Abstract
Background
The purpose of this study was to investigate the association between weekday of surgery and survival following cardiac surgery.
Methods and Results
In a nationwide cohort study, we included all patients who underwent cardiac surgery in 1999–2013 from the SWEDEHEART (Swedish Web‐system for Enhancement and Development of Evidence‐based care in Heart disease Evaluated According to Recommended Therapies) register. All‐cause mortality until March 2014 was obtained from national registers. The association between weekday of surgery and mortality was estimated using Cox regression, and reported as hazard ratios with 95% CI. We used the restricted mean survival time difference to estimate loss of life related to weekday of surgery. Among 106 473 patients, 25 221 (24%), 24 471 (23%), 22 977 (22%), 20 189 (19%), 9251 (8.7%), and 4364 (4.1%) underwent surgery during a Monday, Tuesday, Wednesday, Thursday, Friday, and a Saturday/Sunday, respectively. More patients were operated on urgently during Friday to Sunday, and unadjusted analyses showed higher early and late mortality in those patients. The adjusted hazard ratios (95% CI) were 1.00 (0.89–1.13), 1.00 (0.88–1.12), 1.02 (0.90–1.16), 1.17 (1.01–1.37), and 1.05 (0.86–1.29) in patients who underwent surgery during a Tuesday, Wednesday, Thursday, Friday, and Saturday/Sunday compared to a Monday, after 1 year of follow‐up conditional on 30‐day survival. In elective surgery (n=46 146), the 1‐year restricted mean survival time difference (95% CI) was −0.5 (−1.8–0.8), −0.5 (−1.9–0.8), −1.0 (−2.6–0.5), 0.02 (−2.2–2.3), and −1.2 (−6.3–3.9) days in patients who underwent surgery during a Tuesday, Wednesday, Thursday, Friday, and a Saturday/Sunday, respectively, compared to a Monday.
Conclusions
We found no evidence of a clinically relevant weekday effect in patents who underwent cardiac surgery in Sweden during a 15‐year period. These data suggest that the early risk and long‐term prognosis following cardiac surgery was not affected by the weekday of surgery.
Clinical Trial Registration
URL: http://www.clinicaltrials.gov. Unique identifier: NCT02276950.
Keywords: cardiac surgery, long‐term outcome, quality of care, risk factors, weekday effect
Subject Categories: Cardiovascular Surgery, Mortality/Survival, Quality and Outcomes, Epidemiology
Clinical Perspective
What Is New?
Prior studies have suggested that weekday of surgery influences survival after major surgery, but whether this “weekday” effect is relevant in cardiac surgery is mostly unknown.
We found no evidence of a clinically relevant weekday effect in patients who underwent cardiac surgery in Sweden, and our findings suggest that the early risk and long‐term prognosis following cardiac surgery were not affected by the weekday of surgery.
What Are the Clinical Implications?
Cardiac surgery can safely be performed on any day of the week without an increase in the risk of short‐ or long‐term mortality.
Introduction
Studies have suggested that weekday of surgery influences survival after major surgery such that patients who undergoe surgery late in the week have worse survival compared with those who had surgery early in the week.1, 2, 3 It has been speculated that this association is driven by poorer quality of care and intensity of specialist staffing during the weekend.1, 3, 4, 5 Others have suggested surgeon fatigue with subsequent impaired surgical precision later in the week as a possible explanation in particularly demanding procedures such as esophageal cancer surgery.2 Therefore, we hypothesized that patients who were operated on later in the work week had worse survival compared with those who were operated on early in the week.
Although most reports showed higher short‐term mortality in patients who underwent surgery late compared with early in the week,1, 3, 6, 7 results were conflicting.8, 9 Notably, previous studies investigated only short‐term mortality in patients who underwent different types of operations. Very few studies have examined the impact of weekday of surgery on long‐term mortality.2, 9 One study found no association between day of the work week and in‐hospital death or major short‐term morbidity in patients who underwent elective isolated coronary artery bypass graft surgery.10 No previous study known to us has examined possible weekday‐related variation in long‐term outcomes in cardiac surgery patients.
Therefore, we included all adults who underwent cardiac surgery in Sweden during a 15‐year period in a nationwide population‐based cohort study to investigate whether outcomes after cardiac surgery are associated with weekday of surgery.
Methods
This was a nationwide population‐based observational cohort study that was approved by the regional Human Research Ethics Committee, Stockholm, Sweden (Approval number: 2016/1241‐32), and the need for informed consent was waived by the committee.
Study Population and Data Sources
We identified all patients who underwent cardiac surgery in Sweden from the SWEDEHEART (Swedish Web‐system for Enhancement and Development of Evidence‐based care in Heart disease Evaluated According to Recommended Therapies) register.11, 12 Using unique personal identity numbers, which are assigned to all residents of Sweden,13 the data from SWEDEHEART were linked with data from other national healthcare registers as previously described.14, 15 The National Patient Register16 was used to determine relevant medical history, and the Longitudinal integration database for health insurance and labor market studies,17 managed by Statistics Sweden, was used to obtain details regarding educational level, household disposable income, country of birth, and marital status. The Swedish National Board of Health and Welfare performed the record linkages. The International Classification of Diseases codes used for extraction of medical history and coexisting diagnoses are listed in Table S1.
