Development of a risk prediction model for delirium following surgery for oral squamous cell cancer

Development of a risk prediction model for delirium following surgery for oral squamous cell cancer | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Research Article Development of a risk prediction model for delirium following surgery for oral squamous cell cancer Matthias Jakob Posch, Johannes Kalbhenn, Stefan Schlager, Damian Sommer, and 6 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-7687245/v1 This work is licensed under a CC BY 4.0 License Status: Under Review Version 1 posted 9 You are reading this latest preprint version Abstract Oral squamous cell carcinoma (OSCC) is a common malignancy that is commonly treated by surgical resection and - if necessary - with adjuvant radio(chemo)therapy. Depending on pre-existing patient related conditions and extent of surgery, postoperative intensive care may be indicated. Postoperative delirium (POD) is a frequent and serious complication in the intensive care unit (ICU), characterized by confusion and cognitive disturbances of the patient. This single-centre, retrospective study (2013–2023, University of Freiburg) analysed 299 OSCC patients admitted to ICU post-surgery to identify risk factors and develop a risk assessment tool for POD. A bias-reducing automated variable selection method identified ICU stay length, the use of percutaneous endoscopic gastrostomy (PEG) tubes, age, cancer stage, use of psychotropic and antihypertensive drugs to be associated with POD. Key predictors were used to create an online assessment tool, which predicted POD with about 85% accuracy. Notably, longer ICU stays increased POD risk, while the use of PEG tubes appeared to have a protective effect. Prospective studies are needed to validate the tool and to assess its clinical practicability. Postoperative delirium prediction model oral cancer risk factors Introduction Oral squamous cell cancer (OSCC) is among the most frequent malignancies and causes more than 180,000 deaths per year worldwide. OSCC mostly occurs in patients of advanced age and substantial alcohol and tobacco consumption histories (Bray et al., 2024 ). In most stages of the disease, surgical resection in combination with neck dissection and often primary reconstruction with microvascular free flaps, is crucial to achieve the best possible functional and oncological outcome (Chamoli et al., 2021 ). The extent of neck dissection is determined by the radiologically and clinically detected lymph node involvement and by the tumor location. Depending on the patient’s condition and the extent of the surgical procedure, admission to the intensive care unit (ICU) is required to ensure optimal perioperative care and to avoid complications. Postoperative delirium (POD) is a common complication among such patients at the ICU, with substantial medical and economic relevance (Lander et al., 2025 ; Sahle et al., 2022 ; Witlox et al., 2010 ). POD is an acute, fluctuating neurocognitive disorder characterized by acute confusion, hallucinations, disorientation, inappropriate behaviour and other symptoms. Depending on the clinical presentation, delirium can be categorised as hyperactive, hypoactive, or mixed (Dilmen et al., 2024 ). Validated tools such as the Diagnostic and Statistical Manual of Mental Disorders (DSM) 5 criteria, the Confusion Assessment Method and the Nursing Delirium Screening Scale are recommended for diagnosing delirium (Aldecoa et al., 2024 ). It is worth noting that high incidences of POD have been reported in patients receiving surgery for OSCC, (Hasegawa et al., 2015 ; Makiguchi et al., 2018 ; Takahashi et al., 2021 ; Ying et al., 2025 ). Treatment of POD is challenging and therefore individualized preventive measures based on risk factors are highly recommended (Aldecoa et al., 2024 ). Several studies investigated risk factors for POD following surgery for OSCC and revealed a high heterogeneity with regard to the identified risk factors (Hasegawa et al., 2015 ; Makiguchi et al., 2018 ; Obermeier et al., 2022 ; Shiiba et al., 2009 ; Takahashi et al., 2021 ; Ying et al., 2025 ). Next to the multifactorial aetiology, limited sample size and the lack of adjustment for risk bias may contribute to the inhomogeneous results observed regarding POD following surgery for OSCC (Steyerberg and Van Calster, 2020 ; Vach, 2013 ). Therefore, in this study we used an automatic bias reducing variable selection method to avoid potential confounding caused by pre-selection of factors. Identified risk factors should be used to establish a structured risk assessment for POD in daily clinical practice. However, this appears to be challenging even in modern health care systems (Falegnami et al., 2021 ). Thus, our study had two objectives. Firstly, the identification of risk factors following surgery for OSCC. Secondly, the development of a freely available online risk assessment tool that would allow healthcare personnel to evaluate an individual's risk of POD based on the identified risk factors. Materials & Methods In this retrospective study, electronic patient charts of patients were reviewed in order to identify risk factors for POD and generate an according risk assessment tool starting January 1st, 2013 and ending December 31st, 2020 at the Department of Oral and Maxillofacial Surgery, Medical Center, University of Freiburg, Germany. Surgical and perioperative treatment was conducted following the German guidelines for the treatment of oral cavity cancer (Wolff et al., 2021 ). No intervention was performed due to study implementation. Ethical clearance The study was approved by the local ethics committee of the University of Freiburg, Germany (No. 127/15 and No. 30/20). Informed consent was waived due to the retrospective nature of this study. The study was registered in the German Clinical Trials Register (Deutsches Register klinischer Studien, DRKS) No. DRKS00038043. Inclusion & exclusion criteria All patients that underwent primary resection of OSCC and who were scheduled for intensive care unit as part of our standard postoperative care in the time between January 1st, 2013 and ending December 31st, 2020 were included. No secondary exclusion of patients due to specific surgical reconstruction procedures was performed. All included patients were of legal age, revealed cancer-free resection margins (R0) after histopathological analysis and were administered to the ICU postoperatively. A postoperative follow-up of at least 12 months had to be ensured and consistent documentation of the postoperative ICU and in-ward stay had to be available. Patients with malignancies other than OSCC, secondary surgery for recurrent OSCC, evidence of distant metastases at initial diagnosis, and neo-adjuvant therapy prior to surgery were excluded. Variables examined The primary outcome variable was the development of postoperative delirium (POD). POD was diagnosed either during the stay in the ICU or after transferring the patient back to the oral and maxillofacial ward using the Nurse Delirium Screening Scale (NU-DESC). To be diagnosed with postoperative delirium a score of 2 or higher had to be clearly documented in the patient file. These patients were classified into Group 1. Group 2 was composed of the control group with patients who did not develop POD. In addition, 29 patients- and treatment-specific factors (boolean, categorical and continuous) were collected to examine their association with POD and to account for possible confounding. The factors related to the patients' general health status, pre-existing comorbidities and medications, as well as to specifications regarding OSCC and its surgical treatment. Statistical analysis To find the variables that best explain the occurrence of POD, without allowing for human pre-selection, an automatic variable selection method was developed. This method involves evaluating a generalised linear model (GLM) formulated as a logistic regression model. The model is then extended stepwise by adding additional predictor variables. If an added variable contributes significantly (p-value < 0.05) to the quality of the model, it is retained, otherwise it is discarded. Variable selection is performed using the R-package “ glmtoolbox” (Vanegas et al., 2021 ). To estimate each variable’s contribution to the model, we computed the marginal R 2 for GLM using the R-package http://glmm.hp (Vanegas et al., 2021 ) that estimates the percentage of individual effects towards total (marginal) R 2 for each predictor. Using a logistic regression model for predictions returns (log) probabilities rather than boolean values used to encode the presence of POD. We selected a threshold of 0.5 for interpreting the result as true (1.0) or false (0.0). That means if the estimated probability for a given patient is above 50%, we assign the Boolean value true for having a POD to that patient and false for not having a POD otherwise. Confidence intervals regarding the percentage of correct or false predictions were computed using a cross-validation approach where the data is repeatedly (1000 rounds) split into training and testing sets. The training set consists of 200 randomly selected patients to compute the model, which is then used to predict the remaining 99 patients. We will report the 10th, 50th and 90th percentiles of correct assignments. These are then compared to the baseline by randomly reassigning the POD variable throughout the sample, which is the same as guessing based on the prior probability of a patient experiencing POD from the data. Additionally, the percentage of false positives and negatives obtained by the cross-validation will be reported. Development of a risk assessment tool Using the R-package shiny (Chang et al., 2012 ), we developed a web application with a graphical user interface to easily compute the probability for the occurrence of delirium given the chosen predictor value. Results Study Group Characteristics Two hundred ninety-nine patients (130 female and 169 male) received primary surgery of OSCC between January 1st, 2013 and December 31st, 2020 at the Department of Oral and Maxillofacial Surgery, University Hospital Freiburg, Germany and met the inclusion criteria. An overall mean age of 65.5 ± 12.1 years was observed. POD occurred in 64 (21.4%) cases in total. POD rates of men (23.1%) and women (19.2%) were comparable. All demographic and clinical characteristics of the study patients are provided in Table 1 . Table 1 All characteristics of the study group are plotted in dependence of the occurrence of POD. Variable Total (n) POD Group (n) Non-POD Group (n) 299 64 235 Age (%) 65 years 149 (49.8) 48 (75.0) 101 (43.0) Male sex (%) 169 (56.5) 39 (60.9) 130 (55.3) Nicotine abuse (%) 150 (50.2) 27 (42.2) 123 (52.3) Alcohol abuse (%) 103 (34.4) 24 (37.5) 79 (33.6) BMI average in kg/m 2 (SD) 24.5 (6.1) 24.0 (6.7) 24.6 (6.0) ASA PS Classification 1 (%) 3 (1) 0 3 (1.8) 2 (%) 133 (44.5) 19 (29.7) 114 (48.5) 3 (%) 152 (50.8) 41 (64.1) 111 (47.2) 4 (%) 11 (3.7) 4 (6.3) 7 (3.0) Cardiovascular disease (%) 181 (60.5) 49 (76.6) 132 (56.2) Pulmonary disease (%) 81 (27.1) 24 (37.5) 57 (24.3) Psychiatric disease (%) 29 (9.7) 7 (10.9) 22 (9.4) Thyroid disease (%) 44 (14.7) 10 (15.6) 34 (14.5) Hepatic disease (%) 19 (6.4) 3 (4.7) 16 (6.8) Renal disease (%) 20 (6.7) 10 (15.6) 10 (4.3) Diabetes mellitus (%) 51 (17.1) 15 (23.4) 36 (15.3) Platelet aggregation inhibitors (%) 21 (7.0) 6 (9.4) 15 (6.4) Other anticoagulants (%) 22 (7.4) 5 (7.8) 17 (7.2) Psychotropic drugs (%) 86 (28.8) 14 (21.9) 72 (30.6) Antidiabetics (%) 23 (7.7) 6 (9.4) 17 (7.2) Thyroid medication (%) 14 (4.7%) 4 (6.3) 10 (4.3) Antihypertensive drugs (%) 77 (25.8) 24 (37.5) 53 (22.6) Analgesics (%) 55 (18.4) 12 (18.8) 43 (18.3) PEG (%) 34 (11.4) 7 (10.9) 27 (11.5) Tracheostomy (%) 16 (5.4) 5 (7.8) 11 (4.7) Surgery time average in hours (SD) 8.0 (3.7) 9.1 (3.7) 7.8 (3.6) Free Flap Reconstruction (%) 129 (43.1) 42 (65.6) 87 (37.0) Localisation of OSCC Floor of the mouth (%) 141 (47.2) 31 (48.4) 110 (46.7) Tongue (%) 70 (23.4) 5 (7.8) 65 (27.7) Oropharynx (%) 5 (1.7) 2 (3.1) 3 (1.3) Cheek and lips (%) 18 (6.0) 6 (9.4) 12 (5.1) Upper jaw (%) 35 (11.7) 13 (20.3) 22 (9.4) Multiple localisations (%) 30 (10.0) 7 (10.9) 23 (9.8) Hospital stay average in days (SD) 18.2 (12.9) 24.9 (15.7) 16.4 (11.2) ICU stay average in days (SD) 3.2 (4.8) 6.8 (8.3) 2.2 (2.4) Perioperative transfusion (%) 86 (28.8) 33 (51.6) 53 (22.6) UICC Stage 1 (%) 82 (27.4) 7 (10.9) 75 (31.9) 2 (%) 66 (22.0) 10 (15.6) 56 (23.8) 3 (%) 60 (20.1) 16 (25) 44 (18.7) 4 (%) 91 (30.4) 31 (48.4) 60 (25.5) Prediction model Automatic variable selection resulted in the following set of variables, that contribute significantly to the model (marginal R 2 in brackets): ICU stay (46.1%), PEG (14.6%), age (12.4%), UICC (8.4%), psychotropic drugs (6.1%), antihypertensive drugs (4.0%). Table 2 indicates the R 2 , p-value and regression coefficient of these variables. Prediction accuracy was at 85.3% correct assignments with the cross-validated confidence interval of 80.8% (10th percentile) and 88.9% (90th percentile). This has to be evaluated against the reference value for resampling (guessing based on prior probabilities) which is between 63.8% (10th percentile) and 