Influence of team composition on turnover and efficiency of total hip and knee arthroplasty

Abstract

Aims

Surgical costs are a major component of healthcare expenditures in the USA. Intraoperative communication is a key factor contributing to patient outcomes. However, the effectiveness of communication is only partially determined by the surgeon, and understanding how non-surgeon personnel affect intraoperative communication is critical for the development of safe and cost-effective staffing guidelines. Operative efficiency is also dependent on high-functioning teams and can offer a proxy for effective communication in highly standardized procedures like primary total hip and knee arthroplasty. We aimed to evaluate how the composition and dynamics of surgical teams impact operative efficiency during arthroplasty.

Methods

We performed a retrospective review of staff characteristics and operating times for 112 surgeries (70 primary total hip arthroplasties (THAs) and 42 primary total knee arthroplasties (TKAs)) conducted by a single surgeon over a one-year period. Each surgery was evaluated in terms of operative duration, presence of surgeon-preferred staff, and turnover of trainees, nurses, and other non-surgical personnel, controlling cases for body mass index, presence of osteoarthritis, and American Society of Anesthesiologists (ASA) score.

Results

Turnover among specific types of operating room staff, including the anaesthesiologist (p = 0.011), circulating nurse (p = 0.027), and scrub nurse (p = 0.006), was significantly associated with increased operative duration. Furthermore, the presence of medical students and nursing students were associated with improved intraoperative efficiency in TKA (p = 0.048) and THA (p = 0.015), respectively. The presence of surgical fellows (p > 0.05), vendor representatives (p > 0.05), and physician assistants (p > 0.05) had no effect on intraoperative efficiency. Finally, the presence of the surgeon’s ‘preferred’ staff did not significantly shorten operative duration, except in the case of residents (p = 0.043).

Conclusion

Our findings suggest that active management of surgical team turnover and composition may provide a means of improving intraoperative efficiency during THA and TKA.

Introduction

Surgical time represents one of the most valuable resources in a healthcare system in terms of its influence on both outcome and upfront cost.¹ Prolonged surgical time for a given procedure is strongly associated with an increased likelihood of an adverse event.² Consequently, excessive operative duration contributes directly, via facility and physician service fees, and indirectly, via adverse outcomes, to rising medical expenses in the USA.³ This relationship is particularly important in orthopaedic surgery, for which increasing operating room costs and surgical volume are outpacing those in other surgical subspecialties.^4,5

A growing body of evidence suggests that surgical outcome may depend on the operating team’s dynamics and the effectiveness of that team’s communication. One study found that communication failures occurred during 30% of exchanges between team members, and that in one-third of cases with at least one communication failure, patient safety was jeopardized.⁶ Gawande et al⁷ reported that intraoperative communication breakdowns caused 43% of errors committed by surgical teams.

Factors related to the dynamics of operating teams include intraoperative staff turnover (i.e. personnel changes within a given role during surgery), types of staff present, and interpersonal familiarity, among others.^8–10 For instance, Cassera et al¹¹ found that in laparoscopic procedures conducted by a general surgeon, after controlling for case complexity, each additional team member beyond four, excluding the anaesthesiologist, resulted in an additional 15 minutes of surgical time. A more recent analysis of cardiothoracic surgery focusing on nursing staff reported that increased nursing staff turnover was associated with a 29% increase in surgical time.¹² Meanwhile, an increasing body of evidence suggests that personnel decisions not just limited to nursing staff, such as overall staff turnover, have significant effects on operative efficiency outcomes.¹³ Such decisions have postoperative consequences as well, and frequency of preventable adverse events, such as surgical site infections, have been linked to operative staff turnover.¹⁴

The number of trainees present during a surgery (including fellows, residents, medical students, and nursing students) has been proposed to affect operative duration.^15,16 A 2001 study quantified the cost of this “educational investment” at $662 per case for orthopaedic residents, requiring an additional 40 minutes of operating time and 60 minutes of anaesthesia time per case.¹⁷ Similar increases in operative length due to the presence of residents and fellows are observed in other surgical sub-specialties, such as plastic surgery.¹⁸ Perhaps in an effort to offload labour and reduce costs, the prevalence of physician assistants in surgical teams has increased in recent years, but the effect on surgical duration has not been re-investigated in depth.¹⁹

The mutual familiarity of the operating room staff is also thought to play a significant role in operative efficiency.^20,21 For instance, initiatives dedicated to surgical team consistency demonstrate improvements in operative time.²² Team consistency also yields improvement in outcomes that depend on operative time, such as surgical site infection.²³ In systems without consistent operative teams, surgeons often describe ‘preferred’ staff members. We hypothesized that the presence of a ‘preferred’ team would affect operative efficiency. Hence, we investigate operating team composition and dynamics and its influence on surgical efficiency. We also investigate the impact of staff turnover, trainee presence, and availability of preferred teammates on incision time and total procedure time for total hip and knee arthroplasties.

