Data sources

SPOR was initiated by the Swedish Society for Anaesthesia and Intensive Care, set up in 2011, and started including data in 2011. All data is managed by Uppsala Clinical Research Centre, which runs many of the largest nationwide quality registers in Sweden. The purpose of SPOR is to provide data for national and local quality assurance projects and research. It is not used for billing purposes. Data are uploaded daily through smooth integration with local surgery planning and billing systems. A process for validation of data is ongoing since several years.

The Swedish Intensive Care Register (SIR) has a similar design and was initiated in 2001 by the same professional organization. The purpose of SIR is to promote and develop quality in Swedish intensive care. Both the above registers have Certification Level 1 and publish open access reports continuously on their respective websites.

Inclusion and exclusion criteria

Only children <16 years of age at the date of surgery were included, and centers that had not started reporting January 1^st 2019 or stopped reporting to SPOR after 2019 were excluded, to enable comparisons with pre-pandemic conditions. We decided early in the planning of the study to only include centers that reported data all four years. All hospitals connected to SPOR upload their data automatically to the register monthly. The uploading process has been validated for all data fields used in this study [8]. Data consist of patient ID, type of surgery, diagnoses, time stamps during the perioperative process (from the decision to operate to the time of discharge from the postoperative recovery area) as well as quality measures.

The year 2019 was defined as baseline for analyses of changes in caseload. The absolute number of cases as well as percentage changes relative to the corresponding baseline periods were reported.

Outcomes

The primary outcome was reductions in total number of procedures during the pandemic years of 2020, 2021 and 2022 compared to corresponding pre-pandemic numbers, i.e., during 2019. Secondary outcomes were differences in caseload reduction regarding 1) level of care (university, district, and county hospital, as indicated in SPOR; university hospitals are tertiary centers with 24–7 services for most pediatric surgical and medical services, district hospitals have no separate pediatric anesthesia or surgery but capacity to treat children 24–7, county hospitals have only core medical and surgical services and the term smaller units is reserved for private clinics typically focused on elective surgery in a single speciality), 2) surgical specialty according to the national surgical procedure code system KVA-97 [9], 3) emergency vs elective surgery, and 4) age at the time of surgery. For the primary outcome, we determined the weekly caseload and compared to the corresponding week in 2019 to study the evolution of the effects of the pandemic over time. To put these changes into context of the fluctuating intensity of the pandemic we retrieved the number of new admissions to all Swedish Intensive Care Units (ICU) from the Swedish ICU Register [10]. Furthermore, a sub-analysis of the changes during the four different waves of the pandemic was performed, comparing them to baseline numbers of 2019 and displaying differences in number of procedures per day including separate data on emergency and elective procedures.

Finally, out of a total of 3,276 different registered procedures we analyzed mean age at the time of surgery for the 20 most common procedures in 2019 for each of the different years of the study period, to provide more detail and probe for potentially delayed scheduling. Regional distribution as well as changes in waiting times during the pandemic were investigated.

Confounders

Possible confounders are, e.g., demographic changes such as increasing immigration and variations in nativity, the increasing cost of health care both before and during the pandemic, and possible changes in political policy. These were not accounted for.

Statistical methods

The dataset is a convenience sample based on data available at the time of study planning. Differences are displayed as absolute and percentage reductions. Poisson regression was used to analyze the main categories during the different years adjusted for risk population based on the population <16 years of age in Sweden during each year using data from Statistics Sweden [11]. Student’s t-test was used to investigate changes in patient age at surgery; differences displayed with 95% confidence intervals. A p-value of less than 0.05 was considered statistically significant. Data was managed and analyzed using Microsoft Excel (Version 2210 Build 16.0.15726.20188) and R version 4.1.1.

