Improving physician in training well-being and reducing fatigue are an increasing priority in Canada, prompting regulatory changes limiting resident duty hours. In Ontario, the Professional Association of Interns and Residents of Ontario-Council of Academic Hospitals of Ontario (PAIRO-CAHO) has terms of employment that limit in-hospital overnight calls to a maximum of one in four nights and a maximum of 24 hours consecutive work.1 A crossover period for adequate patient care handover also is required.1
Many other factors affect the resident hospital workforce. For example, the resident workforce has not increased, and fewer housestaff are available in hospital to provide continuity in patient care, particularly when a resident is postcall. Furthermore, increasing formal education obligations for residents have further taken residents away from the hospital. Lastly, although computerized physician order entry has been shown to decrease medical error and may improve organizational efficiency, computerized orders take more provider time at point of entry, thus residents need more time to enter computerized patient orders than handwritten ones.2,3 Restricted resident work hours have already caused significant workforce issues in Canada, and coupled with an aging population and increasing prevalence of chronic disease suggests that PAs and NPs will be increasingly necessary to maintain quality care and patient safety. These challenges have led many academic hospitals to consider using physician assistants (PAs). PAs are less well-known in Canada than in the United States, although the NP is well-known in Canada. PA practice in the United States is guided by the national professional society, the American Academy of Physician Assistants (AAPA), but governed and regulated in each state. The AAPA recommends that state PA practice laws include:
- Licensure as the regulatory term
- Full prescriptive authority
- Scope of practice determined at the practice level
- Adaptable collaboration requirements
- Chart cosignature requirements determined at the practice
- Number of PAs a physician may supervise determined at the practice level.4
PAs were introduced into the Ontario healthcare system in 2007. They are not a regulated healthcare profession, and have a more restrictive scope of practice until regulation can be achieved.5 Meantime, medical directives are implemented, specifying the medical procedures and treatments that may be performed by the PA on a defined patient population under the supervising physician(s); these directives are specific to each institution.6 In Ontario, the typical PA model is one PA under the direct supervision of one staff physician in primary care internal or emergency medicine within a community setting. Little data exist as to the effect of PAs in this model of care.
Recently, the Division of General Surgery at Mount Sinai Hospital in Toronto employed two PAs as members of the academic general surgical inpatient care team. The team consists of surgeons, clinical fellows, residents, and medical students. Because this was both a new PA model and unique PA position, we wanted to describe the effect of the PA in the team setting. The objectives of this study were to:
- determine if PAs reduced the number of late hospital discharges (defined as occurring after 10 a.m.)
- determine if PAs decreased the number of hours residents spent on the electronic medical record (EMR)
- elicit residents' perceptions of the PA as a member of the surgical team.
Two PAs were employed in the Division of General Surgery at Mount Sinai Hospital starting in November 2012. They were assigned to one of three general surgery teams; each team consists of two to three staff surgeons, one clinical fellow, one senior resident, and one junior resident. Each team provides care for about 20 acute surgical patients undergoing breast, colorectal, or multivisceral resections. The PA's work hours were 6:30 a.m. to 2:30 p.m. Monday through Friday. The PAs attended morning care rounds with the residents; gave patient report to nurses; attended interdisciplinary rounds; planned for patient discharge; assisted in clinic with patient interviews and physical examinations; assisted in the OR with retraction, closure of incisions, and patient transfer to postoperative recovery; and assessed on the unit and in the ED by triaging patients and alerting the surgical team about urgent issues.
Over a 13-month period (December 2012 to December 2013), the PAs prospectively collected discharge data onto data abstraction sheets developed by the PAs and their supervising physicians. The data collected included timely completion of home care referrals, time of discharge orders entered into the EMR, prescriptions, and any reasons for delay of discharge. Data were documented in a binary fashion (“yes” or “no”) to state whether the PAs and resident teams were able to complete the home care referral the day before discharge, enter the discharge order before 10 a.m., and have the prescription written (if patient required one) by a resident for the 10 a.m. discharge. Reasons for delay of discharge were documented in free text by the PAs as either delay due to late discharge order, medical reassessment, resolution of outstanding medical issue, or awaiting staff approval of the patient discharge plan. Data from the abstraction sheets were combined and analyzed in a spreadsheet program. A planned discharge was defined as any discharge that met the following criteria:
- anticipated hospital discharge was communicated to the nursing team 12 to 24 hours in advance
- the home care referral was completed the day before discharge
- prescriptions were ready by 10 a.m. on the discharge day
- a discharge order was entered into the EMR by 10 a.m. on the discharge day.
