Potential and limitations of machine meta-learning (ensemble) methods for predicting COVID-19 mortality in a large inhospital Brazilian dataset.

The majority of early prediction scores and methods to predict COVID-19 mortality are bound by methodological flaws and technological limitations (e.g., the use of a single prediction model). Our aim is to provide a thorough comparative study that tackles those methodological issues, considering multiple techniques to build mortality prediction models, including modern machine learning (neural) algorithms and traditional statistical techniques, as well as meta-learning (ensemble) approaches. This study used a dataset from a multicenter cohort of 10,897 adult Brazilian COVID-19 patients, admitted from March/2020 to November/2021, including patients [median age 60 (interquartile range 48-71), 46% women]. We also proposed new original population-based meta-features that have not been devised in the literature. Stacking has shown to achieve the best results reported in the literature for the death prediction task, improving over previous state-of-the-art by more than 46% in Recall for predicting death, with AUROC 0.826 and MacroF1 of 65.4%. The newly proposed meta-features were highly discriminative of death, but fell short in producing large improvements in final prediction performance, demonstrating that we are possibly on the limits of the prediction capabilities that can be achieved with the current set of ML techniques and (meta-)features. Finally, we investigated how the trained models perform on different hospitals, showing that there are indeed large differences in classifier performance between different hospitals, further making the case that errors are produced by factors that cannot be modeled with the current predictors.

Prevention and Intervention in the Risk of Suicide in the Elderly

The risk of suicide in the elderly is higher than in the general population. Therefore, we aim to provide information that supports nursing decision-making in assessing this risk. Methodologically, we conducted a literature review. We analysed nine articles that supported the model developed in the flowchart divided into three phases: anamnesis, risk analysis and nursing interventions according to weighted risk. As a consequence of COVID-19, it is believed that there is an increase in suffering and the risk of suicide, making it pertinent to disseminate information of quick consultation that supports decision-making in nursing.

Anxiety and Depression in Patients with Prostate Cancer, at Cancer Diagnosis and after a One-Year Follow-Up.

Prostate cancer (PCa) is the most prevalent among men, and psychological symptoms may affect many patients. This study aims to describe the prevalence of probable anxiety and depression before PCa treatments and after one year and to identify sociodemographic and clinical factors associated with these outcomes. Between February 2018 and March 2020, 292 patients recently diagnosed with PCa were recruited at the Instituto Português de Oncologia-Porto. The Hospital Anxiety and Depression Scale (HADS) was used to define probable anxiety and depression (cutoff = 11). The prevalence of probable anxiety remained stable from baseline to one year (7.8% vs. 8.5%, p = 0.866) while there was an increase in probable depression (3.1% vs. 6.8%, p = 0.012). After one year, probable depression persisted in 55.6% of patients with probable depression at baseline and 47.8% of those with probable anxiety at the first assessment had normal anxiety scores. At baseline, anxiety was more frequent among dwellers in rural areas (adjusted odds ratio-aOR, 95%CI: 2.80, 0.91-8.58) and less frequent in patients with body mass index 25-29.9 kg/m2 (aOR, 95%CI: 0.33, 0.12-0.91) compared to 18.5-24.9 Kg/m2, while those living alone had higher odds of depression (aOR, 95%CI: 6.35, 1.43-28.30). The frequency of anxiety and depression fluctuated during the course of treatment. Monitoring these symptoms would identify the most affected patients, contributing for a better use of mental health services.

