Analysis Of Treatment Effects Research Articles

Impact of Heterogenous Capabilities on Export Performance amid the Digital Transformation

We investigate the effects of adopting digital technology on export performance of Korean manufacturing firms amidst the digital transformation. We recognize firm-level capabilities should be closely associated with adoption of rapidly progressing digital technology but also with export performance. And our data indicates that the most common purpose of digital technology adoption is to launch new products to the market. Hence, we consider that technology adoption is a strategic and purposeful decision to gain competitive edge mostly by producing new product, and treating technology adoption exogenous can be misleading. Due to the endogeneity and selection issues in technology adoption, the endogenous switching regression is applied to this study as Coad et al. (2020). On technology adoption decision, we find external innovative resources from strategic alliance in addition to internal innovative capabilities stand out. Given technology adoption decision, internal capabilities like patent rights and international affiliation are complementary factors to export growth. The treatment effect analysis has implications as follows: the result on contribution of technology adoption to export growth for actually adopting firms is rather small, and shows heterogenous innovation, organizational, and external capabilities are still critical factors as much as new products embedding high-end digital technology; the result on non-adopting firms indicates potential of advanced digital technology to improve export performance by helping to produce such new products if they were actually capable to do it. In sum, our findings provide another evidence that structurally positive interaction between innovative activities and export performance as Aw et al. (2011), since adoption of digital technology to products itself is innovative in the digital transformation. Furthermore, the result is consistent with the capability theory in that heterogenous innovative and complementary capabilities determine strategic technology adoption and export performance simultaneously. Finally, our findings indicate that the digital transformation might be still at the early stage. The fact that Korean firms have adopted advanced digital technology mainly for new products can be interpreted as an indicator of early development stage of transformation, since firms concentrate on product innovation than process innovation to gain competitive edge at the early stage (Utterback and Abernathy, 1975). Thus, we expect that further evolution of digital transformation can facilitate process innovation, then contribute to firm performance by improving cost efficiency that should be tackled in the upcoming study.

Safety and Effectiveness Analysis of Endovascular Treatment vs Standard Medication Treatment in Patients with Progressive AIS-LVO Stroke in the Ultra-late Time Window: A Propensity Score Matched Cohort Study

The time from onset to symptom deterioration in ischemic stroke often exceeds 24 hours, and this ultra-late time window is excluded from the endovascular treatment (EVT) guideline. This study aimed to explore the safety and efficacy of EVT in progressive acute ischemic stroke with large vessel occlusion (AIS-LVO) stroke patients with onset to symptom deterioration times of 24h-7 days. Progressive stroke patients with time window of 24h-7 days treated at our hospital over the past 6 years were retrospectively collected. Patients were categorized into EVT and standard medication treatment (SMT) group based on the treatment approach. Patients were matched using propensity score matching (PSM). Safety outcomes primarily included 3-month mortality and symptomatic intracranial hemorrhage (sICH), efficacy outcome primarily included functional independence (3-month mRS≤2). A total of 396 patients were included in the study, with 86(21.7%) in EVT and 310(78.3%) in SMT group. There were 140 remaining after PSM, with 70 in each group (50%). Compared to SMT group, EVT group had higher functional independence (52.9% vs 15.7%, OR=7.504, 95% CI 2.141-14.093, P<0.001) and lower 3-month mortality (14.3% vs 40.0%, OR=0.412, 95% CI 0.099-0.856, P<0.001). EVT was also associated with higher sICH (25.7% vs 5.7%, OR=9.926, 95% CI 1.874-36.547, P<0.001). For patients with progressive AIS-LVO in the ultra-late time window, EVT remains a viable treatment approach.

Design of explosively welded Fe–Al multilayer laminated composite pipes: A critical microscopy analysis of stand-off distance and post-weld heat treatment effects on interface properties

