BWBrian Harvey Willis

Brian Harvey Willis

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

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article

Previous Organizations / Institutes

The organizations / institutes from the scholar's other articles in the past ten years


Brian Harvey Willis has a total of 168 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Open Access

    Information theory and diagnostic reasoning in primary care: a framework for navigating uncertainty

    by Brian H Willis

    This is an overview of how information theory offers a rigorous and intuitive framework for understanding diagnostic reasoning in primary care. Primary care clinicians frequently encounter patients early in their illness trajectory, when symptoms are vague, evolving, or non-specific, creating high diagnostic uncertainty. Information theory, through its core concepts of entropy, information content, and information gain, provides a structured way to quantify this uncertainty and understand how it changes as new clinical information is gathered. Entropy reflects the uncertainty associated with the breadth of the differential diagnosis; information content captures how surprising or discriminating a finding is; and information gain describes how much a particular question, sign, or test reduces uncertainty. Applied to the consultation, these principles explain why rare but specific features are disproportionately informative, why sequences of low-yield tests may add little value, and how targeted history-taking can outperform broad investigation. The framework complements Bayesian reasoning, hypothetico-deductive approaches, and intuitive pattern recognition by clarifying the informational value that underpins each. Through clinical examples, the paper illustrates how information theory sharpens diagnostic strategy, supports efficient use of investigations, and helps clinicians navigate uncertainty with greater clarity and safety. Ultimately, it positions diagnosis as a scientific process of uncertainty reduction central to primary care practice.

  • Open Access

    Less continuity with more complaints: a repeated cross-sectional study of the association between relational continuity of care and patient complaints in English general practice

    by Jinyang Chen, Panos Kasteridis, Zecharias Anteneh, Sheila Greenfield, Fiona Scheibl, Kamil Sterniczuk, Brian H Willis, Iestyn Williams, Tom Marshall

    Objective Relational continuity of care is associated with better patient experience and health outcomes. In England, relational continuity of primary care has been declining over a decade, coinciding with an increase in patient complaints. This study investigates the relationship between relational continuity of care and patient complaints. Methods Cross-sectional analysis of linked practice-level data in the English National Health Service (NHS) (2016/2017–2022/2023) obtained from NHS Digital and General Practice Patient Survey (GPPS). A negative binomial model was used to investigate the association between the proportion of patients never or almost never seeing their preferred general practitioner (GP) and new written complaints per 10 000 patients, with adjustment for patient demographics, socioeconomic status, care experiences, practice care capacity and care quality. Mediation analysis was further conducted to examine patients’ lost trust and unmet clinical needs as potential mechanisms. Results A 10 percentage point increase in the proportion of patients reporting low continuity was associated with 1.34 more new complaints per 10 000 patients (95% CI 1.23 to 1.46). The association may be stronger after than before the pandemic, among general practices with historically better continuity, and in more deprived areas. The findings were robust in using different measures of relational continuity, adjusting for primary case demand–supply mismatches, implementing a Poisson model with practice fixed effects and excluding ethnicity from the model specification. Mediation analysis showed that neither lost trust nor unmet care needs were important mediators of the effects of low continuity. Conclusion Self-reported low continuity of primary care is associated with more patient complaints in England. Future research should explore potential underlying mechanisms and establish whether the same relationship exists between objectively measured relational continuity and patient complaints.


  • Open Access

    A nonparametric clustering stopping rule based on spatial median

    by Hend Gabr, Brian H. Willis, Mohammed Baragilly

    In this work, we introduce a nonparametric clustering stopping rule algorithm based on the spatial median. Our proposed method aims to achieve a balance between the homogeneity within the clusters and the heterogeneity between clusters. The proposed algorithm maximises the ratio of the variation between clusters and the variation within clusters while adjusting for the number of clusters and number of observations. The proposed algorithm is robust against distributional assumptions and the presence of outliers. Simulations were used to validate the algorithm. We further evaluated the stability and the efficacy of the proposed algorithm using three real-world datasets. Moreover, we compared the performance of our model with 13 other traditional algorithms for determining the number of clusters. We found that the proposed algorithm outperformed 11 of the algorithms considered for comparison in terms of clustering number determination. The finding demonstrates that the proposed method provides a reliable alternative to determine the number of clusters for multivariate data.


