Explore the vast range of titles available.

Background Confounding bias is a common concern in epidemiological research. Its presence is often determined by comparing exposure effects between univariable- and multivariable regression models, using an arbitrary threshold of a 10% difference to indicate confounding bias. However, many clinical researchers are not aware that the use of this change-in-estimate criterion may lead to wrong conclusions when applied to logistic regression coefficients. This is due to a statistical phenomenon called noncollapsibility, which manifests itself in logistic regression models. This paper aims to clarify the role of noncollapsibility in logistic regression and to provide guidance in determining the presence of confounding bias. Methods A Monte Carlo simulation study was designed to uncover patterns of confounding bias and noncollapsibility effects in logistic regression. An empirical data example was used to illustrate the inability of the change-in-estimate criterion to distinguish confounding bias from noncollapsibility effects. Results The simulation study showed that, depending on the sign and magnitude of the confounding bias and the noncollapsibility effect, the difference between the effect estimates from univariable- and multivariable regression models may underestimate or overestimate the magnitude of the confounding bias. Because of the noncollapsibility effect, multivariable regression analysis and inverse probability weighting provided different but valid estimates of the confounder-adjusted exposure effect. In our data example, confounding bias was underestimated by the change in estimate due to the presence of a noncollapsibility effect. Conclusion In logistic regression, the difference between the univariable- and multivariable effect estimate might not only reflect confounding bias but also a noncollapsibility effect. Ideally, the set of confounders is determined at the study design phase and based on subject matter knowledge. To quantify confounding bias, one could compare the unadjusted exposure effect estimate and the estimate from an inverse probability weighted model.

Compare meta-analysis in a distributed network to individual-level analysis for assessment of time trends of health services utilization with health administrative data. We used administrative data from Ontario, Canada to analyze temporal trends in pediatric inflammatory bowel disease health services use. Beta coefficients were obtained using negative binomial, logistic, and Cox proportional hazards regression models. We replicated the individual-level analyses in each Ontario Local Health Integration Network (LHIN), then meta-analyzed aggregate trends using both fixed and random effects meta-analysis. We compared the pooled estimates of effect with individual-level analysis. Beta coefficients, summary effect estimates, and 95% confidence intervals (CIs) from the meta-analysis of data from distributed networks were not different than those from individual-level data, regardless of meta-analytic approach used. For example, the 5-year odds ratio of colectomy in ulcerative colitis using individual-level analysis was 0.978 (95% CI 0.950 to 1.007) compared to distributed network fixed effects meta-analysis: 0.982 (95% CI 0.950 to 1.015), and random effects meta-analysis: 0.982 (95% CI 0.950 to 1.015). Meta-analysis of multi-jurisdictional estimates were similar to estimates obtained from individual-level analysis. This method is a valid alternative for analysis of multi-jurisdictional data when individual-level data cannot be shared.

To evaluate whether the systemic inflammation score (SIS) could predict postoperative outcomes for patients undergoing video-assisted thoracoscopic surgery (VATS) lobectomy for early-stage non-small-cell lung cancer (NSCLC). This retrospective study was conducted on the prospectively maintained database in our institution between January 2016 and December 2017. Preoperative SIS comprising serum albumin (sALB) and lymphocyte-to-monocyte ratio (LMR) was graded into 0, 1 and 2, and then utilized to distinguish patients at high surgical risks. Multivariable logistic-regression analysis was conducted to determine independent risk factors for postoperative outcomes. There were 1,025 patients with TNM-stage I-II NSCLC included, with an overall morbidity rate of 31.1% and mortality rate of 0.3%. We applied the sALB at 40 g/L and the median LMR of our series at 4.42 as dichotomized cutoffs for modified SIS scoring criteria. Both minor and major morbidity rates in patients with SIS=2 were significantly higher than those in patients with SIS=0 and with SIS=1 ( <0.001). No difference was found in overall morbidity rate between patients with SIS=1 and with SIS=0 ( =0.20). No significant difference was found in the mortality rate between these 3 groups. Patients with SIS=2 had the highest probability to experience most of individual complications. Finally, multivariable logistic-regression analysis suggested that preoperative SIS=2 could independently predict the morbidity risks following VATS lobectomy (OR=1.73; 95% CI=1.11-2.71; =0.016). The SIS scoring system can be employed as a simplified, effective and routinely operated risk stratification tool in patients undergoing VATS lobectomy.

