Comparison between the WHO-CFICPS and the PRISMA classification of safety-related events in a radiation oncology department.

INTRODUCTION: Describing Safety-Related Events (SREs) in a radiotherapy (RT) department and comparing WHO-CFICPS (World Health Organization's Conceptual Framework For The International Classification For Patient Safety) and PRISMA (Prevention and Recovery Information System for Monitoring and Analysis) methods for classifying SREs. METHODS: From February 2017 to October 2020, two Quality Managers (QMs) randomly classified 1173 SREs using 13 incident types of WHO-CFICPS. The same two QMs, reclassified the same SREs according to 20 PRISMA incident codes. Statistical analysis was performed to assess the association between the 13 incident types of WHO-CFICPS and the 20 PRISMA codes. The chi-squared and post-hoc tests using adjusted standardized residuals were applied to detect the association between the two systems. RESULTS: There was a significant association between WHO-CFICPS incident types and PRISMA codes (P < 0.001). Ninety-two percent of all SREs were categorized using 4 of 13 WHO-CFICPS incident types including Clinical Process/Procedure (n = 448, 38.2%), Clinical Administration (n = 248, 21.1%), Documentation (n = 226, 19.2%) and Resources/Organizational Management (n = 15,613.3%). According to PRISMA classification, 14 of the 20 codes were used to describe the same SREs. PRISMA captured 41 Humans Skill Slips from 226 not better defined WHO-CFICPS Documentation Incidents, 38 Human Rule-based behaviour Qualification from not better defined 447 Clinical Process/Procedure and 40 Organization Management priority events from 156 not better defined WHO-CFICPS Resources/Organizational Management events (P < 0.001). CONCLUSION: Although there was a significant association between WHO-CFICPS and PRISMA, The PRISMA method provides a more detailed insight into SREs compared to WHO-CFICPS in a RT department.

Compliance with Barrier Gestures during COVID-19 Pandemic as a Function of the Context: A Longitudinal Observational Survey at the University of Liège.

During the COVID-19 pandemic, barrier gestures such as mask wearing, physical distancing, greetings without contact, one-way circulation flow, and hand sanitization were major strategies to prevent the spread of SARS-CoV-2, but they were only useful if consistently applied. This survey was a follow-up of the first survey performed in 2020 at the University of Liège. We aim to evaluate the compliance with these gestures on campuses and examine differences in the extent of the compliance observed in different educational activities and contexts. During 3.5 months, the counting of compliant and non-compliant behaviors was performed each week in randomly selected rooms. Using data collected during both surveys (2020 and 2021), binomial negative regression models of compliance depending on periods (teaching periods and exam sessions), type of rooms, and campuses were conducted to evaluate prevalence ratios of compliance. The percentage of compliance in this second survey was the highest for mask wearing and physical distancing during educational activities (90% and 88%, respectively) and lowest for physical distancing outside educational activities and hand sanitization (45% and 52%, respectively). Multivariate analyses revealed that the compliance with most gestures was significantly higher in teaching rooms than in hallways and restaurants and during exam sessions. The compliance with physical distancing was significantly higher (from 66%) in auditoriums, where students had to remain seated, than during practical works that allowed or required free movement. Therefore, the compliance with barrier gestures was associated with contextual settings, which should be considered when communicating and managing barrier gestures. Further studies should specify and confirm the determining contextual characteristics regarding the compliance with barrier gestures in times of pandemic.

Patients' self-triage for unscheduled urgent care: a preliminary study on the accuracy and factors affecting the performance of a Belgian self-triage platform.

BACKGROUND: Management of unscheduled urgent care is a complex concern for many healthcare providers. Facing the challenge of appropriately dispatching unscheduled care, primary and emergency physicians have collaboratively implemented innovative strategies such as telephone triage. Currently, new original solutions tend to emerge with the development of new technologies. We created an interactive patient self-triage platform, ODISSEE, and aimed to explore its accuracy and potential factors affecting its performance using clinical case scenarios. METHODS: The ODISSEE platform was developed based on previously validated triage protocols for out-of-hours primary care. ODISSEE is composed of 18 icons leading to algorithmic questions that finally provide an advised orientation (emergency or primary care services). To investigate ODISSEE performance, we used 100 clinical case scenarios, each associated with a preestablished orientation determined by a group of experts. Fifteen volunteers were asked to self-triage with 50 randomly selected scenarios using ODISSEE on a digital tablet. Their triage results were compared with the experts' references. RESULTS: The 15 participants performed a total of 750 self-triages, which matched the experts references regarding the level of care in 85.6% of the cases. The orientation was incorrect in 14.4%, with an undertriage rate of 1.9% and an overtriage rate of 12.5%. The tool's specificity and sensitivity to advise participants on the appropriate level of care were 69% (95% CI: 64-74) and 97% (95% CI: 95-98) respectively. When combined with advice on the level of urgency, the tool only found the correct orientation in 68.4% with 9.2% of undertriages and 22.4% of overtriages. Some participant characteristics and the types of medical conditions demonstrated a significant association with the tool performance. CONCLUSION: Self-triage apps, such as the ODISSEE platform, could represent an innovative method to allow patients to self-triage to the most appropriate level of care. This study based on clinical vignettes highlights some positive arguments regarding ODISSEE safety, but further research is needed to assess the generalizability of such tools to the population without equity issues.