  • Open Access

    Quantifying, Understanding and Enhancing Relational Continuity of Care (QUERCC): a mixed-methods protocol

    by Tom Marshall, Fiona Scheibl, Iestyn Williams, Krishnarajah Nirantharakumar, Brian H Willis, Panagiotis Kasteridis, Kamil Sterniczuk, Jinyang Chen, Zecharias Fetene Anteneh, Sheila Greenfield

    Introduction Relational continuity of care is where patients see the same clinicians over time. Evidence suggests relational continuity of care is valued by patients and clinicians and results in better health. While current National Health Service policy aims to maintain relational continuity of care, it has been declining in recent years, which may be linked to the growth in practice size, increased staff turnover, part-time working and the focus on patient access. Our research aims to develop resources to help clinicians measure, manage and improve relational continuity of care. Methods and analysis A mixed-methods approach in UK primary care commencing with two workshops drawing patients, clinicians and researchers together to establish an agreed approach on the measurement of continuity of care. Second, analysis of national data will provide insight into how staff turnover, part time working, practice size and funding per patient affects continuity. Third, case studies in a sample of high-performing practices will document the barriers and facilitators to the establishment and maintenance of continuity of care. Fourth, an economic analysis of resource costs and health outcomes using linked primary and secondary care data will show whether costs influence continuity for different patient groups (by age, sex, deprivation status and chronic disease status). Fifth, we will develop practical guidance for clinicians to improve continuity of care, based on the findings from each stage of the research. Ethics and dissemination The study has approval from HRA Health and Care Research Wales Research Ethics Committee (HCRW). Findings will be disseminated through peer-reviewed publications, participatory workshops, podcasts, clinical networks and academic conferences.


  • Open Access

    Functional Data Clustering Based on Weighted Functional Spatial Ranks With Clinical Applications

    by Mohammed Baragilly, Hend Gabr, Brian H. Willis

    Functional data analysis is receiving increasing attention in several scientific disciplines. However, identifying and classifying clusters of data that are essentially curves that map into an infinite dimensional space poses a significant challenge for existing methods. Here, we introduce weighted functional spatial ranks (WFSRs) as part of a nonparametric clustering approach for functional data analysis. A two‐stage or filtering method is used to approximate the curves into some basis functions and reduce the dimension of the data using functional principle components analysis (FPCA). The curves are then ranked based on WFSRs to create a contour map. This allows the visualization of the cluster structure and the size and content of each cluster to be ascertained. The effectiveness of the methods in functional data analysis is evaluated using numerical examples from simulated and two real medical datasets. Compared with several other cluster methods, the WFSR algorithm records the lowest misclassification rates over the two real datasets.


  • Open Access

    Clustering Analysis of Multivariate Data: A Weighted Spatial Ranks-Based Approach

    by Mohammed H. Baragilly, Hend Gabr, Brian H. Willis

    Determining the right number of clusters without any prior information about their numbers is a core problem in cluster analysis. In this paper, we propose a nonparametric clustering method based on different weighted spatial rank (WSR) functions. The main idea behind WSR is to define a dissimilarity measure locally based on a localized version of multivariate ranks. We consider a nonparametric Gaussian kernel weights function. We compare the performance of the method with other standard techniques and assess its misclassification rate. The method is completely data-driven, robust against distributional assumptions, and accurate for the purpose of intuitive visualization and can be used both to determine the number of clusters and assign each observation to its cluster.


