Publications
87 results found
Monje-Garcia L, Vairale J, Watt H, et al., 2023, Psychological well-being of people living with a colorectal cancer predisposition syndrome: evidence from a systematic review, Annual Meeting of the British-Society-of-Gastroenterology (BSG), Publisher: BMJ PUBLISHING GROUP, Pages: A43-A43, ISSN: 0017-5749
Creese H-M, Creese H-M, 2023, Risk factors of persistent adolescent thinness: findings from the UK millennium cohort study, BMC Public Health, Vol: 23, Pages: 1-12, ISSN: 1471-2458
BackgroundThinness during adolescence can increase the risk of adverse health outcomes across the life-course and impede development. There is limited research examining the prevalence and determinants of persistent adolescent thinness in the United Kingdom (UK). We used longitudinal cohort data to investigate determinants of persistent adolescent thinness.MethodsWe analyzed data from 7,740 participants in the UK Millennium Cohort Study at ages 9 months, 7, 11, 14 and 17 years. Persistent thinness was defined as thinness at ages 11, 14 and 17; thinness was defined as an age- and sex-adjusted Body Mass Index (BMI) of less than 18.5 kg/m2. In total, 4,036 participants, classified either as persistently thin or at a persistent healthy weight, were included in the analyses. Logistic regression analyses were conducted to examine associations between 16 risk factors and persistent adolescent thinness by sex.ResultsThe prevalence of persistent thinness among adolescents was 3.1% (n = 231). Among males (n = 115), persistent adolescent thinness was significantly associated with non-white ethnicity, low parental BMI, low birthweight, low breastfeeding duration, unintended pregnancy, and low maternal education. Among females (n = 116), persistent adolescent thinness was significantly associated with non-white ethnicity, low birthweight, low self-esteem, and low physical activity. However, after adjusting for all risk factors, only low maternal BMI (OR: 3.44; 95% CI:1.13, 10.5), low paternal BMI (OR: 22.2; 95% CI: 2.35, 209.6), unintended pregnancy (OR: 2.49; 95% CI: 1.11, 5.57) and low self-esteem (OR: 6.57; 95% CI: 1.46,29.7) remained significantly associated with persistent adolescent thinness among males. After adjustment for all risk factors, not reaching the recommended physical activity levels (OR: 4.22; 95% CI: 1.82, 9.75) remained significantly associated with persistent adolescent thinness among females. No appreciable associa
Watt H, 2023, Data Wrangling Recipes in R, Publisher: bookdown
Preparing messy data for analysis using R software. Practical resources structured around the workflow with easy-to-modify code. Includes check list for preparing to merge data and crucial checks, plus recipe book-style instructions for pivoting/ restructuring datasets.
Watt H, 2023, Data wrangling recipes in R: software online resources structured around the workflow to prepare data for analysis, Royal Statistical Society annual conference 2023
Watt H, 2023, Teaching confidence intervals: strategies designed to enhance conceptual understand and to guard against the most serious misconceptions, Royal Statistical Society conference proceedings 2023
Watt H, 2023, Data wrangling recipes in R: merits of resources aligned around the workflow., Burwalls medical statistics conference 2023
Watt H, 2023, Graphs and images to each conceptual understanding of confidence intervals whilst steering away from major misconceptions., Burwalls medical statistics conference 2023
Amati F, Green J, Kitchin L, et al., 2023, Ethnicity as a predictor of outcomes of psychological therapies for anxiety and depression: a retrospective cohort analysis, Behavioural and Cognitive Psychotherapy, Vol: 51, Pages: 164-173, ISSN: 1352-4658
Background:Studies on predictors of outcomes of treatment for common mental health disorders (CMDs) in community mental health settings are scarce, and sample sizes are often small. Research on the impact of identifying as a member of an ethnic minority group on treatment outcomes is limited.Aims:To ascertain whether ethnicity is an independent predictor of outcome and the extent to which any association is mediated by other sociodemographic factors.Method:Retrospective observational study of anonymised treatment data collected for routine clinical purposes. Data were analysed from nine Improving Access to Psychological Therapy (IAPT) services from 2009 to 2016. Social functioning, ethnic group, age, gender, occupation and baseline severity of the mental health disorder were analysed as predictors of outcome.Results:Outcomes varied with ethnic group. Levels of occupation, social deprivation, initial morbidity and social functioning varied between ethnic groups at baseline. After adjustment for these factors the impact of ethnicity was attenuated and only some ethnic groups remained as significant independent predictors of treatment outcome.Conclusions:Ethnic minority status is a marker for multiple disadvantages. Some of the differences in outcome seen between ethnic groups may be the result of more general factors present in all ethnic groups but at greater intensity in some ethnic minority groups.
