Bring More Data!-A Good Advice? Removing Separation in Logistic Regression by Increasing Sample Size

Influence of various factors on the legal outcome of cases of child abuse-experiences gathered at an interdisciplinary forensic examination center in Vienna, Austria

BACKGROUND AND OBJECTIVE

To improve the currently low conviction rate in cases of child abuse a forensic examination center for children and adolescents (FOKUS) was established in Vienna, Austria. Besides a state of the art treatment combined with forensic documentation, one of FOKUS' key goals is to identify potential areas for improvements within the process legal proceedings in cases of child abuse through constant scientific monitoring. The accompanying study at hand includes all patients referred to FOKUS within a two year timeframe (n = 233), monitoring their progression from first contact with the medical professionals from FOKUS to the end of criminal proceedings. A detailed analysis of case files was performed in those cases that were reported to the legal authorities by the clinicians of FOKUS (n = 87). Aim of the study is to investigate which factors contribute to the initiation of legal proceedings and a successful conviction.

RESULTS

Multivariate logistic regression analyses showed that main proceedings were opened more often in cases where the offender was an adult (p < 0.001) or admitted his guilt (p < 0.001) and if digital traces were available (p = 0.001) or trial support (p = 0.024) present. Furthermore, the combined occurrence of medical documentation and victim disclosure was related to a higher probability of opening main trials.

CONCLUSION

These findings underline how challenging the successful persecution of an offender in cases of child abuse is.

Seroepidemiology of Lassa virus in pregnant women in Southern Nigeria: A prospective hospital-based cohort study

BACKGROUND

There is limited epidemiological evidence on Lassa fever in pregnant women with acute gaps on prevalence, infection incidence, and risk factors. Such evidence would facilitate the design of therapeutic and vaccine trials and the design of control programs. Our study sought to address some of these gaps by estimating the seroprevalence and seroconversion risk of Lassa fever in pregnant women.

METHODOLOGY/PRINCIPAL FINDINGS

We conducted a prospective hospital-based cohort between February and December 2019 in Edo State, Southern Nigeria, enrolling pregnant women at antenatal clinic and following them up at delivery. Samples were evaluated for IgG antibodies against Lassa virus. The study demonstrates a seroprevalence of Lassa IgG antibodies of 49.6% and a seroconversion risk of 20.8%. Seropositivity was strongly correlated with rodent exposure around homes with an attributable risk proportion of 35%. Seroreversion was also seen with a seroreversion risk of 13.4%.

CONCLUSIONS/SIGNIFICANCE

Our study suggests that 50% of pregnant women were at risk of Lassa infection and that 35.0% of infections might be preventable by avoiding rodent exposure and conditions which facilitate infestation and the risk of human-rodent contact. While the evidence on rodent exposure is subjective and further studies are needed to provide a better understanding of the avenues of human-rodent interaction; public health measures to decrease the risk of rodent infestation and the risk of spill over events may be beneficial. With an estimated seroconversion risk of 20.8%, our study suggests an appreciable risk of contracting Lassa fever during pregnancy and while most of these seroconversions may not be new infections, given the high risk of adverse outcomes in pregnancy, it supports the need for preventative and therapeutic options against Lassa fever in pregnancy. The occurrence of seroreversion in our study suggests that the prevalence obtained in this, and other cohorts may be an underestimate of the actual proportion of women of childbearing age who present at pregnancy with prior LASV exposure. Additionally, the occurrence of both seroconversion and seroreversion in this cohort suggests that these parameters would need to be considered for the development of Lassa vaccine efficacy, effectiveness, and utility models.

An exact regression-based approach for the estimation of natural direct and indirect effects with a binary outcome and a continuous mediator

In the causal mediation framework, a number of parametric regression-based approaches have been introduced in recent years for estimating natural direct and indirect effects for a binary outcome in an exact manner, without invoking simplifying assumptions based on the rareness or commonness of the outcome. However, most of these works have focused on a binary mediator. In this article, we aim at a continuous mediator and introduce an exact approach for the estimation of natural effects on the odds ratio, risk ratio, and risk difference scales. Our approach relies on logistic and linear models for the outcome and mediator, respectively, and uses numerical integration to calculate the nested counterfactual probabilities underlying the definition of natural effects. Formulas for the delta method standard errors for all effects estimators are provided. The performance of our proposed exact estimators was evaluated in simulation studies that featured scenarios with different levels of outcome rareness/commonness, including a marginally but not conditionally rare outcome scenario. Furthermore, we evaluated the merit of Firth's penalization to mitigate the bias in the logistic regression coefficients estimators for the smallest outcome prevalences and sample sizes investigated. Using a SAS macro provided, we implemented our approach to assess the effect of placental abruption on low birth weight mediated by gestational age. We found that our exact natural effects estimators worked properly in both simulated and real data applications.

