Subpopulation-specific machine learning prognosis for underrepresented patients with double prioritized bias correction.

Background: Many clinical datasets are intrinsically imbalanced, dominated by overwhelming majority groups. Off-the-shelf machine learning models that optimize the prognosis of majority patient types (e.g., healthy class) may cause substantial errors on the minority prediction class (e.g., disease class) and demographic subgroups (e.g., Black or young patients). In the typical one-machine-learning-model-fits-all paradigm, racial and age disparities are likely to exist, but unreported. In addition, some widely used whole-population metrics give misleading results. Methods: We design a double prioritized (DP) bias correction technique to mitigate representational biases in machine learning-based prognosis. Our method trains customized machine learning models for specific ethnicity or age groups, a substantial departure from the one-model-predicts-all convention. We compare with other sampling and reweighting techniques in mortality and cancer survivability prediction tasks. Results: We first provide empirical evidence showing various prediction deficiencies in a typical machine learning setting without bias correction. For example, missed death cases are 3.14 times higher than missed survival cases for mortality prediction. Then, we show DP consistently boosts the minority class recall for underrepresented groups, by up to 38.0%. DP also reduces relative disparities across race and age groups, e.g., up to 88.0% better than the 8 existing sampling solutions in terms of the relative disparity of minority class recall. Cross-race and cross-age-group evaluation also suggests the need for subpopulation-specific machine learning models. Conclusions: Biases exist in the widely accepted one-machine-learning-model-fits-all-population approach. We invent a bias correction method that produces specialized machine learning prognostication models for underrepresented racial and age groups. This technique may reduce potentially life-threatening prediction mistakes for minority populations.

Antibiotics in the Community During the COVID-19 Pandemic: A Qualitative Study to Understand Users' Perspectives of Antibiotic Seeking and Consumption Behaviors in Bangladesh.

INTRODUCTION: The COVID-19 pandemic is thought to have led to increased "inappropriate" or "unjustified" seeking and consumption of antibiotics by individuals in the community. However, little reference has been made to antibiotic seeking and using behaviors from the perspectives of users in Bangladesh during this health crisis. PURPOSE: This study seeks to document how antibiotic medicines are sought and used during a complex health crisis, and, within different contexts, what are the nuanced reasons why patients may utilize these medicines sub-optimally. METHODS: We used an exploratory, qualitative design. Forty semi-structured telephone interviews were conducted with people diagnosed with COVID-19 (n=20), who had symptoms suggestive of COVID-19 (n=20), and who had received care at home in two cities between May and June 2021 in Bangladesh. In this study, an inductive thematic analysis was performed. RESULTS: The analysis highlighted the interlinked relationships of antibiotic seeking and consumption behaviors with the diversity of information disseminated during a health crisis. Antibiotic-seeking behaviors are related to previous experience of use, perceived severity of illness, perceived vulnerability, risk of infection, management of an "unknown" illness and anxiety, distrust of expert advice, and intrinsic agency on antimicrobial resistance (AMR). Suboptimal adherence, such as modifying treatment regimes and using medication prescribed for others, were found to be part of care strategies used when proven therapeutics were unavailable to treat COVID-19. Early cessation of therapy was found to be a rational practice to avoid side effects and unknown risks. CONCLUSION: Based on the results, we highly recommend the take up of a pandemic specific antimicrobial stewardship (AMS) program in the community. To deliver better outcomes of AMS, incorporating users' perspectives could be a critical strategy. Therefore, a co-produced AMS intervention that is appropriate for a specific cultural context is an essential requirement to reduce the overuse of antibiotics during the COVID-19 pandemic and beyond.

Knowledge, Attitudes, and Common Practices of Livestock and Poultry Veterinary Practitioners Regarding the AMU and AMR in Bangladesh.

Current evidence indicates that more than half of all antimicrobials are used in the animal food-producing sector, which is considered a significant risk factor for the development, spread, and existence of antimicrobial resistance (AMR) pathogens in animals, humans, and the environment. Among other factors, clinical etiology and the level of knowledge, attitudes, and practices (KAP) of veterinarians are thought to be responsible for inappropriate prescriptions in the animal-source protein production sector in lower-resource settings like Bangladesh. We performed this cross-sectional study to assess factors associated with veterinarians' antimicrobial prescription behavior and their KAP on antimicrobial use (AMU) and AMR in Bangladesh. Exploratory and multivariate logistic models were used to describe an association between knowledge, attitudes, and practices of AMU and AMR and demographic characteristics of veterinarians. The results demonstrated that when selecting an antimicrobial, there was no to minimal influence of culture and susceptibility tests and patients' AMU history but moderate to high influence of the farmer's economic condition and drug instructions among the veterinarians. The results also demonstrated that more than half of the veterinarians had correct KAP regarding AMU and AMR, while the rest had moderate or lower levels of KAP. The factor score analysis revealed that age, level of education, years of experience, gender, and previous training on AMU and AMR were the key influencing factors in their level of KAP. Adjusted logistic regression analysis showed that respondents' age, current workplace, and previous training on AMU and AMR had a positive association with increased KAP. Considering the results, it is imperative to include AMR issues on vet curricula, and to provide post-education training, awareness campaigns, easy access to, and dissemination of AMR resources. Increasing the veterinary services to the outreach areas of the country and motivating veterinarians to follow the national AMR guidelines could be some other potential solutions to tackle the over-prescriptions of antimicrobials.