Defining acceptable data collection and reuse standards for queer artificial intelligence research in mental health: protocol for the online PARQAIR-MH Delphi study.

INTRODUCTION: For artificial intelligence (AI) to help improve mental healthcare, the design of data-driven technologies needs to be fair, safe, and inclusive. Participatory design can play a critical role in empowering marginalised communities to take an active role in constructing research agendas and outputs. Given the unmet needs of the LGBTQI+ (Lesbian, Gay, Bisexual, Transgender, Queer and Intersex) community in mental healthcare, there is a pressing need for participatory research to include a range of diverse queer perspectives on issues of data collection and use (in routine clinical care as well as for research) as well as AI design. Here we propose a protocol for a Delphi consensus process for the development of PARticipatory Queer AI Research for Mental Health (PARQAIR-MH) practices, aimed at informing digital health practices and policy. METHODS AND ANALYSIS: The development of PARQAIR-MH is comprised of four stages. In stage 1, a review of recent literature and fact-finding consultation with stakeholder organisations will be conducted to define a terms-of-reference for stage 2, the Delphi process. Our Delphi process consists of three rounds, where the first two rounds will iterate and identify items to be included in the final Delphi survey for consensus ratings. Stage 3 consists of consensus meetings to review and aggregate the Delphi survey responses, leading to stage 4 where we will produce a reusable toolkit to facilitate participatory development of future bespoke LGBTQI+-adapted data collection, harmonisation, and use for data-driven AI applications specifically in mental healthcare settings. ETHICS AND DISSEMINATION: PARQAIR-MH aims to deliver a toolkit that will help to ensure that the specific needs of LGBTQI+ communities are accounted for in mental health applications of data-driven technologies. The study is expected to run from June 2024 through January 2025, with the final outputs delivered in mid-2025. Participants in the Delphi process will be recruited by snowball and opportunistic sampling via professional networks and social media (but not by direct approach to healthcare service users, patients, specific clinical services, or via clinicians' caseloads). Participants will not be required to share personal narratives and experiences of healthcare or treatment for any condition. Before agreeing to participate, people will be given information about the issues considered to be in-scope for the Delphi (eg, developing best practices and methods for collecting and harmonising sensitive characteristics data; developing guidelines for data use/reuse) alongside specific risks of unintended harm from participating that can be reasonably anticipated. Outputs will be made available in open-access peer-reviewed publications, blogs, social media, and on a dedicated project website for future reuse.

Clinical Prompt Learning With Frozen Language Models.

When the first transformer-based language models were published in the late 2010s, pretraining with general text and then fine-tuning the model on a task-specific dataset often achieved the state-of-the-art performance. However, more recent work suggests that for some tasks, directly prompting the pretrained model matches or surpasses fine-tuning in performance with few or no model parameter updates required. The use of prompts with language models for natural language processing (NLP) tasks is known as prompt learning. We investigated the viability of prompt learning on clinically meaningful decision tasks and directly compared this with more traditional fine-tuning methods. Results show that prompt learning methods were able to match or surpass the performance of traditional fine-tuning with up to 1000 times fewer trainable parameters, less training time, less training data, and lower computation resource requirements. We argue that these characteristics make prompt learning a very desirable alternative to traditional fine-tuning for clinical tasks, where the computational resources of public health providers are limited, and where data can often not be made available or not be used for fine-tuning due to patient privacy concerns. The complementary code to reproduce the experiments presented in this work can be found at https://github.com/NtaylorOX/Public_Clinical_Prompt.

Artificial intelligence for dementia-Applied models and digital health.

INTRODUCTION: The use of applied modeling in dementia risk prediction, diagnosis, and prognostics will have substantial public health benefits, particularly as "deep phenotyping" cohorts with multi-omics health data become available. METHODS: This narrative review synthesizes understanding of applied models and digital health technologies, in terms of dementia risk prediction, diagnostic discrimination, prognosis, and progression. Machine learning approaches show evidence of improved predictive power compared to standard clinical risk scores in predicting dementia, and the potential to decompose large numbers of variables into relatively few critical predictors. RESULTS: This review focuses on key areas of emerging promise including: emphasis on easier, more transparent data sharing and cohort access; integration of high-throughput biomarker and electronic health record data into modeling; and progressing beyond the primary prediction of dementia to secondary outcomes, for example, treatment response and physical health. DISCUSSION: Such approaches will benefit also from improvements in remote data measurement, whether cognitive (e.g., online), or naturalistic (e.g., watch-based accelerometry).

