Asthma caused around 436,000 deaths globally in 2021. Predicting asthma attacks can save lives, reduce costs, and strengthen healthcare systems. However, risk prediction requires extensive data, which is often unavailable. This study aims to develop an approach that minimizes data requirements, reducing complexity and enabling broader applicability in data-limited settings. We employed the CRISP-DM methodology, including combinations of feature selection strategies, machine learning algorithms, and data imbalance handling techniques. Altogether, 120 models were constructed. Model performance was evaluated based on accuracy and efficiency. The top models were externally validated. The key finding is that the LR and XGB models, when combined with an under-sampling technique, perform better, given that the feature selection is conducted based on the feature importance generated by the XGB-RUS model. Even a few features, including asthma attacks in the past year and SABA_ICS ratio, can achieve a reasonable level of performance. Through rigorous experimentation, we achieved high accuracy in multiple scenarios. We have presented 6 feature sets. Among those, FS1 achieved the best performance, while FS6 yielded a reasonable performance with only two features. This paper presents a comprehensive analysis of the data requirements for predicting the risk of asthma attacks. This will enhance the acceptability of the prediction models using simplified feature sets in clinical settings.

Despite advances in digital health technologies, the integration of clinical decision support systems (CDSSs) into routine primary care for asthma has been extremely limited. Asthma CDSSs employ diverse approaches, from prescribing support to risk prediction, but it remains unclear which are most likely to achieve sustained, real-world impact. Our objectives were to determine the mechanisms of action, clinical practice integration approaches, outputs and outcomes of asthma CDSSs in recent literature. Five electronic databases (Embase via Ovid, PubMed, CENTRAL, the Health Technology Assessment (HTA) Database, and the ISRCTN registry of clinical trials) were searched to identify papers published between 2012 and 2024 describing pilot studies, feasibility studies, or clinical trials, of primary care-based asthma CDSS. Two independent reviewers screened the retrieved literature and extracted the data on study designs, interventions, outcomes and mechanisms of action, and results. Across 18 included trials, interventions demonstrated substantial heterogeneity in mechanisms, integration methods, and targeted clinical behaviours. Although some studies showed improvements in adherence to prescribing best practices and the delivery of personalised action plans, most reported modest or declining system use over time and inconsistent effects on asthma control or severe attack outcomes. Continued progress will depend on integrating behavioural theory, improving workflow compatibility, and generating rigorous evidence to guide the development of CDSSs with genuine potential for sustainable impact.

More than 8 million people in the UK have been diagnosed with asthma, a chronic respiratory condition which leads to the death of more than 25 people per week. Clinical decision support software (CDSS) can be used to improve patient care by improving clinical accuracy or by increasing the efficiency of clinical practice. This study aimed to identify desired functionalities of a software tool for asthma care, from both primary and specialist health care providers. Qualitative data were collected from semi-structured interviews with 19 participants. Eight functionalities were identified for prioritisation for the development of clinical decision support system in asthma, including diagnosis support, medical history retrieval, and adherence to best practice guidelines. Parallels were drawn with successful CDSS applications in other medical fields, such as radiology, highlighting the value of interdisciplinary learning and adaptation. Prioritising user-centred design, and aligning processes with existing workflows, may allow software tools to lighten administrative burdens, promote proactive patient care, and improve patient outcomes. Healthcare software development must focus on creating tools which are intuitive, reliable, and co-designed, in order to increase uptake and sustained impact.

Background: People with asthma are recommended to have regular reviews in primary care, with assessment of symptoms, adjustment of treatment and self-management processes, and the delivery of a written action plan for emergencies. Aim: To investigate the incidence and factors associated with attendance of annual asthma reviews. Design & setting: This observational study used electronic health records for 49 307 patients in Scotland with asthma between 1 January 2000 and 31 March 2017. The analysis population of 13 726 patients had at least five asthma-related encounters between 2008 and 2016. Method: Multivariable logistic regression was employed, using linked primary care prescription data and primary care registration demographic data. Results: There was a median of 381 days between subsequent reviews. Reviews in the index year were strongly associated with reviews in the following year (odds ratio [OR] 1.76, 95% confidence interval [CI] 1.68 to 1.84). In contrast, asthma consultations (excluding reviews) in the index year were associated with lower odds of having a review in the following year (OR 0.48, 95% CI = 0.46 to 0.51). Those aged 18-35 years in the index year or those with missing addresses in the practice registration data were the least likely groups to have an asthma review in the following year. Conclusion: Reviewing the delivery of asthma care identifies patients who may be slipping through the gaps by receiving only reactive asthma care rather than the structured, preventive care that can be delivered through annual reviews. Understanding the risk factors for not receiving an annual review can be leveraged to create more effective review invitations, such as explaining the specific content of reviews, introducing new contact methods to improve health equity, and reviewing the algorithm used to determine who is invited.

