Task #10: Delivery of major update proposals for NordDRG 2015
Splitting of the 900-series DRGs
|Target version:||Expert Group 2014|
|Case type:||Owner / responsible:||Nordic Casemix Centre|
|MDC:||GEN||Old forum status:|
National ID: CPK ID 476
Sent to NordDRG Forum: 2014-01-24
In the Swedish expert network for NordDRG (SWEX) we have discussed how to improve the system’s ability to describe outpatient care. We then came to the conclusion that a substantial improvement is achieved by splitting the conservative outpatient DRGs (initially called the 900 series DRGs) because today there is only one such DRG per MDC which makes them rather heterogeneous and these groups are also the most frequent (in average more than 160 000 cases per group in the Swedish cost database for 2012).
We first tried to group the material by principal diagnosis at the three-digit code level. We then got more than 2000 groups and still 785 groups if we included only groups with 500 or more cases. The need for further aggregation was evident.
We therefore tried the alternative to group the material according to the existing inpatient grouping logic (we just added one day to the hospital stay and then regrouped the material.) but aggregated the complicated and non-complicated subgroups. We then got totally only 191 groups (see ) even if we include also the smallest groups.
We presented the results for the Swedish specialist associations and asked them what type of split they preferred. Six out of 19 associations answered and four of them preferred the split based on inpatient logic with the motivation that it then is easy to compare inpatient and outpatient materials and to study the degree of policlinization over time. (The Association for Phoniatrics did not prefer the inpatient logic which is quite natural since they never treat inpatients and the Association for Urology obviously misunderstood the problem – they answered that a split on new visits and return visits is enough.)
Sweden has decided to split the conservative outpatient DRGs according to the existing inpatient grouping logic as described above, with the exception for the groups in MDC 15, 19 and 23 since these splits have to be discussed further with the specialists. Some smaller groups will also be aggregated (see “Not motivated due to low frequency” in Appendix CPK ID 476/”analysis”) and the group for radiation therapy (R40) is not needed since we already have DRGs for that. Rest groups in each MDC, with a grouping logic almost identical to the logic of the present outpatient DRGs, will also be omitted because there is no point in having double residual groups. After these adjustments the number of new groups will be 128. The new groups will have the same code as the corresponding inpatient groups at a three-digit code level but with the letter O in the end. The new DRGs are listed in and the new rules for these DRG can be seen in .
NPK, Sweden – 2014-01-23
We think that this model is an interesting improvement in describing outpatient care and want the Nordic Expert Network to consider an introduction also in the Common version.
Detailed technical changes for the Swedish version are described in appendix .
NPK, Sweden – 2014-01-23
Sweden has decided to introduce the change described above in NordDRG for 2015.
Approximately 70 % of the outpatient cases that used to be grouped to the conservative groups for single doctor visits (one per MDC) will instead be grouped to one of the new groups, 2-19 per MDC.
128 new DRGs are added to the table drgnames and 209 rules for these DRGs are inserted in the table Drglogic. See “+Decision CPK ID 476+”.
NordDRG  [Sweden]
#8 Updated by Martti Virtanen over 5 years ago
2014-04-30 Martti Virtanen
The technical changes as described from Sweden are otherwise OK, but at some points the ORD-numbering needs to be considered. The ORD -codes should not be part of the follow codes (like 000D1050 and 000D10501 because these may cause problems in some applications. Instead the coding should be 000D10500 and 000D10501.
This is especially important when a group new codes is added like for obstetrics - see attached table. Therefore in the process some of the ORD codes may be changed (and even some old ORD codes may change).
#9 Updated by Ralph Dahlgren over 5 years ago
Mats Fernström NPK SWE 2014-05-06
SWE accepts Marttis suggestions ord values in case 215 but DRG A39O should not be included as Martti suggested at the meeting. We noticed in the file rows (two of them) about DRG P99O marked with green colour and no Case ID. To which case do these belong?