Articles Effect of COVID-19 pandemic lockdowns on planned cancer surgery for 15 tumour types in 61 countries: an international, prospective, cohort study COVIDSurg Collaborative* Summary Background Surgery is the main modality of cure for solid cancers and was prioritised to continue during COVID-19 Lancet Oncol 2021; 22: 1507–17 outbreaks. This study aimed to identify immediate areas for system strengthening by comparing the delivery of Published Online elective cancer surgery during the COVID-19 pandemic in periods of lockdown versus light restrictions. October 5, 2021 https://doi.org/10.1016/ S1470-2045(21)00493-9 Methods This international, prospective, cohort study enrolled 20 006 adult (≥18 years) patients from 466 hospitals in *Collaborating authors are listed 61 countries with 15 cancer types, who had a decision for curative surgery during the COVID-19 pandemic and were in the appendix followed up until the point of surgery or cessation of follow-up (Aug 31, 2020). Average national Oxford COVID-19 Correspondence to: Stringency Index scores were calculated to define the government response to COVID-19 for each patient for the Mr Aneel Bhangu, NIHR Global period they awaited surgery, and classified into light restrictions (index <20), moderate lockdowns (20–60), and full Health Research Unit on Global lockdowns (>60). The primary outcome was the non-operation rate (defined as the proportion of patients who did not Surgery, Institute of Translational Medicine, University of undergo planned surgery). Cox proportional-hazards regression models were used to explore the associations between Birmingham, Birmingham, UK lockdowns and non-operation. Intervals from diagnosis to surgery were compared across COVID-19 government a.a.bhangu@bham.ac.uk response index groups. This study was registered at ClinicalTrials.gov, NCT04384926. or Mr James Glasbey, NIHR Global Findings Of eligible patients awaiting surgery, 2003 (10·0%) of 20 006 did not receive surgery after a median follow-up Health Research Unit on Global of 23 weeks (IQR 16–30), all of whom had a COVID-19-related reason given for non-operation. Light restrictions were Surgery, Institute of Translational Medicine, University of associated with a 0·6% non-operation rate (26 of 4521), moderate lockdowns with a 5·5% rate (201 of 3646; adjusted Birmingham, Birmingham, UK hazard ratio [HR] 0·81, 95% CI 0·77–0·84; p<0·0001), and full lockdowns with a 15·0% rate (1775 of 11 827; HR 0·51, j.glasbey@bham.ac.uk 0·50–0·53; p<0·0001). In sensitivity analyses, including adjustment for SARS-CoV-2 case notification rates, moderate See Online for appendix lockdowns (HR 0·84, 95% CI 0·80–0·88; p<0·001), and full lockdowns (0·57, 0·54–0·60; p<0·001), remained independently associated with non-operation. Surgery beyond 12 weeks from diagnosis in patients without neoadjuvant therapy increased during lockdowns (374 [9·1%] of 4521 in light restrictions, 317 [10·4%] of 3646 in moderate lockdowns, 2001 [23·8%] of 11 827 in full lockdowns), although there were no differences in resectability rates observed with longer delays. Interpretation Cancer surgery systems worldwide were fragile to lockdowns, with one in seven patients who were in regions with full lockdowns not undergoing planned surgery and experiencing longer preoperative delays. Although short-term oncological outcomes were not compromised in those selected for surgery, delays and non-operations might lead to long-term reductions in survival. During current and future periods of societal restriction, the resilience of elective surgery systems requires strengthening, which might include protected elective surgical pathways and long- term investment in surge capacity for acute care during public health emergencies to protect elective staff and services. Funding National Institute for Health Research Global Health Research Unit, Association of Coloproctology of Great Britain and Ireland, Bowel and Cancer Research, Bowel Disease Research Foundation, Association of Upper Gastrointestinal Surgeons, British Association of Surgical Oncology, British Gynaecological Cancer Society, European Society of Coloproctology, Medtronic, Sarcoma UK, The Urology Foundation, Vascular Society for Great Britain and Ireland, and Yorkshire Cancer Research. Copyright © 2021 The Author(s). Published by Elsevier Ltd. This is an Open Access article under the CC BY-NC-ND 4.0 license. Introduction public and media attention has largely focussed on the During the COVID-19 pandemic, government restrictions economic impact of lockdowns, the broader effects on have aimed to control community SARS-CoV-2 trans­ general health are poorly understood.3 Lockdowns might mission and included reducing population movement, have had collateral effects beyond controlling community closing public facilities, and restricting gatherings.1 SARS-CoV-2 rates alone, due to changes in both public Restrictions have varied worldwide in stringency, with the behaviour and health system performance.4 These most severe leading to so-called lockdowns.2 Although might have disproportionate effects on vulnerable and www.thelancet.com/oncology Vol 22 November 2021 1507 Articles Research in context Evidence before this study prospectively. It included data from the 15 most common solid Guidance from health ministries and national surgical cancer types across all country-income settings, providing wide associations prioritised time-dependent cancer surgery to generalisability to global policy. The analysis allowed a direct continue during societal restrictions related to COVID-19. comparison of full and moderate lockdowns to light We searched PubMed and Embase on Feb 12, 2021, without date restrictions, accounting for their dynamic nature, where limits, for prospective, multicountry studies describing non- different patients from the same country were exposed to operation rates for patients planned to undergo elective surgery different lockdown states. during national or regional COVID-19 lockdowns using primary Implications of all the available evidence data. We used the search terms “COVID-19”, “SARS-CoV-2”, This study has direct policy, organisational, and clinical “coronavirus”, “lockdown” and “pandemic”, in combination with implications. It has revealed the fragility of elective cancer surgery “surgery” and “non-operation”, “cancellation”, “postponement” systems to lockdowns, particularly health systems in lower- or “delays”, and applied no language restrictions. Several middle-income countries. This study demonstrates the need for modelling studies estimated total elective surgeries cancelled due system strengthening in elective surgery across all settings to to COVID-19, but no primary studies described the impact of mitigate against impending COVID-19 lockdowns and future lockdowns on non-operation rates for patients due to undergo