Watt H, 2023, Teaching Conceptual Understanding of p-Values and of ConfidenceIntervals, Whilst Steering Away from Common Misinterpretations., Burwalls teaching medical statistics., Editors: Medeiros Mirra, Farnell
Teaching strategies aiming to address:Concern that statistical inference is hard to understand. Distributions of sample statistics underpinning calculation methods arenotoriously challenging concepts. Definitions are based on repeated random samples, not practical study designs. P-values areoften interpreted using language that does not accurately reflect the information they contain.Concern over standards of statistical interpretation in the applied literature. Methods of statistical education and resulting varyingunderstanding of statistical inference contribute to this. Excessive focus on whether p < 0.05 leads to some people erroneouslydeclaring studies with very similar odds ratios as incompatible, merely because only one has p < 0.05.Setting. Master’s in public health statistics course at Imperial College London.Strategy. Refer regularly to calculated results amongst study participants and clarify that p-values and confidence intervals (CIs)reflect imprecision resulting from random choices in selection of participants. Select CI and p-value “definitions” that clearlyreflect their purpose within them. Report the assumption of random sampling from some greater population throughout.Include practical exercises that focus on graphs that further support conceptual understanding.Conclusions. Statistics educators should steer away from common misconceptions by focussing on conceptual understanding ofCIs and of p-values. Research is warranted that assesses the subsequent impact of these methods on student understanding.
Hoang K, Watt H, Golemme M, et al., 2022, Noradrenergic add-on therapy with extended-release guanfacine in alzheimer’s disease: study protocol for a randomised clinical trial (NorAD) and COVID-19 amendments, Trials, Vol: 23, ISSN: 1745-6215
Background:Guanfacine is a α2A adrenergic receptor agonist approved for treating Attention Deficit Hyperactivity Disorder (ADHD). It is thought to act via postsynaptic receptors in the prefrontal cortex, modulating executive functions including the regulation of attention. Attention is affected early in Alzheimer’s Disease (AD), and this may relate to pathological changes within the locus coeruleus, the main source of noradrenergic pathways within the brain. Given that cholinergic pathways, also involved in attention, are disrupted in AD, the combination of noradrenergic and cholinergic treatments may have a synergistic effect in symptomatic AD. The primary objective of the NorAD trial is to evaluate change in cognition with 12 weeks treatment of extended-release guanfacine (GXR) against a placebo as a combination therapy with cholinesterase inhibitors in participants with mild to moderate Alzheimer’s Disease.Methods/Design:NorAD is a 3-month, single-centre, randomised, double-blind, placebo-controlled, phase III trial of extended-release guanfacine (GXR) in participants with mild to moderate Alzheimer’s Disease. A total of 160 participants will be randomised to receive either daily guanfacine or placebo in combination with approved cholinesterase treatment for 12 weeks. The primary outcome is change in cognition, as measured by the Alzheimer’s Disease Assessment Scale-Cognitive subscale (ADAS-Cog), from baseline to follow-up in the treatment group compared to the placebo group. Secondary outcomes include change in additional cognitive measures of attention (Tests of Attention: Trails A and B, Digit-symbol substitution, Test of Everyday attention and CANTAB-RVP), neuropsychiatric symptoms (Neuropsychiatric Inventory), caregiver burden (Zarit Burden Interview) and activities of daily living (Alzheimer’s Disease Co-operative Study – Activities of Daily Living Inventory). From July 2020 observation of change following cessation
