Original Research

Anesthesia Care Practice Models in the Veterans Health Administration

Although the VHA primarily relies on teams for anesthesia care, unsupervised certified registered nurse anesthetists also are used to meet veterans’ surgical care needs.

Author and Disclosure Information



Anesthesia care is provided by physician anesthesiologists, certified registered nurse anesthetists (CRNAs), anesthesiology residents, and anesthesiologist assistants. These providers may practice alone (anesthesiologists or CRNAs) or in various combinations of supervised roles and teams. Previous studies reveal mixed findings regarding whether patient outcomes differ by anesthesia practice models.1-7However, little is known about the prevalence of various anesthesia models in the US.


In recent years, anesthesiology has undergone substantial expansion in its scope of services provided, the settings in which it is provided, and the diversity of its workforce.8As the field continues to evolve, especially within the context of value-based health care reform, it is imperative to evaluate how anesthesia care models are used in health systems and how these models may optimize care delivery.

The Veterans Health Administration (VHA) is the largest integrated health care system in the US, providing surgical care in 110 inpatient medical centers and 27 ambulatory surgery centers. Despite national integration, anesthesia practices vary widely among facilities. The question of which model of anesthesia care is associated with the best outcomes and offers the most value is widely debated.1,5,7,9 As an important first step in understanding anesthesia care delivery, a baseline assessment of the practice patterns of anesthesia providers is necessary and may benefit future studies of the impact of these care models on outcomes. Thus, the aim of this work was to understand and describe the previously unassessed landscape of anesthesia care delivery within the VHA.


As part of a larger evaluation of anesthesia care delivery in the VHA, an observational assessment of anesthesia provider practice patterns was conducted using retrospective surgical data. This project complies with VHA policy pertaining to nonresearch operational activities and did not require institutional review board approval and adheres to the EQUATOR Network guidelines described in Strengthening the Reporting of Observational Studies in Epidemiology (STROBE).10

Data were obtained from the VHA Managerial Cost Accounting National Data Extract for Surgery package for all surgical procedures (n = 726,706) between October 1, 2013 and March 31, 2015. There were 420 facilities represented in these surgical data. The VHA facility records were used to specifically identify inpatient and ambulatory surgery facilities for inclusion. Additionally, to ensure facilities were valid surgical sites with sufficient surgical volume, those with 100 or fewer cases during the period were excluded. In total, 288 facilities with 9,434 surgical cases (representing 1% of cases) were excluded. These excluded facilities included nursing homes (38%), domiciliaries (26%), outpatient clinics (11%), rehabilitation programs (9%), other nonsurgical facilities (8%), and medical centers (8%). The majority (80%) of excluded medical centers had 30 or fewer surgical cases.

In 6 instances, data from subfacilities were combined with their organizationally affiliated main facilities. The final sample included 125 facilities. The VHA assigns a complexity level designation to facilities, defined as follows: 1a (most complex), 1b, 1c, 2, and 3 (least complex).11 Facilities with 1a designation perform the most complex surgical cases, such as cardiovascular surgery or neurosurgery and have more staff and resource support, whereas levels 2 and 3 facilities perform fewer and less complex cases.

Surgical records were excluded when the primary Current Procedural Terminology (CPT) code was missing (n = 85,748, or 12% of cases). This resulted in 631,524 remaining cases. The surgical CPT codes were mapped to anesthesia CPT codes to obtain the associated base unit (BU) values via a published crosswalk by the American Society of Anesthesiologists (ASA).12 A higher number of associated BUs indicates a more complex procedure. For example, procedures such as biopsies, arthroscopies, and laparoscopies receive 3 to 4 BUs, whereas a venous thrombectomy of the leg and a transurethral resection of the prostate are both 5 BUs, a total knee arthroplasty is 7 BUs, a craniotomy is 10 BUs, and a coronary artery bypass receives 18 BUs. Surgical case complexity was defined as low (3 or 4 BUs), medium (5 BUs), and high (≥ 6 BUs). Although the VHA has an existing case complexity assignment process based on CPT codes, it defines complexity differently for inpatient facilities and ambulatory surgery centers. Thus, the BU-defined complexity permitted a standardized complexity categorization across all facilities. Categorization of BUs similar to this has previously been used in the literature as a proxy for case complexity.13,14

Patient-level information included the ASA physical status classification, a measure of overall health status determined by an anesthesia provider preoperatively.15 These classifications included ASA I (healthy), ASA II (mild systemic disease), ASA III (severe systemic disease), ASA IV (severe systemic disease that is a constant threat to life), and ASA V (moribund patient who is not expected to survive without surgery). The last classification, ASA VI: brain-dead with planned organ donation, was excluded. The “E” subcategory denoting “emergency” was subsumed within the corresponding ASA category (eg, ASA V-E was combined with ASA V).

Provider data identified the principal and supervising (if present) anesthetists involved in the case. The provision of anesthesia care was categorized into 3 models: Model 1—a physician anesthesiologist supervising a CRNA; Model 2—a physician anesthesiologist practicing independently or supervising an anesthesiology resident; and Model 3—a CRNA without supervision. Surgical cases were excluded when there was no anesthesia provider (n = 95,795, or 15% of remaining cases), or a nonanesthesia provider (n = 51,647, or 8% of remaining cases) on record. The final sample was 484,082 surgical cases conducted at 125 facilities.

Related: Improving Care and Reducing Length of Stay in Patients Undergoing Total Knee Replacement

Next Article:

2018: A banner year for hematology drug approvals

Related Articles