Safety and feasibility of a single-surgeon cardiac surgery program during workforce shortage: a propensity-matched cohort study

Abstract

Objectives

This study compared clinical outcomes of cardiac surgeries performed by a single surgeon (operating alone [OA]) versus those performed by multiple surgeons (operating together [OT]) to evaluate the feasibility and safety of different surgical staffing models.

Methods

We retrospectively reviewed adult patients who underwent cardiac surgery at Chung-Ang University Hospital from September 2016 to August 2021. Between September 2016 and August 2018, two or three surgeons assisted each other (OT), while from September 2018 to August 2021, a single surgeon operated alone (OA). After propensity score matching, 79 patients were selected in each group for analysis.

Results

Major postoperative adverse events, including mortality, stroke, new-onset atrial fibrillation, and low cardiac output, did not statistically significant differ between groups. Other postoperative outcomes such as, acute kidney injury, ICU stay length, hospital days, CPB time, ACC time was not statistically different, whereas operation time was shorter in the OT group than in the OA group (median 300 vs. 379 min, P < 0.001).

Conclusions

Except for differences in operation time, there were not significantly different short-term postoperative outcomes in OT and OA groups. Given the various limitations of the present study, the result should be interpreted cautiously.

Introduction

In a typical cardiac surgery setting, three medical personnel are involved — the operator, a licensed cardiac surgeon; the first assistant, who may be a trainee or licensed cardiac surgeon; and the second assistant, typically a trainee or surgical assistant. While variations may exist depending on the medical center’s policy, it is generally recommended that a minimum of two cardiac surgeons participate in a cardiac surgery program [1]. It is uncommon that a single cardiac surgeon performs cardiac surgeries.

Contemporary cardiac surgery is undergoing a structural change by increasing patient complexity, limited perioperative resources (e.g., ICU capacity) and interactions with evolving interventional cardiology. Patients referred to cardiac surgery are older and have more comorbidities. At the same time, the evolving transcatheter and percutaneous interventions have changed referral patterns and patient’s characteristics, which cardiac surgery performed to more complex or higher-risk patients. These changes have increased the need for sustainable cardiac surgical staffing models that can maintain patient safety.

South Korea is a country of more than 50 million people and about 70 hospitals with cardiac surgery programs, performing about 14 thousand cardiac surgeries annually [2, 3]. About 50% of the cardiac surgeries in South Korea are performed at five big volume centers, all in the Seoul area. About 20 to 30 residents begin thoracic and cardiovascular surgery training every year; while the number of retiring cardiothoracic surgeons is expected to exceed the number of newly trained cardiac surgeons over the next decade. The uneven geographic distribution and imbalanced turnover in staffing within the cardiac surgery field pose significant challenges for many hospitals in recruiting an adequate number of cardiac surgeons. Moreover, in 2024, due to the Korean government’s aggressive expansion of medical school enrollments, many residents have resigned en masse, raising serious concerns about the future supply of cardiac surgeons.

Chung-Ang University Hospital, located in Seoul, South Korea, is a medium-volume center approximately 900 beds. Historically, the cardiac surgery program of Chung-Ang had three cardiac surgeons. However, following the resignation of one surgeon and the retirement of another, the program was left with only one cardiac surgeon operating alone from September 2018 to August 2021.

This study aimed to evaluate the safety and feasibility of a single-surgeon cardiac surgery program during a period of workforce shortage. We compared short-term postoperative outcomes between operations performed by a single attending cardiac surgeon without physician assistant and those performed by an attending cardiac surgeon assisted by another cardiac surgeon, fellow or resident physician.