"Debriefing and Organizational Lessons Learned" (DOLL): A Qualitative Study to Develop a Classification Framework for Reporting Clinical Debriefing Results.

Background: The COVID-19 crisis has radically affected our healthcare institutions. Debriefings in clinical settings provide a time for the clinicians to reflect on the successes (pluses) and difficulties (deltas) encountered. Debriefings tend to be well-received if included in the broader management of the unit. The goal of this study was to develop a framework to categorize these debriefings and to assess its worthiness. Methods: A qualitative approach based on a grounded theory research method was adopted resulting in the "Debriefing and Organizational Lessons Learned" (DOLL) framework. Debriefings were conducted within two Emergency Departments of a Belgian University Hospital during an 8-week period. In the first step, three researchers used debriefing transcripts to inductively develop a tentative framework. During the second step, these three researchers conducted independent categorizations of the debriefings using the developed framework. In step 3, the team analyzed the data to understand the utility of the framework. Chi-square was conducted to examine the associations between the item types (pluses and deltas) and the framework's dimensions. Results: The DOLL is composed of seven dimensions and 13 subdimensions. Applied to 163 debriefings, the model identified 339 items, including 97 pluses and 242 deltas. Results revealed that there was an association between the frequency of pluses and deltas and the dimensions (p < 0.001). The deltas were mainly related to the work environment (equipment and maintenance) (p < 0.001) while the pluses identified tended to be related to the organization of the unit (communication and roles) (p < 0.001). With leadership's support and subsequent actions, clinicians were more enthusiastic about participating and the researchers anecdotally detected a switch toward a more positive organizational learning approach. Conclusion: The framework increases the potential value of clinical debriefings because it organizes results into actionable areas. Indeed, leadership found the DOLL to be a useful management tool. Further research is needed to investigate how DOLL may work in non-crisis circumstances and further apply the DOLL into incident reporting and risk management process of the unit.

Consensus on the Objectives of an Educational Intervention for Patients with Oropharyngeal Dysphagia and Their Informal Caregivers: A Delphi Study.

Purpose: In the absence of literature allowing for an evidence-based approach to therapeutic patient education (TPE) in Oropharyngeal Dysphagia (OD), this study aims to reach a consensus of experts on the content of a competency framework of an educational program for OD patients and their informal caregivers (ICGs). Methods: We used the Delphi consensus-building method. Four categories of experts were recruited: 12 patients, 17 ICGs, 46 healthcare professionals (HCP) (experienced in OD, not necessarily certified in TPE), and 19 experts in TPE (trained individuals to set up and run TPE programs not necessarily HCPs). The content of the questionnaire of the first round (R) was established according to the result of a scoping review and the opinion of an expert committee. We carried out three rounds. In R1 and R2, we collected the opinions on the relevance (7-point Likert-type scale) and on comprehensiveness (YES/No question and asking participants to propose additional content). Participants were also invited to leave comments on each objective. In R3, we asked the participants to give their opinion about the relevance of the objectives again and asked them to rank the themes from highest to lowest priority. Results: Objectives were considered relevant for all participants if they reached consensus when the interquartile (IQR) ≤ 1, and if the median indicated agreement (Mdn ≥ 6) (6= appropriate, 7 = totally appropriate). Following three rounds, the final content of the educational program is composed of 23 educational objectives organized in 13 themes with an agreement about relevance amongst all participants (Mdn ≥ 6; IQR ≤ 1). The comprehensiveness criterion received also a consensus (IQR ≤ 1). The participants ranked the theme "normal swallowing vs difficulty swallowing" as the highest priority. Conclusion: This Delphi study resulted in a consensus, on the content of a competency framework of an educational program for OD patients and their ICGs. Further steps are needed to construct learning activities based on these objectives before testing their feasibility and efficacy.

