Logistic Regression Spline Research Articles

Maternal Steroid Hormone Levels in Early Pregnancy and Autism in the Offspring: A Population-Based, Nested Case-Control Study

BackgroundA role for prenatal steroid hormones in the etiology of autism has been proposed, but evidence is conflicting. MethodsHere, we examined serum levels of maternal estradiol, testosterone, 17-hydroxyprogesterone (OHP), and cortisol from the first trimester of gestation (mean = 10.1 weeks) in relation to the odds of diagnosed autism with and without co-occurring intellectual disability (ID) in the offspring (n = 118 autism with ID, n = 249 autism without ID, n = 477 control). Levels of maternal hormones were measured using highly sensitive liquid chromatography tandem mass spectrometry, standardized according to gestational timing of sample collection, and analyzed with restricted cubic spline logistic regression models adjusting for child’s sex and maternal health, demographic, and socioeconomic factors. ResultsWe observed significant nonlinear associations between maternal estradiol, 17-OHP, and cortisol with autism, which varied with the presence of co-occurring ID. Compared to mean levels, lower levels of estradiol were associated with higher odds of autism with ID (odds ratio for concentrations 1 SD below the mean = 1.66; 95% CI, 1.24–2.11), while higher cortisol levels were associated with lower odds (odds ratio for 1 SD above the mean = 0.55; 95% CI, 0.36–0.88). In contrast, higher 17-OHP was associated with increased odds of autism without ID (odds ratio for 1 SD above the mean = 1.49; 95% CI, 1.11–1.99). We observed no evidence for interaction with sex of the child. ConclusionsThese findings support the notion that the maternal steroid hormonal environment in early pregnancy may contribute to autism, but also emphasize the complex relationship between early-life steroid exposure and autism.

Abstract PO1-14-12: Quantitative Biomarkers and 21-Gene Assay Results Predict Breast-Conserving Surgery Following Neoadjuvant Therapy in Early-Stage, Hormone Receptor-Positive, HER2-Negative Breast Cancer