Materials and methods

Study populations

This study included patients older than 18 years of age who underwent cardiac surgery at the Chung-Ang University Hospital from September 2016 to August 2021. Patients were categorized into two groups according to the number of surgical staffs: surgeries performed between September 2016 and August 2018 by two or three cardiac surgeons operating together (operating together [OT] group), and surgeries performed between September 2018 and August 2021 by a single cardiac surgeon operating alone (operating alone [OA] group). Operating together (OT) was defined as cardiac surgery performed by an attending cardiac surgeon with a physician first assistant. The physician’s first assistant could be another attending cardiac surgeon, a fellow, or a resident in cardiovascular surgery department. In the OT group, the physician first assistant primarily assisted the first operator, in addition to support from non-physician surgical assistants and the standard perioperative team, including anesthesiologists, perfusionists, operating room nurses, and other surgical staff. Operating alone (OA) was defined as cardiac surgery performed by a single attending cardiac surgeon without another physician serving as the first assistant. In the OA group, the attending cardiac surgeon was supported by non-physician surgical assistants and the standard perioperative team as OT group. During the OA period, the single attending cardiac surgeon was responsible for all the cardiac surgery program and was on call 24 h a day, 7 days a week. After regular working hours, initial inpatient monitoring and routine patient care were supported by an on-duty intern physician. While important clinical decision, cardiac surgery–related decisions, postoperative surgical issues, and emergency operative decisions remained the responsibility of the one attending cardiac surgeon.

Study endpoint

The primary outcome was major post-operative adverse events such as 72-hour mortality, 30-day mortality, stroke, new-onset atrial fibrillation, low cardiac output and reoperation for bleeding, respectively. The secondary outcomes were other post-operative outcomes such as acute kidney injury (AKI), severe AKI requiring renal replacement therapy, ventilation time, operation time, intensive care unit (ICU) and hospital stay duration, and sternal wound infections. Low cardiac output was defined as a cardiac index < 2.0 L/min/m² and the requirement for prolonged inotropic support (> 24 h) or mechanical circulatory support to maintain adequate systemic perfusion. Postoperative acute kidney injury was defined according to the Kidney Disease: Improving Global Outcomes (KDIGO) criteria as an increase in serum creatinine ≥ 0.3 mg/dL within 48 h or ≥ 1.5 times baseline within 7 days, or urine output < 0.5 mL/kg/h for at least 6 h. Severe AKI requiring renal replacement therapy was defined as AKI requiring initiation of renal replacement therapy, including continuous renal replacement therapy or intermittent hemodialysis.

Statistical analysis

Continuous variables are expressed as the mean ± standard deviation (SD) or median and inter-quartile ranges according to the normal distribution of the data. Categorical variables are reported as the sample number and percentage. The mean differences for the continuous variables between the OT and OA groups were analyzed using independent two-sample t-tests or Wilcoxon rank-sum tests according to the normal distribution of the data. A chi-squared test with Yates’ correction or Fisher’s exact test was performed to analyze the association between operation groups and categorical variables. Spearman’s correlation analysis was performed between the number of off-pump coronary artery bypass (OPCAB) anastomosis and operation time in OPCAP patients. OPCAB was defined as coronary artery bypass grafting (CABG) performed without the use of cardiopulmonary bypass. The subgroup analysis in patients with undergone OPCAB was intention-to-treat analysis. Actually, one cases of OPCAB was converted to conventional CABG. For descriptive analyses of cardiopulmonary bypass (CPB) and aortic cross clamp (ACC) times in the overall cohort, cases in which these variables were not applicable were excluded from the calculation. Differences in baseline characteristics between the OT and OA groups were adjusted using propensity score matching (PSM). The propensity score was calculated using a logistic regression model that included patients’ characteristics in Table 1, including age, sex, body surface area, left ventricular ejection fraction, estimated glomerular filtration rate, euro SCORE II, hypertension, diabetes mellitus, pulmonary hypertension, and surgical priority, and type of surgery presented in Table 2 as categorical variable. PS matching was performed on the logit of the PS within caliper of 0.2 standard deviation. Standardized mean differences (SMD) were presented to assess group balance before and after propensity score matching. After matching, no substantial imbalance (SMD < 0.1) was observed in the matched variables (age, sex, body surface area, left ventricular ejection fraction, estimated glomerular filtration rate, euro SCORE II, hypertension, diabetes, pulmonary hypertension, surgical priority, and type of surgical procedure) between the OT and OA groups. Logistic regression analyses were performed for each binary postoperative outcome to estimate adjusted odds ratios and 95% confidence intervals using clinically relevant variable (age, sex, body surface area, estimated glomerular filtration rate, euro SCORE II, hypertension, diabetes, pulmonary hypertension, surgical priority, and type of surgical procedure). Statistical significance was defined as a P-value less than 0.05. Statistical analyses were performed using R Statistical Software (version 4.1.2; R Foundation for Statistical Computing, Vienna, Austria) and T&F program, ver. 4.0 (YooJin BioSoft, Korea) (Table 3).

