Explore the vast range of titles available.

Anxiety and depression are closely associated. However, they are typically treated separately and there is a dearth of information on tackling them together. The study's purpose was to establish how best to treat co-occurring anxiety and depression in a routine clinical service-specifically, to compare cognitive behaviour therapy (CBT) focusing only on depression (CBT-D) to a broader CBT focusing on both depression and anxiety (CBT-DA). Case notes of 69 patients with equally severe clinical levels of depression and anxiety seen in a routine clinical service were randomly selected to review from a pool of 990 patients. The mean age was 44.61 years (SD = 12.97). 65% of the sample were female and 88% reported their ethnicity white. The content of electronic records reporting techniques used and scores on a measure of depression (The Patient Health Questionnaire) and anxiety (The Generalized Anxiety Disorder Assessment) were reviewed to categorise therapy as CBT-D or CBT-DA. Results indicated significant overall improvement with CBT; 70% and 77% of the sample met criteria for reliable improvement on The Patient Health Questionnaire and The Generalized Anxiety Disorder Assessment respectively. Fewer patients who received CBT-DA met The Generalized Anxiety Disorder Assessment recovery criteria at the end of treatment than those who received CBT-D. Mean post treatment PHQ-9 and GAD-7 scores remained above threshold for those receiving CBT_DA but not those receiving CBT-D. There was no evidence suggesting CBT-DA was superior to CBT-D. In patients with equally severe clinical levels of depression and anxiety, a broader treatment addressing both anxiety and depression does not appear to be associated with improved outcomes compared to treatment focused on depression.

BackgroundDespite the increasing availability of electronic healthcare record (EHR) data and wide availability of plug-and-play machine learning (ML) Application Programming Interfaces, the adoption of data-driven decision-making within routine hospital workflows thus far, has remained limited. Through the lens of deriving clusters of diagnoses by age, this study investigated the type of ML analysis that can be performed using EHR data and how results could be communicated to lay stakeholders.MethodsObservational EHR data from a tertiary paediatric hospital, containing 61 522 unique patients and 3315 unique ICD-10 diagnosis codes was used, after preprocessing. K-means clustering was applied to identify age distributions of patient diagnoses. The final model was selected using quantitative metrics and expert assessment of the clinical validity of the clusters. Additionally, uncertainty over preprocessing decisions was analysed.FindingsFour age clusters of diseases were identified, broadly aligning to ages between: 0 and 1; 1 and 5; 5 and 13; 13 and 18. Diagnoses, within the clusters, aligned to existing knowledge regarding the propensity of presentation at different ages, and sequential clusters presented known disease progressions. The results validated similar methodologies within the literature. The impact of uncertainty induced by preprocessing decisions was large at the individual diagnoses but not at a population level. Strategies for mitigating, or communicating, this uncertainty were successfully demonstrated.ConclusionUnsupervised ML applied to EHR data identifies clinically relevant age distributions of diagnoses which can augment existing decision making. However, biases within healthcare datasets dramatically impact results if not appropriately mitigated or communicated.

It can be overwhelming and frightening to deal with health issues for children and young people. They are often intimidated by hospitals and information given by clinicians that is beyond their comprehension. The fear of unknowing often associates hospitals with anxiety and uncertainty in children and young people. The purpose of this study is to identify needs, so that it is possible to address concerns and reduce anxiety in children and young people to make their healthcare journey well informed.We conducted scenario-based interviews and co-design sessions with a group of children (n=9) aged between 12-17 who were members of GOSH Young Persons’ Advisory Group (YPAG). The sessions focused on gathering thoughts and emotions they feel when visiting a hospital and identifying their information needs.Our findings revealed that the entire group associated negative emotions when given a scenario about diagnosis and treatment of a long-term health condition. Many participants admitted they’ll feel ‘nervous, stressed, confused and discouraged’ before going for a treatment or procedure, while some expressed concerns about its effect on their mental and physical health and how painful it would be. They also probed about the short term and long-term prognosis of the condition, if they can refuse treatment and how long will they have to miss school. When asked about what will help them prepare better for their hospital visits, 86% of the participants voted for data visualizations of their condition and risk levels and 71% expressed interest in knowing details about the procedure and an ideal recovery timeline. 57% of the participants agreed that knowing the environment of the hospital and the people they’ll be meeting beforehand will help in reducing anxiety.These findings support the use of digital information systems for children undergoing medical treatment to reduce pre-operative anxiety and negative emotions.