Abstract Background: Neoadjuvant chemotherapy (NACT) and neoadjuvant endocrine therapy (NET) are often used to downstage tumors and/or enable breast-conserving surgery (BCS) in patients (pts) with early-stage, hormone receptor-positive/HER2-negative (HR+/HER2-) breast cancer. Given the increasing use of neoadjuvant therapy for HR+ breast cancer, studies are needed to better quantify the success of NACT/NET in downstaging tumors as a function of estrogen (ER), progesterone (PR), and Ki-67 expression, and 21-gene recurrence score (RS) in this patient population. Methods: Data were from pts with stage I-III, HR+/HER2- breast cancer diagnosed from 2010-2020 in the National Cancer Database (NCDB). Quantitative ER, PR, and Ki-67 data became available in 2018. We included pts who received either NACT or NET, with surgical plan categorized as mastectomy or BCS. Pts who received NACT for 12–30 weeks or NET for 4–36 months prior to surgery were eligible. Predicted rates of BCS by quantitative biomarkers (ER, PR, Ki-67 percentage, and RS) were estimated using restricted cubic spline logistic regression. Multivariable logistic regression models were fit to examine the associations between surgical plan and quantitative biomarkers for the two neoadjuvant cohorts. Adjusted odds ratios (aOR) and 95% confidence intervals (CI) were calculated per 10% increase in percentages of ER, PR, and Ki-67 expression. Results: Of 44,589 pts treated with NACT, the mean age was 53 years, the median amount of time they were on NACT before surgery was 21 weeks, and 65% had cT1-2 disease. Of 10,466 pts treated with NET, the mean age was 68 years, the median amount of time they were on NET before surgery was 6 months, and 76% had cT1-2 disease. Overall, 42% of pts underwent BCS after NACT; while 64% did after NET. In the NACT cohort, 13,752 of the pts with available quantitative biomarker data were included in regression analyses; in the NET cohort, 4,003 were included. After adjusting for demographic and clinicopathologic factors, increasing ER% (aOR=0.96, 95% CI: 0.94–0.97) and PR% (aOR=0.98, 95% CI: 0.96–0.99) were associated with lower odds of BCS after NACT. Increasing Ki-67% was associated with greater odds of BCS after NACT (aOR=1.07, 95% CI: 1.04–1.10). Pts with a low (aOR=0.50, 95% CI: 0.29–0.88) or intermediate (aOR=0.58, 95% CI: 0.41–0.81) RS were significantly less likely than pts with a high RS to undergo BCS after NACT. Asian pts were less likely than White pts to have BCS (aOR=0.75, 95% CI: 0.56–0.99), while rates of BCS after NACT were similar to pts from other racial and ethnic groups. In the NET cohort, increasing ER% was associated with greater odds of BCS (aOR=1.09, 95% CI: 1.01–1.17). There was no significant association between baseline PR% or Ki-67% and BCS after NET. Pts with a low (aOR=0.92, 95% CI: 0.60–1.43) or intermediate (aOR=0.96, 95% CI: 0.65–1.41) RS were numerically less likely than pts with a high RS to have BCS after NET, though not statistically significant. Compared with White pts, Asian pts were less likely to have BCS after NET (aOR=0.60, 95% CI: 0.40–0.91). Conclusions: In this analysis of pts from the NCDB who received neoadjuvant therapy for early-stage, HR+/HER2- breast cancer, we found that BCS after NACT was higher for patients with a high RS or high Ki-67, suggesting that NACT is unlikely to downstage tumors with a low/intermediate RS or low Ki-67. Most pts receiving NET underwent BCS (likely due to a smaller clinical tumor size); BCT after NET was most dependent on ER expression. Asian pts were less likely to undergo BCS after either NACT or NET, which is consistent with previous reports. These data could facilitate appropriate neoadjuvant treatment selection based on tumor biology and improve patient counseling on the likelihood of successful BCS. Citation Format: Jincong Freeman, Sarah Shubeck, Nan Chen, Rita Nanda, Dezheng Huo, Frederick Howard. Quantitative Biomarkers and 21-Gene Assay Results Predict Breast-Conserving Surgery Following Neoadjuvant Therapy in Early-Stage, Hormone Receptor-Positive, HER2-Negative Breast Cancer [abstract]. In: Proceedings of the 2023 San Antonio Breast Cancer Symposium; 2023 Dec 5-9; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2024;84(9 Suppl):Abstract nr PO1-14-12.

Postpartum loneliness predicts future depressive symptoms: a nationwide Japanese longitudinal study.

The COVID-19 pandemic has intensified feelings of loneliness, especially among postpartum women. This nationwide Japanese longitudinal study assessed the impact of such feelings on depressive symptoms and mother-to-infant bonding difficulties (MIBD), two pivotal determinants of maternal and infant well-being. Starting with a baseline survey conducted between July and August 2021, we tracked 1254 postpartum Japanese women who initially reported minimal depressive symptoms (i.e., Edinburgh Postnatal Depression Scale < 9) and MIBD (i.e., Mother-to-Infant Bonding Scale < 5), over a follow-up period of approximately 6months. Baseline loneliness was evaluated with the UCLA Loneliness Scale Short-Form (UCLA-LS3-SF3). Forty-nine percent of the sample reported the presence of baseline feelings of loneliness. After propensity score matching on sociodemographics and various pregnancy, childbirth, and COVID-19-related aspects, baseline loneliness was associated with increased risks of later depressive symptoms but not MIBD. Using restricted cubic spline logistic regression and considering loneliness as a continuous variable, we found a positive increasing quadratic relationship with depressive symptoms. As loneliness increased, so did the risk of later depressive symptoms. However, there was no significant association between loneliness and MIBD. These results were confirmed through a sensitivity analysis using inverse probability weighting to address attrition bias. Feelings of postpartum loneliness are associated with future risks of depressive symptoms. The data suggests that addressing loneliness in postpartum women early is crucial to safeguarding their well-being and that of their infants.

Body mass index, frailty, and outcomes in heart failure with preserved ejection fraction.