This study examined the microstructure and mechanical properties of explosively welded (EXW) Fe–Al multilayer laminated composite pipes, specifically SS321/AA1050/AA5083, with varying stand-off distances (SD) and post-weld heat treatments (PWHT). Findings revealed that higher collision forces at higher-SD produced a wavy interface with finer grains at the AA1050/AA5083 interface. In contrast, lower-SD yielded a smooth, flat interface. At the SS321/AA1050 interface, collision energy resulted in the formation of an uneven composite reaction layer comprising multiple intermetallics. Analysis of the SS321/AA1050 interface revealed that an increase in the PWHT temperature results in diffusion toward the reaction layer, transforming the unevenly Al-rich and Fe-rich phases into a uniform Al85(Fe,Cr,Ni)15 solid solution. The increase in the fraction of this phase made the reaction layer more brittle. At the AA1050/AA5083 interface, PWHT at higher temperatures caused a significant hardness decline extending further from the interface in samples with higher-SD. On the SS321 side of the laminated composite, higher SD resulted in higher kernel average misorientation (KAM) and increased hardness. Despite a decrease in hardness after PWHT at 350 °C, the hardness gradient from the matrix to the interface in laminates with higher SD remained relatively unaffected, whereas the lower SD counterparts experienced a noticeable stress relief.

Primary care payment models and avoidable hospitalizations in Ontario, Canada: A multivalued treatment effects analysis.

Improving access to primary care physicians' services may help reduce hospitalizations due to Ambulatory Care Sensitive Conditions (ACSCs). Ontario, Canada's most populous province, introduced blended payment models for primary care physicians in the early- to mid-2000s to increase access to primary care, preventive care, and better chronic disease management. We study the impact of payment models on avoidable hospitalizations due to two incentivized ACSCs (diabetes and congestive heart failure) and two non-incentivized ACSCs (angina and asthma). The data for our study came from health administrative data on practicing primary care physicians in Ontario between 2006 and 2015. We employ a two-stage estimation strategy on a balanced panel of 3710 primary care physicians (1158 blended-fee-for-service (FFS), 1388 blended-capitation models, and 1164 interprofessional team-based practices). First, we account for the differences in physician practices using a generalized propensity score based on a multinomial logit regression model, corresponding to three primary care payment models. Second, we use fractional regression models to estimate the average treatment effects on the treated outcome (i.e., avoidable hospitalizations). The capitation-based model sometimes increases avoidable hospitalizations due to angina (by 7 per 100,000 patients) and congestive heart failure (40 per 100,000) relative to the blended-FFS-based model. Switching capitation physicians into interprofessional teams mitigates this effect, reducing avoidable hospitalizations from congestive heart failure by 30 per 100,000 patients and suggesting better access to primary care and chronic disease management in team-based practices.

Effectiveness, Utility, and Economic Analysis of Three-drug Treatment vs Two-drug Treatment in Chronic Obstructive Pulmonary Disease Patients

Effectiveness, Utility, and Economic Analysis of Three-drug Treatment vs Two-drug Treatment in Chronic Obstructive Pulmonary Disease Patients

Treatment effect analysis of the Frailty Care Bundle (FCB) in a cohort of patients in acute care settings

PurposeThe aim of this study is to explore the feasibility of using machine learning approaches to objectively differentiate the mobilization patterns, measured via accelerometer sensors, of patients pre- and post-intervention.MethodsThe intervention tested the implementation of a Frailty Care Bundle to improve mobilization, nutrition and cognition in older orthopedic patients. The study recruited 120 participants, a sub-group analysis was undertaken on 113 patients with accelerometer data (57 pre-intervention and 56 post-intervention), the median age was 78 years and the majority were female. Physical activity data from an ankle-worn accelerometer (StepWatch 4) was collected for each patient during their hospital stay. These data contained daily aggregated gait variables. Data preprocessing included the standardization of step counts and feature computation. Subsequently, a binary classification model was trained. A systematic hyperparameter optimization approach was applied, and feature selection was performed. Two classifier models, logistic regression and Random Forest, were investigated and Shapley values were used to explain model predictions.ResultsThe Random Forest classifier demonstrated an average balanced accuracy of 82.3% (± 1.7%) during training and 74.7% (± 8.2%) for the test set. In comparison, the logistic regression classifier achieved a training accuracy of 79.7% (± 1.9%) and a test accuracy of 77.6% (± 5.5%). The logistic regression model demonstrated less overfitting compared to the Random Forest model and better performance on the hold-out test set. Stride length was consistently chosen as a key feature in all iterations for both models, along with features related to stride velocity, gait speed, and Lyapunov exponent, indicating their significance in the classification.ConclusionThe best performing classifier was able to distinguish between patients pre- and post-intervention with greater than 75% accuracy. The intervention showed a correlation with higher gait speed and reduced stride length. However, the question of whether these alterations are part of an adaptive process that leads to improved outcomes over time remains.

