Search

Abstract Nonischaemic myocardial fibrosis is associated with cardiac dysfunction, malignant arrhythmias and sudden cardiac death. In the absence of a specific aetiology, its finding as late gadolinium enhancement (LGE) on cardiac magnetic resonance imaging is often attributed to preceding viral myocarditis. Athletes presenting with ventricular arrhythmias often have nonischaemic LGE. Previous studies have demonstrated an adverse effect of exercise on the course of acute viral myocarditis. In this study, we have investigated, for the first time, the impact of endurance training on longer-term outcomes such as myocardial fibrosis and arrhythmogenicity in a murine coxsackievirus B3 (CVB)-induced myocarditis model. Male C57BL/6J mice (n = 72) were randomly assigned to 8 weeks of forced treadmill running (EEX) or no exercise (SED). Myocarditis was induced 2 weeks later by a single intraperitoneal injection with CVB, versus vehicle in the controls (PBS). In a separate study, mice (n = 30) were subjected to pretraining for 13 weeks (preEEX), without continuation of exercise during myocarditis. Overall, continuation of exercise resulted in a milder clinical course of viral disease, with less weight loss and better preserved running capacity. CVB-EEX and preEEX-CVB mice tended to have a lower mortality rate. At sacrifice (i.e. 6 weeks after inoculation), the majority of virus was cleared from the heart. Histological assessment demonstrated prominent myocardial inflammatory infiltration and cardiomyocyte loss in both CVB groups. Inflammatory lesions in the CVB-EEX group contained higher numbers of pro-inflammatory cells (iNOS-reactive macrophages and CD8+ T lymphocytes) compared to these in CVB-SED. Treadmill running during myocarditis increased interstitial fibrosis [82.4% (CVB-EEX) vs. 56.3% (CVB-SED); P = 0.049]. Additionally, perivascular and/or interstitial fibrosis with extensive distribution was more likely to occur with exercise [64.7% and 64.7% (CVB-EEX) vs. 50% and 31.3% (CVB-SED); P = 0.048]. There was a numerical, but not significant, increase in the number of scars per cross-section (1.9 vs. 1.2; P = 0.195), with similar scar distribution and histological appearance in CVB-EEX and CVB-SED. In vivo electrophysiology studies did not induce sustained monomorphic ventricular tachycardia, only nonsustained (usually polymorphic) runs. Their cumulative beat count and duration paralleled the increased fibrosis between CVB-EEX and CVB-SED, but the difference was not significant (P = 0.084 for each). Interestingly, in mice that were subjected to pretraining only without continuation of exercise during myocarditis, no differences between pretrained and sedentary mice were observed at sacrifice (i.e. 6 weeks after inoculation and training cessation) with regard to myocardial inflammation, fibrosis, and ventricular arrhythmogenicity. In conclusion, endurance exercise during viral myocarditis modulates the inflammatory process with more pro-inflammatory cells and enhances perivascular and interstitial fibrosis development. The impact on ventricular arrhythmogenesis requires further exploration.

Arterial hypertension is a major cause of mortality and morbidity worldwide. Medical therapy is the most common treatment. However, in some cases there is a persistent high blood pressure despite medical therapy. These patients with medication refractory arterial hypertension can be treated by renal denervation. Until now an endovascular approach has been used. There are however limitations in eligibility based on vascular or anatomical anomalies. For these patients, as well as other patients eligible for renal denervation, robot-assisted renal denervation has the potential to become a surgical treatment option based on our findings.