A planned late discharge did not meet these four requirements; for example, if the prescription was not written before 10 a.m. During the early data collection phase of the study, unplanned early discharge emerged as an unexpected but important outcome measure that was included in the study. Unplanned early discharges were defined as patients deemed not ready for discharge during morning rounds but who were subsequently discharged later the same day after reassessment by the surgical team. Surgical reassessment often involved follow-up of adequate oral intake, blood work, imaging results, or liaison with a consulting service
These data collected by the PAs were compared with data previously collected over 2 months by nursing unit administrators on the surgical units, who examined the reasons for delayed patient discharges. These data identified that late physician order entry contributed to delayed patient discharges and represented an area for improvement in patient care.
Resident computer time
All orders for patients admitted to the surgical unit at Mount Sinai Hospital must be entered into the EMR. Test results, interdisciplinary notes, and active medications also are documented in the EMR. The hospital informatics department provided the number of hours that each junior resident logged onto the EMR system each week to provide direct patient care. We used these hours as a surrogate measure of resident workload. Users are automatically logged out of the system after being idle for 20 minutes. The junior resident is typically the team member who does the electronic order entry for inpatients because after morning rounds, senior residents immediately report to the OR for patient preparation and staff surgeon report. Junior residents also are the first to be called for unit and on-call issues. Because of this, we excluded senior residents from the analysis and only compared junior resident hours logged on the EMR for teams with and without a PA.
Resident perceptions of PAs on the team
We developed a resident satisfaction survey to elicit resident perceptions of the effect PAs had on patient care and resident workload. The unvalidated survey consisted of 24 items rated on a 5-point Likert scale. The survey was e-mailed to all fellows and residents on the two teams with a PA at the end of their rotation and was reviewed anonymously. One e-mail reminder including the survey link was sent to all nonresponders 2 weeks after the initial e-mail.
Descriptive statistics were used to present means and standard deviations for continuous variables and proportions for dichotomous variables. Differences were tested using the student's t-test for continuous data. Resident survey data and any commentary were examined but not analyzed on software.
Over the 13-month study period, the two surgical teams with PAs had 848 planned discharges (Table 1). Of these, 0.5% (4 of 848) did not have discharge orders by 10 a.m. and were considered late discharges using the study criteria for late discharges. Discharge data gathered before PAs joined the teams showed a late discharge rate of 20% (25 of 126) over 2 months, with the most common reason for delay being home care forms that were not completed on time.
Teams with PAs had 132 unplanned early discharges, primarily because the PAs ensured that patients were reassessed before 2:30 p.m. to determine if they were ready for discharge. Similarly, the PAs could resolve any outstanding medical issues that were preventing discharge in an ongoing fashion throughout the day until 2:30 p.m. Surgical teams that did not have a PA often were unable to perform reassessments until the end of the day when the residents returned to the units after finishing their clinical duties in the OR or clinic. By this time in the evening, discharge was difficult to coordinate. For example, allied health staff and home care coordinators were no longer in hospital and the patient had to wait until the following day for safe discharge.
Over the study period, 41 junior residents completed a general surgery rotation at Mount Sinai Hospital. Junior residents on the team without a PA spent about 10 hours more per week on the EMR compared with junior residents on teams with PAs (31.3 hours compared with 21 hours, P<0.05). No significant difference was found in the number of hours the junior residents logged on the EMR between the two teams in which a PA was assigned (19.7 hours versus 21.6 hours, P>0.05).
Resident perceptions of PAs on the team
The survey to elicit resident perceptions effect of PAs on resident workload was sent to all 72 residents who completed a general surgery rotation at Mount Sinai Hospital during the study period. Of these, 45 returned the completed survey, for a response rate of 63%.
In terms of resident workload, residents' perceptions were that having a PA on the team significantly decreased the time residents spent completing home care forms (91%) and discharge summaries (91%), decreased the number of pages that they received (77%), and made it significantly easier to get to the OR (78%) and educational sessions (79%), and leave postcall by noon (71%) (Table 2). Optional comments by the residents on the survey suggested that one of the biggest gains was letting them get to the OR in the morning on time and letting junior residents get to the OR more often. Other positive statements received were: “Before the PA arrived I was struggling to keep up with basic patient care... I think most importantly it improved patient safety,” and “As constant team members they eased orientation into the surgical system ... enhanced patient flow and communication with the nurses.” Importantly, 90% (38 of 42) agreed that the PAs significantly improved the overall quality of their surgical rotation when compared with other surgical rotations without a PA.