To identify risk factors for surgical site infections (SSIs) following C-sections in an underserved, urban population. Retrospective case-control study and multivariable regression analyses. Multicenter urban hospital system. All women undergoing C-sections during 2023. To identify risk factors for SSIs, patients suffering SSIs were compared to a propensity-matched control group (controlled for the following variables: age, body mass index, diabetes mellitus, American Society of Anesthesia (ASA) score, wound class, and duration of surgery). In addition, multivariable logistic regression analysis was performed to identify independent risks for SSIs. Of 4,642 C-sections performed, 73 SSIs were identified; 90% were detected after hospital discharge. Compared to a propensity-matched group, more patients in the SSI cohort received gentamicin with clindamycin (vs a cefazolin-based regimen); gentamicin dosing was consistently below recommended levels. Also, significantly more patients in the SSI group were recent immigrants to the United States compared to the control group (20.5% vs 4.1%, = .004). Multivariate regression analysis revealed 3 independent risk factors for SSIs: ASA score, surgery at a hospital without an Obstetrics-Gynecology residency program, and residence in the borough of the Bronx, NY. For women living in areas of low socioeconomic status, most SSIs after C-sections are detected following hospital discharge. Women who are recent immigrants and living in areas of high poverty are particularly at higher risk. Addressing the broader social determinants of health, particularly in underserved areas, will be crucial in reducing SSIs and improving overall maternal health outcomes.

Background Confounding is a common issue in epidemiological research. Commonly used confounder-adjustment methods include multivariable regression analysis and propensity score methods. Although it is common practice to assess the linearity assumption for the exposure-outcome effect, most researchers do not assess linearity of the relationship between the confounder and the exposure and between the confounder and the outcome before adjusting for the confounder in the analysis. Failing to take the true non-linear functional form of the confounder-exposure and confounder-outcome associations into account may result in an under- or overestimation of the true exposure effect. Therefore, this paper aims to demonstrate the importance of assessing the linearity assumption for confounder-exposure and confounder-outcome associations and the importance of correctly specifying these associations when the linearity assumption is violated. Methods A Monte Carlo simulation study was used to assess and compare the performance of confounder-adjustment methods when the functional form of the confounder-exposure and confounder-outcome associations were misspecified (i.e., linearity was wrongly assumed) and correctly specified (i.e., linearity was rightly assumed) under multiple sample sizes. An empirical data example was used to illustrate that the misspecification of confounder-exposure and confounder-outcome associations leads to bias. Results The simulation study illustrated that the exposure effect estimate will be biased when for propensity score (PS) methods the confounder-exposure association is misspecified. For methods in which the outcome is regressed on the confounder or the PS, the exposure effect estimate will be biased if the confounder-outcome association is misspecified. In the empirical data example, correct specification of the confounder-exposure and confounder-outcome associations resulted in smaller exposure effect estimates. Conclusion When attempting to remove bias by adjusting for confounding, misspecification of the confounder-exposure and confounder-outcome associations might actually introduce bias. It is therefore important that researchers not only assess the linearity of the exposure-outcome effect, but also of the confounder-exposure or confounder-outcome associations depending on the confounder-adjustment method used.