Methods

All primary total hip arthroplasties (THAs) and primary total knee arthroplasties (TKAs) performed by a single surgeon (DFA) at a large, academic, tertiary medical centre between the dates of 1 January 2017 and 31 December 2017 were extracted. Only arthroplasties performed to treat osteoarthritis were included. In all, 112 surgeries (70 THAs and 42 TKAs) met the criteria for inclusion during the 12-month window. A review of operative notes was conducted, and details on surgical staffing and operative milestone times were extracted from operating notes updated throughout surgery and contained within an electronic health record. Specifically, ‘room time’ was defined as the entire interval the patient was in the operating room, ‘procedure time’ was defined from anaesthetic induction to rousal, and ‘incision time’ was defined from incision to closure. Staff turnover was derived from the number of individuals listed under specific roles. Since one individual occupies a defined role on the operating team at any point in time, listing multiple individuals for a given role represented the intraoperative exchange of staff members. For example, if three names were listed as ‘scrub nurse’, that would imply three staff swapped in and out of the procedure. Trainees were excluded from these calculations. A list of preferred operating room staff was solicited from the surgeon. The names of all staff participating in surgical procedures during the study window were provided retrospectively to the surgeon; for each role, including orthopaedic resident, circulating nurse, scrub nurse, vendor representative, and anaesthesiologist, one preferred staff member was selected. In order to account for operative complexity, all cases were controlled for body mass index and American Society of Anesthesiologists (ASA) score.²⁴ Additionally, all cases were controlled for presence of osteoarthritis, which was extracted from the qualitative section of operative notes.

Statistical analysis.

Descriptive statistics (frequencies and percentages) were reported for categorical variables, while means and SDs or medians and interquartile ranges (IQRs) were reported for continuous variables, such as room times. Paired t-tests or Mann-Whitney U tests were used to test for associations between room, incision, and procedure times and the presence of certain staff members, contingent on normality of observations. Multivariate linear regression was used to model room time, incision time, and procedure time including all staff member variables and adjusting for patient body mass index (BMI), ASA scores, and the presence of arthritic changes in the joint of interest. Per minute operating room costs were derived from publicly available literature, ranging from $22 to $133 per minute.²⁵ All statistical analyses were completed using RStudio (Boston, Massachusetts, USA) using a two-sided level of significance of 0.05.

Results

Composition of surgical teams during total joint replacement.

THA and TKA procedures had a mean of 9.1 (SD 1.6) and 8.4 (SD 1.1) staff members (excluding trainees) present in the operating room, respectively. The distribution of residents, circulating nurses, scrub nurses, vendor representatives, and attending anaesthesiologists were approximately equivalent for both THAs and TKAs (Table I). Trainees were present in at least half of all surgeries, and medical students were the most common, attending approximately 40% of both THA and TKA cases (Table II). Nursing students, present in 27% of THAs and 21% of TKAs, were the next most common. The presence of arthroplasty fellows and physician assistants varied substantially by procedure type. Arthroplasty fellows and physician assistants were present in 16% and 27% of THAs, respectively, but both appeared in fewer than 5% of TKAs.

Table I.

Number of core staff per total hip arthroplasty (THA) and total knee arthroplasty (TKA) procedure.

Variable	THA	TKA
Mean residents, n (SD)	1.7 (0.7)	1.6 (0.5)
Mean circulating nurses, n (SD)	2.3 (0.6)	2.2 (0.5)
Mean scrub nurses, n (SD)	1.7 (0.6)	1.7 (0.7)
Mean vendor representatives, n (SD)	0.9 (0.5)	0.7 (0.5)
Mean anaesthesiologists, n (SD)	1.8 (0.6)	1.8 (0.4)
Total, n (SD)	9.1 (1.6)	8.4 (1.1)

Table II.

Percentage of cases in which trainee and advanced practice staff were present during total hip arthroplasty (THA) and total knee arthroplasty (TKA) procedures.