Weekday and Public Holidays
Patients were assigned to a weekday category corresponding to the date of surgery. In Sweden, ≈17 days per year are established public holidays that are practically equivalent to Sundays. Some public holidays always occur on a Sunday (eg, Easter Sunday), and others (eg, New Year's Day) can occur on any day of the week, depending on the year. For the purpose of this study, we categorized all operations that were performed on a public holiday as a surgery performed on a Sunday. With very few exceptions, only nonelective surgery is performed during weekends (from Friday afternoon to Monday morning).
Outcome Measure
The main outcome measure was time to death from any cause. The Cause of Death Register was used to obtain dates of death.
Statistical Analyses
We categorized patients into groups according to weekday of surgery, with surgery performed on Saturday/Sunday combined with surgery performed on public holidays. Baseline characteristics were described with frequencies and percentages for categorical variables and means and standard deviations for continuous variables. Person–time in days was counted from the date of surgery until the date of death or end of follow‐up (March 24, 2014). We reported crude incidence rates and 95% CI, and the Kaplan–Meier method was used to calculate cumulative survival. We used Cox proportional hazards regression with and without multivariable adjustment to estimate hazard ratios (HR) and 95% CI for the association between weekday categories and survival with varying follow‐up times using patients having surgery on a Monday as the reference category. All Cox regression models were stratified by calendar year of surgery and hospital. Patient age was modeled using restricted cubic splines, and all other variables were included as categorical terms. We reported the restricted mean survival time18 by weekday at 1 and 3 years as absolute measures of survival time, and calculated the difference in restricted mean survival time with 95% CI. The difference in restricted mean survival time is a model‐free measure that is useful in order to quantify the contrast between 2 survival curves, and can from a clinical perspective be interpreted as loss of life expectancy.
Missing data (left ventricular ejection fraction [31%], emergent operation [30%], body mass index [12%], renal function [11%], educational level [5.7%], waiting time ≤10 days [4.9%], and disposable income [0.4%]) were handled by multiple imputation by chained equations.19 The imputation models included all variables in Table 1, year of surgery, hospital, and also the event indicator and the Nelson‐Aalen estimator of the cumulative baseline hazard.20 Ten data sets were imputed and estimates from these data sets were combined according to Rubin's rules. We also repeated the analyses using a missing value indicator category, and finally we performed a complete case analysis.
Table 1.
Baseline Characteristics in 106 473 Patients Who Underwent Cardiac Surgery in Sweden During 1999 to 2013 According to Weekday of Surgery
| Total Population | Monday | Tuesday | Wednesday | Thursday | Friday | Saturday–Sunday | |
|---|---|---|---|---|---|---|---|
| N (%) | 106 473 (100%) | 25 221 (24%) | 24 471 (23%) | 22 977 (22%) | 20 189 (19%) | 9251 (8.7%) | 4364 (4.1%) |
| Age, y—mean, SD | 66.7 (10.9) | 67.1 (10.5) | 66.9 (10.9) | 66.8 (11.1) | 66.5 (11.2) | 66.2 (10.8) | 65.4 (11.6) |
| Sex | |||||||
| Men | 77 806 (73%) | 18 482 (73%) | 17 776 (73%) | 16 633 (72%) | 14 785 (73%) | 6955 (75%) | 3175 (73%) |
| Women | 28 667 (27%) | 6739 (27%) | 6695 (27%) | 6344 (28%) | 5404 (27%) | 2296 (25%) | 1189 (27%) |
| Birth region | |||||||
| Nordic countries | 97 625 (92%) | 23 263 (92%) | 22 363 (91%) | 21 057 (92%) | 18 457 (91%) | 8493 (92%) | 3992 (91%) |
| Other | 8848 (8%) | 1958 (8%) | 2108 (9%) | 1920 (8%) | 1732 (9%) | 758 (8%) | 372 (9%) |
| Education | |||||||
| <10 y | 44 162 (44%) | 10 505 (44%) | 10 133 (44%) | 9577 (44%) | 8385 (44%) | 3853 (44%) | 1709 (41%) |
