TDThomas C. Draper

Thomas C. Draper

*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


Thomas C. Draper has a total of 41 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Open Access

    AI-generated clinical summaries: errors and susceptibility to speech and speaker variability

    by Thomas C Draper, Jason Leake, Timothy Cox, Kathryn Lamb-Riddell, Benjamin E Johns, John McCormick, Stephen Trowell, Janice Kiely, Richard Luxton

    Objectives: To evaluate whether variability in patients’ communication style (personality), international English-accents (human and synthetic) and speech impairments affects the accuracy of a Clinical AI Scribe (CAIS) and identify where performance degrades to inform pre-deployment validation and monitoring. Methods: We conducted simulated primary-care consultations using trained actors. For personality types, four scenarios were enacted, each with five patient-personality types. For accents, transcripts of consultations were used to generate combinations of seven accents across five scenarios. The CAIS produced summaries that were compared with transcripts, and errors classified as omissions, factual inaccuracies or hallucinations. For speech impairments, public recordings representing five profiles were transcribed and word-recognition accuracy calculated. Results: Personality types showed no statistically significant differences in errors (all p>0.05). Extraversion had the highest total errors (median 3.5). Across accents, comparisons were non-significant for both patient and doctor voices (patients: p=0.851; doctors: p=0.980). Omissions predominated, with low rates of hallucinations and factual inaccuracies. Omissions were slightly higher for Chinese-accented and Indian-accented doctors (both medians 3.0). Conversely, speech impairments differed: cleft palate and vowel disorders were near-perfect, whereas phonological impairment markedly reduced recognition (p<0.001). Discussion: Operationally, CAIS deployment should include clinician-in-the-loop verification, subgroup performance monitoring (accents, impairments) and predefined ‘switch-off’ criteria for severe phonological patterns. High-quality synthetic voices are a pragmatic proxy for accent testing when balanced corpora are unavailable. Conclusions: Under controlled conditions CAIS performance was broadly stable across communication styles and most accents, but vulnerable to specific speech characteristics, particularly phonological impairment, in this single-system simulation study.


  • Open Access

    AI-Generated Clinical Summaries: Errors and Susceptibility to Speech and Speaker Variability

    by Thomas C. Draper, Jason Leake, Timothy Cox, Kathryn Lamb-Riddell, Benjamin E. Johns, John McCormick, Stephen Trowell, Janice Kiely, Richard Luxton

    Summary Box: What is already known on this topic: Clinical AI Scribe outputs can contain errors, and the impact of human factors (e.g. communication style, accents, speech impairments) in clinical contexts remains under-characterised. What this study adds: In controlled simulations, patient personality and accent did not significantly alter total CAIS errors, with omissions predominating and hallucinations/inaccuracies remaining low. Speech-impairment effects were highly varied, with near-perfect recognition for cleft palate and vowel disorders, whereas phonological impairment substantially reduced accuracy. How this study might affect research, practice or policy: Supports clinician-in-the-loop deployment with local validation across representative accents and impairment profiles, prioritising detection of clinically critical errors. Routine governance should include subgroup performance reporting (accents, impairments) and ongoing audit of error rates. Objective: The study aims to evaluate whether variability in patients’ communication style (personality, international English accents, and speech impairments) affects the accuracy of a Clinical AI Scribe (CAIS), and to identify where performance degrades. Method and Analysis: We conducted simulated primary-care consultations in a purpose-built lab using trained actors. To investigate personality types, four scenarios were enacted, each with five patient-personality types. For accents, human-verified transcripts of consultations were used to generate all doctor/patient combinations of seven different accents (including a synthetic reference voice) across five scenarios. The CAIS produced SOAP-structured summaries that were compared with the transcripts. Errors were classified as omissions, factual inaccuracies, or hallucinations. For speech impairments, public recordings representing five profiles were transcribed and word-recognition accuracy was calculated. Results: Personality types showed no statistically significant differences in errors (all p >0.05). Extraversion had the highest total errors (median 3.5), while conscientiousness and agreeableness were lower (1.5 and 2.0, respectively). Across accents, both pairwise tests and group comparisons were non-significant for both patient and doctor voices (patients: p =0.851; doctors: p =0.98). Omissions predominated, with low rates of hallucinations and factual inaccuracies. Omissions were slightly higher for Chinese- and Indian-accented doctors (both medians 3.0). In contrast, speech impairments differed: cleft palate and vowel disorders were near-perfect, whereas phonological impairment markedly reduced recognition ( p <0.001). Conclusions: Under controlled conditions, CAIS performance was broadly stable across communication styles and most accents but remained vulnerable to specific speech characteristics, particularly phonological impairment. Future evaluations using real-world, multi-speaker clinical audio are needed to confirm performance.