69.2% (90th percentile). Table 2 Variables that contribute significantly to the model's marginal R 2 are listed on the left. R 2 is specified in percentage. All selected variables have a significant effect (p 0) or negatively with the occurrence of POD (value < 0). Variable R 2 p-value Regression coefficient Length of ICU stay 53.5 < 0.001 0.32 PEG 17.9 < 0.001 (-)2.34 Age 10.8 < 0.001 0.05 UICC stage 11 < 0.01 0.47 Psychotropic drugs 4.1 < 0.05 (-) 0.98 Antihypertensive drugs 2.7 < 0.05 0.74 Development of a risk assessment tool Due to the low R 2 value of the last two variables (psychotropic drugs R 2 = 4.1%, antihypertensive drugs = 2.7%) we did not include them in the development of the risk assessment tool. The prediction accuracy of the condensed model only dropped from 85% correct assignments to 84% and therefore, were not considered clinically relevant. The confidence of the accuracy for the model lies between 79% (10th percentile) and 87% (90th percentile). The distribution of false negatives lies between 9% (10th percentile) and 18% (90th percentile) with a mean of 13.2%. False positives range far lower with a mean of 3.3% (10th percentile = 1%; 90th percentile = 6%). It is deployed and available online ( https://zkm-uni-freiburg.shinyapps.io/DelirmodelApp/ ). Discussion Main results In recent decades, numerous studies have investigated risk factors for postoperative delirium (POD) in general, as well as in patients following oral cancer surgery. The incidence of POD in our cohort was 21.4%, which is comparable to the reported incidences in the literature (Hasegawa et al., 2015 ; Makiguchi et al., 2018 ; Takahashi et al., 2021 ; Ying et al., 2025 ). Unlike the vast majority of available literature, this study is one of the few studies that omitted manual selection of factors, instead developing an automated variable selection process to eliminate selection bias. The most significant factors correlating with the occurrence of POD were determined to be length of ICU stay, PEG, age, UICC stage, and the use of antipsychotic and antihypertensive drugs. To our knowledge, our study is the largest investigating risk factors for POD following surgery for OSCC in western countries. Clinical and scientific relevance of significant factors Of all variables, length of ICU stay contributed most to the developed risk prediction model (R 2 = 46.1%). This is in line with recent findings of Ying et al. in a large cohort of Chinese patients following oral cancer surgery (Ying et al., 2025 ), but has also been described in other patient cohorts (Kirfel et al., 2022 ). A high noise level and unfavourable light conditions on the ICU lead to disturbance of circadian rhythm and a disrupted sleep of poor quality (Elliott et al., 2013 ; Showler et al., 2024 ), which is thought to be related to POD (Kamdar et al., 2016 ). Furthermore, critical illness and a prolonged ICU stay are associated with a variety of physical and cognitive impairments (Hiser et al., 2023 ; Lee et al., 2025 ). Of course, it has also to be considered the presence of delirium itself may have influenced the length of ICU stay (Ma et al., 2023 ; Perry et al., 2024 ), which might explain the comparatively strong statistical influence of this factor. The presence of a PEG-tube was identified as the second most significant factor (R 2 = 14.6%). In contrast to most of the investigated factors, the presence of a PEG-tube revealed to be a protective factor (coefficient = -2.34). It should be noted that PEG tube placement was performed at different times among the study group. While some patients received PEG tube placement preoperatively, others received it during their stay in the ICU or surgical ward. The protective effect of PEG-tubes might be explained by the improvement of perioperative nutrition and by avoidance of a nasogastric tube which both have been identified as predisposing factors for POD (Fields et al., 2018 ; Moellmann et al., 2024 ). Recent studies evaluating the effects of PEG tube placement in both stroke and neurosurgical patients found PEG feeding to be a safe and well-tolerated method of ensuring enteral nutrition. However, no correlation with an improvement in the patients' neurological status was found (Koc et al., 2007 ; Rajalbandi et al., 2024 ). Based on our observations, we therefore encourage practitioners involved in the perioperative care of OSCC patients to consider preoperative PEG tube placement to ensure safe enteral feeding particularly for patients with expected postoperative impairment of swallowing and chewing. Age also contributed significantly to the prediction model (R 2 = 12.4%). Advanced age is often accompanied by frailty, an increased vulnerability to external stressors, which trigger pathologic changes in cognition and attention (Evered et al., 2022 ; Quinlan et al., 2011 ). Furthermore, disturbed perioperative cerebral blood flow seems to be a relevant factor increasing the incidence of neurocognitive disorders in elderly patients (NeuroVISION Investigators, 2019 ). Indeed, age has been identified as a highly relevant risk factor in previous studies on risk factors for POD in general (Mevorach et al., 2023 ) but also in patients following surgery for OSCC (Hasegawa et al., 2015 ; Shiiba et al., 2009 ; Ying et al., 2025 ). Higher UICC stage was also identified as a predictor for POD in our cohort (R 2 = 8.4%). This association was equally found following head and neck surgery (Densky et al., 2019 ), but was not included in most studies following oral cancer surgery (Hasegawa et al., 2015 ; Makiguchi et al., 2018 ; Ying et al., 2025 ). Higher UICC stages usually require a more invasive approach leading to an increased surgical inflammatory stress response, known to cause neuro-inflammation and trigger perioperative neurocognitive disorders (Evered et al., 2022 ). (Cerejeira et al., 2013 ). Two other factors contributed significantly, albeit to lesser extent, to our prediction model. Whereas therapy with psychotropic drugs had a protective effect, therapy with antihypertensive drugs had a risk-increasing effect for the development of POD. Due to low R 2 values, both factors were not included in the developed risk assessment tool. Psychiatric disorders in contrast showed no significant association with POD, whereas this was reported in patients following head and neck surgery (Choi et al., 2015 ). It is worth noting that in our cohort less than 60% of patients with psychiatric disorders received psychotropic drugs. Assuming that untreated psychiatric disorder particularly increases the risk for POD, the protective influence of the treatment with psychotropic drugs seems plausible. The association of POD with arterial hypertension and antihypertensives has been described before in patients undergoing maxillofacial tumour surgery (Kong et al., 2021 ), but also in a wide variety of patient collectives and in different types of neurocognitive disorders (Huang and Aronow, 2024 ; Ormseth et al., 2023 ). The underlying mechanisms of this association are complex and bidirectional. Disturbed cerebral autoregulation of blood flow in hypertensive patients might play a role in this context (Futier et al., 2017 ; Saugel and Sessler, 2021 ). Statistical analysis A common approach that has been used to identify factors associated with POD is to search statistically significantly different distribution in patients with and without delirium by a univariate approach and then perform a subsequent logistic regression analysis (Hasegawa et al., 2015 ; Makiguchi et al., 2018 ; Takahashi et al., 2021 ; Ying et al., 2025 ). Yet, pre-selection of variables from a very large set of potential predictors based on their significance in a univariate model leads to small p-values and overestimation of effects in a multivariate model (Steyerberg and Van Calster, 2020 ; Vach, 2013 ). Thus, the use of bias reducing automatic variable selection may be a feasible tool to avoid overestimation of effects. Our study is also assessing the model’s strengths and shortcomings against the backdrop of results obtained by guessing (random assignment) based on the prior probabilities as well as reporting confidence intervals based on bootstrapping procedures. This is important as the sample is unbalanced with only about 21% patients being affected by POD and simply reporting prediction accuracy might lead to overconfidence regarding the model’s predictive power. However, our statistical approach requires a complete dataset and does not permit any missing values. This poses a certain challenge especially for retrospective studies. Therefore, some variables that have been reported as risk factors before could not be included in our analysis. Limitations Several limitations of the study including, but not limited to its retrospective design have to be considered. Patients were predominantly screened with NU-DESC which is a validated tool known to have a high sensitivity even without special training, yet has a lower specificity compared to other Delirium screening tools (Aldecoa et al., 2024 ; Neufeld et al., 2013 ). Furthermore, patients were not screened systematically for preoperative cognitive disorders, which increases the risk for POD (Weiss et al., 2023 ) and thus represents a neglected factor in our study. Prospective studies are needed to specify risk profiles in oral cancer patients and may offer opportunities to optimize our prediction tool for a future establishment in clinical practice. Conclusion Our study revealed four factors increasing and two factors reducing the risk for POD by the use of a bias reducing automatic variable selection method. The resulting online risk assessment tool might offer an easy-to-implement option to identify patients at risk for POD, which however has to be validated in clinical practice. Declarations Funding Statement The authors declare that no external funding was received for this study. Data availability statement Due to institutional, ethical, and privacy data-protection policies, the underlying individual patient-level data cannot be made publicly available. Access may be provided upon reasonable request, subject to approval by the relevant institutional review board and in compliance with applicable regulations. Author contribution statement Conceptualization & Study design: Matthias Jakob Posch, Johannes Kalbhenn, Marc Metzger, Julia Vera Brandenburg, Leonard Simon Brandenburg. Methodology: Stefan Schlager, Leonard Simon Brandenburg, Matthias Jakob Posch. Data curation: Konstantin Hasel, Julia Vera Brandenburg, Leonard Simon Brandenburg. Formal analysis & Statistics: Stefan Schlager, Matthias Jakob Posch, Leonard Simon Brandenburg. Software / Risk prediction tool development: Stefan Schlager, Matthias Jakob Posch, Leonard Simon Brandenburg. Visualization (tables, figures, online tool): Stefan Schlager, Matthias Jakob Posch. Writing – original draft: Matthias Jakob Posch, Julia Vera Brandenburg, Leonard Simon Brandenburg. Writing – review & editing: Johannes Kalbhenn, Julia Brandenburg, Jonas Wüster, Damian Sommer, Marc Metzger Rainer Schmelzeisen. Supervision: Johannes Kalbhenn, Marc Metzger, Rainer Schmelzeisen. Competing Interests The authors declare that they have no competing interests. Ethics approval and consent to participate The study was approved by the Ethics Committee of the University of Freiburg, Germany (No. 127/15 and No. 30/20). The requirement for informed consent was waived due to the retrospective design of the study. Consent for publication This manuscript does not contain any individual person’s data. Therefore, consent for publication was not required. References Aldecoa, C., Bettelli, G., Bilotta, F., Sanders, R.D., Aceto, P., Audisio, R., Cherubini, A., Cunningham, C., Dabrowski, W., Forookhi, A., Gitti, N., Immonen, K., Kehlet, H., Koch, S., Kotfis, K., Latronico, N., MacLullich, A.M.J., Mevorach, L., Mueller, A., Neuner, B., Piva, S., Radtke, F., Blaser, A.R., Renzi, S., Romagnoli, S., Schubert, M., Slooter, A.J.C., Tommasino, C., Vasiljewa, L., Weiss, B., Yuerek, F., Spies, C.D., 2024. Update of the European Society of Anaesthesiology and Intensive Care Medicine evidence-based and consensus-based guideline on postoperative delirium in adult patients. Eur J Anaesthesiol 41, 81–108. https://doi.org/10.1097/EJA.0000000000001876 Bray, F., Laversanne, M., Sung, H., Ferlay, J., Siegel, R.L., Soerjomataram, I., Jemal, A., 2024. Global cancer statistics 2022: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries. CA Cancer J Clin 74, 229–263. https://doi.org/10.3322/caac.21834 Cerejeira, J., Batista, P., Nogueira, V., Vaz-Serra, A., Mukaetova-Ladinska, E.B., 2013. The stress response to surgery and postoperative delirium: evidence of hypothalamic-pituitary-adrenal axis hyperresponsiveness and decreased suppression of the GH/IGF-1 Axis. J Geriatr Psychiatry Neurol 26, 185–194. https://doi.org/10.1177/0891988713495449 Chamoli, A., Gosavi, A.S., Shirwadkar, U.P., Wangdale, K.V., Behera, S.K., Kurrey, N.K., Kalia, K., Mandoli, A., 2021. Overview of oral cavity squamous cell carcinoma: Risk factors, mechanisms, and diagnostics. Oral Oncol 121, 105451. https://doi.org/10.1016/j.oraloncology.2021.105451 Chang, W., Cheng, J., Allaire, J., Sievert, C., Schloerke, B., Xie, Y., Allen, J., McPherson, J., Dipert, A., Borges, B., 2012. shiny: Web Application Framework for R. https://doi.org/10.32614/CRAN.package.shiny Choi, G.J., Baek, C.W., Kang, H., Park, Y.H., Yang, S.Y., Shin, H.Y., Jung, Y.H., Woo, Y.C., Lee, U.L., 2015. Emergence agitation after orthognathic surgery: a randomised controlled comparison between sevoflurane and desflurane. Acta Anaesthesiol Scand 59, 224–231. https://doi.org/10.1111/aas.12435 Densky, J., Eskander, A., Kang, S., Chan, J., Tweel, B., Sitapara, J., Ozer, E., Agrawal, A., Carrau, R., Rocco, J., Teknos, T.N., Old, M., 2019. Risk Factors Associated With Postoperative Delirium in Patients Undergoing Head and Neck Free Flap Reconstruction. JAMA Otolaryngol Head Neck Surg 145, 216–221. https://doi.org/10.1001/jamaoto.2018.3820 Dilmen, O.K., Meco, B.C., Evered, L.A., Radtke, F.M., 2024. Postoperative neurocognitive disorders: A clinical guide. J Clin Anesth 92, 111320. https://doi.org/10.1016/j.jclinane.2023.111320 Elliott, R., McKinley, S., Cistulli, P., Fien, M., 2013. Characterisation of sleep in intensive care using 24-hour polysomnography: an observational study. Crit Care 17, R46. https://doi.org/10.1186/cc12565 Evered, L., Atkins, K., Silbert, B., Scott, D.A., 2022. Acute peri-operative neurocognitive disorders: a narrative review. Anaesthesia 77 Suppl 1, 34–42. https://doi.org/10.1111/anae.15613 Falegnami, A., Patriarca, R., Costantino, F., Di Gravio, G., Bilotta, F., 2021. Surveying work-as-done in post-operative delirium risk factors collection and diagnosis monitoring. Appl Ergon 92, 103347. https://doi.org/10.1016/j.apergo.2020.103347 Fields, A., Huang, J., Schroeder, D., Sprung, J., Weingarten, T., 2018. Agitation in adults in the post-anaesthesia care unit after general anaesthesia. Br J Anaesth 121, 1052–1058. https://doi.org/10.1016/j.bja.2018.07.017 Futier, E., Lefrant, J.