Relationship between body mass index (BMI), frailty, and clinical adverse events remains unclear in patients with heart failure (HF) with preserved ejection fraction (HFpEF) in different patient populations. We aimed to compare the association of BMI, frailty, and clinical adverse events between a US cohort from the Treatment of Preserved Cardiac Function Heart Failure with an Aldosterone Antagonist (TOPCAT) study and a Chinese cohort from the Heart Failure Registry of Patient Outcomes (HERO) study. We used data of 1715 participants enrolled from America in the TOPCAT study and 1487 patients with HFpEF in the Chinese registry study, the HERO. We evaluated the relationship between BMI and frailty using multivariate restricted cubic spline logistic regression. Association between frailty and BMI categories and primary outcomes including HF hospitalization, aborted sudden death, and cardiovascular death, all-cause mortality, and HF hospitalization were analysed by Cox proportional hazards models. The patients' mean age was 72±11years for both study populations, with 50% and 46% female for the TOPCAT study and the HERO study, respectively. Patients in the TOPCAT study had a higher mean BMI (33.9 vs. 24kg/m2), with 72.3% vs. 52.9% defined as moderately to severely frail (frailty index>0.3). In the TOPCAT study, risk of frailty rose as BMI increased, but not in the HERO study. Patients with frailty were at significant higher risk for the primary composite outcomes [hazard ratio (HR) 1.84 (95% confidence interval: 1.46-2.32)], all-cause mortality [HR 1.73 (1.34-2.25)], and HF hospitalization [HR 1.83 (1.40-2.40)] in the TOPCAT study. The corresponding numbers in the HERO study were 1.26 (1.01-1.57), 2.21 (1.45-3.35), and 1.15 (0.81-1.37), respectively. The association of frailty with clinical outcomes did not vary with BMI categories in the two studies. BMI distribution and association between BMI and frailty risk were different between the two study populations. Frailty was associated with clinical adverse events and this association was consistent across different BMI categories in both studies.

Liquidity ratios and corporate failures

AbstractOne of the most widespread claims in financial statement analysis is that liquidity ratios are useful for predicting failures. However, academic research has found surprisingly little empirical support for this claim. Using logistic regression splines, a non‐parametric method, this paper finds that the relation between the current ratio and failures differs significantly depending on the level of the current ratio. At low, but not high levels, the current ratio is significantly negatively related to failure. Incorporating such context provides statistically and economically significant predictive power about failures. These findings help resolve the discrepancy between practitioners and the academic literature.

Physical activity is inversely associated with overall cancer risk among college students in the United States: Results from the National College Health Assessment.

To our knowledge, this is the first epidemiologic study to examine the association between physical activity (PA) and cancer using data from the American College Health Association-National College Health Assessment (ACHA-NCHA). The goal of the study was to understand the dose-response relation between PA and cancer, as well as the associations between meeting US PA guidelines and overall cancer risk in US college students. The ACHA-NCHA provided self-reported information on demographic characteristics, PA, body mass index, smoking status, and overall cancer during 2019-2022 (n = 293,682; 0.08% cancer cases). To illustrate the dose-response relationship, a restricted cubic spline logistic regression analysis was used to evaluate the association of overall cancer with moderate-to-vigorous PA (MVPA) on a continuous basis. Logistic regression models were used to calculate odds ratios (ORs) and 95% confidence intervals for the associations between meeting the three U.S. PA guidelines and overall cancer risk. The cubic spline observed that MVPA was inversely associated with the odds of overall cancer risk after adjusting for covariates; a one hour/week increase in moderate and vigorous PA was associated with a 1% and 5% reduced overall cancer risk, respectively. Multivariable-adjusted logistic regression analyses showed that meeting the US guidelines for aerobic PA for adults (≥150 minutes/week of moderate aerobic PA or ≥75 minutes of vigorous PA) (OR: 0.85), for PA for adults (≥2 days of muscle strengthening activity in addition to aerobic MVPA) (OR: 0.90), and for highly active adults (≥2 days of muscle strengthening activity and ≥300 minutes/week of aerobic moderate PA or 150 minutes/week of vigorous PA) (OR: 0.89) were statistically significant and inversely associated with cancer risk. MVPA, especially meeting US guidelines, may be inversely associated with overall cancer among college students in the US. To reduce cancer risks, multilevel interventions to promote US physical activity guidelines among college students are warranted.

