On January 24th, DRG’s Garima Kaul, Sr. Director Biopharma Insights, was one of four invited panelists who spoke on the application of artificial intelligence at AI Conclave 2020. The conference was organized by BML Munjal University at the India Habitat Centre, New Delhi, India and included representatives from industries such as banking, finance, and consulting. Below, Garima recaps the discussion that took place on the application of AI in the healthcare industry.
The Application of Artificial Intelligence (AI)
A person may order an item through an e-commerce website and be guided to other frequently bought items and then subsequently log-on to a social media account and see similar recommendations in their feed. What may seem like an invisible hand in the background moving items into view, is AI running pattern-based algorithms, curating data and information based on consumer preferences. This application happens in most industries, but there is a significant gap for the healthcare industry. For example, if you order paracetamol there is not an ad that appears to recommend cough syrup. Similarly, if you are interested in a drug that targets a receptor, you are not receiving recommendations for five other similar drugs. Why is the healthcare industry slow to adapt and what lies on the horizon?
The Reality of AI in Healthcare
Artificial intelligence is one of the world’s fastest-growing technologies with a forecast that is expected to grow from about $600 million in 2014 to $150 billion by 2026 across industries. Although the healthcare industry has been one of the slowest to adapt to artificial intelligence given regulations, it has countless applications in the $8.5 trillion healthcare market. The success rate of bringing one molecule to the market is less than 12% and takes 12+ years to bring a molecule from bench to bed side, with an increasing cost of no less than $2 billion. Drugs are priced to pay for the thousand molecules that fail to reach the market due to different scientific, pharmacokinetic, pharmacodynamic, or regulatory reasons. Times are changing quickly, and artificial intelligence is no longer a science fiction, but is becoming a reality for the healthcare industry. It has begun to impact the roles of providers and caregivers, bringing tremendous benefits to healthcare providers and patients.
The Challenges for AI in Healthcare
For AI to produce a useful output, it needs a significantly large amount of data beyond what can be handled manually in terms of volume and speed. The AI programs have an inherent bias (for sex, race, etc.), either due to how the program is built or because of the input data. As these are pattern-based models, it becomes difficult to generalize if the data set is narrow or the differences are minor word choices in the input data. It is important to remember that the output of data is only as good as the quality of the data inputted. The application of AI becomes challenging in a restricted and regulated industry like healthcare where there are various legal liabilities, ethical issues, data privacy concerns, patient choices, and ownership apprehensions if errors occur due to negligence. In case of a legal issue, who is to be challenged; the robot, programmer, manufacturer, or healthcare practitioner?
Success of AI in an Organization
The success of AI in any industry depends a lot on organizational behavior and orientation. A symbiotic relationship between technology, data, people, and processes is needed. Artificial intelligence has the capability to transform businesses only if it is incorporated as part of the broader strategy, across functions, departments, and geographies, rather than being limited to only technology or data scientists. Also, AI should not be taken as an approach to cut costs and increase productivity. It is a catalyst to trigger and accelerate growth and revenue in the business. Data is inherently dumb, but intelligence can bring out insights from it.
The Future of AI in Healthcare
The amalgamation of AI platforms, predictive analytics, and trillions of health record data points are helping to personalize plans and treatment for patients. Data hemorrhaging, the loss of data (clinical, genetic sequencing, patient, to image recognition) or inability to connect data points, could significantly slow down the drug development process, impact the patient journey, or cause mis-diagnosis. However, the application of AI is expected to improve the patient experience by shifting the balance from repetitive administrative responsibilities such as updating electronic records and writing reports to patient-focused tasks like consultations, better diagnosis, preventive medicine, and performing surgeries.
Artificial intelligence is truly evolving the experience for providers, payers, and patients. It is a disruptive change and there is no escape from it. All big companies are investing in technology to adapt and evolve to this vast change as it is expected to increase the capabilities in healthcare making an extension and not an extinction of healthcare professionals. Helping them to do their job more efficiently and effectively and leading to better outcomes for treatment.
To learn more about the trends happening in 2020 and beyond, please check-out our recent thought leadership pieces here. Have a question for Garima or want to learn more on the application of AI in healthcare?