Blood utilization in hip and knee arthroplasty: a cost-minimization study
A. GOWER, A. I. HUSSEIN, P. J. BRIGGS AND M. S.
DEWAR*
Departments of Orthopaedics and *Haematology, Wansbeck General Hospital, Woodhorn Lane,
Ashington, Northumberland, UK
The non-utilization of crossmatched blood is an expensive waste of resources. We have audited blood utilization for all primary total hip and knee arthroplasty patients. We compared routine pre-operative crossmatching (Phase 1) with a policy of group, screen and save (G & S) only (Phase 2). The patient groups were similar in both phases. Pre-and post-operative haemoglobin results were not significantly different between Phase 1 and 2. No adverse transfusion reactions occurred. In Phase 1,213 units were crossmatched pre-operatively, but only 127(60%) were transfused. In Phase 2, 117 units were requested and all transfused. The G & S only policy proved to be a safe and practical option which improved the efficiency and cost-effectiveness of blood ordering. Based on a handling charge of £37.50 per unit of blood by the Regional Transfusion Centre, an estimated annual saving of over £8000 can be made in our directorate.
Keywords: arthroplasty, audit, blood transfusion.
The non-utilization of crossmatched blood represents an expensive waste of scarce resources. The maximum surgical blood order schedule (MSBOS) is widely used in elective surgical practice.1,2 Blood is ordered for patients scheduled for an operation according to a locally agreed tariff. For example, in our department 3 units of blood were routinely crossmatched pre-operatively for patients undergoing a total hip arthroplasty (THA). The British Committee for Standards in Haematology endorsed the use of MSBOS as a desirable level of practice in UK hospitals in 1991. 2,3 This practice still involves unnecessary crossmatching and consequently over-ordering of blood.4,5
A crossmatch involves incubation of the patient’s serum with donor red blood cells (RBCs); if no reaction occurs, the donor unit is labelled and reserved exclusively for the identified patient for up to 72 h. An alternative method of ensuring the availability of blood for elective surgery is the group, screen and save (G & S) technique, whereby the patient’s serum is incubated with pooled RBCs containing all common RBC antigens; if no reaction occurs the patient can safely have any ABO-and Rhesus-compatible blood ‘off the shelf’. If the G & S is performed pre-operatively, blood can be provided rapidly on request following a 2-min spin crossmatch (to exclude an ABO labelling error by the Regional Transfusion Service). If the RBC pool screen is positive then the antibody responsible is identified and a conventional crossmatch performed. This ‘antibody screen and rapid-spin’ method avoids the ordering of blood destined not to be transfused. The result is the improved efficiency of blood ordering and handling. The method has been thoroughly validated and has recently been sanctioned by the British Blood Transfusion Society.6-8
The aim of this audit was to determine the safety and practicality of replacing the pre-operative crossmatch with a G & S only policy for all primary total hip and knee arthroplasty patients.
A prospective audit of blood ordering for all patients undergoing primary elective total hip replacement (THA) or primary elective total knee replacement (TKA) was performed between 01.05.95 and 3 1.07.95 (Phase 1). According to our agreed protocol, 3 units of blood were requested pre-operatively for each patient. Significant over-ordering of blood was identified. Following discussion with our anaesthetic and haematology colleagues a G & S policy was introduced. The audit was then repeated between 01.01.96 and 31.03.96 (Phase 2).
In Phase 1,73 patients underwent surgery (38 THAs; 35 TKAs). Mean age was 69.2 years (range 45—88); 26 were male and 47 female (Table 1). In Phase 2, 67 patients underwent surgery (33 THAs; 34 TKAs). Mean age was 67.2 years (range 40—84); 31 were male and 36 female (Table 1). Pre-operative and 48-h postoperative haemoglobins were recorded as well as the number of units of blood requested and transfused during the hospital admission. Any complications of the transfusions were also recorded. The data was analysed on a personal computer and statistical analysis was by Student’s t-test or chi-squared test, where appropriate.
Table 1 Patient groups
Phase 1 (MSBOS) |
Phase 2 (G & S) | |
|---|---|---|
| Age (years; mean range) | 69.2 (45-88) |
67.2 (40-84) |
| Sex (male: female) | 26:47 |
31: 36 |
| Operation | ||
| THA | 38 |
33 |
| TKA | 35 |
34 |
| Surgeon | ||
| Consultant | 50 |
56 |
| Registrar | 23 |
11 |
| Length of admission days (range) | 12 (7—74) | 9.5(4—28) |
The patients in the two phases of the audit were comparable for age and sex distribution, and grade of operating surgeon (Table 1). The length of admission was a mean of 12 days in Phase 1, compared with 9.5 days in Phase 2. This was statistically significant (P< 0.05).