Background The proximal humeral fracture (PHF) is the third most common fracture in older individuals. Prior level of care (LoC) and associated comorbidities may have an impact on patient outcome and prognosis. Methods Retrospective German health insurance data from patients with PHF aged 65 years and older between 01/17 to 09/22 were analysed. The primary endpoints included overall survival (OS), major adverse events (MAEs), thromboembolic events (TEs), and surgery- or injury-related complications. All endpoints were analysed using multivariable models. Results A total of 55,798 patients (median age 79 years; 84% female) were included. Prior to PHF, 68% had no LoC (LoC I 3%, LoC II 12%, LoC III 11%, LoC IV 6%, LoC V 1%), and 8% were living in a nursing home. With increasing LoC, the proportion of patients receiving non-operative treatment (no LoC 52%, LoC I 53%, LoC II 62%, LoC III 64%, LoC IV 71%, LoC V 76%) and the likelihood of a worse outcome increased. Both, mortality rates (1-year mortality: no LoC 4%, LoC I 12%, LoC II 19%, LoC III 29%, LoC IV 41%, LoC V 50%) and rates of MAEs increased drastically with increasing LoC. Multivariable analyses confirmed that increasing LoC was associated with a greater risk of death, MAEs, and TEs (all p  < 0.001). Conclusion Prior LoC has a significant effect on the course of PHF and the choice of treatment method in older individuals. This should be considered when making treatment decisions. Level of evidence Level III, retrospective comparative study.

Digital health literacy describes an individual's ability to use digital information and tools to improve their own health. Understanding how digital health literacy varies across populations could help improve health equity. However, the determinants of digital health literacy have been scarcely evaluated. This study aims to assess the levels of digital health literacy in 4 countries (United Kingdom, Sweden, Italy, and Germany) and explore potential associations between digital health literacy and demographic characteristics and self-perceived health status. A cross-sectional online survey was disseminated to participants from the United Kingdom, Italy, Germany, and Sweden in December 2020. Digital health literacy was self-reported using the validated eHealth Literacy Scale (eHEALS; range: 0-40); low digital health literacy has been previously defined as an eHEALS score<26. Participant characteristics collected were sex, age group, ethnicity, country, and perceived overall health status. A multivariable linear regression analysis was performed to explore associations between these variables and digital health literacy. A total of 6331 participants were included (51.7% female, n=3272). The mean eHEALS score was 29.2 (SD 6.8). Participant age, sex, health status, and country of residence were included in the final multivariable model. Compared to the 45- to 54-year age group, the 55 years and older age group had lower digital health literacy (β=-1.0; 95% CI -1.4 to -0.5; P<.001), while digital health literacy was higher in those aged 25-34 years (β=0.9; 95% CI 0.3-1.5; P=.002) and 35-44 years (β=0.6; 95% CI 0.1-1.2; P=.03). Better health status was associated with greater digital health literacy (β=0.3; 95% CI 0.2-0.4; P<.001). Compared to participants from Germany, those from the United Kingdom (β=2.1; 95% CI 1.7-2.5; P<.001) and Sweden (β=2.9; 95% CI 2.4-3.4; P<.001) had higher digital health literacy scores, while there was no difference with Italian participants (P=.399). Sex and ethnicity did not have any significant effect on digital health literacy. This study found significant variations in digital health literacy by age, health status, and country of residence. Targeted educational programs for vulnerable groups, particularly those of older age and poorer health status, are essential. Policies fostering accessible digital health solutions and mitigating health technology-related uncertainties for these populations are crucial for achieving optimal health outcomes.