A 2-month field cohort study of SARS-CoV-2 in saliva of BNT162b2 vaccinated nursing home workers.

Background: Nursing home (NH) residents have been severely affected during the COVID-19 pandemic because of their age and underlying comorbidities. Infection and outbreaks in NHs are most likely triggered by infected workers. Screening for asymptomatic NH workers can prevent risky contact and viral transmission to the residents. This study examined the effect of the BNT162b2 mRNA COVID­19 (Comirnaty®; BioNTech and Pfizer) vaccination on the saliva excretion of SARS-CoV-2 among NH workers, through weekly saliva RT-qPCR testing. Methods: A 2-month cohort study was conducted among 99 NHs in the Walloon region (Belgium), at the start of February 2021. Three groups of workers, i.e., non-vaccinated (n = 1618), one-dosed vaccinated (n = 1454), and two-dosed vaccinated (n = 2379) of BNT162b2 mRNA COVID­19 vaccine, were followed-up weekly. Their saliva samples were used to monitor the shedding of SARS-CoV-2. All positive samples were sequenced and genotyped to identify the circulating wild-type virus or variants of concern. Results: The protection fraction against the excretion of the SARS-CoV-2 in the saliva samples of the workers after the second dose is estimated at 0.90 (95% CI: 0.18; 0.99) at 1 week and 0.83 (95% CI: 0.54; 0.95) at 8 weeks. We observe more circulating SARS-CoV-2 and a greater variability of viral loads in the unvaccinated group compared to those of the vaccinated group. Conclusions: This field cohort study advances our knowledge of the efficacy of the mRNA BNT162b2 COVID-19 vaccine on the viral shedding in the saliva specimens of vaccinated NH workers, contributing to better decision-making in public health interventions and management.

Clinical prediction models for diagnosis of COVID-19 among adult patients: a validation and agreement study.

BACKGROUND: Since the beginning of the pandemic, hospitals have been constantly overcrowded, with several observed waves of infected cases and hospitalisations. To avoid as much as possible this situation, efficient tools to facilitate the diagnosis of COVID-19 are needed. OBJECTIVE: To evaluate and compare prediction models to diagnose COVID-19 identified in a systematic review published recently using performance indicators such as discrimination and calibration measures. METHODS: A total of 1618 adult patients present at two Emergency Department triage centers and for whom qRT-PCR tests had been performed were included in this study. Six previously published models were reconstructed and assessed using diagnostic tests as sensitivity (Se) and negative predictive value (NPV), discrimination (Area Under the Roc Curve (AUROC)) and calibration measures. Agreement was also measured between them using Kappa's coefficient and IntraClass Correlation Coefficient (ICC). A sensitivity analysis has been conducted by waves of patients. RESULTS: Among the 6 selected models, those based only on symptoms and/or risk exposure were found to be less efficient than those based on biological parameters and/or radiological examination with smallest AUROC values (< 0.80). However, all models showed good calibration and values above > 0.75 for Se and NPV but poor agreement (Kappa and ICC < 0.5) between them. The results of the first wave were similar to those of the second wave. CONCLUSION: Although quite acceptable and similar results were found between all models, the importance of radiological examination was also emphasized, making it difficult to find an appropriate triage system to classify patients at risk for COVID-19.

Distinction of Lymphoma from Sarcoidosis on 18F-FDG PET/CT: Evaluation of Radiomics-Feature-Guided Machine Learning Versus Human Reader Performance.