Development of the Escalation of Therapy or Intervention (ETI) Calculator for Patients with Ulcerative Colitis Using ePROMs.

BACKGROUND AND AIMS: Digital collection of patient-reported outcome measures [PROMs] is largely unexplored as a basis for follow-up for patients with ulcerative colitis [UC]. Our aim was to develop a model to predict the likelihood of escalation of therapy or intervention at an outpatient appointment that may be used to rationalize follow-up. METHODS: TrueColours-IBD is a web-based, real-time, remote monitoring software that allows longitudinal collection of ePROMs. Data for prediction modelling were derived from a Development Cohort, guided by the TRIPOD statement. Logistic regression modelling used ten candidate items to predict escalation of therapy or intervention. An Escalation of Therapy or Intervention [ETI] calculator was developed, and applied in a Validation Cohort at the same centre. RESULTS: The Development Cohort [n = 66] was recruited in 2016 and followed for 6 months [208 appointments]. From ten items, four significant predictors of ETI were identified: SCCAI, IBD Control-8, faecal calprotectin, and platelets. For practicality, a model with only SCCAI and IBD Control-8, both entered remotely by the patient, without the need for faecal calprotectin or blood tests was selected. Between 2018 and 2020, a Validation Cohort of 538 patients [1188 appointments] was examined. A 5% threshold on the ETI calculator correctly identified 343/388 [88%] escalations and 274/484 [57%] non-escalations. CONCLUSIONS: A calculator based on digital, patient-entered data on symptoms and quality of life can predict whether a patient with UC requires escalation of therapy or intervention at an outpatient appointment. This may be used to streamline outpatient appointments for patients with UC.

Harnessing the potential of machine learning and artificial intelligence for dementia research.

Progress in dementia research has been limited, with substantial gaps in our knowledge of targets for prevention, mechanisms for disease progression, and disease-modifying treatments. The growing availability of multimodal data sets opens possibilities for the application of machine learning and artificial intelligence (AI) to help answer key questions in the field. We provide an overview of the state of the science, highlighting current challenges and opportunities for utilisation of AI approaches to move the field forward in the areas of genetics, experimental medicine, drug discovery and trials optimisation, imaging, and prevention. Machine learning methods can enhance results of genetic studies, help determine biological effects and facilitate the identification of drug targets based on genetic and transcriptomic information. The use of unsupervised learning for understanding disease mechanisms for drug discovery is promising, while analysis of multimodal data sets to characterise and quantify disease severity and subtype are also beginning to contribute to optimisation of clinical trial recruitment. Data-driven experimental medicine is needed to analyse data across modalities and develop novel algorithms to translate insights from animal models to human disease biology. AI methods in neuroimaging outperform traditional approaches for diagnostic classification, and although challenges around validation and translation remain, there is optimism for their meaningful integration to clinical practice in the near future. AI-based models can also clarify our understanding of the causality and commonality of dementia risk factors, informing and improving risk prediction models along with the development of preventative interventions. The complexity and heterogeneity of dementia requires an alternative approach beyond traditional design and analytical approaches. Although not yet widely used in dementia research, machine learning and AI have the potential to unlock current challenges and advance precision dementia medicine.

Explainable artificial intelligence for mental health through transparency and interpretability for understandability.

The literature on artificial intelligence (AI) or machine learning (ML) in mental health and psychiatry lacks consensus on what "explainability" means. In the more general XAI (eXplainable AI) literature, there has been some convergence on explainability meaning model-agnostic techniques that augment a complex model (with internal mechanics intractable for human understanding) with a simpler model argued to deliver results that humans can comprehend. Given the differing usage and intended meaning of the term "explainability" in AI and ML, we propose instead to approximate model/algorithm explainability by understandability defined as a function of transparency and interpretability. These concepts are easier to articulate, to "ground" in our understanding of how algorithms and models operate and are used more consistently in the literature. We describe the TIFU (Transparency and Interpretability For Understandability) framework and examine how this applies to the landscape of AI/ML in mental health research. We argue that the need for understandablity is heightened in psychiatry because data describing the syndromes, outcomes, disorders and signs/symptoms possess probabilistic relationships to each other-as do the tentative aetiologies and multifactorial social- and psychological-determinants of disorders. If we develop and deploy AI/ML models, ensuring human understandability of the inputs, processes and outputs of these models is essential to develop trustworthy systems fit for deployment.