| 10 to 12 y | 38 386 (38%) | 9059 (38%) | 8784 (38%) | 8268 (38%) | 7295 (38%) | 3331 (38%) | 1649 (40%) |
| >12 y | 17 870 (18%) | 4287 (18%) | 4116 (18%) | 3831 (18%) | 3346 (18%) | 1529 (18%) | 761 (18%) |
| Civil status | |||||||
| Married | 67 807 (64%) | 16 072 (64%) | 15 688 (64%) | 14 621 (64%) | 12 864 (64%) | 5925 (64%) | 2637 (60%) |
| Other | 38 666 (36%) | 9149 (36%) | 8783 (36%) | 8356 (36%) | 7325 (36%) | 3326 (36%) | 1727 (40%) |
| Household disposable income | |||||||
| Quartile 1 (lowest) | 26 512 (25%) | 6167 (25%) | 6027 (25%) | 5815 (25%) | 5028 (25%) | 2355 (26%) | 1120 (26%) |
| Quartile 2 | 26 498 (25%) | 6367 (25%) | 6141 (25%) | 5691 (25%) | 5063 (25%) | 2222 (24%) | 1014 (23%) |
| Quartile 3 | 26 505 (25%) | 6205 (25%) | 6151 (25%) | 5655 (25%) | 5110 (25%) | 2326 (25%) | 1058 (24%) |
| Quartile 4 (highest) | 26 505 (25%) | 6398 (25%) | 6061 (25%) | 5724 (25%) | 4891 (24%) | 2294 (25%) | 1137 (26%) |
| Body mass index, kg/m2—mean, SD | 26.9 (4.2) | 26.9 (4.1) | 26.9 (4.2) | 26.9 (4.2) | 26.9 (4.2) | 27.0 (4.2) | 26.6 (4.2) |
| Diabetes mellitus | 20 883 (20%) | 4980 (20%) | 4795 (20%) | 4474 (19%) | 4025 (20%) | 1805 (20%) | 804 (18%) |
| Hypertension | 32 462 (30%) | 7808 (31%) | 7501 (31%) | 6984 (30%) | 6118 (30%) | 2818 (30%) | 1233 (28%) |
| Hyperlipidemia | 19 079 (18%) | 4641 (18%) | 4350 (18%) | 4104 (18%) | 3541 (18%) | 1797 (19%) | 646 (15%) |
| Peripheral vascular disease | 10 398 (10%) | 2371 (9%) | 2392 (10%) | 2404 (10%) | 1965 (10%) | 802 (9%) | 464 (11%) |
| eGFR, mL/min per 1.73 m2 | |||||||
| >60 | 70 314 (74%) | 16 788 (74%) | 16 228 (74%) | 15 131 (74%) | 13 351 (74%) | 6186 (75%) | 2630 (73%) |
| 45 to 60 | 16 234 (17%) | 3953 (17%) | 3724 (17%) | 3536 (17%) | 3105 (17%) | 1346 (16%) | 570 (16%) |
| 30 to 45 | 6104 (6%) | 1421 (6%) | 1444 (7%) | 1352 (7%) | 1110 (6%) | 502 (6%) | 275 (8%) |
| 15 to 30 | 1255 (1%) | 305 (1%) | 275 (1%) | 261 (1%) | 221 (1%) | 125 (2%) | 68 (2%) |
| <15a | 976 (1%) | 176 (1%) | 249 (1%) | 217 (1%) | 206 (1%) | 69 (1%) | 59 (2%) |
| Chronic pulmonary disease | 7357 (7%) | 1762 (7%) | 1692 (7%) | 1579 (7%) | 1394 (7%) | 605 (7%) | 325 (7%) |
| Prior myocardial infarction | 41 405 (39%) | 9544 (38%) | 9083 (37%) | 8553 (37%) | 7818 (39%) | 4235 (46%) | 2172 (50%) |
| Prior PCI | 12 594 (12%) | 2993 (12%) | 2861 (12%) | 2697 (12%) | 2409 (12%) | 1085 (12%) | 549 (13%) |
| Heart failure | 16 566 (16%) | 3794 (15%) | 3874 (16%) | 3750 (16%) | 3262 (16%) | 1319 (14%) | 567 (13%) |
| LV ejection fraction, % | |||||||
| >50 | 51 057 (69%) | 12 448 (70%) | 11 806 (69%) | 11 034 (70%) | 9700 (69%) | 4048 (67%) | 2021 (65%) |
| 30 to 50 | 18 486 (25%) | 4415 (25%) | 4254 (25%) | 3929 (25%) | 3515 (25%) | 1576 (26%) | 797 (26%) |
| <30 | 4291 (6%) | 905 (5%) | 931 (5%) | 904 (6%) | 861 (6%) | 397 (7%) | 293 (9%) |
| Stroke | 9208 (9%) | 2185 (9%) | 2085 (9%) | 2004 (9%) | 1731 (9%) | 801 (9%) | 402 (9%) |
| Atrial fibrillation | 13 203 (12%) | 3187 (13%) | 3268 (13%) | 2932 (13%) | 2549 (13%) | 877 (9%) | 390 (9%) |
| Cancer | 6491 (6%) | 1536 (6%) | 1469 (6%) | 1467 (6%) | 1211 (6%) | 538 (6%) | 270 (6%) |
| Alcohol dependency | 2116 (2%) | 495 (2%) | 477 (2%) | 441 (2%) | 397 (2%) | 205 (2%) | 101 (2%) |
| Emergent operation | 4994 (7%) | 619 (3%) | 709 (4%) | 732 (5%) | 790 (6%) | 790 (13%) | 1354 (43%) |
| Waiting time ≤10 d | 35 109 (35%) | 6689 (28%) | 6168 (26%) | 6813 (31%) | 7378 (38%) | 4836 (55%) | 3225 (81%) |
| Operation | |||||||
| Isolated CABG | 66 319 (62%) | 15 880 (63%) | 14 821 (61%) | 13 773 (60%) | 12 324 (61%) | 6611 (71%) | 2910 (67%) |
| Isolated valve | 13 557 (13%) | 3413 (14%) | 3428 (14%) | 3082 (13%) | 2551 (13%) | 866 (9%) | 217 (5%) |
| Valve+CABG | 16 857 (16%) | 4146 (16%) | 4018 (16%) | 3921 (17%) | 3371 (17%) | 1012 (11%) | 389 (9%) |
| Other | 9740 (9%) | 1782 (7%) | 2204 (9%) | 2201 (10%) | 1943 (10%) | 762 (8%) | 848 (19%) |
Data are n (%) unless otherwise noted. CABG indicates coronary artery bypass grafting; eGFR, estimated glomerular filtration rate; LV, left ventricular; PCI, percutaneous coronary intervention.