Aim We aimed to identify enablers and barriers of using primary care routine data for healthcare research, to formulate recommendations for improving efficiency in knowledge discovery.Background Data recorded routinely in primary care can be used for estimating the impact of interventions provided within routine care for all people who are clinically eligible. Despite official promotion of ‘efficient trial designs’, anecdotally researchers in the Asthma UK Centre for Applied Research (AUKCAR) have encountered multiple barriers to accessing and using routine data.Methods Using studies within the AUKCAR portfolio as exemplars, we captured limitations, barriers, successes, and strengths through correspondence and discussions with the principal investigators and project managers of the case studies.Results We identified 14 studies (8 trials, 2 developmental studies and 4 observational studies). Investigators agreed that using routine primary care data potentially offered a convenient collection of data for effectiveness outcomes, health economic assessment and process evaluation in one data extraction. However, this advantage was overshadowed by time-consuming processes that were major barriers to conducting efficient research. Common themes were multiple layers of information governance approvals in addition to the ethics and local governance approvals required by all health service research; lack of standardisation so that local approvals required diverse paperwork and reached conflicting conclusions as to whether a study should be approved. Practical consequences included a trial that over-recruited by 20% in order to randomise 144 practices with all required permissions, and a 5-year delay in reporting a trial while retrospectively applied regulations were satisfied to allow data linkage.Conclusions Overcoming the substantial barriers of using routine primary care data will require a streamlined governance process, standardised understanding/application of regulations and adequate National Health Service IT (Information Technology) capability. Without policy-driven prioritisation of these changes, the potential of this valuable resource will not be leveraged.

BACKGROUND:Sotrovimab is a neutralising monoclonal antibody (nMAB) currently available to treat extremely clinically vulnerable COVID-19 patients in England. Trials have shown it to have mild to moderate side effects, however, evidence regarding its safety in real-world settings remains insufficient. METHODS:Descriptive and multivariable logistic regression analyses were conducted to evaluate uptake, and a self-controlled case series analysis performed to measure the risk of hospital admission (hospitalisation) associated with 49 pre-specified suspected adverse outcomes in the period 2-28 days post-Sotrovimab treatment among eligible patients treated between December 11, 2021 and May 24, 2022. RESULTS:Here we show that among treated and untreated eligible individuals, the mean ages (54.6 years, SD: 16.1 vs 54.1, SD: 18.3) and sex distribution (women: 60.9% vs 58.1%; men: 38.9% vs 41.1%) are similar. There are marked variations in uptake between ethnic groups, which is higher amongst individuals categorised ethnically as Indian (15.0%; 95%CI 13.8, 16.3), Other Asian (13.7%; 95%CI 11.9, 15.8), white (13.4%; 95%CI 13.3, 13.6), and Bangladeshi (11.4%; 95%CI 8.8, 14.6); and lower amongst Black Caribbean individuals (6.4%; 95%CI 5.4, 7.5) and Black Africans (4.7%; 95%CI 4.1, 5.4). We find no increased risk of any of the suspected adverse outcomes in the period 2-28 days post-treatment. CONCLUSIONS:We find no safety signals of concern for possible adverse outcomes in the period 2-28 days post treatment with Sotrovimab. However, there is evidence of unequal uptake of Sotrovimab treatment across ethnic groups.