In adult practice there is recognition that average patient complexity is increasing, with a greater proportion of patients having multiple diagnoses or comorbidities. This study aims to examine whether there has been a change in number of recorded coexisting diagnoses per patient over a 24-year period for children attending as in-patients to a specialist children's hospital in England. Following all in-patient admissions, patient episodes are allocated specific diagnosis codes (ICD-10) by a specialist clinical coding team according to standard NHS criteria and guidance. We examine the number of coexisting diagnoses allocated per patient admission over a 24-year period. From a total of 278,579 overnight in-patient admissions during the study period (2000-2023) there were 1,023,276 ICD-10 patient diagnoses. The mean number of diagnoses per admission increased from 2.72 to 10.43 over the period (Kendall's tau statistic of 0.93; p-value < 0.001), an increase of 284% (95% confidence interval 275% - 293%). Over recent decades, the recorded complexity of patients attending a specialist children's hospital appear to have increased significantly, with an almost 3-fold increase in the number of coexisting diagnoses present per admission. The cause of this finding cannot be determined from the data; however, it appears to be gradual and consistent, and across all speciality areas suggesting biological or referral factors rather than artefactual coding issues. Recognition of such a trend is important when interpreting retrospective data for AI, research, and planning purposes.

IntroductionDigital health data forms the foundation for most research and many patient treatment decisions. Whether developing a new clinical decision-making tool or carrying out data-driven research, it is essential that the data itself can be trusted and relied upon.Laboratory tests are a common data source, which generate results based on patient samples, with minimal user input. Point of Care Tests (POCT) rely more heavily on user-entered values, increasing the potential for introducing human error, systematic biases, and approximation. We analysed a statistical method to detect tests that show the most non-random variation, indicative that the data may be less credible and require further investigation.MethodsLaboratory data for 2020-2021 at the Great Ormond Street Hospital were extracted from the electronic patient record system. These data were explored with the programming language R using the secure Aridhia Digital Research Environment (DRE). Using runs analysis on a per-test per-day basis, the Anhøj criteria were used to quantitatively generate a ‘suspicion index’. The average suspicion index per test over the year was used for comparison between tests. A high suspicion index indicated that the values observed were less likely due to random variation and therefore potentially less credible for further analyses.ResultsBased on 4,749,278 test results across 2,371 distinct test types, when ranked in order of credibility, the 23 least credible tests were all POCT. The suspicion indices ranged from 0.8 to 15.6. There was a clear separation of POCT from the laboratory tests, with the latter being generally more credible.ImpactLow credibility tests could adversely affect clinical tools designed to utilise them. By calculating the credibility of tests, low scoring tests could be identified and targeted for improvement or excluded from, or modified for, clinical tool development.

IntroductionMedications for ‘Anaemia and other Blood Disorders’ (BNF 9.1) are commonly administered to patients diagnosed with Chronic Kidney Disease (CKD) to help control associated conditions. In this study, we used routinely collected data of Great Ormond Street Hospital (GOSH) patients to describe the medication administration characteristics while gaining a better understanding of this rich data set, and how it can be leveraged to provide an informative summary to clinicians and researchers as part of a clinical informatics consultation tool.MethodsWe performed a retrospective study of GOSH patients diagnosed with CKD between 2005 and 2021 that were administered medications for anaemia and other blood disorders. We conducted descriptive analyses using R scripts. Data sets were extracted using processes developed by the GOSH Digital Research Environment (DRE) group. Analyses focused on the number of patients receiving each medication and how long after diagnosis a medication was administered. We also looked for patterns in the order that medications were administered, based on when each medication was first administered.ResultsOur data set consisted of 14 different medications and Electronic Health Records (EHRs) of 795 CKD patients. Darbepoetin alfa and sodium feredetate had the highest number of patients (n=361), followed by epoetin beta (n=286). Darbepoetin alfa was administered on average 10 days after a stage 5 diagnosis (n=255). The median (Q1-Q3) number of medications per patient was 2 (1-3).ConclusionsWe identified patterns in the order these medications were administered. Processes and tools developed by the DRE enabled us to efficiently acquire rich patient data. The analysis presented in this study could be implemented as a component of a clinical informatics tool that supports automated and reproducible data-driven medical research.

BackgroundPatients at Great Ormond Street Hospital (GOSH) are often diagnosed with multiple complex diseases and medical conditions. The implementation of the Epic Electronic Health Record (EHR) has made detailed coded diagnosis data available to clinicians and researchers. To support clinical care and research, we have developed a general framework for analysing comorbidities in GOSH patients based on this data. We envisage that this framework can be applied to data from multiple clinical specialties and can be implemented as a component of a clinical informatics consultation tool providing us with a better understanding of diagnoses.MethodsWe present a retrospective study of GOSH patients seen at the Nephrology specialty between 2015 and 2021 as a use case for the framework developed. In this study, we examined patterns of medical condition diagnosis over time and visualised the time that a diagnosis occurs after an ‘index event’ while stratifying by demographic attributes. We define an ‘index event’ as a particular diagnosis, procedure, or some other event of interest. All analyses and visualisations were carried out using R. Data sets were extracted using processes developed by the GOSH Digital Research Environment (DRE) group.Results and DiscussionWe analysed EHRs of 1,314 patients seen at the Nephrology specialty at GOSH. We identified patterns of diagnosis including common comorbidities and their timing. Processes and tools developed by DRE enabled us to efficiently acquire detailed diagnosis data and work towards the development of a set of methods that can be embedded to a clinical informatics tool to automatically generate a comorbidity analysis output for any clinical specialty.

