Presentation (CTI)

OBJECTIVES: Health economic and outcomes research (HEOR) systematic reviews rely on labour-intensive manual extraction of tabular data. Rapid evidence production is increasingly critical for reimbursement and policy decisions, yet data extraction remains the most time-consuming stage of the review process. This study aimed to quantify the time savings achieved using an AI-assisted table-extraction workflow (Laser AI) compared with conventional manual extraction in Excel during HEOR systematic reviews.
METHODS: Nine tables representing diverse HEOR-related data types: (costs, resource use, treatment patterns, utilities, cost-effectiveness results, epidemiology, population characteristics, and transition probabilities for clinical and safety outcomes) were purposefully sampled from published reviews. Three reviewers of differing experience extracted each table once with Laser AI and, in a separate session, once with Microsoft Excel. Table-reviewer assignments were alternated to avoid learning effects. Table-level extraction time was recorded. Per-cell extraction speed (seconds/value) was calculated post-hoc by dividing table time by value count.
RESULTS: The nine tables contained 5 058 discrete values (median 378; range 112 - 1995). Median extraction time decreased from 21:43 min/table (7:10 - 77:25) with Excel to 10:58 min/table (4:19 - 43:38) using Laser AI, yielding a 50% relative reduction and an absolute saving of 10:15 min/table. Estimated per-cell speed improved from 2.6 s to 1.3 s. Because timing was captured only at table level, per-cell metrics are illustrative rather than directly measured. Observed savings were consistent across reviewers despite divergent familiarity with the software. Completeness of cell extraction was 100% in both workflows.
CONCLUSIONS: An AI-assisted extraction halved the time required to collect tabular HEOR data without loss of completeness, demonstrating clear operational benefits for systematic review teams. Future work will audit accuracy and explore learning-curves. As we transition toward a fully automated extraction process, we anticipate further improvements in time savings and efficiency.