Emergency vs elective surgery

Elective surgery was severely affected while there was no decline in emergency surgery. There are several plausible explanations for elective surgery cancellations during the pandemic. For example, patients or health care workers having symptoms of Covid-19, personnel relocation or simply postponement of non-urgent surgery to focus on the pandemic. In addition, recommendations for social distancing may have led to reductions in common pediatric respiratory illnesses [12–15] which could possibly have affected the number of procedures. One Dutch review reported that the Covid-19 pandemic caused significant decrease in pediatric admissions and emergency department (ED) utilization in both the Netherlands and the rest of the world [16]. Furthermore, a German study found that ED consultation during each phase of the pandemic was lower than in pre-pandemic conditions, with larger reductions displayed for less urgent patient consultations and younger patients [17]. This is in line with a British study from the beginning of the pandemic displaying large reductions in ED-visits with the most marked reductions reported in school-age children [18]. A US study found a 42% decline in ED visits during the early pandemic compared to 2019 and found that this reduction was especially pronounced for children and females [19]. Regarding surgical volume during the pandemic a cohort study from a large hospital in Massachusetts found a decline of 44,6% in the early weeks of the pandemic, with the largest differences being displayed in laryngeal, plastic, oral maxillofacial and general surgery while emergent/urgent surgery was less severely affected [20]. This is in line with our results showing more marked reductions in the elective category. There are large spikes in elective surgery in Fig 1 during the last two weeks of 2020, 2021 and 2022. These outliers may be artifacts, explained by the customary low rate of elective surgery during the Christmas season.

ENT surgery most severely affected

Ear, nose and throat/oral surgery was the group most significantly affected. There could be several explanations for this. Firstly, ENT procedures in children can be postponed for months or even years without serious harm to the children. This is generally the case for, e.g., grommets insertion, tonsillectomy and adenoidectomy, all of which are on the list of the 20 most common procedures (Table 3). However, some children will suffer harm from the consequences of postponing both grommets insertion and tonsil surgery, respectively.

Tonsillectomy has two main indications: sleep-disordered breathing (SBD) and recurrent tonsillitis. In a meta-analysis, SDB was associated with later subsequent development of neurobehavioral deficits (pooled RR 3.24, 95%CI 1.25–8.41) [21]. Furthermore, it has been shown that children with SDB have higher antibiotic usage rates, more hospital visits and more healthcare visits due to upper respiratory infections [22]. The benefits of surgical treatment due to the latter indication are less clear. According to a Cochrane review from 2014, adenoid-/tonsillectomy leads to a reduction in the number of episodes of sore throat and days with sore throat in children in the first year after surgery compared to (initial) non-surgical treatment, although the size of effect of surgery was described as “modest” [23]. Secondly, aerosol-producing otolaryngologic procedures were identified as high-risk early in the pandemic due to the high viral load in the oropharyngeal cavity and adjacent structures, and otolaryngologists were among the first physician casualties [24]. This coupled with the fact that children are more prone to respiratory infections which can be hard to distinguish from a Covid-19 infection made surgical planning difficult and elective procedures in children with upper respiratory tract infections (even when Covid-19 was not the culprit) were therefore postponed at least 7 weeks during the pandemic [25].Thirdly, a large quantity of the surgeries are procedures on the tonsils, where infection is one indication for surgery and infections of the upper respiratory tract have been less common during the pandemic because of social distancing, as shown by the Swedish National Board of Health and Welfare and the Public Health Agency of Sweden [14,15].

Potential costs and benefits of postponing surgery

Postponing surgery may be problematic for several reasons, such as advancement of illness or increased burden of disease. Especially relevant in children is a window of opportunity being lost, e.g., for cochlear implants, orchidopexy or hypospadias surgery. Fortunately, we found no convincing evidence of significantly higher age at surgery for the sum of the 20 most common procedures except for grommets and tonsillectomy (Table 3). The risk of postponing grommets in children is chronic and recurrent ear infections with fluid in the middle ear leading to hearing loss that in turn affects the child’s speech and language development [26]. In line with the discussion above about the consequences of postponing the common ENT procedures, surgery for undescended testis should ideally be performed before the age of 1 to avoid increased risk of infertility and malignancy, although the evidence for the latter is a contentious issue [27]. While the overall age at operation for, e.g., undescended or ectopic testis did not increase significantly, it is likely that many of these operations were postponed beyond the age of 12 months in some if not all centers.

