Today I had the pleasure to give a brief insight for AI applications in medical practice and clinic for Hessischer Chirurgenverband e. V. The good news: there is always something new! Bad news? There is always something new. What do we learn out of it?
When talking about AI, it’s crucial to understand that it is directed to the past and can generated content based on existant information. This differentiates AI from humans, which are blessed with the power of immagination. At least until the day of writing this article. When considering AI, this circuumstand is important. AI and humans have their strength and both should be used where they can be played full-out: in workflows!
There are tons of AI engines and best-of-breed solutions. These dominate one another every day and it does not matter if it is based on quality, costs, usability efficiency or on any other criteria. Therefore, it is crucial to be open for any AI tool at their proper location. Your businesses workflows are the relevant part of your company and you can attach to those for integrating AI. Afterwards you can easily exchange these, based on your needs.
After finding proper spaces for AI – and defining interactions with human inputs – you can start generating, using AI, for enhancing your productivity. For instance:
- A doctor talks to a patient and creates an anamnesis.
The human interaction is expected by patientes and is key for the appointment. It must not be exchanged.
KI might listen and generate related documents. This gives a doctors more time for important things: the patient. - A doctor might do an operation, while notes are taken in the computer system.
The human is crucial for the operation.
KI might automatically generate a doctors letter and surgical report, once the operation documentation is finalized. It can provide if for a review and – once approved or corrected by the doctor – share it to relevant parties, incl. archiving it. - The bill needs to be created.
AI can create bills, based on GOÄ, EBM and h-DRG and request approval for automatic billing with insurances or stakeholders. The human is accountable and required for a approval. - Invoices need to be addressed.
During the operation, staff and rooms of supplier (clinics) might be used. These need to be payed. Therefore, AI can allign incomming invoices with operations and review, e.g. dates for proper invoicing. Also contracts existing can be considered for proper correlation. The human can approve invoices, as invoices are enriched with necessary contexts.
As the example shows: There are multiple interactions with AI possible. Billing might require another AI as creating reports, as billing requires more actual data. Invoices require higher sensitivity due to processing of sensitive internal documents. The compination as part of a workflow provides the required flexibility.