-Y., Guinot, P.-G., Godet, T., Lorne, E., Cuvillon, P., Bertran, S., Leone, M., Pastene, B., Piriou, V., Molliex, S., Albanese, J., Julia, J.-M., Tavernier, B., Imhoff, E., Bazin, J.-E., Constantin, J.-M., Pereira, B., Jaber, S., INPRESS Study Group, 2017. Effect of Individualized vs Standard Blood Pressure Management Strategies on Postoperative Organ Dysfunction Among High-Risk Patients Undergoing Major Surgery: A Randomized Clinical Trial. JAMA 318, 1346–1357. https://doi.org/10.1001/jama.2017.14172 Hasegawa, T., Saito, I., Takeda, D., Iwata, E., Yonezawa, N., Kakei, Y., Sakakibara, A., Akashi, M., Minamikawa, T., Komori, T., 2015. Risk factors associated with postoperative delirium after surgery for oral cancer. J Craniomaxillofac Surg 43, 1094–1098. https://doi.org/10.1016/j.jcms.2015.06.011 Hiser, S.L., Fatima, A., Ali, M., Needham, D.M., 2023. Post-intensive care syndrome (PICS): recent updates. J Intensive Care 11, 23. https://doi.org/10.1186/s40560-023-00670-7 Huang, L., Aronow, W.S., 2024. Association of Hypertension with Different Cognitive Disorders. J Clin Med 13, 6029. https://doi.org/10.3390/jcm13206029 Kamdar, B.B., Martin, J.L., Needham, D.M., Ong, M.K., 2016. Promoting Sleep to Improve Delirium in the ICU. Crit Care Med 44, 2290–2291. https://doi.org/10.1097/CCM.0000000000001982 Kirfel, A., Guttenthaler, V., Mayr, A., Coburn, M., Menzenbach, J., Wittmann, M., 2022. Postoperative delirium is an independent factor influencing the length of stay of elderly patients in the intensive care unit and in hospital. J Anesth 36, 341–348. https://doi.org/10.1007/s00540-022-03049-4 Koc, D., Gercek, A., Gencosmanoglu, R., Tozun, N., 2007. Percutaneous endoscopic gastrostomy in the neurosurgical intensive care unit: complications and outcome. JPEN J Parenter Enteral Nutr 31, 517–520. https://doi.org/10.1177/0148607107031006517 Kong, S., Wang, J., Xu, H., Wang, K., 2021. Effect of hypertension and medication use regularity on postoperative delirium after maxillofacial tumors radical surgery. Oncotarget 12, 1811–1820. https://doi.org/10.18632/oncotarget.28048 Lander, H.L., Dick, A.W., Joynt Maddox, K.E., Oldham, M.A., Fleisher, L.A., Mazzeffi, M., Lustik, S.J., Shang, J., Stone, P.W., Gloff, M.S., Nadler, J., Wu, I., Zollo, R., Glance, L.G., 2025. Postoperative Delirium in Older Adults Undergoing Noncardiac Surgery. JAMA Netw Open 8, e2519467. https://doi.org/10.1001/jamanetworkopen.2025.19467 Lee, S.Y., Huh, J.W., Hong, S.-B., Lim, C.-M., Ahn, J.H., 2025. Physical and Cognitive Impairments at ICU Discharge are Associated with High Long-Term Mortality in ICU Survivors with Solid Malignancies: A Retrospective Cohort Study. Ther Clin Risk Manag 21, 1121–1133. https://doi.org/10.2147/TCRM.S520206 Ma, Y., Li, C., Peng, W., Wan, Q., 2023. The influence of delirium on mortality and length of ICU stay and analysis of risk factors for delirium after liver transplantation. Front Neurol 14, 1229990. https://doi.org/10.3389/fneur.2023.1229990 Makiguchi, T., Yokoo, S., Kurihara, J., 2018. Risk factors for postoperative delirium in patients undergoing free flap reconstruction for oral cancer. Int J Oral Maxillofac Surg 47, 998–1002. https://doi.org/10.1016/j.ijom.2018.03.011 Mevorach, L., Forookhi, A., Farcomeni, A., Romagnoli, S., Bilotta, F., 2023. Perioperative risk factors associated with increased incidence of postoperative delirium: systematic review, meta-analysis, and Grading of Recommendations Assessment, Development, and Evaluation system report of clinical literature. Br J Anaesth 130, e254–e262. https://doi.org/10.1016/j.bja.2022.05.032 Moellmann, H.L., Alhammadi, E., Boulghoudan, S., Kuhlmann, J., Mevissen, A., Olbrich, P., Rahm, L., Frohnhofen, H., 2024. Risk of sarcopenia, frailty and malnutrition as predictors of postoperative delirium in surgery. BMC Geriatr 24, 971. https://doi.org/10.1186/s12877-024-05566-1 Neufeld, K.J., Leoutsakos, J.S., Sieber, F.E., Joshi, D., Wanamaker, B.L., Rios-Robles, J., Needham, D.M., 2013. Evaluation of two delirium screening tools for detecting post-operative delirium in the elderly. Br J Anaesth 111, 612–618. https://doi.org/10.1093/bja/aet167 NeuroVISION Investigators, 2019. Perioperative covert stroke in patients undergoing non-cardiac surgery (NeuroVISION): a prospective cohort study. Lancet 394, 1022–1029. https://doi.org/10.1016/S0140-6736(19)31795-7 Obermeier, K.T., Kraus, M., Smolka, W., Henkel, J., Saller, T., Otto, S., Liokatis, P., 2022. Postoperative Delirium in Patients with Oral Cancer: Is Intraoperative Fluid Administration a Neglected Risk Factor? Cancers (Basel) 14, 3176. https://doi.org/10.3390/cancers14133176 Ormseth, C.H., LaHue, S.C., Oldham, M.A., Josephson, S.A., Whitaker, E., Douglas, V.C., 2023. Predisposing and Precipitating Factors Associated With Delirium: A Systematic Review. JAMA Netw Open 6, e2249950. https://doi.org/10.1001/jamanetworkopen.2022.49950 Perry, H., Alight, A., Wilcox, M.E., 2024. Light, sleep and circadian rhythm in critical illness. Curr Opin Crit Care 30, 283–289. https://doi.org/10.1097/MCC.0000000000001163 Quinlan, N., Marcantonio, E.R., Inouye, S.K., Gill, T.M., Kamholz, B., Rudolph, J.L., 2011. Vulnerability: the crossroads of frailty and delirium. J Am Geriatr Soc 59 Suppl 2, S262-268. https://doi.org/10.1111/j.1532-5415.2011.03674.x Rajalbandi, R.S., Qureshi, A., Bains, N., Gillani, S., Maqsood, H., 2024. Effect of Percutaneous Endoscopic Gastrostomy Tube Placement on Confusion/Delirium in Stroke Patients with Dysphagia (P7-5.007). Neurology 102. https://doi.org/10.1212/wnl.0000000000205070 Sahle, B.W., Pilcher, D., Litton, E., Ofori-Asenso, R., Peter, K., McFadyen, J., Bucknall, T., 2022. Association between frailty, delirium, and mortality in older critically ill patients: a binational registry study. Ann Intensive Care 12, 108. https://doi.org/10.1186/s13613-022-01080-y Saugel, B., Sessler, D.I., 2021. Perioperative Blood Pressure Management. Anesthesiology 134, 250–261. https://doi.org/10.1097/ALN.0000000000003610 Shiiba, M., Takei, M., Nakatsuru, M., Bukawa, H., Yokoe, H., Uzawa, K., Tanzawa, H., 2009. Clinical observations of postoperative delirium after surgery for oral carcinoma. Int J Oral Maxillofac Surg 38, 661–665. https://doi.org/10.1016/j.ijom.2009.01.011 Showler, L., Deane, A.M., Litton, E., Ankravs, M.J., Wibrow, B., Barge, D., Goldin, J., Hammond, N., Saxena, M.K., Young, P.J., Venkatesh, B., Finnis, M., Abdelhamid, Y.A., 2024. A multicentre point prevalence study of nocturnal hours awake and enteral pharmacological sleep aids in patients admitted to Australian and New Zealand intensive care units. Crit Care Resusc 26, 192–197. https://doi.org/10.1016/j.ccrj.2024.06.009 Steyerberg, E.W., Van Calster, B., 2020. Redefining significance and reproducibility for medical research: A plea for higher P-value thresholds for diagnostic and prognostic models. Eur J Clin Invest 50, e13229. https://doi.org/10.1111/eci.13229 Takahashi, N., Hiraki, A., Kawahara, K., Nagata, M., Yoshida, R., Matsuoka, Y., Tanaka, T., Obayashi, Y., Sakata, J., Nakashima, H., Arita, H., Shinohara, M., Nakayama, H., 2021. Postoperative delirium in patients undergoing tumor resection with reconstructive surgery for oral cancer. Mol Clin Oncol 14, 60. https://doi.org/10.3892/mco.2021.2222 Vach, W., 2013. Regression models as a tool in medical research, A Chapman & Hall book. CRC Press, Boca Raton, Fla. Vanegas, L.H., Rondón, L.M., Paula, G.A., 2021. glmtoolbox: Set of Tools to Data Analysis using Generalized Linear Models. CRAN: Contributed Packages. https://doi.org/10.32614/cran.package.glmtoolbox Weiss, Y., Zac, L., Refaeli, E., Ben-Yishai, S., Zegerman, A., Cohen, B., Matot, I., 2023. Preoperative Cognitive Impairment and Postoperative Delirium in Elderly Surgical Patients: A Retrospective Large Cohort Study (The CIPOD Study). Ann Surg 278, 59–64. https://doi.org/10.1097/SLA.0000000000005657 Witlox, J., Eurelings, L.S.M., de Jonghe, J.F.M., Kalisvaart, K.J., Eikelenboom, P., van Gool, W.A., 2010. Delirium in elderly patients and the risk of postdischarge mortality, institutionalization, and dementia: a meta-analysis. JAMA 304, 443–451. https://doi.org/10.1001/jama.2010.1013 Wolff, K.D., Rau, A., Weitz, J., 2021. S3-Leitlinie Mundhöhlenkarzinom, AWMF Registernummer: 007/100OL (No. AWMF Registernummer: 007/100OL). Ying, C., Xiaona, L., Aili, Z., Zengxiang, W., Ying, W., Yu, P., Hongbo, Z., Danni, W., Meiping, J., Hongyuan, D., 2025. Development and validation of a nomogram model for predicting postoperative delirium in elderly patients with oral cancer: a retrospective study. BMC Oral Health 25, 990. https://doi.org/10.1186/s12903-025-06167-z Additional Declarations No competing interests reported. Cite Share Download PDF Status: Under Review Version 1 posted Reviews received at journal 26 Nov, 2025 Reviews received at journal 23 Nov, 2025 Reviewers agreed at journal 18 Nov, 2025 Reviewers agreed at journal 15 Nov, 2025 Reviewers invited by journal 10 Nov, 2025 Editor invited by journal 09 Oct, 2025 Editor assigned by journal 08 Oct, 2025 Submission checks completed at journal 08 Oct, 2025 First submitted to journal 22 Sep, 2025 You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-7687245","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":546888641,"identity":"89e05e52-5e1d-4406-af87-f58719b76e0c","order_by":0,"name":"Matthias Jakob Posch","email":"","orcid":"","institution":"Medical Center - University of Freiburg, University of Freiburg, Freiburg I.Br","correspondingAuthor":false,"prefix":"","firstName":"Matthias","middleName":"Jakob","lastName":"Posch","suffix":""},{"id":546888642,"identity":"ec7bf828-0b42-4cee-a377-99846927b4c4","order_by":1,"name":"Johannes Kalbhenn","email":"","orcid":"","institution":"Medical Center - University of Freiburg, University of Freiburg, Freiburg I.Br","correspondingAuthor":false,"prefix":"","firstName":"Johannes","middleName":"","lastName":"Kalbhenn","suffix":""},{"id":546888643,"identity":"3f952538-697d-4cfc-9ca4-fbc5dc3e208e","order_by":2,"name":"Stefan Schlager","email":"","orcid":"","institution":"Medical Center—University of Freiburg, University of Freiburg","correspondingAuthor":false,"prefix":"","firstName":"Stefan","middleName":"","lastName":"Schlager","suffix":""},{"id":546888644,"identity":"51930851-5359-4c8d-aa01-d563c9556e3f","order_by":3,"name":"Damian Sommer","email":"","orcid":"","institution":"Medical Center—University of Freiburg, University of Freiburg","correspondingAuthor":false,"prefix":"","firstName":"Damian","middleName":"","lastName":"Sommer","suffix":""},{"id":546888645,"identity":"ed9e8192-e462-4c71-a582-8c952e2a6503","order_by":4,"name":"Konstantin Hasel","email":"","orcid":"","institution":"Oralchirurgie am Schlossberg","correspondingAuthor":false,"prefix":"","firstName":"Konstantin","middleName":"","lastName":"Hasel","suffix":""},{"id":546888646,"identity":"df148f6c-96e0-4243-9814-82cf68a53d5c","order_by":5,"name":"Julia Brandenburg","email":"","orcid":"","institution":"Medical Center—University of Freiburg, University of Freiburg","correspondingAuthor":false,"prefix":"","firstName":"Julia","middleName":"","lastName":"Brandenburg","suffix":""},{"id":546888647,"identity":"e3d297a4-02d0-4f91-bd3f-c6f1b0649b09","order_by":6,"name":"Jonas Wüster","email":"","orcid":"","institution":"Medical Center—University of Freiburg, University of