There has been a surge of interest in artificial intelligence and machine learning (AI/ML)-based medical devices. However, it is poorly understood how and which AI/ML-based medical devices have been approved in the USA and Europe. We searched governmental and non-governmental databases to identify 222 devices approved in the USA and 240 devices in Europe. The number of approved AI/ML-based devices has increased substantially since 2015, with many being approved for use in radiology. However, few were qualified as high-risk devices. Of the 124 AI/ML-based devices commonly approved in the USA and Europe, 80 were first approved in Europe. One possible reason for approval in Europe before the USA might be the potentially relatively less rigorous evaluation of medical devices in Europe. The substantial number of approved devices highlight the need to ensure rigorous regulation of these devices. Currently, there is no specific regulatory pathway for AI/ML-based medical devices in the USA or Europe. We recommend more transparency on how devices are regulated and approved to enable and improve public trust, efficacy, safety, and quality of AI/ML-based medical devices. A comprehensive, publicly accessible database with device details for Conformité Européene (CE)-marked medical devices in Europe and US Food and Drug Administration approved devices is needed.

No abstract

As COVID-19 spreads across the globe, crowdsourced digital technology harbours the potential to improve surveillance and epidemic control, primarily through increased information coverage, higher information speed, fast case tracking and improved proximity tracing. Targeting those aims, COVID-19-related smartphone and webbased health applications are continuously emerging, leading to a multitude of options, raising ethical and legal challenges and potentially overwhelming end users.Building on an existing trustworthiness checklist for digital health applications, we searched the literature and developed a framework to guide the assessment of smartphone and web-based applications that aim to contribute to controlling the current epidemic or mitigating its effects. It further integrates epidemiological subject knowledge and a legal analysis, outlining the mechanisms through which new applications can support the fight against COVID-19.The resulting framework includes 40 questions across 8 domains on "purpose", "usability", "information accuracy", "organisational attributes / reputation", "transparency", "privacy" and "user control / self-determination". All questions should be primarily answerable from publicly available data, as provided by application manufacturers. The framework aims to guide end users in choosing a transparent, safe and valuable application and suggests a set of information items that developers ideally make available to allow a balanced judgement and facilitate the trustworthiness of their products.

ObjectiveTo characterize the therapeutic value of new drugs approved by the US Food and Drug Administration (FDA) and European Medicines Agency (EMA) and the association between these ratings and regulatory approval through expedited programs.DesignRetrospective cohort study.SettingNew drugs approved by the FDA and EMA between 2007 and 2017, with follow-up through 1 April 2020.Data sourcesTherapeutic value was measured using ratings of new drugs by five independent organizations (Prescrire and health authorities of Canada, France, Germany, and Italy).Main outcome measuresProportion of new drugs rated as having high therapeutic value; association between high therapeutic value rating and expedited status.ResultsFrom 2007 through 2017, the FDA and EMA approved 320 and 268 new drugs, respectively, of which 181 (57%) and 39 (15%) qualified for least one expedited program. Among 267 new drugs with a therapeutic value rating, 84 (31%) were rated as having high therapeutic value by at least one organization. Compared with non-expedited drugs, a greater proportion of expedited drugs were rated as having high therapeutic value among both FDA approvals (45% (69/153) v 13% (15/114); P<0.001) and EMA approvals (67% (18/27) v 27% (65/240); P<0.001). The sensitivity and specificity of expedited program for a drug being independently rated as having high therapeutic value were 82% (95% confidence interval 72% to 90%) and 54% (47% to 62%), respectively, for the FDA, compared with 25.3% (16.4% to 36.0%) and 90.2% (85.0% to 94.1%) for the EMA.ConclusionsLess than a third of new drugs approved by the FDA and EMA over the past decade were rated as having high therapeutic value by at least one of five independent organizations. Although expedited drugs were more likely than non-expedited drugs to be highly rated, most expedited drugs approved by the FDA but not the EMA were rated as having low therapeutic value.

Several sources of bias can affect the performance of machine learning systems used in medicine and potentially impact clinical care. Here, we discuss solutions to mitigate bias across the different development steps of machine learning-based systems for medical applications.

No abstract

No abstract

The high cost of cancer medicines is a public health challenge. Policy makers in the US and Europe are debating reforms to drug pricing that would cover both the prices of new medicines when entering the market and price increases after they are launched.OBJECTIVE To assess launch prices, postlaunch price changes, and clinical benefit of cancer drugs in the US compared with 3 European countries (England, Germany, and Switzerland). DESIGN, SETTING, AND PARTICIPANTSThis economic evaluation identified all new drugs that were approved for use in the US, England, Germany, and Switzerland with initial indications for treatment of adult solid tumor and hematologic cancers. Analysis included drugs approved by the US Food and Drug Administration between January 1, 2009, and December 31, 2019, and by the European Medicines Agency and Swissmedic until December 31, 2019. Prices were adjusted for currency and inflation. Clinical benefit of drugs indicated for solid tumors was assessed using the American Society of Clinical Oncology Value Framework and European Society for Medical Oncology Magnitude of Clinical Benefit Scale. Using Spearman rank correlation coefficients, correlations between clinical benefit and launch prices and postlaunch price changes for each country were evaluated.MAIN OUTCOMES AND MEASURES Launch prices, postlaunch price changes, and clinical benefit of cancer drugs. RESULTSThe cohort included 65 drugs: 47 (72%) approved for solid tumors and 18 (28%) for hematologic cancers. In all countries, the lowest median monthly treatment costs at launch