Sarcoidosis and lymphoma often share common features on 18F-FDG PET/CT, such as intense hypermetabolic lesions in lymph nodes and multiple organs. We aimed at developing and validating radiomics signatures to differentiate sarcoidosis from Hodgkin lymphoma (HL) and diffuse large B-cell lymphoma (DLBCL). Methods: We retrospectively collected 420 patients (169 sarcoidosis, 140 HL, and 111 DLBCL) who underwent pretreatment 18F-FDG PET/CT at the University Hospital of Liege. The studies were randomly distributed to 4 physicians, who gave their diagnostic suggestion among the 3 diseases. The individual and pooled performance of the physicians was then calculated. Interobserver variability was evaluated using a sample of 34 studies interpreted by all physicians. Volumes of interest were delineated over the lesions and the liver using MIM software, and 215 radiomics features were extracted using the RadiomiX Toolbox. Models were developed combining clinical data (age, sex, and weight) and radiomics (original and tumor-to-liver TLR radiomics), with 7 different feature selection approaches and 4 different machine-learning (ML) classifiers, to differentiate sarcoidosis and lymphomas on both lesion-based and patient-based approaches. Results: For identifying lymphoma versus sarcoidosis, physicians' pooled sensitivity, specificity, area under the receiver-operating-characteristic curve (AUC), and accuracy were 0.99 (95% CI, 0.97-1.00), 0.75 (95% CI, 0.68-0.81), 0.87 (95% CI, 0.84-0.90), and 89.3%, respectively, whereas for identifying HL in the tumor population, it was 0.58 (95% CI, 0.49-0.66), 0.82 (95% CI, 0.74-0.89), 0.70 (95% CI, 0.64-0.75) and 68.5%, respectively. Moderate agreement was found among observers for the diagnosis of lymphoma versus sarcoidosis and HL versus DLBCL, with Fleiss κ-values of 0.66 (95% CI, 0.45-0.87) and 0.69 (95% CI, 0.45-0.93), respectively. The best ML models for identifying lymphoma versus sarcoidosis showed an AUC of 0.94 (95% CI, 0.93-0.95) and 0.85 (95% CI, 0.82-0.88) in lesion- and patient-based approaches, respectively, using TLR radiomics (plus age for the second). To differentiate HL from DLBCL, we obtained an AUC of 0.95 (95% CI, 0.93-0.96) in the lesion-based approach using TLR radiomics and 0.86 (95% CI, 0.80-0.91) in the patient-based approach using original radiomics and age. Conclusion: Characterization of sarcoidosis and lymphoma lesions is feasible using ML and radiomics, with very good to excellent performance, equivalent to or better than that of physicians, who showed significant interobserver variability in their assessment.

University population-based prospective cohort study of SARS-CoV-2 infection and immunity (SARSSURV-ULiège): a study protocol.

INTRODUCTION: For a safe and sustainable return to normal functioning of academic activities in higher education, objective-driven testing strategies that are flexible and rapidly adaptable are essential to effectively monitor and respond to new developments of the COVID-19 pandemic. To date, prospective longitudinal research on SARS-CoV-2 antibody testing in saliva and seroprevalence in higher education contexts is substantially lacking, limiting our understanding of COVID-19 prevalence, incidence and nature of the immune response to SARS-CoV-2 at various stages of the infection and vaccination. To address this lack of evidence, a prospective population-based cohort study (SARSSURV-ULiège) has recently been started. METHODS AND ANALYSIS: Students (n=1396) and staff members (n=1143) of the University of Liège are followed up over more than 1 year. All participants are required to complete anamnestic, clinical and vaccine hesitancy questionnaires for medical histories and undertaken treatments. Previous proven or suspected SARS-CoV-2 infection is also registered. In phase 1, weekly saliva samples to perform RT-qPCR to detect SARS-CoV-2 and monthly COVID-19 serological rapid test results are collected. Once being positive to either saliva RT-qPCR assay for SARS-CoV-2 presence or to serological test, the participant is invited to enter phase 2. If participants get vaccinated during the study period, they are invited to phase 2. In this second phase, besides weekly saliva self-test, depending on the participants' profiles, both gargle and blood samples are collected to obtain various biological data to measure the presence of neutralising antibodies against SARS-CoV-2, determine the magnitude and the duration of antibody responses over time. ETHICS AND DISSEMINATION: The study has received the approval from the University Hospital of Liège Ethics Committee (reference number 2021/96, dated 26 March 2021). Potential protocol amendments will be presented to the Research Ethics Committee. The findings of the present study will be presented at scientific conferences and the results published in peer-review publications. Weekly reports will be submitted to the risk assessment group and the risk management group against COVID-19 of the university to enable a timely public health action if necessary.

Use of Vandenberg and Kuse Mental Rotation Test to Predict Practical Performance of Sinus Endoscopy.

