Covariate Balance Research Articles

Metabolomic profiling identifies signatures and biomarkers linking air pollution to dementia risk: a prospective cohort study

Exposure to air pollution has been associated with increased dementia. However, it remains unknown what specific metabolic mechanisms play a role in this relationship. We included 192,300 dementia-free participants from the UK Biobank cohort study. Annual concentrations of air pollution were assessed based on the residential address. Elastic net regression was performed to identify air pollution-related metabolites, and metabolic score was constructed. Cox regression models and covariate balancing generalized propensity scores (CBGPS) regression models were conducted to explore the longitudinal associations between air pollution/metabolic signatures and dementia risk. The underlying mechanisms between air pollution and dementia driven by metabolic signature or specific metabolites were also investigated. A total of 2,592 incident dementia cases were documented. We identified the metabolite profiles in response to air pollution exposure, including 87 metabolites for PM2.5, 65 metabolites for PM10, 76 metabolites for NO2, and 71 metabolites for NOx. The air pollution-related metabolic signatures were associated with increased risk of dementia, with hazard ratios (HR) of 1.17 (95% CI: 1.12, 1.22), 1.06 (95% CI: 1.02, 1.11), 1.16 (95% CI: 1.10, 1.21), and 1.17 (95% CI: 1.12, 1.22) for PM2.5, PM10, NO2 and NOx, respectively. The associations persisted using causal models. Metabolic signatures mediated the associations between air pollution exposure and dementia risk, with mediation proportions ranging from 6.57% to 12.71%. Additionally, we observed that a metabolite known as free cholesterol in medium VLDL (M-VLDL-FC) played a crucial mediating role. Our study provides novel insights into the metabolic mechanisms linking air pollution exposure to dementia risk.

Abstract 4138721: Outcomes of Coronary Revascularization in Patients with Contemporaneous Abdominal Organ Transplant

Background: Clinical evaluation surrounding liver and kidney transplantation may reveal obstructive coronary artery disease (CAD). However, little is known about outcomes after revascularization in this population. Our aim was to examine coronary artery bypass grafting (CABG) and percutaneous intervention (PCI) in peri-transplant patients compared to the general CABG/PCI populations. Methods: We conducted a matched retrospective cohort study between 2014-2022 at a single academic center. Matching was done by optimal full matching of a logistic regression-based propensity score adjusting for age, chronic kidney disease, diabetes, tobacco use, chronic lung disease, hypertension, prior myocardial infarction, and valvular disease. Covariate balance was assessed as standardized mean difference <0.2. The primary outcome was a composite of in-hospital death, stroke, and myocardial infarction, and a key secondary outcome was incident arrhythmia. Results: The study population comprised 2655 patients: 24 CABG/transplant (median age, 64 [IQR, 59-66]; 4 female [17%]), 1037 CABG/non-transplant (age, 66 [59-73]; 190 female [18%]), 27 PCI/transplant (age, 62 [IQR, 55-67]; 6 female [22%]), and 1567 PCI/non-transplant (age, 70 [IQR, 60-78]; 399 female [25%]). CABG/PCI preceded transplant by a median 321/324 days, respectively, and followed transplant by a median 461/36 days, respectively. For the primary outcome in CABG patients, there were no events in the transplant group and an estimated potential outcome in 6.5% of the matched non-transplant group (marginal RR, <0.001; p<0.001). For the primary outcome in PCI patients, there were no events in the transplant group and an estimated potential outcome in 2.4% of the matched non-transplant group (marginal RR, <0.001; p<0.001). For incident arrhythmia in CABG patients, the estimated potential outcome was 12% and 17.5% in the transplant and matched non-transplant groups, respectively (marginal RR, 0.98; 95% CI, 0.37-2.6; p=0.97). For incident arrhythmia in PCI patients, the estimated potential outcome was <0.001% and 3.6% in the transplant and matched non-transplant groups, respectively (marginal RR, 0.89; 95% CI, 0.12-6.5; p=0.91). Conclusions: Patients with contemporaneous abdominal organ transplant undergoing coronary revascularization suffered no in-hospital mortality, stroke, or myocardial infarction, and were not at greater risk of incident arrhythmia than similar non-transplant patients.