Watt H, Leedham-Green K, Farnell D, et al., 2022, Survey of statistics educators' confidence interval “definitions”: errors in definitions, perception that they are "not practical" & non-standard definitions., Royal Statistical Society 2022 international conference
Watt H, Leedham-Green K, Farnell D, et al., 2022, Round Table: Survey of Statistics Educators: Many fail to recognise that p-values are not relevant to results amongst study participant’s values., Burwalls 2022: Annual Meeting for Teachers of Statistics in Medicine and Allied Health Sciences
Watt H, Leedham-Green K, Farnell D, et al., 2022, Round Table: Discussion on choice of confidence interval (CI) “definition”, Burwalls 2022: Annual meeting for teachers of statistics in medicine and allied health sciences
Watt H, Leedham-Green K, Farnell D, et al., 2022, Denying knowledge of differences amongst subjects when p>0.05; people are less likely to make this error when their p-value interpretation includes “population”, Royal Statistical Society 2022 International conference
Watt H, Leedham-Green K, Farnell D, et al., 2022, Survey of language used to interpret confidence intervals by statistics educators: features of interpretations that develop conceptual understanding, Royal Statistical Society 2022 international conference
Khanzada A, Mousa A, Thompson R, et al., 2021, The Association between GDP and non-communicable disease mortality in the WHO European Region, EUROPEAN JOURNAL OF PUBLIC HEALTH, Vol: 31, ISSN: 1101-1262
Loreto F, Gunning S, Golemme M, et al., 2021, COGNITIVE PERFORMANCE AND AFFECTIVE SYMPTOMS IN PATIENTS UNDERGOING CLINICAL AMYLOID PET IMAGING, JOURNAL OF NEUROLOGY NEUROSURGERY AND PSYCHIATRY, Vol: 92, ISSN: 0022-3050
O'Farrelly C, Barker B, Watt H, et al., 2021, A video-feedback parenting intervention to prevent enduring behaviour problems in at-risk children aged 12-36 months: the Healthy Start, Happy Start RCT, HEALTH TECHNOLOGY ASSESSMENT, Vol: 25, Pages: 1-+, ISSN: 1366-5278
- Author Web Link
- Cite
- Citations: 1
Loreto F, Gunning S, Golemme M, et al., 2021, Evaluating cognitive profiles of patients undergoing clinical Amyloid-PET Imaging, Brain Communications, ISSN: 2632-1297
Loreto F, Gunning S, Golemme M, et al., 2021, Evaluating cognitive profiles of patients undergoing clinical Amyloid-PET Imaging, Brain Communications, ISSN: 2632-1297
Episodic memory impairment and brain amyloid-beta (Aβ) are two of the main hallmarks of Alzheimer’s Disease (AD). In patients with suspected AD, these are often evaluated through neuropsychological testing and amyloid PET imaging (API), respectively. Crucially, the use of amyloid PET in clinical practice is only indicated in patients with substantial diagnostic uncertainty due to atypical clinical presentation, multiple comorbidities and/or early age of onset. The relationship between Aβ and cognition has been previously investigated, but no study has examined how neuropsychological features relate to the presence of amyloid pathology in the clinical population meeting the appropriate use criteria for API. In this study, we evaluated a clinical cohort of patients (n=107) who presented at the Imperial Memory Clinic and were referred for clinical API and neuropsychological assessment as part of their diagnostic workup. We compared the cognitive performance of amyloid-positive patients (Aβ-pos, n=47) with that of stable amyloid-negative (stableAβ-neg, n=26) and progressive amyloid-negative (progAβ-neg, n=34) patients. The Aβ-pos group performed significantly worse than both the amyloid-negative groups in the visuospatial and working memory domains. Episodic memory performance, instead, effectively differentiated the Aβ-pos group from the stableAβ-neg but not the progAβ-neg group. On affective questionnaires, the stableAβ-neg group reported significantly higher levels of depression than the Aβ-pos group. In our clinical cohort, visuospatial dysfunction and working memory impairment were better indicators of amyloid positivity than episodic memory dysfunction. These findings highlight the limited value of isolated cognitive scores in patients with atypical clinical presentation, comorbidities and/or early age of onset.