Association of systolic, diastolic, mean, and pulse pressure with morbidity and mortality in septic ICU patients: a nationwide observational study

BackgroundIntensivists target different blood pressure component values to manage intensive care unit (ICU) patients with sepsis. We aimed to evaluate the relationship between individual blood pressure components and organ dysfunction in critically ill septic patients.MethodsIn this retrospective observational study, we evaluated 77,328 septic patients in 364 ICUs in the eICU Research Institute database. Primary exposure was the lowest cumulative value of each component; mean, systolic, diastolic, and pulse pressure, sustained for at least 120 min during ICU stay. Primary outcome was ICU mortality and secondary outcomes were composite outcomes of acute kidney injury or death and myocardial injury or death during ICU stay. Multivariable logistic regression spline and threshold regression adjusting for potential confounders were conducted to evaluate associations between exposures and outcomes. Sensitivity analysis was conducted in 4211 patients with septic shock.ResultsLower values of all blood pressures components were associated with a higher risk of ICU mortality. Estimated change-points for the risk of ICU mortality were 69 mmHg for mean, 100 mmHg for systolic, 60 mmHg for diastolic, and 57 mmHg for pulse pressure. The strength of association between blood pressure components and ICU mortality as determined by slopes of threshold regression were mean (− 0.13), systolic (− 0.11), diastolic (− 0.09), and pulse pressure (− 0.05). Equivalent non-linear associations between blood pressure components and ICU mortality were confirmed in septic shock patients. We observed a similar relationship between blood pressure components and secondary outcomes.ConclusionBlood pressure component association with ICU mortality is the strongest for mean followed by systolic, diastolic, and weakest for pulse pressure. Critical care teams should continue to follow MAP-based resuscitation, though exploratory analysis focusing on blood pressure components in different sepsis phenotypes in critically ill ICU patients is needed.

The relative and interactive effects of urinary multiple metals exposure on hyperuricemia among urban elderly in China.

Independent and interactive effects of multiple metals levels in urine on the risk of hyperuricemia (HUA) in the elderly were investigated. A total of 6,508 individuals from the baseline population of the Shenzhen aging-related disorder cohort were included in this study. We detected urinary concentrations of 24 metals using inductively coupled plasma mass spectrometry, fitted unconditional logistic regression models, and the least absolute shrinkage and selection operator regression models for the selection of metals as well as unconditional stepwise logistic regression models and restricted cubic spline logistic regression models for assessing the associations of urinary metals and HUA risk, and finally applied generalized linear models to determine the interaction with urinary metals on the risk of HUA. Unconditional stepwise logistic regression models showed the association between urinary vanadium, iron, nickel, zinc, or arsenic and HUA risk (all P < 0.05). We revealed a negative linear dose-response relationship between urinary iron levels and HUA risk (P overall < 0.001, P nonliner = 0.682), a positive linear dose-response relationship between urinary zinc levels and HUA risk (P overall < 0.001, P nonliner = 0.513), and an additive interaction relationship between urinary low-iron and high-zinc levels and HUA risk (RERI = 0.31, 95% CI: 0.03-0.59; AP = 0.18, 95%CI: 0.02-0.34; S = 1.76, 95%CI: 1.69-3.49). Urinary vanadium, iron, nickel, zinc, or arsenic levels were associated with HUA risk, and the additive interaction of low-iron (<78.56 μg/L) and high-zinc (≥385.39 μg/L) levels may lead to a higher risk of HUA.

Ability of ordinal spline logistic regression model in the classification of nutritional status data

Ability of ordinal spline logistic regression model in the classification of nutritional status data

The use of the binary spline logistic regression model on the nutritional status data of children

The nutritional status of children can be grouped into two, namely normal and abnormal nutrition so that the data are analyzed using a binary logistic regression model. In this study, the use of binary logistic regression was developed on the use of the spline estimator as part of the nonparametric regression approach. This model is able to model qualitative response data by showing several trends that may occur in the data. Data on the nutritional status of children were analyzed based on the child's weight through a linear spline estimator and the optimal model was obtained at the use of a one-knot point, which was 4.7 kg. These results indicate that if a child has a body weight of 4.7 kg and above then there is a chance that the child has an abnormal nutritional status of 9.013 times compared to a child with a body weight below 4.7 kg. This could be due to the fact that children are no longer getting breast milk, so they need to get attention regarding their nutritional status.