Abstract 4137905: The association between prehospital epinephrine administration and short-term outcomes in patients with shockable out-of-hospital cardiac arrest and extracorporeal cardiopulmonary resuscitation: a propensity matched analysis

Background: In out-of-hospital cardiac arrest (OHCA) patients with an initial shockable rhythm, epinephrine increases the likelihood of return of spontaneous circulation (ROSC), but its effect on neurological outcome remains uncertain. Epinephrine administered before extracorporeal resuscitation (ECPR), which is a non-pharmacological method to obtain ROSC, may have worsened neurological outcome. Aim: To evaluate the impact of prehospital epinephrine administration on the prognosis of OHCA patients undergoing ECPR. Methods: This is a retrospective analysis of a cohort study from a multicenter, prospective registry of 81234 OHCA patients between 2014 and 2021. After the application of exclusion criteria, 1061 OHCA patients with an initial shockable rhythm and who underwent ECPR were eligible for this study. The primary outcome was favorable neurological outcome at 30 days after the OHCA and the secondary outcome was ROSC during transfer. Patients who did and did not receive prehospital epinephrine were propensity score-matched on the basis of age, gender, witness arrest, bystander-initiated CPR, dispatcher-assisted CPR, advanced airway management, call-to-defibrillation interval, year and district. Results: Among 1061 eligible patients, 442 patients received epinephrine and 619 patients did not. Matching was successful in achieving covariate balance as shown by a standardized difference of <10% for all variables (329 matched pairs). The proportion of patients with prehospital ROSC was significantly higher in those who received prehospital epinephrine than those who did not (36 [8%] vs 31 [5%]; adjusted odds ratio [OR] after propensity score-matching, 1.96 [95% CI 1.05-3.67], p=0.03). The proportion of patients with favorable neurological outcome at 30 days was not significantly lower in those who received prehospital epinephrine than those who did not (34 [8%] vs 78 [13%]; adjusted OR after propensity score-matching, 0.78 [95% CI 0.45-1.37], p=0.39). Conclusion: In patients with shockable OHCA and ECPR, prehospital epinephrine administration was significantly associated with increased ROSC before hospital arrival and not significantly with decreased favorable neurological outcome at 30 days.

Using Genomic Heterogeneity to Inform Therapeutic Decisions for Metastatic Colorectal Cancer: An Application of the Value of Heterogeneity Framework.

Mutations in KRAS and NRAS are predictive of poor response to cetuximab and panitumumab, two anti-epidermal growth factor receptor (EGFR) monoclonal antibodies used in metastatic colorectal cancer (mCRC). Our objective was to explore the value of using KRAS and NRAS mutation status to inform third-line anti-EGFR therapy for mCRC using the value of heterogeneity (VOH) framework. We used administrative data to identify mCRC patients who were potentially eligible for third-line therapy in 2006-2019 in British Columbia (BC), Canada. We compared three alternative stratification policies in place during the study period: the unstratified policy where anti-EGFR therapy was not offered (2006-2009), stratification by KRAS mutation (2009-2016), and stratification by KRAS+NRAS mutation (2016-2019). We used inverse-probability-of-treatment weighting to balance covariates across the three groups. Cost and survival time were calculated using a 3-year time horizon and adjusted for censoring, with bootstrapping to characterize uncertainty. Mean net monetary benefit (NMB) was calculated at a range of threshold values. The VOH of using KRAS and NRAS mutation status to inform treatment selection was calculated as the change in NMB with increasing stratification, under current (static VOH) or perfect (dynamic VOH) information. We included 2664 patients in the analysis. At a willingness-to-pay of CA$100,000/ life-year gained (LYG), stratification on KRAS mutation status provided a static VOH of CA$1565 per patient; further stratification on KRAS+NRAS provided additional static VOH of CA$594. The static VOH exceeded the marginal cost of genomic testing under both policies. Stratification of anti-EGFR therapy by KRAS and NRAS mutation status can provide additional value at a threshold of CA$100,000/LYG. There is diminishing marginal value and increasing marginal costs as the policy becomes more stratified. The VOH framework can illustrate the value of subgroup-specific decisions in a comprehensive way, to better inform targeted treatment policies.

COVID-19 vaccination modified the effect of nirmatrelvir-ritonavir on post-acute mortality and rehospitalization: a retrospective cohort study.

