Routine Postoperative Laboratory Testing: Necessity or Habit?

Introduction

Since 1960, health care expenditures have continued to climb, decade after decade. In 2017, the United States national health care spending exceeded 3.5 trillion dollars. This aggregate is estimated to be 17.9% of the national gross domestic product. ¹ Increased spending creates pressure on hospitals to practice medicine with cost-containment strategies in mind. Furthermore, there is a necessity for providers to practice evidence-based medicine for the well-being of their patients. To aid in the increasing cost of health care, novel cost-saving measures should be identified.

Laboratory tests are becoming a habitual part of clinical practice as opposed to supplementing histories and physical examinations. Studies have projected that more than 70% of routine laboratory tests are of no clinical utility and do not change clinical decision making.^2,3 It is unclear whether testing has become standard for providers, is due to inexperienced physicians, or is being used as a legal safeguard to prevent liabity. ³ Regardless, this is one area of medicine in which changes can be made to save money for patients, hospitals, and the United States.

Multiple studies have analyzed the necessity of postoperative laboratory testing.^4–6 Paynter et al. provided evidence that routine postoperative laboratory tests were unnecessary after primary reverse-shoulder arthroplasty: While postoperative laboratory tests were performed on all patients, there was no change in patient management in 87.3% of patients based on their laboratory results. ⁴ McKenna et al. also supported dismissing postoperative serum laboratory tests following elective colorectal surgery: Of the 8205 laboratory tests in that study, only 3.8% of the POD1 labs were abnormal and only 1% of the values warranted clinical intervention. ⁶ Both studies showed that POD1 laboratory tests can, and should be, eliminated, as they do not improve clinical outcomes and are economically wasteful.^4,6

There is a lack of data demonstrating that routine postoperative hemoglobin and hematocrit studies improve clinical outcomes in urogynecologic surgical patients. While there are many different types of surgical procedures within this field, the majority are low-risk and rarely result in complications such as stroke, venous thromboembolism, and death. ⁷ There is anecdotal evidence suggesting that needing a transfusion is a rare complication of gynecologic surgery; however, there are not many data-driven studies to support this. ⁷ In particular, there is little evidence suggesting that the decision to transfuse patients is based solely on postoperative hemoglobin and hematocrit levels. It would appear that the decision to transfuse patients is determined by a combination of preoperative laboratory values and clinical signs of anemia in addition to postoperative laboratory tests. Howell et al. established that preoperative anemia was the strongest predictor of an abnormal postoperative complete blood count (CBC) and the need for transfusion. ⁵ As such, it seems appropriate to order postoperative hemoglobin and hematocrit levels only on patients with signs and symptoms of anemia, who are not recovering as expected or who had anemia revealed on preoperative tests.^8,9

The objective of this study was to investigate the cost-effectiveness and utility of routine postoperative day–1 (POD1) hemoglobin and hematocrit testing in patients who underwent urogynecologic surgery. It was hypothesized that these studies are not cost effective and do not lead to interventions that are not otherwise anticipated by clinical suspicion and/or preoperative laboratory values.

Materials and Methods

This original research project was reviewed and approved by the local institutional review board. A two-institution retrospective review was completed, investigating the cases of 312 patients who underwent urogynecologic surgeries over 3 years. The procedures included anterior and posterior repair, hysterectomy, midurethral slings, and myomectomies that were performed between 2015 and 2018 at 2 different hospitals in Denver, CO. A detailed chart review was completed by hand to collect patient demographics, estimated blood losses (EBLs), and pre- and postoperative hemoglobin and hematocrit levels. Patients were assessed for clinically significant outcomes, including increases in hospital lengths of stays (LOS) or blood transfusions, with any correlations between EBLs and postoperative hemoglobin levels. Risk factors for transfusions were assessed. Cases were excluded if there were no preoperative laboratory tests in the patients' charts.

For the economic assessment, the laboratory charges at 1 hospital institution were used. Statistical analysis was completed using Student's t-tests, assuming unequal variance. A p-value <0.001 was considered to be statistically significant.

Results

The retrospective review of 312 charts was completed using charts ranging from 2015 to 2018. The youngest patient was 20 years' old and the oldest patient was 88 years' old. The average patient age was 54, with all patients being females. All procedures were urogynecologic surgeries including, but not limited to, total vaginal hysterectomy, total abdominal hysterectomy, anterior repair, posterior repair, bilateral salpingo-oophorectomy, dilation and curettage, abdominal sacrocolpopexy, paravaginal repair, enterocele repair, and rectocele repair. The most-common procedure performed was anterior colporrhaphy. Data points evaluated were as follows: female patient; urogynecologic surgery within the last 5 years; preoperative hemoglobin and hematocrit; postoperative hemoglobin and hematocrit; EBL, discharge date; discharge disposition; barriers to discharge; transfusion requirement, and reason for transfusion.