Radiomic Analysis of Treatment Effect for Patients with Radiation Necrosis Treated with Pentoxifylline and Vitamin E.

The combination of oral pentoxifylline (Ptx) and vitamin E (VitE) has been used to treat radiation-induced fibrosis and soft tissue injury. Here, we review outcomes and perform a radiomic analysis of treatment effects in patients prescribed Ptx + VitE at our institution for the treatment of radiation necrosis (RN). A total of 48 patients treated with stereotactic radiosurgery (SRS) had evidence of RN and had MRI before and after starting Ptx + VitE. The radiation oncologist's impression of the imaging in the electronic medical record was used to score response to treatment. Support Vector Machine (SVM) was used to train a model of radiomics features derived from radiation necrosis on pre- and 1st post-treatment T1 post-contrast MRIs that can classify the ultimate response to treatment with Ptx + VitE. A total of 43.8% of patients showed evidence of improvement, 18.8% showed no change, and 25% showed worsening RN upon imaging after starting Ptx + VitE. The median time-to-response assessment was 3.17 months. Nine patients progressed significantly and required Bevacizumab, hyperbaric oxygen therapy, or surgery. Patients who had multiple lesions treated with SRS were less likely to show improvement (p = 0.037). A total of 34 patients were also prescribed dexamethasone, either before (7), with (16), or after starting (11) treatment. The use of dexamethasone was not associated with an improved response to Ptx + VitE (p = 0.471). Three patients stopped treatment due to side effects. Finally, we were able to develop a machine learning (SVM) model of radiomic features derived from pre- and 1st post-treatment MRIs that was able to predict the ultimate treatment response to Ptx + VitE with receiver operating characteristic (ROC) area under curve (AUC) of 0.69. Ptx + VitE appears safe for the treatment of RN, but randomized data are needed to assess efficacy and validate radiomic models, which may assist with prognostication.

A Longitudinal Treatment Effect Analysis of Antipsychotics on Behavior of Residents in Long-Term Care

ObjectiveThe proportion of long-term care (LTC) residents being treated with antipsychotic medication is high, and these medications may exacerbate behavioral symptoms. We used propensity scores to investigate the effect of antipsychotic use on the worsening of behavioral symptoms among residents in LTC facilities. DesignA retrospective study. Setting and participantsResidents in LTC in 8 provinces and 1 territory in Canada, without severe aggressive behavior at baseline and reassessed at follow-up, between March 2000 and March 2022. MethodsWe used propensity score matching and weighting to balance baseline covariates and logistic regression to estimate the effect of antipsychotics on the worsening of behavioral symptoms in the original, matched, and weighted cohorts. The treatment variable was use of antipsychotic medication at baseline and the outcome was worsening of behavior at follow-up. ResultsA total of 494,215 participants were included (318,234 women and 175,981 men; mean age 82.8 years [SD 10.1; range 18–112]).130 558 (26.4%) used antipsychotics at baseline and 88,632 (17.9%) had worsening behavior in follow-up. In the matched cohort, there were 249,698 participants, and 124,849 were matched (1:1) in each treatment group. There was a significant association between antipsychotic use at baseline and worsening in behavior at follow-up in the adjusted regression models (OR 1.27 [95% CI 1.25–1.29], <0.0001) as well as in matched (OR 1.20 [95% CI 1.17–1.21], <0.0001) and weighted (OR 1.26 [95% CI 1.24–1.28], <0.0001) cohorts. Conclusions and implicationsThis study further evidence to support the cautious use of antipsychotics in LTC facilities. Future research in LTC facilities could include a more granular analyses of behavior change, including bidirectional analyses between different symptom severity classifications.

Comparison of perioperative bleeding risk between direct oral anticoagulants in transurethral resection of prostate.

