Case #672

C671 Deceased during daycare visit

Added by Ralph Dahlgren 9 months ago. Updated 6 months ago.

Status:Further activeStart date:2020-01-30
Priority:MinorSpent time:-
Target version:-
Initiator:Sweden Target year:2021
Case type:Major Owner / responsible:National organisations
MDC:GEN Old forum status:
Target Grouper:SWE


2020-01-30 Ralph Dahlgren
Deceased during daycare visits are now grouped to in care DRGs. These patients in Sweden will mostly be from the emergency wards.
When out patients are at daycare visits and dies during the visit there are a problem how to track these patients in PAR (Swedish patient register) and KPP (Cost Data register). They cannot be found actually.

When some patients arrive at the Emergency Ward they are already dead at arrival but is still admitted to the Emergency Ward, other dies in the Emergency Ward even after treatment.
These patients are handled the same in our administration system no-matter where death occurs. They will be handled like in patients. The Swedish administration system do not allow a patient to be signed out as dead at the moment due to the rules in the Swedish DRG-system.
The reason for handling these patients in different ways is multifactorial. Partly because it looks different at our Emergency Wards. It is different for big city hospitals compared to small rural hospitals. Added to that in certain situations, one ‘must make', from an ethical perspective, different decisions.

As far as I understand it was decided, this was before I started to work with DRG, that the patient that dies during an out patient visit was as resource consuming as an in-patient. This might be true in some cases but it might also not be true in others.
As it is now in Sweden we cannot even begin to investigate these patients.
Swedish Department of Health and Welfare 2020-01-28
Sweden cannot get any secure data from either the Swedish Patient register (PAR) or our Cost data base (KPP) concerning the amount of patients involved or the resources that is used.
It is actually also an administrative problem at the moment.
In Sweden there are different ways of handling these patients in different counties, even within different counties. I mostly involves the Emergency ward and is managed differently in different Regions / County councils in Sweden the DRG system must be adapted to the reality we have.
So this is a problem that should be solved.
In the Swedish DRG system, if a daycare patient is discharged as deceased after a visit the patient is grouped into a DRG for in care patient.
In our data bases we can get the numbers involved, to divide in complication levels are not possible. The numbers that we can find right now to low. We think that will be possible in the future with the new DRG.

Swedish Department of Health and Welfare 2020-01-28
Sweden wants to create a new DRG for out patients who dies during a daycare visit.
The DRG should be located high up in the hierarchy.
If located high up in the Swedish DRGlogic all patient that dies during the daycare visit will be grouped to this DRG.
For the time being we want everyone of these patients to be grouped to one DRG when they are discharged as Deceased.
This also means that the very few patients that dies during day-surgery also will be grouped to this new DRG if not major procedures are done, they are excluded.
Mostly the cases will be at the emergency wards. In Sweden this is where the patients mostly are pronounced dead at outpatient situations.
For Technical changes see ‘Technical changes C761 Deceased during daycare visit.xlsx
Decided changes
Swedish Department of Health and Welfare 2020-01-28.
Sweden has made three rows for the Swedish DRGlogic to make sure that all the patients outpatients that dies during a visit will be grouped to this new DRG (When not having OR 1 procedures done).
Sweden will implement this new DRG from 2021.

DRG change
All out-patients that dies during a daycare (except those with OR 1) will be grouped to this new DRG X09O Deceased during open daycare visit, specialized care visit. It will effect some in care DRGs which we cannot foresee right now. Resources that will be allocated cannot be found right now.