Background: The DAWN antivirals trial was a multicentric, randomised placebo-controlled trial evaluating antiviral medication for COVID-19 in general practice. The trial was prematurely terminated because of insufficient recruitment. Aim: To explore which factors contributed to the premature termination. Design & setting: General practice in Belgium. Method: Patients were randomised to camostat or placebo (patients and physicians blinded) between June 2021 and July 2022; a third arm evaluating molnupiravir (open label) was opened in March 2022. The outcome assessor was blinded for all comparisons except for the patient reported outcomes in case of molnupiravir. The authors analysed available trial data and evaluated trial context, implementation, and mechanisms of impact based on semi-structured interviews with trial stakeholders. Results: The trial recruited 44 participants; 19 were allocated to camostat (median age 55 years), 8 to molnupiravir (median age 60 years), and 17 to placebo (median age 56 years). There were no serious adverse events in either group. Most difficulties were related to the pandemic context: disruption to routine clinical services; multiple changes to the service model for COVID-19 patients; overwhelmed clinical staff; delays of trial medication; and staff shortages in the sponsor and clinical team. In addition, regulatory approval processes were lengthy and led to additional study procedures. It was felt that the trial started too late, when vaccinations had already begun. Conclusion: The DAWN antivirals trial was stopped prematurely. Although many barriers were related to the pandemic itself, hurdles such as a small and inexperienced sponsor and clinical teams, delays in regulatory processes, and research capacity in routine settings could be overcome by established research infrastructure and standardisation of processes.

Summary: Protein-protein interactions play an important biological role in every aspect of cellular homeostasis and functioning. Proximity labeling mass spectrometry-based proteomics overcomes challenges typically associated with other methods and has quickly become the current state of the art in the field. Nevertheless, tight control of proximity-labeling enzymatic activity and expression levels is crucial to accurately identify protein interactors. Here, we leverage a T2A self-cleaving peptide and a non-cleaving mutant to accommodate the protein of interest in the experimental and control TurboID setup. To allow easy and streamlined plasmid assembly, we built a Golden Gate modular cloning system to generate plasmids for transient expression and stable integration. To highlight our T2A Split/link design, we applied it to identify protein interactions of the glucocorticoid receptor and severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) nucleocapsid and non-structural protein 7 (NSP7) proteins by TurboID proximity labeling. Our results demonstrate that our T2A split/link provides an opportune control that builds upon previously established control requirements in the field. Motivation: In proximity labeling proteomics, protein-protein interactions are identified by in vivo biotinylation. However, the current lack of a universally applicable negative control for differential analysis affects accurate mapping of the interactome. To bridge this gap, we conceptualized a system based on the T2A self-cleaving peptide to match expression levels between control and bait protein setups while using the same bait protein. In addition, we implemented a versatile modular cloning system to build mammalian expression vectors for, but not limited to, proximity labeling assays.

Introduction Although sexual health has been holistically defined to include sexual satisfaction, it has been largely absent in health services and sexual and reproductive health and rights programmes in many parts of the world. We propose sexual satisfaction as a useful indicator, as one of the proxy measures for sexual health and well-being and as a component of well-being in general.Methods The Sialon II project is a multicentre biological and behavioural cross-sectional community-based survey implemented across 13 European cities during 2013–2014 among men who have sex with men. Sexual satisfaction was explored using one single item: ‘How satisfied are you with your sex life?’ A multivariable multilevel logistic random-intercept model was estimated to identify factors associated with reporting positive sexual satisfaction versus negative sexual satisfaction.Results Age, the number of partners and self-reported HIV status were not significantly associated with sexual satisfaction in the multivariate model. Participants reporting an insertive role or reported both an insertive and receptive role during the last anal intercourse were more likely to be sexually satisfied, compared with a receptive role. Participants reporting anal intercourse with a condom were more likely to be satisfied than those declaring no anal intercourse in the last 6 months, but no significant association was found compared with anal intercourse without condom. Knowledge of HIV-serostatus concordance with the last sexual partner was positively correlated with sexual satisfaction. Having had sexual intercourse with non-steady partners only in the last 6 months was negatively correlated. The more positive participants perceived their work/school, parents and friends/acquaintances’ attitudes towards gay or bisexual persons, the higher the odds they were satisfied with their sexual life.Conclusion Using a single item on sexual satisfaction in a bio-behavioural study, our analysis has shown that it is associated with individual, interpersonal and social/structural factors and has proven its usefulness as a sexual health indicator among men who have sex with men.

