It’s been years since AI proponents began promising big returns on healthcare providers’ investments in the technology. The results have yet to catch up with the pitches. What’s the problem?
Analysts at McKinsey & Company may have the answer.
“Organizations are caught between enthusiasm to quickly seize the opportunity and lack of alignment on where to start, as well as general caution given the potential risks of deploying AI,” they write in a report published on November 15.
To help accelerate the use of AI to improve the consumer experience, while improving supplier outcomes, the authors propose five steps. Here are key excerpts.
1. Solve the 70% data preparation problem.
Delivering tangible value to healthcare consumers through AI requires integrated, ready-to-consume data – a “difficult task that represents, on average, 70% of the work when developing AI-based solutions,” write the subject matter experts at McKinsey. More:
“To gain meaningful insights, AI users can supplement their clinical and patient data with information on social determinants of health, patient-reported outcomes, retail purchases and health tracking tools. well-being.”
2. Focus on customer experience priorities to ensure AI success.
This is a crucial step to avoid trying to do too much at once, which could limit meaningful progress, the authors point out. “To prioritize areas of focus, it is imperative to engage cross-functional leaders across the organization,” they add.
“Clinical leaders, in particular, have direct insight into patient issues and what exactly is not working in care delivery and the consumer experience. »
3. Optimize real-time insights for AI-driven interventions.
By analyzing details such as a patient’s appointment preferences and how or when they responded to outreach, notes McKinsey, AI can tailor the timing, frequency and themes of messages to provide the most appropriate recommendations. more likely to resonate. More:
“Generation AI can further improve the effectiveness of these timed interventions through hyperpersonalized message content.”
4. Map AI risks in healthcare and develop mitigation plans.
In addition to providing transparency into data usage, organizations “can provide consumers with clear logs and documentation of AI systems, including bias mitigation strategies and training protocols such as details on population profiles used”. More:
“Mature, integrated data repositories designed to power AI can become valuable targets for cyberattacks: 2023 broke the record for healthcare data breaches, recording some 725 breaches of 500 records or more, more than the double what was reported in 2017.”
5. Improve your team’s AI capabilities.
“One way to increase the chances of successful AI implementation is to use a co-pilot model, in which employees work alongside AI tools to make incremental process improvements,” the analysts write from McKinsey. “This capitalizes on the speed and capability of AI with the checks and balances of human skills and intuition to mitigate errors and risks.” More:
“Importantly, this process includes periods of testing capabilities and collecting learnings within a small group of users before being rolled out across the enterprise. Such a test-and-learn tactic allows organizations to de-risk scaling and measure impact and adoption within existing workflows.