Abstract
Perioperative hypothermia is associated with increased morbidity and mortality. Systematic temperature monitoring is considered important for early detection of hypothermia and timely intervention, yet incidence remains high. This study examined monitoring practices across the perioperative course. We conducted a 1-week retrospective observational study of 1261 patients undergoing surgery lasting more than 30 min. Temperature monitoring data, including hypothermia incidence and the use of monitoring devices, were extracted from electronic records. Only 11.1% of patients had complete temperature documentation across all stages, and hypothermia was frequently recorded. Preoperative temperature measurements were documented in 31.8% of patients with an average 7 h delay between ward measurement and arrival in the operating theatre, potentially limiting clinical utility of these measurements. Intraoperative monitoring occurred in 26.1% of patients and in 29.0% at the post-anesthesia care unit. We found intraoperative hypothermia in 38% of patients. The most commonly used device preoperatively was an infrared aural canal thermometer, and in the operating theatre, the most common device was urinary catheter-based invasive temperature monitoring. These findings indicate suboptimal continuity of temperature monitoring across the perioperative pathway. Our findings highlight opportunities to improve the consistency of perioperative temperature monitoring and warrant further research into factors influencing current practice.
Introduction
Ensuring patient safety in the operating room includes prevention of perioperative hypothermia, defined as a core body temperature below 36.0°C.1–4 Multiple factors contribute to unintentional hypothermia during surgery, including redistribution of body heat caused by anesthesia, prolonged exposure of skin and body cavities, extended procedure durations, age, and low ambient temperatures.4–6 Maintaining normothermia is an essential nursing responsibility and is fundamental to physiological stability.7–9 Temperature fluctuations are associated with adverse outcomes such as myocardial events, surgical site infections, delayed wound healing, increased intraoperative blood loss, patient discomfort, and prolonged length of stay in the post-anesthesia care unit (PACU).4–6
Despite widespread awareness of preventive measures, perioperative hypothermia persists as a serious clinical concern, with incidence rates reported between 25% and 90%.10–13 In Norway, rates are similarly high with incidences of 32.5% and 24% reported in recent studies.14,15 Nurses play a key role in perioperative temperature monitoring and documentation, 8 and their adherence to recommended practices is essential to guide and evaluate treatment, maintaining normothermia, and ensuring patient safety.1,16 Failure to implement monitoring and active warming delays detection and treatment of hypothermia and may contribute to perioperative hypothermia and related complications.17–19
The National Institute for Health and Care Excellence (NICE) guidelines recommend strict temperature monitoring throughout the perioperative period, including measurement prior to transfer from the ward, before anesthetic induction, every 30 min intraoperatively, every 15 min in the PACU, and at least every 4 h postoperatively on the ward. 1 When active warming is required, monitoring should occur every 30 min.1,5 Active thermal management before, during, and after surgery is consistently shown to reduce risk of hypothermia.18,20,21 NICE defines active warming as a process that transfers heat to the patient. 1 Nevertheless, temperature monitoring often receives insufficient attention.5,17 An Australian study of perioperative temperature monitoring across five hospitals found that over half of surgical patients received two or fewer temperature assessments perioperatively, while one-third had no documented temperature recorded before anesthetic induction. 17 Implementation of temperature management, including continuous monitoring and preventive strategies, remains inconsistent across healthcare systems.6,8,22
Few studies have examined the quality and consistency of temperature monitoring across the entire perioperative pathway.10,14,17 Strengthening knowledge regarding temperature assessment practices is therefore critical to improving clinical outcomes and ensuring standardized implementation of preventive strategies.
Aim
The primary aim of this study was to evaluate perioperative temperature monitoring practices across all stages of the perioperative pathway in Norwegian surgical patients. Secondary aims were to assess the prevalence of perioperative hypothermia and the use of specific temperature-monitoring devices.
Method
This study employed an observational design, utilizing electronic patient records to estimate the proportion of patients in a 1-week surgical sample who underwent perioperative temperature monitoring. Patient record data provide valuable real-world insights by allowing researchers to analyze and cross-reference information across multiple hospitals, thereby enhancing the breadth and depth of the analysis.23,24 The study used a protocol from an Australian study by Munday and colleagues 17 published in 2023, which we adapted to the Norwegian context. The study is reported in accordance with the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) statement. 25
Study settings
Participating hospitals included three hospitals in Southeast Norway: one large university health service (located at four hospital campuses) and two regional health services (one with two hospital campuses, the other with three hospital campuses). See Figure 1 for an illustration of the Norwegian specialist health care system. The four university hospital campuses comprise approximately 100 operating rooms and perform about 55,000 operations annually, 29% of which are classified as emergencies. The two regional hospitals across five campuses have approximately 25 and 30 operating rooms, respectively, with annual surgical volumes of approximately 17,000 and 20,000 procedures.
