AI 中文总结
研究临床质量语言（CQL）引擎的两种评估架构，介绍将CQL评估视为数据库问题的Mercury引擎，通过在相同硬件上用Blaze基准测试套件评估Mercury 2.0.1与Blaze 1.10.1，对比两者性能。
AI 中文摘要
临床质量语言（CQL）引擎服务于两个方向：决策支持（评估单个患者，当前）和质量测量（对人群进行评分）。综合评分很少是最终产物，因为其背后的个体判定才使得评分能够被检查、归因和采取行动。根据聚合发生的位置有两种架构：一种在引擎内对整个存储进行综合计算（无数据移动）；另一种评估单个记录并在外部进行综合汇总，保留每个中间结果。Mercury是第二种类型的专用CQL数据库引擎。它将CQL评估视为数据库问题——以患者优先的紧凑二进制编码存储FHIR，根据CQL检索内容进行索引过滤，规划器选择访问路径，CQL作为查询语言——而非对通用FHIR存储进行内存解释。我们使用Blaze自己发布的基准测试套件，在相同的AWS硬件上，针对100000名患者的Synthea语料库（1.123亿资源），将Mercury 2.0.1与最快的CQL评估器之一Blaze 1.10.1进行评估。
英文摘要
Clinical Quality Language (CQL) engines serve two axes: decision support (evaluate one patient, now) and quality measurement (score a population). The composite score is rarely the end product, because the individual determinations behind it are what let a score be inspected, attributed, and acted on. Two architectures follow from where aggregation happens: one computes the composite inside the engine over the whole store (no data movement); the other evaluates individual records and lets the composite be totaled externally, keeping every intermediate available. Mercury is a purpose-built CQL database engine of the second kind. It treats CQL evaluation as a database problem -- FHIR stored in a compact binary encoding keyed patient-first, indexes derived from what CQL retrieves filter on, a planner selecting access paths, CQL as the query language -- rather than as in-memory interpretation over a generic FHIR store. We evaluate Mercury 2.0.1 against Blaze 1.10.1, among the fastest CQL evaluators, using Blaze's own published benchmark suite on identical AWS hardware over a 100,000-patient Synthea corpus (112.3M resources).
Comments12 pages, 5 figures. More details on Engine and benchmark harness at getcql.com