Benign Hysterectomy Operative Times and 30-Day Complications: A Cohort Study

To evaluate the 30-day complication rate among different hysterectomy routes and operative times. A retrospective cohort study. American College of Surgeons National Surgical Quality Improvement Program database from 2011 to 2019. A total of 216 621 total cases including total abdominal hysterectomies (TAHs), total vaginal hysterectomies, total laparoscopic-assisted vaginal hysterectomies, and total laparoscopic hysterectomies. Eligible cases included benign hysterectomies with operative times between 20 minutes and 500 minutes. We excluded cases involving disseminated cancer, emergency surgery, supracervical approaches, or concomitant procedures. Multivariable logistic regression was used to evaluate the relationship between postoperative complications and operative time for each operative route. Multivariable logistic regression with a linear spline term was used to analyze differences in the association between postoperative complications and operative time below and above threshold operative times. Multivariable logistic regression demonstrated a significant association between operative time and overall complication rates for all hysterectomy routes. Spline logistic regression demonstrated a significant increase in adjusted odds of a complication occurring at or above the thresholds of 100 minutes for TAH. Patients undergoing a TAH for benign indications had a significantly increased odds of developing a complication within 30 days when operative time exceeded 100 minutes. Operative time may play a larger role in preoperative, intraoperative, and postoperative management than previously recognized for TAH in contrast to other hysterectomy routes.

Association Between the Prevalence of Frailty and Doubly Labeled Water-Calibrated Energy Intake Among Community-Dwelling Older Adults.

Appropriate energy intake (EI) is essential to prevent frailty. Because self-reported EI is inaccurate and has systematic errors, adequate biomarker calibration is required. This study examined the association between doubly labeled water (DLW)-calibrated EI and the prevalence of frailty among community-dwelling older adults. A cross-sectional study was performed using baseline data of 7,022 older adults aged ≥65 years in the Kyoto-Kameoka Study. EI was evaluated using a validated food frequency questionnaire (FFQ), and calibrated EI was obtained from a previously established equation using the DLW method. Physical and comprehensive frailty were defined by the Fried phenotype (FP) model and the Kihon Checklist (KCL), respectively. We used multivariable-adjusted restricted cubic spline logistic regression analysis. The prevalence of physical frailty was 14.8% and 13.6% in women and men, respectively. The spline models showed significant reverse J-shaped or U-shaped relationships between the prevalence of physical or comprehensive frailty against the DLW-calibrated EI, respectively. The lowest prevalence of both types of frailty was found at 1,900-2,000 kcal/d in women and 2,400-2,500 kcal/d in men, which corresponded to approximately 40 kcal/d/kg IBW (ideal body weight = 22 × height2) with DLW-calibrated EI. Uncalibrated EI underestimated approximately 20% compared with calibrated EI; underestimated EI were attenuated by calibration approach. This study suggests that low EI has a greater detrimental effect compared with excessive EI, particularly on physical frailty. Using biomarkers to calibrate EI holds promise for providing accurate energy requirements to establish guidelines used in public health and clinical nutrition.

A U-Shaped Relationship Between the Prevalence of Frailty and Body Mass Index in Community-Dwelling Japanese Older Adults: The Kyoto-Kameoka Study.

The relationship between body mass index (BMI) and frailty remains unclear. Using two validated frailty assessment tools, this study aimed to investigate the relationship between the prevalence of frailty and BMI in Japanese older adults. This cross-sectional study used baseline data of 7191 individuals aged ≥65 years, living in Kameoka City, Kyoto, Japan. The BMI was calculated based on self-reported height and body weight, and classified into six categories. Frailty was defined using two validated assessment tools, the Fried phenotype (FP) model and Kihon Checklist (KCL). We evaluated the relationship between frailty and BMI using a multivariate restricted cubic spline logistic regression. The prevalence of frailty defined using the FP model was 25.3%, 19.6%, 14.3%, 12.4%, 12.6%, and 19.4% for each BMI category of <18.5, 18.5–19.9, 20.0–22.4, 22.5–24.9, 25.0–27.4, and ≥27.5 kg/m2, respectively. The spline model showed a significant U-shaped relationship between BMI and the prevalence of frailty defined using both, KCL and FP models. This study found that the BMI range corresponding to lowest prevalence of frailty defined using both tools was 21.4–25.7 kg/m2. Thus, a healthy BMI may reduce the prevalence of frailty, and the risk of frailty needs to be evaluated in individuals who are underweight or overweight.