Norwegian specialist healthcare system.
The study population represented a diverse mix of surgical specialties, and day surgeries, elective procedures, and emergency procedures were included.
Data collection
Retrospective data extraction of historical cases from hospital patients’ electronic health records was conducted manually by three researchers and recorded in a predefined digital format (SurveyXact™). Patients at the regional hospitals and 14 of 16 university hospital departments who underwent procedures of more than 30 min duration under any type of anesthesia between 3 and 9 February 2020 were included. (Two departments were not registered owing to organizational issues). Historical data were extracted from the same week as Munday and colleagues 17 to facilitate comparison of the findings. Data extraction took place between 2 June 2022 and 22 April 2024. In total, 1261 of 1434 procedures were included and analyzed. The process was approved by hospital administrations and carried out in full compliance with ethical regulations.
Patient characteristics included age, sex, body mass index (BMI), American Society of Anesthesiologists (ASA) physical status classification, type of admission (elective or emergency), surgical specialty, and duration of anesthesia (minutes). The primary outcome was documentation of temperature monitoring at predefined perioperative time points: (a) preoperative ward prior to transfer to the operative department; (b) first temperature recorded after arriving in the operating theatre; (c) intraoperatively during surgery; (d) first temperature recorded after arriving in the PACU; and (e) first temperature recorded on the surgical ward. The selection of monitoring intervals followed the NICE recommendations 1 and the Australian College of Perioperative Nurses Standard for Managing Hypothermia in the Perioperative Environment 26 (Table 1). To be able to answer the secondary aim of this study, we collected data from patient records on temperature measurement at all perioperative stages, and the usage of devices for temperature measurement. In addition, we also collected data on recorded documentation of active warming.
Overview of sample (n = 1261).
American Society of Anesthesiologists (ASA).
On “other modes of anesthesia,” the most prevalent reported modes were regional blockades (n = 42/164 (25.6%)) and sedation (n = 11/164 (6.7%)).
Gastrointestinal, gynecology, and obstetric surgery.
Ear/nose/throat and jaw.
Data analysis
Variables were summarized using descriptive statistics, reported as number (n) and percentage (%), as mean with standard deviation (SD), or median with interquartile range (IQR), as appropriate based on distribution. Chi-square tests were performed to describe differences between hospitals and to assess temperature monitoring between pediatric and older patients. Missing values were not imputed; all analyses were conducted using complete-case data for the variables included in each comparison. A two-sided p value <0.05 was considered statistically significant. Statistical analyses were performed using SPSS versions 29 and 30.
Ethical approval
All data were managed in accordance with the principles of the World Medical Association Declaration of Helsinki. 27 Ethical approval, including a waiver of patient consent, was granted by the Norwegian Regional Committees for Medical and Health Research Ethics 28 (application no. 347745), with special permission to include pediatric patients. The study was registered with the Norwegian Centre for Research Data (NSD) (project no. 624423). NSD is the national authority responsible for research supervision and data protection. 29
Additionally, the Faculty Research Ethics Committee (FEK) at the Faculty of Health and Sports Sciences, University of Agder, Norway, confirmed that all necessary approvals were secured for implementation of the study (ethical clearance no. RITM0220661). Furthermore, data protection officers at the participating hospitals established a data transfer agreement specifying each institution's responsibilities to ensure compliance with the granted approvals.
Result
Of the identified 1261 patients, 739 (58.6%) were from the university hospitals, while the regional hospitals contributed 214 (17.0%) and 308 (24.4%) patients, respectively. Patients were divided into three age groups (<2 years, 2–12 years, and >12 years) based on the American Academy of Pediatrics age classification. 30 Among all patients, 35 (2.8%) patients were younger than 2 years, 71 (5.6%) patients were between 2 and 12 years, and 1155 (91.6%) were older than 12 years. Most procedures were classified as ASA II 536 (47.4%) or ASA I 313 (27.7%), and 934 (76.6%) were elective admissions. Orthopedic and abdominal surgeries were the most frequent specialties, each accounting for approximately one-third of the sample (Table 1).