Technical change
Sweden will create three new rows in our DRGlogic table. These three rows will lead to the same DRG.
Four Swedish procedure codes: GD006 (WXYB00) Konstaterande av dödsfall (i), XV011 (ZYZX94) Visning av avliden (i), XV013 (WXYX90) Vård av avliden inför eventuell organdonation, QX007 (WXYX90) Omhändertagande av avliden (i) will get a new PROCPR 40V60 Deceased during daycare visit.
Nine Swedish procedures codes: DF010 (FXXA00) Elkonvertering, defibrillering UNS (i), DF012 (WAA902) Toraxkompression, DF017 (FXXA05) Mekanisk toraxkompression (ej manuell), DF025 (WAA902 ) Elkonvertering, defibrillering av kammararytmi, DF026 (WAA902) Elkonvertering, defibrillering av förmaksflimmer, planerat, DF027 (WAA902) Elkonvertering, defibrillering av förmaksflimmer, akut, DF028 () Hjärtlungräddning (i), DG017 (WAA900) Trakeal intubation, DG018 (WAA704) Trakeal intubation, fiberendoskopisk will get a new PROCPR 40V62 Intervention for deceased during daycare visit.
Two new proc pro 40V60 and 40V62
The rows will look like below:
ID drg english rtc icd or procpro1 secproc1 disch dur
000D0000007 X09O Avlidna u öppenvårdsbesök O 0 + 40V60 40V62 <1
000D0000008 X09O Avlidna u öppenvårdsbesök O 0 + 40V60 <1
000D0000009 X09O Avlidna u öppenvårdsbesök O 0 N E <1
The last row has OR N and disch E, this to include all daycare out-patients but not those with major surgery.
See the file: ‘Technical changes C761 Deceased during daycare visit.xlsx’

Technical changes C761 Deceased during daycare visit.xlsx (18 KB) Ralph Dahlgren, 2020-01-30 13:45

Technical changes case #672.xlsx (40.1 KB) Martti Virtanen, 2020-03-25 08:43

Technical changes case #672-2.xlsx (40.2 KB) Martti Virtanen, 2020-03-31 16:13

2020-04-22_672_TC_from_NDMS.xlsx (19.6 KB) Kristiina Kahur, 2020-04-22 16:14


#1 Updated by Martti Virtanen 8 months ago

  • File Technical changes case #672.xlsx added

2020-03-04 Martti Virtanen
The basic idea of this proposal is interesting and something that has bothered at least me from the beginning of the use of outpatient DRG’s. Sometimes patient who die early are quite expensive, sometimes they cause almost no cost.

Patient administrative data includes the information of the death of the patient as a form of discharge. This should also be available for outpatients.

There are intervention codes (used mostly by Swe) that clearly indicate death. These are the (3 NCSP+) codes listed by Swe (See technical changes NCSP+ sheet)

However, there are also (14 ICD+) dx’s that indicate death (See technical changes ICD+ sheet)

If we rely on patient administrative data, we do not need any other indication of death. The last of the proposed rules will cover all cases excluding the cases with real surgery.

If we do not rely on patient administrative data, we can use other indicators of death in the data. The cheapest deaths are those patients who arrive nearly death at the hospital and should be recorded with one of the dx’s listed in ICD+ sheet of technical changes. They do not necessarily get any other coding.

The intervention codes listed by Swe also indicate death. If a patient dies from a known cause (not MI that has specific codes for that situation) the dx may be selected in a way that does not reveal death. Thus these codes may add some cases to the cases that died.
Resuscitation is a resource intensive activity. Thus the cases with resuscitation before death are more resource intensive than cases without resuscitation. However, if the case are assigned (by the first proposed rule) to the same group as patient without resuscitation (by the second proposed rule) the first rule is unnecessary since the second rule will assign the deaths to the same case in any case.

Do countries have any economic data on patients who die at admission?
Swe proposes that death after surgery is assigned to inpatient DRG’s as before. However, this is valid only for cases detected by administrative data. Is that OK?
Swe proposes that patients who die at admission with resuscitation or without resuscitation are assigned to the same new DRG. Should resuscitated cases have their own DRG?
Is the patient administrative data reliable for death and can it be used without secondary check of possible deaths?

#2 Updated by Martti Virtanen 8 months ago

  • Status changed from Active to Further active

2020-03-09 Expert group
Expert group accepted the change for Swe.
The rules will detect different cases although the last rule would probably detect all cases belonging to this group. Sweden plans to analyze the effects by rule and later remove unnnecessary rules or create subgroupes.
The list of interventions and dx associated will also be controlled.
The case will retain open to remind for the need of check-up.