Watt H, 2020, Reflections on modern methods: statistics education beyond “significance”: novel plain English interpretations to deepen understanding of statistics and to steer away from misinterpretations, International Journal of Epidemiology, Vol: 49, Pages: 2083-2088, ISSN: 0300-5771
Concerns have been expressed over standards of statistical interpretation. Results with p<0.05 are often referred to as “significant” which, in plain English, implies important. This leads some people directly into the misconception that this provides proof that associations are clinically relevant. There are calls for statistics educators to respond to these concerns. This article provides novel plain English interpretations that are designed to deepen understanding. Experience teaching post-graduates at Imperial College is discussed.A key issue with focusing on “significance”, is the common inappropriate practice of implying no association exists, simply because p>0.05. Referring to strengths of association in “study participants” gives them gravitas, which may help to avoid this. This contrasts with the common practice of focusing on imprecision, by referring to the “sample” and to “point estimates”.Unlike formal statistical definitions, interpretations developed and presented here are rooted in the application of Statistics. They are based on one set of study participants (not many random samples). Precision of strengths of association are based on using strengths in study participants to estimate strengths of association in the population (from which participants were selected by probability random sampling). Reference to “compatibility with study data, dependent on statistical modelling assumptions”, reminds us of the importance of data quality and modelling assumptions. A straight-forward graph shows the relationship between p-values and test statistics. This figure and associated interpretations were developed to illuminate the continuous nature of p-values. This is designed to discourage focus on whether p<0.05, and encourage interpretation of exact p-values.
Boshari T, Sharpe C, Poots A, et al., 2020, An observational study of the association between diverse licensed premises types and alcohol-related violence in an inner-London borough, Journal of Epidemiology and Community Health, Vol: 74, Pages: 1016-1022, ISSN: 0143-005X
Background: An ecological correlation has been observed between licensed premises and alcohol-related violence (ARV). In the United Kingdom to date, no evidence directly connects alcohol-related harm to a single premises type. Recent policies have called for a diversified alcohol offer yet quantitative evidence in support remains sparse. This study aims to inform policy by determining whether diversification of the alcohol economy is desirable, and to inform the licensing process and submission of public health evidence. Methods: Using 11-years of local licensing data from the London Borough of Southwark, alcohol availability over time was approximated by the number of extant alcohol licences, categorised by outlet type: drinking establishments, eateries, takeaways, off-sales, and ‘other’. Harm was quantified drawing on law enforcement intelligence that recorded ARV. A linked dataset was analysed using negative binomial regression, contrasting cumulative impact zones (CIZ) – a common alcohol control policy – with non-CIZ geographies. Results: Each licensed drinking establishment was associated with a 1.6% (95% CI 0.7% to 2.6%; p=0.001) increase in ARV, respectively. ‘Other’ outlets had a protective effect and were associated with a 1.8% (95% CI 1.0% to 2.5%; p<0.001) decrease in ARV. Conclusion: This study provides direct evidence for an association between alcohol-related harm and licensed premises. The varying associations between outlet type and ARV provide local public health stakeholders with an evidence base upon which to advocate for licensing policies that diversify alcohol availability.
O'Farrelly C, Watt H, Babalis D, et al., 2020, A brief home-based parenting intervention (VIPP-SD) to reduce behaviour problems in young children: A pragmatic randomised clinical trial., JAMA Pediatrics, ISSN: 1072-4710
Sharpe CA, Poots A, Watt H, et al., 2019, An observational study to examine how Cumulative Impact Zones influence alcohol availability from different types of licensed outlets in an inner London Borough, BMJ Open, Vol: 9, Pages: 1-8, ISSN: 2044-6055
OBJECTIVES: Cumulative Impact Zones (CIZs) are a widely implemented local policy intended to restrict alcohol availability in areas proliferated with licensed outlets. Limited previous research has questioned their effectiveness and suggested they may play a more nuanced role in shaping local alcohol environments. This study evaluates the association between CIZ implementation and the number of licence applications made and the number issued, relative to a control region. DESIGN: A quantitative observational study.SETTING: The inner London Borough of Southwark, which currently enforces three CIZs. POPULATION: Licence applications received by Southwark Council’s Licensing Authority between 1 April 2006 and 31 March 2017 (N = 1254).INTERVENTIONS: CIZ implementation. PRIMARY OUTCOME MEASURES: Five outlet types were categorised and evaluated: Drinking Establishments, Eateries, Takeaways, Off Sales, and Other Outlets. Primary outcome measures were the number of applications received and the number of licences issued. These were analysed using Poisson regression of counts over time.RESULTS: Across all CIZs, implementation was associated with greater increases in the number of eateries in CIZ regions (IRR = 1.58, 95% CI: 1.02 – 2.52, P = 0.04) and number of takeaway venues (IRR = 3.89, 95% CI: 1.32 – 11.49, P = 0.01), relative to the control area. No discernible association was found for the remaining outlet types. Disaggregating by area indicated a 10-fold relative increase in the number of new eateries in Peckham CIZ (IRR = 10.38, 95% CI: 1.39 – 77.66, P = 0.02) and a four-fold relative increase in the number of newly licensed takeaways in Bankside CIZ (IRR = 4.38, 95% CI: 1.20 – 15.91, P = 0.03).CONCLUSIONS: Cumulative Impact Zones may be useful as policy levers to shape local alcohol environments to support the licensing goals of specific geographical areas and diversify the night-time economy.
