Participant flow diagrams for health equity in AI.

Selection bias can arise through many aspects of a study, including recruitment, inclusion/exclusion criteria, input-level exclusion and outcome-level exclusion, and often reflects the underrepresentation of populations historically disadvantaged in medical research. The effects of selection bias can be further amplified when non-representative samples are used in artificial intelligence (AI) and machine learning (ML) applications to construct clinical algorithms. Building on the "Data Cards" initiative for transparency in AI research, we advocate for the addition of a participant flow diagram for AI studies detailing relevant sociodemographic and/or clinical characteristics of excluded participants across study phases, with the goal of identifying potential algorithmic biases before their clinical implementation. We include both a model for this flow diagram as well as a brief case study explaining how it could be implemented in practice. Through standardized reporting of participant flow diagrams, we aim to better identify potential inequities embedded in AI applications, facilitating more reliable and equitable clinical algorithms.

Clinical characteristics governing treatment adjustment in COPD patients: results from the Swiss COPD cohort study.

BACKGROUND: Chronic obstructive pulmonary disease (COPD) is a widespread chronic disease characterised by irreversible airway obstruction [1]. Features of clinical practice and healthcare systems for COPD patients can vary widely, even within similar healthcare structures. The Global Initiative for Chronic Obstructive Lung Disease (GOLD) strategy is considered the most reliable guidance for the management of COPD and aims to provide treating physicians with appropriate insight into the disease. COPD treatment adaptation typically mirrors the suggestions within the GOLD guidelines, depending on how the patient has been categorised. However, the present study posits that the reasons for adjusting COPD-related treatment are hugely varied. OBJECTIVES: The objective of this study was to assess the clinical symptoms that govern both pharmacological and non-pharmacological treatment changes in COPD patients. Using this insight, the study offers suggestions for optimising COPD management through the implementation of GOLD guidelines. METHODS: In this observational cohort study, 24 general practitioners screened 260 COPD patients for eligibility from 2015-2019. General practitioners were asked to collect general information from patients using a standardised questionnaire to document symptoms. During a follow-up visit, the patient's symptoms and changes in therapy were assessed and entered into a central electronic database. Sixty-five patients were removed from the analysis due to exclusion criteria, and 195 patients with at least one additional visit within one year of the baseline visit were included in the analysis. A change in therapy was defined as a change in either medication or non-medical treatment, such as pulmonary rehabilitation. Multivariable mixed models were used to identify associations between given symptoms and a step up in therapy, a step down, or a step up and a step down at the same time. RESULTS: For the 195 patients included in analyses, a treatment adjustment was made during 28% of visits. In 49% of these adjustments, the change in therapy was a step up, in 33% a step down and in 18% a step up (an increase) of certain treatment factors and a step down (a reduction) of other prescribed treatments at the same time. In the multivariable analysis, we found that the severity of disease was linked to the probability of therapy adjustment: patients in GOLD Group C were more likely to experience an increase in therapy compared to patients in GOLD Group A (odds ratio [OR] 3.43 [95% confidence interval {CI}: 1.02-11.55; p = 0.135]). In addition, compared to patients with mild obstruction, patients with severe (OR 4.24 [95% CI: 1.88-9.56]) to very severe (OR 5.48 [95% CI: 1.31-22.96]) obstruction were more likely to experience a therapy increase (p <0.0001). Patients with comorbidities were less likely to experience a treatment increase than those without (OR 0.42 [95% CI: 0.24-0.73; p = 0.002]). A therapy decrease was associated with both a unit increase in COPD Assessment Test (CAT) score (OR 1.07 [95% CI: 1.01-1.14; p = 0.014]) and having experienced an exacerbation (OR 2.66 [95% CI: 1.01-6.97; p = 0.047]). The combination of steps up as well as steps down in therapy was predicted by exacerbation (OR 8.93 [95% CI: 1.16-68.28; p = 0.035]) and very severe obstruction (OR 589 [95% CI: 2.72 - >999; p = 0.109]). CONCLUSIONS: This cohort study provides insight into the management of patients with COPD in a primary care setting. COPD Group C and airflow limitation GOLD 3-4 were both associated with an increase in COPD treatment. In patients with comorbidities, there were often no treatment changes. Exacerbations did not make therapy increases more probable. The presence of neither cough/sputum nor high CAT scores was associated with a step up in treatment.