While previous research examined coronavirus disease 2019 (COVID-19) antiviral-vaccine interactions through exploratory subgroup analysis, none specifically designed for examining this interaction or its impact on post-acute outcomes. This study examined the interaction between nirmatrelvir-ritonavir and complete COVID-19 vaccination on reducing the risk of post-acute outcomes among COVID-19 patients. We followed COVID-19 patients hospitalized between 11 March 2022 and 10 October 2023, until 31 October 2023 in Hong Kong. Exposure groups were based on nirmatrelvir-ritonavir usage and vaccination status (fully or not fully vaccinated). Post-acute death and all-cause rehospitalization were the study outcomes. Propensity score weighting was applied to balance covariates among exposure groups, including age, sex, Charlson Comorbidity Index, and concomitant treatments. Multiplicative and additive interactions between nirmatrelvir-ritonavir and vaccination status were assessed. A total of 50,438 COVID-19 patients were included in this study and arranged into four exposure groups. Significant additive interaction on post-acute rehospitalization was observed (relative excess risk, 0.10; 95% CI, 0.02-0.19; p-value, 0.018; attributable proportion, 0.07; 95% CI, 0.01-0.12; p-value, 0.017; synergy index, 1.26; 95% CI, 1.02-1.55; p-value, 0.032). The interaction on post-acute mortality was marginally significant. In the subgroup analysis, the interaction effect is more pronounced in older adults, female, and CoronaVac recipients. In conclusion, our study demonstrated an additive interaction between nirmatrelvir-ritonavir and complete vaccination on post-acute outcomes, suggesting greater long-term benefits of the antiviral for fully vaccinated individuals compared to not fully vaccinated patients.

Covariate balancing for high-dimensional samples in controlled experiments

In controlled experiments, achieving covariate balancing across all groups is crucial as it ensures that the estimated treatment effects are not confounded by the effects of covariates. This study proposes a mixed-integer nonlinear programming model to address the covariate balancing problem. Specifically, we introduce a new covariate imbalance measure, which is the maximum discrepancy in both the first and second central moments between any two groups. The second central moment can effectively capture the correlation of covariates in a physical sense, which is crucial for partitioning high-dimensional samples. A mixed-integer nonlinear programming model is constructed to minimize the proposed measure to obtain the optimal partitioning results. The nonlinear model is then linearized to accelerate the optimization process. We conduct computational experiments based on simulated datasets, including one-dimensional, two-dimensional, and three-dimensional Gaussian distributed samples, and a real clinic trial dataset. Compared to the conventional discrepancy-based method, our method achieves a 54.81% and a 40.6% reduction in the maximum discrepancy of partitioning results in the two-dimensional simulated Gaussian samples and the real clinic trial dataset, respectively. These results demonstrate the superiority of the proposed model in partitioning high-dimensional samples with correlated covariates compared with the conventional discrepancy-based method.

Machine learning methods for propensity and disease risk score estimation in high-dimensional data: a plasmode simulation and real-world data cohort analysis.

Machine learning (ML) methods are promising and scalable alternatives for propensity score (PS) estimation, but their comparative performance in disease risk score (DRS) estimation remains unexplored. We used real-world data comparing antihypertensive users to non-users with 69 negative control outcomes, and plasmode simulations to study the performance of ML methods in PS and DRS estimation. We conducted a cohort study using UK primary care records. Further, we conducted a plasmode simulation with synthetic treatment and outcome mimicking empirical data distributions. We compared four PS and DRS estimation methods: 1. Reference: Logistic regression including clinically chosen confounders. 2. Logistic regression with L1 regularisation (LASSO). 3. Multi-layer perceptron (MLP). 4. Extreme Gradient Boosting (XgBoost). Covariate balance, coverage of the null effect of negative control outcomes (real-world data) and bias based on the absolute difference between observed and true effects (for plasmode) were estimated. 632,201 antihypertensive users and nonusers were included. ML methods outperformed the reference method for PS estimation in some scenarios, both in terms of covariate balance and coverage/bias. Specifically, XgBoost achieved the best performance. DRS-based methods performed worse than PS in all tested scenarios. We found that ML methods could be reliable alternatives for PS estimation. ML-based DRS methods performed worse than PS ones, likely given the rarity of outcomes.

Fisher Meets Bayes: The Value of Randomisation for Bayesian Inference of Causal Effects

SummaryFor a Bayesian agent with beliefs about the relationship between covariates and potential outcomes, deterministically selecting an assignment that yields optimal covariate balance rationally dominates randomisation. However, randomisation—by enabling control over the probabilities of erroneous causal conclusions due to unknown covariate imbalances—offers insurance against the possibility that an agent's beliefs may be misleading. For the most part, such rational justifications for optimum assignment have presupposed the framework of Bayesian inference, while such epistemic justifications for randomisation have presupposed the framework of significance testing. In this paper, I build on a conception of balance that seems inextricable from the significance testing framework, Fisherian balance, to show that it implies an analogous epistemic justification for randomisation within the framework of Bayesian inference. Consequently, for the choice between optimum and random assignment, this paper shows that epistemic justifications need not be wedded to significance testing nor must Bayesian inference be wedded to rational justifications.