This category included patients on dialysis.
Data management and statistical analyses were performed using Stata 14.2 (Stata Corp, College Station, TX) and R version 3.3.1 (R Foundation for Statistical Computing, Vienna, Austria).
Results
Study Population and Patient Characteristics
Patient characteristics according to weekday category are presented in Table 1. The number of patients operated on each day of the week is shown in Figure 1. A total of 106 473 patients with a mean age of 67 years were included. Of these, 27% were women. The majority (87%) of patients underwent surgery during Monday to Thursday. Patients were categorized according to weekday of surgery: 25 221 (24%), 24 471 (23%), 22 977 (22%), 20 189 (19%), 9251 (8.7%), and 4364 (4.1%) underwent surgery during a Monday, Tuesday, Wednesday, Thursday, Friday, and a Saturday/Sunday, respectively. Patient characteristics were similar across groups with the exception that in relative terms, semi‐urgent (waiting time 10 days or less) and emergent surgery were more common during Friday–Sunday (Figure 1).
Figure 1.
Number of operations per weekday. Number of heart operations performed per weekday in Sweden during 1999 to 2013.
Follow‐Up and Mortality
During a median follow‐up time of 6.8 years, a total of 29 806 (28%) patients died: 27%, 28%, 28%, 28%, 28%, and 32% of patients who were operated during a Monday, Tuesday, Wednesday, Thursday, Friday, and Saturday/Sunday, respectively. The incidence rate, and crude and adjusted relative risks for all‐cause mortality are shown in Table 2. The crude HR increased toward the end of the week, although this effect was attenuated with increasing length of follow‐up (Figure 2). After multivariable adjustment, only operations carried out on Saturday/Sunday were significantly associated with early mortality (follow‐up until 30 days); HR (95% CI) 1.53 (1.32–1.77). For patients who had surgery on Fridays, there was a significantly increased risk of death; adjusted HR (95% CI): 1.17 (1.01–1.37), after 1 year of follow‐up conditional on 30‐day survival, but this was attenuated after restricting the analytical sample to elective patients operated on during Monday to Friday; adjusted HR (95% CI): 1.08 (0.80–1.46). There was no association between weekday of surgery and mortality after multivariable adjustment in patients followed beyond 1 year and until 15 years (Figure 2).
Table 2.