Table 1.

Baseline characteristics

Variable	Before PSM				After PSM
	OT group (n = 100)	OA group (n = 100)	P-value	SMD	OT group (n = 79)	OA Group (n = 79)	P-value	SMD
Age, years	67.50 (54.25–76.00)	65.00 (56.25–74.75)	0.559	0.039	67.00 (54.50–75.00)	65.00 (56.00–73.00)	0.562	0.027
Sex (female)	68 (68.0%)	56 (56.0%)	0.109	0.120	29 (36.7%)	31 (39.2%)	0.578	0.025
BSA	2.88 (2.46–3.39)	2.83 (2.31–3.17)	0.137	0.243	2.83 (2.37–3.31)	2.86 (2.43–3.21)	0.992	0.036
LVEF, %	60.00 (45.00–65.00)	57.00 (44.00–64.00)	0.502	0.057	61.00 (51.00–65.00)	55.00 (44.00–64.00)	0.159	0.092
GFR, mL/min	80.25 (50.09–98.75)	85.97 (65.32–100.16)	0.261	0.173	81.22 (55.59–99.02)	86.04 (65.37-100.14]	0.564	0.045
euroSCORE II, %	0.03 (0.02–0.07)	0.03 (0.01–0.06)	0.573	0.151	0.03 (0.01–0.06)	0.03 (0.01–0.06)	0.846	0.014
Hypertension, n (%)	64 (64.0%)	53 (53.0%)	0.151	0.110	44 (55.7%)	46 (58.2%)	0.872	0.025
Diabetes mellitus, n (%)	31 (31.0%)	36 (36.0%)	0.549	0.050	26 (32.9%)	27(34.2%)	1.000	0.013
COPD, n (%)	7 (7.1%)	5 (5.0%)	0.738	0.021	5 (6.5%)	2 (2.5%)	0.419	0.039
Pulmonary hypertension, n (%)	36 (38.3%)	40 (40.0%)	0.924	0.017	28 (37.8%)	31 (39.2%)	0.990	0.014
Stroke, n (%)	23 (23.0%)	23 (23.0%)	1.000	0.000	19 (24.0%)	17 (21.5%)	0.850	0.025
Atrial fibrillation, n (%)	16 (16.0%)	16 (16.0%)	1.000	0.000	12 (15.2%)	14 (17.7%)	0.830	0.025
Prior cardiac surgery, n (%)	2 (2.0%)	2 (2.0%)	1.000	0.000	2 (2.5%)	0 (0%)	0.477	0.025
NYHA Class, n (%)			0.003				0.001
I	21 (21.0%)	9 (9.0%)		0.120	19 (24.1%)	6 (7.6%)		0.165
II	37 (37.0%)	44 (44.0%)		0.070	28 (35.4%)	36 (45.6%)		0.101
III	33 (33.0%)	23 (23.0%)		0.100	27 (34.2%)	18 (22.8%)		0.113
IV	9 (9.0%)	24 (24.0%)		0.150	5 (6.3%)	19 (24.1%)		0.172
Priority, n (%)			0.253	0.070			0.825	0.025
Elective	80 (80.0%)	87 (87.0%)			66 (83.5%)	68 (86.1%)
Non-elective	20 (20.0%)	13 (13.0%)			13 (16.5%)	11 (13.9%)