MP47-03 IS THERE A LEARNING CURVE PLATEAU FOR ACHIEVING TRIFECTA AND MINIMIZING OPERATIVE TIME (OT) FOR ROBOTIC-ASSISTED PARTIAL NEPHRECTOMY (RAPN)?

You have accessJournal of UrologySurgical Technology & Simulation: Training & Skills Assessment II (MP47)1 Apr 2020MP47-03 IS THERE A LEARNING CURVE PLATEAU FOR ACHIEVING TRIFECTA AND MINIMIZING OPERATIVE TIME (OT) FOR ROBOTIC-ASSISTED PARTIAL NEPHRECTOMY (RAPN)? Essa M. Bajalia*, Kevin A. Parikh, Daniela A. Haehn, Amanda E. Kahn, Colleen T. Ball, and David D. Thiel Essa M. Bajalia*Essa M. Bajalia* More articles by this author , Kevin A. ParikhKevin A. Parikh More articles by this author , Daniela A. HaehnDaniela A. Haehn More articles by this author , Amanda E. KahnAmanda E. Kahn More articles by this author , Colleen T. BallColleen T. Ball More articles by this author , and David D. ThielDavid D. Thiel More articles by this author View All Author Informationhttps://doi.org/10.1097/JU.0000000000000902.03AboutPDF ToolsAdd to favoritesDownload CitationsTrack CitationsPermissionsReprints ShareFacebookLinked InTwitterEmail Abstract INTRODUCTION AND OBJECTIVE: The complexities of RAPN can lead to a steep learning curve. We assessed the surgical learning curve and plateau point for achieving trifecta and minimizing OT in RAPN. METHODS: We evaluated 418 consecutive RAPN from a fellowship trained robotic surgeon between February 2008 and April 2019. The cases were separated into increments of 50. Cases were excluded from analysis if the patient had a prior RAPN performed at our institution (7 cases), if the patient was treated for calyceal diverticulum (4 cases), or if it was classified as a combo case (1 case). Trifecta was defined as warm ischemia time <25 minutes, negative surgical margins, and no grade 3 or higher postoperative complications. For the primary analysis we explored the correlation between surgical case number and patient outcomes using the Kendall correlation test. Age, sex, body mass index, tumor size, Mayo Adhesive Probability score, and R.E.N.A.L. score were all adjusted for as potential confounding variables in the spline regression methods. All statistical tests were two-sided. P values < 0.05 were considered statistically significant. RESULTS: Among the 406 eligible patients included in the study, 252 (62.1%) were male, median age was 63 years (range, 22 to 84), and median body mass index was 29 kg/m2 (IQR 26, 33). 272 patients achieved trifecta and average OT was 200 minutes. We found that RAPN experience (higher case number) was associated with shorter OT (P<0.001). OT (minutes) for the case increments were as follows: 1-50 (222.3±43.6), 51-100 (204.1±47.2), 101-150 (202.1±30.2), 151-200 (201.7±35.3), 201-250 (196.5±46.0), 251-300 (188.2±37.6), 301-350 (194.0±40.7), and 401-418 (186.1±40.4). Trifecta achievement for the increments were as follows: 1-50 (63%), 51-100 (82%), 101-150 (66%), 151-200 (67%), 201-250 (54%), 251-300 (71%), 301-350 (84%), 351-400 (74%), and 401-418 (72%). Although we did not find statistically significant overall increasing or decreasing trends in trifecta with surgeon experience, the restricted cubic spline logistic regression suggests peak performance with slightly under 100 cases (Figure 1). CONCLUSIONS: Maximizing OT performance and achievement of trifecta in RAPN appears to occur at a learning curve of about 100 cases. Source of Funding: None © 2020 by American Urological Association Education and Research, Inc.FiguresReferencesRelatedDetails Volume 203Issue Supplement 4April 2020Page: e684-e684 Advertisement Copyright & Permissions© 2020 by American Urological Association Education and Research, Inc.MetricsAuthor Information Essa M. Bajalia* More articles by this author Kevin A. Parikh More articles by this author Daniela A. Haehn More articles by this author Amanda E. Kahn More articles by this author Colleen T. Ball More articles by this author David D. Thiel More articles by this author Expand All Advertisement PDF downloadLoading ...

