BACKGROUND: COVID-19 protective behaviors are key interventions advised by the World Health Organization (WHO) to prevent COVID-19 transmission. However, achieving compliance with this advice is often challenging, particularly among socially vulnerable groups.
OBJECTIVE: We developed a social vulnerability index (SVI) to predict individuals\' propensity to adhere to the WHO advice on protective behaviors against COVID-19 and identify changes in social vulnerability as Omicron evolved in African countries between January 2022 and August 2022 and Asia Pacific countries between August 2021 and June 2022.
METHODS: In African countries, baseline data were collected from 14 countries (n=15,375) during the first Omicron wave, and follow-up data were collected from 7 countries (n=7179) after the wave. In Asia Pacific countries, baseline data were collected from 14 countries (n=12,866) before the first Omicron wave, and follow-up data were collected from 9 countries (n=8737) after the wave. Countries\' socioeconomic and health profiles were retrieved from relevant databases. To construct the SVI for each of the 4 data sets, variables associated with COVID-19 protective behaviors were included in a factor analysis using polychoric correlation with varimax rotation. Influential factors were adjusted for cardinality, summed, and min-max normalized from 0 to 1 (most to least vulnerable). Scores for compliance with the WHO advice were calculated using individuals\' self-reported protective behaviors against COVID-19. Multiple linear regression analyses were used to assess the associations between the SVI and scores for compliance to WHO advice to validate the index.
RESULTS: In Africa, factors contributing to social vulnerability included literacy and media use, trust in health care workers and government, and country income and infrastructure. In Asia Pacific, social vulnerability was determined by literacy, country income and infrastructure, and population density. The index was associated with compliance with the WHO advice in both time points in African countries but only during the follow-up period in Asia Pacific countries. At baseline, the index values in African countries ranged from 0.00 to 0.31 in 13 countries, with 1 country having an index value of 1.00. The index values in Asia Pacific countries ranged from 0.00 to 0.23 in 12 countries, with 2 countries having index values of 0.79 and 1.00. During the follow-up phase, the index values decreased in 6 of 7 African countries and the 2 most vulnerable Asia Pacific countries. The index values of the least vulnerable countries remained unchanged in both regions.
CONCLUSIONS: In both regions, significant inequalities in social vulnerability to compliance with WHO advice were observed at baseline, and the gaps became larger after the first Omicron wave. Understanding the dimensions that influence social vulnerability to protective behaviors against COVID-19 may underpin targeted interventions to enhance compliance with WHO recommendations and mitigate the impact of future pandemics among vulnerable groups.

To improve prediction performance and reduce artifacts in Raman spectra, we developed an eXtreme Gradient Boosting (XGBoost) preprocessing method to preprocess the Raman spectra of glucose, glycerol and ethanol mixtures. To ensure the robustness and reliability of the XGBoost preprocessing method, an X-LR model (which combined XGBoost preprocessing and a linear regression (LR) model) and a X-MLP model (which combined XGBoost preprocessing and a multilayer perceptron (MLP) model) were developed. These two models were used to quantitatively analyze the concentrations of glucose, glycerol and ethanol in the Raman spectra of mixed solutions. The proportion map of hyperparameters was firstly used to narrow down the search range of hyperparameters in the X-LR and the X-MLP models. Then the correlation coefficients (R2), root mean square of calibration (RMSEC), and root mean square error of prediction (RMSEP) were used to evaluate the models\' performance. Experimental results indicated that the XGBoost preprocessing method achieved higher accuracy and generalization capability, with better performance than those of other preprocessing methods for both LR and MLP models.

Vegetable quality parameters are established according to standards primarily based on visual characteristics. Although knowledge of biochemical changes in the secondary metabolism of plants throughout development is essential to guide decision-making about consumption, harvesting and processing, these determinations involve the use of reagents, specific equipment and sophisticated techniques, making them slow and costly. However, when non-destructive methods are employed to predict such determinations, a greater number of samples can be tested with adequate precision. Therefore, the aim of this work was to establish an association capable of modeling between non-destructive-physical and colorimetric aspects (predictive variables)-and destructive determinations-bioactive compounds and antioxidant activity (variables to be predicted), quantified spectrophotometrically and by HPLC in \'Nanicão\' bananas during ripening. It was verified that to predict some parameters such as flavonoids, a regression equation using predictive parameters indicated the importance of R2, which varied from 83.43 to 98.25%, showing that some non-destructive parameters can be highly efficient as predictors.