  • Open Access

    Clinical AI Scribes in primary care: accuracy, error severity and implications for clinical practice

    by Thomas C Draper, Timothy Cox, Kathryn Lamb-Riddell, Luigi Andrea Moretti, John McCormick, Stephen Trowell, Janice Kiely, Richard Luxton

    Objectives: To investigate the performance of commercially available Clinical Artificial Intelligence Scribes (CAISs), assessing their accuracy, potential clinical impact of errors, and documentation quality, given growing concerns around errors and safety. Methods and analysis: Seven CAIS products were investigated, using eight standardised clinical consultation scenarios recorded as audio. CAIS-generated summaries were assessed against a human-validated transcript and evaluated for errors (omissions, factual inaccuracies and hallucinations). Error severity was rated by medical doctors, generating a novel severity-weighted mpact Score (linear and exponential variants), to quantify potential clinical impact. Further analysis using the Physician Documentation Quality Instrument (PDQI-10) (a validated clinical note quality score) reinforced the findings. Results: Omissions dominated error counts (83.8%, p<<0.001), with CAISs varying widely in error frequency and severity, and a median of 1–6 omissions per consultation (depending on CAIS). Although less frequent, hallucinations and factual inaccuracies were more often clinically serious. No tested CAIS produced error-free summaries. The Impact Score highlighted clinical severity, notably amplifying the significance of less frequent but high-severity errors. PDQI-10 analysis indicated summaries were weakest in succinctness and organisation, but strong in consistency and clinical usefulness. Conclusions: The CAISs demonstrate high levels of summarisation accuracy. However, there is great disparity between the currently available CAIS products and, while some perform well, none are perfect. Clinicians should therefore maintain vigilance, particularly checking omitted psychosocial details and medications, and scrutinising plausible-sounding insertions. Purchasers and regulators should be aware of the significant performance disparities identified, reinforcing the need for careful evaluation and selection of CAIS products.


  • Open Access

    Impact of acoustic and informational noise on AI-generated clinical summaries

    by Thomas C Draper, Jason Leake, Kathryn Lamb-Riddell, Timothy Cox, John McCormick, Stephen Trowell, Janice Kiely, Richard Luxton

    Objectives: To investigate the effect of introducing environmental noise, microphone identity and position, and informational noise on the accuracy of a commercial Clinical AI Scribe (CAIS). Methods and analysis: Consultations on five medical conditions (memory loss, diarrhoea, headaches, skin rash or prostate symptoms) were recorded in a simulated primary care setting. A ground-truth audio file was produced, along with a ground-truth clinical summary via the CAIS. Different types of ‘noise’ at different levels were then investigated by (i) varying the type and distance of the microphone, (ii) introducing background noise (rain, baby, construction, toddler) at various levels and (iii) introducing informational noise (irrelevant discussions, medical and non-medical). These were then compared with the ground-truth summaries. The degradation of clinical summary accuracy was measured (omissions, hallucinations and inclusions), relative to the ground-truth summary. Results: Errors increased for all noise levels and noise types, principally being omissions. At 4.5 m from the consultation, for all microphones tested, all facts were omitted. At up to 2 m, all microphones, except the laptop, only introduced a small number (<5) of omissions. For background noise, the type of noise was found to correlate with omissions, with toddler and heavy rain being particularly deleterious. The CAIS was remarkably apt at rejecting additional informational noise. Conclusions: The CAIS generally excelled at capturing a doctor-patient consultation and producing accurate clinical summaries; however, error rate (especially omissions) increased notably when acoustic noise was introduced. The system was better able to handle informational noise.


