It is not to be underestimated that many of the issues that are slowing down the adoption of AI in healthcare such as fragmented data, misaligned economic incentives, legacy IT systems, and more are currently leading to less than ideal practice of medicine. That means patient outcomes that are far from optimal, waste of resources with duplicative tests and variability in decision making, inefficiencies in delivering and receiving care, care that is not longitudinal and does not assist patients in following the treatment plan fully outside the care setting, and more.
What can explain this massive waste of resources? It is the lack of standardization and the imbalance of resources that we discussed in the last section. It is also that the healthcare system was never set up to care for patients outside of the clinical setting. Increasingly, we have realized that in order to achieve the best outcomes, patients need to be assisted in following the prescribed treatment plans on an ongoing basis in their daily lives. Well, how does a model of care delivery that was set up to care for patients when they’re ill start to take on the extra task of caring for patients outside of the clinical setting?
The fact is that there is already a shortage of human resources to do what the healthcare systems have traditionally done. So, how does it take on this new model of care? Well, that is where technology and automation comes in. Emerging technologies such as sensors, software, analytics, and AI can perform many of the tasks that would be unrealistic for the current healthcare force to do at scale to improve the quality of patient outcomes. These technologies can collect data from patients, analyze it, identify opportunities to intervene, and involve the healthcare providers only when their input is needed. This would not only assist in getting more done but also would introduce intelligence and consistency to how those things are done.
Currently, any analysis of the healthcare system would show high cost of care with poor outcomes as a result of a number of factors: mistakes in diagnosis, wrong or too many treatments, poor decision making by physicians due to lack of time, variable quality of training, difficulty taking statistical factors in mind (Baysian theorem,) inadequate data (no genetic data, fragmented data at the point of care), treatments not individualized based on clinical trials or not studied for best responders, and more. (3) There are more than 12 million significant missed diagnosis per year on medical imaging. An alarmingly high percentage of imaging is unnecessary; and healthcare resources are badly stretched, especially now with COVID.
This is why there is significant opportunity to improve care using AI. The ability of AI to improve patient outcomes, reducing the healthcare costs, facilitate precision medicine, and ongoing patient management outside of the clinical setting sets it apart from other promising technologies in healthcare.