Patient-reported Outcomes: the ICHOM Standard Set for Inflammatory Bowel Disease in Real-life Practice Helps Quantify Deficits in Current Care.

BACKGROUND: Patient-reported outcome measures [PROMs] are key to documenting outcomes that matter most to patients and are increasingly important to commissioners of health care seeking value. We report the first series of the ICHOM Standard Set for Inflammatory Bowel Disease [IBD]. METHODS: Patients treated for ulcerative colitis [UC] or Crohn's disease [CD] in our centre were offered enrolment into the web-based TrueColours-IBD programme. Through this programme, e-mail prompts linking to validated questionnaires were sent for symptoms, quality of life, and ICHOM IBD outcomes. RESULTS: The first 1299 consecutive patients enrolled [779 UC, 520 CD] were studied with median 270 days of follow-up (interquartile range [IQR] 116, 504). 671 [52%] were female, mean age 42 years (standard deviation [sd] 16), mean body mass index [BMI] 26 [sd 5.3]. At registration, 483 [37%] were using advanced therapies. Median adherence to fortnightly quality of life reporting and quarterly outcomes was 100% [IQR 48, 100%] and 100% [IQR 75, 100%], respectively. In the previous 12 months, prednisolone use was reported by 229 [29%] patients with UC vs 81 [16%] with CD, p <0.001; 202 [16%] for <3 months; and 108 [8%] for >3 months. An IBD-related intervention was reported by 174 [13%] patients, and 80 [6%] reported an unplanned hospital admission. There were high rates of fatigue [50%] and mood disturbance [23%]. CONCLUSIONS: Outcomes reported by patients illustrate the scale of the therapeutic deficit in current care. Proof of principle is demonstrated that PROM data can be collected continuously with little burden on health care professionals. This may become a metric for quality improvement programmes or to compare outcomes.

Personalised treatment for cognitive impairment in dementia: development and validation of an artificial intelligence model.

BACKGROUND: Donepezil, galantamine, rivastigmine and memantine are potentially effective interventions for cognitive impairment in dementia, but the use of these drugs has not been personalised to individual patients yet. We examined whether artificial intelligence-based recommendations can identify the best treatment using routinely collected patient-level information. METHODS: Six thousand eight hundred four patients aged 59-102 years with a diagnosis of dementia from two National Health Service (NHS) Foundation Trusts in the UK were used for model training/internal validation and external validation, respectively. A personalised prescription model based on the Recurrent Neural Network machine learning architecture was developed to predict the Mini-Mental State Examination (MMSE) and Montreal Cognitive Assessment (MoCA) scores post-drug initiation. The drug that resulted in the smallest decline in cognitive scores between prescription and the next visit was selected as the treatment of choice. Change of cognitive scores up to 2 years after treatment initiation was compared for model evaluation. RESULTS: Overall, 1343 patients with MMSE scores were identified for internal validation and 285 [21.22%] took the drug recommended. After 2 years, the reduction of mean [standard deviation] MMSE score in this group was significantly smaller than the remaining 1058 [78.78%] patients (0.60 [0.26] vs 2.80 [0.28]; P = 0.02). In the external validation cohort (N = 1772), 222 [12.53%] patients took the drug recommended and reported a smaller MMSE reduction compared to the 1550 [87.47%] patients who did not (1.01 [0.49] vs 4.23 [0.60]; P = 0.01). A similar performance gap was seen when testing the model on patients prescribed with AChEIs only. CONCLUSIONS: It was possible to identify the most effective drug for the real-world treatment of cognitive impairment in dementia at an individual patient level. Routine care patients whose prescribed medications were the best fit according to the model had better cognitive performance after 2 years.