Green K, Cooke O'Dowd N, Watt H, et al., 2019, Prescribing trends of gabapentin, pregabalin and oxycodone; a secondary analysis of primary care prescribing patterns in England, BJGP Open, Vol: 3, Pages: 1-10, ISSN: 2398-3795
BackgroundThe risk of iatrogenic harm from prescription drug use, misuse and abuse of drugssuch as gabapentin, pregabalin and oxycodone is substantial. In recent years, deathsassociated with these drugs in England have increased.AimsTo characterise general practice prescribing trends for gabapentin, pregabalin andoxycodone – termed dependence forming medicines (DFM) – in England and describepotential drivers of unwarranted variation.Design and SettingThis study is a retrospective secondary analysis of open source, publicly availablegovernment data from various sources pertaining to primary care demographics andprescriptions.MethodsThis study used five consecutive years (April 2013 – March 2018) of aggregate datato investigate longitudinal trends of prescribing and variation in prescribing trends atpractice and clinical commissioning group (CCG) level.ResultsAnnual prescriptions of gabapentin, pregabalin and oxycodone have increased eachyear over the period. Variation in prescribing trends were associated with GP practicedeprivation quintile, where the most deprived GP practices prescribe 313% (p<0.001)and 238% (p<0.001) greater volumes of gabapentin and pregabalin per person thanpractices in the least deprived quintile. The highest prescribing CCGs of each of thesedrugs were predominantly in northern and eastern regions of England.ConclusionsSubstantial increases in gabapentin, pregabalin and oxycodone prescriptions areconcerning and will increase iatrogenic harm from drug-related morbidity and mortality.More research is needed to understand the large variation in prescribing between general practices; and to develop and implement interventions to reduce unwarrantedvariation and increase the appropriateness of prescribing of these drugs.
Boshari T, Sharpe C, Poots A, et al., 2018, Public health and alcohol licensing policy in local government: an observational study of licensed premises and alcohol-related violence in London, UK, Public Health Science 2018, Publisher: Elsevier, Pages: S12-S12, ISSN: 0140-6736
Shather Z, Laverty A, Bottle RA, et al., 2018, Sustained socio-economic inequalities in hospital admissions for cardiovascular events among people with diabetes in England, The American Journal of Medicine, Vol: 131, Pages: 1340-1348, ISSN: 0002-9343
ObjectiveThis study aimed to determine changes in absolute and relative socio-economic inequalities in hospital admissions for major cardiovascular events and procedures among people with diabetes in England.MethodsWe identified all patients with diagnosed diabetes aged ≥45 years admitted to hospital in England between 2004-2005 and 2014-2015 for acute myocardial infarction, stroke, percutaneous coronary intervention (PCI) and coronary artery bypass graft (CABG). Socio-economic status was measured using Index of Multiple Deprivation. Diabetes-specific admission rates were calculated for each year by deprivation quintile. We assessed temporal changes using negative binomial regression models.ResultsMost admissions for cardiovascular causes occurred among people aged ≥65 years (71%) and men (63.3%), and the number of admissions increased steadily from the least to the most deprived quintile. People with diabetes in the most deprived quintile had 1.94-fold increased risk of acute myocardial infarction (95% CI 1.79-2.10), 1.92-fold risk of stroke (95% CI 1.78-2.07), 1.66-fold risk of CABG (95% CI 1.50-1.74), and 1.76-fold risk of PCI (95% CI 1.64-1.89) compared with the least deprived group. Absolute differences in rates between the least and most deprived quintiles did not significantly change for acute myocardial infarction (P=0.29) and were reduced for stroke, CABG and PCI (by 17.5, 15 and 11.8 per 100,000 people with diabetes, respectively, P≤0.01 for all).ConclusionsSocio-economic inequalities persist in diabetes-related hospital admissions for major cardiovascular events in England. Besides improved risk stratification strategies considering socio-economically defined needs, wide-reaching population-based policy interventions are required to reduce inequalities in diabetes outcomes.