Peripheral Blood Single-Cell Sequencing Uncovers Common and Specific Immune Aberrations in Fibrotic Lung Diseases.

Rationale and Objectives: The extent and commonality of peripheral blood immune aberrations in fibrotic interstitial lung diseases are not well characterized. In this study, we aimed to identify common and distinct immune aberrations in patients with idiopathic pulmonary fibrosis (IPF) and fibrotic hypersensitivity pneumonitis (FHP) using cutting-edge single-cell profiling technologies. Methods: Single-cell RNA sequencing was performed on patients and healthy controls' peripheral blood and bronchoalveolar lavage samples using 10X Genomics 5' gene expression and V(D)J profiling. Cell type composition, transcriptional profiles, cellular trajectories and signaling, and T and B cell receptor repertoires were studied. The standard Seurat R pipeline was followed for cell type composition and differential gene expression analyses. Transcription factor activity was imputed using the DoRothEA-VIPER algorithm. Pseudotime analyses were conducted using Monocle3, while RNA velocity analyses were performed with Velocyto, scVelo, and CellRank. Cell-cell connectomics were assessed using the Connectome R package. V(D)J analyses were conducted using CellRanger and Immcantation frameworks. Across all analyses, disease group differences were assessed using the Wilcoxon rank-sum test. Measurements and Main Results: 327,990 cells from 83 samples were profiled. Overall, changes in monocytes were common to IPF and FHP, whereas lymphocytes exhibited disease-specific aberrations. Both diseases displayed enrichment of CCL3 hi /CCL4 hi CD14+ monocytes (p<2.2e-16) and S100A hi CD14+ monocytes (p<2.2e-16) versus controls. Trajectory and RNA velocity analysis suggested that pro-fibrotic macrophages observed in BAL originated from peripheral blood monocytes. Lymphocytes exhibited disease-specific aberrations, with CD8+ GZMK hi T cells and activated B cells primarily enriched in FHP patients. V(D)J analyses revealed unique T and B cell receptor complementarity-determining region 3 (CDR3) amino acid compositions (p<0.05) in FHP and significant IgA enrichment in IPF (p<5.2e-7). Conclusions: We identified common and disease-specific immune mechanisms in IPF and FHP; S100A hi monocytes and SPP1 hi macrophages are common to IPF and FHP, whereas GMZK hi T lymphocytes and T and B cell receptor repertoires were unique in FHP. Our findings open novel strategies for the diagnosis and treatment of IPF and FHP.

Risk Factors for Recurrent Exacerbations in the General-Practitioner-Based Swiss Chronic Obstructive Pulmonary Disease (COPD) Cohort.

BACKGROUND: Patients with chronic obstructive pulmonary disease (COPD) often suffer from acute exacerbations. Our objective was to describe recurrent exacerbations in a GP-based Swiss COPD cohort and develop a statistical model for predicting exacerbation. METHODS: COPD cohort demographic and medical data were recorded for 24 months, by means of a questionnaire-based COPD cohort. The data were split into training (75%) and validation (25%) datasets. A negative binomial regression model was developed using the training dataset to predict the exacerbation rate within 1 year. An exacerbation prediction model was developed, and its overall performance was validated. A nomogram was created to facilitate the clinical use of the model. RESULTS: Of the 229 COPD patients analyzed, 77% of the patients did not experience exacerbation during the follow-up. The best subset in the training dataset revealed that lower forced expiratory volume, high scores on the MRC dyspnea scale, exacerbation history, and being on a combination therapy of LABA + ICS (long-acting beta-agonists + Inhaled Corticosteroids) or LAMA + LABA (Long-acting muscarinic receptor antagonists + long-acting beta-agonists) at baseline were associated with a higher rate of exacerbation. When validated, the area-under-curve (AUC) value was 0.75 for one or more exacerbations. The calibration was accurate (0.34 predicted exacerbations vs 0.28 observed exacerbations). CONCLUSION: Nomograms built from these models can assist clinicians in the decision-making process of COPD care.

Strategies and solutions to address Digital Determinants of Health (DDOH) across underinvested communities.

