Case #696

DRG assignment for COVID-19

Added by Kristiina Kahur 3 months ago. Updated about 1 month ago.

Status:AcceptedStart date:2020-03-27
Priority:Error correctionSpent time:-
Assignee:Kristiina Kahur
Target version:-
Initiator:Nordic Casemix Centre Target year:2020, 2021
Case type: Owner / responsible:Nordic Casemix Centre
MDC: Old forum status:
Target Grouper:


This case is related to case #677.

Given the latest changes in ICD10 regarding Covid-19 codes, there is need to make the corrections also in the definition tables.
According to WHO information*, there are two codes for Covid-10 in ICD10:
U07.1 – COVID-19, virus identified (added to WHO ICD10 online browser)
U07.2 - COVID-19, virus not identified (as of March 27 am, no updates in online browser yet).

Countries may have different coding standards and therefore the country specific features of new codes (or codepairs consisting of new codes) may differ. NDMS can technically handle the differences between countries.

At this point the following features will be added to Finnish 2020 version (as Finland as announced they are going to update the current 2020 version) for ICD+ code U0710 (U0720 will be added as soon as Finnish authorities will make it official in Finland):
MDC18 Infectious and parasitic disease
DGCAT 18M04 Viral illness
PDGPROP 05P10 Circulatory and respiratory problem

Codepairs will be added to Diagnosis Features table:
J9980 * U0710
J9980 * U0720

The following features will be added to each codepair based on the current features of code J8000 Adult respiratory distress syndrome (ARDS):
MDC04 Diseases and disorders of the respiratory system
DGCAT 04M06 Pulmonary edema and respiratory failure
COMPL 04G24 Respiratory failure
DGPROP 04X99 Diseases and disorders of the respiratory system
DGPROP 15X98 Other significant problem of neonate
PDGPROP 05P10 Circulatory and respiratory problem
COMPLEX 04C24 Respiratory failure
COMPLEX 04C29 Respiratory arrest

If the codepair J1710*U0710 or J1710*U0720 is used for grouping the features of J1710 will be used.
If the main diagnosis is J1280 and secondary diagnosis whether U071 or U072, the existing features of J1280 (Other viral pneumonia) will be used.
If the main diagnosis is J8000 and the secondary diagnosis is whether U071 or U072, the existing features for ARDS will be used.

Countries can still update their 2020 tables by sending the request to NCC via this ticket (as Finland did at #677). Otherwise the changes will become effective in 2021 tables based on the countries interest and request.


TC_Covid-19_696.xlsx (16.7 KB) Kristiina Kahur, 2020-04-14 11:52

TC_Covid-19_#696_C831_SWE.xlsx (27.4 KB) Mats Fernström, 2020-04-23 16:44

TC_Covid-19_#696_C831_SWE_2020-05-07.xlsx (30.5 KB) Mats Fernström, 2020-05-07 18:09

TC_Covid-19_696_2.xlsx (28.2 KB) Kristiina Kahur, 2020-05-12 09:20

TC_Covid-19_696_3.xlsx (28.1 KB) Kristiina Kahur, 2020-05-14 09:55


#1 Updated by Malle Avarsoo 3 months ago

Estonia would like to get this update.
Codes for COVID-19 have been implemented to our IT systems already, as WHO has suggested.

#2 Updated by Kristīne Putniņa 3 months ago

Latvia would like to get this update as well.
U07.2 code for COVID-19 is being implemented to our IT system also. The U07.1 code already is there.

#3 Updated by Kristiina Kahur 3 months ago

Nordic Casemix Centre/Kristiina Kahur 27-3-2020

Could you provide the LAT name for U072.
The WHO ones are currently (not visible in online browser yet):
U07.1 – COVID-19, virus identified
U07.2 - COVID-19, virus not identified

#4 Updated by Kristīne Putniņa 3 months ago

The names for U071 und U072 in LAT:
U07.1 Covid-19, ja viruss identificets
U07.2 Covid-19, ja viruss nav identificets

#5 Updated by Kristin Dahlen 3 months ago

The Norwegian coding guidelines for Covid-19 are briefly presented here. The guidelines deviates from the rules that Sweden has implemented.

Main rules:
The code U07.1 "Covid-19 with detected virus" should only be used for detected SARS-CoV-2 after testing.
The code U07.2 "Covid-19 without detected virus" is used in a clinical or epidemiological diagnosis setting (also in probable or suspected cases of covid-19) when laboratory examination is not available or the result is inconclusive.