OBJECTIVES: The aim of this study was to assess the predictive value of the Vandenberg and Kuse Mental Rotation Test (MRT) on performance of novice medical students for manipulation of a nasal endoscope on a cadaveric model. MATERIAL AND METHOD: We randomly selected 39 medical students who had never handled a nasal endoscope and subjected them to the MRT. General information including experience in manual, technical, or surgical activities and testing of anatomical knowledge were collected to exclude possible confounding factors. They were then asked to perform series of cadaveric model exercises using a nasal endoscope. Their cadaver performance was evaluated by 2 blinded observers, using a standardized scale. RESULTS: We found that medical students with higher mental rotation skills had significantly increased endoscopic sinus performance (P = .0002 using multivariate regression adjusted for specialty choice, previous surgical exposure, and anatomy knowledge). Higher anatomy knowledge was also associated with better endoscopic sinus performance (P = .0141). Other parameters had no impact on endoscopic sinus performance measured by the endoscopic scale (P > .05). CONCLUSION: The score obtained on the MRT was correlated with the practical performance of manipulating the nasal endoscope in cadaver. It could therefore be a useful spatial ability tool for directing targeted training in rhinology.

Repetitive saliva-based mass screening as a tool for controlling SARS-CoV-2 transmission in nursing homes.

Nursing home (NH) residents and staff have been severely affected by the COVID-19 pandemic. The aim of this study was to examine the use of weekly saliva RT-qPCR testing for SARS-CoV-2 detection among NH workers as a strategy to control disease transmission within NHs in Belgium. From 16 November to 27 December 2020, a voluntary and anonymous weekly screening was implemented in a cohort of 50,000 workers across 572 NHs in the Walloon region of Belgium to detect asymptomatic cases of SARS-CoV-2 via saliva RT-qPCR testing and using the Diagenode saliva sample collection device. Positive workers were isolated to avoid subsequent infections in residents and other staff. RT-qPCR testing was based on pooled saliva sampling techniques from three workers, followed by individual testing of each positive or inconclusive pool. The majority of NHs (85%) and 55% of their workers participated. Pooling did not affect sensitivity as it only induced a very decrease in sensitivity estimated as 0.33%. Significant decreases in the prevalence (34.4-13.4%) and incidence of NHs with either single (13.8-2%) or multiple positive workers (3.7-0%) were observed over time. In addition, deaths among NH residents and NH worker absences decreased significantly over time. Weekly saliva RT-qPCR testing for SARS-CoV-2 demonstrated large-scale feasibility and efficacy in disrupting the chain of transmission. Implementation of this testing strategy in NHs could also be extended to other settings with the aim to control viral transmission for maintaining essential activities.

Validation of a French-language version of TeamSTEPPS® T-TPQ and T-TAQ questionnaires.

Teamwork training and evaluation are essential to enhance safety and quality of care. The lack of the psychometric testing of the TeamSTEPPS® Teamwork Attitudes Questionnaire (T-TAQ) and Teamwork Perceptions Questionnaire (T-TPQ) across different language and cultural settings has questioned their widespread use because such attitudes and perceptions are highly subjective and context-bound. The present study aims to translate the T-TAQ and T-TPQ into the French language and validate the psychometric properties of the two questionnaires in a public health context. A forward-backward translation process, panel reviewing, and pilot testing in two rounds were followed to develop the French versions. Confirmatory factor analysis (CFA) and Cronbach's alpha were used to examine the factor structure and internal consistency, whereas two-way mixed Intraclass Correlation Coefficient (ICC) was performed to assess test-retest reliability. A total of 235 healthcare professionals in the French-speaking community of Belgium completed the T-TAQ and T-TPQ. After two to four weeks, 102 participants took part in the second round. Despite good fit indices as revealed by the CFA and Cronbach's alpha from 0.53 to 0.75 for the five dimensions of the T-TAQ and 0.76 to 0.79 for the T-TPQ, the squared correlations among the constructs were higher than the average variance extracted. Two-way mixed ICCs indicated fair to good test-retest reliability for all the five constructs of the two questionnaires, except the leadership scale of the T-TAQ. The French-language versions of the T-TAQ and T-TPQ were semantically equivalent and culturally relevant with adequate test-retest reliability as compared to the English versions. These two instruments might be used to capture the overall attitude toward teamwork and perceptions of team skills and behaviors. Yet, further research is advisable to refine the scales to establish the discriminant validity of the different dimensions and discriminative power of the instruments.

Evaluation of Two Rapid Antigenic Tests for the Detection of SARS-CoV-2 in Nasopharyngeal Swabs.