The average change in hemoglobin was 2.89 g/dL. The average change in hematocrit was 7.45 g/dL. The average EBL was 171.3 mL. There were no increases in hospital LOS nor a correlation between EBL from the operation and change in hemoglobin level.

Of 312 patients, 12 patients received blood transfusions while 300 did not, which translated to a total of 3.8% of pts who required transfusions). Some patients received more than 1 transfusion during their hospital stays, which meant there were a total of 17 transfusion events. Eight patients had preoperative anemia based on their preoperative CBCs. Three of the 8 anemic patients required intraoperative transfusions. The other 4 patients received blood transfusions based on their clinical findings. Clinical findings included hypotension, tachycardia, and bleeding. Transfused patients had a lower preoperative hemoglobin (11.58; standard deviation [SD] = 3.08) and hematocrit (35.42; SD = 7.79). Four patients received transfusions based on routine POD1 hemoglobin and hematocrit values in combination with their clinical findings. This correlated to 1.3% of patients requiring changes in their management based on supporting POD1 laboratory tests. No patients received transfusions based solely on their POD1 laboratory test results.

Charts were also examined for differences in preoperative laboratory test results to determine if differences in those tests were predictive of needing blood transfusions during the procedures. Patients not receiving transfusions had a mean age of 53.53, with a preoperative hemoglobin level of 14.22 (SD = 1.188), and a preoperative hematocrit of 42.1 (SD = 3.065). Patients receiving transfusions were on average 55.94 years' old, with a preoperative hemoglobin level of 11.59 (SD = 3.081) and a preoperative hematocrit of 35.42 (SD = 7.795). Hemoglobin and hematocrit readings were normally distributed, allowing parametric t-tests to be performed to determine if these differences in preoperative tests were statistically significant. Differences in these test results were statistically significant (p < 0.001) for both hemoglobin and hematocrit results, using t-tests, assuming unequal variance.

Discussion

While every patient at the 2 hospital institutions is required to undergo routine POD1 hemoglobin and hematocrit testing, it does not appear that universal ordering of these laboratory tests affects clinical outcomes. If anything, these tests could subject patients to unnecessary anxiety and the discomfort associated with blood draws.

Routine postoperative hemoglobin and hematocrit testing on urogynecologic surgical patients appears to be unwarranted and not cost-effective. With the low rate of laboratory-associated interventions, the current study suggests that routine hemoglobin studies on POD1, while habitually ordered, are not necessary, evidence-based, nor cost effective, and can be discontinued. Only 1.3% of patients required an intervention using POD1 hemoglobin and hematocrit test results. While the laboratory values might have played a role in the decision of the surgeon, clinical suspicion and physical examination findings were likely the main indicators when deciding on initiation of a blood transfusion. A similar conclusion was made for both oncologic and benign gynecologic surgery patients in a different study. ⁸

Given the above results, the current authors suggest that ordering postoperative hemoglobin and hematocrit level tests only on patients for whom these tests would influence patient management. This includes patients with clinical symptoms of anemia, who are anemic on preoperative laboratory tests, or who have underlying conditions putting the patients at higher risk for postoperative complications. In this setting, obtaining postoperative hemoglobin and hematocrit levels can contribute to patient management and may indicate the need for blood transfusions. The current authors do not recommend ordering postoperative hemoglobin and hematocrit levels on low-risk, asymptomatic patients who are otherwise recovering well from surgery and whose preoperative laboratory values are within normal limits.

There were a couple of limitations to this study. First, the patient charts evaluated from 2 different institutions included patients being treated by only 1 physician. Clinical practice and decision making among surgical providers may differ, but it is expected that physicians are practicing according to the standard of care. Second, the patients evaluated included urogynecologic-specific operative surgeries, which do not normally produce large amounts of blood loss. No other kinds of surgical procedures were evaluated.

Based on the results of this study, the next step would be to implement an institutional analysis to evaluate the utility of POD1 laboratory testing at each institution. This would likely produce results similar to this study but would allow for more cumulative data. Institutions could then implement a preoperative anemia-laboratory protocol. The culmination of this systems-improvement project would decrease routine ordering of laboratory tests, resulting in a significant decrease in health care costs.

Conclusions

Further studies on transfusion rates associated with urogynecologic surgeries are needed. The current study data revealed that 1.3% of patients needed transfusion, but that decision was not based solely on postoperative laboratory values. As such, routine postoperative hemoglobin and hematocrit tests on every gynecologic surgery patient can be discontinued without resulting in a negative effect on patient outcomes. Further studies at each institution could be useful to confirm this conclusion. Individual institutions may want to complete a quality-improvement study on the efficacy and cost-effectiveness of their routine postoperative laboratory tests.