  • Open Access

    Optimising a coordinate ascent algorithm for the meta-analysis of test accuracy studies

    by Mohammed Baragilly, Brian H Willis

    Meta-analysis may be used to summarise a test’s accuracy. Often the sensitivity and specificity are the measures of interest and as these are correlated a bivariate random effects model is commonly used to fit the data. This model has five parameters and it may be optimised using a Newton-Raphson based algorithm providing adequate initial values of the parameters are identified. Numerical methods may be used to estimate robust initial values but estimating these is computationally expensive and it is not clear whether they provide a significant advantage over closed form methods in terms of reducing bias, mean square error, average relative error, and coverage probability. Here we consider six closed form methods for estimating the initial values of the parameters for a co-ordinate ascent algorithm used to fit the bivariate model and compare them with numerically derived robust initial values. Using simulation studies we demonstrate that all the closed form methods lead to a reduction in computation time of around 80% and rank higher overall across the metrics when compared with the robust initial values method. Although no initial values estimator dominated the others across all parameters and metrics, the two-step Hedges-Olkin estimator ranked highest overall across the different scenarios.

  • Open Access

    Risk of developing gallbladder cancer in patients with gallbladder polyps detected on transabdominal ultrasound: a systematic review and meta-analysis

    by Kieran G Foley, Zena Riddell, Bernadette Coles, S Ashley Roberts, Brian H Willis

    Objective: To estimate the risk of malignancy in gallbladder polyps of incremental sizes detected during transabdominal ultrasound (TAUS).Methods: We searched databases including MEDLINE, Embase, and Cochrane Library for eligible studies recording the polyp size from which gallbladder malignancy developed, confirmed following cholecystectomy, or by subsequent follow-up. Primary outcome was the risk of gallbladder cancer in patients with polyps. Secondary outcome was the effect of polyp size as a prognostic factor for cancer. Risk of bias was assessed using the Quality in Prognostic Factor Studies (QUIPS) tool. Bayesian meta-analysis estimated the median cancer risk according to polyp size. This study is registered with PROSPERO (CRD42020223629).Results: 82 studies published since 1990 reported primary data for 67,837 patients. 67,774 gallbladder polyps and 889 cancers were reported. The cumulative median cancer risk of a polyp measuring 10 mm or less was 0.60% (99% credible range 0.30–1.16%). Substantial heterogeneity existed between studies (I2= 99.95%, 95% credible interval 99.86–99.98%). Risk of bias was generally high and overall confidence in evidence was low. 13 studies (15.6%) were graded with very low certainty, 56 studies (68.3%) with low certainty, and 13 studies (15.6%) with moderate certainty. In studies considered moderate quality, TAUS monitoring detected 4.6 cancers per 10,000 patients with polyps less than 10 mm.Conclusion: Malignant risk in gallbladder polyps is low, particularly in polyps less than 10 mm, however the data are heterogenous and generally low quality. International guidelines, which have not previously modelled size data, should be informed by these findings.Advances in knowledge: This large systematic review and meta-analysis has shown that the mean cumulative risk of small gallbladder polyps is low, but heterogeneity and missing data in larger polyp sizes (>10 mm) means the risk is uncertain and may be higher than estimated. Studies considered to have better methodological quality suggest that previous estimates of risk are likely to be inflated.


  • Open Access

    Measuring and exploring mental health determinants: a closer look at co-residents’ effect using a multilevel structural equations model

    by Hend Gabr, Mohammed Baragilly, Brian H. Willis

    Objective: Previous research has demonstrated that individual risk of mental illness is associated with individual, co-resident, and household risk factors. However, modelling the overall effect of these risk factors presents several methodological challenges. In this study we apply a multilevel structural equation model (MSEM) to address some of these challenges and the impact of the different determinants when measuring mental health risk. Study design and setting: Two thousand, one hundred forty-three individuals aged 16 and over from 888 households were analysed based on the Household Survey for England-2014 dataset. We applied MSEM to simultaneously measure and identify psychiatric morbidity determinants while accounting for the dependency among individuals within the same household and the measurement errors. Results: Younger age, female gender, non-working status, headship of the household, having no close relationship with other people, having history of mental illness and obesity were all significant (p < 0.01) individual risk factors for psychiatric morbidity. A previous history of mental illness in the co-residents, living in a deprived household, and a lack of closeness in relationships among residents were also significant predictors. Model fit indices showed a very good model specification (CFI = 0.987, TLI = 0.980, RMSEA = 0.023, GFI = 0.992). Conclusion: Measuring and addressing mental health determinants should consider not only an individual’s characteristics but also the co-residents and the households in which they live.


