Physically Meaningful Surrogate Data for COPD.

The rapidly increasing prevalence of debilitating breathing disorders, such as chronic obstructive pulmonary disease (COPD), calls for a meaningful integration of artificial intelligence (AI) into respiratory healthcare. Deep learning techniques are "data hungry" whilst patient-based data is invariably expensive and time consuming to record. To this end, we introduce a novel COPD-simulator, a physical apparatus with an easy to replicate design which enables rapid and effective generation of a wide range of COPD-like data from healthy subjects, for enhanced training of deep learning frameworks. To ensure the faithfulness of our domain-aware COPD surrogates, the generated waveforms are examined through both flow waveforms and photoplethysmography (PPG) waveforms (as a proxy for intrathoracic pressure) in terms of duty cycle, sample entropy, FEV1/FVC ratios and flow-volume loops. The proposed simulator operates on healthy subjects and is able to generate FEV1/FVC obstruction ratios ranging from greater than 0.8 to less than 0.2, mirroring values that can observed in the full spectrum of real-world COPD. As a final stage of verification, a simple convolutional neural network is trained on surrogate data alone, and is used to accurately detect COPD in real-world patients. When training solely on surrogate data, and testing on real-world data, a comparison of true positive rate against false positive rate yields an area under the curve of 0.75, compared with 0.63 when training solely on real-world data.

Identifying benefits and concerns with using digital health services during COVID-19: Evidence from a hospital-based patient survey.

Despite large-scale adoption during COVID-19, patient perceptions on the benefits and potential risks with receiving care through digital technologies have remained largely unexplored. A quantitative content analysis of responses to a questionnaire (N = 6766) conducted at a multi-site acute trust in London (UK), was adopted to identify commonly reported benefits and concerns. Patients reported a range of promising benefits beyond immediate usage during COVID-19, including ease of access; support for disease and care management; improved timeliness of access and treatment; and better prioritisation of healthcare resources. However, in addition to known risks such as data security and inequity in access, our findings also illuminate some less studied concerns, including perceptions of compromised safety; negative impacts on patient-clinician relationships; and difficulties in interpreting health information provided through electronic health records and mHealth apps. Implications for future research and practice are discussed.

Smartphone-Based Remote Monitoring in Heart Failure With Reduced Ejection Fraction: Retrospective Cohort Study of Secondary Care Use and Costs.

BACKGROUND: Despite effective therapies, the economic burden of heart failure with reduced ejection fraction (HFrEF) is driven by frequent hospitalizations. Treatment optimization and admission avoidance rely on frequent symptom reviews and monitoring of vital signs. Remote monitoring (RM) aims to prevent admissions by facilitating early intervention, but the impact of noninvasive, smartphone-based RM of vital signs on secondary health care use and costs in the months after a new diagnosis of HFrEF is unknown. OBJECTIVE: The purpose of this study is to conduct a secondary care health use and health-economic evaluation for patients with HFrEF using smartphone-based noninvasive RM and compare it with matched controls receiving usual care without RM. METHODS: We conducted a retrospective study of 2 cohorts of newly diagnosed HFrEF patients, matched 1:1 for demographics, socioeconomic status, comorbidities, and HFrEF severity. They are (1) the RM group, with patients using the RM platform for >3 months and (2) the control group, with patients referred before RM was available who received usual heart failure care without RM. Emergency department (ED) attendance, hospital admissions, outpatient use, and the associated costs of this secondary care activity were extracted from the Discover data set for a 3-month period after diagnosis. Platform costs were added for the RM group. Secondary health care use and costs were analyzed using Kaplan-Meier event analysis and Cox proportional hazards modeling. RESULTS: A total of 146 patients (mean age 63 years; 42/146, 29% female) were included (73 in each group). The groups were well-matched for all baseline characteristics except hypertension (P=.03). RM was associated with a lower hazard of ED attendance (hazard ratio [HR] 0.43; P=.02) and unplanned admissions (HR 0.26; P=.02). There were no differences in elective admissions (HR 1.03, P=.96) or outpatient use (HR 1.40; P=.18) between the 2 groups. These differences were sustained by a univariate model controlling for hypertension. Over a 3-month period, secondary health care costs were approximately 4-fold lower in the RM group than the control group, despite the additional cost of RM itself (mean cost per patient GBP £465, US $581 vs GBP £1850, US $2313, respectively; P=.04). CONCLUSIONS: This retrospective cohort study shows that smartphone-based RM of vital signs is feasible for HFrEF. This type of RM was associated with an approximately 2-fold reduction in ED attendance and a 4-fold reduction in emergency admissions over just 3 months after a new diagnosis with HFrEF. Costs were significantly lower in the RM group without increasing outpatient demand. This type of RM could be adjunctive to standard care to reduce admissions, enabling other resources to help patients unable to use RM.