  • Open Access

    The Impact of Acoustic and Informational Noise on AI-Generated Clinical Summaries

    by Thomas C. Draper, Jason Leake, Kathryn Lamb-Riddell, Timothy Cox, John McCormick, Stephen Trowell, Janice Kiely, Richard Luxton

    Objectives: To investigate the effect of the introduction of environmental noise, microphone identity & position, and informational noise on the accuracy of a commercial Clinical AI Scribe (CAIS). Design: Consultations on five medical conditions (memory loss, diarrhoea, headaches, skin rash, or prostate symptoms) were recorded. A ground-truth audio file was produced for a noise free consultation. This in turn produced a ground-truth clinical summary via the AI scribe, typically containing 30 facts. Different types of ‘noise’ at different levels were then investigated by (i) increasing the distance of the consultation from the microphone, and changing microphone type (ii) introducing background noise: heavy rain, baby crying, construction noise and a toddler at a range of levels between -10 to +5dB relative to the consultation audio and (iii) introducing informational noise: the medical consultations were modified to include additional conversations between the clinician and the patient to introduce irrelevant discussions (medical and non-medical, shorter and longer). The resulting 160 summaries were then compared with the corresponding ground-truth summaries and the errors introduced were classified and enumerated. Setting: Simulated primary care setting. Participants: The roles of the doctor and patient were played by actors, rotating between roles. Main outcome measures: Outcomes measured were the quantification of the degradation of accuracy of the clinical summary, i.e. omissions, hallucinations and inclusions, relative to the ground-truth summary, following the introduction of noise of all types at different levels. Results: Error rates and absolute error counts increased for all noise levels and noise types; the principal error being omissions. At 4.5 m from the consultation, for all microphones tested, all facts were omitted. At up to 2 m all microphones, except the laptop, only introduced a small number (<5) of omissions. For background noise, the type of noise was found to correlate with omissions – toddler and heavy rain were particularly deleterious, 7 and 15 omissions at 0 and +5dB respectively, i.e. when the noise loudness was comparable to or greater than that of the consultation. The CAIS was remarkably apt at rejecting additional informational noise. This was true for the addition of either short or long, medical or non-medical, information. Conclusions: The CAIS was very good at capturing a doctor-patient consultation and producing accurate clinical summaries, however error rate (especially omissions) increased notably when acoustic noise was introduced. The system was better able to handle informational noise. Summary Box: What is already known on this topic: The use of AI is rapidly increasing, with many commercial products available, both specific to clinical use and not. Commercial Clinical AI Scribes (generating clinical summaries from consultation audio) are being used in healthcare settings, sometimes with little guidance or knowledge of employers. These tools can operate well in perfect conditions, but there is limited data on nosier environments. What this study adds: The commercial Clinical AI Scribe only introduced a small number of errors when introduced to informational noise, however large numbers of errors (especially omissions) were caused by acoustic noise (dependant on the particular sound type and volume), especially with low-quality poorly-located microphones. Users of Clinical AI Scribes must be particularly vigilant in checking the summary when the environment contains certain deleterious background noises, even at lower volumes, to ensure errors are not introduced to the patient records.