Validation of UK Biobank data for mental health outcomes: A pilot study using secondary care electronic health records.

UK Biobank (UKB) is widely employed to investigate mental health disorders and related exposures; however, its applicability and relevance in a clinical setting and the assumptions required have not been sufficiently and systematically investigated. Here, we present the first validation study using secondary care mental health data with linkage to UKB from Oxford - Clinical Record Interactive Search (CRIS) focusing on comparison of demographic information, diagnostic outcome, medication record and cognitive test results, with missing data and the implied bias from both resources depicted. We applied a natural language processing model to extract information embedded in unstructured text from clinical notes and attachments. Using a contingency table we compared the demographic information recorded in UKB and CRIS. We calculated the positive predictive value (PPV, proportion of true positives cases detected) for mental health diagnosis and relevant medication. Amongst the cohort of 854 subjects, PPVs for any mental health diagnosis for dementia, depression, bipolar disorder and schizophrenia were 41.6%, and were 59.5%, 12.5%, 50.0% and 52.6%, respectively. Self-reported medication records in UKB had general PPV of 47.0%, with the prevalence of frequently prescribed medicines to each typical mental health disorder considerably different from the information provided by CRIS. UKB is highly multimodal, but with limited follow-up records, whereas CRIS offers a longitudinal high-resolution clinical picture with more than ten years of observations. The linkage of both datasets will reduce the self-report bias and synergistically augment diverse modalities into a unified resource to facilitate more robust research in mental health.

Associations Between Brain Volumes and Cognitive Tests with Hypertensive Burden in UK Biobank.

BACKGROUND: Mid-life hypertension is an established risk factor for cognitive impairment and dementia and related to greater brain atrophy and poorer cognitive performance. Previous studies often have small sample sizes from older populations that lack utilizing multiple measures to define hypertension such as blood pressure, self-report information, and medication use; furthermore, the impact of the duration of hypertension is less extensively studied. OBJECTIVE: To investigate the relationship between hypertension defined using multiple measures and length of hypertension with brain measure and cognition. METHODS: Using participants from the UK Biobank MRI visit with blood pressure measurements (n = 31,513), we examined the cross-sectional relationships between hypertension and duration of hypertension with brain volumes and cognitive tests using generalized linear models adjusted for confounding. RESULTS: Compared with normotensives, hypertensive participants had smaller brain volumes, larger white matter hyperintensities (WMH), and poorer performance on cognitive tests. For total brain, total grey, and hippocampal volumes, those with greatest duration of hypertension had the smallest brain volumes and the largest WMH, ventricular cerebrospinal fluid volumes. For other subcortical and white matter microstructural regions, there was no clear relationship. There were no significant associations between duration of hypertension and cognitive tests. CONCLUSION: Our results show hypertension is associated with poorer brain and cognitive health however, the impact of duration since diagnosis warrants further investigation. This work adds further insights by using multiple measures defining hypertension and analysis on duration of hypertension which is a substantial advance on prior analyses-particularly those in UK Biobank which present otherwise similar analyses on smaller subsets.

Effect of trazodone on cognitive decline in people with dementia: Cohort study using UK routinely collected data.

OBJECTIVES: Evidence in mouse models has found that the antidepressant trazodone may be protective against neurodegeneration. We therefore aimed to compare cognitive decline of people with dementia taking trazodone with those taking other antidepressants. METHODS: Three identical naturalistic cohort studies using UK clinical registers. We included all people with dementia assessed during 2008-16 who were recorded taking trazodone, citalopram or mirtazapine for at least 6 weeks. Linear mixed models examined age, time and sex-adjusted Mini-mental state examination (MMSE) change in people with all-cause dementia taking trazodone compared with those taking citalopram and mirtazapine. In secondary analyses, we examined those with non-vascular dementia; mild dementia; and adjusted results for neuropsychiatric symptoms. We combined results from the three study sites using random-effects meta-analysis. RESULTS: We included 2,199 people with dementia, including 406 taking trazodone, with mean 2.2 years follow-up. There was no difference in adjusted cognitive decline in people with all-cause or non-vascular dementia taking trazodone, citalopram or mirtazapine in any of the three study sites. When data from the three sites were combined in meta-analysis, we found greater mean MMSE decline in people with all-cause dementia taking trazodone compared to those taking citalopram (0·26 points per successive MMSE measurement, 95% CI 0·03-0·49; p = 0·03). Results in sensitivity analyses were consistent with primary analyses. CONCLUSIONS: There was no evidence of cognitive benefit from trazodone compared to other antidepressants in people with dementia in three naturalistic cohort studies. Despite preclinical evidence, trazodone should not be advocated for cognition in dementia.