Freiburg","correspondingAuthor":false,"prefix":"","firstName":"Jonas","middleName":"","lastName":"Wüster","suffix":""},{"id":546888648,"identity":"baba011c-3817-4027-9d64-d640e1b99a76","order_by":7,"name":"Marc Metzger","email":"","orcid":"","institution":"Medical Center—University of Freiburg, University of Freiburg","correspondingAuthor":false,"prefix":"","firstName":"Marc","middleName":"","lastName":"Metzger","suffix":""},{"id":546888649,"identity":"2445a340-f340-4abc-9d3b-0c3e54868a52","order_by":8,"name":"Rainer Schmelzeisen","email":"","orcid":"","institution":"Center—University of Freiburg, University of Freiburg","correspondingAuthor":false,"prefix":"","firstName":"Rainer","middleName":"","lastName":"Schmelzeisen","suffix":""},{"id":546888650,"identity":"f6c37f8b-2733-4a41-b808-6090497e46f1","order_by":9,"name":"Leonard Simon Brandenburg","email":"data:image/png;base64,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","orcid":"","institution":"Center—University of Freiburg, University of Freiburg","correspondingAuthor":true,"prefix":"","firstName":"Leonard","middleName":"Simon","lastName":"Brandenburg","suffix":""}],"badges":[],"createdAt":"2025-09-23 01:53:11","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-7687245/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-7687245/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":96310761,"identity":"3cb104f0-ff5c-4107-8743-cca116196df5","added_by":"auto","created_at":"2025-11-19 16:21:48","extension":"docx","order_by":0,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":107446,"visible":true,"origin":"","legend":"","description":"","filename":"ManuscriptDeliriumRiskPrediction20251005.docx","url":"https://assets-eu.researchsquare.com/files/rs-7687245/v1/84a624515488e2177f3b783e.docx"},{"id":96310760,"identity":"cfae8e11-4213-4032-a38e-ccf735c63ad1","added_by":"auto","created_at":"2025-11-19 16:21:48","extension":"json","order_by":1,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":10968,"visible":true,"origin":"","legend":"","description":"","filename":"769ffe1a63de42e289c5bc28233a8ff6.json","url":"https://assets-eu.researchsquare.com/files/rs-7687245/v1/db4d3d9ad3230c02fc04ce6c.json"},{"id":96365293,"identity":"aa13e5c9-4243-4726-97d3-e92e5609dcdb","added_by":"auto","created_at":"2025-11-20 10:10:14","extension":"xml","order_by":2,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":137525,"visible":true,"origin":"","legend":"","description":"","filename":"769ffe1a63de42e289c5bc28233a8ff61enriched.xml","url":"https://assets-eu.researchsquare.com/files/rs-7687245/v1/9f7336d5aaea8b56b037e95c.xml"},{"id":96310762,"identity":"878659b7-b896-4a36-b373-218c9f891bb0","added_by":"auto","created_at":"2025-11-19 16:21:48","extension":"xml","order_by":3,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":134536,"visible":true,"origin":"","legend":"","description":"","filename":"769ffe1a63de42e289c5bc28233a8ff61structuring.xml","url":"https://assets-eu.researchsquare.com/files/rs-7687245/v1/5614f2157538a65e0dc7db86.xml"},{"id":96310764,"identity":"b6515785-f7da-436c-bb49-8f38d18b8378","added_by":"auto","created_at":"2025-11-19 16:21:48","extension":"html","order_by":4,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":142369,"visible":true,"origin":"","legend":"","description":"","filename":"earlyproof.html","url":"https://assets-eu.researchsquare.com/files/rs-7687245/v1/c7e2bada6f7b54c53ddd4a18.html"},{"id":96708060,"identity":"2eb0771e-09c4-4ccb-a8b1-5a7581c22fd5","added_by":"auto","created_at":"2025-11-25 09:54:40","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":790151,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-7687245/v1/56c60a3b-cccd-4109-a5e5-ab4b58333530.pdf"}],"financialInterests":"No competing interests reported.","formattedTitle":"Development of a risk prediction model for delirium following surgery for oral squamous cell cancer","fulltext":[{"header":"Introduction","content":"\u003cp\u003eOral squamous cell cancer (OSCC) is among the most frequent malignancies and causes more than 180,000 deaths per year worldwide. OSCC mostly occurs in patients of advanced age and substantial alcohol and tobacco consumption histories (Bray et al., \u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2024\u003c/span\u003e). In most stages of the disease, surgical resection in combination with neck dissection and often primary reconstruction with microvascular free flaps, is crucial to achieve the best possible functional and oncological outcome (Chamoli et al., \u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e2021\u003c/span\u003e). The extent of neck dissection is determined by the radiologically and clinically detected lymph node involvement and by the tumor location. Depending on the patient\u0026rsquo;s condition and the extent of the surgical procedure, admission to the intensive care unit (ICU) is required to ensure optimal perioperative care and to avoid complications. Postoperative delirium (POD) is a common complication among such patients at the ICU, with substantial medical and economic relevance (Lander et al., \u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e2025\u003c/span\u003e; Sahle et al., \u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e2022\u003c/span\u003e; Witlox et al., \u003cspan citationid=\"CR43\" class=\"CitationRef\"\u003e2010\u003c/span\u003e). POD is an acute, fluctuating neurocognitive disorder characterized by acute confusion, hallucinations, disorientation, inappropriate behaviour and other symptoms. Depending on the clinical presentation, delirium can be categorised as hyperactive, hypoactive, or mixed (Dilmen et al., \u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e2024\u003c/span\u003e). Validated tools such as the Diagnostic and Statistical Manual of Mental Disorders (DSM) 5 criteria, the Confusion Assessment Method and the Nursing Delirium Screening Scale are recommended for diagnosing delirium (Aldecoa et al., \u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e2024\u003c/span\u003e). It is worth noting that high incidences of POD have been reported in patients receiving surgery for OSCC, (Hasegawa et al., \u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e2015\u003c/span\u003e; Makiguchi et al., \u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e2018\u003c/span\u003e; Takahashi et al., \u003cspan citationid=\"CR39\" class=\"CitationRef\"\u003e2021\u003c/span\u003e; Ying et al., \u003cspan citationid=\"CR45\" class=\"CitationRef\"\u003e2025\u003c/span\u003e). Treatment of POD is challenging and therefore individualized preventive measures based on risk factors are highly recommended (Aldecoa et al., \u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e2024\u003c/span\u003e).\u003c/p\u003e\u003cp\u003eSeveral studies investigated risk factors for POD following surgery for OSCC and revealed a high heterogeneity with regard to the identified risk factors (Hasegawa et al., \u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e2015\u003c/span\u003e; Makiguchi et al., \u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e2018\u003c/span\u003e; Obermeier et al., \u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e2022\u003c/span\u003e; Shiiba et al., \u003cspan citationid=\"CR36\" class=\"CitationRef\"\u003e2009\u003c/span\u003e; Takahashi et al., \u003cspan citationid=\"CR39\" class=\"CitationRef\"\u003e2021\u003c/span\u003e; Ying et al., \u003cspan citationid=\"CR45\" class=\"CitationRef\"\u003e2025\u003c/span\u003e).\u003c/p\u003e\u003cp\u003eNext to the multifactorial aetiology, limited sample size and the lack of adjustment for risk bias may contribute to the inhomogeneous results observed regarding POD following surgery for OSCC (Steyerberg and Van Calster, \u003cspan citationid=\"CR38\" class=\"CitationRef\"\u003e2020\u003c/span\u003e; Vach, \u003cspan citationid=\"CR40\" class=\"CitationRef\"\u003e2013\u003c/span\u003e). Therefore, in this study we used an automatic bias reducing variable selection method to avoid potential confounding caused by pre-selection of factors.\u003c/p\u003e\u003cp\u003eIdentified risk factors should be used to establish a structured risk assessment for POD in daily clinical practice. However, this appears to be challenging even in modern health care systems (Falegnami et al., \u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e2021\u003c/span\u003e). Thus, our study had two objectives. Firstly, the identification of risk factors following surgery for OSCC. Secondly, the development of a freely available online risk assessment tool that would allow healthcare personnel to evaluate an individual's risk of POD based on the identified risk factors.\u003c/p\u003e"},{"header":"Materials \u0026 Methods","content":"\u003cp\u003eIn this retrospective study, electronic patient charts of patients were reviewed in order to identify risk factors for POD and generate an according risk assessment tool starting January 1st, 2013 and ending December 31st, 2020 at the Department of Oral and Maxillofacial Surgery, Medical Center, University of Freiburg, Germany.\u003c/p\u003e\u003cp\u003eSurgical and perioperative treatment was conducted following the German guidelines for the treatment of oral cavity cancer (Wolff et al., \u003cspan citationid=\"CR44\" class=\"CitationRef\"\u003e2021\u003c/span\u003e). No intervention was performed due to study implementation.\u003c/p\u003e\u003cdiv id=\"Sec3\" class=\"Section2\"\u003e\u003ch2\u003eEthical clearance\u003c/h2\u003e\u003cp\u003eThe study was approved by the local ethics committee of the University of Freiburg, Germany (No. 127/15 and No. 30/20). Informed consent was waived due to the retrospective nature of this study. The study was registered in the German Clinical Trials Register (Deutsches Register klinischer Studien, DRKS) No. DRKS00038043.\u003c/p\u003e\u003c/div\u003e\n\u003ch3\u003eInclusion \u0026 exclusion criteria\u003c/h3\u003e\n\u003cp\u003eAll patients that underwent primary resection of OSCC and who were scheduled for intensive care unit as part of our standard postoperative care in the time between January 1st, 2013 and ending December 31st, 2020 were included. No secondary exclusion of patients due to specific surgical reconstruction procedures was performed. All included patients were of legal age, revealed cancer-free resection margins (R0) after histopathological analysis and were administered to the ICU postoperatively. A postoperative follow-up of at least 12 months had to be ensured and consistent documentation of the postoperative ICU and in-ward stay had to be available.\u003c/p\u003e\u003cp\u003ePatients with malignancies other than OSCC, secondary surgery for recurrent OSCC, evidence of distant metastases at initial diagnosis, and neo-adjuvant therapy prior to surgery were excluded.\u003c/p\u003e\n\u003ch3\u003eVariables examined\u003c/h3\u003e\n\u003cp\u003eThe primary outcome variable was the development of postoperative delirium (POD). POD was diagnosed either during the stay in the ICU or after transferring the patient back to the oral and maxillofacial ward using the Nurse Delirium Screening Scale (NU-DESC). To be diagnosed with postoperative delirium a score of 2 or higher had to be clearly documented in the patient file. These patients were classified into Group 1. Group 2 was composed of the control group with patients who did not develop POD.\u003c/p\u003e\u003cp\u003eIn addition, 29 patients- and treatment-specific factors (boolean, categorical and continuous) were collected to examine their association with POD and to account for possible confounding. The factors related to the patients' general health status, pre-existing comorbidities and medications, as well as to specifications regarding OSCC and its surgical treatment.\u003c/p\u003e\u003cdiv id=\"Sec6\" class=\"Section2\"\u003e\u003ch2\u003eStatistical analysis\u003c/h2\u003e\u003cp\u003eTo find the variables that best explain the occurrence of POD, without allowing for human pre-selection, an automatic variable selection method was developed. This method involves evaluating a generalised linear model (GLM) formulated as a logistic regression model. The model is then extended stepwise by adding additional predictor variables. If an added variable contributes significantly (p-value\u0026thinsp;\u0026lt;\u0026thinsp;0.05) to the quality of the model, it is retained, otherwise it is discarded. Variable selection is performed using the R-package \u0026ldquo;\u003cem\u003eglmtoolbox\u0026rdquo;\u003c/em\u003e(Vanegas et al., \u003cspan citationid=\"CR41\" class=\"CitationRef\"\u003e2021\u003c/span\u003e). To estimate each variable\u0026rsquo;s contribution to the model, we computed the marginal R\u003csup\u003e2\u003c/sup\u003e for GLM using the R-package \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttp://glmm.hp\u003c/span\u003e\u003cspan address=\"http://glmm.hp\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (Vanegas et al., \u003cspan citationid=\"CR41\" class=\"CitationRef\"\u003e2021\u003c/span\u003e) that estimates the percentage of individual effects towards total (marginal) R\u003csup\u003e2\u003c/sup\u003e for each predictor. Using a logistic regression model for predictions returns (log) probabilities rather than boolean values used to encode the presence of POD. We selected a threshold of 0.5 for interpreting the result as true (1.0) or false (0.0). That means if the estimated probability for a given patient is above 50%, we assign the Boolean value true for having a POD to that patient and false for not having a POD otherwise.