OBJECTIVE: Midwives are susceptible to burnout due to the physically and emotionally demanding nature of their job. Burnout is an occupational phenomenon with far-reaching consequences.
OBJECTIVE: This study aimed to assess the magnitude of burnout and predictors among midwives working at public hospitals in northwest Ethiopia.
METHODS: An institutional-based cross-sectional study was conducted from February 7 to April 30, 2022. A simple random sampling method was employed to include 640 study participants. Data were collected using a self-administered questionnaire, entered into Epi-data 4.6 software, and exported to SPSS version 25 for analysis. A multivariable linear regression analysis model was fitted to identify factors contributing to midwives\' burnout.
RESULTS: The overall prevalence of midwives\' burnout was 55.3 % (95 % CI = 51.7-58.9). The prevalence of personal, work-related, and client-related burnout was 58.3 %, 60.3 %, and 55.5 %, respectively. Factors that were significantly associated with burnout includes workplace violence (β = 5.02, CI: 2.90, 7.13), not receiving training (β = 4.32 CI: 1.81, 6.80), being exposed to blood and body fluids or needle stick injuries (β = 5.13 CI: 3.12, 7.13), low superior support (β = 5.13 CI: 1.94, 5.30), working in tertiary hospitals (β = 12.77 CI: 9.48, 16.06), and job rotation of six months or less (β = 16.75, CI: 13.12, 20.39).
CONCLUSIONS: This study found that the prevalence of burnout among midwives was significantly high. Addressing burnout requires implementing effective burnout prevention measures including enhancing management support, offering professional training, creating a conducive working environment, and adhering to standard precautions.

Heart failure (HF) is defined as the inability of the heart to meet body oxygen demand requiring an elevation in left ventricular filling pressures (LVP) to compensate. LVP increase can be assessed in the cardiac catheterization laboratory, but this procedure is invasive and time-consuming to the extent that physicians rather rely on non-invasive diagnostic tools. In this work, we assess the feasibility to develop a novel machine-learning (ML) approach to predict clinically relevant LVP indices. Synchronized invasive (pressure-volume tracings) and non-invasive signals (ECG, pulse oximetry, and cardiac sounds) were collected from anesthetized, closed-chest Göttingen minipigs. Animals were either healthy or had HF with reduced ejection fraction and circa 500 heartbeats were included in the analysis for each animal. The ML algorithm showed excellent prediction of LVP indices estimating, for instance, the end-diastolic pressure with a R2 of 0.955. This novel ML algorithm could assist clinicians in the care of HF patients.

Background Gastrointestinal stromal tumors (GISTs) represent the most common mesenchymal neoplasms of the gastrointestinal tract, arising from the interstitial cells of Cajal. These tumors bridge the nervous system and muscular layers of the gastrointestinal tract, playing a crucial role in the digestive process. The incidence of GISTs demonstrates notable variations across different racial and ethnic groups, underscoring the need for in-depth analysis to understand the interplay of genetic, environmental, and socioeconomic factors behind these disparities. Linear regression analysis is a pivotal statistical tool in such epidemiological studies, offering insights into the temporal dynamics of disease incidence and the impact of public health interventions. Methodology This investigation employed a detailed dataset from 2009 to 2020, documenting GIST incidences across Asian, African American, Hispanic, and White populations. A meticulous preprocessing routine prepared the dataset for analysis, which involved data cleaning, normalization of racial terminologies, and aggregation by year and race. Linear regression models and Pearson correlation coefficients were applied to analyze trends and correlations in GIST incidences across the different racial groups, emphasizing an understanding of temporal patterns and racial disparities in disease incidence. Results The study analyzed GIST cases among four racial groups, revealing a male predominance (53.19%) and an even distribution of cases across racial categories: Whites (27.66%), Hispanics (25.53%), African Americans (24.47%), and Asians (22.34%). Hypertension was the most common comorbidity (32.98%), followed by heart failure (28.72%). The linear regression analysis for Asians showed a decreasing trend in GIST incidences with a slope of -0.576, an R-squared value of 0.717, and a non-significant p-value of 0.153. A significant increasing trend was observed for Whites, with a slope of 0.581, an R-squared value of 0.971, and a p-value of 0.002. African Americans exhibited a moderate positive slope of 0.277 with an R-squared value of 0.470 and a p-value of 0.201, indicating a non-significant increase. Hispanics showed negligible change over time with a slope of -0.095, an R-squared value of 0.009, and a p-value of 0.879, suggesting no significant trend. Conclusions This study examines GIST incidences across racial groups, revealing significant disparities. Whites show an increasing trend (p = 0.002), while Asians display a decreasing trend (p = 0.153), with stable rates in African Americans and Hispanics. Such disparities suggest a complex interplay of genetics, environment, and socioeconomic factors, highlighting the need for targeted research and interventions that address these differences and the systemic inequalities influencing GIST outcomes.

This article proposes a powerful alternative to the t-test of the null hypothesis that a coefficient in a linear regression is equal to zero when a regressor is mismeasured. We assume there are two contaminated measurements of the regressor of interest. We allow the two measurement errors to be nonclassical in the sense that they may both be correlated with the true regressor, they may be correlated with each other, and we do not require any location normalizations on the measurement errors. We propose a new maximal t-statistic that is formed from the regression of the outcome onto a maximally weighted linear combination of the two measurements. The critical values of the test are easily computed via a multiplier bootstrap. In simulations, we show that this new test can be significantly more powerful than t-statistics based on OLS or IV estimates. Finally, we apply the proposed test to a study of returns to education based on twin data from the UK. With our maximal t-test, we can discover statistically significant returns to education when standard t-tests do not.