Probabilistic bisection with spatial metamodels

Probabilistic Bisection Algorithms perform root finding based on knowledge acquired from noisy oracle responses. We consider the generalized PBA setting (G-PBA) where the statistical distribution of the oracle is unknown and location-dependent, so that model inference and Bayesian knowledge updating must be performed simultaneously. To this end, we propose to leverage the spatial structure of a typical oracle by constructing a statistical surrogate for the underlying logistic regression step. We investigate several surrogates, including Binomial Gaussian Processes (B-GP), Polynomial, Kernel, and Spline Logistic Regression. In parallel, we develop sampling policies that adaptively balance learning the oracle distribution and learning the root. One of our proposals mimics active learning with B-GPs and provides a novel look-ahead predictive variance formula. The resulting gains of our Spatial PBA algorithm relative to earlier G-PBA models are illustrated with synthetic examples and a challenging stochastic root finding problem from Bermudan option pricing.

Immigrant apprehensions and birth outcomes: Evidence from California birth records 2008-2015.

While several studies find adverse birth outcomes among Latina mothers after discrete immigration enforcement events, it is unknown whether day-to-day enforcement activities precede adverse birth outcomes. This study examines birth outcomes among Latinas following local immigrant apprehensions-48-h holds on suspected undocumented immigrants by local law enforcement-over an 8-year period. County-level apprehensions, scaled to 1000 of the population, were averaged across the third trimester. We analyzed the association between county-level apprehensions and low birth weight (LBW) and preterm birth (PTB) between 2008 and 2015 in California using spline logistic regression models with knots at the quartiles of apprehensions and included covariates, county fixed-effects, and a time propensity variable to account for trend and seasonality in LBW and PTB. We performed these regressions for non-Hispanic White, all Latina, and foreign-born (FB) Latina mothers. There were no significant associations between apprehensions and LBW. For all Latina women, there were no differences in PTB between zero apprehensions and levels that were below the median. Latina mothers exposed to moderately high apprehensions had lower odds for PTB compared to zero apprehensions. At the highest levels of apprehensions, however, Latina women showed significantly elevated odds of PTB (at 0.71 apprehensions: OR=1.06, 95% CI 1.01-1.12). PTB responded differentially to localized enforcement threats. At lower and moderate levels, Latino communities may have mobilized in ways that reduced the risk for PTB. Risks emerged at the highest levels of enforcement, underlying the intense fear and anxiety in hyper-surveilled areas.

Out-of-hospital cardiac arrest: Probability of bystander defibrillation relative to distance to nearest automated external defibrillator

AimsDespite wide dissemination of automated external defibrillators (AEDs), bystander defibrillation rates remain low. We aimed to investigate how route distance to the nearest accessible AED was associated with probability of bystander defibrillation in public and residential locations. MethodsWe used data from the nationwide Danish Cardiac Arrest Registry and the Danish AED Network to identify out-of-hospital cardiac arrests and route distances to nearest accessible registered AED during 2008–2013. The association between route distance and bystander defibrillation was described using restricted cubic spline logistic regression. ResultsWe included 6971 out-of-hospital cardiac arrest cases. The proportion of arrests according to distance in meters (≤100, 101–200, >200) to the nearest accessible AED was: 4.6% (n=320), 5.3% (n=370), and 90.1% (n=6281), respectively. For cardiac arrests in public locations, the probability of bystander defibrillation at 0, 100 and 200m from the nearest AED was 35.7% (95% confidence interval 28.0%–43.5%), 21.3% (95% confidence interval 17.4%–25.2%), and 13.7% (95% confidence interval 10.1%–16.8%), respectively. The corresponding numbers for cardiac arrests in residential locations were 7.0% (95% confidence interval −2.1%–16.1%), 1.5% (95% confidence interval 0.002%–2.8%), and 0.9% (95% confidence interval 0.0005%–1.7%), respectively. ConclusionsIn public locations, the probability of bystander defibrillation decreased rapidly within the first 100m route distance from cardiac arrest to nearest accessible AED whereas the probability of bystander defibrillation was low for all distances in residential areas.