Depression and Anxiety, Stigma, and Social Support Among Women in the Postpartum Period

OBJECTIVE

To examine the relationships among depression and anxiety symptomatology, stigma of mental illness, levels of social support, and select demographics among hospitalized women in the postpartum period.

DESIGN

Descriptive, cross-sectional, correlational.

SETTING

A convenience sample of 105 English-speaking and Spanish-speaking women was recruited and enrolled from a 208-bed free-standing Southern California women's community hospital postpartum unit serving a diverse community.

MEASUREMENTS

Data were collected on potential covariates including participants' characteristics, depression and anxiety symptomatology, social support, and stigma.

RESULTS

Sixteen participants were classified as high risk for depression or anxiety based on the Edinburgh Postnatal Depression Scale and Generalized Anxiety Disorder-7. Anxiety was significantly associated with all stigma subscales; the strongest association was with Internal Stigma (r = .46, p < .001, moderate effect), followed by Disclosure Stigma (r = .36, p < .001, moderate effect) and External Stigma (r = .30, p = .002, moderate effect). All social support subscales were negatively associated with depression and anxiety; the Friends subscale had the strongest correlations with depression (r = -.27, p = .006, small effect) and anxiety (r = .34, p = .001, moderate effect). Firth (penalized likelihood) logistic regression analysis was conducted to ascertain the effects of study covariates on the likelihood of participants being at risk for postnatal depression or anxiety. The significant factor that increased the odds of participants being in the high-risk group was decreased social support (adjusted OR = 0.46, 95% CI [0.24, 0.76], p = .003).

CONCLUSION

These results show the enduring prevalence of postnatal depression and anxiety and the relevance of stigma and social support in aggravating or mitigating symptomatology. There is a need for comprehensive standardized screening to ensure the identification of and referral to treatment for women at risk.

Reliability-Based Design Optimization of Structures Considering Uncertainties of Earthquakes Based on Efficient Gaussian Process Regression Metamodeling

The complexity of earthquakes and the nonlinearity of structures tend to increase the calculation cost of reliability-based design optimization (RBDO). To reduce computational burden and to effectively consider the uncertainties of ground motions and structural parameters, an efficient RBDO method for structures under stochastic earthquakes based on adaptive Gaussian process regression (GPR) metamodeling is proposed in this study. In this method, the uncertainties of ground motions are described by the record-to-record variation and the randomness of intensity measure (IM). A GPR model is constructed to obtain the approximations of the engineering demand parameter (EDP), and an active learning (AL) strategy is presented to adaptively update the design of experiments (DoE) of this metamodel. Based on the reliability of design variables calculated by Monte Carlo simulation (MCS), an optimal solution can be obtained by an efficient global optimization (EGO) algorithm. To validate the effectiveness and efficiency of the developed method, it is applied to the optimization problems of a steel frame and a reinforced concrete frame and compared with the existing methods. The results show that this method can provide accurate reliability information for seismic design and can deal with the problems of minimizing costs under the probabilistic constraint and problems of improving the seismic reliability under limited costs.

Aging through the time of COVID-19: a survey of self-reported healthcare access

BACKGROUND

Chronic conditions are common and require ongoing continuous management and preventive measures. The COVID-19 pandemic may have affected the management of chronic conditions by delaying care. We sought to understand the impact of personal characteristics (i.e., age) and healthcare factors (i.e., access to a provider) on healthcare access in a sample of Americans 50 years of age or older during COVID-19.

