E-mail from David Robinson:

Hi all

I’m sorry I missed the call last night

I like to try to keep things as simple as I can - out of necessity as I’m not a deep thinker - but like Phil I’m struggling here, but probably in different ways - which in itself is an issue

To me the crucial statement for co-occurrence Bruce’s document is

Thus, in deciding whether CLPa and CLPb are co-occurrent, the question becomes can either life phase persists after the other has resolved
If the answer is no, then CLPa and CLPb are co-occurrent.

In the embolus & infarction example I think it is clear that the infarct can persist after the embolus has been treated so Causation 1 - not co-occurrent

Equally I find Intestinal obstruction co-occurrent and due to decreased peristalsis to be straightforward, when the peristalsis resumes so the obstruction resolves

But there are harder cases

Hydronephrosis due to bladder obstruction - can the hydronephrosis persist after obstruction is resolved - I am inclined to say yes as might not resolve immediately but I suspect people may disagree and if that can persist could the obstruction in the previous example also persist even if only for a short while

Where I also have a problem is in chronic disease many of which can relapse & remit & diseases that are potentially curable.

To date I have been guided by the Diabetic retinopathy example from the earlier document so this was once incurable but is now curable - & the statement in that document 'We should regard pattern 2 as the 'safest' option (co-occurrence is neither assumed nor necessary)’ pattern 2 I think equates to Causation 1 - does this still hold? I have probably erred on the side of not stating co-occurrence

For the former can the manifestation persist in periods of remission - for example

Myelitis co-occurrent and due to Sjogren syndrome or Myelitis due to Sjogren syndrome - I really do not know

For the latter taking some examples from our ICD11 work

Coronary artery disease co-occurrent with and due to diabetes mellitus - should not be co-occurrent?

But other diabetes example probably should be co-occrurrent e.g.

Acidosis co-occurrent with and due to diabetes mellitus
Coma co-occurrent with and due to type 2 diabetes mellitus

Question arising from that is do we assume every disease is potentially curable? A good example is HIV infection - treatable but not currently curable but given the effort going into it potentially curable in the future - do we state co-occurent & change them if/when a cure is found or take the ‘safe option' as in the earlier document & simply state due to?

I am slightly reassured that I am not alone by the statement in the original document 'Difficulty in distinguishing pattern 2 from pattern 3 - Probably we can never close this off..’ And I have some sympathy with Phil’s question of the value of distinguishing

Finally while writing this it has occurred to me that our work might be a useful test case - we have a little over 400 concepts that we have looked at & either provisionally modelled on the due_to attribute or flagged primitive but with ‘due to’ in our proposed FSN that are likely to be added to support the first release of ICD11 - would it be feasible to pass them by this group so that we are confident - we still have some work to do on tidying up our proposed terms although we could provide the ICD11 rubrics now

Thoughts welcome & please forward to whoever & apologies if I’ve missed any answers to my questions

My response

David, I will attempt to address your questions individually by indicating my thought processes. Please send your 400 concepts. We are planning to develop domain-specific heuristics (e.g. for AIDS, DM, etc..) and seeing this content would help.

Hydronephrosis due to bladder obstruction - can the hydronephrosis persist after obstruction is resolved - I am inclined to say yes as might not resolve immediately but I suspect people may disagree and if that can persist could the obstruction in the previous example also persist even if only for a short while

Singh S, Bahl A, Kasyouhanan FT. Rapid Resolution of Hydronephrosis after Spontaneous Passage of Kidney Stone. J Emerg Med Critical
Care. 2015;1(2): 2
Abstract
Obstructing kidney stones can present with sudden onset of
pain with varying degree of hydronephrosis. The timing of resolution
of hydronephrosis after relief of obstruction has not been clearly
demonstrated but can take days or even weeks for complete
resolution. Long-standing hydronephrosis is associated with obstructive
nephropathy and renal failure. We present a case of a patient with
renal colic due to presumed obstructing stone with immediate
complete resolution of hydronephrosis after presumed passage of a
stone.

Thus one case report of immediate, spontaneous resolution likely related to acute obstruction. For long term obstruction, resolution would be expected to be prolonged and thus as a whole, not necessarily co-occurrent

Diabetic retinopathy
Diabetes (in all forms) should now be considered potentially curable (e.g. type II with bariatric surgery, type I with islet cell transplant, secondary with removal of causative agent). In terms of retinopathy, there are several published series demonstrating that retinal disease most often persists unchanged after resolution of type II DM following bariatric surgery. Thus not necessarily co-occurrent

Myelitis co-occurrent and due to Sjogren? syndrome or Myelitis due to Sjogren? syndrome - I really do not know

Sjogren syndrome is a chronic autoimmune inflammatory syndrome that can affect multiple body sites, the spinal cord being one. In this case the myelopathy is a manifestation of the Sjogren syndrome and would not exist independent of the syndrome - thus necessarily co-occurrent

Coronary artery disease co-occurrent with and due to diabetes mellitus - should not be co-occurrent?
Correct. Atherosclerotic lesions may persist after DM II resolution with weight reduction

Acidosis co-occurrent with and due to diabetes mellitus
Coma co-occurrent with and due to type 2 diabetes mellitus
These are co-occurrent and due to. Both ketoacidosis and coma are self-limited and would not be expected to exist independently after the DM has resolved. Thus the causative condition (DM) is necessarily present during the entire resultant life phase.
What is beginning to concern me here is that dependent upon the nature of the complication, different modeling patterns may be appropriate and thus some diabetic complications will be descendants of diabetes while others will not. One advantage of the combined disorder model is the ability to capture an additional parent for purposes of data retrieval and analysis. Perhaps Stefan and Ed can comment?

Bruce