pandemics. This should include both global reorganisation to curative cancer surgery. provide protected COVID-19-free elective surgical pathways Added value of this study (and staffing) that sustainably allow safe surgery to continue, There is limited evidence of the collateral effects of COVID-19 and improved surge capacity for acute care during public health pandemic lockdowns outside of modelling studies. Uniquely, emergencies. The potential long-term effects for patients who this international study prospectively enrolled patients with a underwent delayed surgery may require closer follow-up for decision for curative surgery awaiting surgery during the metastatic disease. This study could inform policy makers’ SARS-CoV-2 pandemic and tracked their care pathways planning regarding the collateral effects of societal restrictions. marginalised communities.5,6 Understanding these effects Methods will justify expenditure on targeted system strengthening, Study design and participants as further societal restrictions are predicted at a global This international, multicentre, prospective cohort study level. included adult patients (≥18 years) with a diagnosis of a In the first COVID-19 waves (ie, the 12 weeks of peak surgically curable cancer during the COVID-19 disruption), at least 21 million elective operations were pandemic. The study was conducted in accordance with a For the protocol see https:// cancelled globally, partly due to concerns over post­ preregistered protocol (NCT04384926). Local principal globalsurg.org/cancercovidsurg/ operative SARS-CoV-2 infection and partly due to capacity investigators were responsible for obtaining clinical issues within hospitals.7,8 There was, however, general audit, institutional review board, or ethical approval in guidance from health ministries and national surgical line with local and national regulations. In most settings, associations that time-dependent surgery should a waiver of individual patient consent was obtained. In continue.9 This included curative cancer surgery, which other countries, formal written or verbal consent was is a priority among the oncology com­munity and a high- required based on recommendations of local ethics and value topic for society.10,11 Surgery remains the primary governance committees. Data were collected online and method of cure for most solid cancers. stored on a secure server running the Research Electronic Since surgical databases and cancer registries do not Data Capture (REDCap) web application.14 capture prospective decision making, they lack fidelity to Any hospital worldwide that performed elective cancer detect patients who did not undergo planned surgery. In surgery in an area affected by the COVID-19 pandemic the case of curative surgery, these are the patients who was eligible to participate. Patients listed for surgery to might have suffered the most harm.12,13 Resection cure a solid cancer were included in each centre for margins alone are an inadequate marker of success, as 3 months from local emergence of COVID-19, defined on selection bias in patients who are able to undergo surgery a centre-by-centre basis as the date where first notification risks underestimating harm from treatment delays, and of SARS-CoV-2 cases occurred in the local area (between neglects whole-system effects. We planned the Jan 21 and April 14, 2020). Participating centres identified prospective COVIDSurg Cancer study to address these all patients with a decision for surgery (or would have areas and provide an accurate, whole-system analysis of had a decision for surgery under normal, prepandemic the impact of COVID-19 on planned cancer surgery. circumstances) from multidisciplinary team meetings, Understanding any harms could allow for immediate tumour board, outpatient clinics, or local equivalents. local and national policy changes, in preparation for Previous international outcomes studies from our group future societal restrictions. have shown that this method achieves greater than 1508 www.thelancet.com/oncology Vol 22 November 2021 Articles 95% case ascertainment and greater than 98% data light restrictions (index <20), moderate lockdowns accuracy during external validation.15 If a specialty within (20–60), and full lockdowns (>60). These groups allowed a hospital was unable to confirm consecutive enrolment, a direct com­parison of full and moderate lockdowns to their data were excluded from analysis. Patients’ care light restrictions, accounting for their dynamic nature, pathways were followed up until the point of surgery or whereby different patients from the same country were until cessation of follow-up at Aug 31, 2020. This date exposed to different lockdown states. For each patient, a was selected to ensure all patients had a minimum of median average score while waiting for surgery and the 12-weeks follow-up. Where a patient underwent surgery, number of weeks in full lockdown were calculated outcome data was collected up to 30 days after surgery. and used in analyses. Full details are given in the Where patients remained non-operated, their last known appendix (pp 2, 57–58). status was recorded. The 15 most common solid cancer types were included Definition of SARS-CoV-2 rates in this study, including colorectal, oesophageal, gastric, The case notification rate was calculated at an individual head and neck (oral, oropharyngeal, laryngeal, hypo­ patient level as a median average between the date of pharyngeal, salivary, thyroid, paranasal sinus, skin), local emergence of COVID-19 and the date of surgery or thoracic (lung, pleural, mediastinal, chest wall), liver, cessation of follow-up via the Our World in Data portal.16 pancreatic, prostate, bladder, renal and upper urinary A high COVID-19 burden area was classified as a median tract urothelial, gynaeco­logical (uterine, ovarian, cervical, of at least 25 cases per 100 000 per 14 days, representing vulval, vaginal), breast, soft-tissue sarcoma, bony sarcoma, WHO recommendations at the time of the study (ie, in and intracranial malignancies. Participating centres could keeping with first pandemic wave levels). Case rates were contribute data for either single or multiple cancers. Early used for exploratory analyses only, and stratified by World cancers that were planned to be managed with endo­ Bank income tertile (appendix p 58).17,18 scopic surgery alone (eg, transurethral resection of bladder tumour, transanal endoscopic micro­ surgery) Other definitions were excluded. Patients who were suspected to have an The World Bank index (2019/20 update) was used to operable cancer, but were later identified to have a classify countries and patients into three groups based non-cancerous condition (eg, on postoperative histo­ on Gross National Income per capita (US$) calculated pathology), or were treated as benign and unexpectedly