Sodium zirconium cyclosilicate treatment and rates of emergency interventions for hyperkalaemia: a propensity-score weighted case-control study

Abstract Background Sodium zirconium cyclosilicate (SZC) reduces serum potassium in patients with chronic hyperkalaemia in clinical trials, but its role in the emergency treatment of hyperkalaemia is unproven. We hypothesised that SZC use for emergent hyperkalaemia would be associated with a reduction in rates of emergency interventions for hyperkalaemia. Methods This was a single centre, propensity-score weighted case-control study of patients admitted with hyperkalaemia to a specialist renal centre. We randomly selected 250 patients admitted between April 2021 and September 2022 (post-SZC era) with a potassium ≥ 5.5 mmol/L treated with at least one ≥ 10g dose of SZC (treatment group). We randomly selected a comparator group of 250 patients admitted between January 2018 and December 2019 (pre-SZC era) with a potassium ≥ 5.5 mmol/L (control group). Baseline demographic and clinical characteristics were recorded and used as covariates for propensity scoring (PS) and inverse probability treatment weighting (IPTW). Our primary outcome measure, rates of emergency haemodialysis, were tested using unadjusted models and multivariable logistic regression models on unweighted data in addition to unadjusted models on weighted data. We also reviewed rates of emergency temporary central venous access as a secondary outcome. Results 59% were male, the mean age was 67 (SD 14) years and 149 (30%) were receiving maintenance dialysis. IPTW achieved satisfactory balance of covariates between the treatment and control groups. In the treatment group, patients were 77% less likely to need emergency haemodialysis (OR 0.23; CI 0.17 to 0.31); this result was consistent following analysis of weighted and unweighted data. Similarly, patients treated with SZC were 73% less likely to require emergency temporary central venous access (OR 0.27; CI 0.20–0.36). Conclusion SZC was associated with signification reduction in rates of emergency haemodialysis and emergency temporary central venous access in patients admitted to a specialised renal centre with emergent hyperkalaemia.

Higher fraction of inspired oxygen during anesthesia increase the risk of postoperative pulmonary complications in patients undergoing non-cardiothoracic surgery: a retrospective cohort study.

The ideal intra-operative inspired oxygen concentration remains controversial. We aimed to investigate the association between the intraoperative fraction of inspired oxygen (FiO2) and the incidence of postoperative pulmonary complications (PPCs) in patients undergoing non-cardiothoracic surgery. This was a retrospective cohort study of elderly patients who underwent non-cardiothoracic surgery between April 2020 and January 2022. According to intraoperative FiO2, patients were divided into low (≤60%) and high (>60%) FiO2 groups. The primary outcome was the incidence of a composite of pulmonary complications (PPCs) within the first seven postoperative days. Propensity score matching (PSM) and inverse probability treatment weighting (IPTW) were conducted to adjust for baseline characteristic differences between the two groups. Multivariate logistic regression analysis was used to calculate the odds ratios (OR) for FiO2 and PPCs. Among the 3,515 included patients with a median age of 70years (interquartile range: 68-74), 492 (14%) experienced PPCs within the first 7 postoperative days. Elevated FiO2 was associated with an increased risk of PPCs in all the logistic regression models. The OR of the FiO2 > 60% group was 1.252 (95%CI, 1.015-1.551, P = 0.038) in the univariate analysis. In the multivariate logistic regression models, the ORs of the FiO2 > 60% group were 1.259 (Model 2), 1.314 (Model 3), and 1.32 (model 4). A balanced covariate distribution between the two groups was created using PSM or IPTW. The correlation between elevated FiO2 and an increased risk of PPCs remained statistically significant with PSM analysis (OR, 1.393; 95% CI, 1.077-1.804; P = 0.012) and IPTW analysis (OR, 1.266; 95% CI, 1.086-1.476; P = 0.003). High intraoperative FiO2 (>60%) was associated with the postoperative occurrence of pulmonary complications, independent of predefined risk factors, in elderly non-cardiothoracic surgery patients. High intraoperative FiO2 should be applied cautiously in surgical patients vulnerable to PPCs.