Survival and health economic outcomes in heart failure diagnosed at hospital admission versus community settings: a propensity-matched analysis.

BACKGROUND AND AIMS: Most patients with heart failure (HF) are diagnosed following a hospital admission. The clinical and health economic impacts of index HF diagnosis made on admission to hospital versus community settings are not known. METHODS: We used the North West London Discover database to examine 34 208 patients receiving an index diagnosis of HF between January 2015 and December 2020. A propensity score-matched (PSM) cohort was identified to adjust for differences in socioeconomic status, cardiovascular risk and pre-diagnosis health resource utilisation cost. Outcomes were stratified by two pathways to index HF diagnosis: a 'hospital pathway' was defined by diagnosis following hospital admission; and a 'community pathway' by diagnosis via a general practitioner or outpatient services. The primary clinical and health economic endpoints were all-cause mortality and cost-consequence differential, respectively. RESULTS: The diagnosis of HF was via hospital pathway in 68% (23 273) of patients. The PSM cohort included 17 174 patients (8582 per group) and was matched across all selected confounders (p>0.05). The ratio of deaths per person-months at 24 months comparing community versus hospital diagnosis was 0.780 (95% CI 0.722 to 0.841, p<0.0001). By 72 months, the ratio of deaths was 0.960 (0.905 to 1.020, p=0.18). Diagnosis via hospital pathway incurred an overall extra longitudinal cost of £2485 per patient. CONCLUSIONS: Index diagnosis of HF through hospital admission continues to dominate and is associated with a significantly greater short-term risk of mortality and substantially increased long-term costs than if first diagnosed in the community. This study highlights the potential for community diagnosis-early, before symptoms necessitate hospitalisation-to improve both clinical and health economic outcomes.

Wearable In-Ear PPG: Detailed Respiratory Variations Enable Classification of COPD.

An ability to extract detailed spirometry-like breathing waveforms from wearable sensors promises to greatly improve respiratory health monitoring. Photoplethysmography (PPG) has been researched in depth for estimation of respiration rate, given that it varies with respiration through overall intensity, pulse amplitude and pulse interval. We compare and contrast the extraction of these three respiratory modes from both the ear canal and finger and show a marked improvement in the respiratory power for respiration induced intensity variations and pulse amplitude variations when recording from the ear canal. We next employ a data driven multi-scale method, noise assisted multivariate empirical mode decomposition (NA-MEMD), which allows for simultaneous analysis of all three respiratory modes to extract detailed respiratory waveforms from in-ear PPG. For rigour, we considered in-ear PPG recordings from healthy subjects, both older and young, patients with chronic obstructive pulmonary disease (COPD) and idiopathic pulmonary fibrosis (IPF) and healthy subjects with artificially obstructed breathing. Specific in-ear PPG waveform changes are observed for COPD, such as a decreased inspiratory duty cycle and an increased inspiratory magnitude, when compared with expiratory magnitude. These differences are used to classify COPD from healthy and IPF waveforms with a sensitivity of 87% and an overall accuracy of 92%. Our findings indicate the promise of in-ear PPG for COPD screening and unobtrusive respiratory monitoring in ambulatory scenarios and in consumer wearables.

