2.12 Coding Audit
The success and fairness of casemix funding is based on accurate reporting of diagnostic information. In order to measure the accuracy of hospital ICD9-CM coding and the resultant assignment of AN-DRGs, a second coding audit was conducted by Coopers and Lybrand using 1995-96 data.
In addition to a random audit, which involved the examination of 3,928 records across 50 hospitals, audits were conducted in regard to target areas (3,270 records). Two audits primarily focused on chemotherapy and rehabilitation episodes and the appropriateness of the admission and the care type assigned. Another two dealt with newborns, looking specifically at clinical criteria and admission weight. A further targeted audit examined 17 selected AN-DRGs with complications and/or co-morbidities. Finally three hospitals originally examined in the random audit were "followed up" with a specific audit using larger sample sizes.
The random audit coded 88.3 per cent of episodes into the same AN-DRG's as coded by the hospital. Five per cent were originally coded by the hospital into an AN-DRG with a higher weight and 6.5 per cent into an AN-DRG with a lower weight. This represents an improvement on the 1993-94 coding audit where the equivalent results were 86.5 per cent, 5.9 per cent and 7.6 per cent respectively.
Whilst the overall percentage change to average sample weight for 1995-96 suggests slight undercoding (1.8 per cent), the greater divergence of results, when compared to 1993-94, is cause for some concern. Of the metropolitan hospitals audited the largest level of overcoding was 9.1 per cent. Significant overcoding disadvantages peer hospitals in two ways. In the short term it shifts resources from hospitals coding within acceptable bounds to those overcoding in excess of that. In the longer term it reduces the cost weight for the AN-DRGs involved, disadvantaging all providers of the services concerned.
A third audit is currently being planned for 1997-98 data which will comprise a smaller number of hospitals with larger random sample sizes. A supplementary audit may be undertaken in hospitals where the primary audit identifies significant coding anomalies. The cost of the supplementary audit, VIMD data correction and WIES adjustment will be borne by the hospital. Further policy on penalties for overcoding will be developed during 1997-98.
Hospitals who participated in the coding audit have been provided with an individual report. Detail of 1995-96 coding audit methodology and a summary of results are contained in Coopers and Lybrand's Summary of the Final Report, which is available from the Department's Health Data Standards and Systems Unit
telephone (03) 9616 8141 fax (03) 9616 7629.
To:
Top of Section 2
Purchasing PolicyTo:
Victoria - Public Hospitals
Policy and Funding Guidelines
1997-1998 - Table of Contents