Abstract Background A previous individual participant data meta-analysis (IPD-MA) of antibiotics for adults with clinically diagnosed acute rhinosinusitis (ARS) showed a marginal overall effect of antibiotics, but was unable to identify patients that are most likely to benefit from antibiotics when applying conventional (i.e. univariable or one-variable-at-a-time) subgroup analysis. We updated the systematic review and investigated whether multivariable prediction of patient-level prognosis and antibiotic treatment effect may lead to more tailored treatment assignment in adults presenting to primary care with ARS. Methods An IPD-MA of nine double-blind placebo-controlled trials of antibiotic treatment (n=2539) was conducted, with the probability of being cured at 8–15 days as the primary outcome. A logistic mixed effects model was developed to predict the probability of being cured based on demographic characteristics, signs and symptoms, and antibiotic treatment assignment. Predictive performance was quantified based on internal-external cross-validation in terms of calibration and discrimination performance, overall model fit, and the accuracy of individual predictions. Results Results indicate that the prognosis with respect to risk of cure could not be reliably predicted (c-statistic 0.58 and Brier score 0.24). Similarly, patient-level treatment effect predictions did not reliably distinguish between those that did and did not benefit from antibiotics (c-for-benefit 0.50). Conclusions In conclusion, multivariable prediction based on patient demographics and common signs and symptoms did not reliably predict the patient-level probability of cure and antibiotic effect in this IPD-MA. Therefore, these characteristics cannot be expected to reliably distinguish those that do and do not benefit from antibiotics in adults presenting to primary care with ARS.

When COVID-19 vaccines were implemented, nursing home residents (NHRs) and staff (NHS) in Belgium were prioritized for vaccination. To characterize the vaccine response over time in this population and to identify poorly responding groups, we assessed antibody concentrations two (T1), four (T2) and six months (T3) after primary course BNT162b2 vaccination in six groups of infection-naive/infection-primed NHRs/NHS, with/without comorbidity (NHRs only). Participant groups (N = 125 per group) were defined within a national serosurveillance study in nursing homes, based on questionnaire data. Dried blood spots were analyzed using ELISA for the quantification of SARS-CoV-2 S1RBD IgG antibodies. Among all groups, antibody concentrations significantly decreased between T1 and T2/T3, all with a ≥70% decrease at T3, except for infection-primed staff (−32%). Antibody concentrations among infection-naive NHRs were 11.96 times lower than those among infection-primed NHR, while the latter were comparable (x1.05) to infection-primed NHS. The largest proportion [13% (95% CI: 11–24%)] of vaccine non-responders was observed in the group of infection-naive NHRs with comorbidities. A longer interval between infection and vaccination (≥3 months) elicited higher antibody responses. Our data retrospectively show the necessity of timely COVID-19 booster vaccination. Infection-naive NHRs require special attention regarding immune monitoring in future epidemics or pandemics.

Background: There is an unmet medical need for the early detection of immune checkpoint inhibitor (ICI)-induced cardiovascular (CV) adverse events due to a lack of adequate biomarkers. This study aimed to provide insights on the incidence of troponin elevations and echocardiographic dynamics during ICI treatment in cancer patients and their role as potential biomarkers for submyocardial damage. In addition, it is the first study to compare hs-TnT and hs-TnI in ICI-treated patients and to evaluate their interchangeability in the context of screening. Results: Among 59 patients, the mean patient age was 68 years, and 76% were men. Overall, 25% of patients received combination therapy. Although 10.6% [95% CI: 5.0–22.5] of the patients developed troponin elevations, none experienced a CV event. No significant changes were found in 3D left ventricular (LV) ejection fraction nor in global longitudinal strain f (56 ± 6% vs. 56 ± 6%, p = 0.903 and −17.8% [−18.5; −14.2] vs. −17.0% [−18.8; −15.1], p = 0.663) at 3 months. There were also no significant changes in diastolic function and right ventricular function. In addition, there was poor agreement between hs-TnT and hs-TnI. Methods: Here, we present a preliminary analysis of the first 59 patients included in our ongoing prospective clinical trial (NCT05699915) during the first three months of treatment. All patients underwent electrocardiography and echocardiography along with blood sampling at standardized time intervals. This study aimed to investigate the incidence of elevated hs-TnT levels within the first three months of ICI treatment. Elevations were defined as hs-TnT above the upper limit of normal (ULN) if the baseline value was normal, or 1.5 ≥ times baseline if the baseline value was above the ULN. Conclusions: Hs-TnT elevations occurred in 10.6% of the patients. However, no significant changes were found on 3D echocardiography, nor did any of the patients develop a CV event. There were also no changes found in NT-proBNP. The study is still ongoing, but these preliminary findings do not show a promising role for cardiac troponins nor for echocardiographic dynamics in the prediction of CV events during the early stages of ICI treatment.

