Diagnosis and Severity Assessment of COPD Using a Novel Fast-Response Capnometer and Interpretable Machine Learning.

INTRODUCTION: Spirometry is the gold standard for COPD diagnosis and severity determination, but is technique-dependent, nonspecific, and requires administration by a trained healthcare professional. There is a need for a fast, reliable, and precise alternative diagnostic test. This study's aim was to use interpretable machine learning to diagnose COPD and assess severity using 75-second carbon dioxide (CO2) breath records captured with TidalSense's N-TidalTM capnometer. METHOD: For COPD diagnosis, machine learning algorithms were trained and evaluated on 294 COPD (including GOLD stages 1-4) and 705 non-COPD participants. A logistic regression model was also trained to distinguish GOLD 1 from GOLD 4 COPD with the output probability used as an index of severity. RESULTS: The best diagnostic model achieved an AUROC of 0.890, sensitivity of 0.771, specificity of 0.850 and positive predictive value (PPV) of 0.834. Evaluating performance on all test capnograms that were confidently ruled in or out yielded PPV of 0.930 and NPV of 0.890. The severity determination model yielded an AUROC of 0.980, sensitivity of 0.958, specificity of 0.961 and PPV of 0.958 in distinguishing GOLD 1 from GOLD 4. Output probabilities from the severity determination model produced a correlation of 0.71 with percentage predicted FEV1. CONCLUSION: The N-TidalTM device could be used alongside interpretable machine learning as an accurate, point-of-care diagnostic test for COPD, particularly in primary care as a rapid rule-in or rule-out test. N-TidalTM also could be effective in monitoring disease progression, providing a possible alternative to spirometry for disease monitoring.

A descriptive cohort study of withdrawal from inhaled corticosteroids in COPD patients.

Inhaled corticosteroid (ICS) therapy is widely prescribed without a history of exacerbations and consensus guidelines suggest withdrawal of ICS in these patients would reduce the risk of side effects and promote cost-effective prescribing. The study describes the prescribing behaviour in the United Kingdom (UK) in relation to ICS withdrawal and identifies clinical outcomes following withdrawal using primary and secondary care electronic health records between January 2012 and December 2017. Patients with a history ≥12 months' exposure who withdrew ICS for ≥6 months were identified into two cohorts; those prescribed a long-acting bronchodilator maintenance therapy and those that were not prescribed any maintenance therapy. The duration of withdrawal, predictors of restarting ICS, and clinical outcomes were compared between both patient cohorts. Among 76,808 patients that had ≥1 prescription of ICS in the study period, 11,093 patients (14%) withdrew ICS therapy at least once during the study period. The median time without ICS was 9 months (IQR 7-14), with the majority (71%) receiving subsequent ICS prescriptions after withdrawal. Patients receiving maintenance therapy with a COPD review at withdrawal were 28% less likely to restart ICS (HR: 0.72, 95% CI 0.61, 0.85). Overall, 69% and 89% of patients that withdrew ICS had no recorded exacerbation event or COPD hospitalisation, respectively, during the withdrawal. This study provides evidence that most patients withdrawing from ICS do not experience COPD exacerbations and withdrawal success can be achieved by carefully planning routine COPD reviews whilst optimising the use of available maintenance therapies.

Blood eosinophils to guide inhaled maintenance therapy in a primary care COPD population.