Primary care consultations provide an opportunity for patients and clinicians to assess asthma attack risk. Using a data-driven risk prediction tool with routinely collected health records may be an efficient way to aid promotion of effective self-management, and support clinical decision making. Longitudinal Scottish primary care data for 21,250 asthma patients were used to predict the risk of asthma attacks in the following year. A selection of machine learning algorithms (i.e., Naïve Bayes Classifier, Logistic Regression, Random Forests, and Extreme Gradient Boosting), hyperparameters, training data enrichment methods were explored, and validated in a random unseen data partition. Our final Logistic Regression model achieved the best performance when no training data enrichment was applied. Around 1 in 3 (36.2%) predicted high-risk patients had an attack within one year of consultation, compared to approximately 1 in 16 in the predicted low-risk group (6.7%). The model was well calibrated, with a calibration slope of 1.02 and an intercept of 0.004, and the Area under the Curve was 0.75. This model has the potential to increase the efficiency of routine asthma care by creating new personalized care pathways mapped to predicted risk of asthma attacks, such as priority ranking patients for scheduled consultations and interventions. Furthermore, it could be used to educate patients about their individual risk and risk factors, and promote healthier lifestyle changes, use of self-management plans, and early emergency care seeking following rapid symptom deterioration.

Exploring factors that increase the risk of asthma attacks is crucial for timely patient management. Machine learning (ML) techniques are increasingly used for risk prediction. This study aimed to identify risk factors for asthma attacks in New Zealand and evaluate ML algorithms' performance in predicting these risks. National health datasets from 355,113 patients aged 6 years and older with asthma were analyzed from 2008 to 2016. The outcome was the occurrence of an asthma attack within 3 months. Two ML models, XGBoost and Random Forest, and a statistical model, Logistic Regression (LR), were developed and performance compared. Key risk predictors included prior asthma attacks, length of winter exposure, and the number of ICS and SABA inhalers. XGB with random under-sampling performed slightly better (AUROC=0.76, F1 score=0.33). ML models performed slightly better than LR-RUS (AUROC=0.75, F1 score=0.32) in predicting asthma attacks. Future research should explore other ML and data imbalance handling techniques to enhance risk prediction.

The UK Health Data Research Alliance presents five recommendations for improving data collection for inclusive health research.

Introduction The role of female sex hormones and their influence on asthma’s development and natural history remain uncertain. Our study aims to enhance understanding of exogenous sex hormones’ role in asthma development and manifestation, considering phenotypic heterogeneity and focusing on metabolic syndrome-linked asthma that has shown increased severity in females.Methods and analysis A cohort study using primary care data from the Clinical Practice Research Datalink (CPRD) databases linked with additional data sources (Hospital Episode Statistics, ethnicity and deprivation) will include individuals aged 16–70 years, spanning 1 January 2005 to 31 December 2019. We will use appropriate statistical learning methods depending on the outcome: extended Cox regression for late-onset asthma; Poisson or negative binomial regression for asthma exacerbations; binary logistic regression for asthma control; and ordered logistic regression for asthma severity. Asthma exacerbation will be defined based on the American Thoracic Society/European Respiratory Society Task Force definition as the presence of either one of an asthma-related accident and emergency department visit, an asthma-related (unscheduled) hospital admission or an acute course of oral corticosteroids (OCS) with evidence of asthma-related medical event and/or review within 2 weeks of OCS prescription. Poor asthma control in any given month will be defined by the occurrence of an exacerbation episode or use of short-acting beta agonist. Asthma severity will be defined based on the British Thoracic Society asthma severity steps. Asthma phenotypes will be identified using k-means clustering. Analyses will be undertaken using both GOLD and Aurum to ensure coverage across UK nations.Ethics and dissemination CPRD has received ethics approval from the Health Research Authority (East Midlands—Derby, REC reference number 21/EM/065) to support research using anonymised data. Approval to conduct this study was obtained through CPRD’s Research Data Governance process. The results will be disseminated through academic publications and conference presentations, contributing to the understanding and practice of asthma management, particularly in the context of the impacts of exogenous sex steroid hormones.

Background While clinical coding is intended to be an objective and standardized practice, it is important to recognize that it is not entirely the case. The clinical and bureaucratic practices from event of death to a case being entered into a research dataset are important context for analysing and interpreting this data. Variation in practices can influence the accuracy of the final coded record in two different stages: the reporting of the death certificate, and the International Classification of Diseases (Version 10; ICD-10) coding of that certificate.Methods This study investigated 91,022 deaths recorded in the Scottish Asthma Learning Healthcare System dataset between 2000 and 2017. Asthma-related deaths were identified by the presence of any of ICD-10 codes J45 or J46, in any position. These codes were categorized either as relating to asthma attacks specifically (status asthmatic; J46) or generally to asthma diagnosis (J45).Results We found that one in every 200 deaths in this were coded as being asthma related. Less than 1% of asthma-related mortality records used both J45 and J46 ICD-10 codes as causes. Infection (predominantly pneumonia) was more commonly reported as a contributing cause of death when J45 was the primary coded cause, compared to J46, which specifically denotes asthma attacks.Conclusion Further inspection of patient history can be essential to validate deaths recorded as caused by asthma, and to identify potentially mis-recorded non-asthma deaths, particularly in those with complex comorbidities.