U07.1 or U07.2 should always be used as code for secondary diagnosis, following a specific code for the infection or symptom the patient has as the main condition.

Rules for selecting the main diagnosis are the same for these cases as for other cases.

A link to the guidelines (in Norwegian) can be found here [[file:///C:/Users/kda/Downloads/Koding%20i%20spesialisthelsetjenesten%20ved%20p%C3%A5vist%20eller%20mistenkt%20covid-19-infeksjon.pdf]]

We want the following properties for code U07.1 and U07.2 in the Norwegian version:

MDC 18 Infectious and parasitic disease
DGCAT 18M04 Viral illness
COMPL 18C99 Complicated infection

#7 Updated by Martti Virtanen 3 months ago

2020-04-01 Martti Virtanen
The guidelines for coding Covid-19 disease vary in the countries. In addition it may happen that coders (and doctors coding) do not follow the instructions. Therefore we have constructed a way to assign the cases if coding done using U071 or U072 as principal dx.

The background facts are:
Covid-19 disease may vary from very subtle respiratory (or gastrointestinal) symptoms to severe septic infection with ARDS (Adult Respiratory Distress Syndrome)
  • Most of the cases are mild
  • Probably most of the infection with the virus are asymptomatic.
  • Thus detecting the virus or coding Covid-19 does not indicate high resource use – the mean cost of these cases will be quite low. Hospitalized cases are more expensive but still the variation is huge.

For these reasons it is necessary to indicate a severe condition by coding also that condition. It can be (viral) pneumonia or it can be severe respiratory insufficiency.

If the severe condition is coded as principal dx, we have no problem with NordDRG. Then Covid-19 (U071/U072) as principal dx indicates the least severe form of the disease.

If the coding is done in ‘wrong’ order, we have to use the secondary dx’s to indicate the severity.

Property changes needed:

For this we need to give U071 and U072 a new principal dx property 04P02 ‘Covid-19’.

The codes indicating viral or bacterial pneumonia (with diagnosis category 04M08) are given a new grouping property (formerly diagnosis property) 04X10 ‘Pneumonia associated with covid-19’. The codes for viral pneumonia are J1280 and J1290 (and J1710), for bacterial pneumonia all codes in groups J13, J14 and J15 and the codes J1700, J1800, J1810, J1880 and J1890)

Of the codes with diagnosis category 04M06 ‘Pulmonary edema and respiratory failure’ J9600, J9610 and J9690 (‘Respiratory failure,…’) already have the grouping property 04X03 ‘Respiratory failure’. It seems that the codes J8000 ‘Adult respiratory distress syndrome’ and J8100 ‘ Pulmonary oedema’ should for good reason have also this property. The current use with group J44 codes would only be enhanced by this change.

In the logic we need a few new rules (depending on the version):

04P02 + 04X10 (U071/U072+pneumonia) will be assigned to pneumonia categories (DRG 089-091 depending on the version, In Swe D47 -group). Each DRG needs it’s own rule! The new rules need to be placed after the current rule of each DRG.

04P02+04X03 (U071/U072 + J800/J8100/J9600/J9610/J9690) will be assigned to DRG’s for pulmonary oedema and respiratory distress (DRG 087x, in Swe D45x). Also here each DRG needs it’s own rule! The new rules need to be placed after the current rule of each DRG.

This model fits also with the Swedish recommendation with covid-19 coding. It results in better discrimination of the different level cases than just giving the U071/U072 the pneumonia property (04M06) as was done with SARS.

This will be performed as an error correction immediately for the Finnish version. Other countries may join.

#8 Updated by Elina Hermiö 3 months ago

Finnish ICD10 is now updated and the coding standards approved and we would like to receive the updated 2020 Finnish tables.

#9 Updated by Kristiina Kahur 3 months ago

Nordic Casemix Centre/Kristiina Kahur, 14-4-2020

This is to sum up the situation with the changes regarding Covid-19.
Based on the countries' feedback, the 2020 tables were updated in EST, FIN, ICE, LAT and NOR versions. Same changes have been implemented also into 2021 tables.
The changes in national versions have slight variations, the differences are visible in attached TC, 'nat_ver' field refers to national version concerned.
This ticket remains active/open. Countries can (re)consider the changes for 2021 version and inform NCC about the updates.

#10 Updated by Ralph Dahlgren 3 months ago

2020-04-16 Ralph D
Sweden will not make new definitiontables for 2020.
We will update our definitiontables with grouping properties for the new Covid-19 virus/disease. We will let everyone know how we're thinking.