Blood eosinophils are a potentially useful biomarker for guiding inhaled corticosteroid (ICS) treatment decisions in COPD. We investigated whether existing blood eosinophil counts predict benefit from initiation of ICS compared to bronchodilator therapy. We used routinely collected data from UK primary care in the Clinical Practice Research Datalink. Participants were aged ≥40 years with COPD, were ICS-naïve and starting a new inhaled maintenance medication (intervention group: ICS; comparator group: long-acting bronchodilator, non-ICS). Primary outcome was time to first exacerbation, compared between ICS and non-ICS groups, stratified by blood eosinophils ("high" ≥150 cells·µL-1 and "low" <150 cells·µL-1). Out of 9475 eligible patients, 53.9% initiated ICS and 46.1% non-ICS treatment with no difference in eosinophils between treatment groups (p=0.71). Exacerbation risk was higher in patients prescribed ICS than those prescribed non-ICS treatment, but with a lower risk in those with high eosinophils (hazard ratio (HR) 1.04, 95% CI 0.98-1.10) than low eosinophils (HR 1.19, 95% CI 1.09-1.31) (p-value for interaction 0.01). Risk of pneumonia hospitalisation with ICS was greatest in those with low eosinophils (HR 1.26, 95% CI 1.05-1.50; p-value for interaction 0.04). Results were similar whether the most recent blood eosinophil count or the mean of blood eosinophil counts was used. In a primary care population, the most recent blood eosinophil count could be used to guide initiation of ICS in COPD patients. We suggest that ICS should be considered in those with higher eosinophils and avoided in those with lower eosinophils (<150 cells·µL-1).

Is stratification testing for treatment of chronic obstructive pulmonary disease exacerbations cost-effective in primary care? an early cost-utility analysis.

OBJECTIVES: Patients with chronic obstructive pulmonary disease (COPD) who experience acute exacerbations usually require treatment with oral steroids or antibiotics, depending on the etiology of the exacerbation. Current management is based on clinician's assessment and judgement, which lacks diagnostic accuracy and results in overtreatment. A test to guide these decisions in primary care is in development. We developed an early decision model to evaluate the cost-effectiveness of this treatment stratification test in the primary care setting in the United Kingdom. METHODS: A combined decision tree and Markov model was developed of COPD progression and the exacerbation care pathway. Sensitivity analysis was carried out to guide technology development and inform evidence generation requirements. RESULTS: The base case test strategy cost GBP 423 (USD 542) less and resulted in a health gain of 0.15 quality-adjusted life-years per patient compared with not testing. Testing reduced antibiotic prescriptions by 30 percent, potentially lowering the risk of antimicrobial resistance developing. In sensitivity analysis, the result depended on the clinical effects of treating patients according to the test result, as opposed to treating according to clinical judgement alone, for which there is limited evidence. The results were less sensitive to the accuracy of the test. CONCLUSIONS: Testing may be cost-saving in primary care, but this requires robust evidence on whether test-guided treatment is effective. High quality evidence on the clinical utility of testing is required for early modeling of diagnostic tests generally.

Clinical Characteristics of Pertussis-Associated Cough in Adults and Children: A Diagnostic Systematic Review and Meta-Analysis.

BACKGROUND: Pertussis (whooping cough) is a highly infective cause of cough that causes significant morbidity and mortality. Existing case definitions include paroxysmal cough, whooping, and posttussive vomiting, but diagnosis can be difficult. We determined the diagnostic accuracy of clinical characteristics of pertussis-associated cough. METHODS: We systematically searched CINAHL, Embase, Medline, and SCI-EXPANDED/CPCI-S up to June 2016. Eligible studies compared clinical characteristics in those positive and negative for Bordetella pertussis infection, confirmed by laboratory investigations. Two authors independently completed screening, data extraction, and quality and bias assessments. For each characteristic, RevMan was used to produce descriptive forest plots. The bivariate meta-analysis method was used to generate pooled estimates of sensitivity and specificity. RESULTS: Of 1,969 identified papers, 53 were included. Forty-one clinical characteristics were assessed for diagnostic accuracy. In adult patients, paroxysmal cough and absence of fever have a high sensitivity (93.2% [CI, 83.2-97.4] and 81.8% [CI, 72.2-88.7], respectively) and low specificity (20.6% [CI, 14.7-28.1] and 18.8% [CI, 8.1-37.9]), whereas posttussive vomiting and whooping have low sensitivity (32.5% [CI, 24.5-41.6] and 29.8% [CI, 8.0-45.2]) and high specificity (77.7% [CI, 73.1-81.7] and 79.5% [CI, 69.4-86.9]). Posttussive vomiting in children is moderately sensitive (60.0% [CI, 40.3-77.0]) and specific (66.0% [CI, 52.5-77.3]). CONCLUSIONS: In adult patients, the presence of whooping or posttussive vomiting should rule in a possible diagnosis of pertussis, whereas the lack of a paroxysmal cough or the presence of fever should rule it out. In children, posttussive vomiting is much less helpful as a clinical diagnostic test.