Abstract Background Primary care is a critical partner for antimicrobial stewardship efforts given its high human antibiotic usage. Peer comparison audit and feedback (A&F) is often used to reduce inappropriate antibiotic prescribing. The design and implementation of A&F may impact its effectiveness. There are no best practice guidelines for peer comparison A&F in antibiotic prescribing in primary care. Objective To develop best practice guidelines for peer comparison A&F for antibiotic prescribing in primary care in high income countries by leveraging international expertise via the Joint Programming Initiative on Antimicrobial Resistance—Primary Care Antibiotic Audit and Feedback Network. Methods We used a modified Delphi process to achieve convergence of expert opinions on best practice statements for peer comparison A&F based on existing evidence and theory. Three rounds were performed, each with online surveys and virtual meetings to enable discussion and rating of each best practice statement. A five-point Likert scale was used to rate consensus with a median threshold score of 4 to indicate a consensus statement. Results The final set of guidelines include 13 best practice statements in four categories: general considerations (n = 3), selecting feedback recipients (n = 1), data and indicator selection (n = 4), and feedback delivery (n = 5). Conclusion We report an expert-derived best practice recommendations for designing and evaluating peer comparison A&F for antibiotic prescribing in primary care. These 13 statements can be used by A&F designers to optimize the impact of their quality improvement interventions, and improve antibiotic prescribing in primary care.

In Belgium, nursing home (NH) staff (NHS) and residents were prioritised for the initial COVID-19 vaccination and successive booster doses. The vaccination campaign for the first booster started in September 2021 in Belgian NH. Our first study about vaccine hesitancy towards the COVID-19 vaccine in Belgian NHS already showed a degree of fear for the primary vaccination course (T1). This new study aims to evaluate vaccine hesitancy to get the first booster (T2) in a population of fully vaccinated (with two doses) NHS. A random stratified sample of NHS who received the primary vaccination course (N = 954) completed an online questionnaire on COVID-19 booster hesitancy (between 25/11/2021 and 22/01/2022). NHS who hesitated or refused the booster were asked for the main reason for their hesitation/refusal. Overall, 21.0 % of our population hesitated before, were still hesitating or refused the booster, NHS that were not hesitant at T1 being 5.7 times less likely to hesitate to get the first booster dose (Adjusted OR 0.179, 95 % CI: 0.120, 0.267). Although there was a slight reduction (23.5 % to 20.1 %) in the proportion of NHS who hesitated/refused vaccination at T1 compared to T2 (p = 0.034), the fear of unknown effects was the principal reason for hesitation/refusal, already mentioned in our first study. NHS were not reassured concerning their initial fears. Given the likelihood that booster vaccinations will be necessary over the coming years, a communication strategy specific to NHS should be implemented.

