Background: Resources are increasingly spent on artificial intelligence (AI) solutions for medical applications aiming to improve diagnosis, treatment, and prevention of diseases. While the need for transparency and reduction of bias in data and algorithm development has been addressed in past studies, little is known about the knowledge and perception of bias among AI developers. Objective: This study’s objective was to survey AI specialists in health care to investigate developers’ perceptions of bias in AI algorithms for health care applications and their awareness and use of preventative measures. Methods: A web-based survey was provided in both German and English language, comprising a maximum of 41 questions using branching logic within the REDCap web application. Only the results of participants with experience in the field of medical AI applications and complete questionnaires were included for analysis. Demographic data, technical expertise, and perceptions of fairness, as well as knowledge of biases in AI, were analyzed, and variations among gender, age, and work environment were assessed. Results: A total of 151 AI specialists completed the web-based survey. The median age was 30 (IQR 26-39) years, and 67% (101/151) of respondents were male. One-third rated their AI development projects as fair (47/151, 31%) or moderately fair (51/151, 34%), 12% (18/151) reported their AI to be barely fair, and 1% (2/151) not fair at all. One participant identifying as diverse rated AI developments as barely fair, and among the 2 undefined gender participants, AI developments were rated as barely fair or moderately fair, respectively. Reasons for biases selected by respondents were lack of fair data (90/132, 68%), guidelines or recommendations (65/132, 49%), or knowledge (60/132, 45%). Half of the respondents worked with image data (83/151, 55%) from 1 center only (76/151, 50%), and 35% (53/151) worked with national data exclusively. Conclusions: This study shows that the perception of biases in AI overall is moderately fair. Gender minorities did not once rate their AI development as fair or very fair. Therefore, further studies need to focus on minorities and women and their perceptions of AI. The results highlight the need to strengthen knowledge about bias in AI and provide guidelines on preventing biases in AI health care applications.

The European project ORCHESTRA intends to create a new pan-European cohort to rapidly advance the knowledge of the effects and treatment of COVID-19. Establishing processes that facilitate the merging of heterogeneous clusters of retrospective data was an essential challenge. In addition, data from new ORCHESTRA prospective studies have to be compatible with earlier collected information to be efficiently combined. In this article, we describe how we utilized and contributed to existing standard terminologies to create consistent semantic representation of over 2500 COVID-19-related variables taken from three ORCHESTRA studies. The goal is to enable the semantic interoperability of data within the existing project studies and to create a common basis of standardized elements available for the design of new COVID-19 studies. We also identified 743 variables that were commonly used in two of the three prospective ORCHESTRA studies and can therefore be directly combined for analysis purposes. Additionally, we actively contributed to global interoperability by submitting new concept requests to the terminology Standards Development Organizations.

On May 3rd, 2022, the European Commission published its legislative proposal to create a European Health Data Space (EHDS) enabling citizens of the European Union to gain secure access to their electronic health data by establishing a market for digital health. This market will feature the primary and secondary use of electronic health records by digital products and services. The articles of the proposal address many aspects of ensuring health data interoperability. That includes the creation of a European Electronic Health Record Exchange Format for defined data categories including patient summaries and electronic prescriptions, the development of a central platform to provide a cross-border digital infrastructure and that each Member State institutes a digital health authority and a national point of contact. In addition, the Commission will define common specifications that electronic health record systems and medical devices will have to meet as interoperability requirements. In its current form, the proposal does not stipulate specific standards that need to be universally adopted to ensure semantic and syntactical interoperability. Considering that many datasets are not internationally harmonized and lack standardization, these specifications will need to be provided for example by existing standards like the International Patient Summary.

The objective was to develop a dataset definition, information model, and FHIR^® specification for key data elements contained in a German molecular genomics (MolGen) report to facilitate genomic and phenotype integration in electronic health records. A dedicated expert group participating in the German Medical Informatics Initiative reviewed information contained in MolGen reports, determined the key elements, and formulated a dataset definition. HL7’s Genomics Reporting Implementation Guide (IG) was adopted as a basis for the FHIR^® specification which was subjected to a public ballot. In addition, elements in the MolGen dataset were mapped to the fields defined in ISO/TS 20428:2017 standard to evaluate compliance. A core dataset of 76 data elements, clustered into 6 categories was created to represent all key information of German MolGen reports. Based on this, a FHIR specification with 16 profiles, 14 derived from HL7^®’s Genomics Reporting IG and 2 additional profiles (of the FamilyMemberHistory and RiskAssessment resources), was developed. Five example resource bundles show how our adaptation of an international standard can be used to model MolGen report data that was requested following oncological or rare disease indications. Furthermore, the map of the MolGen report data elements to the fields defined by the ISO/TC 20428:2017 standard, confirmed the presence of the majority of required fields. Our report serves as a template for other research initiatives attempting to create a standard format for unstructured genomic report data. Use of standard formats facilitates integration of genomic data into electronic health records for clinical decision support.

ORCHESTRA (“Connecting European Cohorts to Increase Common and Effective Response To SARS-CoV-2 Pandemic”) is an EU-funded project which aims to help rapidly advance the knowledge related to the prevention of the SARS-CoV-2 infection and the management of COVID-19 and its long-term sequelae. Here, we describe the early results of this project, focusing on the strengths of multiple, international, historical and prospective cohort studies and highlighting those results which are of potential relevance for vaccination strategies, such as the necessity of a vaccine booster dose after a primary vaccination course in hematologic cancer patients and in solid organ transplant recipients to elicit a higher antibody titer, and the protective effect of vaccination on severe COVID-19 clinical manifestation and on the emergence of post-COVID-19 conditions. Valuable data regarding epidemiological variations, risk factors of SARS-CoV-2 infection and its sequelae, and vaccination efficacy in different subpopulations can support further defining public health vaccination policies.

Glycosaminoglycans are long polysaccharidic chains, which are mostly present in connective tissues. Modified GAG expression in tissues surrounding malignant cells has been shown to contribute to tumor progression, aggressive status and metastasis in many types of cancer. Ovarian cancer is one of the most lethal gynecological malignancies due to its late diagnosis because of the absence of clear symptoms and unavailability of early disease markers. We investigated for the first time GAG changes at the molecular level as a novel biomarker for primary epithelial ovarian cancer. To this end, serum of a cohort of 68 samples was digested with chondroitinase ABC, which releases chondroitin sulfate into disaccharides. After labeling and purification, they were measured by HPLC, yielding a profile of eight disaccharides. We proposed a novel GAG-based score named “CS- bio” from the measured abundance of disaccharides present that were of statistical relevance. CS-bio’s performance was compared with CA125, the clinically used serum tumor marker in routine diagnostics. CS-bio had a better sensitivity and specificity than CA125. It was more apt in differentiating early-stage patients from healthy controls, which is of high interest for oncologists.