Case #296

Task #11: Delivery of minor update proposals for NordDRG 2015

Case #295: Difference in the properties of corresponding dx’s in different countries

Code pair problems

Added by Anonymous over 7 years ago. Updated about 5 years ago.

Status:AcceptedStart date:2014-03-11
Priority:MinorSpent time:-
Assignee:Martti Virtanen
Target version:Expert Group 2014
Initiator:Nordic Casemix Centre Target year:2015
MDC: Owner / responsible:Nordic Casemix Centre
Target Grouper: Old forum status:


There are 2562 cases with code pair problems that need to be adjusted, to avoid unnecessary conflicts between the country versions in NDMS.

Code pairs or dagger-asterisk pairs are a system developed by WHO to create possibility for specific coding to a number of problems by post coordinate mechanism. The dagger code indicates the etiology of the condition at issue and the asterisk code the manifestation or the symptoms. The system increases the number of possible conditions that can be coded with ICD-10 immensely.

The system is built so that, there are marked etiological codes (asterisk codes) that cannot be used alone. They must according to WHO always be combined with another code that indicates the etiology. Some codes are marked with a dagger (in practice +) to be very suitable for coding of the etiology in a pair. These codes can be used alone without any indicated manifestation. The manifestation may also (at least in principle) be any other code from the system. External causes of health problems (chapter XX) is however excluded. Some pairs are ‘fixed’ i.e. there is a asterisk code that is indicated to be used with a specified dagger code and vice versa. What is the difference of using just one of these codes is not indicated.

Especially problematic is the Danish situation. Denmark decided not to use any code pairs and added instead a number of pre coordinated codes to the Danish ICD-10. All dagger codes are retained (of cause as they be used alone) and all asterisk codes are retained (not very logical because it creates double coding options for a number of conditions). The pre coordinated codes where created sometimes based on manifestation and sometimes on etiology and sometimes both – creating again double coding. Most of the obvious errors have been removed by now.

Finnish and Swedish ICD-10 have also a few codes that correspond to a pair in the original ICD-10.
In NordDRG we have decided to define specifically properties to a number of code pairs. They have usually been listed in some of the national versions but may at this stage been dropped from the official listings. That does not make then unusable for DRG, because the ICD-10 system allows the user to define new pairs when needed. Thus these pairs may still occur in the data but they are obviously very rare as they have always been. One has to have in mind that the basic idea was to code only the less common combinations with this somewhat complicated arrangement.

In NordDRG if a code pair is used and no special rule is defined, the diagnosis category (and MDC) are defined by the manifestation (asterisk) code. In our systems manifestation code is usually coded first and followed by the etiology. With the exception of Finland the systems have no special way to indicate that the coding is intended to be a code pair. Thus it is quite exceptional that the information comes to the grouper, and usually the etiological code is handled as a secondary dg. This is not a main problem, because manifestation should define the diagnosis category anyway. Diagnosis properties come from both components of the diagnosis and are neither a special problem.


In some instances it has been decided to give the pair specific properties that neither of the components have or to use the diagnosis category (and MDC) of the etiological component. These are cases are discussed here, since in the process quite many of these situations have been (by error) dealt with in different ways in different national versions.

In the Danish setting the pre coordinated codes inherit properties from both parents. The diagnosis category (and MDC) is then inherited from the manifestation code even when the code format comes from the etiological component. This creates a situation where there seems to be a difference in the Danish properties, although it does not exist. These are the cases marked with green text in the note column. They will most probably not cause any major problem even when common NordDRG versions are used for Danish data. The Finnish and Swedish pre coordinated codes behave similarly. No change is proposed for these cases.

In other cases there effect on DRG is usually minimal. First the usage of code pairs is very low. The model of code pairs is intended for rare situations and it seems that clinicians have not adopted this model even when it would give some benefit. Second in most cases the clinical situation is expressed with other type of combination of codes.

Technical change

There are altogether 5919 code pair property diffences. In 3306 cases the one national code is linked to a pair and change is proposed. 51 are obvious errors with clear proposal for correction. In 2480 the difference is proposed to be corrected by removal of the property. Most of these situations affect ICE and FIN versions. Both FIN and ICE are involved in more than 2400 cases where as SWE is involved in 21 and NOR 29 cases.
In 82 cases a missing property is proposed to be added to one or more versions. Because the most of the versions have the property, it is proposed that all versions should have it.

DRG change

The effect is minimal. Because the use of code pairs is very rare, no significant effect is expected. Additionally a number of the properties are no more in use and for those cases no theoretical effect is expected.


#1 Updated by Martti Virtanen over 7 years ago

2014-03-17 Martti Virtanen
See also case #233 and case #294

#2 Updated by Anonymous over 7 years ago

  • Owner / responsible deleted (National organisations)

Comment Expert Group 2014-03-25

Proposal accepted.

#3 Updated by Anonymous over 7 years ago

Comment Expert Group 2014-03-25

Decision made in case #233.
There will be no change in any version.
Proposal is accepted.

#4 Updated by Martti Virtanen over 7 years ago

2014-05-05 Martti Virtanen
The comments from SWE resulted in some changes to this case. The table has been adjusted for the decissions in other cases (#250, #251, #290/300 and #292)
There are 5 additional changes (one with two rows) that I have felt necessary to comment. These cases have yellow background for NPK comment -column that can be filtered. I propose that all are done as proposed by Sweden.

#5 Updated by Anonymous over 6 years ago

  • Status changed from Active to Further active

#6 Updated by Martti Virtanen over 5 years ago

2016-03-15 Expert group (MV)
The additional changes were accepted without further discussion.

#7 Updated by Anonymous about 5 years ago

  • Status changed from Further active to Accepted

Also available in: Atom PDF