using the Atlas method: high-income countries, UMICs, identified to be malignant on postoperative histopathology and LMICs (including patients from both low-income were also excluded. countries and LMICs). Data on baseline patient status was collected for the purpose of adjustment for case-mix Definition of lockdowns in exploring associations between lockdowns and We used the Oxford COVID-19 Stringency Index to surgical capacity (appendix p 59). define each country’s national government response to Patients were classified into three groups according to COVID-19. This index is a composite of 19 indicators their neoadjuvant treatment group: (1) no neoadjuvant including measures and behavioural interventions therapy (ie, straight to surgery); (2) neoadjuvant therapy, related to containment and closure, economic response, standard care (where the treating clinician administered and health systems. Each indicator is scored using an neoad­juvant treatment in accordance with their usual ordinal scale (0 to 2, 3, 4, or 5), with an overall score care); (3) neoadjuvant therapy, COVID-19 decision calculated by adding together individual indicator scores (where the treating clinician administered a neoadjuvant (appendix p 56). Total scores can range from 0 (no treatment where this would not typically be indicated). To restrictions) to 100 (most stringent restrictions). The estimate the impact of lockdown on treatment delays, the index has been previously validated by demonstrating relationship between lockdowns and the interval from associations with population SARS-CoV-2 infection rates diagnosis to decision for surgery to surgery was and mobile phone mobility data.1 measured. The interval from date of diagnosis to the date The average national Oxford COVID-19 Stringency of surgery was calculated in whole weeks to identify Index scores were calculated for each patient for the points of system friction (appendix p 60). period they waited for surgery. To define cutoffs that were reflective of real-world policy, we sampled reported Outcomes lockdown dates from a sample of high-income countries, A resilient elective surgical care system is defined as a upper-middle-income countries (UMICs), and lower- hospital or network of hospitals that is able to maintain middle-income countries (LMICs; appendix p 55). Dates both its capacity and safety during public health crises.19 were taken from national policy, media, and press As a measure of the ability of surgical systems to sources. On the date of transition into lockdown, the maintain their capacity, the primary outcome measure point estimate for the COVID-19 Stringency Index score was the non-operation rate. This non-operation rate was was extracted (appendix p 55). This point estimate was defined as an eligible patient (ie, with a plan to undergo used to classify patients into three stringency groups: surgery) not undergoing their planned operation during v www.thelancet.com/oncology Vol 22 November 2021 1509 Articles 27 700 patients identified from multidisciplinary team meeting (or equivalent) 7694 excluded 764 surgery planned with palliative intent 2466 diagnosis made after surgery 840 diagnosis made after decision for surgery 1859 carcinoma in situ (Tis/Ta) 222 planned redo resection 883 intracranial tumours 660 local resection only 20 006 adult patients with decision for curative cancer surgery (61 countries, 466 centres) 18 003 operated 2003 non-operated 15 622 no neo- 1736 neoadjuvant 645 neoadjuvant 1353 no neoadjuvant 164 neoadjuvant 486 neoadjuvant adjuvant therapy therapy therapy therapy therapy therapy (standard care) (COVID-19 (standard care) (COVID-19 decision) decision) 4727 <4 weeks from diagnosis to 5597 5–8 weeks from diagnosis to 2601 9–12 weeks from diagnosis to 2697 >12 weeks from diagnosis to surgery surgery surgery surgery 173 unresectable disease* 87 unresectable disease* 46 unresectable disease* 52 unresectable disease* Figure 1: Flowchart of included patients *Found clinically, radiologically, or during surgery. the study window. Patients were classified as being complex interplay with treatment interval, an a-priori operated if they underwent surgery, regardless of whether decision was made to only include patients who went there was a change to the planned urgency (eg, from straight to surgery (no neoadjuvant therapy) in exploration elective to emergency) or intent (eg, from curative of the effects of treatment delay on secondary outcomes to palliative). Patients who died or progressed to (appendix p 62). unresectable disease before surgery were classified as non-operated, so these did not act as competing risks. Statistical analysis For patients who did not receive their surgery as The full statistical method is presented in the planned during the follow-up window, the treating appendix (p 63). Cox proportional hazards regression clinical team selected one or more reasons that the modelling was used to explore associations between patient had not had surgery. These reasons included lockdowns and the primary outcome, presented as those explicitly COVID-19 related (eg, decision to delay adjusted hazard ratios (HRs) and 95% CIs. Operation surgery due to patient risk during COVID-19), and those was included as the outcome event, and no censoring non-COVID-19 related (eg, delay due to other unrelated was performed for death or progression to unresectable medical or surgical condition). More than one reason for disease to deal with competing risks, given individuals non-operation could be selected for each patient, had the same follow-up time (ie, describing sub­ representing the complexity of decision making (full distribution rather than cause-specific hazards). An details are given in the appendix p 61). α level was set at 0·05 (5%) for interpretation of Secondary outcome measures related to safety of surgery significance. Several preplanned sensitivity analyses were presented for patients who underwent surgery during were conducted for the primary analysis to examine the follow-up period: (1) resection margin status for those robustness of findings; namely, (1) including elective selected for surgery; (2) resectable disease at the time of operations only in the definition of the primary outcome; surgery; (3) preoperative cancer compli­ cation requiring (2) accounting for an interaction effect between World emergency surgery; (3) 30-day post­operative SARS-CoV-2 Bank income group and COVID-19 stringency index infection rate; (4) 30-day post­ operative mortality rate; group; and (3) accounting for local SARS-CoV-2 case (5) new detection of metastatic disease up to a maximum notification rates, stratified by World Bank income group. of 30 days after surgery. As neoadjuvant therapy has a Two further sensitivity analyses were performed to 1510 www.thelancet.com/oncology Vol 22 November 2021 Articles ensure