Point-of-care screening for heart failure with reduced ejection fraction using artificial intelligence during ECG-enabled stethoscope examination in London, UK: a prospective, observational, multicentre study.

BACKGROUND: Most patients who have heart failure with a reduced ejection fraction, when left ventricular ejection fraction (LVEF) is 40% or lower, are diagnosed in hospital. This is despite previous presentations to primary care with symptoms. We aimed to test an artificial intelligence (AI) algorithm applied to a single-lead ECG, recorded during ECG-enabled stethoscope examination, to validate a potential point-of-care screening tool for LVEF of 40% or lower. METHODS: We conducted an observational, prospective, multicentre study of a convolutional neural network (known as AI-ECG) that was previously validated for the detection of reduced LVEF using 12-lead ECG as input. We used AI-ECG retrained to interpret single-lead ECG input alone. Patients (aged ≥18 years) attending for transthoracic echocardiogram in London (UK) were recruited. All participants had 15 s of supine, single-lead ECG recorded at the four standard anatomical positions for cardiac auscultation, plus one handheld position, using an ECG-enabled stethoscope. Transthoracic echocardiogram-derived percentage LVEF was used as ground truth. The primary outcome was performance of AI-ECG at classifying reduced LVEF (LVEF ≤40%), measured using metrics including the area under the receiver operating characteristic curve (AUROC), sensitivity, and specificity, with two-sided 95% CIs. The primary outcome was reported for each position individually and with an optimal combination of AI-ECG outputs (interval range 0-1) from two positions using a rule-based approach and several classification models. This study is registered with ClinicalTrials.gov, NCT04601415. FINDINGS: Between Feb 6 and May 27, 2021, we recruited 1050 patients (mean age 62 years [SD 17·4], 535 [51%] male, 432 [41%] non-White). 945 (90%) had an ejection fraction of at least 40%, and 105 (10%) had an ejection fraction of 40% or lower. Across all positions, ECGs were most frequently of adequate quality for AI-ECG interpretation at the pulmonary position (979 [93·3%] of 1050). Quality was lowest for the aortic position (846 [80·6%]). AI-ECG performed best at the pulmonary valve position (p=0·02), with an AUROC of 0·85 (95% CI 0·81-0·89), sensitivity of 84·8% (76·2-91·3), and specificity of 69·5% (66·4-72·6). Diagnostic odds ratios did not differ by age, sex, or non-White ethnicity. Taking the optimal combination of two positions (pulmonary and handheld positions), the rule-based approach resulted in an AUROC of 0·85 (0·81-0·89), sensitivity of 82·7% (72·7-90·2), and specificity of 79·9% (77·0-82·6). Using AI-ECG outputs from these two positions, a weighted logistic regression with l2 regularisation resulted in an AUROC of 0·91 (0·88-0·95), sensitivity of 91·9% (78·1-98·3), and specificity of 80·2% (75·5-84·3). INTERPRETATION: A deep learning system applied to single-lead ECGs acquired during a routine examination with an ECG-enabled stethoscope can detect LVEF of 40% or lower. These findings highlight the potential for inexpensive, non-invasive, workflow-adapted, point-of-care screening, for earlier diagnosis and prognostically beneficial treatment. FUNDING: NHS Accelerated Access Collaborative, NHSX, and the National Institute for Health Research.

Determinants of Shielding Behavior During the COVID-19 Pandemic and Associations With Well-being Among National Health Service Patients: Longitudinal Observational Study.

