Case #502

DRG 316 Kidney failure – splitting into two groups– with and without cc and dialysis

Added by Kristiina Kahur almost 5 years ago. Updated over 2 years ago.

Status:AcceptedStart date:2017-01-16
Priority:MajorSpent time:-
Assignee:Kristiina Kahur
Target version:Expert Group 2017
Initiator:Finland Target year:2018
MDC:MDC11 Owner / responsible:National organisations
Target Grouper:COMMON Old forum status:


There are two rules to get the case grouped in DRG 316: a) main diagnosis with DGCAT Renal failure (11M01), or b) main procedure with PROCPROP Dialysis (11S09).
The grouping does not take into account the presence of complications/co-morbidities (cc).

DRG 316 in Finland is a heterogeneous group with variation coefficient 153% in 2013 (# of cases 1 043), 146% in 2014 (# of cases 1 094) and 213% in 2015 (# of cases 1 019). The latter was the final call for further analysis of DRG 316.

The aim of the analysis was to find the reason for high heterogeneity throughout the years. The focus was on the cases with cc and with dialysis.
The results reveal that the cases with both, cc and dialysis, have ca 2,5 times higher mean cost than the cases without cc and dialysis – 13 062 EUR versus 5 418 EUR in 2014, and 14 038 EUR versus 4 981 EUR in 2015.
The variation coefficient decreases after split but remains relatively high though.
After trimming 1% of high-end cases the result gets better, i.e. in 2014 the V% of the cases with cc and dialysis is 102% and without 103%, in 2015 108% and 87% respectively.

Given the remarkable difference of mean cost between two groups of cases within DRG 316, and improvement of V% after the changes we suggest splitting DRG 316 into two:
a) Kidney failure with cc and dialysis
b) Kidney failure without cc and dialysis

DRG change
After the suggested change the cases in DRG 316 will be grouped to separate DRGs depending on presence of cc and dialysis.

Technical changes
Appendix 2. The technical changes may contain mistakes and have to be used with caution.

Cost analysis
Appendix 1. Cost analysis is based on five university hospital data of 2015, grouped with Finnish 2017 grouper, and data of 2014, grouped with 2016 grouper.

MDC11_Nefrology_DRG316_split_Appendix1.xlsx - Appendix 1 (11.4 KB) Kristiina Kahur, 2017-01-16 15:15

MDC11_Nefrology_DRG316_split_Appendix2.xlsx - Appendix 2 (14.9 KB) Kristiina Kahur, 2017-01-16 15:18

Appendix #502 SWE 2017_1.xlsx (15.4 KB) Mats Fernström, 2017-03-02 10:13

316_follow-up_FIN.xls - 316_follow-uo_FIN (33 KB) Kristiina Kahur, 2017-03-29 14:52

Technical changes case #502.xlsx (19.6 KB) Martti Virtanen, 2017-05-05 13:23

Appendix #502 SWE 2017_1 - modified.xlsx (16.2 KB) Martti Virtanen, 2017-05-05 13:26


#1 Updated by Kristiina Kahur over 4 years ago

  • Target Grouper COMMON added

#2 Updated by Mats Fernström over 4 years ago

Mats Fernström, NPK, Sweden 2017-03-01 (NPK ID C675)
DRG 316 corresponds in SWE to DRG M31 which is divided in "very complicated" (M31A), "complicated" (M31C) and "not complicated" (M31E). The rules can be seen in Appendix #502 SWE 2017_1.xlsx (sheet “Rules”). Our cost data with this logic can be seen in the sheet “Present data”. The cost variation in our trimmed data looks much better (59-66 %) but then we have trimmed the data rather heavily (14-23 %) compared to your 1 %. Our cost variation untrimmed is 117-146 % which is better than what you have in the present DRG 316 (compared to your data from 2015) but it is still rather high so we are interested in your suggestion. In a preliminary analysis we divided our three M31 groups in “with dialysis” and “without dialysis” (see the sheet “Divided on dialysis”). We can confirm the finding that the cases with dialysis are much more expensive. Of course we cannot have six groups for cases with renal failure, but even if so, the cost variations are rather high. We will go on with this problem and try to find out if some of the procedures with 11S09 are less cost driving. If so, perhaps they should not be grouped to complicated or very complicated groups. Furthermore we have to look at the principal diagnoses. Perhaps there are cases that should be grouped to other DRGs according to the principal diagnosis. Meanwhile we do not want any changes in the Swedish version.

