MSMansour Taghavi Azar Sharabiani

Mansour Taghavi Azar Sharabiani

*This is an algorithmically generated author record based on publicly available information.

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article


Mansour Taghavi Azar Sharabiani has a total of 93 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Open Access

    Estimating the potential cardiovascular benefit of lifestyle risk factors using causal machine learning: Insights from UK Biobank

    by Christophe A.T. Stevens, Julia Brandts, Amany Elshorbagy, Fotios Barkas, Didac Llop Paredes, Kanika I. Dharmayat, Irene Karungi, Sunganani Kwilasi, Alireza Mahani, Mansour T.A. Sharabiani, et al.

  • Open Access

    Multimorbidity, health service use, and health insurance by socioeconomic groups in 31 countries: A multi-cohort study

    by Yanshang Wang, Chang Cai, Zhenyu Shi, Qian Gao, Alex Bottle, Mansour Taghavi Azar Sharabiani, Joshua Stott, Benedict Hayhoe, Ping He

    Background: The prevalence of physical, psychological, and cognitive multimorbidity is characterised by marked socioeconomic status (SES) inequalities. However, the relationships between multimorbidity patterns—particularly those involving cognitive conditions—and healthcare utilisation, as well as the role of health insurance, remain poorly understood. This study aims to explore healthcare-seeking behaviour among individuals with multimorbidity and assess whether these associations vary by SES and health insurance coverage. Methods and findings: This multicohort study analysed harmonised data from six longitudinal studies across 31 countries, including participants aged 50 years and older. Multimorbidity was defined as the coexistence of two or more disorders across physical, psychological, or cognitive disorders. Outpatient and inpatient healthcare utilisation were measured. Random-effects logistic regression models were used to estimate associations with healthcare utilisation, and random-effects negative binomial models were applied to analyse visit frequencies. All models were adjusted for age, gender, educational attainment, work status, marital status, and SES, as well as lifestyle factors. Country-specific estimates were pooled using multinational meta-analysis to generate overall effect sizes. Compared with individuals without any conditions, those with the most complex multimorbidity pattern (physical-psychological-cognitive multimorbidity) were more likely to use outpatient care (OR 3.21, 95% CI [2.39, 4.03]; p < 0.001) but not as high as those with physical-psychological multimorbidity (OR 7.84, 95% CI [6.59, 9.10]; p < 0.001). Additionally, the association varied across socioeconomic groups, individuals of lower SES experiencing more pronounced disparities in healthcare use. For inpatient care, adding a cognitive disorder to an existing physical or psychological condition was not associated with increased inpatient utilisation. Among individuals with health insurance coverage, the association between multimorbidity and outpatient care utilisation was generally attenuated. This was especially evident for those with physical-psychological-cognitive multimorbidity: insured individuals had an OR of 6.22 (95% CI [5.33, 7.25]; p < 0.001), compared with 3.40 (95% CI [3.03, 3.82]; p < 0.001) among uninsured individuals. A limitation of this study is that healthcare utilisation measures differed across cohorts and were harmonised retrospectively. Conclusions: Cognitive disorders further complicate the relationship between multimorbidity and health service use, indicating potential unmet healthcare needs, especially among individuals with lower SES. Our study highlights the potential role of health insurance in reducing socioeconomic disparities in healthcare utilisation associated with multimorbidity.


  • Open Access

    TabuLLM: Feature Extraction and Interpretation of Text Columns in Tabular Data Using Large Language Models

    by Mansour T.A. Sharabiani, Alex Bottle, Alireza Mahani

    TabuLLM is a Python package for extracting and interpreting LLM-based text features from datasets that combine free-text and structured columns. Motivated by a clinical study predicting postoperative acute kidney injury in paediatric cardiopulmonary bypass patients from electronic health records, the package integrates large language model (LLM) embeddings into scikit-learn-compatible pipelines for tabular datasets containing text columns alongside structured features. The package provides three core components: 1) TextColumnTransformer wraps the LangChain embedding ecosystem in a scikit-learn transformer interface with native multi-column handling; 2) GMMFeatureExtractor extends scikit-learn's Gaussian mixture model with a transform() method that returns per-cluster log-joint features for dimensionality reduction; and 3) ClusterExplainer generates natural language cluster descriptions using LLMs with recursive summarization, optional outcome-based statistical testing, and narrative synthesis. On a fraud detection benchmark, incorporating text features via TabuLLM improves test AUC from 0.92 to 0.98, with stacking ensemble variants reaching 0.99. TabuLLM thus lowers the barrier to text-tabular integration: state-of-the-art embeddings become a drop-in addition to existing scikit-learn pipelines, with interpretability tools to understand what the text features capture. The package is available via pip install tabullm.