Vamos E, Shather Z, Laverty A, et al., 2018, Socio-economic inequalities in hospital admissions for major cardiovascular events in people with diabetes in England, 54th Annual Meeting of the European-Association-for-the-Study-of-Diabetes (EASD), Publisher: SPRINGER, Pages: S576-S576, ISSN: 0012-186X
Sharpe CA, Poots AJ, Watt H, et al., 2018, Controlling alcohol availability through local policy: an observational study to evaluate Cumulative Impact Zones in a London borough, Journal of Public Health, Vol: 40, Pages: e260-e268, ISSN: 1741-3842
BackgroundCumulative impact zones (CIZs) are a discretionary policy lever available to local government, used to restrict the availability of alcohol in areas deemed already saturated. Despite little evidence of their effect, over 200 such zones have been introduced. This study explores the impact of three CIZs on the licensing of venues in the London Borough of Southwark.MethodsUsing 10 years of licensing data, we examined changes in the issuing of licences on the introduction of three CIZs within Southwark, relative to control areas. The number of licence applications made (N = 1110), the number issued, and the proportion objected to, were analysed using negative binomial regression.ResultsIn one area tested, CIZ implementation was associated with 119% more licence applications than control areas (incidence rate ratios (IRR) = 2.19, 95% confidence intervals (CI): 1.29–3.73, P = 0.004) and 133% more licences granted (IRR = 2.33, 95% CI: 1.31–4.16, P = 0.004). No significant effect was found for the other two areas. CIZs were found to have no discernible effect on the relative proportion of licence applications receiving objections.ConclusionsCIZs are proposed as a key lever to limit alcohol availability in areas of high outlet density. We found no evidence that CIZ establishment reduced the number of successful applications in Southwark.
Norheim G, Mueller JE, Njanpop-Lafourcade B-M, et al., 2018, Natural immunity against capsular group X N. meningitidis following an outbreak in Togo, 2007, Vaccine, Vol: 36, Pages: 1297-1303, ISSN: 0264-410X
BACKGROUND: Capsular group X N. meningitidis (MenX) has emerged as a cause of localized disease outbreaks in sub-Saharan Africa, but the human immune response following exposure to MenX antigens is poorly described. We therefore assessed the natural immunity against MenX in individuals who were living in an area affected by a MenX outbreak during 2007 in Togo, West Africa. During 2009, 300 healthy individuals (100 aged 3-5 years, 100 aged 13-19 years and 100 aged 20-25 years) were included in the study, and serum responses were compared with sera from age-matched controls from the U.K. and Burkina Faso. METHODS: MenX serum bactericidal antibody (SBA) was measured using rabbit complement, and antibodies against MenX polysaccharide (XPS) and outer membrane vesicles (XOMVs) were quantified by ELISA. RESULTS: The proportion of Togolese individuals with an SBA titer of ≥8 against the MenX strain was 29% (95% confidence interval (CI) 18-41) among those aged 3-5 years, 34% (95% CI 9-60) among those aged 13-19 years and 32% (95% CI 24-40) among those aged 20-25 years. These were significantly higher than observed in the control populations from the U.K (range 13-16%) and Burkina Faso (range 2-6%). CONCLUSION: In Togolese individuals, the concentration of serum IgG against XPS was higher among the two older age groups as compared to the youngest age group. Antibody concentrations against MenX PS correlated significantly with SBA titers. This supports further development of a MenX PS based conjugate vaccine. Further studies are needed to verify the ability of MenX PS to induce SBA in humans.
This data is extracted from the Web of Science and reproduced under a licence from Thomson Reuters. You may not copy or re-distribute this data in whole or in part without the written consent of the Science business of Thomson Reuters.