AimsWe performed an analysis of GOSH’s Electronic Patient Record (EPR) data derived insights relating to patients diagnosed with an ICD-10 anxiety code. This analysis examined trends in frequency distribution and duration of anxiety diagnoses over time, alongside the medication administrations and procedures performed on these patients.MethodsRoutine data for all patients diagnosed with an ICD-10 anxiety code (F41-) from July 2019 to March 2022 were extracted, de-identified, and analysed in the secure GOSH Digital Research Environment (DRE). The Python package Pandas was used to clean and analyse data. Interactive visualisations were created using Plotly. Medication drug classes for these patients (Hypnotics, Anxiolytics, and Antidepressants), and OPCS-4-classification-identified-procedures were analysed.ResultsAcross 1573 patients in the cohort, ‘Anxiety disorder, unspecified’ (F419) was the most common anxiety type until 2021, gradually being replaced by ‘Other specified anxiety disorders’ (F418). The monthly sum of anxiety diagnoses demonstrated a seasonal variation, peaking in July 2019 and July 2021, with a trough spanning UK COVID-19 lockdowns. Chronic Kidney Disease, Autism, and other developmental disorders were the most prevalent comorbidities. Stratifying by drug class, Hypnotics & Anxiolytic administrations were more popular than Antidepressants for patients diagnosed with anxiety. Melatonin was the most administered medication. The OPCS-4 Class ‘U’ (‘Diagnostic imaging, testing and rehabilitation’) was the most common group of procedures performed on the same day as an anxiety diagnosis. Transthoracic Echocardiography and CSF Injection were most prevalent.ConclusionThis analysis of EPR data found a seasonal variation in anxiety diagnosis frequency, with a gradual change in the specific type. Hypnotics & Anxiolytics are more popular than Antidepressants. Anxiety diagnoses relate mostly to imaging and testing procedures. Further work includes the same analyses and trends on earlier data.

An ‘Informatics Consultation’ is an approach to supporting clinical decision making by providing on-demand evidence based on data from other patients. While high quality trials and academic publications currently provide the best evidence for clinical practice, these trials are inherently generalised, often exclude complex comorbidities and subsequently do not necessarily translate well to the realities of patients with even slightly atypical characteristics or multimorbidities.We present PICTURE, a web-application and data science platform developed at the Great Ormond Street Hospital (GOSH). PICTURE provides informatics consultation support to the hospital, based on the real-world evidence from routinely acquired data stored in the electronic health record. The application allows clinical teams to select characteristics of their individual patient to define tailored cohorts of historical patients upon which evidence-producing analyses, and models are run. For example, the application can generate a personalised prognosis report for a specific patient (in both interactive and static formats) by comparing historical patients with similar demographics and comorbidities.The PICTURE proof-of-concept application has been built in the R programming language with a Shiny based user interface and is hosted on the Aridhia Digital Research Environment. The tool makes use of the existing data ‘extract, transform, load’ pipeline and data-model in the GOSH ‘Data Research, Innovation, and Virtual Environments’ (DRIVE) unit to allow analyses across a full range of patient data, including, diagnoses, procedures, laboratory tests, medications, and admissions.The current application has demonstrated how simple analytics methods can be configured to run across user-defined patient cohorts to provide informative prognostic real-world evidence. Future work will develop the PICTURE tool, in collaboration with clinicians from a range of specialities, to provide more effective and detailed evidence reports. Additionally, the tool will be expanded to incorporate methods from machine learning and allow data linkage with external data sources.

Great Ormond Street Hospital NHS Trust (GOSH) has gathered a range of electronic health record data on diagnoses, interventions, and outcomes for its patients. These resources represent a database of experience that can be used to develop innovative data-based tools that could complement more traditional information sources for clinicians to provide care. We explored how a data processing and analysis pipeline could be created that provides information to clinicians about how the current diagnoses of a patient affect probable future procedures. In particular, which procedures are most likely to be required, and when.Based on dummy data (i.e., not from real patients) of diagnoses and procedures generated by the GOSH Digital Research Environment (DRE), a machine learning model was developed. The model was based on an array of logistic regression models, each trained to predict the probability of one procedure. The weights learned by the models were used to identify which diagnoses made a specific procedure more likely. The pipeline also used survival analysis methods to predict and present the time scale in which a procedure might be required. This information is displayed on plots of how the risk of a procedure changes over time after a diagnosis. The modelling and analyses were carried out in Python on the Aridhia secure DRE.The model was trained on 4000 dummy patients and validated on 1000 dummy patients, achieving an average AUC score of 0.72 in the validation set.This study shows how routinely acquired electronic health record data can produce personalised predictions of which procedures a patient is likely to require, and when, to support clinical decision making, resource planning and patient education. These tools were developed using dummy data that respect real patient data properties, which showcases the potential of dummy data techniques to innovate while keeping high privacy standards.