Background Female sex was reported to be associated with an unfavorable outcome in acute myocardial infarction (AMI). In this nationwide analysis we assessed sex differences in acute outcomes of AMI and recent trends in patient healthcare. Methods We analyzed 875 735 German cases hospitalized with a main diagnosis of ST‐ (STEMI) and non ST‐elevation myocardial infarction (NSTEMI) between January 01 2014 and December 31 2017 regarding morbidity, in‐hospital mortality and treatments. A multivariable logistic regression model was designed to evaluate the use of interventions and their impact on in‐hospital mortality. Results STEMI cases decreased from 72 894 in 2014 to 68 213 in 2017, with 70% assignable to men. Female sex was associated with older age (74 vs. 62 years), and higher prevalence of cardiovascular risk factors such as chronic kidney disease (19.2% vs. 12.5%), hypertension (69.0% vs. 65.0%) and left ventricular heart failure (36.0% vs. 32.1%). In NSTEMI, female sex was also associated with older age (78 vs. 71 years), and higher prevalence of cardiovascular risk factors such as chronic kidney disease (29.7% vs. 23.9%), hypertension (77.4% vs. 74.5%) and left ventricular heart failure (40.5% vs. 36.4%). Overall, 74.3% of female and 81.3% of male STEMI cases received percutaneous coronary intervention (PCI, p < 0.001). In NSTEMI, PCI was performed in 40.8% of female and 52.0% of male cases (p < 0.001). In‐hospital mortality was notably increased in female patients with STEMI (15.0% vs. 9.6%; p < 0.001; OR 1.07; 95% CI 1.03–1.10) and NSTEMI (8.3% vs. 6.3%; p < 0.001; OR 0.91; 95% CI 0.89–0.93) compared to males. Conclusions Our nationwide real‐world data document that in‐patient STEMI cases continue to decrease in women and men. The observed higher in‐hospital mortality in women was largely attributed to a more unfavorable risk and age distribution rather than to female‐intrinsic factors. Women with AMI continue to be less likely to receive revascularization therapies.

After the ore (seam) extraction in longwall mining, the immediate roof layers over the extracted panel are strained and suspended downward. This process expands upward and causes the caving and fracturing of damaged roof rock strata. The combination height of the caved and interconnected fractured zones is considered as the height of caving–fracturing zone (HCFZ) in this research. Precise estimation of this height is crucial to the exact determination of directed loads toward the front and sides abutments. The paper describes an intelligent model based on the artificial neural network (ANN) to predict HCFZ. To validate the ability of ANN model, its results are compared to the multivariable regression analysis (MVRA) results. For models construction and evaluation, a wide range of datasets comprising of geometrical and geomechanical characteristics of mined panel and roof strata have been gathered. Performance evaluation indices including determination coefficient ( R 2 ), variance account for, mean absolute error ( E a ) and mean relative error ( E r ) have been utilized to assess the models’ capability. Comparison results show that the ANN model performance is considerably better than the MVRA model. Moreover, obtained results are further compared with the results of available in situ, empirical, analytical, numerical and physical models reported in the literature. This comparison confirms that a reasonable agreement exists between the ANN model and the previous comparable methods. Finally, the sensitivity analysis of ANN results shows that the overburden depth has the maximum effect, whereas the Poisson’s ratio has the minimum effect on the HCFZ in this research.

This paper presents the application of artificial neural networks (ANN) and multivariable regression analysis (MRA) to predict the bearing capacity and the settlement of multi-edge skirted footings on sand. Respectively, these parameters are defined in terms of the bearing capacity ratio (BCR) of skirted to unskirted footing and the settlement reduction factor (SRF), the ratio of the difference in settlement of unskirted and skirted footing to the settlement of unskirted footing at a given pressure. The model equations for the prediction of the BCR and the SRF of the regular shaped footing were first developed using the available data collected from the literature. These equations were later modified to predict the BCR and the SRF of the multi-edge skirted footing, for which the data were generated by conducting a small scale laboratory test. The input parameters chosen to develop ANN models were the angle of internal friction (ϕ) and skirt depth (Ds) to the width of the footing (B) ratio for the prediction of the BCR; as for the SRF one additional input parameter was considered: normal stress ( ). The architecture for the developed ANN models was 2-2-1 and 3-2-1 for the BCR and the SRF, respectively. The R2 for the multi-edge skirted footings was in the range of 0,940-0,977 for the ANN model and 0,827-0,934 for the regression analysis. Similarly, the R2 for the SRF prediction might have been 0,913-0,985 for the ANN model and 0,739-0,932 for the regression analysis. It was revealed that the predicted BCR and SRF for the multi-edge skirted footings with the use of ANN is superior to MRA. Furthermore, the results of the sensitivity analysis indicate that both the BCR and the SRF of the multi-edge skirted footings are mostly affected by skirt depth, followed by the friction angle of the sand.