Exploring the discrepancies between clinical trials and real-world data: A small-cell lung cancer study.

The potential of real-world data to inform clinical trial design and supplement control arms has gained much interest in recent years. The most common approach relies on reproducing control arm outcomes by matching real-world patient cohorts to clinical trial baseline populations. However, recent studies pointed out that there is a lack of replicability, generalisability, and consensus. In this article, we propose a novel approach that aims to explore and examine these discrepancies by concomitantly investigating the impact of selection criteria and operations on the measurements of outcomes from the patient data. We tested the approach on a dataset consisting of small-cell lung cancer patients receiving platinum-based chemotherapy regimens from a real-world data cohort (n = 223) and six clinical trial control arms (n = 1224). The results showed that the discrepancy between real-world and clinical trial data potentially depends on differences in both patient populations and operational conditions (e.g., frequency of assessments, and censoring), for which further investigation is required. Discovering and accounting for confounders, including hidden effects of differences in operations related to the treatment process and clinical trial study protocol, would potentially allow for improved translation between clinical trials and real-world data. Continued development of the method presented here to systematically explore and account for these differences could pave the way for transferring learning across clinical studies and developing mutual translation between the real-world and clinical trials to inform clinical study design.

Exploring Hospital Overcrowding with an Explainable Time-to-Event Machine Learning Approach.

Emergency department (ED) overcrowding is a complex problem that is intricately linked with the operations of other hospital departments. Leveraging ED real-world production data provides a unique opportunity to comprehend this multifaceted problem holistically. This paper introduces a novel approach to analyse healthcare production data, treating the length of stay of patients, and the follow up decision regarding discharge or admission to the hospital as a time-to-event analysis problem. Our methodology employs traditional survival estimators and machine learning models, and Shapley additive explanations values to interpret the model outcomes. The most relevant features influencing length of stay were whether the patient received a scan at the ED, emergency room urgent visit, age, triage level, and the medical alarm unit category. The clinical insights derived from the explanation of the models holds promise for increase understanding of the overcrowding from the data. Our work demonstrates that a time-to-event approach to the over- crowding serves as a valuable initial to uncover crucial insights for further investigation and policy design.

FEMaLe: The use of machine learning for early diagnosis of endometriosis based on patient self-reported data-Study protocol of a multicenter trial.

INTRODUCTION: Endometriosis is a chronic disease that affects up to 190 million women and those assigned female at birth and remains unresolved mainly in terms of etiology and optimal therapy. It is defined by the presence of endometrium-like tissue outside the uterine cavity and is commonly associated with chronic pelvic pain, infertility, and decreased quality of life. Despite the availability of various screening methods (e.g., biomarkers, genomic analysis, imaging techniques) intended to replace the need for invasive surgery, the time to diagnosis remains in the range of 4 to 11 years. AIMS: This study aims to create a large prospective data bank using the Lucy mobile health application (Lucy app) and analyze patient profiles and structured clinical data. In addition, we will investigate the association of removed or restricted dietary components with quality of life, pain, and central pain sensitization. METHODS: A baseline and a longitudinal questionnaire in the Lucy app collects real-world, self-reported information on symptoms of endometriosis, socio-demographics, mental and physical health, economic factors, nutritional, and other lifestyle factors. 5,000 women with confirmed endometriosis and 5,000 women without diagnosed endometriosis in a control group will be enrolled and followed up for one year. With this information, any connections between recorded symptoms and endometriosis will be analyzed using machine learning. CONCLUSIONS: We aim to develop a phenotypic description of women with endometriosis by linking the collected data with existing registry-based information on endometriosis diagnosis, healthcare utilization, and big data approach. This may help to achieve earlier detection of endometriosis with pelvic pain and significantly reduce the current diagnostic delay. Additionally, we may identify dietary components that worsen the quality of life and pain in women with endometriosis, upon which we can create real-world data-based nutritional recommendations.

Tensions between real-world practices and the digitalization paradigm for data-driven services in eldercare: observations from an ethnographic study in Sweden.