  • Open Access

    Towards goal-directed perfusion - Part II: Temperature dependence of Q 10 , oxygen-demand dependence on age and weight, further clinical insights

    by Mansour T. A. Sharabiani, Alireza S. Mahani, Richard W. Issitt, Yadav Srinivasan, Alex Bottle, Serban Stoica

    Background: Goal-directed perfusion (GDP) during cardiopulmonary bypass (CPB) seeks to match oxygen delivery with patient-specific metabolic demand, yet determinants of oxygen demand and dynamics of oxygen extraction ratio (OER) are not well characterised in paediatric populations. Building on our GARIX (Global AutoRegressive Integrated model with eXogenous variables and an equilibrium force) model from Part 1, we analysed minute-level intraoperative data to derive clinical insights. Methods: GARIX integrates autoregressive memory, exogenous clinical variables (cardiac index, haemoglobin, arterial oxygen saturation, temperature), and an equilibrium component defining indexed oxygen demand (tVO 2 i) as a function of age, weight, and temperature. We applied GARIX to 20,443 minutes of data from 293 paediatric CPB procedures to study: (1) age- and weight-related variation in oxygen demand; (2) temperature dependence of demand, estimating local and interval Q 10 ; (3) age-stratified OER responses to haemoglobin and other variables; and (4) time evolution of the mismatch between oxygen consumption and demand after step changes in oxygen delivery. Results: tVO 2 i followed a non-monotonic pattern, rising from birth to ∼3 years and declining thereafter. Metabolic demand decreased nonlinearly with temperature, with a 17% reduction from 37°C to 32°C and 39% further reduction to 27°C. Local Q 10 ranged from ∼1.3 near normothermia to ∼3.8 in deep hypothermia. OER responsiveness to haemoglobin was blunted in neonates and infants compared with older children. Simulations revealed transient over- and under-oxygenation following step changes in oxygen delivery due to lags in OER adaptation. Conclusions: We identify risks of over- and under-oxygenation during CPB despite normal OER readings, challenge the use of fixed targets and constant Q 10 , and highlight age- and weight-dependent dynamics. These findings support individualised, time-aware perfusion strategies and lay the groundwork for model-guided decision support and control.


  • Open Access

    Are we getting better over time? Clinical and patient-reported outcomes for reverse shoulder arthroplasty: a National Joint Registry cohort study

    by Olivia O’mAlley, Andrew Davies, Mansour Taghavi Azar Sharabiani, Amar Rangan, Sanjeeve Sabharwal, Peter Reilly

    Objectives This study aims to review whether both clinical and Patient Reported Outcome Measures (PROMs) of Reverse Shoulder Arthroplasty have improved over time using the National Joint Registry (NJR). Design This study is a population-based cohort study using the NJR and Hospital Episode Statistics for England. Setting Publicly funded hospitals and procedures in England from 1 January 2013 to 31 December 2021. Participants All patients that received a reverse shoulder arthroplasty (RSA) in the specified time period. Patients were excluded if they had less than 1 year of follow-up. Main outcome measures Primary outcome was revision at one year. Secondary outcomes were non-revision re-operation and mortality at one year, length of stay (LOS) and mean change in Oxford Shoulder Score (OSS) from pre-operatively to 6 months post-operatively. Results There were 24 411 RSA cases available for analysis. There was no significant improvement in revision rates over time; however, there was a significant reduction in non-revision re-operations (OR 0.93 (0.86–0.99) p=0.03) and mortality (0.96 (0.92–1.00) p=0.04). LOS over time improved with an average reduction of 0.24 days per year, ranging from a mean of 3.94 days in 2013 to 2.44 days in 2021 (pConclusion Over the 9-year period recorded in the NJR, revision rates were low and remained similar. There has, however, been an improvement in other clinical outcomes such as non-revision reoperation and mortality as well as functional outcomes and reduced LOS, which demonstrates progress in the quality of care provided to shoulder replacement patients and is suggestive of advancements in surgical techniques, perioperative management and rehabilitation strategies.


  • Open Access

    Towards goal-directed perfusion – Part I: Developing physiology-inspired mathematical framework using high-resolution paediatric CPB data

    by Mansour T. A. Sharabiani, Alireza S. Mahani, Richard W. Issitt, Yadav Srinivasan, Alex Bottle, Serban Stoica