BSA, body surface area; LVEF, left ventricular ejection fraction; GFR, glomerular filtration rate; COPD, chronic obstructive pulmonary disease; NYHA, New York Heart Association

Continuous variables are presented as the mean ± standard deviation or median (inter-quartile range), and P-values were computed using an independent two sample t-test or Wilcoxon rank-sum test according to the normal distribution. Categorical variables are presented as the sample number (%), and P-values were computed using a chi-squared test with Yates’ correction or Fisher’s exact test

Table 2.

Types of operations

Variable	Before PSM				After PSM
	OT group (n = 100)	OA group (n = 100)	P-value	SMD	OT group (n = 79)	OA Group (n = 79)	P-value	SMD
CABG	46 (46.0%)	52 (52.0%)	0.479	0.060	32 (40.5%)	40 (50.6%)	0.263	0.091
Off pump	20 (20.0%)	37 (37.0%)	0.012	0.160	20 (25.3%)	24 (30.4%)	0.594	0.051
On pump	26 (26.0%)	15 (15.0%)	0.080	0.100	12 (15.2%)	16 (20.3%)	0.532	0.013
Valve surgery	45 (45.0%)	38 (38.0%)	0.389	0.070	35 (44.3%)	32 (40.5%)	0.747	0.038
Single value	34 (34.0%)	29 (29.0%)	0.543	0.050	26 (32.9%)	24 (30.4%)	0.864	0.025
Multiple values	11 (11.0%)	9 (9.0%)	0.814	0.020	9 (11.4%)	8 (10.1%)	1.000	0.013
Aorta	11 (11.0%)	4 (4.0%)	0.107	0.070	4 (5.1%)	4 (5.1%)	1.000	0.000
Combined surgery	14 (14.0%	8 (8.0%)	0.258	0.060	6 (7.6%)	8 (10.1%)	0.780	0.025
Valve+CABG	7 (7.0%)	5 (5.0%)	0.766	0.020	3 (3.8%)	5 (6.3%)	0.717	0.025
Aorta+CABG	2 (2.0%)	1 (1.0%)	1.000	0.010	0 (0%)	1 (1.3%)	1.000	0.013
Aorta+valve	9 (9.0%)	2 (2.0%)	0.063	0.070	3 (3.8%)	2 (2.5%)	1.000	0.013
Others	3 (3.0%)	8 (8.0%)	0.215	0.050	3 (3.8%)	7 (8.9%)	0.327	0.051

CABG, coronary artery bypass graft

Variables are presented as the sample number (%), and P-values were computed using a chi-squared test with Yates’ correction or Fisher’s exact test

Table 3.