COVID-19 outcomes in people living with HIV: peering through the waves (preprint)

Background: HIV infection remains a public health concern, especially in low- and middle-income countries. Data regarding exposure of COVID-19 in HIV infected patients remains scarce. We evaluated clinical characteristics and outcomes of COVID-19 patients infected with HIV, and compared with a paired sample without HIV infection. Methods: This is a substudy of a large Brazilian cohort that comprised two periods (2020 and 2021). Data was obtained through the retrospective review of medical records to collect variables of interest and primary outcomes: intensive care admission, mechanical ventilation and death. COVID-19 patients infected with HIV were compared to COVID-19 patients without concomitant diagnosis of HIV infection using the Chi-Square Test and Fisher's exact test for categorical variables and the Wilcoxon test for numerical variables. Both groups were matched for age, sex, number of comorbidities and hospital of origin using the technique of propensity score matching (up to 4:1).Results: Throughout the study, 17,101 COVID-19 patients were hospitalized, 130 (0.76%) of these infected with HIV. The median age was 54 (IQR: 43.0;64.0) years in 2020 and 53 (IQR: 46.0;63.5) years in 2021, with predominance of females in both periods. People living with HIV (PLHIV) and their controls showed similar prevalence for the admission in the ICU and mechanical ventilation requirement in the two periods, with no significant differences. In 2020, in-hospital mortality was higher in the PLHIV compared to the controls (27.9% vs 17.7%; p=0.049), but there was no difference in mortality between groups in 2021 (25.0% vs 25.1%; p>0.999). Conclusion: Our results reiterate that PLHIV were at higher risk of COVID-19 mortality in the early stages of the pandemic, however, this finding did not sustain in 2021, indicating that measures such as large-scale immunization programs have successfully contributed to reducing the excess mortality seen in PLHIV.

Prevalence of Cognitive Impairment before Prostate Cancer Treatment.

Cognitive impairment is common among patients with different types of cancer, even before cancer treatment, but no data were reported among patients with prostate cancer (PCa), who may be at high risk due to advanced age. This study aims to estimate the prevalence of cognitive impairment before PCa treatment. Between February 2018 and April 2021, the NEON-PC cohort recruited 605 patients with PCa proposed for treatment at the Portuguese Institute of Oncology of Porto. The Montreal Cognitive Assessment (MoCA) was used to assess cognitive performance. Participants with a MoCA < 1.5 standard deviations (SD) of age- and education-specific normative values were considered to have probable cognitive impairment (PCI) and were referred for a comprehensive neuropsychological assessment. Data from the population-based cohort EPIPorto (n = 351 men aged ≥40 years, evaluated in 2013-2015) were used for comparison. The prevalence of PCI was 17.4% in EPIPorto and 14.7% in NEON-PC (age- and education-adjusted odds ratio: 0.82, 95%CI: 0.58,1.18). Neuropsychological assessment was performed in 63 patients with PCa: 54.0% had cognitive impairment. These results suggest that the impact of PCa on cognitive performance could be negligible in the short term, contrary to what other studies have reported regarding other types of cancer.

THE IMPACT OF MOTORCYCLE ACCIDENTS IN A REFERENCE HOSPITAL DURING THE COVID-19 PANDEMIC

the number of motorcycle accidents attended at a reference Hospital for trauma in Campinas, state of São Paulo, and the pandemic by COVID-19 during the year 2020. Methods: This is a cross-sectional, descriptive retrospective study carried out at Hospital PUC-Campinas, through the analysis of medical records of patients, victims of motorcycle trauma undergoing orthopedic surgical management in 2020. The phases of the pandemic and the isolation rates according to the São Paulo Plan were studied. Chi-Square tests, and the Least Squares method were applied for statistical calculations. Results: 155 medical records were analyzed, of which 91.61% of the patients were male. Of those admitted, 94.84% suffered fractures and 51.61%, polyfractures. There was a correlation between the average isolation rates and the number of accidents. In the 14-day period, as the average isolation rate increased by 10%, there was an increase of approximately 3 accidents in that period. In the 7-day analysis, 1.7 more traumas were observed for every 10% increase in the average isolation. Conclusion: The results suggest that by increasing the average isolation rate, the number of traumas per motorcycle treated at the institution increased. Level of Evidence III, Comparative Retrospective Study. © 2022, Acta Ortopedica Brasileira. All Rights Reserved.