Background: Routine preoperative testing is discouraged before low-risk surgery because testing does not provide any beneficial effect in terms of patient outcome. However, few studies have assessed the utilization of hospital health care resources in terms of preoperative tests in a real-world setting. Here, we aimed to assess the prevalence and factors associated with preoperative blood tests before low-risk surgery in Japan.
Methods: In this retrospective observational study, we used the nationwide insurance claims data of Japan. Patients who underwent low-risk surgeries between April 1, 2012 and March 31, 2016, were included. Our primary outcome was the receipt of any preoperative tests within 60 days before an index procedure: complete blood count, basic metabolic panel, coagulation tests, and liver function tests. We performed a descriptive analysis to estimate the proportions of preoperative blood tests, and examined the associations between patient-level and institutional-level factors and preoperative blood tests, using multilevel logistic regression analysis. Interinstitutional variation in the utilization of preoperative tests was summarized using the median odds ratio (OR).
Results: The study sample included 59,818 patients (mean [standard deviation] age, 44.0 [11.3] years; 33,574 [56.1%] women) from 9746 institutions. The overall proportion of each test was: complete blood count, 58.7%; metabolic panel, 47.8%; coagulation tests, 36.6%; and liver function tests, 48.5%. The proportion receiving any preoperative tests in the overall sample was 59.5%. Multilevel logistic regression analysis indicated that preoperative blood tests were associated with the Charlson comorbidity index score (score ≥3: adjusted OR, 4.21; 95% confidence interval [CI], 3.69-4.80), anticoagulant use (adjusted OR, 4.12; 95% CI, 2.35-7.22), type of anesthesia (general anesthesia: adjusted OR, 5.69; 95% CI, 4.85-6.68; regional anesthesia: adjusted OR, 3.76; 95% CI, 3.28-4.30), surgical setting (inpatient procedure: adjusted OR, 3.64; 95% CI, 3.30-4.00), and number of beds (≥100 beds: adjusted OR, 3.61; 95% CI, 3.19-4.08). The median institutional-specific proportion of preoperative tests was 40.0% (interquartile range, 0%-100%). The median OR for interinstitutional variation in ordering preoperative tests was 4.34. These findings were consistent across a sensitivity analysis.
Conclusions: Preoperative blood tests were performed before 59.5% of low-risk surgeries. Preoperative tests were associated with the type of anesthesia, patient characteristics, and medical facility status. There was a substantial interinstitutional variation in the utilization of preoperative tests.