Additional comment for resolution on the US Core errata package made available 15 July 2020:

The smoking status value set question caused me to review https://build.fhir.org/ig/HL7/US-Core-R4/StructureDefinition-us-core-smokingstatus.html. The current timing is listed as issued (https://build.fhir.org/ig/HL7/US-Core-R4/StructureDefinition-us-core-smokingstatus-definitions.html#Observation.issued) Issued is define as "The date and time this version of the observation was made available to providers, typically after the results have been reviewed and verified." The definition is most clearly written to address laboratory observation results or imaging study results. However, the definition for issued is most consistent with when the result is reported - not when it was obtained/recorded. Eric provided a link to a previous  documentation for using issued vs applied for smoking from much earlier CCDA to FHIR mapping project:
https://docs.google.com/document/d/1RPGMEhWdEPdtCpA8NKUQ79HkpK0fLp3l2eucQ21w778/edit.  That matrix maps as follows: 
issued = when the observation happened

appliesPeriod = period when the patient smokes or smoked

However, the definition in that discussion is not the definition of issued in US Core or FHIR - "When the observation happened" is most consistent with effectiveTime (dateTime) - when the observation occurred.  The observation may happen as a result of a questionnaireResponse or an individual question to the patient. It doesn't indicated the "appliesPeriod" which would be response to an observable that provide a startTime and stopTime of smoking - not a code - using effectiveTime (period).  Hence, US Core should not be using issued for the observation time. It should be using effectiveTime (dateTime) of the observation.  The is analogous to using effectiveTime for a laboratory test (I.e., physiologic time) and issued for the time the result was reported. Without this change, there will be a problematic disconnect between QI-Core and US Core for the same data element.