To evaluate the perioperative morbidity and mortality associated with direct oral anticoagulants (DOACs) and warfarin for patients receiving transurethral resection of prostate (TURP). This was a single-centre, retrospective cohort analysis of patients who underwent TURP for benign prostate hyperplasia from April 2019 to December 2023. The primary objective was to evaluate the perioperative bleeding and thromboembolic risk between anticoagulated (AC) vs no-AC patients. The secondary objective was to evaluate perioperative bleeding and thromboembolic risk between different formulations of DOACs. Patient demographics, prior treatment, prostate size, baseline bleeding risk, and operative details were collected. Bleeding and thromboembolic-related morbidity were captured within a 3-month postoperative period. Perioperative management of AC therapy was recorded, and all patients had their AC therapy withheld. Cohort characteristic between the AC vs no-AC, and DOAC groups were analysed with two-sided t-test, and chi-square test. Further logistic regression analyses were carried out to identified significant variables between the groups. These significant variables were used for adjustment in inverse probability-weighted treatment effect analysis to evaluate bleeding risk. There were 629 patients in the cohort, and 113 (18%) patients were receiving AC therapy. The AC patients were at 1.6 times statistically significant increased risk of acute bleeding, and 11 times increased risk of prolonged haematuria for >14 days. When compared to apixaban, patients on rivaroxaban conferred a statistically significant increased risk of acute bleeding by 2.21 times. Patients receiving AC therapy had a statistically significant increased risk of stroke in the perioperative setting (no-AC vs AC: 0.4% vs 2.7%, P = 0.01). This is the first study to evaluate risk of bleeding for TURP patients receiving DOACs. The AC patients are more likely to experience haematuria and stroke in the perioperative period despite withholding therapy. Apixaban appears to cause less bleeding-related complications than rivaroxaban.

Exploring the mechanism of path-creating strategy for latecomers: a combined approach of econometrics and causal machine learning

In this paper, the traditional econometrics method and causal machine-learning method are combined to study the mechanism of a path-creating strategy of latecomers to influence latecomers’ catch-up performance. A total of 283 high-tech manufacturing enterprises listed on the Shanghai and Shenzhen stock exchanges from 2007 to 2019 were selected for the study. OLS linear regression model verifies that path-creating has a positive impact on latecomers’ technological catch-up performance, technological capability plays an intermediary role between path-creating and technology catch-up performance, technological innovation appropriability positively moderates the effect of path-creating on technological capability, and technological innovation cumulativeness negatively moderates the effect of path-creating on technological catch-up but positively moderate the effect of technological capability on catch-up performance. Through machine learning, on the one hand, a conclusion basically consistent with the linear regression model is obtained, but on the other hand, a more heterogeneous situation is presented. Through analyses of the individual treatment effect of a path-creating strategy of latecomers, the Shapley value graph shows the complex influence of different features on the treatment effect of the enterprise using the path-creating strategy. Through the decision tree, some more complex patterns are found. In addition, the decision tree model based on causal analysis can also assist enterprises in making strategic decisions.

Effects of electroacupuncture on cognitive symptoms in major depressive disorder: a pilot study and randomized controlled trial

Objectives To investigate the impact of electroacupuncture on cognitive function, quality of life (QoL), and depression severity in patients with major depressive disorder (MDD). Methods This double-blinded randomized controlled trial included 60 participants aged 18-55 with cognitive symptoms related to MDD at Thammasat University Hospital. Participants were divided into two groups: the electroacupuncture group combined with standard antidepressant treatment (EG; n=30) and the control group receiving standard care with placebo acupuncture (CG; n=30). The study assessed 1) executive functions using the Trail making test- B and Stroop Color and Word Test, 2) delayed recall, and 3) subjective cognitive complaint and Quality of life (QoL) using WHODAS 2.0. Depressive symptoms were measured using the Thai version of the Patient Health Questionnaire (PHQ-9). Baseline and post-intervention assessments were conducted over 10 weeks. Mann-Whitney U test analyzed treatment effects by comparing median differences between groups. Results Both groups exhibited similar demographics and cognitive traits. Cognitive improvement was observed in both groups at the endpoint. Intention-to-treat analysis revealed significantly higher median scores for subjective cognitive complaints in the EG compared to the CG (EG: Median = 5.5, CG: Median = 0.0, p=0.049). No serious side effects were identified from either electroacupuncture or placebo acupuncture. Conclusions Electroacupuncture improved subjective complaints in MDD patients with cognitive symptoms, but did not show effects on specific cognitive functions, QoL, or depressive symptoms. This study provides initial evidence supporting the potential of electroacupuncture in MDD patients with cognitive symptoms, suggesting opportunities for further research. Trial registration NCT06239740, February 2, 2024, ClinicalTrials.gov.