Variable	THA	TKA
Presence of surgical fellows, %	15.7	4.8
Presence of medical students, %	41.4	40.5
Presence of nursing students, %	27.1	21.4
Presence of any student, %	60.0	50.0
Presence of physician assistants, %	27.1	4.8

Impact of nursing and anaesthesiology turnover on operative efficiency.

In both THAs and TKAs, turnover among subsets of operating room staff was significantly associated with increased operative duration. For THAs, increased turnover among nurses was associated with a rise in procedure and room time by 12.0 minutes and 12.1 minutes, respectively for each additional circulating nurse (p = 0.027 and p = 0.032, multivariate linear regression); and by 12.2 minutes and 17.0 minutes, respectively for each additional scrub nurse (p = 0.036 and p = 0.006, multivariate linear regression). Scrub nurse turnover was also associated with an increased incision time of 10.4 minutes (p = 0.041, multivariate linear regression; Table III). For TKAs, anaesthesiologist turnover was associated with increases in procedure and incision time by 19.5 and 20.0 minutes, respectively (p = 0.011 and p = 0.025, multivariate linear regression; Table IV). We observed that neither procedure time (r = 0.026, p = 0.511, multivariate linear regression), room time (r = 0.039, p = 0.323, multivariate linear regression), nor incision time (r = 0.034, p = 0.393, multivariate linear regression) were associated with later procedure start time.

Table III.

Estimated influence of number of core staff on total hip arthroplasty operative efficiency. Models were adjusted for body mass index, American Society of Anesthesiologists grade, and presence of arthritic changes.

	Incision time		Procedure time		Room time
Variable	Estimated change in duration, mins	p-value*	Estimated change in duration, mins	p-value*	Estimated change in duration, mins	p-value*
Number of residents	0.1	0.981	−1.6	0.775	−2.3	0.681
Number of circulating nurses	7.7	0.102	12.0	0.027	12.1	0.032
Number of scrub nurses	10.4	0.041	12.2	0.036	17.0	0.006
Number of vendor representatives	8.9	0.140	8.3	0.224	4.7	0.507
Number of anaesthesiologists	6.1	0.211	6.9	0.216	11.8	0.043

^*Multivariate linear regression.

Table IV.

Estimated influence of number of core staff on total knee arthroplasty operative efficiency. Models were adjusted for body mass index, American Society of Anesthesiologists grade, and presence of arthritic changes.

	Incision time		Procedure time		Room time
Variable	Estimated change in duration, mins	p-value*	Estimated change in duration, mins	p-value*	Estimated change in duration, mins	p-value*
Number of residents	−5.0	0.370	−5.4	0.392	−5.5	0.401
Number of circulating nurses	9.7	0.148	9.0	0.233	2.5	0.740
Number of scrub nurses	−0.7	0.882	4.2	0.429	5.9	0.282
Number of vendor representatives	7.4	0.277	6.0	0.435	2.6	0.741
Number of anaesthesiologists	20.0	0.011	19.5	0.025	10.8	0.212

^*Multivariate linear regression.

Impact of trainees and advanced practice staff on operative efficiency.

The presence of non-resident trainees of any type (medical students, nursing students, and surgical fellows) was not associated with increases in THA or TKA in any measure of operative duration. In certain cases, the presence of students was associated with significantly decreased operative time (such as overall room time) (Table V). Furthermore, the presence of a physician assistant was not associated with a significant change in operative time for THAs. Using the midpoint of per-minute operating room cost range ($77.50 per minute), these differences translate to cost reductions for each procedure: an estimated $232 per case for surgical fellows and nursing students in THA; and an estimated $775 per case for medical students in TKA (Table VI).

Table V.

Estimated impact of trainee and advanced practice staff on operative efficiency during arthroplasty. Models were adjusted for body mass index, American Society of Anesthesiologists grade, and presence of arthritic changes.

	THA estimated difference						TKA estimated difference
Presence of:	Mean incision time	p-value*	Mean procedure time	p-value*	Mean room time	p-value*	Mean incision time	p-value*	Mean procedure time	p-value*	Mean room time	p-value*
Surgical fellows†	−8.7	0.281	−16.4	0.077	−8.9	0.349	N/A	N/A	N/A	N/A	N/A	N/A
Medical students	−21.0	0.075	−22.7	0.092	−16.8	0.228	14.4	0.351	25.0	0.157	36.6	0.048
Nursing students	−28.0	0.031	−36.5	0.015	−30.2	0.050	13.1	0.311	14.5	0.324	29.7	0.055
Any student	15.2	0.267	21.3	0.175	11.0	0.496	−14.6	0.400	−18.7	0.343	−39.1	0.060
Physician assistants†	−2.2	0.756	−0.3	0.967	1.0	0.912	N/A	N/A	N/A	N/A	N/A	N/A

^*Multivariate linear regression.