Med7: A transferable clinical natural language processing model for electronic health records.

Electronic health record systems are ubiquitous and the majority of patients' data are now being collected electronically in the form of free text. Deep learning has significantly advanced the field of natural language processing and the self-supervised representation learning and the transfer learning have become the methods of choice in particular when the high quality annotated data are limited. Identification of medical concepts and information extraction is a challenging task, yet important ingredient for parsing unstructured data into structured and tabulated format for downstream analytical tasks. In this work we introduced a named-entity recognition (NER) model for clinical natural language processing. The model is trained to recognise seven categories: drug names, route of administration, frequency, dosage, strength, form, duration. The model was first pre-trained on the task of predicting the next word, using a collection of 2 million free-text patients' records from MIMIC-III corpora followed by fine-tuning on the named-entity recognition task. The model achieved a micro-averaged F1 score of 0.957 across all seven categories. Additionally, we evaluated the transferability of the developed model using the data from the Intensive Care Unit in the US to secondary care mental health records (CRIS) in the UK. A direct application of the trained NER model to CRIS data resulted in reduced performance of F1 = 0.762, however after fine-tuning on a small sample from CRIS, the model achieved a reasonable performance of F1 = 0.944. This demonstrated that despite a close similarity between the data sets and the NER tasks, it is essential to fine-tune the target domain data in order to achieve more accurate results. The resulting model and the pre-trained embeddings are available at https://github.com/kormilitzin/med7.

Using item response theory (IRT) to improve the efficiency of the Simple Clinical Colitis Activity Index (SCCAI) for patients with ulcerative colitis.

BACKGROUND: The SCCAI was designed to facilitate assessment of disease activity in ulcerative colitis (UC). We aimed to interrogate the metric properties of individual items of the SCCAI using item response theory (IRT) analysis, to simplify and improve its performance. METHODS: The original 9-item SCCAI was collected through TrueColours, a real-time software platform which allows remote entry and monitoring of patients with UC. Data were securely uploaded onto Dementias Platform UK Data Portal, where they were analysed in Stata 16.1 SE. A 2-parameter (2-PL) logistic IRT model was estimated to evaluate each item of the SCCAI for its informativeness (discrimination). A revised scale was generated and re-assessed following systematic removal of items. RESULTS: SCCAI data for 516 UC patients (41 years, SD = 15) treated in Oxford were examined. After initial item deletion (Erythema nodosum, Pyoderma gangrenosum), a 7-item scale was estimated. Discrimination values (information) ranged from 0.41 to 2.52 indicating selected item inefficiency with three items < 1.70 which is a suggested discriminatory value for optimal efficiency. Systematic item deletions found that a 4-item scale (bowel frequency day; bowel frequency nocturnal; urgency to defaecation; rectal bleeding) was more informative and discriminatory of trait severity (discrimination values of 1.50 to 2.78). The 4-item scale possesses higher scalability and unidimensionality, suggesting that the responses to items are either direct endorsement (patient selection by symptom) or non-endorsement of the trait (disease activity). CONCLUSION: Reduction of the SCCAI from the original 9-item scale to a 4-item scale provides optimum trait information that will minimise response burden. This new 4-item scale needs validation against other measures of disease activity such as faecal calprotectin, endoscopy and histopathology.

Real-world effectiveness, its predictors and onset of action of cholinesterase inhibitors and memantine in dementia: retrospective health record study.