BACKGROUND: The implementation of a data-driven approach within the health care system happens in a rapid pace; including in the eldercare sector. Within Swedish eldercare, data-driven health approach is not yet widely implemented. In the specific context of long-term care for older adults, quality of care is as much determined by how social care is being performed as it is by what kind medical care that is provided. In particular, relational aspects have been proven to have a crucial influence on the experience of quality of care for the actors involved. Drawing on ethnographic material collected at a Swedish nursing home, this paper explores in what way the relational aspects of care could potentially become affected by the increased use of a data-driven health approach. METHODS: An ethnographic approach was adopted in order to investigate the daily care work at a long-term care facility as it unfolded. Fieldwork was conducted at a somatic ward in a Swedish long-term care facility over 4 months (86 h in total), utilizing the methods of participant observation, informal interviews and document analysis. The material was analyzed iteratively throughout the entire research process adopting thematic analysis. RESULTS: Viewing our ethnographic material through an observational lense problematising the policy discourse around data-driven health approach, two propositions were developed. First, we propose that relational knowledge risk becoming less influential in shaping everyday care, when moving to a data-driven health approach. Second, we propose that quality of care risk becoming more directed on quality of medical care at the expense of quality of life. CONCLUSION: While the implementation of data-driven health approach within long-term care for older adults is not yet widespread, the general development within health care points towards a situation in which this will become reality. Our study highlights the importance of taking the relational aspects of care into consideration, both during the planning and implementation phase of this process. By doing this, the introduction of a data-driven health approach could serve to heighten the quality of care in a way which supports both quality of medical care and quality of life.

Investigating the Connections Between Delivery of Care, Reablement, Workload, and Organizational Factors in Home Care Services: Mixed Methods Study.

BACKGROUND: Home care is facing increasing demand due to an aging population. Several challenges have been identified in the provision of home care, such as the need for support and tailoring support to individual needs. Goal-oriented interventions, such as reablement, may provide a solution to some of these challenges. The reablement approach targets adaptation to disease and relearning of everyday life skills and has been found to improve health-related quality of life while reducing service use. OBJECTIVE: The objective of this study is to characterize home care system variables (elements) and their relationships (connections) relevant to home care staff workload, home care user needs and satisfaction, and the reablement approach. This is to examine the effects of improvement and interventions, such as the person-centered reablement approach, on the delivery of home care services, workload, work-related stress, home care user experience, and other organizational factors. The main focus was on Swedish home care and tax-funded universal welfare systems. METHODS: The study used a mixed methods approach where a causal loop diagram was developed grounded in participatory methods with academic health care science research experts in nursing, occupational therapy, aging, and the reablement approach. The approach was supplemented with theoretical models and the scientific literature. The developed model was verified by the same group of experts and empirical evidence. Finally, the model was analyzed qualitatively and through simulation methods. RESULTS: The final causal loop diagram included elements and connections across the categories: stress, home care staff, home care user, organization, social support network of the home care user, and societal level. The model was able to qualitatively describe observed intervention outcomes from the literature. The analysis suggested elements to target for improvement and the potential impact of relevant studied interventions. For example, the elements "workload" and "distress" were important determinants of home care staff health, provision, and quality of care. CONCLUSIONS: The developed model may be of value for informing hypothesis formulation, study design, and discourse within the context of improvement in home care. Further work will include a broader group of stakeholders to reduce the risk of bias. Translation into a quantitative model will be explored.

Application of Process Mining for Modelling Small Cell Lung Cancer Prognosis.

Process mining is a relatively new method that connects data science and process modelling. In the past years a series of applications with health care production data have been presented in process discovery, conformance check and system enhancement. In this paper we apply process mining on clinical oncological data with the purpose of studying survival outcomes and chemotherapy treatment decision in a real-world cohort of small cell lung cancer patients treated at Karolinska University Hospital (Stockholm, Sweden). The results highlighted the potential role of process mining in oncology to study prognosis and survival outcomes with longitudinal models directly extracted from clinical data derived from healthcare.

A novel analytical framework for risk stratification of real-world data using machine learning: A small cell lung cancer study.

In recent studies, small cell lung cancer (SCLC) treatment guidelines based on Veterans' Administration Lung Study Group limited/extensive disease staging and resulted in broad and inseparable prognostic subgroups. Evidence suggests that the eight versions of tumor, node, and metastasis (TNM) staging can play an important role to address this issue. The aim of the present study was to improve the detection of prognostic subgroups from a real-word data (RWD) cohort of patients and analyze their patterns using a development pipeline with thoracic oncologists and machine learning methods. The method detected subgroups of patients informing unsupervised learning (partition around medoids) including the impact of covariates on prognosis (Cox regression and random survival forest). An analysis was carried out using patients with SCLC (n = 636) with stage IIIA-IVB according to TNM classification. The analysis yielded k = 7 compacted and well-separated clusters of patients. Performance status (Eastern Cooperative Oncology Group-Performance Status), lactate dehydrogenase, spreading of metastasis, cancer stage, and CRP were the baselines that characterized the subgroups. The selected clustering method outperformed standard clustering techniques, which were not capable of detecting meaningful subgroups. From the analysis of cluster treatment decisions, we showed the potential of future RWD applications to understand disease, develop individualized therapies, and improve healthcare decision making.