1. Digitalisation and AI
Connectivity in drug delivery devices has been a popular topic for some time now. We already know that adherence to medication, taken routinely and compliantly with correct device preparation, is a big factor in successful delivery of drugs to the lungs. Device developers are continuing to tackle these aspects with the introduction of digital connectivity and sensing modules. When paired with a device and companion app, these technologies can support the user in achieving compliance. They also offer on-the-go monitoring of device usage, which provides very useful data for healthcare providers.
The annual DDL lecture this year from Dr Sabine Häussermann and Professor David Price went beyond the upfront user benefits of connectivity and discussed the role that Artificial Intelligence (AI) and machine learning have in supporting effective respiratory healthcare outcomes. AI algorithms can make sense of vast quantities of data with an efficiency that far exceeds human cognitive ability. There are already great examples of using AI for the diagnosis of respiratory diseases, for example with the interpretation of spirometry measurements. Here data driven AI analysis can be proven to be 85% accurate, versus human ability to generally only achieve 45% accuracy when interpreting data sets.
Dr Häussermann was clear to point out that “the key to good AI is good data” and “garbage in equals garbage out”. There is a clear emphasis on device sensing and connectivity technologies to ensure that they generate the high-quality input data that is needed to support AI systems. She also noted that for healthcare deployment the most effective use of AI is through a “combination of both the machine and the human”. AI saves valuable time on the laborious number crunching tasks, however, humans add something which the algorithm cannot, which is common sense and empathy with the patient.
AI is the natural progression for digitisation of healthcare. However, as these systems develop, we must be mindful of cybersecurity risks and a regulatory landscape that is currently struggling to keep up with the rapid pace of development. We’ve already discussed some of these cybersecurity risks which you can read here.