AI is making medical information easier to access. According to a recent KFF poll, nearly a third of US adults have used AI for health information or advice in the past year, and nearly 1 in 5 have used it to compare treatment options.
The appeal is clear. People can ask health questions in their own words and get answers almost immediately. But receiving an answer and knowing how and when to use it—and whether they can rely on it—are different things.
Health plans will see an increasing number of members using AI to evaluate their care options. That reality creates a need for plans to fill in the gaps of today’s AI capabilities.
AI introduces a new health literacy challenge
Early research suggests that access to health information from an AI tool does not guarantee understanding or better decision-making.
In a recent study, large language models correctly identified medical conditions in nearly 95% of test scenarios when evaluated on their own. However, participants using those same models identified the relevant condition in less than 35% of cases, performing no better than participants without AI assistance.
As AI becomes a common source of health advice, the challenge is strengthening health literacy—patients’ ability to access, understand and effectively use information to make health decisions.
AI accelerates patient access to information about care options, but making a treatment decision based on an open-ended query is far riskier than making a decision based on reliable, specific and evidence-based decision support within the context of a patient’s health history, goals and preferences.
For example, if a patient starts a conversation with an AI chatbot about their health history, symptoms and goals, the chatbot may suggest surgery as a treatment path. With additional prompting, AI could also explain the trade-offs of proceeding with surgery and expected recovery timelines.
However, the chatbot interaction relies on the patient in two important ways: the chatbot assumes the patient has provided a holistic view of their health history and care journey, while the patient assumes the chatbot’s information is reliable and accurate. If either is untrue, the applicability of the guidance provided is put into question.
Another important component often missing from AI-generated responses is guidance on how patients should partner with their clinical team to make care decisions. Information that prepares them to engage in shared decision-making, such as what questions to ask their doctors, results in more appropriate, less risky treatment choices.
How health plans can guide better care decisions
The value of helping members understand their care options further upstream is well established. Patient decision aids—tools that simplify medical information and explain potential benefits and risks—have been shown to improve patient knowledge and help people make decisions that better reflect their values.
AI makes the need for member-facing decision-support capabilities more urgent because patients can now encounter treatment information before speaking with a clinician. Health plans should respond in three ways:
- Reach members before the decision point: Decision support is most useful when members are still evaluating their options. Health plans can partner with a solution that uses claims data and predictive analytics to identify people approaching a treatment decision and proactively engage them before preferences harden or care is scheduled.
- Put treatment information in context: Members need help understanding how the benefits, risks and trade-offs of an option relate to their health status, goals and priorities. Structured, specific, evidence-based education that is reliable, governed and connected to a member’s care journey leads to better treatment decisions and outcomes.
- Prepare members to partner with their providers: Decision support should provide a framework for discussing treatment considerations with their clinical team. Specifically, it should help members identify questions to ask their providers so they can weigh their options rather than make decisions independently.
Incorporating proactive shared decision support programs for members has been shown to improve clinical and financial outcomes for plans.
Strengthening health literacy in the AI era
When AI puts health information at patients’ fingertips, timing matters. The earlier health plans provide meaningful support, the greater the opportunity to help members make the best choices.
By offering education and shared decision support before treatment decisions occur, health plans can help members weigh their options more fully, have better-informed conversations with their clinicians, and avoid unnecessary costs and complications.
Learn more: October is Health Literacy Month.
Opinions expressed by SmartBrief contributors are their own.
_______________
Subscribe to AI Impact: Healthcare powered by SmartBrief and SmartBrief for Healthcare Leaders, two of our more than 30 healthcare publications.