Trustworthy artificial intelligence and the European Union AI act: On the conflation of trustworthiness and acceptability of risk.

In its AI Act, the European Union chose to understand trustworthiness of AI in terms of the acceptability of its risks. Based on a narrative systematic literature review on institutional trust and AI in the public sector, this article argues that the EU adopted a simplistic conceptualization of trust and is overselling its regulatory ambition. The paper begins by reconstructing the conflation of "trustworthiness" with "acceptability" in the AI Act. It continues by developing a prescriptive set of variables for reviewing trust research in the context of AI. The paper then uses those variables for a narrative review of prior research on trust and trustworthiness in AI in the public sector. Finally, it relates the findings of the review to the EU's AI policy. Its prospects to successfully engineer citizen's trust are uncertain. There remains a threat of misalignment between levels of actual trust and the trustworthiness of applied AI.

Ferric Carboxymaltose in Iron-Deficient Patients with Hospitalized Heart Failure and Reduced Kidney Function.

BACKGROUND: Reduced kidney function is common among patients with heart failure. In patients with heart failure and/or kidney disease, iron deficiency is an independent predictor of adverse outcomes. In the AFFIRM-AHF trial, patients with acute heart failure with iron deficiency treated with intravenous ferric carboxymaltose demonstrated reduced risk of heart failure hospitalization, with improved quality of life. We aimed to further characterize the impact of ferric carboxymaltose among patients with coexisting kidney impairment. METHODS: The double-blind, placebo-controlled AFFIRM-AHF trial randomized 1132 stabilized adults with acute heart failure (left ventricular ejection fraction <50%) and iron deficiency. Patients on dialysis were excluded. The primary end point was a composite of total heart failure hospitalizations and cardiovascular death during the 52-week follow-up period. Additional end points included cardiovascular hospitalizations, total heart failure hospitalizations, and days lost to heart failure hospitalizations or cardiovascular death. For this subgroup analysis, patients were stratified according to baseline eGFR. RESULTS: Overall, 60% of patients had an eGFR <60 ml/min per 1.73 m 2 (the lower eGFR subgroup). These patients were significantly older, more likely to be female and to have ischemic heart failure, and had higher baseline serum phosphate levels and higher rates of anemia. For all end points, event rates were higher in the lower eGFR group. In the lower eGFR group, the annualized event rates for the primary composite outcome were 68.96 and 86.30 per 100 patient-years in the ferric carboxymaltose and placebo arms, respectively (rate ratio, 0.76; 95% confidence interval, 0.54 to 1.06). The treatment effect was similar in the higher eGFR subgroup (rate ratio, 0.65; 95% confidence interval, 0.42 to 1.02; Pinteraction = 0.60). A similar pattern was observed for all end points ( Pinteraction > 0.05). CONCLUSIONS: In a cohort of patients with acute heart failure, left ventricular ejection fraction <50%, and iron deficiency, the safety and efficacy of ferric carboxymaltose were consistent across a range of eGFR values. CLINICAL TRIAL REGISTRY NAME AND REGISTRATION NUMBER: Study to Compare Ferric Carboxymaltose With Placebo in Patients With Acute Heart Failure and Iron Deficiency (Affirm-AHF), NCT02937454 .

Association Between Hemoglobin Levels and Efficacy of Intravenous Ferric Carboxymaltose in Patients With Acute Heart Failure and Iron Deficiency: An AFFIRM-AHF Subgroup Analysis.

