Identifying and measuring palliative care in US-based healthcare claims and administrative data: A scoping review.

316 Background: Palliative care (PC) is a critical component of whole-person healthcare for individuals with serious illnesses. Administrative claims data offer a valuable resource for evaluating the delivery of PC services. However, considerable variation exists in how PC is identified using claims data. Ensuring methodological rigor in claims-based research is critical to advancing a reliable evidence base for PC delivery. This review aimed to characterize the different research approaches used to identify and measure delivery of PC. Methods: Following PRISMA-ScR guidelines, we developed a search strategy combining terms related to (1) “Palliative” and (2) “Medicare,” “claims,” or “billing” across PubMed, Scopus, and Web of Science for studies published between 1996 and 2025. We initially screened titles and abstracts, followed by full-text review for studies that met inclusion criteria. Studies were excluded if they were not in English, not US-based, limited to abstracts, not primary research (e.g., book chapters), or if they did not pertain to analysis of claims data. From the final set of included studies, we extracted citation details (e.g., title, first author’s last name, year), PC service examined, the type of administrative claim used, and the type of billing codes reported. For PC service and billing codes, we synthesized into higher-level categories for analysis. Results: Among the 115 studies included, we identified 129 distinct instances of PC measurement. Each instance was categorized as: (1) PC clinician encounter, (2) specialty PC (sPC) clinician encounter, or (3) palliative-intent therapies (e.g., chemotherapy, radiotherapy). In 69 PC instances, 45 (65%) used diagnosis codes (ICD-9-CM/ICD-10-CM) only. In 11 sPC instances, 5 (46%) used provider taxonomy-based codes (PECOS, NPI taxonomy) only. In 49 therapy instances, 26 (53%) used procedure codes (ICD-9-CM/ICD-10-PCS, CPT/HCPCS) only. In 5 (7%) PC, 3 (27%) sPC, and 14 (29%) therapy instances, measures combined strategies (e.g., diagnosis and provider taxonomy codes). In about one-fifth of the instances, the paper did not offer sufficient details on the measurement approach to classify or reproduce the strategy (PC [20%], sPC [18%], and therapies [16%]). Conclusions: This scoping review identified three types of PC service measures, each with a range of identification and measurement approaches in claims data, and many with insufficient detail for reproducibility. Robust methodological descriptions and increased consistency and transparency in reporting would enhance reproducibility, cross-study comparability, and future meta-analyses.