Changes in the clinical trials landscape have been driven by advancements in digital technology. The use of electronic informed consent to inform research participants and to obtain their consent electronically has the potential to improve participant–researcher interactions over time, facilitate clinical trial participation, and increase efficiency in clinical trial conduct. A personalized electronic informed consent platform that enables long-term interactions with the research team could function as a tool to empower participant engagement in clinical trials. However, significant challenges persist impeding successful and widespread implementation. This Perspective provides insights into the opportunities and challenges for the implementation of electronic informed consent in clinical trials. It sets out key recommendations to promote the implementation of this innovative approach to the informed consent process, including the creation of uniform electronic informed consent platforms at regional and national level.

BackgroundOver the years, there has been increasing interest in electronic informed consent (eIC) in clinical research. The user-friendliness of an eIC application and its acceptance by stakeholders plays a central role in achieving successful implementation. ObjectiveThis study aims to identify insights for the design and implementation of a user-friendly, personalized, and long-term eIC application based on a usability study with (potential) research participants and semistructured interviews with stakeholders on the practical integration of such an application into their daily practice. MethodsAn eIC prototype was evaluated and refined through usability testing among Belgian citizens and iterative redesign. On the basis of a digital literacy questionnaire, a heterogeneous sample of participants was established. Participants needed to complete a series of usability tasks related to personalization and long-term interaction with the research team while using the “think aloud” technique. In addition, usability tests involved completing the System Usability Scale questionnaire and taking part in a semistructured feedback interview. Furthermore, semistructured interviews were conducted with ethics committee members, health care professionals, and pharmaceutical industry representatives active in Belgium and involved in clinical research. Thematic analysis was undertaken using the NVivo software (Lumivero). ResultsIn total, 3 iterations of usability tests were conducted with 10 participants each. Each cycle involved some participants who reported having low digital skills. The System Usability Scale scores related to the tasks on personalization and long-term interaction increased after each iteration and reached 69.5 (SD 8.35) and 71.3 (SD 16.1) out of 100, respectively, which represents above-average usability. Semistructured interviews conducted with health care professionals (n=4), ethics committee members (n=8), and pharmaceutical industry representatives (n=5) identified the need for an eIC system that can be easily set up. For example, a library could be established enabling stakeholders to easily provide background information about a clinical study, presented in the second layer of the interface. In contrast, some functionalities, such as informing participants about new studies through an eIC system, were not considered useful by stakeholders. ConclusionsThis study provides insights for the implementation of a user-friendly personalized and long-term eIC application. The study findings showed that usability testing is key to assessing and increasing the user-friendliness of an eIC application. Although this eIC system has the potential to be usable by a wide audience, participants with low digital literacy may not be able to use it successfully, highlighting the need for additional support for participants or other alternatives to an eIC system. In addition, key lessons emerging from the interviews included ensuring that the application is easy to implement in practice and is interoperable with other established systems.

The presence of Pseudomonas aeruginosa biofilms in cystic fibrosis (CF) patients suffering from chronic lung infections contributes to the failure of antimicrobial therapy. Conventionally, the minimal inhibitory concentration (MIC) is determined to assess the antimicrobial susceptibility of a pathogen, however this parameter fails to predict success in treating biofilm-associated infections. In the present study we developed a high throughput method to determine the antimicrobial concentration required to prevent P. aeruginosa biofilm formation, using a synthetic cystic fibrosis sputum medium (SCFM2).Biofilms were grown in SCFM2 for 24 h in the presence of antibiotics (tobramycin, ciprofloxacin or colistin), whereafter biofilms were disrupted and a resazurin staining was used to quantify the number of surviving metabolically active cells. In parallel, the content of all wells was plated to determine the number of colony forming units (CFU). Biofilm preventing concentrations (BPCs) were compared to MICs and minimal bactericidal concentrations (MBCs) determined according to EUCAST guidelines. Correlations between the resazurin-derived fluorescence and CFU counts were assessed with Kendall’s Tau Rank tests.A significant correlation between fluorescence and CFU counts was observed for 9 out of 10 strains investigated, suggesting the fluorometric assay is a reliable alternative to plating for most P. aeruginosa isolates to determine biofilm susceptibility in relevant conditions. For all isolates a clear difference between MICs and BPCs of all three antibiotics was observed, with the BPCs being consistently higher than the MICs. Additionally, the extent of this difference appeared to be antibiotic-dependent.Our findings suggest that this high throughput assay could be a valuable addition to evaluate the antimicrobial susceptibility in P. aeruginosa biofilms in the context of CF.