    Background: During cardiopulmonary bypass (CPB), goal-directed perfusion (GDP) seeks to match oxygen delivery to metabolic demand, but the dynamics of oxygen extraction and intraoperative oxygen demand remain poorly understood, especially in paediatric populations. Existing models rely on limited data and assume, for example, a linear relationship between log oxygen demand and temperature. Methods: We developed GARIX (Global AutoRegressive Integrated model with eXogenous variables and an equilibrium force) to predict minute-by-minute changes in oxygen extraction ratio (OER) using high-resolution intraoperative data. GARIX combines: (1) an autoregressive term group (ATG) encoding memory of past OER; (2) an exogenous term group (XTG) incorporating recent and planned changes in cardiac index (CI), haemoglobin (Hb), SaO 2 , and temperature; and (3) an equilibrium term group (ETG) that aligns oxygen consumption with demand via nonlinear temperature and patient-specific terms (log age, log weight, and interaction). A baseline model (bGARIX) used a linear temperature term and no patient-specific covariates. We trained on 20,443 min from 293 paediatric CPB procedures and evaluated performance through repeated K-fold cross-validation, simulations, and bootstrap confidence intervals. Results: GARIX reproduced physiologically plausible OER dynamics. Lagged coefficients captured adaptive responses to CI, Hb, SaO 2 , and temperature. Equilibrium analysis estimated a Q 10 of 2.5 with bGARIX, matching textbook values (2.4–2.7). Likelihood ratio tests confirmed GARIX’s improved fit, revealing age- and weight-related heterogeneity and nonlinearity in temperature dependence. Variable importance showed ATG and XTG dominated predictive accuracy, underscoring the role of system dynamics. Simulations indicated slow OER response to CI and Hb and faster adaptation to temperature. Conclusions: GARIX offers an interpretable, physiology-aligned model of oxygenation dynamics in paediatric CPB, enabling estimation of dynamic responses and latent demand. It provides a foundation for clinical insights and future real-time monitoring and intelligent perfusion control.


  • Open Access

    Exploring associations between active school environments and children’s physical activity, mental health and educational performance in Greater London primary schools: the Health and Activity of Pupils in the Primary Years (HAPPY) study protocol

    by Bina Ram, Nancy Gullett, Amina Benkhelfa, Mark Cunningham, Mansour Taghavi Azar Sharabiani, Esther van Sluijs, Nadia Siddiqui, Melvyn Hillsdon, Carolyn Summerbell, Miranda Pallan, et al.

    Introduction School environments that encourage children to be physically active can embed lifelong positive health behaviours and contribute towards reducing health inequalities. The Health and Activity of Pupils in the Primary Years (HAPPY) study aims to: (1) explore the extent to which the WHO criteria for creating active school environments are implemented by primary schools and (2) examine associations between active school environments and children’s physical activity, mental health and educational performance. Methods and analysis The HAPPY study is a quasi-experimental study comprising: (1) a survey of state-funded Greater London primary schools to identify implementation of the WHO’s six criteria and (2) a cross-sectional study to examine associations between schools’ active environment score (derived from the school survey) and pupils’ physical activity, mental health and educational performance. For our cross-sectional study, we will recruit up to 1000 year-three children (aged 7–8 years). Our primary outcome is accelerometer (GENEActiv) assessed physical activity, our secondary outcomes are parent-reported child mental health (Strengths and Difficulties Questionnaire) and teacher-reported educational performance (age-related expectations). Using multilevel mixed-effects regression models, we will examine associations between the active environment score and physical activity. Physical activity will be included as a measure of acceleration and also different intensities (light, moderate, vigorous). We will repeat this analysis to examine associations between the active environment score and mental health and educational performance. We will adjust for school characteristics and area-level deprivation and include pupil characteristics (eg, sex, ethnic group) as covariates. Clustering at the school level will be included as a random effect. Ethics and dissemination Ethical approval has been obtained from Imperial College Research Ethics Committee (ref: 6800895). Findings will be disseminated through a summary report to all participating schools, peer-reviewed publications, presentations at national and international conferences and National Institute for Health and Care Research policy briefings.


  • Open Access

    GenAI exceeds clinical experts in predicting acute kidney injury following paediatric cardiopulmonary bypass

    by Mansour Sharabiani, Alireza Mahani, Alex Bottle, Yadav Srinivasan, Richard Issitt, Serban Stoica

    The emergence of large language models (LLMs) opens new horizons to leverage, often unused, information in clinical text. Our study aims to capitalise on this new potential. Specifically, we examine the utility of text embeddings generated by LLMs in predicting postoperative acute kidney injury (AKI) in paediatric cardiopulmonary bypass (CPB) patients using electronic health record (EHR) text, and propose methods for explaining their output. AKI could be a serious complication in paediatric CPB and its accurate prediction can significantly improve patient outcomes by enabling timely interventions. We evaluate various text embedding algorithms such as Doc2Vec, top-performing sentence transformers on Hugging Face, and commercial LLMs from Google and OpenAI. We benchmark the cross-validated performance of these ‘AI models’ against a ‘baseline model’ as well as an established clinically-defined ‘expert model’. The baseline model includes structured features, i.e., patient gender, age, height, body mass index and length of operation. The majority of AI models surpass, not only the baseline model, but also the expert model. An ensemble of AI and clinical-expert models improves discriminative performance by 23% compared to the baseline model. Consistency of patient clusters formed from AI-generated embeddings with clinical-expert clusters—measured via the adjusted rand index and adjusted mutual information metrics—illustrates the medical validity of LLM embeddings. We create a reverse mapping from the numeric embedding space to the natural-language domain via the embedding-based clusters, generating medical labels for the clusters in the process. We also use text-generating LLMs to summarise the differences between AI and expert clusters. Such ‘explainability’ outputs can increase medical practitioners’ trust in the AI applications, and help generate new hypotheses, e.g., by studying the association of cluster memberships and outcomes of interest.