Household size and depressive symptoms during the COVID-19 pandemic: A gendered analysis

ABSTRACT The study explores the nature of the relationship between household size and depressive symptoms under COVID-19 strict lockdown conditions, when a surge in large-size households observed in South Africa. The study used nationally representative data to implement covariate balancing propensity scores within a multivalued treatment framework (by varying household size) to control for observables. While the study found that both solo living and large-sized households are positively associated with depressive symptoms for males, the depressive symptoms for females are not found to be strongly associated with household size. Further, the analysis indicates that the higher average depressive symptoms observed for males compared to females under the COVID-19 pandemic are largely driven by males in large-size households. The gendered effect points to the different roles that men and women continue to play within the household. The need for policy interventions to consider living arrangements is highlighted through this study.

Comparative retrospective study on surgical outcomes of hand‐sewn anastomosis versus stapling anastomosis for colectomy using a nationwide inpatient database in Japan with propensity score matching

AbstractBackgroundIntestinal anastomosis is performed by two main methods: hand‐sewn anastomosis (HA) and stapling anastomosis (SA). Studies on anastomosis are still being reported and are an ongoing area of interest. The aim of this study was to evaluate the characteristics of each by comparing them with description and statistical analysis using a nationwide inpatient database.MethodsWe collected data for colon cancer patients who underwent colectomy between April 2014 and March 2022 using the Diagnosis Procedure Combination (DPC) database. In the descriptive analysis, we described the background of the study population and the number of each method. In the statistical analysis, we used propensity score matching to achieve balanced covariates and showed the results of the comparative study using risk ratios and risk differences.ResultsA total of 232 155 cases were selected (HA group: 40 764; SA group: 191 391). SA was performed more frequently than HA during the study period, with the proportion of SA increasing over the last few years (~80% in 2015 and 85% in 2021). We obtained 40 760 pairs through propensity score matching. The comparison of postoperative events varied by outcome (reoperation: 517 [1.27%] vs 380 [0.93%]; risk ratio [RR]: 1.36 [95% confidence interval [CI], 1.19 to 1.55], endoscopic intestinal hemostasis: 17 [0.04%] vs 80 [0.20%]; RR: 0.21 [95% CI, 0.13 to 0.36]).ConclusionSA is becoming more common in colectomy. The risk of reoperation could be higher in the HA group, while the risk of postoperative endoscopic intestinal hemostasis could be higher in the SA group.

Evaluation of preoperative antibiotic prophylaxis in clean-wound general surgery procedures: a propensity score-matched cohort study at a regional hospital.

The administration of antibiotic prophylaxis for clean-wound surgeries is controversial among surgeons, despite guidelines suggesting its use. This study aimed to evaluate its effectiveness in preventing surgical site infections (SSIs) in clean-wound surgeries within a regional setting with varied practices regarding prophylaxis. This retrospective cohort study included four types of common general surgeries performed from March 2021 to February 2023 at a tertiary regional hospital in Thailand. The surgeries included skin/subcutaneous excision, thyroidectomy, inguinal hernia repair, and breast surgeries, all of which required regional or general anesthesia. Antibiotic prophylaxis was administered at the discretion of the attending surgeons. SSI diagnosis followed standard diagnostic criteria, involving reviewing medical records and the records of the infection control unit. Infection risk factors were examined. The primary outcome comparison used inverse probability treatment weighting of propensity scores, with covariate balance evaluated. Of the 501 surgeries identified, 84 were excluded, leaving 417 eligible for analysis. Among these patients, 233 received prophylactic antibiotics, for an SSI rate of 1.3%, while 184 did not receive antibiotics, for an SSI rate of 2.2%. A comparative analysis using propensity score weighting revealed no statistically significant difference in the incidence of SSI between the groups (risk ratio [95% confidence interval]: 0.54 (0.11, 2.50), p = 0.427). In this practical setting, with the given study size, antibiotic prophylaxis in common general surgeries involving clean wounds did not significantly prevent SSIs. Routine use recommendations should be re-evaluated. Not applicable as this study is a retrospective cohort study and not a clinical trial.

Comparison of pathologic response and survival outcomes between neoadjuvant chemoradiotherapy (nCRT) and neoadjuvant immunochemotherapy (nICT) in patients with locally advanced esophageal squamous cell carcinoma: a propensity score-matched analysis