Postoperative outcomes

Variable	Before PSM			After PSM
	OT group (n = 100)	OA group (n = 100)	P-value	OT group (n = 79)	OA goup (n = 79)	P-value
Mortality < 72 h, n (%)	2 (2.0%)	2 (2.0%)	1.000	1 (1.3%)	1 (1.3%)	1.000
Mortality < 30 days, n (%)	6 (6.0%)	4 (4.0%)	0.746	5 (6.3%)	4 (5.1%)	1.000
Stroke, n (%)	6 (6.0%)	4 (4.0%)	0.746	5 (6.3%)	4 (5.1%)	0.717
New onset atrial fibrillation, n (%)	26 (26.0%)	22 (22.0%)	0.619	21 (26.6%)	17 (21.5%)	0.577
Low cardiac output syndrome, n (%)	32 (32.0%)	24 (24.0%)	0.270	21 (26.6%)	21 (26.6%)	1.000
Reoperation for bleeding, n (%)	9 (9.0%)	4 (4.0%)	0.251	7 (8.9%)	4 (5.1%)	0.532
Acute kidney injury, n (%)	15 (15.0%)	6 (6.0%)	0.065	13 (16.5%)	5 (6.3%)	0.080
Acute kidney injury requiring dialysis, n (%)	3 (3.0%)	4 (4.0%)	1.000	4 (5.1%)	2 (2.5%)	0.677
Ventilation time (min)	540.00 (265.00–1150.00)	420.00 (300.00–1305.00)	0.958	465.00 (252.50–1130.00)	480.00 (300.00–1260.00)	0.427
Operation time (min)	317.50 (260.00–428.75)	374.00 (320.25–419.00)	0.001	300.00 (250.00–350.00)	379.00 (334.00–429.00)	< 0.001
CPB time (min)	160.50 (115.25–223.00)	160.00 (119.00–214.00)	0.917	158.00 (109.50–213.00)	170.50 (134.50–222.00)	0.268
ACC time (min)	112.00 (80.00–150.00)	96.50 (73.00–136.75)	0.217	97.00 (76.00–133.00)	104.00 (80.00–142.00)	0.664
ICU stay (days)	2.00 (1.00–3.75)	2.00 (1.00–3.00)	0.453	1.00 (1.00–3.00)	2.00 (1.00–4.00)	0.107
Hospital stays (days)	11.00 (9.00–18.00)	11.00 (8.00–17.75)	0.714	10.00 (8.50–15.50)	11.00 (8.00–18.00)	0.826
Sternal wound infection	1 (1.0%)	1 (1.0%)	1.000	1 (1.3%)	1 (1.3%)	1.000

CPB, cardiopulmonary bypass; ACC, aortic cross clamp; ICU, intensive care unit

Continuous variables are presented as the median (inter-quartile range), and P-values were computed using a Wilcoxon rank-sum test for data with a non-normal distribution. Categorical variables are presented as the sample number (%), and P-values were computed using a chi-squared test with Yates’ correction or Fisher’s exact test

Results

A total of 200 consecutive patients were included in this study, 100 in the OT group and 100 in the OA group. The OT group included 100 cases performed over a 2-year period, whereas the OA group included 100 cases performed over a 3-year period, indicating a reduction in annual surgical volume during the OA period. There were statistically significant differences in the baseline characteristics between the groups before PSM. After PSM, 79 paired patients from each group were selected and absolute standardized mean differences were < 0.1 in all matched variables, indicating there were no significant difference in both groups (Table 1). There were more OPCAB in the OA group than in the OT group (37.0% vs. 20.0%, respectively; P = 0.012, SMD = 0.160) before PSM, but there were no significant differences in the type of operation between the groups after PSM (Table 2).

Postoperatively, there were no significant statistically differences in major post-operative adverse events, including 72-hour mortality, 30-day mortality, stroke, new-onset atrial fibrillation, low cardiac output, reoperation for bleeding. And other post-operative outcomes such as AKI, AKI requiring renal replacement, ICU stays, hospital stays, and sternal wound infection were not statistically significant. Cardiopulmonary bypass (CPB) time, aortic cross clamp (ACC) time, and ventilation time were comparable between the OA and OT groups. However, total operative time was longer in the OA group after PSM (300.00 [250.00-350.00] vs. 379.00 [334.00-429.00], P < 0.001).

In patients who underwent isolated OPCAB, the ventilation time was not significantly different and operation time was longer in the OA group than in the OT group after PSM. (359.00 [321.50–409.00] vs. 275.00 [245.00-295.00], respectively, P < 0.001). The number of anastomosis was similar in both groups. In the OA group, there was one case of OPCAB that was converted to conventional CABG, with a CPB time of 178 min and an ACC time of 131 min (Table 4).

Table 4.