  • Open Access

    pH Oscillating System for Molecular Computation as a Chemical Turing Machine

    by Thomas C. Draper, Eszter Poros-Tarcali, Juan Pérez-Mercader

    It has previously been demonstrated that native chemical Turing machines can be constructed by exploiting the nonlinear dynamics of the homogeneous oscillating Belousov–Zhabotinsky reaction. These Turing machines can perform word recognition of a Chomsky type 1 context sensitive language (CSL), demonstrating their high computing power. Here, we report on a chemical Turing machine that has been developed using the H2O2–H2SO4–SO32––CO32– pH oscillating system. pH oscillators are different to bromate oscillators in two key ways: the proton is the autocatalytic agent, and at least one of the reductants is always fully consumed in each turnover─meaning the system has to be operated as a flow reactor. Through careful design, we establish a system that can also perform Chomsky type 1 CSL word recognition and demonstrate its power through the testing of a series of in-language and out-of-language words.


  • Open Access

    Exploring the symbol processing ‘time interval’ parametric constraint in a Belousov–Zhabotinsky operated chemical Turing machine

    by Thomas C. Draper, Marta Dueñas-Díez, Juan Pérez-Mercader

    Chemical reactions are powerful molecular recognition machines. This power has been recently harnessed to build actual instances of each class of experimentally realizable computing automata, using exclusively small-molecule chemistry (i.e. without requiring biomolecules). The most powerful of them, a programmable Turing machine, uses the Belousov–Zhabotinsky oscillatory chemistry, and accepts/rejects input sequences through a dual oscillatory and thermodynamic output signature. The time interval between the aliquots representing each letter of the input is the parameter that determines the time it takes to run the computation. Here, we investigate this critical performance parameter, and its effect not only on the computation speed, but also on the robustness of the accept/reject oscillatory and thermodynamic criteria. Our work demonstrates that the time interval is a non-trivial design parameter, whose choice should be made with great care. The guidelines we provide can be used in the optimization of the speed, robustness, and energy efficiency of chemical automata computations.


  • Contactless sensing of liquid marbles for detection, characterisation & computing

    by Thomas C. Draper, Neil Phillips, Roshan Weerasekera, Richard Mayne, Claire Fullarton, Ben P. J. de Lacy Costello, Andrew Adamatzky

    New liquid marble sensor exploits impedance to detect presence of liquid marbles, whilst also providing characterisation information.


  • Open Access

    Neuromorphic Liquid Marbles with Aqueous Carbon Nanotube Cores

    by Richard Mayne, Thomas C. Draper, Neil Phillips, James G. H. Whiting, Roshan Weerasekera, Claire Fullarton, Ben P. J. de Lacy Costello, Andrew Adamatzky

    Neuromorphic computing devices attempt to emulate features of biological nervous systems through mimicking the properties of synapses, towards implementing the emergent properties of their counterparts, such as learning. Inspired by recent advances in the utilisation of liquid marbles (microlitre quantities of fluid coated in hydrophobic powder) for the creation of unconventional computing devices, we describe the development of liquid marbles with neuromorphic properties through the use of copper coatings and \SI{1.0}{\milli\gram\per\milli\litre} carbon nanotube-containing fluid cores. Experimentation was performed through sandwiching the marbles between two cup-style electrodes and stimulating them with repeated DC pulses at \SI{3.0}{\volt}. Our results demonstrate that `entrainment' of a carbon nanotube filled-copper liquid marble via periodic pulses can cause their electrical resistance to rapidly switch between high to low resistance profiles, upon inverting the polarity of stimulation: the reduction in resistance between high and low profiles was approximately 88\% after two rounds of entrainment. This effect was found to be reversible through reversion to the original stimulus polarity and was strengthened by repeated experimentation, as evidenced by a mean reduction in time to switching onset of 43\%. These effects were not replicated in nanotube solutions not bound inside liquid marbles. Our electrical characterisation also reveals that nanotube-filled liquid marbles exhibit pinched loop hysteresis IV profiles consistent with the description of memristors. We conclude by discussing the applications of this technology to the development of unconventional computing devices and the study of emergent characteristics in biological neural tissue.