BACKGROUND: The UK National Health Service (NHS) classified 2.2 million people as clinically extremely vulnerable (CEV) during the first wave of the 2020 COVID-19 pandemic, advising them to "shield" (to not leave home for any reason). OBJECTIVE: The aim of this study was to measure the determinants of shielding behavior and associations with well-being in a large NHS patient population for informing future health policy. METHODS: Patients contributing to an ongoing longitudinal participatory epidemiology study (Longitudinal Effects on Wellbeing of the COVID-19 Pandemic [LoC-19], n=42,924) received weekly email invitations to complete questionnaires (17-week shielding period starting April 9, 2020) within their NHS personal electronic health record. Question items focused on well-being. Participants were stratified into four groups by self-reported CEV status (qualifying condition) and adoption of shielding behavior (baselined at week 1 or 2). The distribution of CEV criteria was reported alongside situational variables and univariable and multivariable logistic regression. Longitudinal trends in physical and mental well-being were displayed graphically. Free-text responses reporting variables impacting well-being were semiquantified using natural language processing. In the lead up to a second national lockdown (October 23, 2020), a follow-up questionnaire evaluated subjective concern if further shielding was advised. RESULTS: The study included 7240 participants. In the CEV group (n=2391), 1133 (47.3%) assumed shielding behavior at baseline, compared with 633 (13.0%) in the non-CEV group (n=4849). CEV participants who shielded were more likely to be Asian (odds ratio [OR] 2.02, 95% CI 1.49-2.76), female (OR 1.24, 95% CI 1.05-1.45), older (OR per year increase 1.01, 95% CI 1.00-1.02), living in a home with an outdoor space (OR 1.34, 95% CI 1.06-1.70) or three to four other inhabitants (three: OR 1.49, 95% CI 1.15-1.94; four: OR 1.49, 95% CI 1.10-2.01), or solid organ transplant recipients (OR 2.85, 95% CI 2.18-3.77), or have severe chronic lung disease (OR 1.63, 95% CI 1.30-2.04). Receipt of a government letter advising shielding was reported in 1115 (46.6%) CEV participants and 180 (3.7%) non-CEV participants, and was associated with adopting shielding behavior (OR 3.34, 95% CI 2.82-3.95 and OR 2.88, 95% CI 2.04-3.99, respectively). In CEV participants, shielding at baseline was associated with a lower rating of mental well-being and physical well-being. Similar results were found for non-CEV participants. Concern for well-being if future shielding was required was most prevalent among CEV participants who had originally shielded. CONCLUSIONS: Future health policy must balance the potential protection from COVID-19 against our findings that shielding negatively impacted well-being and was adopted in many in whom it was not indicated and variably in whom it was indicated. This therefore also requires clearer public health messaging and support for well-being if shielding is to be advised in future pandemic scenarios.

The Impact of the COVID-19 Pandemic on the Uptake of Influenza Vaccine: UK-Wide Observational Study.

BACKGROUND: In the face of the COVID-19 pandemic, the UK National Health Service (NHS) extended eligibility for influenza vaccination this season to approximately 32.4 million people (48.8% of the population). Knowing the intended uptake of the vaccine will inform supply and public health messaging to maximize vaccination. OBJECTIVE: The objective of this study was to measure the impact of the COVID-19 pandemic on the acceptance of influenza vaccination in the 2020-2021 season, specifically focusing on people who were previously eligible but routinely declined vaccination and newly eligible people. METHODS: Intention to receive the influenza vaccine in 2020-2021 was asked of all registrants of the largest electronic personal health record in the NHS by a web-based questionnaire on July 31, 2020. Of those who were either newly or previously eligible but had not previously received an influenza vaccination, multivariable logistic regression and network diagrams were used to examine their reasons to undergo or decline vaccination. RESULTS: Among 6641 respondents, 945 (14.2%) were previously eligible but were not vaccinated; of these, 536 (56.7%) intended to receive an influenza vaccination in 2020-2021, as did 466 (68.6%) of the newly eligible respondents. Intention to receive the influenza vaccine was associated with increased age, index of multiple deprivation quintile, and considering oneself to be at high risk from COVID-19. Among those who were eligible but not intending to be vaccinated in 2020-2021, 164/543 (30.2%) gave reasons based on misinformation. Of the previously unvaccinated health care workers, 47/96 (49%) stated they would decline vaccination in 2020-2021. CONCLUSIONS: In this sample, COVID-19 has increased acceptance of influenza vaccination in previously eligible but unvaccinated people and has motivated substantial uptake in newly eligible people. This study is essential for informing resource planning and the need for effective messaging campaigns to address negative misconceptions, which is also necessary for COVID-19 vaccination programs.