Comparison of postoperative outcomes in OPCAB patients

Variable	Before PSM			After PSM
	OT group (n = 20)	OA group (n = 37)	P-value	OT group (n = 20)	OA group (n = 24)	P-value
Mortality < 72 h, n (%)	0 (0.0%)	1 (2.7%)	1.000	0 (0.0%)	0 (0.0%)	1.000
Mortality < 30 days, n (%)	1 (5.0%)	1 (2.7%)	1.000	1 (5.0%)	0 (0.0%)	0.926
Stroke, n (%)	1 (5.0%)	1 (2.7%)	1.000	1 (5.0%)	0 (0.0%)	0.926
New onset atrial fibrillation, n (%)	5 (25.0%)	4 (10.8%)	0.253	5 (25.0%)	2 (8.3%)	0.275
Low cardiac output syndrome, n (%)	1 (5.0%)	3 (8.1%)	1.000	1 (5.0%)	1 (4.2%)	1.000
Reoperation for bleeding, n (%)	0 (0.0%)	0 (0.0%)	1.000	0 (0.0%)	0 (0.0%)	1.000
Acute kidney injury, n (%)	3 (15.0%)	1 (2.7%)	0.119	3 (15.0%)	0 (0%)	0.172
Acute kidney injury requiring dialysis, n (%)	1 (5.0%)	0 (0.0%)	0.764	1 (5.0%)	0 (0.0%)	0.926
Ventilation time (min)	182.50 (152.50–298.75)	300.00 (240.00–360.00)	0.047	182.50 (155.00–297.50)	240.00 (180.00–360.00)	0.246
Operation time (min)	275.00 (242.50–297.50)	349.00 (319.00–404.00)	< 0.001	275.00 (245.00–295.00)	359.00 (321.50–409.00)	< 0.001
CPB time (min)	NA (NA–NA)	*178.00 (178.00–178.00)	NA	NA (NA–NA)	NA (NA–NA)	NA
ACC time (min)	NA (NA–NA)	*131.00 (131.00–131.00)	NA	NA (NA–NA)	NA (NA–NA)	NA
Anastomosis number	3.00 (2.00–3.00)	3.00 (2.00–3.00)	0.650	3.00 (2.00–3.00)	3.00 (2.00–3.00)	0.631
ICU stay (days)	1.00 (1.00–1.00)	1.00 (1.00–2.00)	0.509	1.00 (1.00–1.00)	1.00 (1.00–2.00)	0.296
Hospital stay (days)	9.00 (8.00–10.75)	9.00 (8.00–11.00)	0.899	9.00 (8.00–10.00)	8.50 (8.00–11.00)	0.783
Sternal wound infection	0 (0.0%)	0 (0.0%)	1.000	0 (0.0%)	0 (0.0%)	1.000

CPB, cardiopulmonary bypass; ACC, aortic cross clamp; ICU, intensive care unit

Continuous variables are presented as the median (inter-quartile range), and P-values were computed using a Wilcoxon rank-sum test for data with a non-normal distribution. Categorical variables are presented as the sample number (%), and P-values were computed using Fisher’s exact test

* There was one case of OPCAB that was converted to conventional CABG, with a CPB time of 178 min and an ACC time of 131 min

Table 5 showed the logistic regression analyses for binary postoperative outcomes comparing the OA and OT groups. In both the multivariable adjusted analysis and the propensity score-matched cohort analysis, there were no statistically significant differences between the groups in short-term post-operative outcomes, including mortality, stroke, new onset atrial fibrillations, low cardiac output syndrome, reoperation for bleeding, AKI, and renal replace therapy.

Table 5.