Head-to-head studies on psoriasis and psoriatic arthritis

Given the ever-increasing number of approved therapies for the treatment of psoriasis (PsO) and psoriatic arthritis (PsA), head-to-head (H2H) comparative studies are essential. These are aimed primarily at acomparative analysis of treatment effectiveness. In both PsO and PsA, biological disease-modifying antirheumatic drugs (bDMARD) have been shown to be superior to conventional therapies in H2H studies. In PsO interleukin 17 (IL-17) and IL-23 inhibitors proved superiority compared to tumor necrosis factor (TNF) inhibitors (etanercept and adalimumab) in several studies. Ustekinumab was more effective than etanercept, but less effective than IL-17 and IL-23 inhibitors. Only afew H2H studies have been published on the treatment of PsA. In the Spirit H2H study ixekizumab was superior to adalimumab using acombined endpoint of arthritis and psoriasis response (ACR-50 and PASI-100). When looking at arthritic symptoms only (ACR-20), secukinumab was not significantly superior to adalimumab in the EXCEED study but was superior in terms of the effect on skin involvement (PASI90). Other H2H studies focused on the treatment of enthesitis (ECLIPSA study), the efficacy of Janus kinase (JAK) inhibition (SELECT-PSA-1) or the additional administration of methotrexate to bDMARD treatment (MUST study). The H2H data have been incorporated into the treatment guidelines and have led to IL-17 and IL-23 inhibition being preferred over TNF inhibition in cases of relevant skin involvement in PsA.

The Impact of Double Pelvic Osteotomy on Dorsolateral Subluxation in 24 Dogs.

The aim of this study was to determine the effect of double pelvic osteotomy (DPO) on dorsolateral subluxation (DLS) of the femoral head in a weight-bearing position. Retrospective observational study with treatment effect analysis. Twenty-four client-owned dogs that underwent DPO due to hip dysplasia. Computed tomography (CT) studies were performed on 24 dogs that underwent a DPO between 2015 and 2021. The patients were positioned with the femora in a weight-bearing position. The DLS score was measured pre-DPO, immediately post-DPO, and at follow-up to determine if DPO resulted in an increase in the DLS score. A total of 21 dogs underwent single-stage bilateral DPO and 3 dogs underwent unilateral DPO resulting in 45 hip joints available for analysis. Follow-up CT was available for 29 of these at a median of 51 days (range: 35-482 days). The mean DLS score was 36.1% ± 14.6 preoperatively, 71.4% ± 12.5 post-DPO, and 70.0% ± 15.3 at follow-up. The increase between the pre- and postoperative measurement was significant (p < 0.001). DPO results in reduced subluxation of the femoral head in a weight-bearing position. The in vivo results suggest that DPO will result in reduced femoral subluxation during weight-bearing.

Female leaders as environmental saviors: Tackling urban pollution in Chinese cities

In the Chinese political context, local government officials have substantial influence over the implementation of policies. The approach to urban environmental management varies significantly between female and male officials, which in turn affects governance outcomes differently. Consequently, the gender of local officials plays a crucial role in shaping urban environmental conditions. This study investigates the impact of female officials on urban environmental pollution in China over the period from 2003 to 2018. Utilizing data from the China Urban Statistical Yearbook combined with individual profiles of officials at the prefecture level, the analysis reveals a notable decrease in urban environmental pollution associated with female leadership. Notably, the effect is more pronounced among younger female officials compared to their older counterparts, and those with higher education levels are particularly effective in curbing pollution emissions. The pollution abatement effect of female officials is found to be stronger in the central regions compared to the eastern and western regions. The robustness of these findings is confirmed through checks against alternative explanatory variables, permutation tests, and treatment effect analyses. Overall, this research contributes new perspectives on management strategies that can sisgnificantly improve urban environmental quality in China.