BackgroundIn locally advanced, operable esophageal squamous cell carcinoma (ESCC), neoadjuvant immunochemotherapy (nICT) has shown results that are somewhat comparable to those of standard neoadjuvant chemoradiotherapy (nCRT). The impact of these neoadjuvant treatments on survival outcomes, however, has yet to be elucidated.MethodsThis study included 489 patients with locally advanced ESCC who underwent surgery at Sichuan Cancer Hospital after receiving neoadjuvant treatment between June 2017 and September 2023. Patients were categorized into nCRT and nICT groups based on whether they received neoadjuvant treatment. To mitigate potential biases and balance covariates between the two cohorts, 1:2 propensity score matching (PSM) was conducted using a caliper width of 0.05.ResultsAfter PSM, the baseline characteristics of the 360 patients remained balanced between the two groups. The findings indicated a superior pathological response in the nCRT group, as evidenced by significantly greater rates of complete response (32.87% vs 14.58%, P < 0.001) and favorable tumor regression grade (TRG), as well as reduced ypT stages and less perineural and angioinvasion, despite comparable ypN stages. Despite the improvement in complete pathological response (pCR) in the nCRT group, the 3-year disease-free survival (DFS) and overall survival (OS) rates did not significantly differ between the groups (DFS: 58.32% vs 56.16%, P = 0.67; OS: 69.96% vs 71.99%, P = 0.99). Crucially, The nICT group showed a lower incidence of grade 3 and 4 adverse events in Leukopenia (2.8% vs 29%; P < 0.001) and Neutropenia (2.8% vs 24%; P < 0.001) during neoadjuvant treatment, comparing with nCRT group.ConclusionsOur preliminary findings suggest that nICT followed by surgery offers comparable survival rates to nCRT, despite being less effective in pathologic outcomes. Nonetheless, nICT is a safe and feasible strategy for locally advanced ESCC, warranting further exploration to understand its impact on long-term survival.

Prognostic score-based model averaging approach for propensity score estimation

BackgroundPropensity scores (PS) are typically evaluated using balance metrics that focus on covariate balance, often without considering their predictive power for the outcome. This approach may not always result in optimal bias reduction in the treatment effect estimate. To address this issue, evaluating covariate balance through prognostic scores, which account for the relationship between covariates and the outcome, has been proposed. Similarly, using a typical model averaging approach for PS estimation that minimizes prediction error for treatment status and covariate imbalance does not necessarily optimize PS-based confounding adjustment. As an alternative approach, using the averaged PS model that minimizes inter-group differences in the prognostic score may further reduce bias in the treatment effect estimate. Moreover, since the prognostic score is also an estimated quantity, model averaging in the prognostic scores can help identify a better prognostic score model. Utilizing the model-averaged prognostic scores as the balance metric for constructing the averaged PS model can contribute to further decreasing bias in treatment effect estimates. This paper demonstrates the effectiveness of the PS model averaging approach based on prognostic score balance and proposes a method that uses the model-averaged prognostic score as a balance metric, evaluating its performance through simulations and empirical analysis.MethodsWe conduct a series of simulations alongside an analysis of empirical observational data to compare the performances of weighted treatment effect estimates using the proposed and existing approaches. In our examination, we separately provid four candidate estimates for the PS and prognostic score models using traditional regression and machine learning methods. The model averaging of PS based on these candidate estimators is performed to either maximize the prediction accuracy of the treatment or to minimize intergroup differences in covariate distributions or prognostic scores. We also utilize not only the prognostic scores from each candidate model but also an averaged score that best predicted the outcome, for the balance assessment.ResultsThe simulation and empirical data analysis reveal that our proposed model-averaging approaches for PS estimation consistently yield lower bias and less variability in treatment effect estimates across various scenarios compared to existing methods. Specifically, using the optimally averaged prognostic scores as a balance metric significantly improves the robustness of the weighted treatment effect estimates.DiscussionThe prognostic score-based model averaging approach for estimating PS can outperform existing model averaging methods. In particular, the estimator using the model averaging prognostic score as a balance metric can produce more robust estimates. Since our results are obtained under relatively simple conditions, applying them to real data analysis requires adjustments to obtain accurate estimates according to the complexity and dimensionality of the data.ConclusionsUsing the prognostic score as the balance metric for the PS model averaging enhances the performance of the treatment effect estimator, which can be recommended for a wide variety of situations. When applying the proposed method to real-world data, it is important to use it in conjunction with techniques that mitigate issues arising from the complexity and high dimensionality of the data.