Event Rates and Relative Risks for All‐Cause Mortality Following Cardiac Surgery in Sweden During 1999 to 2013 According to Weekday of Surgery
| Number of Deaths/Person–Y | Unadjusted Mortality Rate Per 1000 Person–Y (95% CI) | Crude HR (95% CI) | Multivariablea Adjusted HR (95% CI) | |
|---|---|---|---|---|
| 0 to 30 d follow‐up | ||||
| Monday | 593/2038 | 291 (268–315) | 1.00 (ref) | 1.00 (ref) |
| Tuesday | 637/1974 | 323 (299–349) | 1.11 (0.99–1.24) | 1.07 (0.95–1.19) |
| Wednesday | 598/1853 | 323 (298–350) | 1.11 (0.99–1.25) | 0.99 (0.88–1.11) |
| Thursday | 567/1626 | 349 (321–379) | 1.21 (1.08–1.36) | 1.01 (0.90–1.13) |
| Friday | 289/744 | 389 (346–436) | 1.36 (1.18–1.57) | 0.98 (0.84–1.13) |
| Saturday/Sunday | 355/337 | 1053 (949–1169) | 3.62 (3.17–4.14) | 1.53 (1.32–1.77) |
| 30 d to 1 y follow‐up | ||||
| Monday | 559/21 885 | 26 (24–28) | 1.00 (ref) | 1.00 (ref) |
| Tuesday | 558/21 229 | 26 (24–29) | 1.03 (0.92–1.16) | 1.00 (0.89–1.13) |
| Wednesday | 542/19 922 | 27 (25–30) | 1.07 (0.95–1.20) | 1.00 (0.88–1.12) |
| Thursday | 492/17 452 | 28 (26–31) | 1.11 (0.98–1.25) | 1.02 (0.90–1.16) |
| Friday | 257/7917 | 33 (29–37) | 1.30 (1.12–1.51) | 1.17 (1.01–1.37) |
| Saturday/Sunday | 130/3538 | 37 (31–44) | 1.42 (1.17–1.72) | 1.05 (0.86–1.29) |
| 1 to 15 y follow‐up | ||||
| Monday | 5628/149 442 | 38 (37–39) | 1.00 (ref) | 1.00 (ref) |
| Tuesday | 5633/146 465 | 39 (38–40) | 1.02 (0.98–1.06) | 1.02 (0.98–1.06) |
| Wednesday | 5374/138 946 | 39 (38–40) | 1.02 (0.99–1.06) | 1.01 (0.97–1.05) |
| Thursday | 4635/121 034 | 38 (37–39) | 1.01 (0.97–1.05) | 1.02 (0.98–1.06) |
| Friday | 2043/57 018 | 36 (34–37) | 0.94 (0.89–0.99) | 1.00 (0.95–1.06) |
| Saturday/Sunday | 908/23 476 | 39 (36–41) | 1.03 (0.96–1.10) | 0.98 (0.91–1.06) |
HR indicates hazard ratio; ref, reference category.
Model included all variables reported in Table 1.
Figure 2.
Risk of mortality according to weekday. The crude and multivariable adjusted hazard ratios and 95% CIs for all‐cause mortality following cardiac surgery in Sweden during 1999 to 2013 according to weekday of surgery. Monday was used as the reference category and all variables reported in Table 1 were included in the multivariable adjusted model. For presentational purposes, the hazard ratios and 95% CIs for Saturday/Sunday were omitted from the Figure but are shown in Table 2.
The adjusted 1‐year cumulative mortality (95% CI) was 1.8 (1.7–2.0), 1.9 (1.7–2.1), 1.9 (1.7–2.1), 1.9 (1.7–2.0), 1.9 (1.7–2.1), and 1.9 (1.7–2.2) in patients who were operated on during a Monday, Tuesday, Wednesday, Thursday, Friday, and Saturday/Sunday, respectively (Figure 3).
Figure 3.
Cumulative mortality according to weekday. Crude (upper panel) and multivariable adjusted (lower panel) cumulative mortality following cardiac surgery in Sweden during 1999 to 2013 according to weekday of surgery.
Patient characteristics were very similar across weekdays, but emergent surgery was more common during Friday–Sunday (Table 1 and Figure 1). The reduction in HR observed in the adjusted compared with the crude analyses was mainly driven by urgency status, and the addition of a wealth of other clinical parameters to the regression model had only a negligible effect on the point estimates. A similar pattern was seen in the restricted mean survival time analyses, where we found no significant differences in the restricted mean survival time between weekdays among patients who underwent elective surgery (Table 3). The difference in survival time ranged from −1.2 days to 0.02 days at 1 year, and −4.7 days to −3.2 days at 3 years, and was not statistically significant at any point.
Table 3.
Restricted Mean Survival Time (Days) and Difference in Restricted Mean Survival Time (Days) at 1 and 3 Years Postoperatively in Patients Who Underwent Elective Surgery (n=46 146) According to Weekday of Surgery
| Mean Survival Time, d (95% CI) | Difference in Survival Time vs Monday, d (95% CI) | P Value | |
|---|---|---|---|
| Follow‐up at 1 y | |||
| Monday | 355 (354–356) | … | … |
| Tuesday | 355 (354–356) | −0.5 (−1.8 to 0.8) | 0.48 |
| Wednesday | 355 (354–356) | −0.5 (−1.9 to 0.8) | 0.45 |
| Thursday | 354 (353–355) | −1.0 (−2.6 to 0.5) | 0.18 |
| Friday | 355 (353–357) | 0.02 (−2.2 to 2.3) | 0.99 |
| Saturday/Sunday | 354 (349–359) | −1.2 (−6.3 to 3.9) | 0.65 |
| Follow‐up at 3 y | |||
| Monday | 1046 (1042–1049) | … | … |
| Tuesday | 1042 (1038–1045) | −4.3 (−9.3 to 0.8) | 0.10 |
| Wednesday | 1043 (1039–1047) | −3.2 (−8.4 to 2.0) | 0.23 |
| Thursday | 1041 (1037–1046) | −4.7 (−10 to 1.0) | 0.11 |
| Friday | 1042 (1034–1050) | −4.0 (−13 to 4.7) | 0.37 |
| Saturday/Sunday | 1042 (1024–1061) | −3.6 (−23 to 15) | 0.71 |
Sensitivity Analyses
We repeated the analyses restricted to patients operated on between 2005 and 2013 to evaluate possible secular trends (Table S2). We also repeated the analyses restricted to patients who underwent elective isolated coronary artery bypass graft to investigate the association between weekday of surgery and survival in a low‐risk subset of patients who underwent a single procedure (Table S3). In addition to the main analyses in the imputed data sets, we carried out a complete case analysis (n=61 425) (Table S4). We also applied an alternative missing data strategy in which we assigned missing values an indicator category. All sensitivity analyses gave results that were similar to the main analyses.