that differences in cancer case-mix across income A settings were not responsible for residual confounding; Operated these included, (1) cancer location removed from the No Yes model; and (2) patients older than 50 years only. 100 0·5 4·2 A secondary analysis was used to explore the incremental 13·7 Proportion of patients (%) effect of weeks in lockdown on a patient’s likelihood of 75 High income non-operation. Intervals from diagnosis to surgery were 50 99·5 compared across COVID-19 government response index 95·8 86·3 groups. We only analysed the interval between diagnosis 25 and surgery for patients who did not receive neo-adjuvant therapy to avoid confounding due to legitimate delays to 0 surgery in patients who receive neo-adjuvant therapy. All 0·4 analyses were carried out using R, version 3.1.1 (packages 100 14·5 10·5 Proportion of patients (%) finalfit, tidyverse, ggsurvplot). Upper-middle income 75 Role of the funding source 50 99·6 85·5 89·5 The funders of the study had no role in study design, data collection, data analysis, data interpretation, or 25 writing of the report. 0 Results 100 1·5 11·5 20 006 patients were eligible for inclusion in 466 hospitals Proportion of patients (%) 24·2 Lower-middle income in 61 countries (figure 1). Of these patients, 1891 (9·5%) 75 were from 17 UMICs and 2249 (11·2%) were from 50 98·5 12 LMICs. A wide range of patients, tumours, and 88·5 75·8 operations were included. The most common tumour 25 types included were breast (n=3896; 19·5%), head and neck (n=3517; 17·6%), colon (n=3428; 17·1%), and 0 gynaecological (n=2169; 10·8%). Distribution of patients Light restrictions Moderate lockdown Full lockdown and cancers across income groups and countries is B shown in the appendix (pp 8–9). 100 + + Light restrictions Of patients planned for cancer surgery during the + + + Moderate lockdown + COVID-19 pandemic, 4521 (22·6%) of 20 006 were + Full lockdown + p<0·0001 Proportion non-operated (%) + awaiting surgery during a period of light restrictions, 75 + + 3646 (18·2%) during moderate lockdowns, and + + + + 11 827 (59·1%) during full lockdowns (n=12 missing data; + + + 50 + appendix p 4). The proportion of patients awaiting + + + + surgery in full lockdowns was higher in areas with high + + + + + + than low community SARS-CoV-2 case notification rates 25 + + + + + + + + and in UMICs and LMICs than in high-income countries + + + + + + + + + + + + + (appendix p 10). Patients awaiting surgery during light + + + + + + + + + + + + + + + + restrictions had a lower mean number of weeks in full 0 0 4 8 12 16 20 24 28 lockdown (2·4 weeks [SD 1·7]) compared with patients Time since diagnosis (weeks) awaiting surgery during moderate (5·5 weeks [2·9]) or Number at risk full lockdowns (12·7 weeks [5·4]; p<0·0001, from one- Light restrictions 4520 3080 1214 581 330 195 86 2 Moderate lockdown 3633 2576 1257 720 454 276 170 46 way ANOVA). Full lockdown 11 678 9266 6572 4752 3274 2044 1132 423 Most patients (16 975 [84·8%] of 20 006) had a plan to progress straight to surgery (no neoadjuvant therapy), Figure 2: Effects of lockdowns on surgical capacity (A) Differences in resilience of surgical systems across income settings by COVID-19 stringency index group. with 1900 (9·5%) receiving standard care neoadjuvant Percentages represent proportion operated by group. (B) Kaplan-Meier plot demonstrating proportion of patients therapy, and 1131 (5·7%) receiving a COVID-19 decision remaining non-operated over time from cancer diagnosis grouped by COVID-19 stringency index group. for neoadjuvant therapy. During full lockdowns, patients Plot censored at 28 weeks maximum follow-up from cancer diagnosis. Shading represents this represents the were more likely to have a COVID-19 decision for 95% CI, using the statistical package ggsurvplot. neoadjuvant therapy than during moderate lockdowns or light restrictions (appendix p 3). awaiting surgery during periods of light restrictions had During the COVID-19 pandemic, 2003 (10·0%) of a lower rate of non-operation (26 [0·6%] of 4521) than 20 006 patients did not undergo their planned surgery by those in moderate lockdowns (201 [5·5%] of 3646) or full the end of follow-up (figure 2; appendix p 12). Patients lockdowns (1775 [15·0%] of 11 827; appendix p 5). v www.thelancet.com/oncology Vol 22 November 2021 1511 Articles n/N HR (95% CI) p value COVID-19 stringency index Ref Light restrictions 4371/4391 Moderate lockdown 3364/3509 0·81 (0·77–0·84) p<0·0001 Full lockdown 9862/11 166 0·51 (0·50–0·53) p<0·0001 World bank index High income 14 138/15 115 Ref Upper-middle income 1687/1801 0·97 (0·92–1·02) p=0·2202 Lower-middle income 1772/2150 0·83 (0·78–0·87) p<0·0001 Age, years <50 3228/3527 Ref 50−59 3636/3916 1·10 (1·05–1·15) p=0·0002 60−69 4742/5166 1·01 (1·04–1·14) p=0·0005 70−79 4333/4639 1·17 (1·11–1·23) p<0·0001 ≥80 1658/1818 1·06 (0·99–1·14) p=0·0772 Sex Female 10 175/10 776 Ref Male 7422/8290 0·99 (0·95–1·02) p=0·4569 ASA grade 1−2 12 549/13 565 Ref 3−5 5048/5501 0·99 (0·95–1·03) p=0·5327 ECOG performance status score 0 10 469/11 201 Ref 1 5195/5672 0·96 (0·92–0·99) p=0·0198 ≥2 1933/2193 0·89 (0·84–0·94) p<0·0001 Current smoker No 15 555/16 802 Ref Yes 2042/2264 1·03 (0·99–1·08) p=0·1712 Pre-existing respiratory condition No 15 598/16911 Ref Yes 1999/2155 0·98 (0·94–1·03) p=0·4716 RCRI 0 5413/5682 Ref 1 8992/9839 1·01 (0·95–1·08) p=0·7788 2 2531/2810 0·96 (0·89–1·03) p=0·2449 ≥3 661/735 0·90 (0·81–0·10) p=0·0438 Disease stage Early disease 9670/10453 Ref Advanced or nodal disease 7927/8613 0·96 (0·93–0·10) p=0·0181 Cancer location Head or neck 3142/3469 Ref Figure 3: Multivariable Cox Colon 3208/3359 1·16 (1·08–1·24) p<0·0001 proportional hazards model Rectal 1309/1462 0·59 (0·54–0·64) p<0·0001 of factors associated with Gastric 644/712 0·67 (0·61–0·74) p<0·0001 non-operation during Oesophageal 324/435 0·32 (0·28–0·36) p<0·0001 COVID-19 19 832 in dataframe, Lung 1047/1159 0·82 (0·76–0·90) p<0·0001 19 066 in model, 766 missing. Liver 696/759 0·63 (0·57–0·69) p<0·0001 17 597 (91·8%) of Pancreatic 628/741 0·73 (0·66–0·81) p<0·0001 19 066 patients included in Sarcoma 377/413 0·67 (0·59–0·75) p<0·0001 this model were operated by Prostate 427/504 0·44 (0·40–0·50) p<0·0001 the end of follow-up. Missing Kidney or upper tract urothelial 363/422 0·65 (0·57–0·73) p<0·0001 data are described in the appendix (p 10), as well as the Bladder 104/139 0·51 (0·42–0·62) p<0·0001 full model (p 12). Gynaecological 1884/2048 0·86 (0·79–0·93) p=0·0001 ASA=American Society of Breast 3135/3444 1·00 (0·94–1·05) p=0·8835 Anesthesiologists Physical Status Classification System. 0·4 0·6 0·8 1·0 1·2 ECOG=Eastern Cooperative Favours Favours Oncology Group. RCRI=Revised non-operation operation Cardiac Risk Index. 