Belief of having had unconfirmed Covid-19 infection reduces willingness to participate in app-based contact tracing.

Contact tracing and lockdown are health policies being used worldwide to combat the coronavirus (COVID-19). The UK National Health Service (NHS) Track and Trace Service has plans for a nationwide app that notifies the need for self-isolation to those in contact with a person testing positive for COVID-19. To be successful, such an app will require high uptake, the determinants and willingness for which are unclear but essential to understand for effective public health benefit. The objective of this study was to measure the determinants of willingness to participate in an NHS app-based contact-tracing programme using a questionnaire within the Care Information Exchange (CIE)-the largest patient-facing electronic health record in the NHS. Among 47,708 registered NHS users of the CIE, 27% completed a questionnaire asking about willingness to participate in app-based contact tracing, understanding of government advice, mental and physical wellbeing and their healthcare utilisation-related or not to COVID-19. Descriptive statistics are reported alongside univariate and multivariable logistic regression models, with positive or negative responses to a question on app-based contact tracing as the dependent variable. 26.1% of all CIE participants were included in the analysis (N = 12,434, 43.0% male, mean age 55.2). 60.3% of respondents were willing to participate in app-based contact tracing. Out of those who responded 'no', 67.2% stated that this was due to privacy concerns. In univariate analysis, worsening mood, fear and anxiety in relation to changes in government rules around lockdown were associated with lower willingness to participate. Multivariable analysis showed that difficulty understanding government rules was associated with a decreased inclination to download the app, with those scoring 1-2 and 3-4 in their understanding of the new government rules being 45% and 27% less inclined to download the contact-tracing app, respectively; when compared to those who rated their understanding as 5-6/10 (OR for 1-2/10 = 0.57 [CI 0.48-0.67]; OR for 3-4/10 = 0.744 [CI 0.64-0.87]), whereas scores of 7-8 and 9-10 showed a 43% and 31% respective increase. Those reporting an unconfirmed belief of having previously had and recovered from COVID-19 were 27% less likely to be willing to download the app; belief of previous recovery from COVID-19 infection OR 0.727 [0.585-0.908]). In this large UK-wide questionnaire of wellbeing in lockdown, a willingness for app-based contact tracing over an appropriate age range is 60%-close to the estimated 56% population uptake, and substantially less than the smartphone-user uptake considered necessary for an app-based contact tracing to be an effective intervention to help suppress an epidemic. Difficulty comprehending government advice and uncertainty of diagnosis, based on a public health policy of not testing to confirm self-reported COVID-19 infection during lockdown, therefore reduce willingness to adopt a government contact-tracing app to a level below the threshold for effectiveness as a tool to suppress an epidemic.

Artificial Intelligence, Data Sensors and Interconnectivity: Future Opportunities for Heart Failure.

A higher proportion of patients with heart failure have benefitted from a wide and expanding variety of sensor-enabled implantable devices than any other patient group. These patients can now also take advantage of the ever-increasing availability and affordability of consumer electronics. Wearable, on- and near-body sensor technologies, much like implantable devices, generate massive amounts of data. The connectivity of all these devices has created opportunities for pooling data from multiple sensors - so-called interconnectivity - and for artificial intelligence to provide new diagnostic, triage, risk-stratification and disease management insights for the delivery of better, more personalised and cost-effective healthcare. Artificial intelligence is also bringing important and previously inaccessible insights from our conventional cardiac investigations. The aim of this article is to review the convergence of artificial intelligence, sensor technologies and interconnectivity and the way in which this combination is set to change the care of patients with heart failure.