#11 Updated by Mats Fernström 2 months ago

Mats Fernström, the National Board of Health and Welfare, Sweden 2020-04-23 (Swe ID C831)
As already mentioned, we will not update the SWE definition tables for 2020. This year, patients with Covid-19 will be grouped to the existing DRGs 470J/Z40 ‘New principal diagnosis without specified grouping’ which is good enough. However, in the tables for 2021PL we will have a logic similar to Finland and Iceland but with some differences.
SWE coding rules
If the reason for the contact is Covid-19, the codes U071 & U072 must be used for principal diagnosis and any manifestations, e.g. pneumonia, must be recorded as secondary diagnoses. The codes U071 & U072 can be used as secondary diagnoses when necessary, if the reason for the contact is something else than Covid-19.
Comments on the file TC_Covid-19_696.xlsx
• There is an asterisk-dagger combination (J9980 ‘Respiratory disorders in other diseases classified elsewhere’ & U0710 ‘COVID-19, virus identified’) that we do not need in Sweden. However, we question the property 04G24 because J9980 is very unspecified and Covid-19 can have minor symptoms, so there shouldn’t be any COMPL at all. Probably the patient has to be isolated but then there is a procedure code, ZYZC00 ‘Isolation’ (in SWE DV091 ‘Smittisolering’ or DV092 ‘Skyddsisolering’), for that and if an analysis shows that cases with ZYZC00 are more expensive, we can give ZYZC00 a suitable COMPL value. Such a design is more universal since it could be used regardless of the infectious agent.
• For the same reason we question the property 18C99 for the codes U0710 & U0720.
• Otherwise we support the suggested changes in the tables DRGLOGIC and DG but we also want some extra grouping rules, see the next paragraph.
Extra grouping rules in the SWE version
• All other cases with pulmonary edema or pneumonia can be grouped into the DRGs 475/D20N ‘Respiratory system diagnosis with respirator therapy’/ ’Respiratorbehandling för sjukdomar i andningsorganen’ and 474/D21N ‘Respiratory system diagnosis with PEEP support’/ ‘Behandling med CPAP/BiPAP’ and those DRGs must be possible also for cases with pulmonary edema or pneumonia due to U0710 & U0720. We have such rules for the SWE version in our TC file “TC_Covid-19_#696_C831_SWE.xlsx”. Cases with U0710 & U0720 can be grouped to DRG 483B/ W02N ‘Intensive circulatory support’/ ‘Extrakorporeal/assisterad cirkulation utan hjärt- eller lungtransplantation’ without any extra rules because the existing rules permit any diagnosis.
• In our TC file we also have rules for Covid-19 outpatients in specialized care and primary care. See the attached file “TC_Covid-19_#696_C831_SWE.xlsx” for details.
Further discussion
Some of the patients with Covid-19 will probably develop gangrene or abscess in the lungs. We are thinking on the following diagnoses:
• J8500 Gangrene and necrosis of lung
• J8510 Abscess of lung with pneumonia
• J8520 Abscess of lung without pneumonia
• J8530 Abscess of mediastinum
• J8600 Pyothorax with fistula
• J8690 Pyothorax without fistula
If these manifestations are registered as principal diagnoses they will be grouped to DRG 079 ‘Respiratory infections & inflammations, age > 17 w cc’, DRG 080 ‘Respiratory infections & inflammations, age > 17 w/o cc’ or DRG 081 ‘Respiratory infections & inflammations, age 0-17’ (corresponding SWE DRGs are D31 & D32). What do you think of making special rules leading to these DRGs even though the principal diagnosis is Covid-19, in a similar way as it is done for pneumonia and respiratory failure in the present TC?

#12 Updated by Martti Virtanen 2 months ago

2020-04-24 Martti Virtanen
About the Norwegian model:
If U0710 and U0720 belong to complication category 18C99 they must also belong to COMPLEX 18C99. Since they should not be used as principal dx this is probably not important but technically it is possible to use them as principal dx.

About the Swedish model:
The order of codes s not clearly stated by WHO and thus this solution is legitimate.
The dagger-asterisk combinations are not much used in Sweden, so it is understandable that the idea of using them now is not popular in Sweden. However, the Swedish system allows the use of code pairs, so I suppose the pairs can be set active for Sweden also – it has very little practical significance.