GP views on strategies to cope with increasing workload: a qualitative interview study.

BACKGROUND: The existence of a crisis in primary care in the UK is in little doubt. GP morale and job satisfaction are low, and workload is increasing. In this challenging context, finding ways for GPs to manage that workload is imperative. AIM: To explore what existing or potential strategies are described by GPs for dealing with their workload, and their views on the relative merits of each. DESIGN AND SETTING: Semi-structured, qualitative interviews with GPs working within NHS England. METHOD: All GPs working within NHS England were eligible. Of those who responded to advertisements, a maximum-variation sample was selected and interviewed until data saturation was reached. Data were analysed thematically. RESULTS: Responses were received from 171 GPs, and, from these, 34 were included in the study. Four main themes emerged for workload management: patient-level, GP-level, practice-level, and systems-level strategies. A need for patients to take greater responsibility for self-management was clear, but many felt that GPs should not be responsible for this education. Increased delegation of tasks was felt to be key to managing workload, with innovative use of allied healthcare professionals and extended roles for non-clinical staff suggested. Telephone triage was a commonly used tool for managing workload, although not all participants found this helpful. CONCLUSION: This in-depth qualitative study demonstrates an encouraging resilience among GPs. They are proactively trying to manage workload, often using innovative local strategies. GPs do not feel that they can do this alone, however, and called repeatedly for increased recruitment and more investment in primary care.

GPs' perceptions of workload in England: a qualitative interview study.

BACKGROUND: GPs report the lowest levels of morale among doctors, job satisfaction is low, and the GP workforce is diminishing. Workload is frequently cited as negatively impacting on commitment to a career in general practice, and many GPs report that their workload is unmanageable. AIM: To gather an in-depth understanding of GPs' perceptions and attitudes towards workload. DESIGN AND SETTING: All GPs working within NHS England were eligible. Advertisements were circulated via regional GP e-mail lists and national social media networks in June 2015. Of those GPs who responded, a maximum-variation sample was selected until data saturation was reached. METHOD: Semi-structured, qualitative interviews were conducted. Data were analysed thematically. RESULTS: In total, 171 GPs responded, and 34 were included in this study. GPs described an increase in workload over recent years, with current working days being long and intense, raising concerns over the wellbeing of GPs and patients. Full-time partnership was generally not considered to be possible, and many participants felt workload was unsustainable, particularly given the diminishing workforce. Four major themes emerged to explain increased workload: increased patient needs and expectations; a changing relationship between primary and secondary care; bureaucracy and resources; and the balance of workload within a practice. Continuity of care was perceived as being eroded by changes in contracts and working patterns to deal with workload. CONCLUSION: This study highlights the urgent need to address perceived lack of investment and clinical capacity in general practice, and suggests that managing patient expectations around what primary care can deliver, and reducing bureaucracy, have become key issues, at least until capacity issues are resolved.

Prescribing antibiotics to 'at-risk' children with influenza-like illness in primary care: qualitative study.

OBJECTIVES: National Institute for Health and Care Excellence guidelines recommend immediate antibiotic treatment of respiratory tract infections in 'at-risk' individuals with comorbidities. Observational evidence suggests that influenza particularly predisposes children to bacterial complications. This study investigates general practitioners' (GPs') accounts of factors influencing their decision-making about antibiotic prescribing in the management of at-risk children with influenza-like illness (ILI). DESIGN: Qualitative interview study using a maximum variation sample with thematic analysis through constant comparison. SETTING: Semistructured telephone interviews with UK GPs using a case vignette of a child with comorbidities presenting with ILI. PARTICIPANTS: There were 41 GPs (41.5% men; 40 from England, 1 from Northern Ireland) with a range of characteristics including length of time in practice, paediatrics experience, practice setting and deprivation. RESULTS: There was considerable uncertainty and variation in the way GPs responded to the case and difference of opinion about how long-term comorbidities should affect their antibiotic prescribing pattern. Factors influencing their decision included the child's case history and clinical examination; the GP's view of the parent's ability to self-manage; the GP's own confidence and experiences of managing sick children and assessment of individual versus abstract risk. GPs rarely mentioned potential influenza infection or asked about immunisation status. All said that they would want to see the child; views about delayed prescribing varied in relation to local health service provision including options for follow-up and paediatric services. CONCLUSIONS: The study demonstrates diagnostic uncertainty and wide variation in GP decision-making about prescribing antibiotics to children with comorbidity. Future guidelines might encourage consideration of a specific diagnosis such as influenza, and risk assessment tools could be developed to allow clinicians to quantify the levels of risk associated with different types of comorbidity. However, the wide range of clinical and non-clinical factors involved in decision-making during these consultations should also be considered in future guidelines.

