experts for each species. Cattle, small ruminants, camels and equids are included.
Initiatives to improve animal health are frequently disease or species-specific.
Key goal here ensure that a flexible approach was adopted to allow for the simple inclusion of multiple species and technology transfer to other settings/countries
In the public private partnership model of service delivery one of the challenges is human capacity.
Veterinary paraprofessionals are front line providers for animal health in remote areas. At Brooke we have demonstrated that a small amount of ongoing educational input can have a huge impact. We have mentored 3900 animal health practitioners globally with positive changes in confidence and competence.
While completing the submission form on the app practitioners change behavior to conduct a thorough clinical examination and reach a robust diagnosis. This acts as an exercise in continued professional development. It increases the ability of people to think more critically.
The average number of clinical signs reported is increased compared to paper based reporting (as you can see in the clinical exam column there is not space to record more than 2 or 3 signs and frequently it is less than this)
The application prompts people to collect more information about the clinical signs and this is automatically recorded.
The danger with paper based reporting is that there is a 1 to 2 month delay. As it is not realistic to expect a breakdown of individual cases, out of necessity it is aggregated which hides important information.
Therefore the full value of this clinical detail is not available at national and regional level. Significant animal health events may be missed. With implications for both animal and human health in the case of zoonotic infections.
Submission of cases via the application allows an automated analysis of syndromic and seasonal trends in close to real time (dependent on regular access to internet)
GPS data linked to cases allows visualization of geographical distribution
This is surveillance that works and it gives something back to the practitioner so is more likely to be sustained.
Following diagnosis recommendations on treatment selection are provided, along with dosage advice.
In similar versions of this app in human health settings advice on medicine selection has led to a significant reduction in antimicrobial use.
The next step will be to include an automated dose calculation avoiding the pitfalls of complex arithmetic which can leave to over or under dosing.
The information collected on medicine use could be used as part of an AMU surveillance system – one of the recommendations in the recent No time to wait publication.
This data could also be used for stock control, ensuring that supplies of essential animal health medicines are always available in remote areas with high numbers of livestock.