Logistic regression analyses of OA group for binary postoperative outcomes

Variable	Univariate OR (95% CI)	P-value	*Adjusted OR (95% CI)	P-value	PS matched OR (95% CI)	P-value
Mortality < 72 h	1.00 (0.14–7.24)	1.000	3.37 (0.11–99.49)	0.482	1.00 (0.66–16.27)	1.000
Mortality < 30 days	0.65 (0.18–2.39)	0.519	0.77 (0.18–3.35)	0.728	0.79 (0.20–3.06)	0.732
Stroke	0.65 (0.18–2.39)	0.519	0.73 (0.18–2.89)	0.653	0.58 (0.13–2.53)	0.473
New onset atrial fibrillation	0.80 (0.42–1.54)	0.508	0.90 (0.45–1.81)	0.769	0.76 (0.36–1.58)	0.457
Low cardiac output syndrome	0.67 (0.36–1.25)	0.209	0.86 (0.41–1.77)	0.674	1.00 (0.49–2.03)	1.000
Reoperation for bleeding	0.42 (0.13–1.42)	0.162	0.44 (0.12–1.58)	0.208	0.55 (0.15–1.95)	0.354
Acute kidney injury	0.36 (0.13–0.97)	0.044	0.40 (0.14–1.14)	0.087	0.34 (0.12–1.01)	0.053
Acute kidney injury requiring dialysis	0.74 (0.16–3.41)	0.701	1.63 (0.27–9.85)	0.593	0.49 (0.09–2.74)	0.414

CI, confidence interval; OR, odd ration;

*Adjusted by age, sex, body surface area, estimated glomerular filtration rate, euro SCORE II, hypertension, diabetes, pulmonary hypertension, surgical priority, and type of surgical procedure

Discussion

This study showed that short-term post-operative outcomes including mortality, reoperation for bleeding, stroke, low cardiac output, AKI, and ICU and hospital stay duration were not different statistically in the OT and OA groups. Only operation time was statistically longer in OT group. However, these results do not justify a single-surgeon cardiac surgery program, because this study is a small, single-center study and implies the possibility of various issues arising from operating with only one surgeon. As shown in Table 1, the proportion of non-elective (emergency and urgent) cases was 20% in group OT and 13% in group OA. Following the period when the surgeon was left alone, the number of non-elective cases decreased, although this reduction was not statistically significant. It is plausible that the surgeon may have tended to refuse more challenging emergency cases during this time.

Human factors that have an impact on the cardiac surgery field have been well studied. Teamwork and communication were identified as critical human factors during cardiac surgery [4–6]. In ORs with strong teamwork, a reduction in the number of staff may result in fewer communication errors; however, there is a decreased possibility that errors will be found and corrected, resulting in the possibility that errors increase exponentially in an unusual or emergent situation. Human errors are classified as active failures and latent conditions, where active failures are errors by people in direct contact with patients and latent conditions are errors by the management of an organization [7]. In a single surgeon setting, the surgeon fulfills the roles related to reducing latent conditions and avoiding active failures. Because of this, if a single surgeon fails, critical failures, which can result in major adverse outcomes such as mortality, may occur during cardiac surgery.

Operation times were longer in the OA group than in the OT group after PSM. CBP and ACC time were not different between both groups. In cardiac surgery, outcomes such as renal dysfunction, neurological complications, and mortality are more closely associated with CPB time and ACC time. Comparable CPB and ACC times between the OA and OT groups may partly explain the absence of significant differences in postoperative outcomes. The longer operation times in the OA group were likely to be because a single surgeon performed the procedures resulting in requiring additional time. Although postoperative results such as mortality, ICU and hospital stay duration, and ventilation time were statistically different in present study, longer operation time might result in poor outcomes in a larger volume study.