Identification of children at risk of influenza-related complications in primary and ambulatory care: a systematic review and meta-analysis.

BACKGROUND: Interventions to prevent influenza-related complications are recommended for individuals at the greatest risk of serious clinical deterioration. However, guidelines are based on consensus opinion rather than evidence, and do not specify risk factors in children. We aimed to provide an evidence-based definition of children who are most at risk of such complications. METHODS: In this systematic review, we searched the Medline and Medline In Process, Embase, Science Citation Index, and CINAHL databases for studies published between inception and April 3, 2013. We included studies that reported data for underlying disorders and complications in children presenting in primary or ambulatory care with influenza or influenza-like illness. We requested unpublished data from investigators of studies that had obtained, but not published, relevant data. We analysed data with univariable meta-analysis and individual patient data multivariable meta-analysis methods. The primary outcome was admission to hospital as a proxy for complications of influenza or influenza-like illness. FINDINGS: We included 28 articles that reported data from 27 studies (14 086 children). Strong risk factors for hospital admission were neurological disorders (univariable odds ratio [OR] 4·â€ˆ62, 95% CI 2·82-7·55), prematurity (4·33, 2·47-7·58), sickle cell disease (3·46, 1·63-7·37), immunosuppression (2·39, 1·24-4·61), diabetes (2·34, 1·20-4·58), and age younger than 2 years (2·51, 1·71-3·69). However, reactive airways disease including asthma (1·36, 0·82-2·26) and obesity (0·99, 0·61-1·62) were not found to be risk factors. On the basis of individual patient data multivariable analysis (1612 children, four studies), the risk of hospital admission was higher in children with more than one risk factor than in children with just one risk factor, when age younger than 2 years was included as a risk factor (92 [74%] of 124 vs 428 [52%] of 817; difference 22%, 95% CI 13-30%, p<0·0001). INTERPRETATION: We identified prematurity as a new strong risk factor for influenza-related complications in children. Our findings also support the inclusion of neurological disorders, sickle cell disease, immunosuppression, diabetes, and age younger than 2 years as risk factors in existing guidelines. Interventions to prevent influenza-related complications should be prioritised in these groups, but should also be considered for other children, especially those with more than one risk factor or severe underlying comorbidities. FUNDING: UK National Institute for Health Research.

Diagnostic accuracy study of three alcohol breathalysers marketed for sale to the public.