(1) Background: In the current context of the COVID-19 crisis, there is a need for fast, easy-to-use, and sensitive diagnostic tools in addition to molecular methods. We have therefore decided to evaluate the performance of newly available antigen detection kits in "real-life" laboratory conditions. (2) Methods: The sensitivity and specificity of two rapid diagnostic tests (RDT)-the COVID-19 Ag Respi-Strip from Coris Bioconcept, Belgium (CoRDT), and the coronavirus antigen rapid test cassette from Healgen Scientific, LLC, USA (HeRDT)-were evaluated on 193 nasopharyngeal samples using RT-PCR as the gold standard. (3) Results: The sensitivity obtained for HeRDT was 88% for all collected samples and 91.1% for samples with Ct ≤ 31. For the CoRDT test, the sensitivity obtained was 62% for all collected samples and 68.9% for samples with Ct ≤ 31. (4) Conclusions: Despite the excellent specificity obtained for both kits, the poor sensitivity of the CoRDT did not allow for its use in the rapid diagnosis of COVID-19. HeRDT satisfied the World Health Organization's performance criteria for rapid antigen detection tests. Its high sensitivity, quick response, and ease of use allowed for the implementation of HeRDT at the laboratory of the University Hospital of Liège.

A systematic review of prediction models to diagnose COVID-19 in adults admitted to healthcare centers.

BACKGROUND: The COVID-19 pandemic is putting significant pressure on the hospital system. To help clinicians in the rapid triage of patients at high risk of COVID-19 while waiting for RT-PCR results, different diagnostic prediction models have been developed. Our objective is to identify, compare, and evaluate performances of prediction models for the diagnosis of COVID-19 in adult patients in a health care setting. METHODS: A search for relevant references has been conducted on the MEDLINE and Scopus databases. Rigorous eligibility criteria have been established (e.g., adult participants, suspicion of COVID-19, medical setting) and applied by two independent investigators to identify suitable studies at 2 different stages: (1) titles and abstracts screening and (2) full-texts screening. Risk of bias (RoB) has been assessed using the Prediction model study Risk of Bias Assessment Tool (PROBAST). Data synthesis has been presented according to a narrative report of findings. RESULTS: Out of the 2334 references identified by the literature search, 13 articles have been included in our systematic review. The studies, carried out all over the world, were performed in 2020. The included articles proposed a model developed using different methods, namely, logistic regression, score, machine learning, XGBoost. All the included models performed well to discriminate adults at high risks of presenting COVID-19 (all area under the ROC curve (AUROC) > 0.500). The best AUROC was observed for the model of Kurstjens et al (AUROC = 0.940 (0.910-0.960), which was also the model that achieved the highest sensitivity (98%). RoB was evaluated as low in general. CONCLUSION: Thirteen models have been developed since the start of the pandemic in order to diagnose COVID-19 in suspected patients from health care centers. All these models are effective, to varying degrees, in identifying whether patients were at high risk of having COVID-19.

Development and validation of a predictive model to determine the level of care in patients confirmed with COVID-19.

BACKGROUND: The COVID-19 pandemic has imposed significant challenges on hospital capacity. While mitigating unnecessary crowding in hospitals is favourable to reduce viral transmission, it is more important to prevent readmissions with impaired clinical status due to initially inappropriate level of care. A validated predictive tool to assist clinical decisions for patient triage and facilitate remote stratification is of critical importance. METHODS: We conducted a retrospective study in patients with confirmed COVID-19 stratified into two levels of care, namely ambulatory care and hospitalization. Data on socio-demographics, clinical symptoms, and comorbidities were collected during the first (N = 571) and second waves (N = 174) of the pandemic in Belgium (2 March to 6 December 2020). Univariate and multivariate logistic regressions were performed to build and validate the prediction model. RESULTS: Significant predictors of hospitalization were old age (OR = 1.08, 95%CI:1.06-1.10), male gender (OR = 4.41, 95%CI: 2.58-7.52), dyspnoea (OR 6.11, 95%CI: 3.58-10.45), dry cough (OR 2.89, 95%CI: 1.54-5.41), wet cough (OR 4.62, 95%CI: 1.93-11.06), hypertension (OR 2.20, 95%CI: 1.17-4.16) and renal failure (OR 5.39, 95%CI: 1.00-29.00). Rhinorrhea (OR 0.43, 95%CI: 0.24-0.79) and headache (OR 0.36, 95%CI: 0.20-0.65) were negatively associated with hospitalization. A receiver operating characteristic (ROC) curve was constructed and the area under the ROC curve was 0.931 (95% CI: 0.910-0.953) for the prediction model (first wave) and 0.895 (95% CI: 0.833-0.957) for the validated dataset (second wave). CONCLUSION: With a good discriminating power, the prediction model might identify patients who require ambulatory care or hospitalization and support clinical decisions by Emergency Department staff and general practitioners.