AbstractBackground Nursing home residents (NHR) and staff have been disproportionally affected by the COVID-19 pandemic and were therefore prioritised in the COVID-19 vaccination strategy. However, frail older adults, like NHR, are known to have decreased antibody responses upon vaccination targeting other viral antigens.Objectives As real-world data on vaccine responsiveness, we assessed the prevalence of SARS-CoV-2 antibodies among Belgian NHR and staff during the primary COVID-19 vaccination campaign.Methods In total, we tested 1629 NHR and 1356 staff across 69 Belgian NHs for the presence of SARS-CoV-2 IgM/IgG antibodies using rapid tests. We collected socio-demographic and COVID-19-related medical data through questionnaires. Sampling occurred between 1 February and 24 March 2021, in a randomly sampled population that received none, one or two BNT162b2 vaccine doses.Results We found that during the primary vaccination campaign with 59% of the study population fully vaccinated, 74% had SARS-CoV-2 antibodies. Among fully vaccinated individuals only, fewer residents tested positive for SARS-CoV-2 antibodies (77%) than staff (98%), suggesting an impaired vaccine-induced antibody response in the elderly, with lowest seroprevalences observed among infection naïve residents. COVID-19 vaccination status and previous SARS-CoV-2 infection were predictors for SARS-CoV-2 seropositivity. Alternatively, age ≥ 80 years old, the presence of comorbidities and high care dependency predicted SARS-CoV-2 seronegativity in NHR.Conclusion These findings highlight the need for further monitoring of SARS-CoV-2 immunity upon vaccination in the elderly population, as their impaired humoral responses could imply insufficient protection against COVID-19.Trial registration This study was retrospectively registered on ClinicalTrials.gov (NCT04738695).

Abstract Background Self-management support is seen as a cornerstone of good diabetes care and many countries are currently engaged in initiatives to integrate self-management support in primary care. Concerning the organisation of these programs, evidence is growing that engagement of health care professionals, in particular of GPs, is critical for successful application. This paper reports on a study exploring why a substantial number of GPs was (initially) reluctant to refer patients to a self-management education program in Belgium. Methods Qualitative analysis of semi-structured face-to-face interviews with a purposive sample of 20 GPs who were not regular users of the service. The Greenhalgh diffusion of innovation framework was used as background and organising framework. Results Several barriers, linked to different components of the Greenhalgh model, emerged from the interview data. One of the most striking ones was the limited readiness for innovation among GPs. Feelings of fear of further fragmentation of diabetes care and frustration and insecurity regarding their own role in diabetes care prevented them from engaging in the innovation process. GPs needed time to be reassured that the program respects their role and has an added value to usual care. Once GPs considered referring patients, it was not clear enough which of their patients would benefit from the program. Some GPs expressed the need for training in motivational skills, so that they could better motivate their patients to participate. A practical but often mentioned barrier was the distance to the centre where the program was delivered. Further, uncertainty about continuity interfered with the uptake of the offer. Conclusions The study results contribute to a better understanding of the reasons why GPs hesitate to refer patients to a self-management education program. First of all, the role of GPs and other health care providers in diabetes care needs to be clarified before introducing new functions. Feelings of security and a basic trust of providers in the health system are a prerequisite for participation in care innovation. Moreover, some important lessons regarding the implementation of an education program in primary care have been learned from the study.