Voluntary sustainability standards and technical efficiency of Honduran smallholder coffee producers

Sustainable coffee production promises to improve production techniques and enhance the socioeconomic conditions of smallholder farmers. Using primary survey data from 659 coffee producers in Honduras, this study assesses the impact of voluntary sustainability standards (VSS) on the technical efficiency of smallholder coffee production. The article uses the Stochastic Production Frontier Analysis model to analyze and compare the technical efficiency (TE) of certified and non-certified coffee producers. To provide reliable comparability between groups, the dataset was balanced using Covariate Balancing Propensity Score (CBPS). The results show that the mean technical efficiency was 52.86% for pooled certified farmers and 55.56% for non-certified smallholder coffee producers. Specifically, the technical efficiency of 4C farmers was 51.58%, 53.82% for the Fairtrade group, 60.56% for RA farmers, and 60.15% for the UTZ group, indicating substantial inefficiencies in the coffee production of the different certified groups. Results from Tobit’s model for the determinants of TE indicated that variables such as the age of the household head, access to credit, and training attendance are among the main factors that significantly drive the technical efficiency of certified and non-certified farmers. Based on the findings, enhancing education opportunities, improving infrastructure for better market access and farm management, and expanding credit access are recommended to improve efficiency in the study area. Honduran smallholder coffee producers have considerable potential to increase output with existing technology by improving their technical efficiency. Therefore, stakeholders’ efforts should focus on enhancing efficiency levels and capitalizing on potential gains for both certified and non-certified farmers, to ultimately improve the farmers’ livelihoods.

Timing of induction of labor in suspected macrosomia: retrospective cohort study, systematic review and meta-analysis.

Large-for-gestational age (LGA) is associated with several adverse maternal and neonatal outcomes. Although many studies have found that early induction of labor (IOL) in case of a LGA fetus reduces the incidence of shoulder dystocia, no current guidelines recommend this particular clinical strategy, owing to concerns about increased rates of Cesarean delivery (CD) and neonatal complications. The purpose of this study was to assess whether the timing of IOL in LGA fetuses affected maternal and neonatal outcomes in a single center, and to combine these results with evidence reported in the literature. This study comprised two parts. The first part was a retrospective cohort study that included consecutive patients with a singleton pregnancy and an estimated fetal weight ≥ 90th percentile on ultrasound between 35 + 0 and 39 + 0 weeks' gestation, who were eligible for normal vaginal delivery. The second part of the study was a systematic review of the literature and meta-analysis, including the results of our cohort study as well as those of previous studies that compared IOL with expectant management in patients with a LGA fetus. The perinatal outcomes of the study were CD, operative vaginal delivery, shoulder dystocia, brachial plexus palsy, anal sphincter injury, postpartum hemorrhage, Apgar score, umbilical artery pH, admission to the neonatal intensive care unit, use of continuous positive airway pressure, intracranial hemorrhage, need for phototherapy and bone fracture. Of the 547 patients included in this retrospective cohort study, 329 (60.1%) underwent IOL and 218 (39.9%) experienced spontaneous labor. Following covariate balancing, the odds of CD were significantly higher in the IOL group compared with the spontaneous-labor group. This difference only became apparent beyond 40 weeks' gestation (hazard ratio, 1.90; P = 0.030). The difference between the IOL and spontaneous-labor groups for the rate of shoulder dystocia was not statistically significant (hazard ratio, 1.57; P = 0.200). Seventeen studies, in addition to our own results, were included in the systematic review and meta-analysis, giving a total population of 111 300 participants. Although there was no significant difference in the rate of CD between IOL and expectant management after pooling the results of included studies, the risk for shoulder dystocia was significantly lower in the IOL group (odds ratio (OR), 0.64 (95% CI, 0.42-0.98); I2 = 19% from 12 studies) when considering only IOL performed before 40 + 0 weeks. When the studies in which IOL was carried out exclusively before 40 + 0 weeks were removed from the analysis, the risk for CD in the remaining studies was significantly higher in the IOL group (OR, 1.46 (95% CI, 1.02-2.09); I2 = 56%). There were no statistically significant differences between the IOL and expectant-management groups for the remaining perinatal outcomes. Nulliparity, history of CD and low Bishop score, but not method of induction, were independent risk factors for intrapartum CD in patients that underwent IOL for LGA. The timing of IOL in patients with suspected macrosomia significantly impacts on perinatal adverse outcomes. IOL has no impact on rates of shoulder dystocia but increases the odds of CD when considered irrespective of gestational age; in contrast, IOL may decrease the risk of shoulder dystocia without increasing the risk of other adverse maternal outcomes, in particular CD, when performed before 40 + 0 weeks (GRADE: low/very low). © 2024 International Society of Ultrasound in Obstetrics and Gynecology.