Residents also agreed that the PAs significantly improved patient care and communication. Residents indicated that the PAs significantly increased the efficiency of the discharge process (93%) and morning rounds (98%); improved consistency and quality of morning report (79%); improved quality of communication between the surgical team, nursing, and other healthcare professionals (93%); decreased wait time for patients to be assessed on the unit (79%); and significantly improved the overall quality of care provided to patients (93%).
Our study suggests that PAs can be successfully integrated on a surgical resident team in a tertiary academic setting in Canada. Overall, the introduction of PAs onto our surgical teams
- decreased late planned discharges after 10 a.m.
- increased unplanned early discharges with earlier reassessment and/or coordination of care
- decreased resident workload
- improved resident satisfaction with their rotation.
Implementation of PAs on our general surgery service led to a significant decrease in late discharges from 20% to 0.5% and increase in unplanned early discharges from 0% to 16%. This resulted in improved bed flow and possibly translated into a 1-day decrease in length of stay on our service, according to the quarterly hospital audit. Length of stay was not measured in this study; however, the audit data were brought forth to the Division of General Surgery and coincided with implementation of the PAs. These results are in keeping with other studies that have reported a decrease in length of stay by 0.7 to 3 days with the use of PAs and NPs.7-10 The PAs on our service were able to effect this change primarily by ensuring that home care referrals were completed before the day of discharge and that prescriptions were written in advance for signature by the physician.
Another significant advantage of having PAs was that they were available during the day, when staff and residents were in the OR, to coordinate discharge planning with the nursing team and other healthcare professionals. PAs acted as a liaison with these services and addressed all patient issues, allowing for coordinated discharge planning. These observations are supported by Moote and colleagues, who surveyed organizational leaders at major academic medical centers across the United States for their opinions on the effects of integrating PAs or NPs on resident teams.7 Respondents ranked accessibility of the PAs and NPs to other members of the interprofessional team as the most valuable contribution of the PA to the team and organization.
This study showed that PAs significantly reduced the electronic order entry workload for junior residents. Junior residents with a PA on the team were logged onto the EMR 10 hours less on average per week than residents without a PA. This was despite the fact that PAs during this study time period were not able to enter medical orders themselves. Most likely, the time saved was due to the PAs reviewing blood work and imaging and summarizing pertinent results for the residents, entering orders for quicker resident cosignature, and completing home care referrals. Although these results show a significant reduction in the number of hours that junior residents were logged onto the EMR each week, these data indicate that even with PAs, junior residents were spending 3 to 4 hours per day on the EMR. Since this study was completed, medical directives have been developed to let PAs enter orders without physician cosignature for blood work, blood cultures, ECGs, and nasogastric tube insertion after demonstration of appropriate training. This could further reduce the number of hours per day that junior residents enter electronic orders.
The survey also found that 80% of residents reported that PAs significantly decreased the time spent completing home care forms and discharge summaries compared with experiences on other surgical teams without PAs, decreased the number of pages that they received, and made it significantly easier to get to the OR and educational sessions and leave postcall by noon. Residents also felt that integration of the PAs onto the service significantly improved patient care primarily by facilitating more effective communication between the surgical team, nursing, and other healthcare professionals.
These findings are similar to other studies on the integration of PAs or NPs that have been conducted in the United States, where the PA role is well-established. These studies also found that resident workload is decreased, based on a variety of measures including number of pages received, hours slept on-call, hours spent in hospital, hours spent in the OR, amount of resident time saved, and improved resident written test scores.6,11-16
Other studies also have shown high resident satisfaction with PAs.17-19 One Canadian group collected qualitative surveys from residents to explore their opinions about PAs on an orthopedic surgery service.20 Of the six residents interviewed, five reported that PAs reduced their workload by decreasing their clinical responsibilities and three reported that their training was improved by the staff surgeon providing closer supervision during surgery when the PAs acted as second assistants.