^†Not analyzed for TKA due to insufficient cohort size.

N/A, not applicable; THA, total hip arthroplasty; TKA, total knee arhroplasty.

Table VI.

Estimated economic impact of trainee and advanced practice staff on operative efficiency during arthroplasty.

	THA			TKA
Presence of	Median room time, mins (IQR)	Median cost, $ (IQR)	Difference in cost per case, $	Median room time, mins (IQR)	Median cost, $ (IQR)	Difference in cost per case, $
Surgical fellows *
No	168 (152 to 193.5)	13,020.00 (11,780.00 to 14996.25)	N/A	N/A	N/A	N/A
Yes	165 (141.5 to 179)	12,787.50 (10,966.25 to 13,872.50)	232.50	N/A	N/A	N/A
Medical students
No	168 (154 to 194)	13,020.00 (11,935.00 to 15,035.00)	N/A	166 (151 to 178)	12,865.00 (11,702.50, to 13,795.00)	N/A
Yes	167 (149 to 177)	12,942.50 (11,547.50 to 13,717.50)	77.50	156 (146 to 190)	12,090.00 (11,315.00 to 14,725.00)	775.00
Nursing students
No	168 (157 to 188)	13,020.00 (12,167.50 to 14,570.00)	N/A	157 (146 to 174)	12,167.50 (11,315.00 to 13,485.00)	N/A
Yes	165 (140.5 to 193.5)	12,787.50 (10,888.75 to 14,996.25)	232.50	168 (156 to 191)	13,020.00 (12,090.00 to 14,802.50)	852.50
Any students
No	168 (157.5 to 190.5)	13,020.00 (12,206.25 to 14,841.25)	N/A	162 (154 to 178.5)	12,555.00 (11,935.00 to 13,833.75)	N/A
Yes	167.5 (145.5 to 191.5)	12,981.25 (11,276.25 to 14,841.25)	38.75	161.5 (145.25 to 186)	12,516.25 (11,256.86 to 14,415.00)	38.75
Physician assistants *
No	168 (147 to 188.5)	13,020.00 (11,392.50 to 14,608.75)	N/A	N/A	N/A	N/A
Yes	168 (157 to 192.5)	13,020.00 (12,167.50 to 14,918.75)	0.00	N/A	N/A	N/A

^*Not analyzed for TKA due to insufficient cohort size.

IQR, interquartile range; N/A, not applicable; THA, total hip arthroplasty; TKA, total knee arthroplasty.

Impact of preferred staff on operative efficiency.

We evaluated the impact of preferred staff members’ presence on operative duration. At least one preferred staff member was present in 84% of cases, and vendors were the only category in which a preferred staff member attended in a majority of cases (Table VII). Preferred scrub and circulating nurses were the next most common, present in 46% and 41% of cases, respectively. In no surgeries were all preferred staff members present simultaneously. The presence of preferred staff members was associated with variable changes on operative duration. For TKAs, preferred residents was associated with a statistically significant reduction in mean room time. However, in no other cases was the presence of preferred staff members associated with improvements in operative efficiency at any phase of surgery (Tables VIII and IX).

Table VII.

Percentage of total hip arthroplasties (THAs) in which preferred non-fellow residents, circulating nurses, scrub nurses, vendor representatives, and anaesthesiologists were present.

Variable	THAs, %
Preferred resident present	12.9
Preferred circulating nurse present	41.4
Preferred scrub nurse present	45.7
Preferred vendor representative present	64.3
Preferred anaesthesiologist present	14.3
All preferred staff members present	0.0
Any preferred staff present	84.3

Table VIII.

Impact of preferred core staff on total hip arthroplasty incision, procedure, and room times.

Variable	Change in mean incision time	p-value*	Change in mean procedure time	p-value*	Change in mean room time	p-value*
Preferred resident present	4.6	0.602	9.7	0.342	−0.4	0.972
Preferred circulating nurse present	6.5	0.312	0.2	0.983	−1.6	0.832
Preferred scrub nurse present	6.9	0.295	−0.1	0.985	−1.9	0.805
Preferred anaesthesiologist present	5.6	0.472	10.8	0.225	−1.0	0.912

^*Mann-Whitney U test.

Table IX.