The lack of statistically significant differences in clinical outcomes between the OA and OT groups may be attributable to several factors. One is that changes in case selection and surgical procedure type during the OA period may have influenced the observed outcomes. The reduction in annual surgical volume and the increased proportion of OPCAB and decreased proportion of combined surgery suggest that the case mix may shifted during the OA period. In addition, the proportion of elective surgery is also increased during OA period. So, more complex or emergent cases may have been delayed, referred to other hospitals, or managed through alternative treatment strategies. These unmeasured changes could have contributed to the comparable postoperative outcomes among patients who underwent surgery at our institution. Second, the relatively small sample size of the study populations limited the statistical power to detect differences in rare but clinically important postoperative outcomes, such as mortality, stroke, or reoperation for bleeding. Because these adverse events rarely occur, even clinically meaningful differences may not have reached statistical significance in this study. Therefore, the absence of statistically significant differences should not be interpreted as evidence of true clinical equivalence between the two staffing models. Third, Because the OA period was later than the OT period, temporal improvements in perioperative care, surgical devices, and postoperative management strategies, including ECMO support strategy, may have favorably influenced outcomes in the OA group. In addition, institutional and team experience is accumulated over time. One attending surgeon was continuously involved throughout the entire study period, and by the time the OA period began, this surgeon had gained additional operative experience. Also, anesthesiologists, perfusionists, and other perioperative staff may also have accumulated experience over the study period, which could have contributed to favorable outcomes in the OA group independent of the staffing model.

In South Korea, cardiac surgery patients and cardiac surgeons are concentrated in a limited number of high-volume centers for various structural and geographic reasons. Although approximately 70 hospitals have cardiac surgery departments, many small and medium-volume centers face challenges in maintaining adequate number of surgeon and surgical. Compared with Germany and United Kingdom (UK) many South Korean cardiac surgery programs may operate with relatively low surgical volume [8, 9]. Because surgical volume and number of medical staffs are closely related to clinical outcomes in cardiac surgery field, the distribution of cardiac surgery resources has important implications for quality of care [10–14]. Therefore, the Korean cardiac surgery system may require further discussion regarding the appropriate balance between regional access, minimum institutional volume, and the number of cardiac surgeons needed to sustain safe and effective cardiac surgery.

This study has several limitations. First, this was a single-center retrospective study, which limits causal interpretation and external validity. Despite statistical adjustment such as propensity score matching or multivariable regression analysis, residual confounding from unmeasured variables which may influence the study outcome cannot be excluded. Selection bias related to the limited number of available cardiac surgeons in OA group may have substantially influenced the observed outcomes. Second, as described above, the relatively small study populations and low incidence of some post-operative outcomes limit the ability to detect differences between two study groups. Third, because the comparison was conducted across two different time periods, there is a possibility of time-related changes in surgical experience, technical or pharmacological advancements that may not be fully adjusted. Given these limitations, the findings of this study should be interpreted with caution. The observed outcomes should not be readily generalized to other institutions or healthcare systems, nor should they be interpreted as evidence supporting broad implementation of a single-surgeon cardiac surgery model.

Conclusions

In this propensity score–matched cohort study, no statistically significant differences were observed between the OT and OA group in short-term clinical outcomes, including one-month mortality, stroke, new-onset atrial fibrillation, low cardiac output syndrome, and acute kidney injury. These findings suggest that, in this selected single-center setting, the OA model was not associated with a statistically significant increase in short-term adverse outcomes compared with the OT model. However, given the various limitations of the present study, these findings should be interpreted cautiously and should not be readily generalized prematurely to other institutions or healthcare systems.

Author contributions

JH Hong designed the study and supervised the project and performed statistical analysis. KT Park, SW Kim, JJ LEE and YJ Son contributed to data collection and interpretation of the results. JH Hong and GT park drafted the initial manuscript. GT park critically revised the manuscript. All authors have read and approved the manuscript.

Funding

None.

Data availability

The datasets analyzed in the current study are not publicly available due to institutional and ethical restrictions but are available from the corresponding author on reasonable request.

Declarations

Ethics approval and consent to participate

This study was approved by the institutional review board of Chung-Ang University Hospital (IRB No. 2103-011-19360). The study was conducted in accordance with the principles of the Declaration of Helsinki. Due to the retrospective nature of the study, the requirement for informed consent was waived by the Institutional Review Board.

Competing interests

The authors declare no competing interests.