AI in Healthcare

Don't Just Rush To Adopt New Technology Solutions

Understand the path to clinical outcomes and monetization by following these rules:

For businesses and organizations across virtually all sectors, a technology revolution is here – including Artificial Intelligence, Machine Learning, Natural Language Processing (like Alexa or Siri), and Blockchain. For the healthcare industry, all these technologies hold promise. The prospect of AI-based diagnostic systems – delivering better, faster, more accurate patient screening – could be transformative.

But caution is in order. In industry after industry, the initial response to these new technologies has been a surge of irrational enthusiasm – followed by no payoff. Leaders make two mistakes. They jump in too fast, in a spirit of “Ready, fire, aim!” – without a clear sense of what problem they’re solving or what value they want to create. And they fail to clearly define the use case.

What’s the alternative?

Technology exists to solve problems. The problems determine the project

The best-known technology projects – like IBM’s “Deep Blue” and Google parent Alphabet’s Alpha Go – were not just ambitious – they also had a clear purpose. But most technology projects aren’t nearly as sharply defined as “winning at chess” or “winning at Go.”

 To establish discipline, ask the questions that define the use case

How to proceed? Ask the questions that define the use case. “What is the problem and where can technology be applied to it?” The questions and the resulting answers define the scope:

 

Use case #1: We want the system to tell us: “Does the patient have cancer?”

 

The threshold for answering that question definitively is high. If the AI system needs to determine with 100 percent accuracy if the patient has cancer, then the tolerance for error is extremely low. Extensive work and experimentation must be done to optimize the algorithm. The FDA will likely want visibility into how the algorithm works, not to mention evidence of outcomes. Oncologists will resist it – they are being asked to surrender their judgment and experience to software.

This project has a long path to realization, with high development costs and many barriers to entry.

Use case #2: But the question could be reframed: “Is there reason to think that this patient might have cancer?  Should we increase surveillance or perform more tests?”

Now the bar is lower, and there is more tolerance for error. It will be easier, faster, and less expensive to create an AI system that conducts an initial screening. The AI provides incremental value – it is efficient at flagging more cases and getting some patients into treatment sooner. The AI doesn’t say, “This patient has cancer.” It says, “This patient might be worth a second look.” This solution will likely need less regulatory review because it supports (and does not replace) medical professionals. And for that reason, there will be less resistance among oncologists. The project, now defined by a less ambitious question, could be saving lives and generating revenue much sooner.

Five steps: a framework for monetizing your technology investment

To answer the right questions and get your technology project on track, you need a systematic framework. We propose this one:

  1. Define the use case. Identify two key variables and assign the project an overall score based on the results

     

    Variable #1: How attractive is the solution to the market? Does it solve a critical business or clinical issue?

     

    Variable #2: How feasible is the solution for technology vendors? What level of technology maturity has it reached?

     

  2. Then, connect the dots between use cases and underlying technology tools and data set

     

    Most technology projects define the data and the technology tools that will be used to leverage it. But few take the next step – defining how the data and the technology tools meet a business need. This failure to connect the dots creates multiple risks – using the wrong data, using the wrong analytic approach, and asking the wrong questions.

     

  3. For each use case, determine the total value created for a customer segment or segments

     

    What value will the solution create?  Our AI diagnostic system creates value if it improves outcomes, reduces costs, and/or frees up medical professionals to focus on more difficult, higher-value problems.

     

  4. Look at the total portfolio of technology projects and set priorities

     

    Organizations generally have several technology projects under development. Which ones can be bundled together to create a sufficiently compelling value proposition? 

     

  5. Create your platform strategy or investment roadmap. Identify the projects that are

 

Feasible in the short term (e.g. 1-2 years), using current technology to address immediate business needs.

Feasible in the medium term (e.g. 3-5 years), using technologies like AI for solutions that will be viable in that time frame and that meet emerging use cases.

Feasible in the long term (e.g. 5 years or more), using technologies that are far from maturity (such as blockchain for medical records security) and that effectively create a new paradigm – an entirely unprecedented use case.

The result will be a rational platform strategy that captures immediate opportunities – and revenue – while still planning for the farther future.

To make best use of the framework, follow these guiding principles

How to apply this framework to your own technology planning? Use these principles

 

  • The highest priority use cases must address the most critical, highest priority business issues.

     

  • The highest priority use cases should not require significantly long development times

     

  • Solving use cases and meeting customer needs is impossible without the right partnerships. You can’t (and shouldn’t want to) build it all yourself.

 

By following our framework and staying true to these guiding principles, you can rationalize, prioritize – and monetize – your technology projects.

Harsha Madannavar is a Managing Director and Partner of global management consulting firm L.E.K. Consulting. Based in L.E.K.’s San Francisco office, he is focused on the firm’s Technology, Healthcare Services, Telecom and Private Equity practices. He advises U.S. and global clients on issues including corporate growth strategy, new product development, corporate finance, and mergers and acquisitions.