Development initiative #545

Requests for development of the NordDRG grouper

Added by Mats Fernström about 2 years ago. Updated about 2 years ago.

Status:ActiveStart date:2017-02-22
Priority:MajorDue date:
Assignee:Mats Fernström% Done:


Spent time:-
Initiator:Sweden Target year:
MDC:GEN Owner / responsible:Board
Target Grouper:COMMON Old forum status:


This message is directed to the Board but it is published on the Forum with the hope of consent, so that the Board gives priority to the matter.

When we now are planning a new NDMS, we also want a development of the NordDRG grouper, including improvements of the definition tables. Among all these requests, the most urgent is to get some more columns in the table Drglogic so that the fields for procedure properties and the fields for diagnosis properties don’t have to be used for the healthcare provider’s profession or for types of healthcare contacts. A strict structure in Drglogic is essential for making a NDRG Explorer that is correct without weekly manual efforts for each new version of NordDRG.

In the table Drglogic, there is a lack of systematic that makes it hard to understand and it is very hard to make a manual based on Drglogic. We need some changes so that the content can be arranged systematically.
1. We have only two fields for procedure properties (PROCPRO1 and SECPROC1) which have forced us to use the fields for diagnosis properties (DGPROP1-4) in the construction of grouping rules for different combinations of procedures. It is confusing to have procedures in the fields for diagnoses. We suggest insertion of another two fields for procedure properties.
2. Initially NordDRG was used only for inpatient care but in the grouping logic for outpatients we need variable types for profession and for type of contact but there are no fields for such variables which have forced us to use the fields for diagnoses and/or the fields for procedures in the construction of grouping rules. It is confusing to have profession and types of visits among diagnosis and procedure criteria. We suggest insertion of one field for type of care and one field for profession. The design of the variable types (VARTYPE) and the variable values (VARVAL) for these new variables can be decided by Martti and the experts.

If the field for the DRG codes is restricted to four characters in the present grouper it should be expanded in the new NDMS to make it easier to construct systematic DRG codes. In the Swedish version of NordDRG we have introduced a kind of systematic DRG codes, where the 1st character, a letter, stands for MDC, the 2nd & 3rd characters is a serial number and the 4th character, a letter, stands for both CC level and type of care/visit. However, there are more MDCs than available letters and there should be one separate character for the CC level and one separate character for the type of care/visit. A serial number with three digits could make it easier to use them in a systematic way. All together we could need at least eight characters for the DRG code field.

The code fields in the tables Proc1 and Dg1 should also be expanded to make it possible to use full ATC codes in the system.

Furthermore, the new NDMS and grouper should be constructed in a way so it is rather easy to add more variables in the future, for example variables for the patients functionality or for severity of illness and other factors that can be of interest to include in the grouping logic in the future.

It has been argued that functions like “AnyOf” for the relation between the DGPROP fields and between the PROC fields should be possible in the new grouper. In the present grouper we only have the function “AllOf”, which means that all the conditions in a grouping rule must be fulfilled. With a function “AnyOf”, the number of rules in Drglogic can be reduced. Personally, I question whether that benefit is worth the effort with a more complex program coding and this development must not cause that the above proposed changes for Drglogic are delayed.

In the future, the new grouper should also be able to handle multiple coding for diagnoses and procedures. In the present grouper, code pairs can be handled for diagnoses but sometimes it is necessary to use three or even four codes for a correct coding of a diagnosis. When multiple procedures are registered, the present grouper is not capable to allocate general qualifiers (NCSP codes beginning with Z) to the right procedure. This is sometimes a great problem, i.e. if a patient has a unilateral hip procedure and bilateral peripheral intravenous catheters, the case will be grouped to the DRG for bilateral hip procedures.


#1 Updated by Martti Virtanen about 2 years ago

2017-03-08 Martti Virtanen

These are interesting ideas, that may be taken to the future NDMS system.
We are currently struckling with the process to create a new NDMS programme and these are issues tha can be discussed.
In the current FoxPro based model I do not expect that we will be able to this type of changes. Currently it looks likely that new version will be based on the existing model but I hope that the flexibility that we have had with FoxPro will continue in the new model.

I hope we will have some time to discuss these items on the expert group meeting in Riga.

#2 Updated by Martti Virtanen about 2 years ago

  • Tracker changed from Questions and answers to Development initiative

Also available in: Atom PDF