METHOD

Participants completed an online survey at the start of the COVID-19 pandemic - the Aging in the Time of COVID Survey. Questions focused on health status, health care access, COVID-19 fear, and social connectedness. Participants were recruited through social media advertisements, list serves, and snowball sampling. Data collection started in early April 2020 and concluded in late May 2020. Logistic regression models examined the results of two key access points: healthcare provider/doctor (n = 481) and medication (n = 765), with 56 and 93% of participants reporting access to a provider and medications, respectively.

RESULTS

Individuals with an established primary care provider were much more likely to obtain access to a healthcare provider, OR = 3.81 (95% CI: 1.69, 8.77), and to receive medication, OR = 4.48 (95% CI: 1.61, 11.48), during the time of COVID-19. In addition, access to medication was (a) higher for those who were older, OR = 1.05 (95% CI: 1.01, 1.09), had a higher income (greater than 100 k compared to less than 50 k, OR = 3.04 (95% CI: 1.11, 8.98), and (b) lower for those having caregiving responsibilities, OR = 0.41 (95% CI: 0.21, 0.78), or greater social isolation, OR = 0.93 (95% CI: 0.87, 0.98).

CONCLUSIONS

Although most participants had access to medication, just over half had access to a healthcare provider when needed. Notably, health-seeking behaviors for individuals who do not have an established primary care providers as well as those who provide unpaid care, are socially isolated, and younger may require more proactive approaches to care monitoring, management, and maintenance.

Artificial intelligence and machine learning assisted drug delivery for effective treatment of infectious diseases

In the era of antimicrobial resistance, the prevalence of multidrug-resistant microorganisms that resist conventional antibiotic treatment has steadily increased. Thus, it is now unquestionable that infectious diseases are significant global burdens that urgently require innovative treatment strategies. Emerging studies have demonstrated that artificial intelligence (AI) can transform drug delivery to promote effective treatment of infectious diseases. In this review, we propose to evaluate the significance, essential principles, and popular tools of AI in drug delivery for infectious disease treatment. Specifically, we will focus on the achievements and key findings of current research, as well as the applications of AI on drug delivery throughout the whole antimicrobial treatment process, with an emphasis on drug development, treatment regimen optimization, drug delivery system and administration route design, and drug delivery outcome prediction. To that end, the challenges of AI in drug delivery for infectious disease treatments and their current solutions and future perspective will be presented and discussed.

To tune or not to tune, a case study of ridge logistic regression in small or sparse datasets

BACKGROUND

For finite samples with binary outcomes penalized logistic regression such as ridge logistic regression has the potential of achieving smaller mean squared errors (MSE) of coefficients and predictions than maximum likelihood estimation. There is evidence, however, that ridge logistic regression can result in highly variable calibration slopes in small or sparse data situations.

METHODS

In this paper, we elaborate this issue further by performing a comprehensive simulation study, investigating the performance of ridge logistic regression in terms of coefficients and predictions and comparing it to Firth's correction that has been shown to perform well in low-dimensional settings. In addition to tuned ridge regression where the penalty strength is estimated from the data by minimizing some measure of the out-of-sample prediction error or information criterion, we also considered ridge regression with pre-specified degree of shrinkage. We included 'oracle' models in the simulation study in which the complexity parameter was chosen based on the true event probabilities (prediction oracle) or regression coefficients (explanation oracle) to demonstrate the capability of ridge regression if truth was known.

RESULTS

Performance of ridge regression strongly depends on the choice of complexity parameter. As shown in our simulation and illustrated by a data example, values optimized in small or sparse datasets are negatively correlated with optimal values and suffer from substantial variability which translates into large MSE of coefficients and large variability of calibration slopes. In contrast, in our simulations pre-specifying the degree of shrinkage prior to fitting led to accurate coefficients and predictions even in non-ideal settings such as encountered in the context of rare outcomes or sparse predictors.

CONCLUSIONS

Applying tuned ridge regression in small or sparse datasets is problematic as it results in unstable coefficients and predictions. In contrast, determining the degree of shrinkage according to some meaningful prior assumptions about true effects has the potential to reduce bias and stabilize the estimates.

Home Visiting for NICU Graduates: Impacts of Following Baby Back Home

BACKGROUND AND OBJECTIVES

The Following Baby Back Home (FBBH) home visiting program supports families of high-risk low birth weight preterm infants after discharge from a hospital NICU. This study compares the health care use, immunization, and infant mortality rate of low birth weight preterm infants enrolled in FBBH with similar infants not in the program.