Discussion
In this nationwide cohort study, we found no evidence of a clinically important weekday or weekend effect in patents who underwent cardiac surgery in Sweden during a 15‐year period. Patients operated on during weekends had increased crude postoperative mortality compared with patients undergoing surgery during weekdays, but this risk increase did not persist when we accounted for the urgency of the procedure. Importantly, addition of other clinical parameters did not affect estimates, indicating that residual confounding should be unlikely and that results are likely to be robust. The difference in loss of life expectancy at 3 years in patients who underwent elective surgery during any weekday compared with Monday was negligible.
Most previous research regarding weekday admission and mortality has focused on acute admissions and the results were contradictory; some studies linked weekend admissions to higher mortality,21, 22 and some did not.23 Very few studies have investigated the possible weekday effect of surgical procedures on adverse outcomes.1, 2, 3 A study of major nonemergent, noncardiac surgery using the Veterans Affairs' database found that 30‐day mortality was increased in patients admitted to non–intensive care unit wards after undergoing surgery on Fridays compared with earlier weekdays,1 although the observed risk was significantly reduced after adjusting for known patient characteristics. Another study that focused on elective surgical procedures in English hospitals found a higher risk of death for patients who underwent surgical procedures during weekends compared with Mondays.3 However, this study was based on hospital administrative data without detailed information regarding individual patient characteristics. A recent Swedish population‐based cohort study included 1748 patients who underwent elective esophagectomy for esophageal cancer, and investigated whether weekday of surgery influenced long‐term all‐cause and disease‐specific mortality.2 That study was strengthened by the nationwide design, large study population, complete follow‐up, and the detailed adjustment for prognostic factors. The authors found better all‐cause and disease‐specific 5‐year survival in patients who had surgery on Monday or Tuesday compared with those who underwent surgery during Wednesday, Thursday, or Friday.2 The authors discussed a possible mechanism that might explain their results in this particular setting; surgery for esophageal cancer is a demanding procedure (mean operating time 6.5 hours) and the alertness and surgical precision of the surgeon, and the surgical team, may be impaired later in the week.2
The reports showing higher mortality in patients admitted or operated on during weekends have discussed 2 potential explanations. First, it has been speculated that this association was mediated through poorer quality of care during the weekend, possibly because of reduced or less experienced weekend staffing.1, 3 Second, it seems clear that patients undergoing surgery or having acute admissions during weekends were more severely ill compared with patients admitted during weekdays.4 Thus, it is important that studies investigating a possible weekday or weekend effect are designed to be able to account for patient risk profile. This was recently illustrated in the setting of stroke care. One study reported lower rates of treatments and worse outcomes in patients admitted for stroke during weekends compared with weekdays.24 However, another study using a specialist clinical database that provided the possibility to adjust for the stroke severity on admission showed no difference in 30‐day mortality related to day of admission.25
The majority of studies in relation to surgical procedures or acute admissions investigating this topic have been conducted using hospital administrative data without detailed information about patient comorbidity or operative risk.3, 4, 5, 24 Therefore, our study extends the findings from previous research because the role of risk factor profile, urgency, and type of procedure could be accounted for in detail through access to SWEDEHEART and other national Swedish health‐data registers.
In Sweden, all patients benefit from universal tax‐financed health care, irrespective of employment status and individual financial situation.15 Insurance status or availability did not influence the indication for surgery. During the study period, cardiac surgery was performed with very similar and stable results in 8 hospitals in Sweden.26 The staffing levels, working hours, access to diagnostics, and physiotherapy were similar between centers.