BACKGROUND: Iron deficiency, with or without anemia, is an adverse prognostic factor in heart failure (HF). In AFFIRM-AHF (a randomized, double-blind placebo-controlled trial comparing the effect of intravenous ferric carboxymaltose on hospitalizations and mortality in iron-deficient subjects admitted for acute heart failure), intravenous ferric carboxymaltose (FCM), although having no significant effect on the primary end point, reduced the risk of HF hospitalization (hHF) and improved quality of life versus placebo in iron-deficient patients stabilized after an acute HF (AHF) episode. These prespecified AFFIRM-AHF subanalyses explored the association between hemoglobin levels and FCM treatment effects. METHODS: AFFIRM-AHF was a multicenter, double-blind, randomized, placebo-controlled trial of FCM in hospitalized AHF patients with iron deficiency. Patients were stratified by baseline hemoglobin level (<12 versus ≥12 g/dL). In each subgroup, the primary composite (total hHF and cardiovascular death) and secondary (total hHF; total cardiovascular hospitalizations and cardiovascular death; time to cardiovascular death, and time to first/days lost due to hHF or cardiovascular death) outcomes were assessed with FCM versus placebo at week 52. Sensitivity analyses using the World Health Organization anemia definition (hemoglobin level <12 g/dL [women] or <13 g/dL [men]) were performed, among others. RESULTS: Of 1108 AFFIRM-AHF patients, 1107 were included in these subanalyses: 464 (FCM group, 228; placebo group, 236) had a hemoglobin level <12 g/dL, and 643 (FCM, 329; placebo, 314) had a hemoglobin level ≥12 g/dL. Patients with a hemoglobin level <12 g/dL were older (mean, 73.7 versus 69.1 years), with more frequent previous HF (75.0% versus 68.7%), serum ferritin <100 µg/L (75.4% versus 68.1%), and transferrin saturation <20% (87.9% versus 81.4%). For the primary outcome, annualized event rates per 100 patient-years with FCM versus placebo were 71.1 and 73.6 (rate ratio, 0.97 [95% CI, 0.66-1.41]), respectively, and 48.5 versus 72.9 (RR, 0.67 [95% CI, 0.48-0.93]) in the hemoglobin levels <12 and ≥12 g/dL subgroups, respectively. No significant interactions between hemoglobin subgroup and treatment effect were observed for primary (Pinteraction=0.15) or secondary outcomes. Changes from baseline in hemoglobin, serum ferritin and transferrin saturation were significantly greater with FCM versus placebo in both subgroups between weeks 6 and 52. Findings were similar using the World Health Organization definition for anemia. CONCLUSIONS: The effects of intravenous FCM on outcomes in iron-deficient patients stabilized after an AHF episode, including improvements in iron parameters over time, did not differ between patients with hemoglobin levels <12 and ≥12 g/dL. REGISTRATION: URL: https://www. CLINICALTRIALS: gov; Unique identifier: NCT02937454.

Impact of ferric carboxymaltose for iron deficiency at discharge after heart failure hospitalization: a European multinational economic evaluation.

AIMS: Iron deficiency (ID) is comorbid in up to 50% patients with heart failure (HF) and exacerbates disease burden. Ferric carboxymaltose (FCM) reduced HF hospitalizations and improved quality of life when used to treat ID at discharge in patients hospitalized for acute HF with left ventricular ejection fraction <50% in the AFFIRM-AHF trial. We quantified the effect of FCM on burden of disease and the wider pharmacoeconomic implications in France, Germany, Poland, Spain and Sweden. METHODS AND RESULTS: The per country eligible population was calculated, aligning with the 2021 European Society of Cardiology (ESC) HF guidelines and the AFFIRM-AHF trial. Changes in burden of disease with FCM versus standard of care (SoC) were represented by disability-adjusted life years (DALYs), hospitalization episodes and bed days, using AFFIRM-AHF data. A Markov model was adapted to each country to estimate cost-effectiveness and combined with epidemiology data to calculate the impact on healthcare budgets. Between 335 (Sweden) and 13 237 (Germany) DALYs were predicted to be avoided with FCM use annually. Fewer hospitalizations and shorter lengths of stay associated with FCM compared to SoC were projected to result in substantial annual savings in bed days, from 5215 in Sweden to 205 630 in Germany. In all countries, FCM was predicted to be dominant (cost saving with gains in quality-adjusted life years), resulting in net savings to healthcare budgets within 1 year. CONCLUSIONS: This comprehensive evaluation of FCM therapy highlights the potential benefits that could be realized through implementation of the ESC HF guideline recommendations regarding ID treatment.

Incidence, Treatment and Clinical Impact of Iron Deficiency in Chronic Heart Failure: A Longitudinal Analysis.

In patients with heart failure (HF), iron deficiency (ID) is a well-recognized therapeutic target; information about its incidence, patterns of iron repletion, and clinical impact is scarce. This single-centre longitudinal cohort study assessed the rates of ID testing and diagnosis in patients with stable HF, patterns of treatment with intravenous iron, and clinical impact of intravenous iron on HF rehospitalization risk. We included 711 consecutive outpatients (4400 visits) with stable chronic HF from 2014 to 2019 (median [interquartile range] visits per patient: 2 [2−7]. ID was defined as serum ferritin <100 µg/L, or 100−299 µg/L with transferrin saturation (TSAT) < 20%. During a median follow-up of 2.20 (1.11−3.78) years, ferritin and TSAT were measured at 2230 (50.7%) and 2183 visits (49.6%), respectively. ID was found at 846 (37.9%) visits, with ferritin and TSAT available (2230/4400), and intravenous iron was administered at 321/4400 (7.3%) visits; 233 (32.8%) patients received intravenous iron during follow-up. After multivariate analyses, iron repletion at any time during follow-up was associated with a lower risk of recurrent HF hospitalization (hazard ratio [HR] = 0.50, 95% confidence interval [CI] = 0.28−0.88; p = 0.016). Thus, ID was a frequent finding in patients with HF, and its repletion reduced the risk of recurrent HF hospitalizations.

Artificial Intelligence and the 'Good Society': the US, EU, and UK approach.