However, potential positive effects have been suggested by Gelardi et al. who reported an improvement in symptomatology for children waiting for adenoidectomy, hinting at a reduced need for surgery during the pandemic [28]. Similarly, Marom et al. reported a reduced incidence in acute otitis media episodes during the pandemic, with high spontaneous resolution rates in children associated with the reduction in grommets insertion procedures [29].

The uneven burdens of a backlog

Cancellations during the pandemic add burden to an already existing health care debt. This might be especially true for procedures with lower priority. This is illustrated in Fig 3, showing waiting times of several years for adenoidectomy in some regions. The regional differences are striking: the waiting times of > 1.5 years ranges from 0 to 51%, in a few regions reaching more than 3 years. We could speculate that the pandemic has aggravated regional issues with unequal access to health care. Moreover, university hospitals were less affected than the smaller district and county hospitals. This could be because of centralization of more complex pediatric cases to university hospitals, while district/county hospitals mainly deal with basic ENT surgery. For example, procedures on neonates and infants are mostly handled by university hospitals in Sweden, and university hospitals display a higher mortality than county and district hospitals [30]. These more complex cases could be less prone to cancellation or postponement.

Health care workers all over the world are still dealing with this post-pandemic backlog while at the same time having to recover from the stress of the pandemic and worker fatigue [31]. Burnout was a problem for healthcare workers even before the pandemic [32] and the increased workload and stress of the pandemic may have added new social and work-related factors that increase risk of burnout. These problems were aggravated in developing countries where vaccination programs had not yet reached the majority of the population. However, these challenges may trigger initiatives to increase workflow [33,34] and reduce late cancellation rates by more rigorous pre-operative screening protocols [12]. Jiang and Carvalho showed an increased OR-efficiency during the pandemic, with lower cancellation rates due to “no show” or “family refusal” [12]. They stipulated mandatory pre-operative Covid-tests causing increased nursing communications and education as a possible explanation. Future studies will be needed to understand the long-term consequences of three years of backlog of more than 10% of the annual number of procedures, possibly adding on to known issues of hospital staffing shortages causing staff relocations, external recruitments, added hours and lost hours due to illnesses, in Sweden during the pandemic [35].

Strengths and limitations

The strength of the present study is that it provides data from a national perioperative database including almost all major Swedish hospitals with pediatric anesthesia. However, some important limitations are noted. The study is retrospective and it was not possible for us to adjust for all possible confounders such as changes in demographics, the increasing cost of health care, etc. This limits the ability to draw conclusions regarding cause-effect. Another limitation is that while presently (2025) SPOR covers 100% of Swedish government-run hospitals a few private clinics perform a small number of pediatric cases without reporting their data to SPOR. It is likely that the private clinics were at least as severely affected by the pandemic as government hospitals but their relative contribution to the results would be negligible since their estimated total caseload is a few percent of the total number of pediatric procedures in Sweden. Moreover, the single year of 2019 serves as baseline in the analysis of the volume for the three following years of the pandemic. The reason for using only 2019 as the baseline year was that this was the first year in which almost all Swedish Hospitals had joined the register. Ideally, we would have had access to complete baseline data from at least three years but since the register was still lacking data from several major centers, inclusion of data reported earlier than 2019 would have been less reliable. Theoretically, the production volume could have been unusually high during 2019, causing an incorrect analysis of reduced volume, but we have no reason to believe that this is the case.

Furthermore, we identified two small hospitals with a total caseload of < 1000 cases per year that stopped their reporting after 2021 and were therefore excluded. We have no information if they stopped reporting due to the pandemic per se or because of increased centralization or re-organization of patient flow during these years. An additional seven small hospitals were excluded since they joined SPOR after 2019 which precluded comparisons with baseline conditions before the pandemic. Finally, the pandemic was not fully contained within the study period, but our data indicates that caseload was returning to base-line numbers at the end of 2022.