Impact of preferred core staff on total knee arthroplasty incision, procedure, and room times.

Variable	Change in mean incision time	p-value*	Change in mean procedure time	p-value*	Change in mean room time	p-value*
Preferred resident present	−13.1	0.171	−17.3	0.113	−23.1	0.043
Preferred circulating nurse present	−1.0	0.886	−3.7	0.635	−3.4	0.667
Preferred scrub nurse present	5.6	0.424	−0.3	0.965	1.3	0.871
Preferred anaesthesiologist present	−4.6	0.585	−11.9	0.219	−5.3	0.588

^*Mann-Whitney U test.

Discussion

Our findings demonstrate a relationship between the composition and dynamics of surgical teams and operative efficiency in arthroplasty. We observed considerable non-resident turnover that was associated with an increased operative duration in both TKAs and THAs. This loss of efficiency was especially pronounced for turnover among anaesthesiologists, circulating nurses, and scrub nurses. The essential role of critical staff members in procedure execution preoperatively (via anaesthetic induction), intraoperatively (via instrument acquisition and exchange), and postoperatively (via room turnover) likely contribute to the impact of their turnover on multiple measures of surgical efficiency. Indeed, these results illustrate that even though handoffs may appear to occur seamlessly when staff replace one another, numerous delays associated with factors such as replacement staff’s unfamiliarity with the nature or status of the case; the availability and layout of instruments; and/or the communication patterns of the team can aggregate to clinically meaningful delays.

Despite previous reports suggesting that the presence of trainees may increase operative duration in non-orthopaedic surgeries, we did not observe such impedance on intraoperative efficiency.¹⁵ Rather than a costly “educational investment,” surgical fellows and other trainee groups may actually represent drivers of cost reduction in arthroplasty. It is likely that the impact of trainees will differ depending on the type of procedure, their level of experience, and the relative level of engagement it requires from non-surgeons. Finally, physician assistants did not have an impact on intraoperative efficiency. It is possible that physician assistants generated efficiencies in other aspects of perioperative and clinical care but that these benefits were not observed when analyzing surgical time alone.

Although many surgeons attest anecdotally to the importance of particular team members, we did not observe a clear relationship between the presence of any preferred staff members and intraoperative efficiency. In the sample, fewer than one preferred operating room team member was present during any given case, and in no cases was the entire preferred team present. Of all staff, preferred residents were least likely to be present in THA. Preferred nursing staff were also present in the minority of procedures. With the exception of residents in TKA, no preferred staff member alone significantly improved efficiency at any phase of surgery, so it is not clear that requesting individual preferred teammates leads to increased surgical efficiency.

Several limitations exist for this study. The sample is relatively small, comprising total joint replacements from a single surgeon at one institution over one year. Additionally, as this study relied upon registration of staff and times in the electronic medical record, it is possible that coding inconsistencies were made which would reduce the reliability of this data. Furthermore, while preferred teams may or may not improve operative efficiency, they could nonetheless improve patient outcomes independent of surgical time. It is also possible that trainees, such as students, selectively joined for less complicated cases. This would imply that the observed shorter case times associated with trainees were an artifact of selection bias, rather than trainee contribution to operative efficiency. Additionally, we considered the possibility that the observed associations between staff turnover and operative duration might be enhanced if complex cases were scheduled later in the day, when routine staff changes may be more common. However, the observed lack of association between procedure start time and any measure of operative duration make this less likely. Lastly, while several measures for case complexity were adjusted for in these models, other qualitative considerations that may effect intraoperative efficiency more dramatically may not have been included here.

In conclusion, our analysis demonstrates that features of operative teams such as the presence and turnover of key individuals plays a significant role in operative efficiency for routine primary total hip and knee arthroplasties. In conjunction with emerging strategies for improving operative outcomes, such as operating theatre optimization,²⁶ our results suggest that protocols for managing surgical staff may play an important role in enhancing operative efficiency and improving patient outcomes. Subsequent investigations tracking larger cohorts across multiple surgeons and monitoring longitudinal postoperative outcomes beyond operative duration can help expand on this field of study.

Take home message

• Features of operative teams such as the presence and turnover of key individuals play a meaningful role in surgical efficiency for routine primary total joint replacements.

• While surgeons often describe the impact of subjectively “preferred” staff on operative efficiency and outcomes, their contribution is not immediately apparent according to objective measures of efficiency.

• Protocols for managing surgical staff may play an important role in enhancing operative efficiency and improving patient outcomes.