Healthcare has long struggled to improve services through technology without further widening health disparities. With the significant expansion of digital health, a group of healthcare professionals and scholars from across the globe are proposing the official usage of the term "Digital Determinants of Health" (DDOH) to explicitly call out the relationship between technology, healthcare, and equity. This is the final paper in a series published in PLOS Digital Health that seeks to understand and summarize current knowledge of the strategies and solutions that help to mitigate the negative effects of DDOH for underinvested communities. Through a search of English-language Medline, Scopus, and Google Scholar articles published since 2010, 345 articles were identified that discussed the application of digital health technology among underinvested communities. A group of 8 reviewers assessed 132 articles selected at random for the mention of solutions that minimize differences in DDOH. Solutions were then organized by categories of policy; design and development; implementation and adoption; and evaluation and ongoing monitoring. The data were then assessed by category and the findings summarized. The reviewers also looked for common themes across the solutions and evidence of effectiveness. From this limited scoping review, the authors found numerous solutions mentioned across the papers for addressing DDOH and many common themes emerged regardless of the specific community or digital health technology under review. There was notably less information on solutions regarding ongoing evaluation and monitoring which corresponded with a lack of research evidence regarding effectiveness. The findings directionally suggest that universal strategies and solutions can be developed to address DDOH independent of the specific community under focus. With the need for the further development of DDOH measures, we also provide a framework for DDOH assessment.

Incidence and Prognostic Significance of Hypoxemia in Fibrotic Interstitial Lung Disease: An International Cohort Study.

BACKGROUND: Hypoxemia is a cardinal feature of fibrotic interstitial lung disease (ILD). The incidence, progression, and prognostic significance of hypoxemia in patients with fibrotic ILD currently is unknown. RESEARCH QUESTION: What are the epidemiologic features of hypoxemia and its additive prognostic value in a current risk prediction model of fibrotic ILD? METHODS: We identified 848 patients with fibrotic ILD (258 with idiopathic pulmonary fibrosis [IPF]) in five prospective ILD registries from Australia, Canada, and Switzerland. Cumulative incidence of exertional and resting hypoxemia from the time of diagnosis was estimated at 1-year intervals in patients with baseline 6-min walk tests, adjusted for competing risks of death and lung transplantation. Likelihood ratio tests were used to determine the prognostic significance of exertional and resting hypoxemia for 1-year mortality or transplantation when added to the ILD-GAP model. The cohort was divided into derivation and validation subsets to evaluate performance characteristics of the extended model (the ILD-GAP-O2 model), which included oxygenation status as a predictor. RESULTS: The 1-, 2-, and 5-year overall cumulative incidence was 6.1%, 17.3%, and 40.1%, respectively, for exertional hypoxemia and 2.4%, 5.6%, and 16.5%, respectively, for resting hypoxemia, which were significantly higher in patients with IPF compared with patients without IPF (P < .001 for both). Addition of exertional or resting hypoxemia to the ILD-GAP model improved 1-year mortality and transplantation prediction (P < .001 for both). The ILD-GAP-O2 model showed improved discrimination (C-index, 0.80 vs 0.75) and model fit (Akaike information criteria, 400 vs 422) in the validation cohort, with comparable calibration. INTERPRETATION: Patients with IPF have higher cumulative incidence of exertional and resting hypoxemia than patients without IPF. The extended ILD-GAP-O2 model provides additional risk stratification for 1-year prognosis in fibrotic ILD.

Predictive factors for exacerbation and re-exacerbation in chronic obstructive pulmonary disease: an extension of the Cox model to analyze data from the Swiss COPD cohort.