Abstract Several population-level studies have described individual clinical risk factors associated with suboptimal antibody responses following COVID-19 vaccination, but none have examined multimorbidity. Others have shown that suboptimal post-vaccination responses offer reduced protection to subsequent SARS-CoV-2 infection; however, the level of protection from COVID-19 hospitalisation/death remains unconfirmed. We use national Scottish datasets to investigate the association between multimorbidity and testing antibody-negative, examining the correlation between antibody levels and subsequent COVID-19 hospitalisation/death among double-vaccinated individuals. We found that individuals with multimorbidity ( ≥ five conditions) were more likely to test antibody-negative post-vaccination and 13.37 [6.05–29.53] times more likely to be hospitalised/die from COVID-19 than individuals without conditions. We also show a dose-dependent association between post-vaccination antibody levels and COVID-19 hospitalisation or death, with those with undetectable antibody levels at a significantly higher risk (HR 9.21 [95% CI 4.63–18.29]) of these serious outcomes compared to those with high antibody levels.

We sought to investigate the incidence of severe COVID-19 outcomes after treatment with antivirals and neutralising monoclonal antibodies, and estimate the comparative effectiveness of treatments in community-based individuals. We conducted a retrospective cohort study investigating clinical outcomes of hospitalisation, intensive care unit admission and death, in those treated with antivirals and monoclonal antibodies for COVID-19 in Scotland between December 2021 and September 2022. We compared the effect of various treatments on the risk of severe COVID-19 outcomes, stratified by most prevalent sub-lineage at that time, and controlling for comorbidities and other patient characteristics. We identified 14,365 individuals treated for COVID-19 during our study period, some of whom were treated for multiple infections. The incidence of severe COVID-19 outcomes (inpatient admission or death) in community-treated patients (81% of all treatment episodes) was 1.2% (n = 137/11894, 95% CI 1.0-1.4), compared to 32.8% in those treated in hospital for acute COVID-19 (re-admissions or death; n = 40/122, 95% CI 25.1-41.5). For community-treated patients, there was a lower risk of severe outcomes (inpatient admission or death) in younger patients, and in those who had received three or more COVID-19 vaccinations. During the period in which BA.2 was the most prevalent sub-lineage in the UK, sotrovimab was associated with a reduced treatment effect compared to nirmaltrelvir + ritonavir. However, since BA.5 has been the most prevalent sub-lineage in the UK, both sotrovimab and nirmaltrelvir + ritonavir were associated with similarly lower incidence of severe outcomes than molnupiravir. Around 1% of those treated for COVID-19 with antivirals or neutralising monoclonal antibodies required hospital admission. During the period in which BA.5 was the prevalent sub-lineages in the UK, molnupiravir was associated with the highest incidence of severe outcomes in community-treated patients.

Prognostic models hold great potential for predicting asthma exacerbations, providing opportunities for early intervention, and are a popular area of current research. However, it is unclear how models should be compared and contrasted, given their differences in both design and performance, particularly with a view to potential implementation in routine practice. This systematic review aimed to identify novel predictive models of asthma attacks in adults and compare differences in construction related to populations, outcome definitions, prediction time horizons, algorithms, validation, and performance estimation. Twenty-five studies were identified for comparison, with varying definitions of asthma attacks and prediction event time horizons ranging from 15 days to 30 months. The most commonly used algorithm was logistic regression (20/25 studies); however, none of the six which tested multiple algorithms identified it as highest performing algorithm. The effect of various study design characteristics on performance was evaluated in order to provide context to the limitations of highly performing models. Models used a variety of constructs, which affected both their performance and their viability for implementation in routine practice. Consultation with stakeholders is necessary to identify priorities for model refinement and to create a benchmark of acceptable performance for implementation in clinical practice.