1512 www.thelancet.com/oncology Vol 22 November 2021 Articles 1775 (88·7%) of 2003 patients who remained non- All patients operated were in regions with full lockdowns COVID-19 related (appendix p 11). The baseline rate of non-operation was Multidisciplinary team decision to delay surgery due to patient risk during COVID-19 1456 (72·8%) low across all income settings during periods of light Change to alternative treatment modality because of COVID-19 533 (26·6%) restrictions (22 [0·5%] of 4089 in high-income countries; one [0·4%] of 228 in UMICs; three [1·5%] of Patient choice to avoid surgery during COVID-19 pandemic 460 (23·0%) 204 in LMICs), and high during periods of full lock­ Ongoing neoadjuvant therapy (COVID decision) 378 (18·9%) down (1188 [13·7%] of 8644 in high-income countries; No bed, critical care bed, or operating room space available due to COVID-19 299 (14·9%) 139 [10·5%] of 1329 in UMICs; 448 [24·2%] of Change of recommendations in society guidelines related to COVID-19 220 (11·0%) 1854 in LMICs). At 12 weeks after diagnosis, 581 [12·8%] Patient unable to travel to hospital related to COVID-19 140 (7·0%) of 4520 patients under light restrictions remained non- Collateral impact on supporting services causing delay 24 (1·2%) operated, 720 (19·9%) of 3622 patients during moderate Patient delayed due to SARS-CoV-2 infection 23 (1·1%) lockdowns, and 4752 (40·7%) of 11 678 patients during Died of COVID-19 while waiting for surgery 14 (0·6%) full lockdowns (figure 2). After multivariable adjustment, Total 2001 (100·0%) both moderate (HR 0·81, 95% CI 0·77–0·84; p<0·0001) Non-COVID-19 related and full lock­ downs (HR 0·51, 0·50–0·53; p<0·0001) Progression to unresectable disease 179 (8·9%) were associated with a lower likelihood of a patient Delay due to other unrelated medical or surgical condition 59 (2·9%) receiving their planned cancer surgery (figure 3; Died unrelated to COVID-19 while waiting for surgery 34 (1·7%) appendix p 10). This was con­ sistent across planned Patient unable to afford surgery 24 (1·2%) sensitivity analyses (appendix pp 13–16). The overall level Patient choice to avoid surgery unrelated to COVID-19 35 (1·7%) of missingness was low (<1%) for all variables included Total 306 (15·3%) in the models. We anticipated that decisions to delay or cancel surgery during COVID-19 would be complex. Therefore, selecting more Being in an LMIC, increasing frailty (Eastern than one reason for non-operation during the follow-up window for each patient was permitted. One patient could Cooperative Oncology Group 1 or ≥2), comorbidity have both one or more COVID-19-related and non-COVID-19-related reasons selected. Where it was unclear whether a reason was directly COVID-related (eg, disease progression) this was classified as not COVID-19-related. Two patients (Revised Cardiac Risk Index ≥3), and having locally (0·1%) had no reasons given for non-operation during the follow-up window selected (missing data). Proportions are advanced or nodal disease, or both, were all independently therefore expressed as a percentage of 2001 non-operated patients and with data available. associated with increased likelihood of non-operation. Table 1: Reasons that patients did not received planned surgery There was significant variability in the likelihood of non-operation by cancer site. Where the primary outcome definition was revised to include elective non-operation compared with light restrictions; surgery only, both UMICs and LMICs were observed to appendix p 20). have a higher adjusted non-operation rates than high- Increasing SARS-CoV-2 case notification rates were income countries (appendix p 13). The effect of lockdown associated with increasing non-operation rates (appendix on non-operation differed by income group, with LMICs p 21). Both moderate and full lockdowns were consistently broadly less likely to operate at a given level of lockdown associated with an increased likelihood of non-operation, compared with the high-income group (appendix p 15). even after adjustment for local SARS-CoV-2 rates Patients waiting for surgery in LMICs during full (appendix p 22). The largest magnitude of effect was seen lockdowns were most likely to remain non-operated when transitioning from light restrictions or moderate compared with patients in LMICs during light lockdowns to full lockdowns across all income and restrictions (HR 0·41, 95% CI 0·38–0·44; p<0·0001; SARS-CoV-2 case notification rate groups. LMICs were appendix p 16). In the secondary analysis, waiting for particularly fragile to increasing SARS-CoV-2 rates and surgery for 5–6 weeks or more in full lockdown was full lockdowns, with a non-operation rate of 381 (58·7%) associated with a reduced likelihood of a patient of 649 (appendix p 21). undergoing their cancer operation compared with By the end of the follow-up (median 23 weeks, 0 weeks in full lockdown (5–6 weeks in full lockdown IQR 16–30), 2003 patients had not undergone planned HR 0·86, 95% CI 0·80–0·93, p<0·0001; appendix p 6). surgery. 453 (22·6%) of 2003 patients had been formally Patients younger than 50 years were less likely than re-staged. Detection of new metastatic disease is shown patients 50 years and older to receive their planned in the appendix (p 26). surgery across several sensitivity analyses. Patients Of non-operated patients for whom data were available planned to have surgery aged younger than 50 years (n=2001; two missing data), all had at least one COVID- were more commonly from LMICs than UMICs 19-related reason provided for non-operation (table 1); or high-income countries commonly from LMICs most commonly this involved a team decision to delay (appendix pp 18–19). In a sensitivity analysis including surgery during COVID-19 due to individual patient risk only patients older than 50 years, the effect of lockdowns (1456 [72·8%] of 2001). 533 (26·6%) of 2001 patients were on surgical capacity remained consistent with the provided an alternative treatment modality as a result of primary result (ie, moderate and full lockdowns were COVID-19. 