The university hospital treated a patient population characterized by younger age, lower BMI, and higher ASA classification compared to the regional hospitals (Table 1). The university hospital also reported a greater proportion of emergency admissions, longer operative durations, and higher use of general anesthesia. Conversely, regional hospitals used spinal and other forms of anesthesia more frequently. Regarding surgical specialties, the university hospital performed fewer orthopedic procedures but a larger number of neurosurgical and cardiothoracic operations relative to the regional hospitals.
Temperature monitoring
Out of 1261 patients, 401 (31.8%) had temperature measurements documented in the surgical ward prior to surgery (Table 2). Of these, 21 (5.2%) patients had a temperature recorded within 1 h before transferring to the operating theatre. The mean (SD) time lag between registered temperature measurements in the ward and arrival at the operating theatre was 413 (405) min. No significant differences were identified in the frequency of preoperative temperature monitoring between pediatric patients (0–2 and 2–12 years) and the rest of the sample (χ2 (2, n = 1258) = 3.204, p = 0.202).
Registered temperature monitoring at preoperative, intraoperative and postoperative timepoints (n = 1261).
Median (25%–75%) = 195 (110–710) min.
Umbrella term for all documented forms of active heating (e.g. hot air blankets, heated beds, and heated operating tables).
In addition, 36 (2.8%) cases warm gas in laparoscopic procedures were applied.
Post-anesthesia care unit.
Median (25%–75%) = 30 (16–88) min.
After arrival in the operating theatre (Table 2), temperature was recorded for 329/1261 (26.1%) patients. Continuous intraoperative temperature monitoring was documented for 283/329 (86.0%) of these patients. Intraoperative temperature measurements were performed in 24/35 (68.6%) children under 2 years of age, 31/71 (43.6%) children aged 2–12 years, and 274/1155 (23.4%) patients older than 12 years. Intraoperative temperature measurements were more frequent in pediatric patients (<12 years) than in older patients (>12 years); this difference was statistically significant (χ2 (1, n = 1258) = 52.571, p < 0.001).
In PACU (Table 2), temperature was measured for 365/1261 (28.9%) patients. The mean interval from the end of anesthesia to measurement was 80 min (SD 119 min). No statistically significant difference was observed between pediatric patients (0–12 years) and older patients (χ2 (1, n = 1258) = 1.423, p = 0.233).
Active warming was administered to 3/1261 (0.2%) patients before arrival in the operating theatre. During surgery, 620/1261 (49.2%) patients received active warming, and 215/1261 (16.5%) patients in the PACU. Active warming on the surgical ward was not recorded (Table 2). The most frequently used device for temperature measurement on the ward, both preoperatively and postoperatively, was the infrared aural canal thermometer. In the operating theatre and PACU, the urinary catheter with integrated bladder thermometer was most used (Table 3).
Devices used in temperature measurement through the perioperative stages.
Post-anesthesia care unit.
With integrated bladder thermometer.
Endotracheal, gastric, or nasal/nasopharyngeal tube.
Incidence of hypothermia
Among patients with temperature measurements in the surgical ward, 64/401 (16.0%) patients were hypothermic. Intraoperatively, 125/329 (38.0%) patients with temperature measurements were hypothermic (Table 2). Nineteen of 216 (8.8%) patients were hypothermic both preoperatively on the ward and intraoperatively. In PACU, 52 of the 365 patients (14.2%) with documented temperature measurements were hypothermic; 54 of 504 (10.7%) remained hypothermic on the surgical ward postoperatively.
Temperature measurements
Temperature was recorded both on the ward and after arrival in the operating theatre for 219/1261 (17.4%) patients (Table 4), and 149/1261 (11.8%) patients were monitored at all stages including the PACU. Overall, 140/1261 (11.1%) patients had temperature recorded consistently throughout the entire perioperative course (Table 4).
Temperature measurement through perioperative stages (n = 1261).
Temperature measurements both preoperatively and after arriving in the operating theatre.
Temperature measurements both preoperatively and after arriving in the operating theatre. intraoperative.
Post-anesthesia care unit.