#3 Updated by Martti Virtanen over 4 years ago

2017-03-09 Martti Virtanen
Both Finland and Sweden seem to have found the same very expensive group: Complicated patients with dialysis. The Swedish data indicate that this really valid only for the MCC cases. Among the CC-level complications only the outliers have a similar tendency.
It seems that the MCC/CC system should be used for this DRG also in other countries than Sweden.

#4 Updated by Kristiina Kahur over 4 years ago

Finnish National DRG-centre 2017-3-29

Based on the comments made by Mats and Martti and discussion during the Expernetwork meeting in Riga we did additional analysis to find out whether or not the MCC/CC system would explain the cost-heterogeneity within DRG 316.
First, we flagged the cases with MCC, CC and non-CC regardless of dialysis and had a look at the cost.
The results were promising:
mean cost of ALL cases in DRG 316 6697 EUR
mean cost of MCC cases 8665 EUR
mean cost of CC cases 7198 EUR
mean cost of non-CC cases 3966 EUR

Second, when we took into account the presence of dialysis (any procedure with PROCPRO 11S09) the difference between MCC and CC was not remarkable anymore:
MCC w/ dialysis 14011 EUR
CC w/ dialysis 14016 EUR

After we had merged the cases with MCC and CC w/ dialysis we got the mean cost 14 014 EUR.
The rest of the cases, i.e non-CC, w/o dialysis got the mean cost 5020 EUR.

All results of cost analysis are Appendix (316_follow-uo_FIN).

In summary, in Finnish data, there is no remarkable difference in mean cost between MCC and CC cases as long as the cases of dialysis are concerned. Therefore we see no reason to use in current case the split of DRG 316 into three new groups based on MCC/CC system but rather take into account the presence of dialysis and then and split it into two new DRGs based on MCC+CC and non-CC.
It means that we keep our original proposal and would like to make the suggested change active in Finnish 2018 grouper version.

#5 Updated by Martti Virtanen over 4 years ago

2017-03-13 Expert group
Accepted in principal.
Finland and Norway support the new split, Sweden has already done a split based on MCC/CC/NoCC model. The result is similar and both models cannot be used simultaneously because the groups become too small.
The change can be done for other version than Swe. Technical changes need some fine tuning.
The countries may decide to use the old model if they so whish. They must react to this.
The Swedish model is also available to everybody.

#6 Updated by Kristiina Kahur over 4 years ago

Kristiina Kahur/Finnish National DRG-centre 2017--5-5

I would just like to clarify the content of the groups in case it has remained unclear. The names in the table may not be the most correct ones.
It should be pretty clear with first new DRG where are the cases with kidney failure+CC+dialysis. All three conditions have to be met and fulfilled.

The name of the second new DRG in technical changes, other Excel tables and also in the text might have been misleading.
All cases in this group have kidney failure, the might have CC but no dialysis, they might have dilaysis but no CC and the might have no CC and no dialysis.

Therefore, I would suggest that the name of the second new DRG would be Kidney failure without cc and/or without dialysis. It still might confusing, therefore all other proposals are welcome for sake of clarity.

As for technical changes, I hope Martti will have a look at them and make corrections (from our recent conversation I understood, that something has to be fixed there.

#7 Updated by Martti Virtanen over 4 years ago

2017-05-05 Martti Virtanen
The change will be performed in NordDRG Fin only.
If the the idea is that all cases that were assigned to DRG 316X will now be assigned to DRG's 316A and 316B we need some modifications to the technical changes. The changes do not affect the other versions.

It is important to note, that the Swedish cost data indicate that a similar change might be usefull for NordDRG Swe also. Especially cases with MCC are clearly more expensive with dialysis than without. In CC and NoCC groups the difference is significant but less. See Appendix #502 SWE 2017_1 - modified.

The case remains further active to allow consideration by Sweden (and other countries).

#8 Updated by Martti Virtanen over 4 years ago

  • Status changed from Active to Further active

#9 Updated by Martti Virtanen over 3 years ago

2018-02-20 Martti Virtanen
If threre is no further interest to this case, it shall be closed at next Expert group meeting.

#10 Updated by Martti Virtanen almost 3 years ago

2018-03-13 Expert group
The expert group confirmed that no futher changes are needed.
Case is closed.

#11 Updated by Martti Virtanen over 2 years ago

  • Status changed from Further active to Accepted

Also available in: Atom PDF