The intermediate results of testing the CC-grouper on Finnish data
|Target version:||Expert Group 2015|
|MDC:||CC||Owner / responsible:||National organisations|
|Target Grouper:||FIN||Old forum status:|
Finnish National DRG-centre has tested the CC grouper which is used in Sweden. The grouper was adapted to some extent to current Finnish 2015 grouping logic but in generally it included the rules of Swedish grouper where the split of DRG is based on three characteristics: no complication (non-CC), complication (CC), major complication (MCC).
General overview of the results of testing
Ca 4,4 million cases from five Finnish university hospitals were used for testing. In order to compare the results two different groupers were used for grouping: 1) adjusted Finnish 2015 CC grouper (2015CC), 2) 2015 Full Finnish grouper (2015F).
With 2015F grouper the cases were signed in total to 862 DRGs, whereas with 2015CC grouper there were 1 060 different DRGs used. In total 211 431 (5% of total) cases changed their DRG with 2015CC grouper.
There was 90 DRGs (10% of all 2015F grouper DRGs) where less than 25 cases were assigned with 2015F and 172 (16% of all 2015CC grouper DRGs) with 2015CC grouper. Among small DRGs were 27 MCC DRGs (in total the cases were assigned to 149 MCC DRGs with 2015CC grouper).
Despite relatively reasonable results of grouping the cases with 2015CC grouper, there are some issues we would like to address and discuss within expert group.
Those are concerning the mean cost of DRGs where the split is based on complication category: MCC, CC, non-CC. Some mean costs did not reveal correctly the differences between the complication categories, e.g. average cost of MCC was lower than CC or non-CC, or average cost of non-CC was higher than CC etc.
There could be different reasons for this, e.g.:
• Differences in coding habits and compliance with coding standards. Coding of secondary diagnoses in Finland is poor. There are no addition nationwide coding guidelines beside WHO coding rules. This has impact on grouping results.
• Problem with grouping logic. It could be that the CC grouping logic needs to be reviewed and if needed some adjustment have to be done so that the cost follow the complication categories of DRGs;
• Use of different/modified clinical protocols and guidelines;
• Differences in cost accounting and pricing;
In Appendix 1 (see separate file) are those DRGs where abovementioned discrepancies between the complication categories and respective mean costs are depicted.
We would be glad to receive the comments from Swedish colleagues (and anyone else) and discuss the results in expert group.
#7 Updated by Ralph Dahlgren over 6 years ago
- File CC_testing_FIN + SWE To Forum.xls added
Mats Fernström, Veronika Stemme, Ralph Dahlgren, Susann Berggren, NPK, Swedish National Board of Health and Welfare, 2015-02-12:
We have tested the CC-grouper on Swedish data, from 2013 with the 2015 grouper. We tested the same DRG-groups as were tested with Finnish data in the case. The results with Swedish data were quite different from the test-results with Finnish data. In all cases except one, the order of the cost-data matched the degree of complication in the respective DRG, i.e. the cost was lower in the non-complicated than in complicated DRG-groups and the cost was lower in the complicated than in the very complicated groups. We conclude that the CC-grouper works well in Sweden. Please see attached Excel-file with the added Swedish data. We added both “All cases” and trimmed data were the outliers are excluded.
We see some possible explanations for the differences:
1.) Different rules for primary classification in Finland and Sweden.
2.) Differences in how costs are calculated in the two countries.
3.) Few cases in some of the Finnish DRG:s.
Suggestion: Use the CC-grouper in the DRG-groups where the cost-data fits with the degree of complication and in the DRG-groups were the cost-data does not fit – use the same DRG-weight for the DRG irrespective of its degree of complication.