Temperature measurements both preoperatively and after arriving in the operating theatre, intraoperative, and PACU.
Temperature measurements at all perioperative stages (including the surgical ward).
Discussion
In this multicenter retrospective observational study, we found substantial gaps in nurses’ perioperative temperature monitoring and hypothermia prevention. Complete documentation across the full pathway was present in only one in 10 patients. Hypothermia occurred at all stages, affecting over one-third of patients intraoperatively and approximately one in seven on arrival in the PACU. These findings suggest that temperature monitoring appeared less consistently documented across the perioperative pathway than recommended by current guidelines.
In this study, only about one-third of patients had their temperature recorded preoperatively on the ward, a minority was monitored 1 h before arriving at the operating theatre, and one-quarter had their temperature recorded after arrival. Only 11.1% of patients had complete temperature documentation throughout the perioperative pathway, highlighting a persistent inconsistency in perioperative temperature monitoring, reflecting previously reported trends.5,20,31 Incomplete temperature documentation may impair the ability to track trends and examine preventive interventions. This may contribute to an increased risk of undetected hypothermia and may create uncertainty about the need for active warming. It could also complicate clinical decision-making, the follow-up, and evaluation of warming interventions, potentially compromising the quality of care and patient safety. This emphasizes the key role of nurses in ensuring systematic temperature monitoring in perioperative care. 6 Furthermore, the observed differences may reflect variation in local practice, implementation processes, adherence to guidelines, or organizational procedures.32,33
According to research, barriers to guideline implementation often arise from a combination of individual, organizational, and contextual factors.20,34 Studies outline barriers to guideline implementation as limited awareness, high workload, time constraints, resource constraints, and weak implementation strategies.32,33 Organizational conditions might play an important role in whether guidelines are successfully translated into practice.32,33 Weak leadership support, along with inadequate training and skills development, is likely to exacerbate these challenges by limiting prioritization, accountability, and integration of guidelines into routine clinical practice.32,33 In the preoperative setting, the low monitoring rate may be due to an underestimation of the importance of temperature monitoring in preventing perioperative hypothermia.31,35 This suggests the need for implementation-focused interventions and future research that addresses organizational and behavioral factors that influence perioperative temperature management. 35
Intraoperative monitoring rates remained low throughout surgery. Notably, the nearly 7 h gap between the ward measurement and arrival in the operating theatre offers nurses limited insight into the patient's temperature at the time of anesthesia induction. The handover from the ward to the operating theatre represents a vulnerable point in the perioperative care process, where responsibilities, information, and clinical judgment are transferred between departments, increasing the risk of loss of clinically relevant data. 36
A delay is a relevant clinical concern, as patients are at increased risk of hypothermia in the preoperative period due to exposure to ambient temperature, fasting, and limited thermal protection during transport and waiting times.4,6,17 Temperature can change rapidly, and reliance on outdated information can lead to clinical decisions based on incomplete data and delay preventive or corrective measures. This in turn may affect the assessment of the need for active warming and increase the risk of perioperative hypothermia.1,6,37
Temperature monitoring in PACU was also inadequate, with fewer than one-third of patients undergoing measurement and an average delay approaching 2 h. This contrasts with the findings of Munday and colleagues, 17 where temperature was measured in over 80% of patients upon arrival at the PACU. Comparable deficiencies in temperature documentation are reported elsewhere.17,38 The low frequency of temperature measurements and the long delay before monitoring in the PACU indicate a notable gap between recommended and actual practice. Inadequate monitoring in the PACU may reflect limited attention to the identification of hypothermia in favor of other vital signs, with potential consequences such as patient discomfort, delayed implementation of warming measures, and an increased risk of postoperative complications.6,22
In the included hospitals, temperatures were entered into different sections of the electronic medical records, leading to inconsistent documentation across perioperative units. This could potentially make it difficult to discover important information and impair clinicians’ ability to make timely and informed decisions regarding temperature management. 31 In Norway there are no national standards for perioperative temperature monitoring; however, we believe that nurses are likely to be familiar with international guidelines. This finding suggests that adherence to guidelines is shaped by more than healthcare professionals’ knowledge of recommendations. Workload and competing priorities may lead to variation in practice and inconsistent integration of temperature monitoring into routine care.39,40
Although temperature assessment is typically included in standard admission procedures, considerable variation in nurses’ recording practices suggests a gap between guideline recommendations and clinical practice. Despite Norwegian specialist nurses having a high level of education and receiving comprehensive training in hypothermia risks and prevention, 41 we identified considerable gaps in temperature monitoring. Consistent with previous reports,38,42 these findings suggest that substantial professional knowledge does not necessarily translate into consistent clinical practice. While clinical judgment remains essential, contextual factors such as time constraints and a strong emphasis on operating theatre efficiency may contribute to the deprioritization of tasks perceived as non-essential. 34 Addressing this lack of consistency provides a clear opportunity for improvement of clinical practice, including examination of institutional processes, procedures for perioperative temperature measurement, and the implementation of systematic temperature monitoring practices.20,21
Preoperative temperature monitoring did not differ significantly between pediatric patients (<12 years) and older patients (>12 years). Intraoperative monitoring was more frequent, particularly in those younger than 2 years, reflecting their increased risk of perioperative hypothermia. 43 However, the limited number of pediatric patients in the dataset warrants caution when interpreting these findings.
