AI Intake / Convert
The next generation in process automation for drug safety
What is it?
AI Intake / Converter is the realization of a fully automatic tool that receives a semi-structured input such as a web-site form, with some text and some free-text input, and returns a full ICSR in E2B(R3) format to the internal case handling team.
How does it work?
It's a clever combination of technologies in play in the background of this service that doesn't even have it's own user interface. You either hook up a web form that already exits or get one from us that you embed on your website. The form submits a message to the API (encrypted of course) and then a number of enrichments are performed, before the final file is delivered to you completely automatic.
Imagine if there was no work left in manually typing information into a safety database and then assessing seriousness, causality etc. That is no longer a dream, it's called "AI Intake / Convert".
1
Reporting of an adverse event report
It all starts with a patient/customer or health-care professional wanting to report an adverse event report. Via a web-form on your web-site or similar, the report is filled in. The form contains some key structured data points, and also some free-text elements, e.g. Medical History. Clicking submit transfers a message with the data to the API. From here, the automation kicks in.
2
Understanding Medical History and other types of free-text
Most adverse event reporting forms have open-ended text boxes, e.g. for entering Medical History or other types of un-structured information that helps the sender of the incident describe the current and past conditions etc. AI Intake is able to understand the medical concepts in the text, and in many cases place the information into structured E2B(R3) elements. This is called natural language processing.
3
Medical Assessment by AI
Seriousness and causality assessments of the product-event pairs in an ICSR is a very important part of pharmacovigilance activities. AI Intake/convert uses big data and artificial intelligence to determine what is the most likely seriousness criteria if any and also assesses the statistical likelihood of a causal relation, based on previously known cases. At the same time, AI Intake/convert prepares signal detection counts that you may use later for your safety surveillance activities
4
Transforming everything to E2B(R3)
The initial message from the web form, the processed free-text information as well as the medical assessments are now assembled into an E2B(R3) xml file. This is done using a series of functions within AI Intake/convert. The format has been tested on major platforms such as HALOPV and Argus, i.e. to ensure that the ICSR is able to import.
5
Delivering the message to you
Finally, AI Intake sends an email or a gateway message to your case handling team, with the E2B(R3) file. This is ready to import or to upload to e.g. EudraVigilance.
In addition, another file is also sent, which includes more information, including the AI confidence scores and statistical information about the product-event pair prevalence.
The AI Intake / convert solution was developed in a public-private collaboration, in partnership with Innovationsfonden as well as with Keycore.
Want to know more, get a demo or take it for a test drive?
We are ready to provide your with further information, and can setup a demo for you, where we take you through a guided tour of the AI Intake / convert API. If you would like to take it a step further, we also have a test account for you to explore in detail.
Write to hello@insife.com - we are looking forward to helping your organization!