In Phase 1, 213 units of blood were crossmatched pre-operatively. Of these, 112 units (5 3%) were transfused within the first 24 h and a further 15 units used later in the first week. Eighty-six units (40%) remained unused (Table 2). Twelve patients required additional crossmatch of 18 units, all of which were transfused within the first week.
Table 2 Blood utilization
| Units requested pre-op | Units transfused day 1 | Units not utilized day 1 | Additional units transfused day 1 | Units transfused day 2—7 | Total units not utilized week 1 | Additional units transfused day 2—7 | Total number of units transfused | |
|---|---|---|---|---|---|---|---|---|
| Phase l | 213 | 112 |
101(47%) |
7 |
15 |
86(40%) |
11 | 145 |
| Phase 2 | 0 |
79 | 0 | 79 | 38 |
0 | 38 | 117 |
In Phase 2, no blood was crossmatched pre-operatively. In the first 24 h, 79 units of blood were requested and transfused in response to peri- operative blood loss and the haemodynamic status. 4 further 38 units were requested and transfused later in the first week, guided by the post-operative haemoglobin. In contrast to Phase 1, all blood requested was transfused.
The post-operative haemoglobin was recorded in 70 patients in Phase 1 and 59 in Phase 2. There was no significant difference between Phase 1 and 2 for pre-operative haemoglobin or postoperative haemoglobin. The mean pre-operative haemoglobin was 129 g/L ( ± 14) in Phase 1 and 131 g/L (±13.7) in Phase 2. The mean 48-h post-operative haemoglobin was 113.6 g/L (± 13.1) in Phase land 113.6g/L(±16.6)inPhase2 (see Figure l). No delay was encountered in the provision of blood for transfusion when it was requested, and no adverse transfusion reaction occurred.
Figure 1 Haemoglobin at 48-h post-op, Phase 1 (n= 70), Phase 2 (n= 59).
Our initial audit of pre-operative blood ordering for primary hip and knee replacement demonstrated significant over-requesting of crossmatched blood. During Phase 1, 40% of units crossmatched were not transfused. In contrast, during Phase 2 all blood requested was transfused. This would seem to demonstrate that a group, screen and save policy resulted in the improved efficiency of blood ordering.
While the National Blood Authority in England and Wales receives some central funding, the Regional Transfusion Centres have to pass on a handling charge for blood to hospitals. The handling charge may be met from the hospitals Department of Pathology budget, or may be devolved to the clinical directorate which ordered the blood. At the time of this study, the Northern Region Transfusion Centre handling charge for one unit of blood was £37.50. Wansbeck General Hospital Blood Bank operates a 24-h shift system; there is therefore no cost differential between work done during ‘office hours’ and work done out of hours. An antibody screen costs £1.20 (labour and materials); crossmatching 3 units of blood costs £1 15.75 (labour, materials and blood handling charge). A request for 3 units of blood will cost the orthopaedic directorate £115.75, regardless of whether the blood ordered from the Region Transfusion Centre is used or not.
We transfused a mean of 44 units per month during the period of study. Routine crossmatching resulted in 40% of units not being transfused (18 units per month). By adopting a G & S policy the orthopaedic directorate was not charged the handling fee for 18 units of blood per month, corresponding to an annual saving of over £8000 per year.
We have demonstrated that the group, screen and save policy for major joint replacement is a safe and acceptable practice. Excess ordering was eradicated, blood was available on request, and there was no change in our practice of actual transfusion of blood.
We recommend the introduction of a G & S policy for ordering blood for major joint replacement surgery, in line with the BCSH blood transfusion service task force guidelines,6 and other authors.4,5 We found it to be a safe and practical option which improves the efficiency of blood ordering and produces a considerable financial saving.
Paper accepted on 11 May 1998
Correspondence: Mr A. Gower, 20 St Aidans Park, Fourstones, Hexham, Northumberland NE475EB, UK
© The Royal College of Surgeons of Edinburgh J.R. Coll. Surg. Edinb., 43, December 1998, 397—399