Shared Decision-Making in Colorectal Cancer Screening for Older Adults

Decisions about whether to stop colorectal cancer (CRC) screening tests in older adults can be difficult and may benefit from shared decision-making (SDM). To evaluate the effect of physician training in SDM and electronic previsit reminders (intervention) vs reminders only (comparator) on receipt of the patient-preferred approach to CRC screening and on overall CRC screening rates of older adults at 12 months. This was a secondary analysis of the Promoting Informed Decisions About Colorectal Cancer Screening in Older Adults (PRIMED) cluster randomized clinical trial. In the PRIMED trial, primary care physicians (PCPs) from 36 primary care practices in Massachusetts and Maine were enrolled between May 1 and August 30, 2019, and were randomized to the intervention group or the comparator group. Patients aged 76 to 85 years who were overdue for CRC screening and did not have a prior diagnosis of CRC enrolled between October 21, 2019, and April 8, 2021. Data analysis was performed between May 24, 2022, and May 10, 2023. Primary care physicians in the intervention group completed an SDM training course and received previsit reminders of patients eligible for CRC testing discussion, whereas PCPs in the comparator group received reminders only. The primary outcome was concordance, or the percentage of patients who received their preferred screening approach. Postvisit surveys were administered to assess patient preference for testing, and electronic health record review was used to assess CRC testing at 12 months. Heterogeneity of treatment effect analyses examined interaction between study groups and different factors on concordance rates. This study included 59 physicians and 466 older adults. Physicians had a mean (SD) age of 52.7 (9.4) years and a mean (SD) of 21.6 (10.2) years in practice; 30 (50.8%) were women and 16 (27.1%) reported prior training in SDM. Patients had a mean (SD) age of 80.3 (2.8) years; 249 (53.4%) were women and 238 (51.1%) reported excellent or very good overall health. Patients preferred stool-based tests (161 [34.5%]), followed by colonoscopy (116 [24.8%]) or no further screening (97 [20.8%]); 75 (16.1%) were not sure. The distribution of patient preferences was similar across groups (P = .36). At 12 months, test uptake was also similar for both the intervention group (29 [12.3%] for colonoscopy, 62 [26.3%] for stool-based tests, and 145 [61.4%] for no testing) and the comparator group (32 [13.9%] for colonoscopy, 35 [15.2%] for stool-based tests, and 163 [70.9%] for no testing; P = .08). Approximately half of patients in the intervention group received their preferred approach vs the comparator group (115 of 226 [50.9%] vs 103 of 223 [46.2%]; P = .47). Heterogeneity of treatment effect analyses found significantly higher rates with the intervention vs the comparator for patients with a strong intention to follow through with the preferred approach (adjusted odds ratio [AOR], 1.79 [95% CI, 1.11-2.89]; P = .02, P = .05 for interaction) and for patients who reported more than 5 minutes (AOR, 3.27 [95% CI, 1.25-8.59]; P = .02, P = .05 for interaction) of discussion with their PCP regarding screening. Higher rates were also observed among patients who reported 2 to 5 minutes of discussion with their PCP, although this finding was not significant (AOR, 1.89 [95% CI, 0.93-3.84]; P = .08, P = .05 for interaction). In this secondary analysis of a cluster randomized clinical trial, approximately half of older patients received their preferred approach to CRC screening. Physician training in SDM did not result in higher concordance rates overall but may have benefitted some subgroups. Future work to refine and evaluate clinical decision support (in the form of an electronic advisory or reminder) as well as focused SDM skills training for PCPs may promote high-quality, preference-concordant decisions about CRC testing for older adults. ClinicalTrials.gov Identifier: NCT03959696.

Do intensive specialized services prevent family separation in parents with intellectual disabilities? A treatment effects analysis

ABSTRACT The current study explores the effectiveness of Project IMPACT, an intensive, in-home, skills-based positive parenting program designed to prevent out-of-home placement for parents with intellectual disabilities. Families participating in Project IMPACT were compared with families in a similar jurisdiction who participated in non-specialized family preservation services using treatment effects analysis. Findings indicate that Project IMPACT is highly effective at keeping families intact. Project IMPACT reduces family separation throughout childhood, the most benefits are seen in early childhood, when the risk of family separation is generally the greatest. Parents with intellectual disability can learn skills to keep children safe at home.

