Discussion
Routine measurement of postoperative hematocrit levels after elective general surgery at NMVAHCS was not necessary. There were 12 transfusions for inpatients (2.8%), which is similar to the findings of a recent study of VA general surgery (2.3%).13 We found that routine postoperative hematocrit measurements to assess anemia had little or no effect on clinical decision-making or clinical outcomes.
According to our results, 88% of initial hematocrit tests after elective partial colectomies could have been eliminated; only 32 of 146 patients demonstrated a clinical reason for postoperative hematocrit testing. Similarly, 36 of 40 postcholecystectomy hematocrit tests (90%) could have been eliminated had the surgeons relied on clinical signs indicating possible postoperative anemia (none were transfused). Excluding patients with major intraoperative blood loss (> 300 mL), only 29 of 288 (10%) patients who had postoperative hematocrit tests had a clinical indication for a postoperative hematocrit test (ie, symptoms of anemia and/or active bleeding). One patient with inguinal hernia surgery who received transfusion was taking an anticoagulant and had a clinically indicated hematocrit test for a large hematoma that eventually required reoperation.
Our study found that routine hematocrit checks may actually increase the risk that a patient would receive an unnecessary transfusion. For instance, one elderly patient, after a right colectomy, had 6 hematocrit levels while on a heparin drip and received transfusion despite being asymptomatic. His lowest hematocrit level prior to transfusion was 23.7%. This patient had a total of 18 hematocrit tests. His EBL was 350 mL and his first postoperative HCT level was 33.1%. In another instance, a patient undergoing abdominoperineal resection had a transfusion on postoperative day 1, despite being hypertensive, with a hematocrit that ranged from 26% before transfusion to 31% after the transfusion. These 2 cases illustrate what has been shown in a recent study: A substantial number of patients with colorectal cancer receive unnecessary transfusions.14 On the other hand, one ileostomy closure patient had 33 hematocrit tests, yet his initial postoperative hematocrit was 37%, and he never received a transfusion. With low-risk surgeries, clinical judgment should dictate when a postoperative hematocrit level is needed. This strategy would have eliminated 206 unnecessary initial postoperative hematocrit tests (72%), could have decreased the number of unnecessary transfusions, and would have saved NMVAHCS about $1600 annually.
Abdominoperineal resections and total abdominal colectomies accounted for a high proportion of transfusions in our study. Inpatient elective operations can be risk stratified and have routine hematocrit tests ordered for patients at high risk. The probability of transfusion was greater in high-risk vs low-risk surgeries; 27% (6 of 22 patients) vs 2% (6 of 408 patients), respectively (P < .001). Since 14 of the 22 patients undergoing high-risk operation already had clinical reasons for a postoperative hematocrit test, we only need to add the remaining 8 patients with high-risk operations to the 74 who had a clinical reason for a hematocrit test and conclude that 82 of 430 patients (19%) had a clinical reason for a hematocrit test, either from signs or symptoms of blood loss or because they were in a high-risk group.
While our elective general surgery cases may not represent many general surgery programs in the US and VA health care systems, we can extrapolate cost savings using the same cost analyses outlined by Kohli and colleagues.1 Assuming 1.9 million elective inpatient general surgeries per year in the United States with an average cost of $21 per CBC, the annual cost of universal postoperative hematocrit testing would be $40 million.11,15 If postoperative hematocrit testing were 70% consistent with our findings, the annual cost for hematocrit tests on 51% of the inpatient general surgeries would be approximately $20.4 million. A reduction in routine hematocrit testing to 25% of all inpatient general surgeries (vs our finding that 19% were deemed necessary) results in an annual savings of $30 million. This conservative estimate could be even higher since there were 4.4 hematocrit tests per patient; therefore, we have about $132 million in savings.
Assuming 181,384 elective VA inpatient general surgeries each year, costing $7.14 per CBC (the NMVAHCS cost), the VA could save $1.3 million annually. If postoperative HCT testing were 70% consistent with our findings, the annual cost for hematocrit tests on 50.4% of inpatient general surgery operations would be about $653,000. A reduction in routine hematocrit testing to 25% of all inpatient general surgeries (vs our 19%) results in annual VA savings of $330,000. This conservative estimate could be even higher since there were on average 4.4 hematocrit levels per patient; therefore, we estimate that annual savings for the VA of about $1.45 million.