While some industries are feeling AI fatigue, the healthcare industry is embracing and accelerating.
AI and machine learning have opened up new possibilities for innovation and improvement in healthcare by analysing large data sets, identifying patterns, and creating models—all things that individual caregivers and researchers struggle to do. By integrating AI, there has been a burst of advancements that can enhance diagnosis, treatment, and overall patient outcomes.
From detecting anomalies during pregnancy to converting medical notes into official documentation, below are several healthcare brands DEPT® is advancing with AI and ML.
Detecting heartbeat anomalies during pregnancy
Kali Healthcare is on a mission to improve pregnancy outcomes, especially for remote or disabled individuals. They do this via sensors that collect data about a woman and baby’s heartbeat and other vital signs. By analysing this data, Kali hopes to “provide insights into predicting pregnancy complications.”
We supported Kali by creating an easy-to-use digital platform experience, which ensures patients and healthcare professionals can access the information they need quickly and efficiently.
More specifically, we embraced machine learning to create a monitoring system with capabilities for real-time data analytics. The ML model can analyse large data sets to detect unusual heartbeat rhythms and then send alerts to doctors immediately.
Converting medical notes into official documentation
Doctors spend as much as 38% of their weekly working hours writing documentation. This excessive workload leads to stress, burnout, and heightened concerns about potential errors in patient records. We recently worked with a startup looking to solve this problem with a purpose-built AI tool. The tool would allow doctors to create and use templates to quickly and easily convert their notes into detailed medical documents.
There were a few things that made building this product a challenge.
- Doctors have a particular style and approach to documentation and value that ownership
- To ensure no sensitive information is at risk, all personal patient data must be de-identified before being processed
We leveraged Gen AI to de-identify data against doctor-defined templates, which minimised the medical context of Gen AI usage. Using ChatGPT 4 and AWS, we are working towards an MVP alongside our client.
Helping clinicians search faster than before
The Clair AI app from Caryheatlh is geared toward clinicians. It helps them access the latest pharmaceutical and medical information.
Considerations for this project included transparency and familiarity with AI. We needed to make sure clinicians knew how to create prompts and had confidence in the results.
To address these challenges, we balanced AI responses with the raw data powering them and added the option of follow-up questions for further justification. We also included a search function in the app that offers open-ended and guided searches to help medical professionals find information quickly.
Innovation with AI
Historically, the healthcare industry has been slow to innovate. And while skeptics abound, we believe that AI is the technology that will encourage and accelerate innovation in this industry.
As you experiment with AI, start with the following criteria for your proof of concept. You want to test AI and ML functionality that is high impact, low risk, and involves communication.