In the present study, the significant association between weekday of surgery and mortality found in the unadjusted analyses was largely explained and eliminated by accounting for the urgency of the procedure. The patient characteristics were very similar across weekdays with the exception that urgent surgery was more common during Friday and during the weekend. Moreover, the loss of life expectancy analyses supported the finding of a trivial difference between weekdays after restricting the analyses to patients who underwent elective surgery. Our results were in line with a previous letter to the editor reporting that weekday of surgery was not associated with early mortality following cardiac surgery in England and Wales after risk factor adjustment.27 Interestingly, in the study by Aylin et al, early mortality in the subset of patients who underwent coronary artery bypass graft was not affected by weekday of surgery.3
Study Strengths and Limitations
The strengths of our study included the large study population and the complete and accurate follow‐up and survival ascertainment because of the high‐quality national Swedish registries, as well as the external validity and generalizability of our findings because of the nationwide complete coverage. Furthermore, we had detailed information regarding patient characteristics, including socioeconomic data, which allowed multivariable adjustment to address confounding. However, there are some important limitations of the study. First, we only had information regarding which weekday a patient underwent surgery, and not the exact time of day. Therefore, we were unable to investigate the association between time of day and outcomes, and also this fact entails a risk for misclassification of exposure because patients assigned to a specific weekday were in fact operated on during the night between 2 weekdays. Second, although we adjusted for a considerable number of established risk factors and risk markers, the possibility remains that residual confounding could have influenced the results. Because this nationwide study was conducted within the Swedish healthcare system, it is unclear whether our findings are valid in other countries where the provision and organization of cardiac surgery may be different.
Conclusions
We found no evidence for a clinically relevant association between weekday of surgery and early or long‐term mortality after cardiac surgery. Patients operated on during weekends had increased crude postoperative mortality, but this risk increase was attenuated after accounting for the urgency of the procedure. The difference in loss of life expectancy at 3 years in patients who underwent elective surgery during any weekday compared with Monday was negligible. These data suggest that early risk and long‐term prognosis after cardiac surgery are not affected by the day of surgery per se.
Sources of Funding
This study was supported by grants from the Swedish Society of Medicine (Dr Sartipy and Dr Holzmann), Karolinska Institutet Foundations and Funds (Dr Sartipy), the Mats Kleberg Foundation (Dr Sartipy), the Swedish Society for Medical Research (Dr Edgren), the Strategic research program in Epidemiology at Karolinska Institutet (Dr Edgren), and the Swedish Heart‐Lung Foundation (Dr Sartipy and Dr Holzmann).
Disclosures
None.
Supporting information
Table S1.
Table S2. Sensitivity Analysis—Recent Era (2005–2013)
Table S3. Sensitivity Analysis—Elective Isolated CABG
Table S4. Sensitivity Analysis—Complete Case Analysis
Acknowledgments
The authors are grateful to the steering committee of SWEDEHEART for providing data for this study.
(J Am Heart Assoc. 2017;6:e005908 DOI: 10.1161/JAHA.117.005908.)28512116
References
- 1. Zare MM, Itani KM, Schifftner TL, Henderson WG, Khuri SF. Mortality after nonemergent major surgery performed on Friday versus Monday through Wednesday. Ann Surg. 2007;246:866–874. [DOI] [PubMed] [Google Scholar]
- 2. Lagergren J, Mattsson F, Lagergren P. Weekday of esophageal cancer surgery and its relation to prognosis. Ann Surg. 2016;263:1133–1137. [DOI] [PubMed] [Google Scholar]
- 3. Aylin P, Alexandrescu R, Jen MH, Mayer EK, Bottle A. Day of week of procedure and 30 day mortality for elective surgery: retrospective analysis of hospital episode statistics. BMJ. 2013;346:f2424. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4. Black N. Is hospital mortality higher at weekends? If so, why? Lancet. 2016;388:108–111. [DOI] [PubMed] [Google Scholar]
- 5. Aldridge C, Bion J, Boyal A, Chen YF, Clancy M, Evans T, Girling A, Lord J, Mannion R, Rees P, Roseveare C, Rudge G, Sun J, Tarrant C, Temple M, Watson S, Lilford R, Hi SC. Weekend specialist intensity and admission mortality in acute hospital trusts in England: a cross‐sectional study. Lancet. 2016;388:178–186. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6. Singla AA, Guy GS, Field JB, Ma N, Babidge WJ, Maddern GJ. No weak days? Impact of day in the week on surgical mortality. ANZ J Surg. 2016;86:15–20. [DOI] [PubMed] [Google Scholar]
- 7. Ruiz M, Bottle A, Aylin PP. Exploring the impact of consultants' experience on hospital mortality by day of the week: a retrospective analysis of hospital episode statistics. BMJ Qual Saf. 2016;25:337–344. [DOI] [PubMed] [Google Scholar]