In October 2016, the White House, the European Parliament, and the UK House of Commons each issued a report outlining their visions on how to prepare society for the widespread use of artificial intelligence (AI). In this article, we provide a comparative assessment of these three reports in order to facilitate the design of policies favourable to the development of a 'good AI society'. To do so, we examine how each report addresses the following three topics: (a) the development of a 'good AI society'; (b) the role and responsibility of the government, the private sector, and the research community (including academia) in pursuing such a development; and (c) where the recommendations to support such a development may be in need of improvement. Our analysis concludes that the reports address adequately various ethical, social, and economic topics, but come short of providing an overarching political vision and long-term strategy for the development of a 'good AI society'. In order to contribute to fill this gap, in the conclusion we suggest a two-pronged approach.

Transparent, explainable, and accountable AI for robotics.

To create fair and accountable AI and robotics, we need precise regulation and better methods to certify, explain, and audit inscrutable systems.

Predictors and impact of non-adherence in adults with attention-deficit/hyperactivity disorder receiving OROS methylphenidate: results from a randomized, placebo-controlled trial.

BACKGROUND: Medication non-adherence has an important impact on treatment efficacy and healthcare burden across a range of conditions and therapeutic areas. The aim of this analysis was to determine predictors of non-adherence and impact of non-adherence on treatment response in adults with attention-deficit/hyperactivity disorder (ADHD). METHODS: Post-hoc analysis of a 13-week randomized, double-blind placebo-controlled study of OROS methylphenidate (MPH) 54 and 72 mg/day. Primary efficacy variable was the Conners' Adult ADHD Rating Scale - Screening Version (CAARS:O-SV). Daily adherence was calculated as average daily adherence (100 × capsules taken/2), with overall adherence calculated as the average daily adherence. Predictors of adherence were assessed using mixed-effects logistic regression. Descriptive statistics were generated for change in CAARS:O-SV score for adherent (> 95% adherence) and non-adherent subjects. Predictors of change were analyzed using a mixed model. RESULTS: Subjects were allocated to OROS MPH (54 mg, n = 87; 72 mg, n = 92) or placebo (n = 97). Mean adherence was 92.6% and 93.3% (OROS MPH 54 and 72 mg/day, respectively), versus 97.5% (placebo). Adherence was higher and less variable in completers. Factors significantly associated with non-adherence included female sex, shorter time since ADHD diagnosis, higher education level (completion of university) and score on the Drug Use Screening Inventory psychiatric disorders subscale. Improvements from baseline in CAARS:O-SV score were numerically greater in subjects defined as adherent than in those who were non-adherent. Significant predictors of CAARS:O-SV change in patients who completed the study included percentage adherence up to the point of assessment (p < 0.0001), baseline score (p < 0.0001) and family history of ADHD (p = 0.0003). CONCLUSION: The results of this analysis suggest that newly diagnosed patients, those with a high score on the DUSI-R psychiatric disorder scale, women, and subjects with high educational degrees may be at increased risk of non-adherence. Clinicians and policymakers should therefore pay special attention to these individuals, as non-adherence is a significant predictor of reduced response to treatment. TRIAL REGISTRATION: EudraCT #: 2007-002111-82.

Long-term administration of L-carnitine to humans: effect on skeletal muscle carnitine content and physical performance.

BACKGROUND: Long-term administration of high oral doses of L-carnitine on the skeletal muscle composition and the physical performance has not been studied in humans. METHODS: Eight healthy male adults were treated with 2 x 2 g of L-carnitine per day for 3 months. Muscle biopsies and exercise tests were performed before, immediately after, and 2 months after the treatment. Exercise tests were performed using a bicycle ergometer for 10 min at 20%, 40%, and 60% of the individual maximal workload (P(max)), respectively, until exhaustion. RESULTS: There were no significant differences between V(O(2)max), RER(max), and P(max) between the three time points investigated. At submaximal intensities, the only difference to the pretreatment values was a 5% increase in V(O(2)) at 20% and 40% of P(max) 2 months after the cessation of the treatment. The total carnitine content in the skeletal muscle was 4.10 +/- 0.82 micromol/g before, 4.79 +/- 1.19 micromol/g immediately after, and 4.19 +/- 0.61 micromol/g wet weight 2 months after the treatment (no significant difference). Activities of the two mitochondrial enzymes citrate synthase and cytochrome oxidase, as well as the skeletal muscle fiber composition also remained unaffected by the administration of L-carnitine. CONCLUSIONS: Long-term oral treatment of healthy adults with L-carnitine is not associated with a significant increase in the muscle carnitine content, mitochondrial proliferation, or physical performance. Beneficial effects of the long-term treatment with L-carnitine on the physical performance of healthy adults cannot be explained by an increase in the carnitine muscle stores.