Abstract Background During a four-year action research project (2003-2007), a program targeting all type 2 diabetes patients was implemented in a well-defined geographical region in Belgium. The implementation of the program resulted in an increase of the overall Assessment of Chronic Illness Care (ACIC) score from 1.45 in 2003 to 5.5 in 2007. The aim of the follow-up study in 2008 was to assess the effect of the implementation of Chronic Care Model (CCM) elements on the quality of diabetes care in a country where the efforts to adapt primary care to a more chronic care oriented system are still at a starting point. Methods A quasi-experimental study design involving a control region with comparable geographical and socio-economic characteristics and health care facilities was used to evaluate the effect of the intervention in the region. In collaboration with the InterMutualistic Agency (IMA) and the laboratories from both regions a research database was set up. Study cohorts in both regions were defined by using administrative data from the Sickness Funds and selected from the research database. A set of nine quality indicators was defined based on current scientific evidence. Data were analysed by an institution experienced in longitudinal data analysis. Results In total 4,174 type 2 diabetes patients were selected from the research database; 2,425 patients (52.9% women) with a mean age of 67.5 from the intervention region and 1,749 patients (55.7% women) with a mean age of 67.4 from the control region. At the end of the intervention period, improvements were observed in five of the nine defined quality indicators in the intervention region, three of which (HbA1c assessment, statin therapy, cholesterol target) improved significantly more than in the control region. Mean HbA1c improved significantly in the intervention region (7.55 to 7.06%), but this evolution did not differ significantly (p = 0.4207) from the one in the control region (7.44 to 6.90%). The improvement in lipid control was significantly higher (p = 0.0021) in the intervention region (total cholesterol 199.07 to 173 mg/dl) than in the control region (199.44 to 180.60 mg/dl). The systematic assessment of long-term diabetes complications remained insufficient. In 2006 only 26% of the patients had their urine tested for micro-albuminuria and only 36% had consulted an ophthalmologist. Conclusion Although the overall ACIC score increased from 1.45 to 5.5, the improvement in the quality of diabetes care was moderate. Further improvements are needed in the CCM components delivery system design and clinical information systems. The regional networks, as they are financed now by the National Institute for Health and Disability Insurance (NIHDI), are an opportunity to explore how this can be achieved in consultation with the GPs. But it is clear that, simultaneously, action is needed on the health system level to realize the installation of an accurate quality monitoring system and the necessary preconditions for chronic care delivery in primary care (patient registration, staff support, IT support). Trial Registration Trial registration number: ClinicalTrials.gov Identifier: NCT00824499

Abstract Background Early distinction between mild and serious infections (SI) is challenging in children in ambulatory care. Clinical prediction models (CPMs), developed to aid physicians in clinical decision-making, require broad external validation before clinical use. We aimed to externally validate four CPMs, developed in emergency departments, in ambulatory care. Methods We applied the CPMs in a prospective cohort of acutely ill children presenting to general practices, outpatient paediatric practices or emergency departments in Flanders, Belgium. For two multinomial regression models, Feverkidstool and Craig model, discriminative ability and calibration were assessed, and a model update was performed by re-estimation of coefficients with correction for overfitting. For two risk scores, the SBI score and PAWS, the diagnostic test accuracy was assessed. Results A total of 8211 children were included, comprising 498 SI and 276 serious bacterial infections (SBI). Feverkidstool had a C-statistic of 0.80 (95% confidence interval 0.77–0.84) with good calibration for pneumonia and 0.74 (0.70–0.79) with poor calibration for other SBI. The Craig model had a C-statistic of 0.80 (0.77–0.83) for pneumonia, 0.75 (0.70–0.80) for complicated urinary tract infections and 0.63 (0.39–0.88) for bacteraemia, with poor calibration. The model update resulted in improved C-statistics for all outcomes and good overall calibration for Feverkidstool and the Craig model. SBI score and PAWS performed extremely weak with sensitivities of 0.12 (0.09–0.15) and 0.32 (0.28–0.37). Conclusions Feverkidstool and the Craig model show good discriminative ability for predicting SBI and a potential for early recognition of SBI, confirming good external validity in a low prevalence setting of SBI. The SBI score and PAWS showed poor diagnostic performance. Trial registration ClinicalTrials.gov, NCT02024282. Registered on 31 December 2013.