This blog summarizes remarks shared by AcademyHealth’s President and CEO, Aaron Carroll, at the latest National Advisory Council Meeting (NAC) for the Agency for Healthcare Research and Quality (AHRQ) meeting on his vision for the future of health services research (HSR). These remarks continue from yesterday’s blog outlining key priorities for HSR that NAC can consider in its recommendations to AHRQ. Today’s focus builds on the discussion of innovation, equity, and incentives with an eye toward how we fund HSR and the ways new technologies like artificial intelligence (AI) may change our work.
Creating More Flexible Funding Mechanisms to Foster Cross-Sector Activities
One of the most innovative initiatives that I have taken pride in learning about during my first three months at AcademyHealth is our Evidence-Informed State Health Policy Institute, which is comprised of three groups: the Medicaid Learning Directors Network, the State University Partnership Learning Network, and the Medicaid Outcomes Research Network.
Through this Institute, AcademyHealth brings together several organizations that are doing HSR and policy research and provides a platform for them to learn from each other, share data, and troubleshoot new interventions. In many ways, this network operates like a force multiplier, allowing participants to leverage insights from work being conducted in one state or region and apply it across other relevant contexts. Although this type of work is much more impactful, it is not usually feasible due to limited allowances in R01s to fund large meetings and convenings. Conference grants have not increased in size in years while the cost of traveling and hosting meetings has risen. These barriers make attendance cost-prohibitive, especially for smaller organizations and programs with limited budgets.
Increasing funding for regular network formation and convening, instead of funding one-off programs, should be a future direction. In the basic sciences for instance, R01s are renewed for years and years, which allows investigators to build labs and continue developing research products. We do not have those kinds of funding structures in health services research. Instead, HSR funding supports one-off programs or secondary database analysis but not much else. As a field we need to envision new and sustainable funding mechanisms and develop robust networks to continue research in the long-term.
Artificial Intelligence and Digital Healthcare Research
Discussions about the future of health services research cannot happen without acknowledging the role of AI and digital technologies in transforming our health care system. We have reached an inflection point with respect to AI. AHRQ has a long history of leadership in health information technology and data‐focused work, including the Clinical Decision Support Innovation Collaborative, which comprises a diverse group of stakeholders at the forefront of using technology to better support care teams. This collaborative, which AcademyHealth supports, provides a good example of the type of learning network we will need as we move forward with artificial intelligence. We must focus on the harms of AI and not just the benefits—which we were not always successful at when it came to early research on information technology, computerized electronic medical records and computerized physician order entry.
We also need to recognize that there are multiple types of AI being deployed now, and clarity on the different types and their functions is critical for precise regulation. As an editor at JAMA Pediatrics, almost all the research that has come across our desks so far has focused on whether ChatGPT can be used to replace physician-patient correspondence. Artificial intelligence is going to be used in those ways, but it is also going to be used in more discrete, yet ubiquitous ways like the use of AI to conduct risk assessments, assist with prognosis decisions, determine how we care for patients, and identify patients ready for discharge. AI may also be leveraged to help health systems spend less while increasing revenue. Although lowering health care spending and making it more efficient is not inherently a bad thing, AI could encourage the wrong incentives and drive decision-making that prioritizes profits over people, which is incongruent with our overarching values for improving population health.
In summary, as we talk about the future of HSR, we need to think about innovation in our methods and creativity in the topics we choose to generate evidence around. These considerations can help reinvigorate our messaging around the importance of HSR. In response to threats to defund health equity initiatives, it is important to maintain a focus on diversity, equity and inclusion in health services research. We need to expand our focus and funding programs to better support community partnerships which could help reimagine the skillsets necessary for HSR and actionable health system reform. Finally, we must be careful not to miss opportunities to inform the use and roll out of AI technologies and think about how we can most effectively study AI in the present to prepare for the future. These areas provide a real opportunity for our field to operationalize evidence and move policy forward that improves health and health care for all.