Endovascular thrombectomy in patients with anterior circulation stroke: an emulated real-world comparison

BackgroundEndovascular thrombectomy (EVT) has been proven effective in anterior circulation stroke due to large vessel occlusion (LVO). However, translation from randomized clinical trials (RCTs) with highly selected patients to real-world requires confirmation, particularly to identify associations outside of strict selection criteria.AimsThis study aims to compare functional outcomes after EVT in real-world with those reported in RCTs, and to identify associations with functional outcome after EVT outside RCT-criteria.MethodsThis study analyzed longitudinal German real-world data from the Stroke Research Consortium in Northern Bavaria (STAMINA) cohort from January, 2015 to June, 2019. We conducted a trial emulation, comparing patients with anterior circulation stroke and LVO meeting selection criteria for RCTs investigating EVT (1) predominantly within 6 hours with those from HERMES meta-analysis, and (2) within 6-24 hours with those from AURORA meta-analysis. We (3) analyzed treatment effects of EVT and association with functional outcome in patients treated outside RCT criteria.ResultsOf 598 patients, 281 (47.0%) met RCT-criteria for treatment within 6 hours (hereinafter STAMINA-HERMES), 74 (12.4%) met RCT-criteria for treatment within 6–24 hours (STAMINA-AURORA), and 277 (46.3%) patients received EVT outside RCT-criteria. We observed no difference in rates of functional independence or mortality, comparing STAMINA-HERMES with HERMES meta-analysis (mRS 0-1: n=120/281 [43%] vs. 291/633 [46%], p=0.36; mortality: n=34/281 [12%] vs. 97/633 [15%], p=0.20), and STAMINA-AURORA with AURORA meta-analysis (mRS 0-1: n=26/74 [35%] vs. 122/266 [46%], p=0.10, mortality: n=10/74 [14%] vs. 45/266 [17%], p=0.48). Patients treated outside RCT-criteria had worse outcome (mRS 0-1: n=38/277 [14%], mortality: n=90/277 [32%], both p<0.001); possibly driven by pre-existing functional dependence (n=172/277 [62%]). Compared to matched controls, EVT outside of RCT-criteria was associated with lower mortality (absolute treatment effect: -14%, 95% Confidence Interval [CI] -23 to -5, p<0.01), but not with recovery to functional independence or premorbid functional status (treatment effect: 4%, CI -4 to 11, p=0.34), which was associated with lower NIHSS (Odds ratio [OR] 0.86, CI 0.80-0.92, p<0.001) and age (OR 0.95, CI 0.93-0.98, p=0.002).ConclusionsTranslation of EVT outcomes reported in RCTs into real-world is possible, however, almost half of patients did not meet trial criteria. Identification of patients who functionally benefit from frequently performed EVT outside RCT-criteria requires further investigation.Trial registrationClinicaltrials.gov, NCT04357899.

Outcome risk model development for heterogeneity of treatment effect analyses: a comparison of non-parametric machine learning methods and semi-parametric statistical methods

BackgroundIn randomized clinical trials, treatment effects may vary, and this possibility is referred to as heterogeneity of treatment effect (HTE). One way to quantify HTE is to partition participants into subgroups based on individual’s risk of experiencing an outcome, then measuring treatment effect by subgroup. Given the limited availability of externally validated outcome risk prediction models, internal models (created using the same dataset in which heterogeneity of treatment analyses also will be performed) are commonly developed for subgroup identification. We aim to compare different methods for generating internally developed outcome risk prediction models for subject partitioning in HTE analysis.MethodsThree approaches were selected for generating subgroups for the 2,441 participants from the United States enrolled in the ASPirin in Reducing Events in the Elderly (ASPREE) randomized controlled trial. An extant proportional hazards-based outcomes predictive risk model developed on the overall ASPREE cohort of 19,114 participants was identified and was used to partition United States’ participants by risk of experiencing a composite outcome of death, dementia, or persistent physical disability. Next, two supervised non-parametric machine learning outcome classifiers, decision trees and random forests, were used to develop multivariable risk prediction models and partition participants into subgroups with varied risks of experiencing the composite outcome. Then, we assessed how the partitioning from the proportional hazard model compared to those generated by the machine learning models in an HTE analysis of the 5-year absolute risk reduction (ARR) and hazard ratio for aspirin vs. placebo in each subgroup. Cochran’s Q test was used to detect if ARR varied significantly by subgroup.ResultsThe proportional hazard model was used to generate 5 subgroups using the quintiles of the estimated risk scores; the decision tree model was used to generate 6 subgroups (6 automatically determined tree leaves); and the random forest model was used to generate 5 subgroups using the quintiles of the prediction probability as risk scores. Using the semi-parametric proportional hazards model, the ARR at 5 years was 15.1% (95% CI 4.0–26.3%) for participants with the highest 20% of predicted risk. Using the random forest model, the ARR at 5 years was 13.7% (95% CI 3.1–24.4%) for participants with the highest 20% of predicted risk. The highest outcome risk group in the decision tree model also exhibited a risk reduction, but the confidence interval was wider (5-year ARR = 17.0%, 95% CI= -5.4–39.4%). Cochran’s Q test indicated ARR varied significantly only by subgroups created using the proportional hazards model. The hazard ratio for aspirin vs. placebo therapy did not significantly vary by subgroup in any of the models. The highest risk groups for the proportional hazards model and random forest model contained 230 participants each, while the highest risk group in the decision tree model contained 41 participants.ConclusionsThe choice of technique for internally developed models for outcome risk subgroups influences HTE analyses. The rationale for the use of a particular subgroup determination model in HTE analyses needs to be explicitly defined based on desired levels of explainability (with features importance), uncertainty of prediction, chances of overfitting, and assumptions regarding the underlying data structure. Replication of these analyses using data from other mid-size clinical trials may help to establish guidance for selecting an outcomes risk prediction modelling technique for HTE analyses.

