After matching, the basic characteristics of patients such as age, sex, CCI, type of insurance, income level, and type of disease showed negligible differences between cohorts (SMD<|0.1|, Table 2). The overall proportion of fusion surgery did not change during the COVID-19 pandemic (p = 0.21). However, there was a slight change in the proportion according to the diagnosis. In HIVD, the proportion of patients who underwent fusion surgery decreased (p < 0.01), while the proportion of patients who underwent endoscopic discectomy increased (p < 0.01). In LSS, the proportion of patients who underwent fusion surgery increased (p < 0.01) (Table 3).
Table 3
Types of surgery. LSS, lumbar spinal stenosis; HIVD, herniated intervertebral disc disease
| Pre-COVID-19 | COVID-19 | p-value |
All patients | n = 75,453 | n = 75,453 | |
Fusion surgery | 18886 (25.03) | 19097 (25.31) | 0.2107 |
Decompression | 56567 (74.97) | 56356 (74.69) | |
Spondylolysis | n = 906 | n = 906 | |
Fusion surgery | 599 (66.11) | 607 (67.00) | 0.6892 |
Decompression | 307 (33.89) | 299 (33.00) | |
Spondylolisthesis | n = 13122 | n = 13122 | |
Fusion surgery | 9395 (71.60) | 9368 (71.39) | 0.7103 |
Decompression | 3727 (28.40) | 3754 (28.61) | |
LSS | n = 50360 | n = 50360 | |
Fusion surgery | 5366 (10.66) | 5912 (11.74) | < .0001 |
Decompression | 44994 (89.34) | 44448 (88.26) | |
HIVD | n = 11065 | n = 11065 | |
Open discectomy | 7473 (67.54) | 7655 (69.18) | 0.0072 |
Endoscopic discectomy | 66 (0.60) | 200 (1.81) | < .0001 |
Fusion surgery | 3526 (31.87) | 3210 (29.01) | < .0001 |
Number of Operations and Monthly Distribution
The total number of operations was 78,420 in the pre-COVID-19 period and 80,165 during the COVID-19 period, indicating a slight increase in surgical procedures during the pandemic (Table 1). Before propensity score matching, the majority of operations took place in primary hospitals, followed by general hospitals, tertiary hospitals, and clinics in both timeframes. After propensity score matching, there was a decrease in the proportion in tertiary hospitals and an increase in general hospitals in the COVID-19 cohort, although the order of prominence remained unchanged (Table 2).
An analysis of the weekly number of operations revealed consistent patterns in both cohorts, showing declines around three major holiday seasons: the Lunar New Year, Chuseok (Korean Thanksgiving), and the Christmas-New Year holidays. Given that the Lunar New Year and Chuseok holidays are determined by the lunar calendar, the specific dates of these holidays differ annually (Fig. 1a). Apart from these holiday seasons, the monthly distribution of surgeries was largely similar between the pre-COVID-19 and COVID-19 cohorts (Table 1). Notably, during the COVID-19 pandemic, there was a decline in surgeries in early March, following the first epidemic wave. This suppressed surgical volume persisted through April, and a compensatory increase in surgeries was observed by May. In August, with the onset of the second epidemic wave, there was a decrease in surgeries, but a compensatory increase occurred more quickly than after the first wave. Despite November witnessing the third and most prominent epidemic wave, there was no significant drop in surgeries or a subsequent compensatory rise. (Fig. 1a)
Joinpoint Regression
By joinpoint regression analysis using a permutation test with Monte Carlo resampling, we identified three key joinpoints: the 9th week in February, the 28th week in July, and the 34th week in August (Fig. 1b). The identified first and second joinpoints were matched with the first and second epidemic waves. The joinpoint plot revealed a marked decline in operations up to the first joinpoint with WPC of -4.70 (p < 0.01). After the first joinpoint, a compensatory increase trend with WPC 1.70 (p < 0.01) was observed, up to the second joinpoint in the 28th week in July. After the second joinpoint, a declining trend emerged with a WPC of -2.55 until the third joinpoint, peak of the second epidemic wave, but this decline was not statistically significant (p = 0.21). After the third joinpoint, the trend shifted towards a consistent increase with a WPC of 0.94 (p < 0.01), without further joinpoints. Remarkably, the period corresponding to the third epidemic wave in November showed neither a decrease nor a compensatory increase (Fig. 1b).
LOH and Hospital Costs
LOH significantly decreased by 1 day or more in patients with spondylolisthesis and HIVD (Table 4). The proportion of LOH exceeding 8 days was 69.2% in the pre-COVID-19 cohort, while it decreased to 65.9% in the COVID-19 cohort (p < 0.01). The mean hospital cost was significantly higher in the COVID-19 cohort than in the pre-COVID-19 cohort, across all diagnoses (from 3,472 to 4,042 USD, p < 0.01) (Table 4).
