Conrad Harrison

Conrad Harrison

Auto-generated profile, based on publicly available information

Conrad Harrison

Surgical Intervention Trials Unit, Nuffield Department of Orthopaedics, Rheumatology and Musculoskeletal Sciences, University of Oxford, Oxford, UK

Research Focus & Metrics

Academic Career Length

11Years (2016-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
29 March 2022 in Radiology
Verified
Journal Article
Closed access
RKRachel Y. L. Kuo
Rachel Y. L. Kuo
CHConrad Harrison
Conrad Harrison
TCTerry-Ann Curran
Terry-Ann Curran
BJBenjamin Jones
Benjamin Jones
AFAlexander Freethy
Alexander Freethy
DFDominic Furniss
Dominic Furniss
Artificial intelligence is noninferior to clinicians in terms of diagnostic performance in fracture detection, showing promise as a useful diagnostic tool. Background Patients with fractures are a common emergency presentation and may be misdiagnosed at radiologic imaging. An increasing number of studies apply artificial intelligence (AI) techniques t...

31 July 2021 in BMC Medical Research Methodology
Verified
Journal Article
Open access
PDF file
CHConrad J. Harrison
Conrad J. Harrison
CSChris J. Sidey-Gibbons
Chris J. Sidey-Gibbons
Background: Unstructured text, including medical records, patient feedback, and social media comments, can be a rich source of data for clinical research. Natural language processing (NLP) describes a set of techniques used to convert passages of written text into interpretable datasets that can be analysed by statistical and machine learning (ML) models. The purpose of this paper is to provide a practical introduction to contemporary techniques for the analysis of text-data, using freely-available software. Methods: We performed three NLP experiments using publicly-available data obtained from medicine review websites. First, we conducted lexicon-based sentiment analysis on open-text patient reviews of four drugs: Levothyroxine, Viagra, Oseltamivir and Apixaban. Next, we used unsupervised ML (latent Dirichlet allocation, LDA) to identify similar drugs in the dataset, based solely on their reviews. Finally, we developed three supervised ML algorithms to predict whether a drug review was associated with a positive or negative rating. These algorithms were: a regularised logistic regression, a support vector machine (SVM), and an artificial neural network (ANN). We compared the performance of these algorithms in terms of classification accuracy, area under the receiver operating characteristic curve (AUC), sensitivity and specificity. Results: Levothyroxine and Viagra were reviewed with a higher proportion of positive sentiments than Oseltamivir and Apixaban. One of the three LDA clusters clearly represented drugs used to treat mental health problems. A common theme suggested by this cluster was drugs taking weeks or months to work. Another cluster clearly represented drugs used as contraceptives. Supervised machine learning algorithms predicted positive or negative drug ratings with classification accuracies ranging from 0.664, 95% CI [0.608, 0.716] for the regularised regression to 0.720, 95% CI [0.664,0.776] for the SVM. Conclusions: In this paper, we present a conceptual overview of common techniques used to analyse large volumes of text, and provide reproducible code that can be readily applied to other research studies using open-source software.