METHODS

From January 2013 to December 2017, 498 children enrolled in FBBH were identified in Arkansas vital statistics records and the Arkansas All-Payer Claims Database. Infants in FBBH were matched with children in a control group on the basis of demographics and medical conditions of the infant. Generalized linear mixed models with double propensity-score adjustment were used to estimate program effects.

RESULTS

In the first year after discharge and compared with a propensity-score matched cohort of control infants, those enrolled in FBBH were significantly more likely to have higher numbers of medical appointments and more compliant immunization history. The odds of dying in the first year of life for control infants was 4.4 times (95% confidence interval: 1.2-20.7) higher than those managed in the program.

CONCLUSIONS

A goal of the FBBH home visiting program is to work with parents to educate and support them as they care for their medically fragile infants. We conclude that education and support was instrumental in the infant health care use and outcome differences we observed during the first year of life.

Modelling hospital outcome: problems with endogeneity

Background

Mortality modelling in the critical care paradigm traditionally uses logistic regression, despite the availability of estimators commonly used in alternate disciplines. Little attention has been paid to covariate endogeneity and the status of non-randomized treatment assignment. Using a large registry database, various binary outcome modelling strategies and methods to account for covariate endogeneity were explored.

Methods

Patient mortality data was sourced from the Australian & New Zealand Intensive Society Adult Patient Database for 2016. Hospital mortality was modelled using logistic, probit and linear probability (LPM) models with intensive care (ICU) providers as fixed (FE) and random (RE) effects. Model comparison entailed indices of discrimination and calibration, information criteria (AIC and BIC) and binned residual analysis. Suspect covariate and ventilation treatment assignment endogeneity was identified by correlation between predictor variable and hospital mortality error terms, using the Stata™ “eprobit” estimator. Marginal effects were used to demonstrate effect estimate differences between probit and “eprobit” models.

Results

The cohort comprised 92,693 patients from 124 intensive care units (ICU) in calendar year 2016. Patients mean age was 61.8 (SD 17.5) years, 41.6% were female and APACHE III severity of illness score 54.5(25.6); 43.7% were ventilated. Of the models considered in predicting hospital mortality, logistic regression (with or without ICU FE) and RE logistic regression dominated, more so the latter using information criteria indices. The LPM suffered from many predictions outside the unit [0,1] interval and both poor discrimination and calibration. Error terms of hospital length of stay, an independent risk of death score and ventilation status were correlated with the mortality error term. Marked differences in the ventilation mortality marginal effect was demonstrated between the probit and the "eprobit" models which were scenario dependent. Endogeneity was not demonstrated for the APACHE III score.

Conclusions

Logistic regression accounting for provider effects was the preferred estimator for hospital mortality modelling. Endogeneity of covariates and treatment variables may be identified using appropriate modelling, but failure to do so yields problematic effect estimates.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12874-021-01251-8.

Elucidating variations in outcomes among older end-stage renal disease patients on hemodialysis in Fukuoka Prefecture, Japan

Variations in health care outcomes and services potentially indicate resource allocation inefficiency. Therefore, this study was conducted to examine variations in mortality and hospitalization cases among end-stage renal disease (ESRD) patients receiving hemodialysis (HD) care from medical facilities located in 13 secondary medical care areas (SMAs) of Fukuoka prefecture, Japan. The research was designed as a retrospective, cross-sectional study using insurance claims data. The subjects of the study were older patients (over 65 years old) insured by the Fukuoka prefecture’s Latter-Stage Elderly Healthcare Insurance. Using an electronic claims database, we identified patients with chronic kidney disease (CKD) who had received HD care from April 1, 2017 to March 31, 2018. The CKD status was identified using International Classification of Disease, 10th revision code, and HD maintenance status was ascertained using specific insurance procedure codes. A total of 5,243 patients met our inclusion criteria and their records were subsequently reviewed. About 73% (n = 3,809) of patients had admission records during the period studied. Thus, the data regarding hospital length of stay (LOS) and admission costs were analyzed separately. Significant differences in terms of increased risks in hospitalization were evident in a number of SMAs. An increase in mortality risk due to heart failure and malignancy was observed in two separate SMAs. Also, analyzed LOS, total hospitalization cost, and cost per day according to SMAs showed statistically significant variations. The findings highlight the magnitude of the burden of CKD and ESRD in the community. The high prevalence of ESRD, associated mortality, and hospitalized HD patients signal the need for clinicians to assume broader roles in measures against chronic kidney disease through involvement in community awareness programs. To improve patient outcomes, improvement of regional health care provision, the level of medical care, and the development of existing human resources are needed.