\u003c/p\u003e\u003cp\u003eConfidence intervals regarding the percentage of correct or false predictions were computed using a cross-validation approach where the data is repeatedly (1000 rounds) split into training and testing sets. The training set consists of 200 randomly selected patients to compute the model, which is then used to predict the remaining 99 patients. We will report the 10th, 50th and 90th percentiles of correct assignments. These are then compared to the baseline by randomly reassigning the POD variable throughout the sample, which is the same as guessing based on the prior probability of a patient experiencing POD from the data. Additionally, the percentage of false positives and negatives obtained by the cross-validation will be reported.\u003c/p\u003e\u003c/div\u003e\n\u003ch3\u003eDevelopment of a risk assessment tool\u003c/h3\u003e\n\u003cp\u003eUsing the R-package shiny (Chang et al., \u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e2012\u003c/span\u003e), we developed a web application with a graphical user interface to easily compute the probability for the occurrence of delirium given the chosen predictor value.\u003c/p\u003e"},{"header":"Results","content":"\u003cdiv id=\"Sec9\" class=\"Section2\"\u003e\u003ch2\u003eStudy Group Characteristics\u003c/h2\u003e\u003cp\u003eTwo hundred ninety-nine patients (130 female and 169 male) received primary surgery of OSCC between January 1st, 2013 and December 31st, 2020 at the Department of Oral and Maxillofacial Surgery, University Hospital Freiburg, Germany and met the inclusion criteria. An overall mean age of 65.5\u0026thinsp;\u0026plusmn;\u0026thinsp;12.1 years was observed. POD occurred in 64 (21.4%) cases in total. POD rates of men (23.1%) and women (19.2%) were comparable. All demographic and clinical characteristics of the study patients are provided in Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e.\u003c/p\u003e\u003cp\u003e\u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab1\" border=\"1\"\u003e\u003ccaption language=\"En\"\u003e\u003cdiv class=\"CaptionNumber\"\u003eTable 1\u003c/div\u003e\u003cdiv class=\"CaptionContent\"\u003e\u003cp\u003eAll characteristics of the study group are plotted in dependence of the occurrence of POD.\u003c/p\u003e\u003c/div\u003e\u003c/caption\u003e\u003ccolgroup cols=\"4\"\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e\u003cthead\u003e\u003ctr\u003e\u003cth align=\"left\" colname=\"c1\"\u003e\u003cp\u003eVariable\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c2\"\u003e\u003cp\u003eTotal (n)\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c3\"\u003e\u003cp\u003ePOD Group (n)\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c4\"\u003e\u003cp\u003eNon-POD Group (n)\u003c/p\u003e\u003c/th\u003e\u003c/tr\u003e\u003c/thead\u003e\u003ctbody\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e299\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e64\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e235\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eAge (%)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e\u0026lt;\u0026thinsp;45 years\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e12 (4)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e1 (1.6)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e11 (4.7)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e45\u0026ndash;65 years\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e138 (46.2)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e15 (23.4)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e123 (52.3)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e\u0026gt;\u0026thinsp;65 years\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e149 (49.8)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e48 (75.0)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e101 (43.0)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eMale sex (%)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e169 (56.5)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e39 (60.9)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e130 (55.3)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eNicotine abuse (%)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e150 (50.2)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e27 (42.2)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e123 (52.3)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eAlcohol abuse (%)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e103 (34.4)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e24 (37.5)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e79 (33.6)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eBMI average in kg/m\u003csup\u003e2\u003c/sup\u003e (SD)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e24.5 (6.1)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e24.0 (6.7)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e24.6 (6.0)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eASA PS Classification\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e1 (%)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e3 (1)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e0\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e3 (1.8)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e2 (%)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e133 (44.5)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e19 (29.7)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e114 (48.5)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e3 (%)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e152 (50.8)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e41 (64.1)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e111 (47.2)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e4 (%)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e11 (3.7)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e4 (6.3)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e7 (3.0)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eCardiovascular disease (%)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e181 (60.5)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e49 (76.6)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e132 (56.2)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003ePulmonary disease (%)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e81 (27.1)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e24 (37.5)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e57 (24.3)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003ePsychiatric disease (%)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e29 (9.7)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e7 (10.9)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e22 (9.4)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eThyroid disease (%)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e44 (14.7)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e10 (15.6)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e34 (14.5)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eHepatic disease (%)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e19 (6.4)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e3 (4.7)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e16 (6.8)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eRenal disease (%)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e20 (6.7)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e10 (15.6)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e10 (4.3)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eDiabetes mellitus (%)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e51 (17.1)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e15 (23.4)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e36 (15.3)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003ePlatelet aggregation inhibitors (%)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e21 (7.0)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e6 (9.4)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e15 (6.4)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eOther anticoagulants (%)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e22 (7.4)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e5 (7.8)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e17 (7.2)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003ePsychotropic drugs (%)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e86 (28.8)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e14 (21.9)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e72 (30.6)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eAntidiabetics (%)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e23 (7.7)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e6 (9.4)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e17 (7.2)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eThyroid medication (%)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e14 (4.7%)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e4 (6.3)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e10 (4.3)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eAntihypertensive drugs (%)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e77 (25.8)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e24 (37.5)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e53 (22.6)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eAnalgesics (%)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e55 (18.4)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e12 (18.8)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e43 (18.3)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003ePEG (%)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e34 (11.4)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e7 (10.9)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e27 (11.5)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eTracheostomy (%)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e16 (5.4)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e5 (7.8)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e11 (4.7)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eSurgery time average in hours (SD)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e8.0 (3.7)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e9.1 (3.7)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e7.8 (3.6)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eFree Flap Reconstruction (%)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e129 (43.1)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e42 (65.6)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e87 (37.0)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eLocalisation of OSCC\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eFloor of the mouth (%)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e141 (47.2)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e31 (48.4)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e110 (46.7)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eTongue (%)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e70 (23.4)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e5 (7.8)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e65 (27.7)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eOropharynx (%)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e5 (1.7)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e2 (3.1)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e3 (1.3)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eCheek and lips (%)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e18 (6.0)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e6 (9.4)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e12 (5.1)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eUpper jaw (%)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e35 (11.7)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e13 (20.3)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e22 (9.4)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eMultiple localisations (%)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e30 (10.0)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e7 (10.9)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e23 (9.8)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eHospital stay average in days (SD)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e18.2 (12.9)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e24.9 (15.7)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e16.4 (11.2)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eICU stay average in days (SD)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e3.2 (4.8)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e6.8 (8.3)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e2.2 (2.4)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003ePerioperative transfusion (%)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e86 (28.8)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e33 (51.6)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e53 (22.6)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eUICC\u0026nbsp;Stage\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e1 (%)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e82 (27.4)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e7 (10.9)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e75 (31.9)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e2 (%)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e66 (22.0)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e10 (15.6)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e56 (23.8)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e3 (%)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e60 (20.1)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e16 (25)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e44 (18.7)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e4 (%)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e91 (30.4)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e31 (48.4)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e60 (25.5)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003c/tbody\u003e\u003c/colgroup\u003e\u003c/table\u003e\u003c/div\u003e\u003c/p\u003e\u003c/div\u003e\n\u003ch3\u003ePrediction model\u003c/h3\u003e\n\u003cp\u003eAutomatic variable selection resulted in the following set of variables, that contribute significantly to the model (marginal R\u003csup\u003e2\u003c/sup\u003e in brackets): ICU stay (46.1%), PEG (14.6%), age (12.4%), UICC (8.4%), psychotropic drugs (6.1%), antihypertensive drugs (4.0%). Table\u0026nbsp;\u003cspan refid=\"Tab2\" class=\"InternalRef\"\u003e2\u003c/span\u003e indicates the R\u003csup\u003e2\u003c/sup\u003e, p-value and regression coefficient of these variables. Prediction accuracy was at 85.3% correct assignments with the cross-validated confidence interval of 80.8% (10th percentile) and 88.9% (90th percentile). This has to be evaluated against the reference value for resampling (guessing based on prior probabilities) which is between 63.8% (10th percentile) and 69.2% (90th percentile).\u003c/p\u003e\u003cp\u003e\u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab2\" border=\"1\"\u003e\u003ccaption language=\"En\"\u003e\u003cdiv class=\"CaptionNumber\"\u003eTable 2\u003c/div\u003e\u003cdiv class=\"CaptionContent\"\u003e\u003cp\u003eVariables that contribute significantly to the model's marginal R\u003csup\u003e2\u003c/sup\u003e are listed on the left. R\u003csup\u003e2\u003c/sup\u003e is specified in percentage. All selected variables have a significant effect (p\u0026thinsp;\u0026lt;\u0026thinsp;0.05). The coefficient indicates, if the variable correlates positively (values\u0026thinsp;\u0026gt;\u0026thinsp;0) or negatively with the occurrence of POD (value\u0026thinsp;\u0026lt;\u0026thinsp;0).\u003c/p\u003e\u003c/div\u003e\u003c/caption\u003e\u003ccolgroup cols=\"4\"\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e\u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e\u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e\u003cthead\u003e\u003ctr\u003e\u003cth align=\"left\" colname=\"c1\"\u003e\u003cp\u003eVariable\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c2\"\u003e\u003cp\u003eR\u003csup\u003e2\u003c/sup\u003e\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c3\"\u003e\u003cp\u003ep-value\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c4\"\u003e\u003cp\u003eRegression coefficient\u003c/p\u003e\u003c/th\u003e\u003c/tr\u003e\u003c/thead\u003e\u003ctbody\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eLength of ICU stay\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e53.5\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e\u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e0.32\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003ePEG\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e17.9\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e\u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e(-)2.34\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eAge\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e10.8\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e\u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e0.05\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eUICC stage\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e11\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e\u003cp\u003e\u0026lt;\u0026thinsp;0.01\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e0.47\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003ePsychotropic drugs\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e4.1\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e\u003cp\u003e\u0026lt;\u0026thinsp;0.05\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e(-) 0.98\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eAntihypertensive drugs\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e2.7\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e\u003cp\u003e\u0026lt;\u0026thinsp;0.05\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e0.74\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003c/tbody\u003e\u003c/colgroup\u003e\u003c/table\u003e\u003c/div\u003e\u003c/p\u003e\u003cdiv id=\"Sec11\" class=\"Section2\"\u003e\u003ch2\u003eDevelopment of a risk assessment tool\u003c/h2\u003e\u003cp\u003eDue to the low R\u003csup\u003e2\u003c/sup\u003e value of the last two variables (psychotropic drugs R\u003csup\u003e2\u003c/sup\u003e\u0026thinsp;=\u0026thinsp;4.1%, antihypertensive drugs\u0026thinsp;=\u0026thinsp;2.7%) we did not include them in the development of the risk assessment tool. The prediction accuracy of the condensed model only dropped from 85% correct assignments to 84% and therefore, were not considered clinically relevant. The confidence of the accuracy for the model lies between 79% (10th percentile) and 87% (90th percentile). The distribution of false negatives lies between 9% (10th percentile) and 18% (90th percentile) with a mean of 13.2%. False positives range far lower with a mean of 3.3% (10th percentile\u0026thinsp;=\u0026thinsp;1%; 90th percentile\u0026thinsp;=\u0026thinsp;6%). It is deployed and available online (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://zkm-uni-freiburg.shinyapps.io/DelirmodelApp/\u003c/span\u003e\u003cspan address=\"https://zkm-uni-freiburg.shinyapps.io/DelirmodelApp/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e).\u003c/p\u003e\u003c/div\u003e"},{"header":"Discussion","content":"\u003cdiv id=\"Sec13\" class=\"Section2\"\u003e\u003ch2\u003eMain results\u003c/h2\u003e\u003cp\u003eIn recent decades, numerous studies have investigated risk factors for postoperative delirium (POD) in general, as well as in patients following oral cancer surgery. The incidence of POD in our cohort was 21.4%, which is comparable to the reported incidences in the literature (Hasegawa et al., \u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e2015\u003c/span\u003e; Makiguchi et al., \u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e2018\u003c/span\u003e; Takahashi et al., \u003cspan citationid=\"CR39\" class=\"CitationRef\"\u003e2021\u003c/span\u003e; Ying et al., \u003cspan citationid=\"CR45\" class=\"CitationRef\"\u003e2025\u003c/span\u003e). Unlike the vast majority of available literature, this study is one of the few studies that omitted manual selection of factors, instead developing an automated variable selection process to eliminate selection bias. The most significant factors correlating with the occurrence of POD were determined to be length of ICU stay, PEG, age, UICC stage, and the use of antipsychotic and antihypertensive drugs. To our knowledge, our study is the largest investigating risk factors for POD following surgery for OSCC in western countries.\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec14\" class=\"Section2\"\u003e\u003ch2\u003eClinical and scientific relevance of significant factors\u003c/h2\u003e\u003cp\u003eOf all variables, length of ICU stay contributed most to the developed risk prediction model (R\u003csup\u003e2\u003c/sup\u003e\u0026thinsp;=\u0026thinsp;46.1%). This is in line with recent findings of Ying et al. in a large cohort of Chinese patients following oral cancer surgery (Ying et al., \u003cspan citationid=\"CR45\" class=\"CitationRef\"\u003e2025\u003c/span\u003e), but has also been described in other patient cohorts (Kirfel et al., \u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e2022\u003c/span\u003e). A high noise level and unfavourable light conditions on the ICU lead to disturbance of circadian rhythm and a disrupted sleep of poor quality (Elliott et al., \u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e2013\u003c/span\u003e; Showler et al., \u003cspan citationid=\"CR37\" class=\"CitationRef\"\u003e2024\u003c/span\u003e), which is thought to be related to POD (Kamdar et al., \u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e2016\u003c/span\u003e). Furthermore, critical illness and a prolonged ICU stay are associated with a variety of physical and cognitive impairments (Hiser et al., \u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e2023\u003c/span\u003e; Lee et al., \u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e2025\u003c/span\u003e).\u003c/p\u003e\u003cp\u003eOf course, it has also to be considered the presence of delirium itself may have influenced the length of ICU stay (Ma et al., \u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e2023\u003c/span\u003e; Perry et al., \u003cspan citationid=\"CR31\" class=\"CitationRef\"\u003e2024\u003c/span\u003e), which might explain the comparatively strong statistical influence of this factor.\u003c/p\u003e\u003cp\u003eThe presence of a PEG-tube was identified as the second most significant factor (R\u003csup\u003e2\u003c/sup\u003e\u0026thinsp;=\u0026thinsp;14.6%). In contrast to most of the investigated factors, the presence of a PEG-tube revealed to be a protective factor (coefficient = -2.34). It should be noted that PEG tube placement was performed at different times among the study group. While some patients received PEG tube placement preoperatively, others received it during their stay in the ICU or surgical ward. The protective effect of PEG-tubes might be explained by the improvement of perioperative nutrition and by avoidance of a nasogastric tube which both have been identified as predisposing factors for POD (Fields et al., \u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e2018\u003c/span\u003e; Moellmann et al., \u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e2024\u003c/span\u003e). Recent studies evaluating the effects of PEG tube placement in both stroke and neurosurgical patients found PEG feeding to be a safe and well-tolerated method of ensuring enteral nutrition. However, no correlation with an improvement in the patients' neurological status was found (Koc et al., \u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e2007\u003c/span\u003e; Rajalbandi et al., \u003cspan citationid=\"CR33\" class=\"CitationRef\"\u003e2024\u003c/span\u003e). Based on our observations, we therefore encourage practitioners involved in the perioperative care of OSCC patients to consider preoperative PEG tube placement to ensure safe enteral feeding particularly for patients with expected postoperative impairment of swallowing and chewing.\u003c/p\u003e\u003cp\u003eAge also contributed significantly to the prediction model (R\u003csup\u003e2\u003c/sup\u003e\u0026thinsp;=\u0026thinsp;12.4%). Advanced age is often accompanied by frailty, an increased vulnerability to external stressors, which trigger pathologic changes in cognition and attention (Evered et al., \u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e2022\u003c/span\u003e; Quinlan et al., \u003cspan citationid=\"CR32\" class=\"CitationRef\"\u003e2011\u003c/span\u003e). Furthermore, disturbed perioperative cerebral blood flow seems to be a relevant factor increasing the incidence of neurocognitive disorders in elderly patients (NeuroVISION Investigators, \u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e2019\u003c/span\u003e). Indeed, age has been identified as a highly relevant risk factor in previous studies on risk factors for POD in general (Mevorach et al., \u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e2023\u003c/span\u003e) but also in patients following surgery for OSCC (Hasegawa et al., \u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e2015\u003c/span\u003e; Shiiba et al., \u003cspan citationid=\"CR36\" class=\"CitationRef\"\u003e2009\u003c/span\u003e; Ying et al., \u003cspan citationid=\"CR45\" class=\"CitationRef\"\u003e2025\u003c/span\u003e).