Cate Lockhart On Data Sources and the Future of AI in RWE Evaluation

Transparency in how real-world evidence (RWE) is sourced, described, and communicated, including its provenance, quality, and limitations, can determine whether payers actually use it, according to Cate Lockhart, PharmD, PhD, AMCP's chief science officer.

In this third and final installment of her conversation with Pharmaceutical Commerce, Lockhart addresses two of the most consequential questions now facing the RWE field: how manufacturers should communicate across increasingly complex data landscapes, and what role artificial intelligence should play in RWE evaluation as the industry works to define appropriate guardrails for its use.

On the data question, payers are often sitting on significant quantities of information, Lockhart notes, but may lack the experience to evaluate the strengths and limitations of the external sources manufacturers are drawing from, whether that's electronic health records, patient-reported outcomes, or real-world registries. The FDA has similarly emphasized that data provenance, fitness-for-purpose, and an honest accounting of limitations are foundational to generating credible RWE — qualities that apply whether the audience is a regulator or a formulary decision-maker.¹ The goal, as Lockhart frames it, isn't just to share findings. It's to present data in a way that allows payers to assess how closely a study population maps to their own members. The GLP-1 category illustrates exactly why that matters: a 2025 study published in Diabetes, Obesity and Metabolism found real-world weight loss with semaglutide and tirzepatide fell meaningfully short of clinical trial benchmarks, with RWD demonstrating that many patients use lower doses and do not stick to their treatment as strictly as participants in a controlled trial might.² At the same time, RWE has surfaced emerging cardiovascular and fracture-risk benefits that controlled study designs weren't built to detect, Lockhart says.

On AI, Lockhart notes the industry has moved past the "use AI" mandate and is now grappling with questions surrounding quality, reliability, consistency, provenance, and traceability. The same standards that govern how RWE gets evaluated, she says, have to govern the AI tools used to analyze it. The hallucination problem alone warrants continued caution. The opportunity ahead isn't to limit AI's role in payer assessment, but to define it carefully before it becomes embedded in decisions that affect formulary access and patient care.

Watch the previous installments of Lockhart’s interview with PC:

1. Payers Want Real-World Evidence. Why Aren't They Using It?
2. The RWE Problem Isn't Quality. It's Communication.

References

1. US Food and Drug Administration. Real-World Evidence. FDA.gov. Updated April 2, 2026. Accessed June 25, 2026. https://www.fda.gov/science-research/science-and-research-special-topics/real-world-evidence
2. Thomsen M, et al. Real-world evidence on the utilization, clinical and comparative effectiveness, and adverse effects of newer GLP-1RA-based weight-loss therapies. Diabetes Obes Metab. 2025. https://dom-pubs.onlinelibrary.wiley.com/doi/full/10.1111/dom.16364