Analyzing Treatment Effect by Integrating Existing Propensity Score and Outcome Regressions with Heterogeneous Covariate Sets

Analyzing treatment or exposure effect is a major research theme in scientific studies. In the current big-data era where multiple sources of data are available, it is of interest to perform a synthesized analysis of treatment effects by integrating information from different data sources or studies. However, studies may contain heterogeneous and incomplete covariate sets, and individual data therein may not be accessible. We apply and extend the generalized meta-analysis method to integrate summary results (e.g., regression coefficients) of outcome and treatment (propensity score, PS) regression analyses across different datasets that may contain heterogeneous covariate sets. The proposed integrated analysis utilizes a reference dataset, which contains data on the complete set of covariates. The asymptotic distribution for the proposed integrated estimator is established. Simulations reveal that the proposed estimator performs well. We apply the proposed method to obtain the causal effect of waist circumference on hypertension by integrating two existing outcomes and PS regression analyses with different sets of covariates.

Impact of linkage level on inferences from big data analyses in health and medical research: an empirical study

BackgroundLinkage errors that occur according to linkage levels can adversely affect the accuracy and reliability of analysis results. This study aimed to identify the differences in results according to personally identifiable information linkage level, sample size, and analysis methods through empirical analysis.MethodsThe difference between the results of linkage in directly identifiable information (DII) and indirectly identifiable information (III) linkage levels was set as III linkage based on name, date of birth, and sex and DII linkage based on resident registration number. The datasets linked at each level were named as databaseIII (DBIII) and databaseDII (DBDII), respectively. Considering the analysis results of the DII-linked dataset as the gold standard, descriptive statistics, group comparison, incidence estimation, treatment effect, and moderation effect analysis results were assessed.ResultsThe linkage rates for DBDII and DBIII were 71.1% and 99.7%, respectively. Regarding descriptive statistics and group comparison analysis, the difference in effect in most cases was “none” to “very little.” With respect to cervical cancer that had a relatively small sample size, analysis of DBIII resulted in an underestimation of the incidence in the control group and an overestimation of the incidence in the treatment group (hazard ratio [HR] = 2.62 [95% confidence interval (CI): 1.63–4.23] in DBIII vs. 1.80 [95% CI: 1.18–2.73] in DBDII). Regarding prostate cancer, there was a conflicting tendency with the treatment effect being over or underestimated according to the surveillance, epidemiology, and end results summary staging (HR = 2.27 [95% CI: 1.91–2.70] in DBIII vs. 1.92 [95% CI: 1.70–2.17] in DBDII for the localized stage; HR = 1.80 [95% CI: 1.37–2.36] in DBIII vs. 2.05 [95% CI: 1.67–2.52] in DBDII for the regional stage).ConclusionsTo prevent distortion of the analyses results in health and medical research, it is important to check that the patient population and sample size by each factor of interest (FOI) are sufficient when different data are linked using DBDII. In cases involving a rare disease or with a small sample size for FOI, there is a high likelihood that a DII linkage is unavoidable.