\u003c/p\u003e\u003cp\u003eHigher UICC stage was also identified as a predictor for POD in our cohort (R\u003csup\u003e2\u003c/sup\u003e\u0026thinsp;=\u0026thinsp;8.4%). This association was equally found following head and neck surgery (Densky et al., \u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e2019\u003c/span\u003e), but was not included in most studies following oral cancer surgery (Hasegawa et al., \u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e2015\u003c/span\u003e; Makiguchi et al., \u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e2018\u003c/span\u003e; Ying et al., \u003cspan citationid=\"CR45\" class=\"CitationRef\"\u003e2025\u003c/span\u003e). Higher UICC stages usually require a more invasive approach leading to an increased surgical inflammatory stress response, known to cause neuro-inflammation and trigger perioperative neurocognitive disorders (Evered et al., \u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e2022\u003c/span\u003e). (Cerejeira et al., \u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e2013\u003c/span\u003e).\u003c/p\u003e\u003cp\u003eTwo other factors contributed significantly, albeit to lesser extent, to our prediction model. Whereas therapy with psychotropic drugs had a protective effect, therapy with antihypertensive drugs had a risk-increasing effect for the development of POD. Due to low R\u003csup\u003e2\u003c/sup\u003e values, both factors were not included in the developed risk assessment tool.\u003c/p\u003e\u003cp\u003ePsychiatric disorders in contrast showed no significant association with POD, whereas this was reported in patients following head and neck surgery (Choi et al., \u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e2015\u003c/span\u003e). It is worth noting that in our cohort less than 60% of patients with psychiatric disorders received psychotropic drugs. Assuming that untreated psychiatric disorder particularly increases the risk for POD, the protective influence of the treatment with psychotropic drugs seems plausible.\u003c/p\u003e\u003cp\u003eThe association of POD with arterial hypertension and antihypertensives has been described before in patients undergoing maxillofacial tumour surgery (Kong et al., \u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e2021\u003c/span\u003e), but also in a wide variety of patient collectives and in different types of neurocognitive disorders (Huang and Aronow, \u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e2024\u003c/span\u003e; Ormseth et al., \u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e2023\u003c/span\u003e). The underlying mechanisms of this association are complex and bidirectional. Disturbed cerebral autoregulation of blood flow in hypertensive patients might play a role in this context (Futier et al., \u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e2017\u003c/span\u003e; Saugel and Sessler, \u003cspan citationid=\"CR35\" class=\"CitationRef\"\u003e2021\u003c/span\u003e).\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec15\" class=\"Section2\"\u003e\u003ch2\u003eStatistical analysis\u003c/h2\u003e\u003cp\u003eA common approach that has been used to identify factors associated with POD is to search statistically significantly different distribution in patients with and without delirium by a univariate approach and then perform a subsequent logistic regression analysis (Hasegawa et al., \u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e2015\u003c/span\u003e; Makiguchi et al., \u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e2018\u003c/span\u003e; Takahashi et al., \u003cspan citationid=\"CR39\" class=\"CitationRef\"\u003e2021\u003c/span\u003e; Ying et al., \u003cspan citationid=\"CR45\" class=\"CitationRef\"\u003e2025\u003c/span\u003e). Yet, pre-selection of variables from a very large set of potential predictors based on their significance in a univariate model leads to small p-values and overestimation of effects in a multivariate model (Steyerberg and Van Calster, \u003cspan citationid=\"CR38\" class=\"CitationRef\"\u003e2020\u003c/span\u003e; Vach, \u003cspan citationid=\"CR40\" class=\"CitationRef\"\u003e2013\u003c/span\u003e). Thus, the use of bias reducing automatic variable selection may be a feasible tool to avoid overestimation of effects. Our study is also assessing the model\u0026rsquo;s strengths and shortcomings against the backdrop of results obtained by guessing (random assignment) based on the prior probabilities as well as reporting confidence intervals based on bootstrapping procedures. This is important as the sample is unbalanced with only about 21% patients being affected by POD and simply reporting prediction accuracy might lead to overconfidence regarding the model\u0026rsquo;s predictive power. However, our statistical approach requires a complete dataset and does not permit any missing values. This poses a certain challenge especially for retrospective studies. Therefore, some variables that have been reported as risk factors before could not be included in our analysis.\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec16\" class=\"Section2\"\u003e\u003ch2\u003eLimitations\u003c/h2\u003e\u003cp\u003eSeveral limitations of the study including, but not limited to its retrospective design have to be considered. Patients were predominantly screened with NU-DESC which is a validated tool known to have a high sensitivity even without special training, yet has a lower specificity compared to other Delirium screening tools (Aldecoa et al., \u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e2024\u003c/span\u003e; Neufeld et al., \u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e2013\u003c/span\u003e). Furthermore, patients were not screened systematically for preoperative cognitive disorders, which increases the risk for POD (Weiss et al., \u003cspan citationid=\"CR42\" class=\"CitationRef\"\u003e2023\u003c/span\u003e) and thus represents a neglected factor in our study. Prospective studies are needed to specify risk profiles in oral cancer patients and may offer opportunities to optimize our prediction tool for a future establishment in clinical practice.\u003c/p\u003e\u003c/div\u003e"},{"header":"Conclusion","content":"\u003cp\u003eOur study revealed four factors increasing and two factors reducing the risk for POD by the use of a bias reducing automatic variable selection method. The resulting online risk assessment tool might offer an easy-to-implement option to identify patients at risk for POD, which however has to be validated in clinical practice.\u003c/p\u003e"},{"header":"Declarations","content":"\n\u003cp\u003e\u003cstrong\u003eFunding Statement\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe authors declare that no external funding was received for this study.\u003c/p\u003e\n\n\u003cp\u003e\u003cstrong\u003eData availability statement\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eDue to institutional, ethical, and privacy data-protection policies, the underlying individual patient-level data cannot be made publicly available. Access may be provided upon reasonable request, subject to approval by the relevant institutional review board and in compliance with applicable regulations.\u003c/p\u003e\n\n\u003cp\u003e\u003cstrong\u003eAuthor contribution statement\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eConceptualization \u0026amp; Study design: Matthias Jakob Posch, Johannes Kalbhenn, Marc Metzger, Julia Vera Brandenburg, Leonard Simon Brandenburg. Methodology: Stefan Schlager, Leonard Simon Brandenburg, Matthias Jakob Posch. Data curation: Konstantin Hasel, Julia Vera Brandenburg, Leonard Simon Brandenburg. Formal analysis \u0026amp; Statistics: Stefan Schlager, Matthias Jakob Posch, Leonard Simon Brandenburg. Software / Risk prediction tool development: Stefan Schlager, Matthias Jakob Posch, Leonard Simon Brandenburg. Visualization (tables, figures, online tool): Stefan Schlager, Matthias Jakob Posch. Writing \u0026ndash; original draft: Matthias Jakob Posch, Julia Vera Brandenburg, Leonard Simon Brandenburg. Writing \u0026ndash; review \u0026amp; editing: Johannes Kalbhenn, Julia Brandenburg, Jonas W\u0026uuml;ster, Damian Sommer, Marc Metzger Rainer Schmelzeisen. Supervision: Johannes Kalbhenn, Marc Metzger, Rainer Schmelzeisen.\u003c/p\u003e\n\n\u003cp\u003e\u003cstrong\u003eCompeting Interests \u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe authors declare that they have no competing interests. \u003c/p\u003e\n\n\u003cp\u003e\u003cstrong\u003eEthics approval and consent to participate\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe study was approved by the Ethics Committee of the University of Freiburg, Germany (No. 127/15 and No. 30/20). The requirement for informed consent was waived due to the retrospective design of the study.\u003c/p\u003e\n\n\u003cp\u003e\u003cstrong\u003eConsent for publication\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThis manuscript does not contain any individual person\u0026rsquo;s data. Therefore, consent for publication was not required.\u003c/p\u003e\n"},{"header":"References","content":"\u003col\u003e\n\u003cli\u003eAldecoa, C., Bettelli, G., Bilotta, F., Sanders, R.D., Aceto, P., Audisio, R., Cherubini, A., Cunningham, C., Dabrowski, W., Forookhi, A., Gitti, N., Immonen, K., Kehlet, H., Koch, S., Kotfis, K., Latronico, N., MacLullich, A.M.J., Mevorach, L., Mueller, A., Neuner, B., Piva, S., Radtke, F., Blaser, A.R., Renzi, S., Romagnoli, S., Schubert, M., Slooter, A.J.C., Tommasino, C., Vasiljewa, L., Weiss, B., Yuerek, F., Spies, C.D., 2024. Update of the European Society of Anaesthesiology and Intensive Care Medicine evidence-based and consensus-based guideline on postoperative delirium in adult patients. Eur J Anaesthesiol 41, 81\u0026ndash;108. https://doi.org/10.1097/EJA.0000000000001876\u003c/li\u003e\n\u003cli\u003eBray, F., Laversanne, M., Sung, H., Ferlay, J., Siegel, R.L., Soerjomataram, I., Jemal, A., 2024. Global cancer statistics 2022: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries. CA Cancer J Clin 74, 229\u0026ndash;263. https://doi.org/10.3322/caac.21834\u003c/li\u003e\n\u003cli\u003eCerejeira, J., Batista, P., Nogueira, V., Vaz-Serra, A., Mukaetova-Ladinska, E.B., 2013. The stress response to surgery and postoperative delirium: evidence of hypothalamic-pituitary-adrenal axis hyperresponsiveness and decreased suppression of the GH/IGF-1 Axis. J Geriatr Psychiatry Neurol 26, 185\u0026ndash;194. https://doi.org/10.1177/0891988713495449\u003c/li\u003e\n\u003cli\u003eChamoli, A., Gosavi, A.S., Shirwadkar, U.P., Wangdale, K.V., Behera, S.K., Kurrey, N.K., Kalia, K., Mandoli, A., 2021. Overview of oral cavity squamous cell carcinoma: Risk factors, mechanisms, and diagnostics. Oral Oncol 121, 105451. https://doi.org/10.1016/j.oraloncology.2021.105451\u003c/li\u003e\n\u003cli\u003eChang, W., Cheng, J., Allaire, J., Sievert, C., Schloerke, B., Xie, Y., Allen, J., McPherson, J., Dipert, A., Borges, B., 2012. shiny: Web Application Framework for R. https://doi.org/10.32614/CRAN.package.shiny\u003c/li\u003e\n\u003cli\u003eChoi, G.J., Baek, C.W., Kang, H., Park, Y.H., Yang, S.Y., Shin, H.Y., Jung, Y.H., Woo, Y.C., Lee, U.L., 2015. Emergence agitation after orthognathic surgery: a randomised controlled comparison between sevoflurane and desflurane. Acta Anaesthesiol Scand 59, 224\u0026ndash;231. https://doi.org/10.1111/aas.12435\u003c/li\u003e\n\u003cli\u003eDensky, J., Eskander, A., Kang, S., Chan, J., Tweel, B., Sitapara, J., Ozer, E., Agrawal, A., Carrau, R., Rocco, J., Teknos, T.N., Old, M., 2019. Risk Factors Associated With Postoperative Delirium in Patients Undergoing Head and Neck Free Flap Reconstruction. JAMA Otolaryngol Head Neck Surg 145, 216\u0026ndash;221. https://doi.org/10.1001/jamaoto.2018.3820\u003c/li\u003e\n\u003cli\u003eDilmen, O.K., Meco, B.C., Evered, L.A., Radtke, F.M., 2024. Postoperative neurocognitive disorders: A clinical guide. J Clin Anesth 92, 111320. https://doi.org/10.1016/j.jclinane.2023.111320\u003c/li\u003e\n\u003cli\u003eElliott, R., McKinley, S., Cistulli, P., Fien, M., 2013. Characterisation of sleep in intensive care using 24-hour polysomnography: an observational study. Crit Care 17, R46. https://doi.org/10.1186/cc12565\u003c/li\u003e\n\u003cli\u003eEvered, L., Atkins, K., Silbert, B., Scott, D.A., 2022. Acute peri-operative neurocognitive disorders: a narrative review. Anaesthesia 77 Suppl 1, 34\u0026ndash;42. https://doi.org/10.1111/anae.15613\u003c/li\u003e\n\u003cli\u003eFalegnami, A., Patriarca, R., Costantino, F., Di Gravio, G., Bilotta, F., 2021. Surveying work-as-done in post-operative delirium risk factors collection and diagnosis monitoring. Appl Ergon 92, 103347. https://doi.org/10.1016/j.apergo.2020.103347\u003c/li\u003e\n\u003cli\u003eFields, A., Huang, J., Schroeder, D., Sprung, J., Weingarten, T., 2018. Agitation in adults in the post-anaesthesia care unit after general anaesthesia. Br J Anaesth 121, 1052\u0026ndash;1058. https://doi.org/10.1016/j.bja.2018.07.017\u003c/li\u003e\n\u003cli\u003eFutier, E., Lefrant, J.-Y., Guinot, P.-G., Godet, T., Lorne, E., Cuvillon, P., Bertran, S., Leone, M., Pastene, B., Piriou, V., Molliex, S., Albanese, J., Julia, J.-M., Tavernier, B., Imhoff, E., Bazin, J.-E., Constantin, J.