Big data and precision medicine: challenges and strategies with healthcare data

Improving acute kidney injury alerts in tertiary care by linking primary care data: An observational cohort using routine care data

Objective

Acute kidney injury (AKI) is easily missed and underdiagnosed in routine clinical care. Timely AKI management is important to decrease morbidity and mortality risks. We recently implemented an AKI e-alert at the University Medical Center Utrecht, comparing plasma creatinine concentrations with historical creatinine baselines, thereby identifying patients with AKI. This alert is limited to data from tertiary care, and primary care data can increase diagnostic accuracy for AKI. We assessed the added value of linking primary care data to tertiary care data, in terms of timely diagnosis or excluding AKI.

Methods

With plasma creatinine tests for 84,984 emergency department (ED) visits, we applied the Kidney Disease Improving Global Outcome guidelines in both tertiary care-only data and linked data and compared AKI cases.

Results

Using linked data, the presence of AKI could be evaluated in an additional 7886 ED visits. Sex- and age-stratified analyses identified the largest added value for women (an increase of 4095 possible diagnoses) and patients ≥60 years (an increase of 5190 possible diagnoses). We observed 398 additional visits where AKI was diagnosed, as well as 185 cases where AKI could be excluded. We observed no overall decrease in time between baseline and AKI diagnosis (28.4 days vs. 28.0 days). For cases where AKI was diagnosed in both data sets, we observed a decrease of 2.8 days after linkage, indicating a timelier diagnosis of AKI.

Conclusions

Combining primary and tertiary care data improves AKI diagnostic accuracy in routine clinical care and enables timelier AKI diagnosis.

AMBAR - Interactive Alteration annotations for molecular tumor boards

MOTIVATION

Personalized decision-making for cancer therapy relies on molecular profiling from sequencing data in combination with database evidence and expert knowledge. Molecular tumor boards (MTBs) bring together clinicians and scientists with diverse expertise and are increasingly established in the clinical routine for therapeutic interventions. However, the analysis and documentation of patients data are still time-consuming and difficult to manage for MTBs, especially as few tools are available for the amount of information required.

RESULTS

To overcome these limitations, we developed an interactive web application AMBAR (Alteration annotations for Molecular tumor BoARds), for therapeutic decision-making support in MTBs. AMBAR is an R shiny-based application that allows customization, interactive filtering, visualization, adding expert knowledge, and export to clinical systems of annotated mutations.

AVAILABILITY

AMBAR is dockerized, open source and available at https://sysbio.uni-ulm.de/?Software:Ambar Contact:hans.kestler@uni-ulm.de.

Strategies of Predictive Schemes and Clinical Diagnosis for Prognosis Using MIMIC-III: A Systematic Review

The prime purpose of the proposed study is to construct a novel predictive scheme for assisting in the prognosis of criticality using the MIMIC-III dataset. With the adoption of various analytics and advanced computing in the healthcare system, there is an increasing trend toward developing an effective prognostication mechanism. Predictive-based modeling is the best alternative to work in this direction. This paper discusses various scientific contributions using desk research methodology towards the Medical Information Mart for Intensive Care (MIMIC-III). This open-access dataset is meant to help predict patient trajectories for various purposes ranging from mortality forecasting to treatment planning. With a dominant machine learning approach in this perspective, there is a need to discover the effectiveness of existing predictive methods. The resultant outcome of this paper offers an inclusive discussion about various available predictive schemes and clinical diagnoses using MIMIC-III in order to contribute toward better information associated with its strengths and weaknesses. Therefore, the paper provides a clear visualization of existing schemes for clinical diagnosis using a systematic review approach.

A privacy-preserving and computation-efficient federated algorithm for generalized linear mixed models to analyze correlated electronic health records data

Large collaborative research networks provide opportunities to jointly analyze multicenter electronic health record (EHR) data, which can improve the sample size, diversity of the study population, and generalizability of the results. However, there are challenges to analyzing multicenter EHR data including privacy protection, large-scale computation resource requirements, heterogeneity across sites, and correlated observations. In this paper, we propose a federated algorithm for generalized linear mixed models (Fed-GLMM), which can flexibly model multicenter longitudinal or correlated data while accounting for site-level heterogeneity. Fed-GLMM can be applied to both federated and centralized research networks to enable privacy-preserving data integration and improve computational efficiency. By communicating a limited amount of summary statistics, Fed-GLMM can achieve nearly identical results as the gold-standard method where the GLMM is directly fitted to the pooled dataset. We demonstrate the performance of Fed-GLMM in numerical experiments and an application to longitudinal EHR data from multiple healthcare facilities.

Healthcare Providers’ Knowledge of Value-Based Care in Germany: An Adapted, Mixed-Methods Approach

Background: Value-Based Care (VBC) is being discussed to provide better outcomes to patients, with an aim to reimburse healthcare providers (HCPs) based on the quality of care they deliver. Little is known about German HCPs’ knowledge of VBC. This study aims to investigate the knowledge of HCPs of VBC and to identify potential needs for further education toward implementation of VBC in Germany. Methods: For evidence generation, we performed a literature search and conducted an online survey among HCPs at 89 hospitals across Germany. The questionnaire was based on published evidence and co-developed with an expert panel using a mixed methods approach. Results: We found HCPs to believe that VBC is more applicable in surgery than internal medicine and that well-defined cycles of care are essential for its application. HCPs believe that VBC can reduce health care costs significantly. However, they also assume that implementing VBC will be challenging. Conclusions: The concept in general is well perceived, however, HCPs do not want to participate in any financial risk sharing. Installing an authority/independent agency that measures achieved value, digital transformation, and that improves the transition between the inpatient and the outpatient sectors are top interests of HCPs.