24 January 2022 in PharmacoEconomics
Verified
Journal Article
Open access
PDF file
Joseph KwonJoseph Kwon
Joseph Kwon
LFLouise Freijser
Louise Freijser
EHElisabeth Huynh
Elisabeth Huynh
MHMartin Howell
Martin Howell
GCGang Chen
Gang Chen
Stavros PetrouStavros Petrou
Stavros Petrou
Background and Aims Patient-reported outcome measures (PROMs) for children (aged ≤ 18 years) present methodological challenges. PROMs can be categorised by their diverse underlying conceptual bases, including functional, disability and health (FDH) status; quality of life (QoL); and health-related quality of life (HRQoL). Some PROMs are designed to be accompanied by preference weights. PROMs should account for childhood developmental differences by incorporating age-appropriate health/QoL domains, guidance on respondent type(s) and design. This systematic review aims to identify generic multidimensional childhood PROMs and synthesise their characteristics by conceptual basis, target age, measurement considerations, and the preference-based value sets that accompany them. Methods The study protocol was registered in the Prospective Register of Systematic Reviews (CRD42021230833), and reporting followed Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines. We conducted systematic database searches for generic multidimensional childhood PROMs covering the period 2012–2020, which we combined with published PROMs identified by an earlier systematic review that covered the period 1992–2011. A second systematic database search identified preference-based value sets for generic multidimensional PROMs. The PROMs were categorised by conceptual basis (FDH status, QoL and HRQoL) and by target age (namely infants and pre-schoolers aged < 5 years, pre-adolescents aged 5–11, adolescents aged 12–18 and multi-age group coverage). Descriptive statistics assessed how PROM characteristics (domain coverage, respondent type and design) varied by conceptual basis and age categories. Involvement of children in PROM development and testing was assessed to understand content validity. Characteristics of value sets available for the childhood generic multidimensional PROMs were identified and compared. Results We identified 89 PROMs, including 110 versions: 52 FDH, 29 QoL, 12 HRQoL, nine QoL-FDH and eight HRQoL-FDH measures; 20 targeted infants and pre-schoolers, 29 pre-adolescents, 24 adolescents and 37 for multiple age groups. Domain coverage demonstrated development trajectories from observable FDH aspects in infancy through to personal independence and relationships during adolescence. PROMs targeting younger children relied more on informant report, were shorter and had fewer ordinal scale points. One-third of PROMs were developed following qualitative research or surveys with children or parents for concept elicitation. There were 21 preference-based value sets developed by 19 studies of ten generic multidimensional childhood PROMs: seven were based on adolescents’ stated preferences, seven were from adults from the perspective of or on behalf of the child, and seven were from adults adopting an adult’s perspective. Diverse preference elicitation methods were used to elicit values. Practices with respect to anchoring values on the utility scale also varied considerably. The range and distribution of values reflect these differences, resulting in value sets with notably different properties. Conclusion Identification and categorisation of generic multidimensional childhood PROMs and value sets by this review can aid the development, selection and interpretation of appropriate measures for clinical and population research and cost-effectiveness-based decision-making.

1 April 2018 in Neuromodulation: Technology at the Neural Interface
Verified
Journal Article
Closed access
PDF file
CHConrad Harrison
Conrad Harrison
SESarah Epton
Sarah Epton
SBStana Bojanic
Stana Bojanic
AGAlexander L. Green
Alexander L. Green
JFJames J. FitzGerald
James J. FitzGerald

3 April 2023 in JAMA Network Open
Verified
Journal Article
Closed access
PDF file
RKRakhshan Kamran
Rakhshan Kamran
LJLiam Jackman
Liam Jackman
CCCynthia Chan
Cynthia Chan
YSYujin Suk
Yujin Suk
CJChloe Jacklin
Chloe Jacklin
JRJeremy Rodrigues
Jeremy Rodrigues
Patient-reported outcome measures (PROMs) are self-report instruments used to measure elements of health from patients’ own perspectives.1 These measures can enable comparisons of clinicians’ performance, driving service improvements.2 By helping to identify bothersome symptoms, PROMs can enhance communication between clinicians and patients,3 enhance patient satisfaction,4 and improve care outcomes,5 including mortality.6 However, recent evidence suggests that the benefits of PROMs are lost without evidence-based implementation strategies to ensure they gain traction and uptake.7,8

29 October 2020 in Journal of Medical Internet Research
Verified
Journal Article
Open access
CHConrad Harrison
Conrad Harrison
BLBao Sheng Loe
Bao Sheng Loe
PLPrzemysław Lis
Przemysław Lis
Christopher GibbonsChristopher Gibbons
Christopher Gibbons
Patient-reported assessments are transforming many facets of health care, but there is scope to modernize their delivery. Contemporary assessment techniques like computerized adaptive testing (CAT) and machine learning can be applied to patient-reported assessments to reduce burden on both patients and health care professionals; improve test accuracy; and provide individualized, actionable feedback. The Concerto platform is a highly adaptable, secure, and easy-to-use console that can harness the power of CAT and machine learning for developing and administering advanced patient-reported assessments. This paper introduces readers to contemporary assessment techniques and the Concerto platform. It reviews advances in the field of patient-reported assessment that have been driven by the Concerto platform and explains how to create an advanced, adaptive assessment, for free, with minimal prior experience with CAT or programming.