Table 4
Length of hospital stay and hospital costs during admission. SD, standard deviation; LSS, lumbar spinal stenosis; HIVD, herniated intervertebral disc disease; USD, US dollar * Consumer price index was set at the year 2020 and the conversion of won against dollars was 1086.3.
| Pre-COVID-19 | COVID-19 | p-value |
| mean ± SD | median [Q1, Q3] | mean ± SD | median [Q1, Q3] | |
Length of hospital stay (Day) |
All patients | 13 ± 10 | 11 [8, 16] | 12 ± 8 | 11 [7, 15] | < .0001 |
Spondylolysis | 16 ± 8 | 15 [10, 19] | 15 ± 9 | 14 [10, 19] | 0.3660 |
Spondylolisthesis | 16 ± 13 | 15 [11, 19] | 15 ± 9 | 14 [10, 18] | < .0001 |
LSS | 12 ± 8 | 10 [7, 15] | 11 ± 8 | 10 [6, 14] | < .0001 |
HIVD | 14 ± 12 | 12 [8, 17] | 13 ± 9 | 11 [8, 16] | < .0001 |
Hospital cost (USD)* |
All patients | 3472 ± 2723 | 2666 [1797, 4257] | 4042 ± 2880 | 3139 [2277, 4962] | < .0001 |
Spondylolysis | 4203 ± 2668 | 3517 [2142, 5851] | 5252 ± 2903 | 4869 [3039, 6875] | < .0001 |
Spondylolisthesis | 5221 ± 3151 | 4974 [2752, 6903] | 5993 ± 322315 ± 9 | 5710 [3545, 7517] | < .0001 |
LSS | 2803 ± 2022 | 2368 [1606, 3338] | 3337 ± 2171 | 2808 [2113, 3847] | < .0001 |
HIVD | 4384 ± 3613 | 3161 [2128, 5632] | 4838 ± 3815 | 3504 [2506, 5927] | < .0001 |
Discharge Disposition
In the pre-COVID-19 cohort, 11.1% of patients were transferred to other hospitals after discharge, while this was the case for 7.97% of patients in the COVID-19 cohort (p < 0.01, Table 5). A significantly higher proportion of patients with longer hospital stays were transferred to another hospital than that of patients with hospital days of 7 days or less (p < 0.01 in both cohorts). The most common type of transfer location was a clinic, followed by a primary hospital, convalescent hospital, oriental medicine hospital, general hospital, tertiary hospital, and public health center (Table 5). The order was the same in the pre-COVID-19 cohort, irrespective of LOH. However, in the COVID-19 cohort, the order of clinics and primary hospitals was switched (Table 5).
Table 5
Type of healthcare centers in transfers and 30-day readmissions. LOH, length of hospital stay
| Pre-COVID-19, Number of patients (%) | COVID-19, Number of patients (%) |
| Total operation (n = 75453) | LOH = < 7 (n = 23261, 30.8%) | LOH > = 8 (n = 52192, 69.2%) | p-value | Total operation (n = 75453) | LOH = < 7 (n = 25760, 34.1%) | LOH > = 8 (n = 49693, 65.9%) | p-value |
Transfer to other hospitals | 8383 (11.11) | 1515 (6.51) | 6868 (13.16) | < .0001 | 6017 (7.97) | 1177 (4.57) | 4840 (9.74) | < .0001 |
Type of healthcare center in transfer | n = 8383 | n = 1515 | n = 6868 | < .0001 | n = 6017 | n = 1177 | n = 4840 | < .0001 |
Convalescent hospital | 1703 (20.31) | 104 (6.86) | 1599 (23.28) | | 1074 (17.85) | 81 (6.88) | 993 (20.52) | |
Public health center | 16 (0.19) | 3 (0.20) | 13 (0.19) | | 2 (0.03) | 0 (0.00) | 2 (0.04) | |
Clinic | 2507 (29.91) | 595 (39.27) | 1912 (27.84) | | 1683 (27.97) | 420 (35.68) | 1263 (26.10) | |
Hospital | 2235 (26.66) | 482 (31.82) | 1753 (25.52) | | 1675 (27.84) | 366 (31.10) | 1309 (27.05) | |
General hospital | 862 (10.28) | 182 (12.01) | 680 (9.90) | | 521 (8.66) | 133 (11.30) | 388 (8.02) | |
Tertiary hospital | 41 (0.49) | 17 (1.12) | 24 (0.35) | | 48 (0.80) | 18 (1.53) | 30 (0.62) | |