The dose-response relationship between tobacco smoking and the risk of lymphomas: a case-control study

BackgroundPrevious studies have provided limited support to the association between tobacco smoking and lymphomas with weak evidence of a dose-response relationship.MethodsWe investigated the relationship between tobacco smoking and risk of non-Hodgkin lymphomas (NHL) and Hodgkin lymphomas (HL) through logistic regression spline models. Data were derived from an Italian hospital-based case-control study (1999–2014), which enrolled 571 NHLs, 188 HLs, and 1004 cancer-free controls. Smoking habits and other lifestyle factors were assessed through a validated questionnaire. Odds ratios (OR) and 95% confidence intervals (CI) were estimated by logistic regression, adjusting for potential confounders.ResultsCompared to never smokers, people smoking ≥15 cigarettes/day showed increased risks of both NHL (OR = 1.42, 95% CI: 1.02, 1.97) and HL (OR = 2.47, 95% CI: 1.25, 4.87); the risk was particularly elevated for follicular NHL (OR = 2.43; 95% CI:1.31–4.51) and mixed cellularity HL (OR = 5.60, 95% CI: 1.31, 23.97). No excess risk emerged for former smokers or people smoking <15 cigarettes/day. Spline analyses showed a positive dose-response relationship with significant increases in NHL and HL risks starting from 15 and 21 cigarettes/day, respectively, with the most evident effects for follicular NHL and mixed cellularity HL. Smoking duration was significantly associated with the HL risk only (OR = 2.15, 95% CI: 1.16, 3.99).ConclusionsThese findings support a role of tobacco smoking in the etiology of both NHL and HL, providing evidence of a direct association of risk with smoking intensity.

Routine biomarkers are strong predictors of short term mortality

Methods A prospective observational cohort study of 6,279 consecutive patients admitted from the Emergency Department of Hillerod University Hospital. All triaged patients with a full biochemical screening (albumin, creatinine, CRP, haemoglobin, Lactate dehydrogenase, leukocytes, potassium, and sodium) were included. Vital status was collected from the Danish Central Office of Civil Registration. The primary endpoint was 30-day mortality. Secondary endpoints were admission to intensive care unit (ICU) and readmission. Univariate logistic regression splines were made for all eight biomarkers with cuts defined after internationally accepted reference intervals. These models were used to create a multivariate logistic regression spline including all eight biomarkers, and discriminative ability was evaluated with receiver operation characteristics and area under the curve (AUC). Ultimately predicted risks of mortality based on the biomarkers were calculated for all patients, and they were divided into four groups: Green 25%.

Dose-response relationships between sedentary behaviour and the metabolic syndrome and its components.

The aim of this study was to examine the relationship among sedentary behaviour (SB) and the metabolic syndrome and its components by age, moderate-to-vigorous physical activity (MVPA) and sex. A cross-sectional analysis was performed on 2003-2006 National Health and Nutrition Examination Survey data from 5,076 adults aged ≥18 years (mean ± SD = 43.8 ± 19.5). SB was measured using ActiGraph accelerometers worn for 1 week and defined as <100 counts/min. Metabolic syndrome was defined using the Adult Treatment Panel III criteria. Natural cubic spline logistic regression models were used to estimate the odds of meeting criteria for the metabolic syndrome and its components by total daily SB time and breaks in SB. Statistical interactions between SB and age, sex and MVPA were explored. The prevalence of the metabolic syndrome was 19% and the average daily SB time was 8.1 ± 2.8 h, with 90 ± 25 breaks/day. The relationship between daily SB time and the metabolic syndrome was linear and characterised by an OR of 1.09 (95% CI 1.01, 1.18) for each hour of SB. Total SB was associated with the following components: high triacylglycerol, low HDL-cholesterol and high fasting glucose. All three associations were modified by MVPA level. No relationship between breaks in SB and the metabolic syndrome was found. There appears to be no SB threshold at which the risk of the metabolic syndrome is elevated. Therefore, an effort should be made to maintain low levels of total time spent in SB and so lessen the risk of the metabolic syndrome.