GPs and nurses in Merton and Wandsworth are exploring the use of an AI-based app to help detect cancer at early stages. The “C the Signs” app analyses combinations of signs, symptoms, and risk factors during patient consultations and helps zero down on patients who are more at risk of contracting cancer. Similarly, the Moorfields Eye Hospital NHS Foundation Trust has been working with DeepMind, a Google-owned firm since 2016 to help clinicians improve the way serious eye conditions are diagnosed and treated.
We’ve seen an explosion in use cases for AI technologies in healthcare, especially since the pandemic, as per a CB Insights report on ‘Healthcare AI Trends to Watch.’ Whether it is through cheaper, faster, and better MRI and CT scans or telepathology that can transform the way that pathology labs function, AI promises to play a big role, as per the report.
Not surprisingly, investments in healthcare AI by private equity players have been going up steadily. With the growing focus on trends such as personalized medicine, the role of technology will remain front and centre in shaping the future of healthcare.
The impact of AI in healthcare extends across areas such as:
-
Improved quality of diagnosis
Given that misdiagnosis could lead to potentially fatal circumstances, the promise of AI to improve the diagnostics process is very exciting. For example, the NHS is using HeartFlow’s AI technology to analyse CT scans of patients who are suspected of having coronary heart disease. Using AI, the scans are developed into a personalised 3D model that helps doctors study blood flow accurately. In radiology, computer vision can help detect anomalies in medical scans and enable better diagnosis.
-
Decision making
By providing much-needed actionable insights based on data analytics, AI and robotics can help support healthcare workers by empowering them to make better decisions. However, one major concern is ensuring that while information must be accessible, it should be protected to prevent misuse. Therefore, there need to be sufficient interventions and audits. Frameworks and governance models must be designed to allow for agility and nimbleness while protecting personal data.
-
Early detection
In serious illnesses such as cancer, delays can mean the difference between life and death. AI-powered solutions can not only aid in early detection of the disease but can also speed up the start of treatment by recommending suitable paths for treatment.
-
Faster diagnosis
A machine-learning algorithm produced by an MIT-led research team is able to analyze 3D scans up to 1,000 times faster than current speeds. Such applications could prove to be extremely valuable as image analysis is currently a tedious and time-consuming process for doctors. The Addenbrooke hospital in Cambridge, for example, is using Microsoft’s InnerEye system to automatically process scans for patients with prostate cancer. AI image analysis can also potentially allow for remote care using tools such as cameras to send information to doctors.
-
Simplifying Administration
Currently, healthcare processes are tedious repetitive, and time-consuming. Technologies such as automatic voice-to-text transcriptions can help drastically simplify or eliminate these tasks for doctors, nurses, and support staff. Of course, rather than blindly automating existing processes, the focus needs to be on eliminating unnecessary or redundant processes by enabling real-time/ straight-through processing.
Maximising AI & Robotics Adoption in Healthcare
While AI brings tremendous potential in healthcare, it also brings certain challenges, especially around privacy and ethics. Given the highly personal nature of healthcare, building trust among patients when it comes to relying on machines for diagnosis and treatment is very vital.
From a broader technology point of view, here are a few ways that healthcare providers can prepare to use AI optimally in their operations.
- Build a Data Lake: Consolidating data in the form of a data lake allows you the flexibility to leverage the data for a variety of use cases. Even if you don’t plan to use it immediately, having the required analytics capabilities in place can be very helpful.
- Establish Data Privacy Protocols: With greater automation, having in place detailed privacy protocols that are established and internalized is key.
- Training: Training all players within the healthcare ecosystem to use the information in the right way, is key. If not, it can have unintended consequences that can cause severe damage.
From the regulatory side too, Government bodies are already in the process of formulating forums, committees etc. to put in place checks and balances and governance mechanisms that can avoid the misuse of data. But it is also important to view AI adoption in healthcare from a strategic perspective and focus on maximizing its positive impact while mitigating risks.
AI, Robotics, and Machine learning are proving fundamental to healthcare for the early diagnosis of critical illness and has radically improved the accuracy of correct diagnosis. The transformative impact to provide faster services through AI enabled tools is positively impacting patients. With a combination of factors such as new innovations in AI, the growing shortfall in trained healthcare professionals and a rapidly ageing global population, AI in healthcare is here to stay.
By Subhro Malik, Senior Vice President & Head Life Sciences, Infosys