306 (15·3%) patients had at least one non- associated with a significant increase in the odds of COVID-19-related reason provided for non-operation. v www.thelancet.com/oncology Vol 22 November 2021 1513 Articles across all neoadjuvant treatment groups (figure 4). In A Interval diagnosis to operation patients who went straight to surgery (no neoadjuvant 0−4 weeks 5−8 weeks 9−12 weeks >12 weeks therapy; 15 622 [86·3%] of 18 003), full lockdown was 100 associated with 2001 [23·8%] of 11 827 patients not receiving surgery within 12 weeks of diagnosis compared Proportion of patients (%) n=34 n=35 n=31 75 with 317 (10·4%) of 3646 patients during moderate No neoadjuvant lockdowns, and 374 (9·1%) of 4521 patients under light therapy n=30 50 n=42 n=39 restrictions (appendix p 25). For these patients, each 25 n=17 additional week in lockdown was associated with treat­ n=15 n=15 ment delay (p<0·0001; figure 4). Increasing SARS-CoV-2 n=22 n=9 n=11 0 case notifi­cation rates were also associated with increased delays beyond 12 weeks across income groups, with the 0·4 n=5 n=4 100 longest delays observed in UMICs and LMICs during n=11 n=9 n=5 Proportion of patients (%) n=7 periods with high SARS-CoV-2 rates (appendix p 23). The n=13 Neoadjuvant therapy, 75 n=12 n=12 point of system friction was different across different standard care 50 income groups (appendix p 7). Full lockdown was asso­ n=84 n=63 n=73 ciated with an increased interval from decision to surgery 25 across all settings compared with both light restrictions 0 and moderate lockdown (figure 4; appendix p 24). In patients who went straight to surgery (n=15 622), 100 n=5 n=2 postoperative histopathological and clinical outcomes n=27 n=10 n=45 during light restrictions, moderate lockdowns, and full Proportion of patients (%) n=36 n=23 Neoadjuvant therapy, n=22 lockdowns were similar (table 2). Characteristics and COVID-19 decision 75 n=14 outcomes of patients by interval from diagnosis to 50 n=27 n=65 operation are in the appendix (pp 25–26). Variation in n=23 25 outcomes by income group are shown in the appendix (p 27). 0 Light restrictions Moderate lockdown Full lockdown Discussion The design of this study allowed a holistic overview of B 25 different health systems’ surgical capacity and outcomes during lockdowns. The analysis allowed a direct Time from diagnosis to operation (weeks) comparison during full and moderate lockdowns to 20 periods with light restrictions, taking account of the dynamic nature of government policies, where different 15 patients from the same country were exposed to different lockdown states. During full lockdowns, one in seven 10 patients did not receive their planned operation, all of whom had a pandemic-related reason for non-operation. 5 This finding was robust, and consistent in sensitivity analyses. In a secondary analysis, awaiting surgery in a 0 full lockdown for greater than 6 weeks was associated 0 5 10 15 20 25 with an increased likelihood of non-operation. These Time in full lockdown, COVID-19 stringency index (weeks) data reveal the fragility of elective cancer surgery to Figure 4: Lockdown and delay to surgery lockdowns, which was independent of both local (A) Delay from diagnosis to surgery during lockdowns (according to COVID-19 stringency index group) by SARS-CoV-2 rates and case-mix. Patients with cancer in neoadjuvant therapy group. Percentages represent proportion of operated patients who were in each interval from LMICs, of increasing frailty, or with advanced disease diagnosis to operation group. (B) Weeks in full lockdown and interval from cancer diagnosis to operation. were most vulnerable to lockdown effects. Capacity for Plot displays patients who went straight to surgery (no neoadjuvant therapy only). Full lockdown defined as a COVID-19 stringency index score of more than 60. Plotted line represents a smoothed conditional mean from a major elective cancer should be part of every country’s fitted generalised additive model. The shaded area denotes bounds of the 95% CI. strategy to address whole-population health needs and prevent further collateral harm. 179 (8·9%) patients progressed to unresectable disease. Identifying at-risk groups allows targeted system 48 (2·4%) patients died before their planned surgery strengthening during both COVID-19 lockdowns and (14 due to COVID-19-related complications, and 34 due to future pandemics. Firstly, vulnerable patient groups non-COVID-19-related causes). (eg, those with a poorer performance score, more cardiac Delays from diagnosis to operation were observed comorbidities, or advanced cancers) were all less likely to during full lockdowns for operated patients (n=18 003) receive surgery. Secondly, certain operation types that 1514 www.thelancet.com/oncology Vol 22 November 2021 Articles require more intensive perioperative care, including Light Moderate Full lockdown Total* p value† those for oesophageal and pancreatic cancer, were at restrictions lockdown (n=8402) (N=15 622) increased risk of cancellation. Thirdly, patients in LMICs (n=4152) (n=3057) were less likely to undergo surgery during lockdowns Margin status and SARS-CoV-2 surges, which included a high R0 3471 (83·7%) 2619 (85·8%) 7238 (86·3%) 13 328 (85·5%) 0·0011 proportion of young patients (<50 years). Protected R1 381 (9·3%) 223 (7·4%) 581 (6·9%) 1185 (7·7%) ·· elective surgical capacity might include protected R2 79 (1·9%) 61 (2·0%) 157 (1·9%) 297 (1·9%) ·· COVID-19-free path­ways (including dedicated surgical, Pathology unavailable 214 (5·2%) 147 (4·8%) 407 (4·8%) 768 (4·9%) ·· anaesthetic, and theatre staff) within larger hospitals, or Missing 7 7 19 33 ·· smaller bespoke elective surgery units that function as Resectable disease at time of surgery part of cancer treatment networks.20 This also requires Resectable 4069 (98·0%) 2967 (97·1%) 8213 (97·8%) 15 249 (97·7%) 0·045 long-term investment in surge capacity for the acute care Unresectable 81 (2·0%) 90 (2·9%) 187 (2·2%) 358 (2·3%) ·· workforce and formal operational planning to manage Unknown 2 0 2 4 ·· public health emergencies without major disruption to Pre-operative cancer-related complication requiring emergency surgery‡ elective care (further details are in the appendix p 64). Elective 4071 (98·2%) 2989 (97·8%) 8199 (97·8%) 15 259 (97·9%) 0·27 Together, protected elective surgical capacity might allow Emergency 74 (1·8%) 67 (2·2%) 185 (2·2%) 326 (2·1%) ·· essential elective surgery to continue despite external 30-day SARS-CoV-2 infection rate‡ system shocks.20 No 4083 (98·3%) 3039 (99·4%) 8362 (99·5%) 15 484 (99·2%) <0·0001 The least resilient systems were in LMIC settings, Yes 69 (1·7%) 18 (0·6%) 40 (0·5%) 127 (0·8%) ·· exacerbating resource scarcity and capacity issues that 30-day postoperative mortality rate‡ were present prepandemic in the management of non-communicable diseases.11,21 Elective cancer surgery No 4080 (98·3%) 3016 (98·7%) 8307 (99·0%) 15 403 (98·8%) 0·0045 systems in LMICs are typically under pressure from a high Yes 70 (1·7%) 41 (1·3%) 84 (1·0%) 195 (1·2%) ·· burden of expedited and emergency presentations.11,15,21 Missing 2 0 11 13 ·· This pressure was seen in our study, where the likelihood New detection of metastatic disease§ of non-operation was higher in both UMICs and LMICs, No 2191 (98·3%) 1625 (98·3%) 4946 (98·2%) 8762 (98·2%) 0·87 despite patients being younger and having fewer Yes 38 (1·7%) 28 (1·7%) 93 (1·8%) 159 (1·8%) ·· comorbidities. These young patients were more frequently Missing 7 5 15 27 ·· affected by financial and geographical causes for non- Data are n (%) or n. Patients with metastatic disease at baseline removed from denominator (N=8957). Percentages operation, revealing a particularly vulnerable group. presented by column total; missing data are excluded. R0=no