BACKGROUND: The efficacy of acetylcholinesterase inhibitors and memantine in the symptomatic treatment of Alzheimer's disease is well-established. Randomised trials have shown them to be associated with a reduction in the rate of cognitive decline. AIMS: To investigate the real-world effectiveness of acetylcholinesterase inhibitors and memantine for dementia-causing diseases in the largest UK observational secondary care service data-set to date. METHOD: We extracted mentions of relevant medications and cognitive testing (Mini-Mental State Examination (MMSE) and Montreal Cognitive Assessment (MoCA) scores) from de-identified patient records from two National Health Service (NHS) trusts. The 10-year changes in cognitive performance were modelled using a combination of generalised additive and linear mixed-effects modelling. RESULTS: The initial decline in MMSE and MoCA scores occurs approximately 2 years before medication is initiated. Medication prescription stabilises cognitive performance for the ensuing 2-5 months. The effect is boosted in more cognitively impaired cases at the point of medication prescription and attenuated in those taking antipsychotics. Importantly, patients who are switched between agents at least once do not experience any beneficial cognitive effect from pharmacological treatment. CONCLUSIONS: This study presents one of the largest real-world examination of the efficacy of acetylcholinesterase inhibitors and memantine for symptomatic treatment of dementia. We found evidence that 68% of individuals respond to treatment with a period of cognitive stabilisation before continuing their decline at the pre-treatment rate.

Utilization of the Signature Method to Identify the Early Onset of Sepsis From Multivariate Physiological Time Series in Critical Care Monitoring.

OBJECTIVES: Patients in an ICU are particularly vulnerable to sepsis. It is therefore important to detect its onset as early as possible. This study focuses on the development and validation of a new signature-based regression model, augmented with a particular choice of the handcrafted features, to identify a patient's risk of sepsis based on physiologic data streams. The model makes a positive or negative prediction of sepsis for every time interval since admission to the ICU. DESIGN: The data were sourced from the PhysioNet/Computing in Cardiology Challenge 2019 on the "Early Prediction of Sepsis from Clinical Data." It consisted of ICU patient data from three separate hospital systems. Algorithms were scored against a specially designed utility function that rewards early predictions in the most clinically relevant region around sepsis onset and penalizes late predictions and false positives. SETTING: The work was completed as part of the PhysioNet 2019 Challenge alongside 104 other teams. PATIENTS: PhysioNet sourced over 60,000 ICU patients with up to 40 clinical variables for each hour of a patient's ICU stay. The Sepsis-3 criteria was used to define the onset of sepsis. INTERVENTIONS: None. MEASUREMENTS AND MAIN RESULTS: The algorithm yielded a utility function score which was the first placed entry in the official phase of the challenge.

Identifying Predictors of Suicide in Severe Mental Illness: A Feasibility Study of a Clinical Prediction Rule (Oxford Mental Illness and Suicide Tool or OxMIS).

BACKGROUND: Oxford Mental Illness and Suicide tool (OxMIS) is a brief, scalable, freely available, structured risk assessment tool to assess suicide risk in patients with severe mental illness (schizophrenia-spectrum disorders or bipolar disorder). OxMIS requires further external validation, but a lack of large-scale cohorts with relevant variables makes this challenging. Electronic health records provide possible data sources for external validation of risk prediction tools. However, they contain large amounts of information within free-text that is not readily extractable. In this study, we examined the feasibility of identifying suicide predictors needed to validate OxMIS in routinely collected electronic health records. METHODS: In study 1, we manually reviewed electronic health records of 57 patients with severe mental illness to calculate OxMIS risk scores. In study 2, we examined the feasibility of using natural language processing to scale up this process. We used anonymized free-text documents from the Clinical Record Interactive Search database to train a named entity recognition model, a machine learning technique which recognizes concepts in free-text. The model identified eight concepts relevant for suicide risk assessment: medication (antidepressant/antipsychotic treatment), violence, education, self-harm, benefits receipt, drug/alcohol use disorder, suicide, and psychiatric admission. We assessed model performance in terms of precision (similar to positive predictive value), recall (similar to sensitivity) and F1 statistic (an overall performance measure). RESULTS: In study 1, we estimated suicide risk for all patients using the OxMIS calculator, giving a range of 12 month risk estimates from 0.1-3.4%. For 13 out of 17 predictors, there was no missing information in electronic health records. For the remaining 4 predictors missingness ranged from 7-26%; to account for these missing variables, it was possible for OxMIS to estimate suicide risk using a range of scores. In study 2, the named entity recognition model had an overall precision of 0.77, recall of 0.90 and F1 score of 0.83. The concept with the best precision and recall was medication (precision 0.84, recall 0.96) and the weakest were suicide (precision 0.37), and drug/alcohol use disorder (recall 0.61). CONCLUSIONS: It is feasible to estimate suicide risk with the OxMIS tool using predictors identified in routine clinical records. Predictors could be extracted using natural language processing. However, electronic health records differ from other data sources, particularly for family history variables, which creates methodological challenges.

