Support minCellCount == 0?

[anthonysena]

We'd like to be able to produce treatment patterns output where there is no minCellCount censoring but this check requires a value of 1 or more:

TreatmentPatterns/R/export.R

Lines 311 to 317 in 7ac1b19

	checkmate::assertIntegerish(
	x = args$minCellCount,
	len = 1,
	lower = 1,
	add = assertCol,
	.var.name = "minCellCount"
	)

Would it be possible to support minCellCount = 0? Happy to put together a PR to help if you agree with this change.

[mvankessel-EMC]

Hi @anthonysena, I'm not necessarily opposed to this change. But I'm not entirely sure why you would want the minCellCount at 0.

A couple of things to consider:

• Pathways with a count of 0 do not exist.
• Persons with no pathways will be under the "None" pathway if nonePaths = TRUE in export(). This can happen when someone does not have (sufficient; minEraDuration, minPostCombinationDuration) time on any event during the target cohort period.

[anthonysena]

Hi @mvankessel-EMC - thanks for this information and your explanation makes good sense to me.

My main motivation for filing this issue arises when running TreatmentPatterns via Strategus. In Strategus, we create the execution settings that are then passed to each module (and respective package). Other HADES packages such as CohortMethod and SCCS set the lower bound for minCellCount to 0 (CM ref, SCCS ref). So, I was hoping we might make this consistent across the packages.

That said, I see your point that censoring a pathway with a count of 0 doesn't make a lot of sense. As you mentioned, there may be specific reasons why a person is excluded based on minEraDuration, minPostCombinationDuration, etc. So, assuming a person is not removed based on those exclusion criteria, a computed pathway can have 1 or more persons contributing to it. If I set the minCellCount == 1, how does the censoring work - does it remove pathways <= minCellCount? Or does it "censor" to use a negative minCellCount like other HADES packages?