#3 Updated by Martti Virtanen 7 months ago

  • File deleted (Technical changes case #672.xlsx)

#5 Updated by Kristiina Kahur 7 months ago

Nordic Casemix Centre/Kristiina Kahur 26-3-2020

There are couple of comments regarding the procedure codes this change is concerning.
In 'proc' sheet, 13 different codes are listed. Given that in NDMS the features of procedure and diagnosis codes are handled through plus (icd+ and ncsp+) codes and not through national ones, all national codes linked to certain plus code get the feature under question too.

While making the changes in NDMS, I found some Swedish codes which should be added PROCPR 40V62 given the current mapping of the codes.
Here is the original list and respective ncsp+ codes :
code code+
GD006 WXYB00
XV011 ZYZX94
XV013 WXYX90
QX007 WXYX90
DF010 WAA902
DF012 FXXA00
DF017 FXXA05
DF025 WAA902
DF026 WAA902
DF027 WAA902
DF028 WAA900
DG017 WAA704
DG018 WAA707

There are two national codes which are mapped into ncsp plus codes but are missing in original list of technical changes:
ZXG05 Sluten hjärtkompression -> FXXA00 Closed cardiac compression
DF030 EECP-behandling (i) -> FXXA05 Mechanical heart compression

If this is a matter of incorrect mapping, it should be corrected as a separate issue.
Otherwise Swedish codes ZXG05 and DF030 should be added PROCPR 40V62 too.

Could you provide the feedback about those two codes (ZXG05 and DF030) at your earlier convenience so that we can proceed with making the changes in NDMS.

In addition, could you also clarify if row #00D000004 is missing dgprop 99X50?

#6 Updated by Martti Virtanen 7 months ago

2020-03-31 Martti Virtanen
The discussion at the expert group was incomplete because the technical changes table I had produced was corrupted.
However, the Swedish proposal was in principle accpeted.

In the technical changes I used grouping propoerty 40X00 'Death' instead of procedure property 40V60 'Intervention on death patient'. This is only technical but it is logical since 'Death' is really a condition, not intervention.
Then I have added this property to the dx codes that indicate death (see dg-table) It would seem logica.

Technically the two codes noted by Kristiina will be added to the list of codes with 40V62. They seem to be of at least similar severity and I think their addition is Ok.

The technical change is logical and allows later correction of the rule if needed. If Swe does not want the dx to affect the grouping, we will not add them now.

I have updated the technical changes accordingly and marked the changes

I hope Sweden will react to this proposal within a week.

#7 Updated by Ralph Dahlgren 7 months ago

2020-04-16 Ralph D
Thank you for your help to make the case better. Sweden agrees too with Kristiina and Martti about adding the ICD-10 codes and also the procedure codes suggested, we want it done for this year.
We do have to remind the Casemix Centre that we also want the new DRG to be implemented in the Swedish Primary Care, we have made rows for that in the original proposal. We also want to change 40X00 to get the text 'Deceased'. It is has just a softer sound in Swedish at least.
Then we want you to add a procedure code to your list also, GD018 Trakeal intubation, fiberendoskopisk, the NCSP+ code is WAA707 Fiberoendoscopic intubation, that can also be included while we're at it.
We also want your technical changes in drglogic-table to switch 40V62 from secproc1 to procpr column, we think that it was just forgotten to be done.
Sorry but this answer is later than a week because of me being on holliday for 2 weeks with the Covid-19 circulating around but so far not finding me.

#8 Updated by Kristiina Kahur 6 months ago

Nordic Casemix Centre/Kristiina Kahur 22-4-2020

This change concerns only Swedish 2021 version.
Based on the original TC and later discussion in expert group meeting and on this ticket, the updated TC (incl testcases) are attached to the ticket.
Please double-check if they look as expected. Based on the testcases, the grouping should work ok but still needs to be checked.
Please provide your feedback within 10 days.

Also available in: Atom PDF