Tightening Blocks in Complementary Analyses of Observational Studies: Optimization Algorithm and Examples

An observational block design has I blocks matched for covariates and J individuals per block, but treatments were not randomly assigned to individuals within blocks, as would have been done in an experiment. Tightening an observational block design means selecting J ′ < J individuals from each block, and possibly I ′ ≤ I blocks, to construct a new observational block design that, in some way, addresses unmeasured biases from nonrandom treatment assignment. Tightening must preserve covariate balance while altering the design to achieve some additional objective. An optimization algorithm is introduced that achieves this while maintaining the block structure by finely balancing covariates across blocks and through optimal subset matching. An example is considered in detail, both to motivate and illustrate the tightening of an observational block design. Two tightened designs are built from a study of light daily alcohol consumption and its possible effects on HDL cholesterol. One tightened design adjusts for an outcome tentatively presuming it was unaffected by the treatment. The second tightened design uses a differential effect to remove bias from an unobserved general disposition that promotes several treatments. An R package tightenBlock implements the method, contains the data, and in that package the help-file for the function tighten reproduces the example.

Impact of Integrative Health and Medicine on Costs Associated with Adult Health System Beneficiaries with Musculoskeletal Conditions: A Retrospective Cohort Study.

Objective: Owing to perceived additional costs, patients may avoid integrative health and medicine (IHM) treatments, while insurers may not cover IHM. We hypothesized that adult beneficiaries of a health system's employee insurance plan with musculoskeletal (MSK) conditions receiving covered outpatient IHM would have reduced total allowed costs over the 1-year follow-up compared with matched controls, secondarily exploring medical and pharmaceutical cost subsets. Methods: We queried medical records and claims spanning 2018-2023 for beneficiaries aged 18-89 years with a new MSK episode. Patients were divided into cohorts: (1) IHM within 3 months after MSK diagnosis and (2) no IHM after initial primary care. After inflation adjustment and trimming, propensity score matching was used to balance cohorts on demographics, comorbidity, health care utilization, and prior 12-month spend. Least-squares mean total, medical, and pharmaceutical allowed costs (United States Dollar) over the 1-year follow-up were analyzed using a linear mixed model. Findings were compared with a generalized linear model without trimming. Results: There were 251 patients per matched cohort, with adequate covariate balance. There was no meaningful between-cohort difference (IHM minus No IHM) in least-squares mean total cost (+703 [95% CI: -314, 1720]). Secondary outcomes included medical cost (+878 [95% CI: 61, 1695]) and pharmaceutical cost (+6 [95% CI: -71, 83]). A generalized linear model revealed no meaningful difference in estimated mean total medical costs (-2561 [95% CI: -7346, +2224]). Conclusions: IHM use among adult health system beneficiaries with MSK conditions was not associated with meaningful differences in 1-year follow-up total health care costs compared with matched controls. Our study was underpowered for secondary outcomes, which should be interpreted with caution. Future research should include a larger sample of patients and examine longitudinal changes in patient-reported outcomes.

Glycaemic control is a modifiable risk factor for hepatocellular carcinoma and liver-related mortality in patients with diabetes.

Optimal glycaemic control has well-established health benefits in patients with diabetes mellitus (DM). It is uncertain whether optimal glycaemic control can benefit liver-related outcomes. To examine the association of optimal glycaemic control with hepatocellular carcinoma (HCC) and liver-related mortality. In a population-based cohort, we identified patients with newly diagnosed DM between 2001 and 2016 in Hong Kong. Optimal glycaemic control was defined as mean haemoglobin A1c (HbA1c) <7% during the 3-year lead-in period after DM diagnosis. By applying propensity score matching to balance covariates, we analysed glycaemic control via competing risk models with outcomes of interest being HCC and liver-related mortality. We identified 146,430 patients (52.2% males, mean age 61.4 ± 11.8 years). During a median follow-up duration of 7.0 years, 1099 (0.8%) and 978 (0.7%) patients developed HCC and liver-related deaths. Optimal glycaemic control, when compared to suboptimal glycaemic control, was associated with reduced risk of HCC (subdistribution hazard ratio [SHR] 0.70, 95% CI 0.61-0.79). The risk of HCC increased with incremental HbA1c increases beyond >7% (SHR 1.29-1.71). Significant associations with HCC were also found irrespective of age (SHR 0.54-0.80), sex (SHR 0.68-0.69), BMI <25 or ≥25 kg/m2 (SHR 0.63-0.75), smoking (SHR 0.61-0.72), hepatic steatosis (SHR 0.67-0.68) and aspirin/statin/metformin use (SHR 0.67-0.75). A lower risk of liver-related mortality in relation to optimal glycaemic control was also observed (SHR 0.70, 95% CI 0.61-0.80). Glycaemic control is an independent risk factor for HCC and liver-related mortality, and should be incorporated into oncoprotective strategies in the general DM population.