- 8. Lagergren J, Mattsson F, Lagergren P. Weekday of oesophageal cancer surgery in relation to early postoperative outcomes in a nationwide Swedish cohort study. BMJ Open. 2016;6:e011097. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9. Baid‐Agrawal S, Martus P, Feldman H, Kramer H. Weekend versus weekday transplant surgery and outcomes after kidney transplantation in the USA: a retrospective national database analysis. BMJ Open. 2016;6:e010482. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10. Tan PJ, Xu M, Sessler DI, Bashour CA. Operation timing does not affect outcome after coronary artery bypass graft surgery. Anesthesiology. 2009;111:785–789. [DOI] [PubMed] [Google Scholar]
- 11. Jernberg T, Attebring MF, Hambraeus K, Ivert T, James S, Jeppsson A, Lagerqvist B, Lindahl B, Stenestrand U, Wallentin L. The Swedish Web‐system for Enhancement and Development of Evidence‐based care in Heart Disease Evaluated According to Recommended Therapies (SWEDEHEART). Heart. 2010;96:1617–1621. [DOI] [PubMed] [Google Scholar]
- 12. Emilsson L, Lindahl B, Koster M, Lambe M, Ludvigsson JF. Review of 103 Swedish healthcare quality registries. J Intern Med. 2015;277:94–136. [DOI] [PubMed] [Google Scholar]
- 13. Ludvigsson JF, Otterblad‐Olausson P, Pettersson BU, Ekbom A. The Swedish personal identity number: possibilities and pitfalls in healthcare and medical research. Eur J Epidemiol. 2009;24:659–667. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14. Dalén M, Ivert T, Holzmann MJ, Sartipy U. Coronary artery bypass grafting in patients 50 years or younger: a Swedish Nationwide Cohort Study. Circulation. 2015;131:1748–1754. [DOI] [PubMed] [Google Scholar]
- 15. Dalén M, Ivert T, Holzmann MJ, Sartipy U. Household disposable income and long‐term survival after cardiac surgery: a Swedish Nationwide Cohort Study in 100 534 patients. J Am Coll Cardiol. 2015;66:1888–1897. [DOI] [PubMed] [Google Scholar]
- 16. Ludvigsson JF, Andersson E, Ekbom A, Feychting M, Kim JL, Reuterwall C, Heurgren M, Olausson PO. External review and validation of the Swedish national inpatient register. BMC Public Health. 2011;11:450. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17. Longitudinal integration database for health insurance and labour market studies (LISA by Swedish acronym). Available at: http://www.scb.se/en_/Services/Guidance-for-researchers-and-universities/SCB-Data/Longitudinal-integration-database-for-health-insurance-and-labour-market-studies-LISA-by-Swedish-acronym. Accessed March 14, 2017.
- 18. Royston P, Parmar MK. Restricted mean survival time: an alternative to the hazard ratio for the design and analysis of randomized trials with a time‐to‐event outcome. BMC Med Res Methodol. 2013;13:152. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19. White IR, Royston P, Wood AM. Multiple imputation using chained equations: issues and guidance for practice. Stat Med. 2011;30:377–399. [DOI] [PubMed] [Google Scholar]
- 20. White IR, Royston P. Imputing missing covariate values for the Cox model. Stat Med. 2009;28:1982–1998. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21. Bell CM, Redelmeier DA. Mortality among patients admitted to hospitals on weekends as compared with weekdays. N Engl J Med. 2001;345:663–668. [DOI] [PubMed] [Google Scholar]
- 22. Cram P, Hillis SL, Barnett M, Rosenthal GE. Effects of weekend admission and hospital teaching status on in‐hospital mortality. Am J Med. 2004;117:151–157. [DOI] [PubMed] [Google Scholar]
- 23. Fonarow GC, Abraham WT, Albert NM, Stough WG, Gheorghiade M, Greenberg BH, O'Connor CM, Nunez E, Yancy CW, Young JB. Day of admission and clinical outcomes for patients hospitalized for heart failure: findings from the Organized Program to Initiate Lifesaving Treatment in Hospitalized Patients With Heart Failure (OPTIMIZE‐HF). Circ Heart Fail. 2008;1:50–57. [DOI] [PubMed] [Google Scholar]
- 24. Palmer WL, Bottle A, Davie C, Vincent CA, Aylin P. Dying for the weekend: a retrospective cohort study on the association between day of hospital presentation and the quality and safety of stroke care. Arch Neurol. 2012;69:1296–1302. [DOI] [PubMed] [Google Scholar]
- 25. Bray BD, Cloud GC, James MA, Hemingway H, Paley L, Stewart K, Tyrrell PJ, Wolfe CD, Rudd AG; Collaboration S . Weekly variation in health‐care quality by day and time of admission: a nationwide, registry‐based, prospective cohort study of acute stroke care. Lancet. 2016;388:170–177. [DOI] [PubMed] [Google Scholar]
- 26. Harnek J, Nilsson J, Friberg Ö, James S, Lagerqvist B, Hambraeus K, Cider A, Svennberg L, Attebring MF, Held C, Johansson P, Jernberg T. The 2011 outcome from the Swedish Health Care Registry on Heart Disease (SWEDEHEART). Scand Cardiovasc J. 2013;47(Suppl 62):1–10. [DOI] [PubMed] [Google Scholar]
- 27. Grant SW, Hickey GL, Taggart DP, Roxburgh J, Cooper G, Bridgewater B. Response to: higher senior staffing levels at weekends and reduced mortality. BMJ. 2012;344:e67. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Table S1.
Table S2. Sensitivity Analysis—Recent Era (2005–2013)
Table S3. Sensitivity Analysis—Elective Isolated CABG
Table S4. Sensitivity Analysis—Complete Case Analysis