The purpose of this study is to examine how the presence of auditors with forensic accounting skills impacts the financial performance of audited companies. Using a quantitative approach, this research employs linear regression analysis and examines a sample of 74 companies from the industrial and service sectors listed on the Amman Stock Exchange between 2012 and 2021. The findings reveal that external auditors with forensic accounting competencies have a positive impact on the financial performance of audited companies. The insights presented in this study could serve as valuable tools for improving the financial performance of companies in the service and industry sectors, highlighting the importance of supporting and promoting the skills and competencies of forensic accountants.

BACKGROUND: Efficacy of smartphone-based interventions depends on intervention content quality and level of exposure to that content. Smartphone-based survey completion rates tend to decline over time; however, few studies have identified variables that predict this decline over longer-term interventions (eg, 26 weeks).
OBJECTIVE: This study aims to identify predictors of survey completion and message viewing over time within a 26-week smoking cessation trial.
METHODS: This study examined data from a 3-group pilot randomized controlled trial of adults who smoke (N=152) and were not ready to quit smoking within the next 30 days. For 182 days, two intervention groups received smartphone-based morning and evening messages based on current readiness to quit smoking. The control group received 2 daily messages unrelated to smoking. All participants were prompted to complete 26 weekly smartphone-based surveys that assessed smoking behavior, quit attempts, and readiness to quit. Compliance was operationalized as percentages of weekly surveys completed and daily messages viewed. Linear regression and mixed-effects models were used to identify predictors (eg, intervention group, age, and sex) of weekly survey completion and daily message viewing and decline in compliance over time.
RESULTS: The sample (mean age 50, SD 12.5, range 19-75 years; mean years of education 13.3, SD 1.6, range 10-20 years) was 67.8% (n=103) female, 74.3% (n=113) White, 77% (n=117) urban, and 52.6% (n=80) unemployed, and 61.2% (n=93) had mental health diagnoses. On average, participants completed 18.3 (71.8%) out of 25.5 prompted weekly surveys and viewed 207.3 (60.6%) out of 345.1 presented messages (31,503/52,460 total). Age was positively associated with overall weekly survey completion (P=.003) and daily message viewing (P=.02). Mixed-effects models indicated a decline in survey completion from 77% (114/148) in the first week of the intervention to 56% (84/150) in the last week of the intervention (P<.001), which was significantly moderated by age, sex, ethnicity, municipality (ie, rural/urban), and employment status. Similarly, message viewing declined from 72.3% (1533/2120) in the first week of the intervention to 44.6% (868/1946) in the last week of the intervention (P<.001). This decline in message viewing was significantly moderated by age, sex, municipality, employment status, and education.
CONCLUSIONS: This study demonstrated the feasibility of a 26-week smartphone-based smoking cessation intervention. Study results identified subgroups that displayed accelerated rates in the decline of survey completion and message viewing. Future research should identify ways to maintain high levels of interaction with mobile health interventions that span long intervention periods, especially among subgroups that have demonstrated declining rates of intervention engagement over time.
BACKGROUND: ClinicalTrials.gov NCT03405129; https://clinicaltrials.gov/ct2/show/NCT03405129.

Linear regression (LR) is vastly used in data analysis for continuous outcomes in biomedicine and epidemiology. Despite its popularity, LR is incompatible with missing data, which frequently occur in health sciences. For parameter estimation, this shortcoming is usually resolved by complete-case analysis or imputation. Both work-arounds, however, are inadequate for prediction, since they either fail to predict on incomplete records or ignore missingness-induced reduction in prediction accuracy and rely on (unrealistic) assumptions about the missing mechanism. Here, we derive adaptive predictor-set linear model (aps-lm), capable of making predictions for incomplete data without the need for imputation. It is derived by using a predictor-selection operation, the Moore-Penrose pseudoinverse, and the reduced QR decomposition. aps-lm is an LR generalization that inherently handles missing values. It is applied on a reference data set, where complete predictors and outcome are available, and yields a set of privacy-preserving parameters. In a second stage, these are shared for making predictions of the outcome on external data sets with missing entries for predictors without imputation. Moreover, aps-lm computes prediction errors that account for the pattern of missing values even under extreme missingness. We benchmark aps-lm in a simulation study. aps-lm showed greater prediction accuracy and reduced bias compared to popular imputation strategies under a wide range of scenarios including variation of sample size, goodness of fit, missing value type, and covariance structure. Finally, as a proof-of-principle, we apply aps-lm in the context of epigenetic aging clocks, linear models that predict a person\'s biological age from epigenetic data with promising clinical applications.