Patient Empowerment During the COVID-19 Pandemic by Ensuring Safe and Fast Communication of Test Results: Implementation and Performance of a Tracking System

Background

Overcoming the COVID-19 crisis requires new ideas and strategies for online communication of personal medical information and patient empowerment. Rapid testing of a large number of subjects is essential for monitoring and delaying the spread of SARS-CoV-2 in order to mitigate the pandemic’s consequences. People who do not know that they are infected may not stay in quarantine and, thus, risk infecting others. Unfortunately, the massive number of COVID-19 tests performed is challenging for both laboratories and the units that conduct throat swabs and communicate the results.

Objective

The goal of this study was to reduce the communication burden for health care professionals. We developed a secure and easy-to-use tracking system to report COVID-19 test results online that is simple to understand for the tested subjects as soon as these results become available. Instead of personal calls, the system updates the status and the results of the tests automatically. This aims to reduce the delay when informing testees about their results and, consequently, to slow down the virus spread.

Methods

The application in this study draws on an existing tracking tool. With this open-source and browser-based online tracking system, we aim to minimize the time required to inform the tested person and the testing units (eg, hospitals or the public health care system). The system can be integrated into the clinical workflow with very modest effort and avoids excessive load to telephone hotlines.

Results

The test statuses and results are published on a secured webpage, enabling regular status checks by patients; status checks are performed without the use of smartphones, which has some importance, as smartphone usage diminishes with age. Stress tests and statistics show the performance of our software. CTest is currently running at two university hospitals in Germany—University Hospital Ulm and University Hospital Tübingen—with thousands of tests being performed each week. Results show a mean number of 10 (SD 2.8) views per testee.

Conclusions

CTest runs independently of existing infrastructures, aims at straightforward integration, and aims for the safe transmission of information. The system is easy to use for testees. QR (Quick Response) code links allow for quick access to the test results. The mean number of views per entry indicates a reduced amount of time for both health care professionals and testees. The system is quite generic and can be extended and adapted to other communication tasks.

Supporting Medical Staff from Psycho-Oncology with Smart Mobile Devices: Insights into the Development Process and First Results

Cancer is a very distressing disease, not only for the patients themselves, but also for their family members and relatives. Therefore, patients are regularly monitored to decide whether psychological treatment is necessary and applicable. However, such monitoring processes are costly in terms of required staff and time. Mobile data collection is an emerging trend in various domains. The medical and psychological field benefits from such an approach, which enables experts to quickly collect a large amount of individual health data. Mobile data collection applications enable a more holistic view of patients and assist psychologists in taking proper actions. We developed a mobile application, FeelBack, which is designed to support data collection that is based on well-known and approved psychological instruments. A controlled pilot evaluation with 60 participants provides insights into the feasibility of the developed platform and it shows the initial results. 31 of these participants received paper-based questionnaire and 29 followed the digital approach. The results reveal an increase of the overall acceptance by 58.5% in the mean when using a digital screening as compared to the paper-based. We believe that such a platform may significantly improve cancer patients’ and relatives’ psychological treatment, as available data can be used to optimize treatment.

Chained correlations for feature selection

Data-driven algorithms stand and fall with the availability and quality of existing data sources. Both can be limited in high-dimensional settings ($$n \gg m$$ n ≫ m ). For example, supervised learning algorithms designed for molecular pheno- or genotyping are restricted to samples of the corresponding diagnostic classes. Samples of other related entities, such as arise in differential diagnosis, are usually not utilized in this learning scheme. Nevertheless, they might provide domain knowledge on the background or context of the original diagnostic task. In this work, we discuss the possibility of incorporating samples of foreign classes in the training of diagnostic classification models that can be related to the task of differential diagnosis. Especially in heterogeneous data collections comprising multiple diagnostic categories, the foreign ones can change the magnitude of available samples. More precisely, we utilize this information for the internal feature selection process of diagnostic models. We propose the use of chained correlations of original and foreign diagnostic classes. This method allows the detection of intermediate foreign classes by evaluating the correlation between class labels and features for each pair of original and foreign categories. Interestingly, this criterion does not require direct comparisons of the initial diagnostic groups and therefore, might be suitable for settings with restricted data access.

Imprecise Data and Their Impact on Translational Research in Medicine

The medical field expects from big data essentially two main results: the ability to build predictive models and the possibility of applying them to obtain accurate patient risk profiles and/or health trajectories. Note that the paradigm of precision has determined that similar challenges need to be faced in both population and individualized studies, namely the need of assembling, integrating, modeling, and interpreting data from a variety of information sources and scales potentially influencing disease from onset to progression. In many cases, data require computational treatment through solutions for otherwise intractable problems. However, as precision medicine remains subject to a substantial amount of data imprecision and lack of translational impact, a revision of methodological inference approaches is needed. Both the relevance and the usefulness of such revision crucially deal with the assimilation of data features dynamically interconnected.

Constraining classifiers in molecular analysis: invariance and robustness

Analysing molecular profiles requires the selection of classification models that can cope with the high dimensionality and variability of these data. Also, improper reference point choice and scaling pose additional challenges. Often model selection is somewhat guided by ad hoc simulations rather than by sophisticated considerations on the properties of a categorization model. Here, we derive and report four linked linear concept classes/models with distinct invariance properties for high-dimensional molecular classification. We can further show that these concept classes also form a half-order of complexity classes in terms of Vapnik-Chervonenkis dimensions, which also implies increased generalization abilities. We implemented support vector machines with these properties. Surprisingly, we were able to attain comparable or even superior generalization abilities to the standard linear one on the 27 investigated RNA-Seq and microarray datasets. Our results indicate that a priori chosen invariant models can replace ad hoc robustness analysis by interpretable and theoretically guaranteed properties in molecular categorization.