Natural language processing for structuring clinical text data on depression using UK-CRIS.

BACKGROUND: Utilisation of routinely collected electronic health records from secondary care offers unprecedented possibilities for medical science research but can also present difficulties. One key issue is that medical information is presented as free-form text and, therefore, requires time commitment from clinicians to manually extract salient information. Natural language processing (NLP) methods can be used to automatically extract clinically relevant information. OBJECTIVE: Our aim is to use natural language processing (NLP) to capture real-world data on individuals with depression from the Clinical Record Interactive Search (CRIS) clinical text to foster the use of electronic healthcare data in mental health research. METHODS: We used a combination of methods to extract salient information from electronic health records. First, clinical experts define the information of interest and subsequently build the training and testing corpora for statistical models. Second, we built and fine-tuned the statistical models using active learning procedures. FINDINGS: Results show a high degree of accuracy in the extraction of drug-related information. Contrastingly, a much lower degree of accuracy is demonstrated in relation to auxiliary variables. In combination with state-of-the-art active learning paradigms, the performance of the model increases considerably. CONCLUSIONS: This study illustrates the feasibility of using the natural language processing models and proposes a research pipeline to be used for accurately extracting information from electronic health records. CLINICAL IMPLICATIONS: Real-world, individual patient data are an invaluable source of information, which can be used to better personalise treatment.

Validation of Plasma Proteomic Biomarkers Relating to Brain Amyloid Burden in the EMIF-Alzheimer's Disease Multimodal Biomarker Discovery Cohort.

We have previously investigated, discovered, and replicated plasma protein biomarkers for use to triage potential trials participants for PET or cerebrospinal fluid measures of Alzheimer's disease (AD) pathology. This study sought to undertake validation of these candidate plasma biomarkers in a large, multi-center sample collection. Targeted plasma analyses of 34 proteins with prior evidence for prediction of in vivo pathology were conducted in up to 1,000 samples from cognitively healthy elderly individuals, people with mild cognitive impairment, and in patients with AD-type dementia, selected from the EMIF-AD catalogue. Proteins were measured using Luminex xMAP, ELISA, and Meso Scale Discovery assays. Seven proteins replicated in their ability to predict in vivo amyloid pathology. These proteins form a biomarker panel that, along with age, could significantly discriminate between individuals with high and low amyloid pathology with an area under the curve of 0.74. The performance of this biomarker panel remained consistent when tested in apolipoprotein E ɛ4 non-carrier individuals only. This blood-based panel is biologically relevant, measurable using practical immunocapture arrays, and could significantly reduce the cost incurred to clinical trials through screen failure.

Named entity recognition in electronic health records using transfer learning bootstrapped Neural Networks.

Neural networks (NNs) have become the state of the art in many machine learning applications, such as image, sound (LeCun et al., 2015) and natural language processing (Young et al., 2017; Linggard et al., 2012). However, the success of NNs remains dependent on the availability of large labelled datasets, such as in the case of electronic health records (EHRs). With scarce data, NNs are unlikely to be able to extract this hidden information with practical accuracy. In this study, we develop an approach that solves these problems for named entity recognition, obtaining 94.6 F1 score in I2B2 2009 Medical Extraction Challenge (Uzuner et al., 2010), 4.3 above the architecture that won the competition. To achieve this, we bootstrap our NN models through transfer learning by pretraining word embeddings on a secondary task performed on a large pool of unannotated EHRs and using the output embeddings as a foundation of a range of NN architectures. Beyond the official I2B2 challenge, we further achieve 82.4 F1 on extracting relationships between medical terms using attention-based seq2seq models bootstrapped in the same manner.