BACKGROUND: The Swiss COPD cohort was established in 2006 to collect data in a primary care setting. The objective of this study was to evaluate possible predictive factors for exacerbation and re-exacerbation. METHODS: In order to predict exacerbation until the next visit based on the knowledge of exacerbation since the last visit, a multistate model described by Therneau and Grambsch was performed. RESULTS: Data of 1,247 patients (60.4% males, 46.6% current smokers) were analyzed, 268 (21.5%) did not fulfill spirometric diagnostic criteria for COPD. Data of 748 patients (63% males, 44.1% current smokers) were available for model analysis. In order to predict exacerbation an extended Cox Model was performed. Mean FEV1/FVC-ratio was 53.1% (±11.5), with a majority of patients in COPD GOLD classes 2 or 3. Hospitalization for any reason (HR1.7; P = 0.04) and pronounced dyspnea (HR for mMRC grade four 3.0; P < 0.001) at most recent visit as well as prescription of short-acting bronchodilators (HR1.7; P < 0.001), inhaled (HR1.2; P = 0.005) or systemic corticosteroids (HR1.8; P = 0.015) were significantly associated with exacerbation when having had no exacerbation at most recent visit. Higher FEV1/FVC (HR0.9; P = 0.008) and higher FEV1 values (HR0.9; P = 0.001) were protective. When already having had an exacerbation at the most recent visit, pronounced dyspnea (HR for mMRC grade 4 1.9; P = 0.026) and cerebrovascular insult (HR2.1; P = 0.003) were significantly associated with re-exacerbation. Physical activity (HR0.6; P = 0.031) and treatment with long-acting anticholinergics (HR0.7; P = 0.044) seemed to play a significant protective role. In a best subset model for exacerbation, higher FEV1 significantly reduced and occurrence of sputum increased the probability of exacerbation. In the same model for re-exacerbation, coronary heart disease increased and hospitalization at most recent visit seemed to reduce the risk for re-exacerbation. CONCLUSION: Our data confirmed well-established risk factors for exacerbations whilst analyzing their predictive association with exacerbation and re-exacerbation. This study confirmed the importance of spirometry in primary care, not only for diagnosis but also as a risk evaluation for possible future exacerbations. TRIAL REGISTRATION: Our study got approval by local ethical committee in 2006 (EK Nr. 170/06) and was registered retrospectively on ClinicalTrials.gov (NCT02065921, 19th of February 2014).

The ADO index as a predictor of two-year mortality in general practice-based chronic obstructive pulmonary disease cohorts.

BACKGROUND: Existing prediction models for mortality in chronic obstructive pulmonary disease (COPD) patients have not yet been validated in primary care, which is where the majority of patients receive care. OBJECTIVES: Our aim was to validate the ADO (age, dyspnoea, airflow obstruction) index as a predictor of 2-year mortality in 2 general practice-based COPD cohorts. METHODS: Six hundred and forty-six patients with COPD with GOLD (Global Initiative for Chronic Obstructive Lung Disease) stages I-IV were enrolled by their general practitioners and followed for 2 years. The ADO regression equation was used to predict a 2-year risk of all-cause mortality in each patient and this risk was compared with the observed 2-year mortality. Discrimination and calibration were assessed as well as the strength of association between the 15-point ADO score and the observed 2-year all-cause mortality. RESULTS: Fifty-two (8.1%) patients died during the 2-year follow-up period. Discrimination with the ADO index was excellent with an area under the curve of 0.78 [95% confidence interval (CI) 0.71-0.84]. Overall, the predicted and observed risks matched well and visual inspection revealed no important differences between them across 10 risk classes (p = 0.68). The odds ratio for death per point increase according to the ADO index was 1.50 (95% CI 1.31-1.71). CONCLUSIONS: The ADO index showed excellent prediction properties in an out-of-population validation carried out in COPD patients from primary care settings.

Endothelial microparticle uptake in target cells is annexin I/phosphatidylserine receptor dependent and prevents apoptosis.

OBJECTIVE: Endothelial microparticles (EMP) are released from activated or apoptotic cells, but their effect on target cells and the exact way of incorporation are largely unknown. We sought to determine the uptake mechanism and the biological effect of EMP on endothelial and endothelial-regenerating cells. METHODS AND RESULTS: EMP were generated from starved endothelial cells and isolated by ultracentrifugation. Caspase 3 activity assay and terminal deoxynucleotidyl transferase dUTP nick end labeling assay showed that EMP protect target endothelial cells against apoptosis in a dose-dependent manner. Proteomic analysis was performed to identify molecules contained in EMP, which might be involved in EMP uptake. Expression of annexin I in EMP was found and confirmed by Western blot, whereas the corresponding receptor phosphatidylserine receptor was present on endothelial target cells. Silencing either annexin I on EMP or phosphatidylserine receptor on target cells using small interfering RNA showed that the uptake of EMP by human coronary artery endothelial cells is annexin I/phosphatidylserine receptor dependent. Annexin I-downregulated EMP abrogated the EMP-mediated protection against apoptosis of endothelial target cells. p38 activation was found to mediate camptothecin-induced apoptosis. Finally, human coronary artery endothelial cells pretreated with EMP inhibited camptothecin-induced p38 activation. CONCLUSIONS: EMP are incorporated by endothelial cells in an annexin I/phosphatidylserine receptor-dependent manner and protect target cells against apoptosis. Inhibition of p38 activity is involved in EMP-mediated protection against apoptosis.