A Framework for Predicting Impactability of Digital Care Management Using Machine Learning Methods.

Digital care management programs can reduce health care costs and improve quality of care. However, it is unclear how to target patients who are most likely to benefit from these programs ex ante, a shortcoming of current "risk score"-based approaches across many interventions. This study explores a framework to define impactability by using machine learning (ML) models to identify those patients most likely to benefit from a digital health intervention for care management. Anonymized insurance claims data were used from a commercially insured population across several US states and combined with inferred sociodemographic data. The approach involves creating 2 models and the comparative analysis of the methodologies and performances therein. The authors first train a cost prediction model to calculate the differences in predicted (without intervention) versus actual (with onboarding onto digital health platform) health care expenditures for patients (N = 5600). This enables classification impactability if differences in predicted versus actual costs meet a predetermined threshold. Several random forest and logistic regression machine learning models were then trained to accurately categorize new patients as impactable versus not impactable. These parameters are modified through grid search to define the parameters that deliver optimal performance, reaching an overall sensitivity of 0.77 and specificity of 0.65 among all models. This approach shows that impactability for a digital health intervention can be successfully defined using ML methods, thus enabling efficient allocation of resources. This framework is generalizable to analyzing impactability of any intervention and can contribute to realizing closed-loop feedback systems for continuous improvement in health care.

Machine learning can accurately predict pre-admission baseline hemoglobin and creatinine in intensive care patients.

Patients admitted to the intensive care unit frequently have anemia and impaired renal function, but often lack historical blood results to contextualize the acuteness of these findings. Using data available within two hours of ICU admission, we developed machine learning models that accurately (AUC 0.86-0.89) classify an individual patient's baseline hemoglobin and creatinine levels. Compared to assuming the baseline to be the same as the admission lab value, machine learning performed significantly better at classifying acute kidney injury regardless of initial creatinine value, and significantly better at predicting baseline hemoglobin value in patients with admission hemoglobin of <10 g/dl.

Cell death markers in patients with cirrhosis and acute decompensation.

The aims of this study were to determine the role of cell death in patients with cirrhosis and acute decompensation (AD) and acute on chronic liver failure (ACLF) using plasma-based biomarkers. The patients studied were part of the CANONIC (CLIF Acute-on-Chronic Liver Failure in Cirrhosis) study (N = 337; AD, 258; ACLF, 79); additional cohorts included healthy volunteers, stable patients with cirrhosis, and a group of 16 AD patients for histological studies. Caspase-cleaved keratin 18 (cK18) and keratin 18 (K18), which reflect apoptotic and total cell death, respectively, and cK18:K18 ratio (apoptotic index) were measured in plasma by enzyme-linked immunosorbent assay. The concentrations of cK18 and K18 increased and the cK18:K18 ratio decreased with increasing severity of AD and ACLF (P < 0.001, respectively). Alcohol etiology, no previous decompensation, and alcohol abuse were associated with increased cell death markers whereas underlying infection was not. Close correlation was observed between the cell death markers and, markers of systemic inflammation, hepatic failure, alanine aminotransferase, and bilirubin, but not with markers of extrahepatic organ injury. Terminal deoxynucleotidyl transferase dUTP nick-end labeling staining confirmed evidence of greater hepatic cell death in patients with ACLF as opposed to AD. Inclusion of cK18 and K18 improved the performance of the CLIF-C AD score in prediction of progression from AD to ACLF (P < 0.05). CONCLUSION: Cell death, likely hepatic, is an important feature of AD and ACLF and its magnitude correlates with clinical severity. Nonapoptotic forms of cell death predominate with increasing severity of AD and ACLF. The data suggests that ACLF is a heterogeneous entity and shows that the importance of cell death in its pathophysiology is dependent on predisposing factors, precipitating illness, response to injury, and type of organ failure. (Hepatology 2018;67:989-1002).