We are at a turning point in how we use data and analytics in health care. Capabilities enabled by artificial intelligence (AI) are no longer just nice to have, and health care leaders agree.
For the second year in a row, we surveyed 500 health industry leaders to gauge their attitudes, perspectives and actions on the use of AI within their sector of health care. Here’s what we found:
- 62% percent of leaders surveyed report actively implementing an AI strategy, compared to just 33% in 2018 — an increase of 88%.
- Nine in 10 leaders are confident they will see a return on investment in AI — half say in less than three years. They also expect to invest more — an average of $39.7 million over the next five years. This amounts to a $7.3 million average increase over the 2018 estimate.
- Significantly, 68% of respondents report having AI/data security policies in place (compared with 48% in 2018).
As more organizations invest in and move forward with AI, the need grows for well-trained AI talent. 91% percent of leaders we asked estimate that 10% to 50% of new roles will require experience working with AI. Leaders are evenly split, however, on solutions for building AI experience within their organizations. Some plan to establish partnerships with organizations that have AI teams, while others plan to create training programs or work with consulting firms.
So how should organizations decide when to build their own AI teams and when to partner instead? Consider this three-step approach to solving the AI talent conundrum.
First: Get focused
Are you trying to tackle inefficiencies? Do you want to detect claims headed for denial? Would you like to predict disease better?
You must first define the fundamental business problem you are trying to solve. At their core, business problems aren’t technology problems. But technology — AI in many cases — can enable solutions.
The type of question you’re asking goes a long way toward informing which resources — data, technology, time, money and people — you’ll need. Open-ended, strategic questions like “where should my health system plan to build a new ambulatory center?” require very different inputs than more tactical ones like “which members would benefit from a COPD care management program?” After all, there is no “general-purpose” AI tool out there that will solve every problem. Today’s AI needs to be tailored to each question you’re trying to answer or task you’re trying to automate.
Once you know what you want to know or solve, it’s time to address one of the most critical and most competitive pieces: talent.
Next: Understand what you are starting with
If you’ve determined that AI will be a business enabler, then it’s time to start building your AI talent plan.
Half (52%) of leaders we surveyed expect AI to create more work opportunities. The majority (87%) agree that hiring candidates who have experience working with AI technology is a priority for their organization.
In terms of talent, do you have a team that understands what you can do with the data you collect? A well-staffed AI team might include:
- Head of analytics (to oversee the program)
- Data scientists (to dive deep into the data)
- A translator (someone who understands the business problem and can translate it for the data scientists)
- Solution architects
- Data architects
If you are looking to build AI capabilities, being thoughtful about where and how to access that talent — or whether to partner for that talent — will be important.
Last: Set your speed
Building an AI talent plan isn’t an overnight task. It’s more like a multi-year roadmap. That doesn’t mean you should plan to wait several years to show value, though.
Partnering early on to conduct pilots of targeted uses cases could help you build success and prove ROI. In turn, this could help as you pursue your broader, longer-term AI talent plan.
Regardless of how quickly you want to move, being able to link your plan to your business strategy will be vital as you think about your path forward.
Why plan now?
It’s true many AI use cases in health care are still considered “emerging.” That shouldn’t stop you from developing an AI talent plan.
We saw earlier that the number of organizations that are actively implementing their AI strategies has nearly doubled in a year. That’s a lot of competition for data science expertise, which is already in short supply. Compound this group with the approximately 40% of the industry that is still in the planning stages and you can see how the demand for AI talent will continue to grow for the foreseeable future.
It’s like that old axiom: The best time to plant a tree was 20 years ago, but the second-best time is today. If you haven’t yet taken the steps to explore how your organization will tap into AI talent, you should consider making it one of your priorities in 2020.
AI in health care may not have reached the same maturity level it has in financial services or retail, but it is on its way to becoming a core capability for health care organizations. It will be a must-have for driving value and taking better care of patients.
The second “OptumIQ Survey on AI in Health Care” uncovers several perspectives of leaders across the health industry — the advantages they anticipate achieving, investments they intend to make and applications they are most confident about. You can see more in the Special Report.
About the Author: Tushar Mehrotra, MBA, is senior vice president of analytics at Optum. Trained in electrical engineering, Mehrotra earned his MBA from the Wharton School at the University of Pennsylvania. He has extensive experience in strategy, growth, digital, analytics, product development and transformation. Before joining Optum, Mehrotra was a core leader in the North America health care digital and analytics practice at McKinsey & Company, a management consulting firm. He oversees 100-plus Optum analytics professionals, including data scientists, health economists and actuaries.