Suitability of casemix classification for output funding of blood services in Victoria. Victorian Blood Users Group.

OBJECTIVE

To study the suitability of diagnosis-related groups (DRGs) to quantify blood use in order to fund hospitals directly for fresh blood components.

DESIGN

(i) Retrospective empirical analysis of hospital inpatient separation data, matching DRG classification with blood usage. (ii) Prediction of blood usage based on actual DRGs and comparison with actual blood products issued from the blood service.

SETTING

Eight large public hospitals, April 1994-March 1995, the Department of Human Services and the Australian Red Cross Blood Service in Victoria, 1994-1997.

MAIN OUTCOME MEASURES

Requirement for transfusion according to DRG; quantity of blood or blood products transfused; and statistical reliability of measuring blood component usage by DRG.

RESULTS

A match between patient records and transfusion records for 287,117 patient separations showed that the patients had received 51,115 units of red cells, 30,451 units of platelet concentrates, 9043 units of fresh frozen plasma and 1273 units of cryoprecipitate. Ten per cent of DRGs (527) accounted for over 70% of blood product usage. The numbers of DRGs in which blood products were used for at least 30 separations (with a relative SEM [for units of blood used per separations using blood] of up to 20% at a 95% confidence level) were 56 for red cells, 8 for platelets and 4 for fresh frozen plasma. Estimates of red cell usage calculated from actual DRGs for three consecutive years (1994-1997) showed that DRGs predicted aggregate issues of red cells by the Red Cross Blood Service (ratio of actual red cells issued to red cell usage estimates were 1.036 for 1994-95, 0.994 for 1995-96, and 1.021 for 1996-97, respectively).

CONCLUSION

DRGs are moderately good predictors of blood product usage, and could be used to allocate funds for blood products to hospitals instead of to the blood service. However, DRG data are not designed to manage blood issue policies, because DRGs do not represent single diagnoses or procedures and some are surrogate descriptions of procedures.