-M., Pereira, B., Jaber, S., INPRESS Study Group, 2017. Effect of Individualized vs Standard Blood Pressure Management Strategies on Postoperative Organ Dysfunction Among High-Risk Patients Undergoing Major Surgery: A Randomized Clinical Trial. JAMA 318, 1346\u0026ndash;1357. https://doi.org/10.1001/jama.2017.14172\u003c/li\u003e\n\u003cli\u003eHasegawa, T., Saito, I., Takeda, D., Iwata, E., Yonezawa, N., Kakei, Y., Sakakibara, A., Akashi, M., Minamikawa, T., Komori, T., 2015. Risk factors associated with postoperative delirium after surgery for oral cancer. J Craniomaxillofac Surg 43, 1094\u0026ndash;1098. https://doi.org/10.1016/j.jcms.2015.06.011\u003c/li\u003e\n\u003cli\u003eHiser, S.L., Fatima, A., Ali, M., Needham, D.M., 2023. Post-intensive care syndrome (PICS): recent updates. J Intensive Care 11, 23. https://doi.org/10.1186/s40560-023-00670-7\u003c/li\u003e\n\u003cli\u003eHuang, L., Aronow, W.S., 2024. Association of Hypertension with Different Cognitive Disorders. J Clin Med 13, 6029. https://doi.org/10.3390/jcm13206029\u003c/li\u003e\n\u003cli\u003eKamdar, B.B., Martin, J.L., Needham, D.M., Ong, M.K., 2016. Promoting Sleep to Improve Delirium in the ICU. Crit Care Med 44, 2290\u0026ndash;2291. https://doi.org/10.1097/CCM.0000000000001982\u003c/li\u003e\n\u003cli\u003eKirfel, A., Guttenthaler, V., Mayr, A., Coburn, M., Menzenbach, J., Wittmann, M., 2022. Postoperative delirium is an independent factor influencing the length of stay of elderly patients in the intensive care unit and in hospital. J Anesth 36, 341\u0026ndash;348. https://doi.org/10.1007/s00540-022-03049-4\u003c/li\u003e\n\u003cli\u003eKoc, D., Gercek, A., Gencosmanoglu, R., Tozun, N., 2007. Percutaneous endoscopic gastrostomy in the neurosurgical intensive care unit: complications and outcome. JPEN J Parenter Enteral Nutr 31, 517\u0026ndash;520. https://doi.org/10.1177/0148607107031006517\u003c/li\u003e\n\u003cli\u003eKong, S., Wang, J., Xu, H., Wang, K., 2021. Effect of hypertension and medication use regularity on postoperative delirium after maxillofacial tumors radical surgery. Oncotarget 12, 1811\u0026ndash;1820. https://doi.org/10.18632/oncotarget.28048\u003c/li\u003e\n\u003cli\u003eLander, H.L., Dick, A.W., Joynt Maddox, K.E., Oldham, M.A., Fleisher, L.A., Mazzeffi, M., Lustik, S.J., Shang, J., Stone, P.W., Gloff, M.S., Nadler, J., Wu, I., Zollo, R., Glance, L.G., 2025. Postoperative Delirium in Older Adults Undergoing Noncardiac Surgery. JAMA Netw Open 8, e2519467. https://doi.org/10.1001/jamanetworkopen.2025.19467\u003c/li\u003e\n\u003cli\u003eLee, S.Y., Huh, J.W., Hong, S.-B., Lim, C.-M., Ahn, J.H., 2025. Physical and Cognitive Impairments at ICU Discharge are Associated with High Long-Term Mortality in ICU Survivors with Solid Malignancies: A Retrospective Cohort Study. Ther Clin Risk Manag 21, 1121\u0026ndash;1133. https://doi.org/10.2147/TCRM.S520206\u003c/li\u003e\n\u003cli\u003eMa, Y., Li, C., Peng, W., Wan, Q., 2023. The influence of delirium on mortality and length of ICU stay and analysis of risk factors for delirium after liver transplantation. Front Neurol 14, 1229990. https://doi.org/10.3389/fneur.2023.1229990\u003c/li\u003e\n\u003cli\u003eMakiguchi, T., Yokoo, S., Kurihara, J., 2018. Risk factors for postoperative delirium in patients undergoing free flap reconstruction for oral cancer. Int J Oral Maxillofac Surg 47, 998\u0026ndash;1002. https://doi.org/10.1016/j.ijom.2018.03.011\u003c/li\u003e\n\u003cli\u003eMevorach, L., Forookhi, A., Farcomeni, A., Romagnoli, S., Bilotta, F., 2023. Perioperative risk factors associated with increased incidence of postoperative delirium: systematic review, meta-analysis, and Grading of Recommendations Assessment, Development, and Evaluation system report of clinical literature. Br J Anaesth 130, e254\u0026ndash;e262. https://doi.org/10.1016/j.bja.2022.05.032\u003c/li\u003e\n\u003cli\u003eMoellmann, H.L., Alhammadi, E., Boulghoudan, S., Kuhlmann, J., Mevissen, A., Olbrich, P., Rahm, L., Frohnhofen, H., 2024. Risk of sarcopenia, frailty and malnutrition as predictors of postoperative delirium in surgery. BMC Geriatr 24, 971. https://doi.org/10.1186/s12877-024-05566-1\u003c/li\u003e\n\u003cli\u003eNeufeld, K.J., Leoutsakos, J.S., Sieber, F.E., Joshi, D., Wanamaker, B.L., Rios-Robles, J., Needham, D.M., 2013. Evaluation of two delirium screening tools for detecting post-operative delirium in the elderly. Br J Anaesth 111, 612\u0026ndash;618. https://doi.org/10.1093/bja/aet167\u003c/li\u003e\n\u003cli\u003eNeuroVISION Investigators, 2019. Perioperative covert stroke in patients undergoing non-cardiac surgery (NeuroVISION): a prospective cohort study. Lancet 394, 1022\u0026ndash;1029. https://doi.org/10.1016/S0140-6736(19)31795-7\u003c/li\u003e\n\u003cli\u003eObermeier, K.T., Kraus, M., Smolka, W., Henkel, J., Saller, T., Otto, S., Liokatis, P., 2022. Postoperative Delirium in Patients with Oral Cancer: Is Intraoperative Fluid Administration a Neglected Risk Factor? Cancers (Basel) 14, 3176. https://doi.org/10.3390/cancers14133176\u003c/li\u003e\n\u003cli\u003eOrmseth, C.H., LaHue, S.C., Oldham, M.A., Josephson, S.A., Whitaker, E., Douglas, V.C., 2023. Predisposing and Precipitating Factors Associated With Delirium: A Systematic Review. JAMA Netw Open 6, e2249950. https://doi.org/10.1001/jamanetworkopen.2022.49950\u003c/li\u003e\n\u003cli\u003ePerry, H., Alight, A., Wilcox, M.E., 2024. Light, sleep and circadian rhythm in critical illness. Curr Opin Crit Care 30, 283\u0026ndash;289. https://doi.org/10.1097/MCC.0000000000001163\u003c/li\u003e\n\u003cli\u003eQuinlan, N., Marcantonio, E.R., Inouye, S.K., Gill, T.M., Kamholz, B., Rudolph, J.L., 2011. Vulnerability: the crossroads of frailty and delirium. J Am Geriatr Soc 59 Suppl 2, S262-268. https://doi.org/10.1111/j.1532-5415.2011.03674.x\u003c/li\u003e\n\u003cli\u003eRajalbandi, R.S., Qureshi, A., Bains, N., Gillani, S., Maqsood, H., 2024. Effect of Percutaneous Endoscopic Gastrostomy Tube Placement on Confusion/Delirium in Stroke Patients with Dysphagia (P7-5.007). Neurology 102. https://doi.org/10.1212/wnl.0000000000205070\u003c/li\u003e\n\u003cli\u003eSahle, B.W., Pilcher, D., Litton, E., Ofori-Asenso, R., Peter, K., McFadyen, J., Bucknall, T., 2022. Association between frailty, delirium, and mortality in older critically ill patients: a binational registry study. Ann Intensive Care 12, 108. https://doi.org/10.1186/s13613-022-01080-y\u003c/li\u003e\n\u003cli\u003eSaugel, B., Sessler, D.I., 2021. Perioperative Blood Pressure Management. Anesthesiology 134, 250\u0026ndash;261. https://doi.org/10.1097/ALN.0000000000003610\u003c/li\u003e\n\u003cli\u003eShiiba, M., Takei, M., Nakatsuru, M., Bukawa, H., Yokoe, H., Uzawa, K., Tanzawa, H., 2009. Clinical observations of postoperative delirium after surgery for oral carcinoma. Int J Oral Maxillofac Surg 38, 661\u0026ndash;665. https://doi.org/10.1016/j.ijom.2009.01.011\u003c/li\u003e\n\u003cli\u003eShowler, L., Deane, A.M., Litton, E., Ankravs, M.J., Wibrow, B., Barge, D., Goldin, J., Hammond, N., Saxena, M.K., Young, P.J., Venkatesh, B., Finnis, M., Abdelhamid, Y.A., 2024. A multicentre point prevalence study of nocturnal hours awake and enteral pharmacological sleep aids in patients admitted to Australian and New Zealand intensive care units. Crit Care Resusc 26, 192\u0026ndash;197. https://doi.org/10.1016/j.ccrj.2024.06.009\u003c/li\u003e\n\u003cli\u003eSteyerberg, E.W., Van Calster, B., 2020. Redefining significance and reproducibility for medical research: A plea for higher P-value thresholds for diagnostic and prognostic models. Eur J Clin Invest 50, e13229. https://doi.org/10.1111/eci.13229\u003c/li\u003e\n\u003cli\u003eTakahashi, N., Hiraki, A., Kawahara, K., Nagata, M., Yoshida, R., Matsuoka, Y., Tanaka, T., Obayashi, Y., Sakata, J., Nakashima, H., Arita, H., Shinohara, M., Nakayama, H., 2021. Postoperative delirium in patients undergoing tumor resection with reconstructive surgery for oral cancer. Mol Clin Oncol 14, 60. https://doi.org/10.3892/mco.2021.2222\u003c/li\u003e\n\u003cli\u003eVach, W., 2013. Regression models as a tool in medical research, A Chapman \u0026amp; Hall book. CRC Press, Boca Raton, Fla.\u003c/li\u003e\n\u003cli\u003eVanegas, L.H., Rond\u0026oacute;n, L.M., Paula, G.A., 2021. glmtoolbox: Set of Tools to Data Analysis using Generalized Linear Models. CRAN: Contributed Packages. https://doi.org/10.32614/cran.package.glmtoolbox\u003c/li\u003e\n\u003cli\u003eWeiss, Y., Zac, L., Refaeli, E., Ben-Yishai, S., Zegerman, A., Cohen, B., Matot, I., 2023. Preoperative Cognitive Impairment and Postoperative Delirium in Elderly Surgical Patients: A Retrospective Large Cohort Study (The CIPOD Study). Ann Surg 278, 59\u0026ndash;64. https://doi.org/10.1097/SLA.0000000000005657\u003c/li\u003e\n\u003cli\u003eWitlox, J., Eurelings, L.S.M., de Jonghe, J.F.M., Kalisvaart, K.J., Eikelenboom, P., van Gool, W.A., 2010. Delirium in elderly patients and the risk of postdischarge mortality, institutionalization, and dementia: a meta-analysis. JAMA 304, 443\u0026ndash;451. https://doi.org/10.1001/jama.2010.1013\u003c/li\u003e\n\u003cli\u003eWolff, K.D., Rau, A., Weitz, J., 2021. S3-Leitlinie Mundh\u0026ouml;hlenkarzinom, AWMF Registernummer: 007/100OL (No. AWMF Registernummer: 007/100OL).\u003c/li\u003e\n\u003cli\u003eYing, C., Xiaona, L., Aili, Z., Zengxiang, W., Ying, W., Yu, P., Hongbo, Z., Danni, W., Meiping, J., Hongyuan, D., 2025. Development and validation of a nomogram model for predicting postoperative delirium in elderly patients with oral cancer: a retrospective study. BMC Oral Health 25, 990. https://doi.org/10.1186/s12903-025-06167-z\u003c/li\u003e\n\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"bmc-oral-health","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"ohea","sideBox":"Learn more about [BMC Oral Health](http://bmcoralhealth.biomedcentral.com/)","snPcode":"","submissionUrl":"https://www.editorialmanager.com/ohea/default.aspx","title":"BMC Oral Health","twitterHandle":"BMC_series","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"em","reportingPortfolio":"BMC Series","inReviewEnabled":true,"inReviewRevisionsEnabled":true},"keywords":"Postoperative delirium, prediction model, oral cancer, risk factors","lastPublishedDoi":"10.21203/rs.3.rs-7687245/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-7687245/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eOral squamous cell carcinoma (OSCC) is a common malignancy that is commonly treated by surgical resection and - if necessary - with adjuvant radio(chemo)therapy. Depending on pre-existing patient related conditions and extent of surgery, postoperative intensive care may be indicated. Postoperative delirium (POD) is a frequent and serious complication in the intensive care unit (ICU), characterized by confusion and cognitive disturbances of the patient. This single-centre, retrospective study (2013\u0026ndash;2023, University of Freiburg) analysed 299 OSCC patients admitted to ICU post-surgery to identify risk factors and develop a risk assessment tool for POD.\u003c/p\u003e\u003cp\u003eA bias-reducing automated variable selection method identified ICU stay length, the use of percutaneous endoscopic gastrostomy (PEG) tubes, age, cancer stage, use of psychotropic and antihypertensive drugs to be associated with POD. Key predictors were used to create an online assessment tool, which predicted POD with about 85% accuracy. Notably, longer ICU stays increased POD risk, while the use of PEG tubes appeared to have a protective effect. Prospective studies are needed to validate the tool and to assess its clinical practicability.\u003c/p\u003e","manuscriptTitle":"Development of a risk prediction model for delirium following surgery for oral squamous cell cancer","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-11-19 16:21:44","doi":"10.21203/rs.3.rs-7687245/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"editorInvitedReview","content":"","date":"2025-11-26T13:38:44+00:00","index":"hide","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2025-11-23T14:52:41+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"250354351713367952000784009172240340000","date":"2025-11-18T11:11:32+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"30236032788203053675312014168534624922","date":"2025-11-15T14:02:20+00:00","index":"hide","fulltext":""},{"type":"reviewersInvited","content":"","date":"2025-11-10T13:55:36+00:00","index":"","fulltext":""},{"type":"editorInvited","content":"","date":"2025-10-09T16:33:42+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2025-10-08T04:56:29+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2025-10-08T04:53:42+00:00","index":"","fulltext":""},{"type":"submitted","content":"BMC Oral Health","date":"2025-09-22T17:49:00+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"bmc-oral-health","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"ohea","sideBox":"Learn more about [BMC Oral Health](http://bmcoralhealth.biomedcentral.com/)","snPcode":"","submissionUrl":"https://www.editorialmanager.com/ohea/default.aspx","title":"BMC Oral Health","twitterHandle":"BMC_series","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"em","reportingPortfolio":"BMC Series","inReviewEnabled":true,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"92b3764e-3fb5-47a3-8402-1b20a3873d00","owner":[],"postedDate":"November 19th, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"under-review","subjectAreas":[],"tags":[],"updatedAt":"2025-11-19T16:21:44+00:00","versionOfRecord":[],"versionCreatedAt":"2025-11-19 16:21:44","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-7687245","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-7687245","identity":"rs-7687245","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}