A retrospective analysis was performed on all patients undergoing cardiac surgery over a 4-year period (N=956) at Enloe Medical Center, a small community hospital in northern California. The Enloe Institutional Review Board approved this study. Patients were divided into two cohorts based on whether their surgical procedure was first assisted by a physician/surgeon (MD) or a physician assistant (PA). Euro scores1 and predicted mortality were calculated for the patients to compare the overall risks of surgery. Surgeon assistants assisted at 78% of all cases (n=748), while PAs assisted at 22% of all cases (n=208). Risk scores for the two cohorts were virtually identical.
Cases were performed by a total of four primary surgeons, although nearly all the cases (n = 947, 98.8%) were performed by two surgeons. Fourteen surgeons first-assisted cases, although the majority of assistance was provided by two surgeons (n = 494, 66.0%; n = 209, 27.9%, respectively). Six PAs functioned as first assistants. Experience in cardiothoracic surgery for these PAs ranged from several months to greater than 10 years. The majority of the assisting was accomplished by three PAs (n= 178, 86%).
Approximately one-third of the cases in the MD cohort did not require extracorporeal bypass. By contrast, approximately two-thirds of the PA cohort cases were off-pump (P < .001) (Table 1). Off-pump bypass is generally considered more technically difficult than bypass utilizing extracorporeal bypass. The use of PAs as first assistants for off-pump cases is in keeping with California Title 22,2 which requires a three-surgeon team for cases requiring extracorporeal bypass.
Analysis of the incidence of surgery demonstrates no significant difference in the types of cases assisted by PAs (P = .057) (Table 2). This would indicate that there was no attempt by the primary surgeon to delay a reoperative case until a physician/surgeon assistant was available and demonstrates confidence by the primary surgeon in the PA's ability to assist even these more difficult cases. Mean Euro scores and predicted mortality were virtually identical for the two cohorts in all case classifications, indicating that the cases assisted by PAs were just as difficult as those assisted by physicians (Table 3).
ANOVA analysis was performed on nine different performance factors to determine differences between the cohorts. OR time was significantly less for PAs (4.09 hours ±1.73; 95% CI 0.394) than for their surgeon counterparts (4.95 hours ± 2.0; 95% CI .211) (P = .002). These results were thought to be most likely attributable to the use of PAs as assistants primarily for off-pump cases, which generally are shorter than on-pump cases. Additional analysis of OR time for only on-pump cases showed no significant difference between the two cohorts (P = .973).
Skin-to-skin incision time was also significantly shorter for PA-assisted cases (167.5 ± 82; 95% CI 11.23) than for surgeon assistants (205.9 ± 6.9; 95% CI 4.80) (P < .001). Again, this is likely the result of a tendency to utilize PAs for shorter off-pump cases. Further analysis of incision times for only on-pump cases showed no statistical significance between the two cohorts (P = .903). All other performance factors examined showed no significant difference between PA and surgeon assistants. Analysis of performance factors is shown in Table 4.
Twenty-six complications including in-hospital and 30-day mortality were analyzed using chi-square testing. Only blood product use was found to be statistically significant. Blood products were utilized for 53.37% of PA-assisted cases and for 73.93% of surgeon-assisted cases (P < .001). Given the significant difference in offpump cases performed by PAs, further analysis was performed for blood product use for on-pump cases only and was found not to be significant (P = .115). Analysis of these complications is depicted in Table 5.
This small sample size, single-site retrospective study strongly suggests that physician assistants provide safe and efficient care as first assistants for cardiac surgery cases but has limited generalizability. A randomized controlled trial should be conducted to further test this hypothesis.
To view the references and tables, please see the online version of this article at www.jaapa.com.
The authors thank Tammy Freeland, STS Database Administrator, and Pablo Zubiate, MD, CCP, Administrator for Cardiac Surgery Program at Enloe Medical Center, for their kind assistance.
1. Rouqes F, Michael P, Goldstone AS, et al. The logistic Euro-SCORE. Eur Heart J.
© 2012 American Academy of Physician Assistants.
2. California Title 22 Division 5, Article 6, §70435, (b) (2).