Perioperative Medicine  |   April 2020
Classification of Current Procedural Terminology Codes from Electronic Health Record Data Using Machine Learning
Author Notes
  • From the Department of Anesthesiology, University of Michigan Medical School, Ann Arbor, Michigan (M.L.B., M.R.M., J.V., X.T., B.L., D.A.C., N.S., S.K., L.S.); Department of Anaesthesiology, University Medical Center Goettingen, Goettingen, Germany (L.S.).
  • Supplemental Digital Content is available for this article. Direct URL citations appear in the printed text and are available in both the HTML and PDF versions of this article. Links to the digital files are provided in the HTML text of this article on the Journal’s Web site (www.anesthesiology.org).
    Supplemental Digital Content is available for this article. Direct URL citations appear in the printed text and are available in both the HTML and PDF versions of this article. Links to the digital files are provided in the HTML text of this article on the Journal’s Web site (www.anesthesiology.org).×
  • This study has been presented in part at the 2017 Annual Meeting of the American Society of Anesthesiologists, October 21 to 25, 2017, in Boston, Massachusetts.
    This study has been presented in part at the 2017 Annual Meeting of the American Society of Anesthesiologists, October 21 to 25, 2017, in Boston, Massachusetts.×
  • Submitted for publication June 8, 2019. Accepted for publication December 17, 2019. Published online first on February 4, 2020.
    Submitted for publication June 8, 2019. Accepted for publication December 17, 2019. Published online first on February 4, 2020.×
  • Address correspondence to Dr. Burns: Department of Anesthesiology, University of Michigan, 1500 East Medical Center Drive, 1H247 UH, SPC 5048, Ann Arbor, Michigan 48109-5048. mlburns@med.umich.edu. Information on purchasing reprints may be found at www.anesthesiology.org or on the masthead page at the beginning of this issue. Anesthesiology’s articles are made freely accessible to all readers, for personal use only, 6 months from the cover date of the issue.
Article Information
Perioperative Medicine / Clinical Science / Practice Management / Technology / Equipment / Monitoring / Quality Improvement
Perioperative Medicine   |   April 2020
Classification of Current Procedural Terminology Codes from Electronic Health Record Data Using Machine Learning
Anesthesiology 4 2020, Vol.132, 738-749. doi:https://doi.org/10.1097/ALN.0000000000003150
Anesthesiology 4 2020, Vol.132, 738-749. doi:https://doi.org/10.1097/ALN.0000000000003150
Abstract

Background: Accurate anesthesiology procedure code data are essential to quality improvement, research, and reimbursement tasks within anesthesiology practices. Advanced data science techniques, including machine learning and natural language processing, offer opportunities to develop classification tools for Current Procedural Terminology codes across anesthesia procedures.

Methods: Models were created using a Train/Test dataset including 1,164,343 procedures from 16 academic and private hospitals. Five supervised machine learning models were created to classify anesthesiology Current Procedural Terminology codes, with accuracy defined as first choice classification matching the institutional-assigned code existing in the perioperative database. The two best performing models were further refined and tested on a Holdout dataset from a single institution distinct from Train/Test. A tunable confidence parameter was created to identify cases for which models were highly accurate, with the goal of at least 95% accuracy, above the reported 2018 Centers for Medicare and Medicaid Services (Baltimore, Maryland) fee-for-service accuracy. Actual submitted claim data from billing specialists were used as a reference standard.

Results: Support vector machine and neural network label-embedding attentive models were the best performing models, respectively, demonstrating overall accuracies of 87.9% and 84.2% (single best code), and 96.8% and 94.0% (within top three). Classification accuracy was 96.4% in 47.0% of cases using support vector machine and 94.4% in 62.2% of cases using label-embedding attentive model within the Train/Test dataset. In the Holdout dataset, respective classification accuracies were 93.1% in 58.0% of cases and 95.0% among 62.0%. The most important feature in model training was procedure text.

Conclusions: Through application of machine learning and natural language processing techniques, highly accurate real-time models were created for anesthesiology Current Procedural Terminology code classification. The increased processing speed and a priori targeted accuracy of this classification approach may provide performance optimization and cost reduction for quality improvement, research, and reimbursement tasks reliant on anesthesiology procedure codes.

Editor’s Perspective:

What We Already Know about This Topic:

  • The ability to process anesthesiology procedure code data in an accurate manner is important for clinical and research considerations. Advanced data science techniques present opportunities to improve coding and develop classification tools.

What This Article Tells Us That Is New:

  • The application of machine learning and natural language processing techniques facilitate a more rapid creation of accurate real-time models for Current Procedural Terminology code classification. The potential benefits of this approach include performance optimization and cost reduction for quality improvement, research, and reimbursement tasks that rely on anesthesiology procedure codes.