microscopic or macroscopic disease. R1=microscopic disease at the margin. R2=macroscopic disease at the margin. *11 missing this data point. †χ2 comparing light versus Measures to strengthen the security of global elective moderate versus full lockdowns for each outcome. ‡Subgroups defined in the appendix (p 62). §Detailed data on cancer surgery must be implemented across all settings, detection of new metastatic disease not collected for liver, pancreatic, breast, and gynaecological cancers. and as a priority in LMICs.6 Despite data demonstrating Table 2: Outcomes across COVID-19 stringency index groups for patients going straight to surgery the safety of neoadjuvant treatment during COVID-19, the (no neoadjuvant therapy) overall rate of neoadjuvant therapy as standard care or a COVID-19 decision was low (15·2%). This low rate might represent safety concerns or highlight capacity issues follow-up for metastatic disease. It is possible that there elsewhere in the cancer care pathway.22 Developing robust will be a reverse trend towards worsening cancer survival pathways from diagnosis through to definitive surgical rates over the next 5 years as a consequence of these treatment, supported by public health teams and financial capacity issues, although the present study was not protection mechanisms, will help to create both pandemic- designed to directly capture these long-term effects. We proof and more equitable systems. acknowledge that for some cancer types, neoadjuvant Although we did not find an increase in the positive therapy has equivalent outcomes to the adjuvant resection margin rate or new metastatic disease application of the same treatment and might be a associated with increasing delays, these were highly reasonable strategy to safely delay treatment where this is selected patients and with short-term follow-up only. The required (eg, endocrine therapy for oestrogen-receptor- high proportion of patients who did not receive planned positive breast cancers).23 The impact of changes to surgery reveals the true extent of potential harm. This neoadjuvant treatment pathways and both short-term part of our analysis focussed on patients who were and long-term oncological outcomes requires further treated without neoadjuvant therapy, who are likely to exploration. represent the group at highest risk from unplanned This study had several further limitations. First, effects delays. Evidence from modelling studies and meta- seen during lockdowns could be interpreted as normal analyses suggest that 4-week incremental delays before practice, which would have occurred outside of the surgery are associated with increased rates of recurrence pandemic era. We dealt with this by including an internal and excess mortality.13 Taken together, patients who comparison (light restrictions), which is akin to normal experienced a delay to surgery during the COVID-19 conditions and carried a non-operation rate of 0·5%. We pandemic might warrant strategies that support closer also collected clinicians’ reasons for non-operation, which v www.thelancet.com/oncology Vol 22 November 2021 1515 Articles were overwhelmingly COVID-19 related. Second, we used At the time of publication, lockdowns of varying the Oxford COVID-19 Stringency Index to define magnitude remain in place across many countries lockdowns,1 calculated for each patient as the median around the world, and further measures might be average during their wait for surgery. Although this index imposed related to novel variants of concern and has been validated, it is not yet widely used, and the variability in vaccine availability around the world COVIDSurg Collaborative is an early adopter of this metric (appendix p 64); however, threats to stable elective for research purposes. This health policy measure surgical systems are not limited to COVID-19; other viral demonstrated association with patient level outcomes in pandemics, seasonal pressures, and natural disasters all our dataset. However, we used an aggregate summary affect surgical patients on an annual and recurring basis. statistic that did not reflect all changes in policy during the The lessons from this study might be used to inform study period. More work is required to understand the best surgical system strengthening both during the COVID-19 method to apply this measure in future epidemiological pandemic and beyond. studies. Third, as part of the exposure period to lockdowns Contributors occurred after study entry (ie, decision for surgery), this The writing group (JCG, AAde, AAdi, EA, APA, FA, JA, AMB, AC-C, JE, might have been subject to future information bias, where ME, MF, CF, GG, DG, EAG, EH, PH, IL, SW, HL, SL, EL, GMAG, HM, EJM, JM, DM, KM, MM, RM, DM, FN, FP, MP, PP, AR-DM, KR, ACR, patients remaining non-operated for a longer time might RKS, RS, JFFS, NS, GDS, RS, SS, ST, EHT, RV, DN, AAB) and the have been more likely to await surgery during different statistical analysis group (JCG, KAM, DN, EH, AAB) contributed to lockdown states, therefore have a central tendency in their writing, data interpretation, and critical revision of the manuscript. median average score.24 Fourth, this study required The writing group, operations committee, and dissemination committee contributed to study conception, protocol development, study delivery, prospective capture of team decision making, which might and management. The collaborators contributed to data collection and have been subject to biases, although the scale and diversity study governance across included sites. All members of the writing of the study mitigated against this. Fifth, definition of group had full access to the data in the study, and JCG, KAM, DN, EH, SARS-CoV-2 rates is dependent on testing performed and and AAB verified the underlying data in the study. AAB or JCG and the writing committee had final responsibility for the decision to submit for reported, so might vary at global scale.17,18 We present publication. Detailed role descriptions of all contributing collaborating exploratory analyses around SARS-CoV-2 rates stratified by authors are shown in the appendix (pp 28–54). income setting and provide sensitivity analyses to Declaration of interests demonstrate that findings were robust. Sixth, we did not All authors declare no competing interests. present more detailed analyses of between-country or Data sharing within-country variation. Despite the large numbers in this Anonymised individual participant data will be made available upon study, individual numbers per country were low enough to request to the corresponding authors after the date of publication, with approval of the operations and dissemination committees, and risk type 1 error through multiple hypothesis testing. completion of a data sharing agreement. Seventh, in this analysis we did not explore different