Summary Artificial intelligence (AI) has an astonishing potential in assisting clinical decision making and revolutionizing the field of health care. A major open challenge that AI will need to address before its integration in the clinical routine is that of algorithmic bias. Most AI algorithms need big datasets to learn from, but several groups of the human population have a long history of being absent or misrepresented in existing biomedical datasets. If the training data is misrepresentative of the population variability, AI is prone to reinforcing bias, which can lead to fatal outcomes, misdiagnoses, and lack of generalization. Here, we describe the challenges in rendering AI algorithms fairer, and we propose concrete steps for addressing bias using tools from the field of open science.
43 Background: Screening for prostate cancer has been and remains contentious with evidence of harms and benefits finely balanced and inadequate to tip the see saw one way or another in order to give men clear, understandable guidance on PSA testing. In recent years the diagnostic pathway has changed significantly delivering a modern pathway that is likely to do far less harm and possibly more benefit. Our work quantifies the levels of harm and benefit at each stage of the entire pathway currently used in the UK from initial PSA through to biopsy-confirmed diagnosis of prostate cancer. Methods: Combining the full results of the most recent high quality diagnosis clinical trials, as well as real world evidence and practice patterns we have mapped each stage of the UK diagnostic pathway. Through this research we are able to quantify the number of men benefitting (i.e. those who move through the pathway correctly with an eventual diagnosis of clinically significant prostate cancer) and the number of men suffering harm (those who suffer biopsy side effects or who are subjected to unnecessary testing as they move through the pathway with no clinically significant cancer). Results: Initial results using clinical trial data and current practice patterns in the UK indicate that 10,000 men entering the diagnostic pathway with a PSA test would generate the following results. Diagnostic outcomes: 463 men would be diagnosed with a clinically significant (Gleason 3+4 or above) localised prostate cancer; 110 men would be diagnosed with a clinically insignificant (Gleason 3+3) prostate cancer; and 646 men would be taken through the entire diagnostic pathway with no diagnosis of prostate cancer. Harm: a maximum of 38 men would suffer clinical side effects from their biopsy. Healthcare costs: 10,000 PSA tests, 1,449 multiparametric MRI scans, and 1,207 biopsies would yield 463 significant cancers diagnosed. An estimated 1,258 clinically insignificant and 168 clinically significant cancers would remain undiagnosed using the current diagnostic pathway due to initial PSA below 3ng/ml. Conclusions: Rates of biopsy-related harm in modern practice are encouragingly low and likely to reduce further. A PSA-initiated diagnostic pathway misses a significant proportion of clinically important cancers. It is interesting to note the high conversion from MRI to biopsy in the data underpinning this model. Next steps will include analysis of the same outcomes in several centres in the UK to determine if actual practice matches expected outcomes (harms and benefits) from previous clinical trials.
Background: Self-medication is defined as the use of drugs to treat self-diagnosed symptoms without the supervision of healthcare physicians. Self-medication is a growing public health phenomenon and is associated with risks such as misdiagnosis and drug toxicity. This study aimed to identify the patterns associated with the practice of self-medication among university students in San José, Costa Rica. Methods: A descriptive cross-sectional study was designed and conducted to identify variables associated. Information was collected on the conditions treated, medications used and their sources. Results: The study found that self-medication is highly prevalent among Costa Rican university students. 91.4% of the sample reported self-medicating, with each student consuming an average of 2.15, ± 1,08 drugs. The most Frequently used active ingredients were paracetamol and Ibuprofen. Results also show a relation between the most consumed types of drugs and the principal causes of drug intoxication reported by Costa Rica’s National Poison Center. 77.8% of the participants considered self-medication a risky practice. Conclusions: Self-medication is common among Costa Rican university students. The prevalence found is higher than that reported in previous studies conducted in the country. These findings suggest the need to implement prevention campaigns and regulatory policies to ensure the safe consumption of medical drugs.
Aim: Explore UK prostate cancer patients’ experiences and preferences for in-person and remote consultations. Materials & methods: In January–March 2021, patients completed a survey of consultation format preferences. Results: Of 971 patients, most preferred in-person consultations when receiving diagnosis and results (92.3% and 66.5%, respectively) and discussing first and further treatment options (92.0 and 84.0%, respectively). Fewer patients considered follow-up (40.9%) or side effect consultations (47.7%) should be in person. Patients with longer travel preferred telephone consultations for receiving test results post-treatment. Patients over 55 preferred in-person consultations for discussing first treatment. Conclusion: To optimize prostate cancer care in the wake of COVID-19, we recommend patients have the option of consultation format, although key decisions should be made in person.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
hi@scite.ai
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.