US CDC publishes provisional counts of deaths by select causes. This year, weekly deaths associated with ICD-10 codes R00-R99 increased from approximately 600 to more than 3,000. Those codes correspond to symptoms, signs and abnormal clinical and laboratory findings, not elsewhere classified. What could be driving this increase?
From the CDC (bold added by me):
Provisional estimates by causes of death are subject to some nonrandom sampling error. This is because the delay in receiving the report of a death depends on the cause of death. The quarterly provisional estimates are based on data that is more incomplete for the most recent months. Causes of death with more delayed reporting tend to be underrepresented in the sample, so the weighting scheme tends to underestimate their rates in the most recent months.
Furthermore, for some deaths, the final cause may not be available at the time the provisional estimates are computed. In those cases, the causes of death may be reported as unknown or pending investigation and coded to the category Other ill-defined and unspecified causes of mortality (ICD–10 code R99), a subcategory of symptoms, signs and abnormal clinical and laboratory findings, not elsewhere classified (ICD–10 codes R00–R99). In the final data, some of the deaths of unknown cause will be reassigned to specific causes if further, more specific cause-of-death information is provided.
I think what you are seeing isn't that R00-R99 deaths are actually increasing, but that in these provisional data, more recent weeks have more deaths that are temporarily listed as R00-R99 because a specific cause has not yet been assigned.
You would expect the chart to look like this (more R00-R99 deaths in recent weeks) any time you view it, not just now. For example if you pulled the data on 2020-12-05, you'd see much higher counts with this code for weeks in Sept-Dec 2020 than you see in these data.
These are provisional data that are changing, not set in stone. If you wait several weeks/months, you'd expect the numbers to decrease and likely eventually reach the steady baseline you see. You can't conclude that there is any increase until those deaths are reclassified.