Considerable variation was observed in the devices used to record temperature throughout the perioperative period. Noninvasive measurement techniques were routinely used on the ward, which is reasonable given their practical utility and minimal patient discomfort. Infrared aural canal thermometry was used in approximately two-thirds of patients preoperatively. This method has been reported to be susceptible to measurement inaccuracies due to imprecise probe positioning. 44 Additionally, patients with unstable circulation may exhibit falsely low temperature readings using this method. 44 In the operating theatre, urinary catheter-based invasive temperature monitoring was used in 85% of patients, often initiated after induction of anesthesia. This approach likely reflects nurses’ efforts to minimize patient discomfort; the method may be influenced by physiological conditions that can reduce its accuracy.44,45 Besides, the different devices are not calibrated against each other. It remains unclear whether nurses consider how device positioning and physiological factors may contribute to variability in temperature measurements. Greater standardization of measurement devices may improve both measurement accuracy and comparability across perioperative settings.
These methodological differences may have reduced the comparability of measurements and increased variability. 44 Consequently, there is a risk that changes in patients’ temperature are not accurately captured, potentially leading to imprecise baselines for clinical assessment. Additionally, it might also prevent early detection of hypothermia, with subsequent rapid warming measures. Moreover, differences in local departmental practices may lead to variation in protocol adherence, with compliance depending on individual nurses rather than being consistent across settings.34,42
Given the critical role that temperature measurement plays in identifying high-risk patients and guiding warming strategies, the continued inadequacy of temperature monitoring is concerning. Ongoing discussions in Norwegian hospitals are focused on determining the timing and methods for temperature measurement, yet differences in device availability and team responsibilities pose barriers to standardization. The importance of accessible devices and resources has been emphasized as a key factor influencing effective temperature monitoring.31,45
The prevalence of intraoperative hypothermia of 38% in this study is consistent with rates reported in comparable research.14,17,46 However, the actual prevalence remains highly uncertain due to the limited frequency of registered temperature measurements across the perioperative period. Furthermore, if device-related differences lead to an underestimation of hypothermia incidence, the condition may receive less clinical attention, increasing the likelihood that clinically important hypothermia remains unrecognized. Temperature monitoring is essential for timely intervention; uncertain, delayed, or absent parameters may lead to under-detection and under-treatment of hypothermia. Consequently, preventive warming strategies may not always be initiated or appropriately evaluated. The findings suggest that the observed gaps are unlikely to be explained by a lack of knowledge alone, given the availability of evidence-based guidelines and specialist nurses trained in hypothermia prevention. This is consistent with implementation research showing that local practice and organizational factors shape adherence to guidelines.32,33 The inconsistent implementation of temperature monitoring in clinical practice may explain why perioperative hypothermia persists despite clinician knowledge of preventive measures.17,20
Active warming was documented intraoperatively in only about half the patients (Table 2), which is concerning given the established benefits of this intervention and inclusion in evidence-based guidelines. 1 However, rates of active warming were higher compared to recent Australian data. 17 The results could indicate that, in some cases, active warming was implemented without knowledge of the initial temperature and without proper evaluation (Table 2). Consequently, the reported warming effects remain uncertain, with the risk of inadequate response or hyperthermia. Furthermore, a substantial proportion of hypothermic patients did not receive active intraoperative warming, which underscores the importance of improving compliance with guidelines. NICE guidelines recommend temperature monitoring on admission to the intensive care unit and continued active warming in hypothermic patients until normothermia is achieved.1,22 Delayed measurement may prevent timely postoperative intervention. The need for systematic temperature monitoring and active warming could be seen as interlinked components of perioperative management, requiring not only individual professional knowledge but also organizational support, clear protocols and consistent implementation.