The code pair J9980*U0710/U0720 is very specific and means ‘Lungdisease caused by Covid-19’ Therefore it can have 04G24. J9980 is otherwise very unspecific and cannot have any complication category.

U0710/U0720 are very unspecific and cannot be considered as complicating diagnoses. As they should be used as primary dx in Sweden, this is rather theoretical anyway. But it is obvious that they must not belong to any complication category.

The typical serious consequences of Covid-19 as listed by Sweden
• J8500 Gangrene and necrosis of lung
• J8510 Abscess of lung with pneumonia
• J8520 Abscess of lung without pneumonia
• J8530 Abscess of mediastinum
• J8600 Pyothorax with fistula
• J8690 Pyothorax without fistula

Typical clinical findings of Covid-19 are not reported to include these manifestations, see f.ex. Is there specific information in Sweden about the occurrence of these problems.

As principal dx the cases are assigned according to Dg category which in all these cases is 04M02 ‘Infections and inflammations of lower respiratory tract’. The cases are assigned to DRG’s 079-081 ‘Respiratory infections & inflammations’ (D31x-D32x in Swe). The basic group is divided by age and CC, in Swe MCC, CC and NoCC and child CC and NoCC. The groups include f. ex. tuberculosis and bacterial pneumonias in addition to the listed conditions. They include only bacterial and fungal diseases with the exception varicella and measles pneumonias that are exceptional and severe complications of these viral diseases.

DRG 087 ‘Pulmonary edema & respiratory failure’ that we used in the Finish (et al) solution consist of really more severe problems that are highly fatal. In Finland it has a broad variation. In Sweden the group is divided to MCC, CC and NoCC (D45x).

However the group 087 is in the logic below DRG 079-081 because 087 is less costly. For example in Finland the mean cost of DRG 079 (CC, adult) is 6’700€, 080 (NoCC, adult) 6’100€ and 081 (child) 7’500€. The mean cost of DRG 087 is 5’200€ (with broad variation).

Viral pneumonias are usually assigned to DRG 089-091 (adult CC, adult No_CC, child CC, child No_CC, in Swe MCC, CC and No_CC, D47x). These groups are clearly least resource demanding, mean cost in Finland is for group 089 (adult, NoCC) 4’100€, 090 (adult CC) 2’200€, 091B (child NoCC) 2’200€, 091A (child CC) 7’800€.

I expect that specific rules for severe covid-19 cases should assign cases to one DRG and not divide them by (somewhat artificial) coding differences to two or more DRG’s. Therefore DRG 087 is appealing. In Sweden, however, the corresponding DRG has 3 levels (MCC, CC, NoCC)

Because special rules are needed anyway, the DRG’s can be selected freely. Viral pneumonia (089-091/D47x) are out of question because of the lowest resource use of current cases. We can choose between group 087/D45x and 079-081/D31x-D32x.
The possible options
1)The proposal for Finland as such (to DRG 087/D45x
2)The proposal for Finland modified so that the DRG’s will be from the group 079-081/D31x-D32x.
3)The dx’s listed by Swe will get property 04X03 and assigned together with the other cases to either of the mentioned groups of DRG’s.
4)The listed dx’s and the current dx’s that currently have 04X03 are given an additional new property (f. ex. 04X04?) and then assigned to either group of DRG’s.
5)The Covid-19 cases can also be assigned to only one CC level (omitting child groups?) of the selected DRG in any of the option above.

These solutions do not demand any change of grouping program and the countries may use different options.