  • Long-term impact of pre-incision antibiotics on children born by caesarean section: a longitudinal study based on UK electronic health records

    by Dana Šumilo, Krishnarajah Nirantharakumar, Brian H Willis, Gavin M Rudge, James Martin, Krishna Gokhale, Rasiah Thayakaran, Nicola J Adderley, Joht Singh Chandan, Kelvin Okoth, et al.

    Background Since changes in the national guidance in 2011, prophylactic antibiotics for women undergoing caesarean section are recommended prior to skin incision, rather than after the baby’s umbilical cord has been clamped. Evidence from randomised controlled trials conducted outside the UK has shown that this reduces maternal infectious morbidity; however, the prophylactic antibiotics also cross the placenta, meaning that babies are exposed to them around the time of birth. Antibiotics are known to affect the gut microbiota of the babies, but the long-term effects of exposure to high-dose broad-spectrum antibiotics around the time of birth on allergy and immune-related diseases are unknown. Objectives We aimed to examine whether or not in-utero exposure to antibiotics immediately prior to birth compared with no pre-incisional antibiotic exposure increases the risk of (1) asthma and (2) eczema in children born by caesarean section. Design This was a controlled interrupted time series study. Setting The study took place in primary and secondary care. Participants Children born in the UK during 2006–18 delivered by caesarean section were compared with a control cohort delivered vaginally. Interventions In-utero exposure to antibiotics immediately prior to birth. Main outcome measures Asthma and eczema in children in the first 5 years of life. Additional secondary outcomes, including other allergy-related conditions, autoimmune diseases, infections, other immune system-related diseases and neurodevelopmental conditions, were also assessed. Data sources The Health Improvement Network (THIN) and the Clinical Practice Research Datalink (CPRD) primary care databases and the Hospital Episode Statistics (HES) database. Previously published linkage strategies were adapted to link anonymised data on mothers and babies in these databases. Duplicate practices contributing to both THIN and the CPRD databases were removed to create a THIN–CPRD data set. Results In the THIN–CPRD and HES data sets, records of 515,945 and 3,945,351 mother–baby pairs were analysed, respectively. The risk of asthma was not significantly higher in children born by caesarean section exposed to pre-incision antibiotics than in children whose mothers received post-cord clamping antibiotics, with an incidence rate ratio of 0.91 (95% confidence interval 0.78 to 1.05) for diagnosis of asthma in primary care and an incidence rate ratio of 1.05 (95% confidence interval 0.99 to 1.11) for asthma resulting in a hospital admission. We also did not find an increased risk of eczema, with an incidence rate ratio of 0.98 (95% confidence interval 0.94 to1.03) and an incidence rate ratio of 0.96 (95% confidence interval 0.71 to 1.29) for diagnosis in primary care and hospital admissions, respectively. Limitations It was not possible to ascertain the exposure to pre-incision antibiotics at an individual level. The maximum follow-up of children was 5 years. Conclusions There was no evidence that the policy change from post-cord clamping to pre-incision prophylactic antibiotics for caesarean sections during 2006–18 had an impact on the incidence of asthma and eczema in early childhood in the UK. Future work There is a need for further research to investigate if pre-incision antibiotics have any impact on developing asthma and other allergy and immune-related conditions in older children. Study registration This study is registered as researchregistry3736. Funding This project was funded by the National Institute for Health and Care Research (NIHR) Health Technology Assessment programme and will be published in full in Health Technology Assessment; Vol. 26, No. 30. See the NIHR Journals Library website for further project information.