Statistical Advances in Epidemiology and Public Health

The key role of statistical modeling in epidemiology and public health is unquestionable [...].

Reasons Behind Generation Z Nursing Students' Intentions to Leave their Profession: A Cross-Sectional Study

This study aims to identify predictors of turnover intention within 2 years of employment among 3rd and 4th grade Generation Z nursing students. Turnover intention is a multi-stage process ranging from mere desire, to serious thoughts, decision-making, and actual planning. Previous studies have focused on identifying the factors affecting turnover intention among practicing nurses. However, undergraduate nursing students also contemplate their decision to stay or leave the nursing profession after graduation. This cross-sectional descriptive study recruited 210 nursing students from 3 colleges in South Korea. A self-administered survey was conducted using the Career Preparation Behavior Scale, the Calling and Vocational Questionnaire, the Social Responsibility Scale, and the Revised Self-Leadership Questionnaire. Four questions were used to explore participants’ turnover intention, their motivation for studying nursing, their major satisfaction, and their clinical experience satisfaction. Descriptive and multiple logistic regression statistics were obtained using SPSS. Of the participants, 17.6% had turnover intention within 2 years of employment. Multiple logistic regression analysis indicates that clinical experience satisfaction is the only significant predictor. In the univariate analysis, nursing students who had turnover intentions were less likely to practice career preparation behaviors and had lower levels of vocational consciousness and social responsibility. To keep a proficient nursing workforce in the profession, professional commitments from universities and hospital institutions are needed to provide quality clinical learning experiences for nursing students. Further prospective study is needed to observe how Generation Z undergraduate students’ turnover intentions change and what factors influence this process.

Adjuvant chemotherapy after radiotherapy or concurrent chemoradiotherapy for pelvic lymph node-positive patients with locally advanced cervical cancer: a propensity score matching analysis

Objective

The benefit of adjuvant chemotherapy after definitive chemoradiotherapy in patients with pelvic lymph node-positive cervical cancer has been poorly studied. This study aimed to test the hypothesis that the addition of adjuvant chemotherapy to definitive radiotherapy or concurrent chemoradiotherapy improves survival in patients with pelvic lymph node-positive cervical squamous cell carcinoma.

Methods

This retrospective study enrolled patients with stage IB–IVA pelvic lymph node-positive cervical squamous cell carcinoma, without para-aortic lymph node metastases and initially treated with definitive radiotherapy or concurrent chemoradiotherapy between March 2007 and February 2018. Patients were classified into the adjuvant chemotherapy (5-fluorouracil or paclitaxel, plus cisplatin) and no-adjuvant chemotherapy groups. Treatment outcomes were compared between the two groups before and after 1:1 ratio propensity score matching.

Results

Medical records of 951 patients were reviewed and 792 patients were excluded. Finally, 159 patients were enrolled for analysis. Of these, 42 patients received a median of two cycles (range, 1–6) of adjuvant chemotherapy and 117 patients under observation after primary treatment. The median follow-up period was 33.8 months (range, 2.9–113.0). Before propensity score matching, no significant difference was observed in survivals between the two groups (P>0.05). After propensity score matching, 37 pairs of patients were selected. The 3-year rates of progression-free survival, overall survival, local control, and distant metastasis-free survival in the adjuvant chemotherapy and no-adjuvant chemotherapy groups were 80.2% and 60.4% (P=0.07), 83.0% and 63.7% (P=0.17), 94.0% and 81.9% (P=0.12), and 85.9% and 60.1% (P=0.04), respectively. The incidences of grade 3–4 acute and late toxicities were comparable between the two groups (P>0.05).

Discussion

Adjuvant chemotherapy significantly improved 3-year distant metastasis-free survival in patients with pelvic lymph node-positive cervical squamous cell carcinoma. Further prospective studies are needed to provide supportive evidence for the therapeutic efficacy of adjuvant chemotherapy.