A, Hospitals are sorted by rates of any preoperative testing for each procedure separately. B, Hospitals are sorted by average rates of any preoperative testing. Rates of preoperative testing for the 3 surgical procedures within each facility were then plotted separately. Median rate of preoperative testing was 50.9% at participating hospitals (n = 63 hospitals). Models were adjusted for patient age and number of comorbidities with hospital identifiers included as random intercepts.
Customize your JAMA Network experience by selecting one or more topics from the list below.
Berlin NL, Yost ML, Cheng B, et al. Patterns and Determinants of Low-Value Preoperative Testing in Michigan. JAMA Intern Med. 2021;181(8):1115–1118. doi:10.1001/jamainternmed.2021.1653
Routine preoperative testing before low-risk surgery has no known benefit and is an important target for deimplementation because it is overused and costly and can lead to downstream care cascades involving invasive diagnostic testing.1-4 Selecting appropriate strategies for deimplementation is limited by an insufficient understanding of patterns and determinants of routine preoperative testing before low-risk surgery. Our objectives were to (1) examine use of preoperative testing before 3 common low-risk, ambulatory surgical procedures across diverse practice settings in Michigan, (2) to assess interhospital and intrahospital variations in testing, and (3) to identify determinants of testing to inform targets for future deimplementation strategies.
We performed a retrospective cohort analysis of administrative claims data from a Blue Cross Blue Shield of Michigan–funded statewide collaborative quality initiative to identify adults who underwent preoperative testing before 1 of 3 common low-risk ambulatory surgeries (lumpectomy, laparoscopic cholecystectomy, or laparoscopic inguinal hernia repair) from January 1, 2015, through June 30, 2019. This study was deemed exempt from human subjects protection review by the University of Michigan Institutional Review Board owing to the use of deidentified data. Patient informed consent was not required owing to the use of deidentified data. This study followed the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guideline.
The primary dependent variable was receipt of at least 1 preoperative test in the 30 days prior to the index procedure (excluding testing on the day of the procedure, during an inpatient stay, or in an emergency setting). Preoperative tests included cardiac testing (ie, electrocardiogram, echocardiogram, cardiac stress tests), laboratory studies (ie, complete blood cell count, basic metabolic panel, coagulation studies, urinalysis), chest radiography, and pulmonary function testing.
Multilevel mixed-effects regression modeling was performed to evaluate associations between patient, practice-level, and hospital characteristics with testing. Statistical analyses were performed using SAS, version 9.4 (SAS Institute) and Stata, version 16 (StataCorp LLC).
The analytic cohort included 9619 patients who underwent lumpectomy, 20 249 patients who underwent laparoscopic cholecystectomy, and 10 172 patients who underwent laparoscopic inguinal hernia repair; mean (SD) age for the cohort was 59.3 (16.1) years, and 24 783 (61.9%) were female. Approximately 51.6% of patients (n = 20 656) received 1 or more preoperative test, 29.4% of patients (n = 11 759) underwent 2 or more tests, and 13.5% (n = 5 395) underwent 3 or more tests. The 3 most common tests were complete blood cell count (n = 13 243 [33.1%]), electrocardiograms (n = 10 078 [25.2%]), and basic metabolic panel (n = 4533 [11.3%]). After adjusting for patient case-mix, there were wide interhospital and intrahospital variations in testing before the surgical procedures (Figure).
In multivariable models, increasing age and comorbidities were associated with preoperative testing (Table). Patients who underwent a separately billed preoperative history and physical examination were more likely to undergo testing (57% vs 35%; P < .001) (Table). Approximately 12% (95% CI, 10% to 18%) of variance in testing was attributed to hospital-specific differences.
In this study of patients undergoing 3 common low-risk surgical procedures, preoperative testing was common, with approximately 52% of patients undergoing at least 1 test and 29% undergoing 2 or more tests. Interhospital and intrahospital variations may be explained by local culture, institutional processes, or protocols facilitating either high or low rates of preoperative testing. Wide intrahospital variations in testing throughout the sample suggests that most hospitals may lack institutionwide protocols to reduce low-value testing or a centralized preoperative process and be more dependent on practice patterns of individual clinicians.
Undergoing separately billed preoperative history and physical examinations, which may be requested to meet the historical Centers for Medicare & Medicaid Services requirement that a preoperative history and physical examination be completed within 30 days of surgery, may be associated with increased low-value preoperative testing. Understanding the rationale for the continued use of these services for low-risk surgery represents a potential deimplementation target for future work. In some cases, insurance claims may not provide sufficient clinical information to discriminate between low-value and appropriate services. While a small portion of testing could be appropriate based on clinical context, it is unclear how this limitation would change our estimated hospital variations, as we could expect this measurement error to be evenly distributed across centers and procedures.
Overall, the findings of this study suggest a role for deimplementation strategies that target specific hospitals. There is a growing consensus that strategies must leverage conceptual frameworks targeting institutions, in addition to individual health care practitioners.5
Accepted for Publication: March 11, 2021.
Published Online: May 17, 2021. doi:10.1001/jamainternmed.2021.1653
Corresponding Author: Lesly A. Dossett, MD, MPH, Department of Surgery, University of Michigan, 1500 E Medical Center Dr, Ann Arbor, MI 48109 (firstname.lastname@example.org).
Author Contributions: Dr Berlin and Ms Yost had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.
Concept and design: Berlin, Kerr, Dossett.
Acquisition, analysis, or interpretation of data: Berlin, Yost, Cheng, Henderson, Nathan, Dossett.
Drafting of the manuscript: Berlin, Yost, Cheng, Dossett.
Critical revision of the manuscript for important intellectual content: Berlin, Henderson, Kerr, Nathan, Dossett.
Statistical analysis: Berlin, Yost, Cheng, Henderson.
Obtained funding: Berlin, Kerr.
Administrative, technical, or material support: Berlin, Nathan, Dossett.
Supervision: Berlin, Nathan, Dossett.
Conflict of Interest Disclosures: Dr Berlin reported receiving funding from the US Department of Veterans Affairs supporting his role as a National Clinician Scholar during the conduct of the study and honoraria from the Research Consortium for Health Care Value Assessment outside the submitted work. Dr Kerr reported receiving funding from the Veterans Affairs Health Services Research & Development Service during the conduct of the study and personal fees for serving on the Clinical Advisory Board for Bind Insurance and serving on the Steering Committee for Choosing Wisely International outside the submitted work. Dr Nathan reported receiving funding from the Agency for Healthcare Research and Quality (K08HS024763) and the National Institutes of Health (R01AG039434) during the conduct of the study. Dr Dossett reported receiving funding from the Agency for Healthcare Research and Quality (K08-HS-026030) during the conduct of the study. Ms Yost, Ms Cheng, and Dr Nathan are supported by a contract from Blue Cross Blue Shield of Michigan Value Partnerships program for the conduct of the Michigan Value Collaborative. No other disclosures were reported.
Disclaimer: This article does not necessarily represent the views of the US Government or Department of Veterans Affairs. The opinions, beliefs, and viewpoints expressed by the authors do not necessarily reflect those of Blue Cross Blue Shield of Michigan or any of its employees.
Additional Contributions: We would like to thank Hannah Lahti, BA, Tony Cuttitta, MPH, and Allen Metz, BA, for editing assistance with the manuscript. We would also like to thank John Syrjamaki, MPH, and Tim Hofer, MD, for additional guidance with the analytic plan. They were not compensated for their contributions.