OBJECTIVES: To assess the diagnostic accuracy of three personal breathalyser devices available for sale to the public marketed to test safety to drive after drinking alcohol. DESIGN: Prospective comparative diagnostic accuracy study comparing two single-use breathalysers and one digital multiuse breathalyser (index tests) to a police breathalyser (reference test). SETTING: Establishments licensed to serve alcohol in a UK city. PARTICIPANTS: Of 222 participants recruited, 208 were included in the main analysis. Participants were eligible if they were 18 years old or over, had consumed alcohol and were not intending to drive within the following 6 h. OUTCOME MEASURES: Sensitivity and specificity of the breathalysers for the detection of being at or over the UK legal driving limit (35 µg/100 mL breath alcohol concentration). RESULTS: 18% of participants (38/208) were at or over the UK driving limit according to the police breathalyser. The digital multiuse breathalyser had a sensitivity of 89.5% (95% CI 75.9% to 95.8%) and a specificity of 64.1% (95% CI 56.6% to 71.0%). The single-use breathalysers had a sensitivity of 94.7% (95% CI 75.4% to 99.1%) and 26.3% (95% CI 11.8% to 48.8%), and a specificity of 50.6% (95% CI 40.4% to 60.7%) and 97.5% (95% CI 91.4% to 99.3%), respectively. Self-reported alcohol consumption threshold of 5 UK units or fewer had a higher sensitivity than all personal breathalysers. CONCLUSIONS: One alcohol breathalyser had sensitivity of 26%, corresponding to false reassurance for approximately three people in four who are over the limit by the reference standard, at least on the evening of drinking alcohol. The other devices tested had 90% sensitivity or higher. All estimates were subject to uncertainty. There is no clearly defined minimum sensitivity for this safety-critical application. We conclude that current regulatory frameworks do not ensure high sensitivity for these devices marketed to consumers for a decision with potentially catastrophic consequences.

Determining immunisation status of children from history: a diagnostic accuracy study.

OBJECTIVES: Children presenting unplanned to healthcare services are routinely asked about previous immunisations as part of their assessment. We aimed to assess the accuracy of screening children for immunisation status by history. DESIGN: Diagnostic accuracy study. We compared information from patient history by a retrospective review of notes and used a central database of child immunisation records as the reference standard. SETTING: Paediatric emergency department in a tertiary hospital in Oxford, UK. PARTICIPANTS: Consecutive children aged 6 months to 6 years presenting over a 2-month period. OUTCOME MEASURES: Proportion of children with documented immunisation history; sensitivity and specificity of detecting overdue immunisations by history compared to central records. RESULTS: 1166 notes were surveyed. 76.3% children were asked about immunisations. The proportion of children who were fully immunised on central records was 93.1%. History had a sensitivity of 41.3% (95% CI 27% to 56.8%) and a specificity of 98.7% (95% CI 97.5% to 99.4%) for detecting those who were overdue. Negative predictive value was 95.8% (95% CI 93.9% to 97.2%). Only around a third of children with overdue immunisations are detected by the current screening methods, and approximately 1 in 20 children stated as being up to date are in fact overdue. CONCLUSIONS: History had poor sensitivity for identifying overdue immunisation. Strategies to improve detection of children overdue with immunisation should focus on alternative strategies for alerting clinicians, such as linkage of community and hospital electronic records.

Pain over speed bumps in diagnosis of acute appendicitis: diagnostic accuracy study.

OBJECTIVE: To assess the diagnostic accuracy of pain on travelling over speed bumps for the diagnosis of acute appendicitis. DESIGN: Prospective questionnaire based diagnostic accuracy study. SETTING: Secondary care surgical assessment unit at a district general hospital in the UK. PARTICIPANTS: 101 patients aged 17-76 years referred to the on-call surgical team for assessment of possible appendicitis. MAIN OUTCOME MEASURES: Sensitivity, specificity, positive and negative predictive values, and positive and negative likelihood ratios for pain over speed bumps in diagnosing appendicitis, with histological diagnosis of appendicitis as the reference standard. RESULTS: The analysis included 64 participants who had travelled over speed bumps on their journey to hospital. Of these, 34 had a confirmed histological diagnosis of appendicitis, 33 of whom reported increased pain over speed bumps. The sensitivity was 97% (95% confidence interval 85% to 100%), and the specificity was 30% (15% to 49%). The positive predictive value was 61% (47% to 74%), and the negative predictive value was 90% (56% to 100%). The likelihood ratios were 1.4 (1.1 to 1.8) for a positive test result and 0.1 (0.0 to 0.7) for a negative result. Speed bumps had a better sensitivity and negative likelihood ratio than did other clinical features assessed, including migration of pain and rebound tenderness. CONCLUSIONS: Presence of pain while travelling over speed bumps was associated with an increased likelihood of acute appendicitis. As a diagnostic variable, it compared favourably with other features commonly used in clinical assessment. Asking about speed bumps may contribute to clinical assessment and could be useful in telephone assessment of patients.