Oriental medicine hospital | 1019 (12.16) | 132 (8.71) | 887 (12.91) | | 1014 (16.85) | 159 (13.51) | 855 (17.67) | |
30-day readmission | 3110 (4.12) | 735 (3.16) | 2375 (4.55) | < .0001 | 2859 (3.79) | 810 (3.14) | 2049 (4.12) | < .0001 |
The same hospital | 1291 (41.51) | 152 (20.68) | 1139 (47.96) | < .0001 | 1129 (39.49) | 155 (19.14) | 974 (47.54) | < .0001 |
Other hospitals | 1819 (58.49) | 583 (79.32) | 1236 (52.04) | | 1730 (60.51) | 655 (80.86) | 1075 (52.46) | |
Type of healthcare center in readmission | n = 3110 | n = 735 | n = 2375 | < .0001 | n = 2859 | n = 810 | n = 2049 | < .0001 |
Convalescent hospital | 172 (5.53) | 8 (1.09) | 164 (6.91) | | 151 (5.28) | 10 (1.23) | 141 (6.88) | |
Public health center | 4 (0.13) | 0 (0.00) | 4 (0.17) | | 0 (0.00) | 0 (0.00) | 0 (0.00) | |
Clinic | 331 (10.64) | 65 (8.84) | 266 (11.20) | | 318 (11.12) | 72 (8.89) | 246 (12.01) | |
Hospital | 1757 (56.50) | 526 (71.56) | 1231 (51.83) | | 1600 (55.96) | 571 (70.49) | 1029 (50.22) | |
General hospital | 482 (15.50) | 89 (12.11) | 393 (16.55) | | 439 (15.36) | 102 (12.59) | 337 (16.45) | |
Tertiary hospital | 114 (3.67) | 26 (3.54) | 88 (3.71) | | 96 (3.36) | 34 (4.20) | 62 (3.03) | |
Oriental medicine hospital | 250 (8.04) | 21 (2.86) | 229 (9.64) | | 255 (8.92) | 21 (2.59) | 234 (11.42) | |
Thirty-day Readmission and Reoperation
The 30-day readmission rate was 4.12% in the pre-COVID-19 cohort and 3.79% in the COVID-19 cohort (OR = 0.92, p < 0.01) (Table 6). In both cohorts, the rates were higher in patients with longer LOH (p < 0.01 in both the pre-COVID-19 and COVID-19 cohorts). Regarding readmission, in both cohorts, approximately 40% of patients were re-admitted to the same hospital, while around 60% of patients were re-admitted at other hospitals. More patients with LOH ≥ 8 days were readmitted to the same hospital (47.96% in the pre-COVID-19 cohort and 47.54% in the COVID-19 cohorts) than patients with LOH ≤ 7 days (20.68% in the pre-COVID-19 cohort [p < 0.01] and 19.14% in the COVID-19 cohort [p < 0.01]). The most common type of healthcare institution was primary hospitals in both the pre-COVID-19 and COVID-19 cohorts (Table 5). The reoperation rate was 1.22% in the pre-COVID-19 cohort and 1.24% in the COVID-19 cohort (OR = 1.02; p = 0.72) (Table 6). The time to reoperation was 23.44 ± 13.49 days in the pre-COVID-19 cohort and 21.65 ± 11.98 days in the COVID-19 cohort.
Table 6
Readmissions, reoperations, and transfers.
| Number of patients | Number of events | Odds ratio (95% CI) | p-value |
30-day readmission | | | | |
Pre-COVID-19 | 75453 | 3110 (4.12) | 1 (ref) | |
COVID-19 | 75453 | 2859 (3.79) | 0.916 (0.870, 0.965) | 0.0009 |
30-day reoperation | | | | |
Pre-COVID-19 | 75453 | 680 (0.90) | 1 (ref) | |
COVID-19 | 75453 | 744 (0.99) | 1.095 (0.986, 1.215) | 0.0889 |
Transfer to other hospital | | | | |
Pre-COVID-19 | 75453 | 8383 (11.11) | 1 (ref) | |
COVID-19 | 75453 | 6017 (7.97) | 0.688 (0.664, 0.713) | < .0001 |
In the COVID-19 cohort, although the rates of 30-day readmission and transfer to other hospitals significantly decreased (OR = 0.92 and OR = 0.67, respectively, p < 0.01), the reoperation rates after surgery and discharge did not show a significant decrease (OR = 1.10 and OR = 1.02, respectively, p = 0.73) (Table 6).