  • Open Access

    Revision of reverse shoulder arthroplasty by indication

    by Olivia O'Malley, Andrew Davies, Mansour T. A. Sharabani, Amar Rangan, Sanjeeve Sabharwal, Peter Reilly

    Aims: Reverse shoulder arthroplasty (RSA), initially designed for cuff tear arthropathy (CTA), is now the most common choice of shoulder arthroplasty in both elective and trauma settings in the UK. Its use has rapidly increased for several indications, including osteoarthritis (OA) with an intact rotator cuff, acute trauma, and trauma sequalae. This study aims to review the revision rates of RSA by indication to assess how the implant is performing for indications for which it was not primarily designed.Methods: Data from the National Joint Registry were obtained from 1 April 2012 to 31 March 2022. Data were linked to Hospital Episode Statistics for England and National Mortality Data. RSAs were identified and sorted into mutually exclusive groups by indication. The primary outcome was first revision and the secondary outcome was non-revision reoperation.Results: The revision rates for RSA for CTA were 1.53% (95% CI 1.31 to 1.78) at one year, 3.21% (95% CI 2.86 to 3.60) at five years, and 4.97% (95% CI 4.23 to 5.84) at nine years. For primary OA, they were 1.21% (95% CI 0.95 to 1.54) at one year, 2.71% (95% CI 2.25 to 3.37) at five years, and 5.00% (95% CI 3.62 to 6.88) at nine years. For trauma, they were 1.51% (95% CI 1.15 to 1.99) at one year and 2.67% (95% CI 2.07 to 3.43) at five years. For trauma sequalae, they were 4.25% (95% CI 3.4 to 5.29) at one year and 7.12% (95% CI 5.90 to 8.59) at five years. Between indications, the revision rates were not statistically different except for trauma sequalae, which had a significantly increased risk of revision (p ≤ 0.001). Incidence of non-revision reoperation across the cohort was 1.1% (n = 283 patients), with the most common being manipulation under anaesthesia with or without capsular release (42.03%, n = 124) followed by subacromial decompression (21.02%, n = 62).Conclusion: This study reports on the range of indications for which RSA is being used in the UK. It demonstrated that, based on the largest analysis of RSA across a range of indications, the revision rates and secondary surgery rates are broadly similar except for trauma sequalae.Cite this article: Bone Jt Open 2025;6(6):691–699.


  • Open Access

    Performance of Continuous Digital Monitoring of Vital Signs with a Wearable Sensor in Acute Hospital Settings

    by Meera Joshi, Fahad M. Iqbal, Mansour Sharabiani, Hutan Ashrafian, Sonal Arora, Kenny McAndrew, Sadia Khan, Graham Cooke, Ara Darzi

    Background: Continuous vital sign monitoring using wearable sensors has gained traction for the early detection of patient deterioration, particularly with the advent of virtual wards. Objective: The objective was to evaluate the reliability of a wearable sensor for monitoring heart rate (HR), respiratory rate (RR), and temperature in acutely unwell hospital patients and to identify the optimal time window for alert generation. Methods: A prospective cohort study recruited 500 patients in a single hospital. Sensor readings were compared to standard intermittent nurse observations using Bland–Altman plots to assess the limits of agreement. Results: HR demonstrated good agreement with nurse observations (intraclass correlation coefficient [ICC] = 0.66, r = 0.86, p < 0.001), with a mean difference of 3.63 bpm (95% LoA: −10.87 to 18.14 bpm). RR exhibited weaker agreement (ICC = 0.20, r = 0.18, p < 0.001), with a mean difference of −2.72 breaths per minute (95% LoA: −10.91 to 5.47 bpm). Temperature showed poor to fair agreement (ICC = 0.30, r = 0.39, p < 0.001), with a mean difference of −0.57 °C (95% LoA: −1.72 to 0.58 °C). A 10 min averaging window was identified as optimal, balancing data retention and real-time alerting. Conclusions: Wearable sensors demonstrate potential for reliable continuous monitoring of vital signs, supporting their future integration into real-world clinical practice for improved patient safety.