#13 Updated by Mats Fernström 2 months ago

Mats Fernström, the National Board of Health and Welfare, Sweden 2020-05-07 (Swe ID C831)
We don’t know if it has been reported Covid-19 cases with lung abscess yet, but it is likely that some such cases will occur (according to our medical experts) and then we think that they should be grouped to the same DRGs as other cases with lung abscess, i.e. DRG D31 ‘Infektioner & inflammationer i andningsorganen, >17 år …’ and DRG D32 ‘Infektioner & inflammationer i andningsorganen, 0-17 år …’ (DRGs 079-081 ‘Respiratory infections & inflammations’ in the combined NordDRG). To achieve that we have introduced a new DGPROP (04X11) for the codes:
• J8500 Gangrene and necrosis of lung
• J8510 Abscess of lung with pneumonia
• J8520 Abscess of lung without pneumonia
• J8530 Abscess of mediastinum
• J8600 Pyothorax with fistula
• J8690 Pyothorax without fistula
and some new rules based on that DGPROP and the PDGPROP 04P02 in our new TC. It is a similar solution as the rules with 04P02 and 04X03 or 04X10 in “TC_Covid-19_696.xlsx” that Kristiina added to the Forum 2020-04-14 and which we also have in our TC.
Martti says “I expect that specific rules for severe covid-19 cases should assign cases to one DRG”. That is of course the easiest way to handle the cases but we think anyhow that it is more logical that cases with lung abscess, respiratory failure or pneumonia are grouped like other cases with lung abscess, respiratory failure or pneumonia and that cases with respirator treatment or PEEP support also are grouped like other cases with respirator treatment or PEEP support, i.e. to DRG D20N ‘Respiratorbehandling för sjukdomar i andningsorganen’ or DRG D21N ‘Behandling med CPAP/BiPAP’ (DRG 475A ‘Respiratory system diagnosis with respirator therapy’, 475B ‘Respiratory system diagnosis with PEEP support’, respectively). A positive side effect of this principle is that we will get the same DRGs as in Norway since they have decided that U07.1 or U07.2 always shall be used as code for secondary diagnosis.
The technical changes that we will do in our version are now updated and attached.

#14 Updated by Kristiina Kahur about 1 month ago

Nordic Casemix Centre/Kristiina Kahur, 12-5-2020

The changes have been made in NDMS according to TC_Covid-19_#696_C831_SWE_2020-05-07.xlsx
Rule ID #00D000831 is probably missing value '1' in the field 'compl' given that it concerns DRG D32S Läkarbesök vid infektioner & inflammationer i andningsorganen, 0-17 år, komplicerat primärvård. The same should concern rule ID #00D000830 which is missing '1' in the field of 'compl'. This too, concerns DRG D32S.
Should this be corrected and value '1' added to both rules?

Updated TC are attached.

Please check the updated TC and provide your feedback asap (within couple of days) so that we can make the final corrections (if needed) and deliver the 2021PL0 tables according to the timetable.

#15 Updated by Mats Fernström about 1 month ago

Mats Fernström, the National Board of Health and Welfare, Sweden 2020-05-13 (Swe ID C831)
Thank you Kristiina. First about the rules with ID #00D000830 & #00D000831. There shall NOT be the value '1' in the field 'compl'. In our logic for primary care, we have some special rules for ‘complicated’ cases and the variable that is making these cases ‘complicated’ is DGPROP 50X55 that you can find in the fields ‘dgprop2’ or ‘dgprop3’. We have 38 diagnosis codes with DGPROP 50X55, all beginning with the letter Q (e.g. Q90 ‘Downs syndrome’) and also one procedure code (ZV020 ‘Användande av tolk = ZZXA00 ‘Use of interpreter’) with DGPROP 50X55. (It is well known that such cases will take much longer time.)
Otherwise, your ‘drg logic’ is OK.
In your ‘dg feat’ I miss the Swedish subcodes to J8000. They are
• J809A Lätt ARDS (oxygeneringsindex (PaO2/FiO2) = 26–40 och PEEP 5 eller mer)
• J809B Måttlig ARDS (oxygeneringsindex (PaO2/FiO2) = 13–25 och PEEP 5 eller mer)
• J809C Svår ARDS (oxygeneringsindex (PaO2/FiO2) mindre än 13 och PEEP 5 eller mer)
• J809X ARDS utan uppgift om svårighetsgrad
And like J8000 they must have DGPROP 04X03. (Unfortunately, I cannot find any corresponding ICD-plus codes in “Icd_10_plus 2020 2019-10-01.xlsx”, which is the latest version we have.)
In your ‘dg feat’ I also miss that MDC 99 & DGCAT 99M01 shall be deleted from the diagnosis codes U0710 & U0720.
Otherwise, your ‘dg feat’ is OK and the other sheets are OK as well.

#16 Updated by Kristiina Kahur about 1 month ago

Nordic Casemix Centre/Kristiina Kahur, 14-5-2020

Thank you Mats for clarification how Swedish primary care logic deals with complicated cases.
I have made corrections in TC and added OUT rows for MDC 99 & DGCAT 99M01 of diagnosis codes U0710 & U0720 and removed '1?' from logic table.
As for Swedish subcodes to J8000, all codes (J809A, J809B, J809C, J809X) have DGPROP 04X03 in NDMS because the codes are mapped to ICD+ code J8000 and there is no need to made any changes in NDMS and TC.
The updated TC are attached.
That said, we now can close this case.

Also available in: Atom PDF