Artificial Intelligence & Real World Data: The Next Steps In Your Big Data Journey

Contributed Commentary By Arun Ghosh 

November 8, 2017 | KPMG’s CEO Survey found that some of the largest U.S. life sciences companies are sending mixed signals about their commitment to technology and trailing other industries in their plans to adopt cutting edge technologies.

On one hand, the 43 U.S. life sciences CEOs didn’t seem too interested in investing in robotic process automation (RPA), a technology that can shorten processes and make routine repetitive tasks much more efficient. Those same CEOs, however, included digitization of their business as one of the top three strategic priorities.

Cognitive technologies, including artificial intelligence and machine learning, seemed to get a stronger endorsement from top executives with 30% seeing a significant investment and 42% planning incremental investment in the next 12 months.

The time is ripe for starting healthcare & life sciences investments in artificial intelligence, cognitive computing, or deep machine learning as an opportunity to reshape business processes to be more efficient while accelerating drug development and enhancing connections to patients and prescribers. In life sciences, this technology can transform how companies operate, but also change the nature of the complex science of developing drugs. 

The back office

While the science underlying medical treatments can be revolutionary, executives at life sciences companies are struggling with administrative challenges, regulatory compliance, and effectively resourcing marketing dollars. If you look at the 2016 income statements for large cap drug and device makers, marketing and administrative costs can account for a quarter to in excess of 37% of revenue. According to our 2017 CEO Outlook survey, about 96% of life sciences CEOs are forecasting annual revenue growth of less than 5% for the next three years. Logic forces executives to look at the SGA line—selling, general, and administrative expenses—to make margin improvements to improve profitability in a low-growth environment.  Cognitive computing and RPA can redefine these tasks. 

Computers are good at taking complex, repetitive, and labor intensive tasks and automating them.   Think about many compliance-driven, back office tasks that are part of doing business as a heavily regulated, pharmaceutical company.  Accounting, regulatory filings, HR functions are filled with tasks that can easily be automated.   Most CEOs do not see a significant change in employment overall from the use of this technology, but it will create opportunities for staff to focus upon more engaging tasks. From a marketing standpoint, opportunities abound to make physician engagement more meaningful and less redundant, as some surveys suggest.

Drug Development and Improving Outcomes

Cognitive computing’s impact in developing medicines will be far reaching.  Drug and medical device makers will have enhanced capabilities to deliver more personalized and predictive medicine.  The ability to sift through structured and unstructured data can generate insights into how patients respond to medications.  From those insights, researchers can narrow patient pools to determine what drugs will be most effective.  Therefore, there is great potential to take what would have been a mediocre treatment and turn it into a great treatment, merely by getting the right drug to the right patient. 

Another area where cognitive computing can have a big impact is overcoming the world of unstructured, real-world data and turning it into meaningful information. Unstructured data comes from a variety of sources:  telemedicine consults, natural language, mobile apps, wireless sensors, and diagnostic images (x-ray, MRI, etc.).  Using these insights from real world data can help drug and device makers with post-marketing studies and developing patient and physician engagement strategies.    

Cognitive computing can lead to more precise treatment patterns and reduce what is sometimes a trial and error approach to treatment. Recent studies have shown where artificial intelligence can either expand access to care or even improve upon diagnostic capabilities of medical professionals.  Some peer reviewed journals are taking note of this, including the Stanford University study on skin cancer detection and a PLOS ONE study that showed how machine learning can help predict heart attacks. 

Why now?

The ability to incorporate a base of knowledge and process unstructured data allows computers to spot patterns create another level of opportunities for machines to “learn.”  A confluence of the need to cut costs, development of sophisticated algorithms to uncover patterns in data, and the falling price of computing power makes an easier business case for this type of investment in artificial intelligence.  As we outlined before, the margin expansion opportunities are available to those who are willing make the investment.

Automating some of the administrative roles and the savings produced can become self-evident pretty quickly.  In fact, many organizations have made small steps in using technology to automate functions.  Think about how human resources has migrated many tasks, such as benefit selection, for example, to the Web.  There are other opportunities to enhance services to employees and make them more empowered from technology.

From a clinical standpoint, evaluating potential savings from artificial intelligence can be done by assessing key therapeutic areas, technology and the potential size of the market you are serving.  These savings can come in the form of reviewing time spent on research, costs associated with the results of the research and also a review of repetitive processes that are time and cost sensitive.  Any organization will need to examine the data repository and the context of the problems to be solved by artificial intelligence.  Pilot programs are the best approach to start, so the hypothesis of their value can be better articulated.

Conclusion

Life sciences companies are admittedly looking at slow revenue growth in the coming years, especially as blockbuster drugs lose patent protection and all companies face pressures in a value-based healthcare environment.   Artificial intelligence, cognitive computing and RPA are available to prepare organizations to better adapt to this market from an efficiency and marketing effectiveness standpoint.  These technologies, when combined with real world evidence, also can stand to make life sciences organizations more effective when it comes to making better products and bringing them to market faster.  Opportunities for digital transformation are available to the organizations bold enough to pursue them.

Arun Ghosh is a Principal in KPMG’s Advisory Digital Enablement Practice and has over 20 years of opportunity identification, business development, engagement delivery and practice building experience. He can be reached at arunghosh@kpmg.com.