Acknowledgments hospital types or delay in care for different cancers. There This research was part-funded by the National Institute for Health might have been hospitals that shutdown completely and Research (NIHR; NIHR 16.136.79) using UK aid from the UK Government did not take part in this study, meaning outcomes might to support global health research. RS receives funding from the Economic have been worse. There might be specialty specific findings and Social Research Council. JCG and AAB are funded by personal awards from the NIHR Academy. The views expressed in this publication are that allow future strategies to become stratified—eg, those of the authors and not necessarily those of the NIHR or the UK patients with rectal and prostate cancer might benefit from Government. scaling up alternative neoadjuvant treatments; breast References and gynaecological cancer surgery might be amenable to 1 Hale T, Angrist N, Goldszmidt R, et al. A global panel database of day case pathways; kidney, bladder, thoracic, and oeso­ pandemic policies (Oxford COVID-19 Government Response Tracker). Nat Hum Behav 2021; 5: 529–38. phagogastric surgery might require the advanced support 2 Han E, Tan MMJ, Turk E, et al. Lessons learnt from easing of surgical units with critical care facilities; colon cancer COVID-19 restrictions: an analysis of countries and regions in Asia could be performed in standalone surgical units. Eighth, Pacific and Europe. Lancet 2020; 396: 1525–34. we did not capture data on delays to diagnosis. Lockdowns 3 Mansfield KE, Mathur R, Tazare J, et al. Indirect acute effects of the COVID-19 pandemic on physical and mental health in the UK: are a system-level issue and high friction in diagnostic a population-based study. Lancet Digit Health 2021; 3: e217–30. pathways was likely to have led to an increasing number of 4 Hamadani JD, Hasan MI, Baldi AJ, et al. Immediate impact of tumours left undetected in the community.25 When stay-at-home orders to control COVID-19 transmission on socioeconomic conditions, food insecurity, mental health, considering resilience of a complete elective surgery and intimate partner violence in Bangladeshi women and their system, there is a vital role of timely diagnostics in families: an interrupted time series. Lancet Glob Health 2020; preventing harm, which might be just as important as 8: e1380–89. 5 Chakrabarti S, Hamlet LC, Kaminsky J, Subramanian SV. delays between diagnosis and surgery. Finally, cancer care Association of human mobility restrictions and race/ethnicity- is just one component of a functioning health system. based, sex-based, and income-based factors with inequities in When making policy decisions about resourcing to well-being during the COVID-19 pandemic in the United States. JAMA Netw Open 2021; 4: e217373. improve resilience, cancer must be balanced with other 6 Haider N, Osman AY, Gadzekpo A, et al. Lockdown measures in high-burden conditions (eg, cardiovascular and cerebro­ response to COVID-19 in nine sub-Saharan African countries. vascular disease).26 BMJ Glob Health 2020; 5: e003319. 1516 www.thelancet.com/oncology Vol 22 November 2021 Articles 7 COVIDSurg Collaborative. Elective surgery cancellations due to the 17 Mohanan M, Malani A, Krishnan K, Acharya A. Prevalence of COVID-19 pandemic: global predictive modelling to inform surgical SARS-CoV-2 in Karnataka, India. JAMA 2021; 325: 1001–03. recovery plans. Br J Surg 2020; 107: 1440–49. 18 Mwananyanda L, Gill CJ, MacLeod W, et al. COVID-19 deaths in 8 Nepogodiev D, Bhangu A, Glasbey JC, et al. Mortality and Africa: prospective systematic postmortem surveillance study. BMJ pulmonary complications in patients undergoing surgery with 2021; 372: n334. perioperative SARS-CoV-2 infection: an international cohort study. 19 Kruk ME, Ling EJ, Bitton A, et al. Building resilient health systems: Lancet 2020; 396: 27–38. a proposal for a resilience index. BMJ 2017; 357: j2323. 9 Burki TK. Cancer guidelines during the COVID-19 pandemic. 20 Glasbey JC, Bhangu A. Elective cancer surgery in COVID-19-free Lancet Oncol 2020; 21: 629–30. surgical pathways during the SARS-CoV-2 pandemic: 10 Nepogodiev D, Moore R, Biccard B, et al. Prioritizing research for an international, multicenter, comparative cohort study. J Clin Oncol patients requiring surgery in low- and middle-income countries. 2020; 39: 66–78. Br J Surg 2019; 106: e113–20. 21 Fitzmaurice C, Abate D, Abbasi N, et al. Global, regional, 11 Sullivan R, Alatise OI, Anderson BO, et al. Global cancer surgery: and national cancer incidence, mortality, years of life lost, years delivering safe, affordable, and timely cancer surgery. Lancet Oncol lived with disability, and disability-adjusted life-years for 29 cancer 2015; 16: 1193–224. groups, 1990 to 2017: a systematic analysis for the Global Burden of 12 Sud A, Jones ME, Broggio J, et al. Collateral damage: the impact on Disease study. JAMA Oncol 2019; 5: 1749–68. outcomes from cancer surgery of the COVID-19 pandemic. 22 Lee LY, Cazier JB, Angelis V, et al. COVID-19 mortality in patients Ann Oncol 2020; 31: 1065–74. with cancer on chemotherapy or other anticancer treatments: 13 Hanna TP, King WD, Thibodeau S, et al. Mortality due to cancer a prospective cohort study. Lancet 2020; 395: 1919–26. treatment delay: systematic review and meta-analysis. BMJ 2020; 23 Dietz JR, Moran MS, Isakoff SJ, et al. Recommendations for 371: m4087. prioritization, treatment, and triage of breast cancer patients during 14 Harris PA, Taylor R, Thielke R, Payne J, Gonzalez N, Conde JG. the COVID-19 pandemic. The COVID-19 pandemic breast cancer Research electronic data capture (REDCap)--a metadata-driven consortium. Breast Cancer Res Treat 2020; 181: 487–97. methodology and workflow process for providing translational 24 Farmer R, Mathur R, Bhaskaran K, Eastwood SV, Chaturvedi N, research informatics support. J Biomed Inform 2009; 42: 377–81. Smeeth L. Promises and pitfalls of electronic health record analysis. 15 GlobalSurg Collaborative and National Institute for Health Research Diabetologia 2018; 61: 1241–48. Global Health Research Unit on Global Surgery. Global variation in 25 Sud A, Torr B, Jones ME, et al. Effect of delays in the 2-week-wait postoperative mortality and complications after cancer surgery: cancer referral pathway during the COVID-19 pandemic on cancer a multicentre, prospective cohort study in 82 countries. Lancet 2021; survival in the UK: a modelling study. Lancet Oncol 2020; 21: 1035–44. 397: 387–97. 26 Kansagra AP, Goyal MS, Hamilton S, Albers GW. Collateral effect 16 Roser M, Ritchie H, Ortiz-Ospina E, Hasell J. Coronavirus of COVID-19 on stroke evaluation in the United States. Pandemic (COVID-19) OurWorldInData.org2020. https:// N Engl J Med 2020; 383: 400–01. ourworldindata.org/coronavirus (accessed Dec 11, 2020). v www.thelancet.com/oncology Vol 22 November 2021 1517