Characterizing the Use of Gastrointestinal Procedures in the Workup of Anemia: Patterns of Care and Areas for Improvement

Background: Anemia is an extremely prevalent disease and contributes to a substantial percentage of worldwide years lived with disability. Clinicians generally agree that investigation of the gastrointestinal (GI) tract with colonoscopy and/or endoscopy is warranted in the presence of either overt clinical evidence of GI bleeding or in the presence of documented iron deficiency. However, little is known about the appropriateness of these procedures in the evaluation of patients without either overt GI bleeding or iron deficiency. Additionally, while a large number of colonoscopies and endoscopies are performed each year, the proportion of such procedures performed in the absence of suspect history or findings has not been characterized.

Methods: We performed a retrospective review of all patients at Mayo Clinic Rochester who underwent endoscopy, colonoscopy, or both procedures between July 1, 2014 and June 30, 2015 for an indication of either "anemia" or "iron deficiency anemia." Sources of data included provider notes, laboratory values, procedure notes, and pathology reports. Information including patient demographic characteristics, medical comorbidities, laboratory values related to anemia, findings at the time of gastrointestinal procedure, and procedure-related complications was collected. Data was entered into an electronic database and data analysis was performed using Microsoft Excel.

Results: 999 procedures were included, including 224 colonoscopies, 306 endoscopies, and 469 combined colonoscopy/endoscopies. Of the 999 total procedures, 78 (8%) had no documented pre-procedure hemoglobin value, 666 (67%) were performed after documented anemia (defined as hemoglobin ≤12 g/dL), and 255 (26%) were performed after a documented hemoglobin of >12 g/dL. Of the 675 procedures performed for "iron deficiency anemia," 462 (68%) patients had documented anemia and documented ferritin prior to the procedure, including only 237 (35%) with both hemoglobin ≤12 and ferritin ≤ 20. 388 (39%) had a documented pre-procedure patient and/or provider report of blood in the stool, and 111 (11%) had a documented pre-procedure positive hemoccult test. 365 (37%) of the 999 procedures identified a probable benign source of bleeding, 54 (5.4%) identified a probable malignant source, and 580 (58%) had no findings consistent with a bleeding source. Patients with documented anemia were more likely to have a probable bleeding source identified than those without documented anemia [46% versus 34%, p<.001]. Of the 83 patients with hemoglobin >12 and ferritin >20, only 1 procedure in 83 demonstrated a malignant finding, and this patient did have blood in the stool at the time of the procedure. There were 46 patients who underwent a procedure without any "warning signs" (either anemia and/or blood in the stool and/or positive hemoccult). Of these patients, 0 had a malignant finding.

Conclusions: A large number of procedures performed for either "anemia" or "iron deficiency anemia" were done without pre-procedure documentation of anemia and/or iron deficiency. Documentation of anemia is associated with an increased chance of finding a source of bleeding during the procedure. There were no missed diagnoses of malignancy among patients undergoing procedures without anemia, blood in the stool, and/or positive hemoccult. Careful attention to documentation should be encouraged, and quality improvement initiatives requiring appropriate documentation prior to performing these procedures may be helpful.