Agree with recommendation.  1. Add the new text (in red) to the paragraphs directly above https://build.fhir.org/ig/HL7/cqf-measures/composite-measures.html#all-or-nothing-scoring : Architecturally, environments that are already capable of calculating measures using the measure scoring methods already described in this implementation guide can readily consume composite measure specifications that use the first approach (individual-based) but would require additional support in order to calculate component-based measures. Specifically, completely generic support for component-based calculation methods would require that an environment be able to evaluate CQL logic in the Population context. However, by restricting composite calculation support to those methods specified by this implementation guide, environments can calculate composites by operating on the results of individual-level scores.  It is important to note that each scoring method should work for any subject for which the measure is developed - e.g patient or encounter. 2. Add a ballot question to solicit examples of composite measures that use a subject other than patient. 3. Change all references of "patient level linear scoring" to "subject level linear scoring"