Strengths and limitations
A key strength of this study is the inclusion of a large, heterogeneous sample from multiple hospitals, covering a wide range of surgical specialties and both elective and emergency procedures. The use of a standardized protocol adapted from previous work supports comparability with international studies. Furthermore, by extracting data from clinical records, we were able to examine temperature monitoring across the full perioperative trajectory rather than focusing on a single unit.
Data collection relied on patient records; as a result, undocumented temperature measurements were not captured, possibly distorting reported monitoring rates. The lack of direct clinical observation precluded detailed assessment of monitoring behaviors and prevented us from identifying the reasons underlying observed practice patterns. Undocumented types of temperature measurement devices restricted assessment of their influence on monitoring accuracy. This observational study cannot establish causal relationships between incomplete temperature documentation and clinical outcomes, nor can it explain how interventions are implemented in practice. However, our relatively large dataset reveals gaps in temperature monitoring across perioperative stages, highlighting the urgent need for better adherence to guideline-based hypothermia prevention.
Broader and more contemporary datasets are needed for more generalizable conclusions. Nevertheless, the sample is considered representative, given its conduct over one week and coverage of nearly all operations at three major Norwegian hospitals. The data were not confounded by the COVID-19 pandemic.
Conclusion
Our findings reveal substantial inconsistencies in temperature monitoring across perioperative care stages, highlighting the importance of stricter adherence to temperature measurement protocols in surgical environments. The average interval of nearly 7 h between ward-based and operating room temperature measurements highlights the limited value of preoperative assessments for informing patient status at anesthesia induction. Our findings suggest that hypothermia may receive less consistent monitoring than recommended in current perioperative guidelines. The high level of inconsistency we observed reveals important gaps in both the documentation and management of perioperative temperature and therefore may have important implications for patient safety. Moreover, active warming protocols remain inconsistently implemented, with approximately one-quarter of hypothermic patients not receiving recommended intraoperative warming interventions.
Implications
The persistent undermonitoring of perioperative hypothermia signals that inconsistent temperature recording is a systemic gap that may have consequences for patient safety. The study highlights the ongoing need for standardized temperature measurement practices across the perioperative care continuum. Variations in documentation practices and diverse temperature measurement techniques complicate efforts to effectively prevent hypothermia. Understanding the factors that influence temperature monitoring practices remains an area for future research.
Footnotes
Acknowledgments
The authors wish to acknowledge additional members of the Australian-Scandinavian research collaborative including Ingrid Gustafsson, Andreas Nilsson, and Clint Douglas.
Ethical considerations
The study was conducted in accordance with the World Medical Association Declaration of Helsinki.
Ethical approval was granted by the Norwegian Regional Committees for Medical and Health Research Ethics (REK) (application no. 347745, approval 30 June 2022). The study was also registered with the Norwegian Centre for Research Data (NSD) (project no. 624423, date of approval 10 January 2022. Faculty Research Ethics Committee (FEK) at the Faculty of Health and Sports Sciences, University of Agder, Norway, confirmed that all required approvals were in place for the study.
Consent to participate
REK approved a waiver of patient consent.
Consent for publication
Not applicable.
Author contributions
Conceptualization: L.I.M.H. and J.M. Methodology: C.H., A.B.W., L.I.M.H., J.M., and T.-I.K. Formal analysis: T.-I.K. Investigation: L.I.M.H. Data curation: C.H., A.B.W., L.I.M.H., and T.-I.K. Writing – original draft preparation: C.H. and A.B.W. Writing – review & editing: C.H., A.B.W., L.I.M.H., J.M., and T.-I.K. Visualization: T.-I.K. Supervision: T.-I.K. Project administration, L.I.M.H.
Funding
The authors received no financial support for the research, authorship, and/or publication of this article.
Declaration of conflicting interests
The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Data availability statement
Relevant and anonymized data can be accessed from Sikt–Norwegian Agency for Shared Services in Education and Research (Reference to be added after acceptance).