One of the main concerns of program directors and department chiefs across Canada about integrating PAs onto resident teams is whether the PA may interfere with resident learning and preparedness.1 Although initially our group shared these concerns, we found the opposite to be true. We found that because the residents needed to supervise the PAs, they needed to know how to manage the units themselves and learned by reviewing the PAs' work and troubleshooting problems when they occurred. The opportunity for residents to work with PAs also prepares them for future practice that will likely include team-based medicine.21 Residents had to be able to communicate and delegate specific tasks to the PAs. In this way, our group felt that the residents not only learned how to do these tasks but also incorporated other CanMEDS competencies beyond medical scholar, including communicator and manager. (CanMEDs, a framework of different roles outlined by the Royal College of Physicians and Surgeons of Canada, sets the standards for expected competencies of resident physicians and is the basis of their formal evaluations.22)
This study had several limitations. Our discharge data did not include weekend discharges. This may have biased the results to higher late discharge and lower unplanned early discharge rates because the data exclude 2 of the 7 potential discharge days per week. However, we do not feel that this would have significantly altered the overall results.
We were unable to control for the amount of call or vacation for the junior residents but feel that this was unlikely to change the overall study results. The amount of call and vacation was likely to have evened out for all residents over the 13-month data collection period. The response rate to the resident survey was 63%, in keeping with other physician surveys. Although this may introduce an element of response bias, perceptions of the nonresponders are unlikely to have significantly affected our survey results because most respondents felt strongly that PAs reduced resident workload and improved patient care.
Finally, our data also are limited by different time frames and different methods that were used to collect the data before and after the implementation of the PAs at our institution.
Our results suggest that PAs can be successfully integrated onto surgical resident teams and lead to positive outcomes for patients while reducing resident workload in Canada. The PA-resident team model described in this study also may be applicable across all fields of medicine.
1. Wu PE, Stroud L, McDonald-Blumer H, Wong BM. Understanding the effect of resident duty hour reform: a qualitative study. CMAJ Open
2. Eslami S, de Keizer NF, Abu-Hanna A. The impact of computerized physician medication order entry in hospitalized patients—a systematic review. Int J Med Inform
3. Shu K, Boyle D, Spurr C, et al. Comparison of time spent writing orders on paper with computerized physician order entry. Stud Health Technol Inform
. 2001;84(Pt 2):1207–1211.
7. Moote M, Krsek C, Kleinpell R, Todd B. Physician assistant and nurse practitioner utilization in academic medical centers. Am J Med Qual
8. Christmas AB, Reynolds J, Hodges S, et al. Physician extenders impact trauma systems. J Trauma
9. Russell D, VorderBruegge M, Burns SM. Effect of an outcomes-managed approach to care of neuroscience patients by acute care nurse practitioners. Am J Crit Care
10. Decloe M, McCready J, Downey J, Powis J. Improving health care efficiency through the integration of a physician assistant into an infectious diseases consult service at a large urban community hospital. Can J Infect Dis Med Microbiol
11. Henderson NC, Cookson DT, Plumley R, et al. The influence of a nurse practitioner on out of hours work intensity for surgical house officers. Scott Med J
12. Holzman MD, Elkins CC, Neuzil DF, Williams LF Jr. Expanding the physician care team: its effect on patient care, resident function, and education. J Surg Res
13. Victorino GP, Organ CH Jr. Physician assistant influence on surgery residents. Arch Surg
14. Jones PE, Cawley JF. Workweek restrictions and specialty-trained physician assistants: potential opportunities. J Surg Educ
15. Spisso J, O'Callaghan C, McKennan M, Holcroft JW. Improved quality of care and reduction of housestaff workload using trauma nurse practitioners. J Trauma
16. Stahlfeld KR, Robinson JM, Burton EC. What do physician extenders in a general surgery residency really do. J Surg Educ
17. Buch KE, Genovese MY, Conigliaro JL, et al. Non-physician practitioners' overall enhancement to a surgical resident's experience. J Surg Educ
18. Reines HD, Robinson L, Duggan M, et al. Integrating midlevel practitioners into a teaching service. Am J Surg
19. Mathur M, Rampersad A, Howard K, Goldman GM. Physician assistants as physician extenders in the pediatric intensive care unit setting-A 5-year experience. Pediatr Crit Care Med
20. Bohm ER, Dunbar M, Pitman D, et al. Experience with physician assistants in a Canadian arthroplasty program. Can J Surg
21. Russell JC, Kaplowe J, Heinrich J. One hospital's successful 20-year experience with physician assistants in graduate medical education. Acad Med