Propensity score and proximity matching using random forest

SURgical vs. PERcutaneous ACCESS in Transfemoral Transcatheter Aortic Valve Implantation (SU-PER-ACCESS Study)

Background: The transfemoral (TF) approach is the most common route in TAVI, but it is still associated with a risk of bleeding and vascular complications. The aim of this study was to compare the clinical outcomes between surgical cut-down (SC) and percutaneous (PC) approach. (2) Methods: Between January 2018 and June 2022, 774 patients underwent a transfemoral TAVI procedure. After propensity matching, 323 patients underwent TAVI in each group. (3) Results: In the matched population, 15 patients (4.6%) in the SC group vs. 34 patients in the PC group (11%) experienced minor vascular complications (p = 0.02), while no difference for major vascular complication (1.5% vs. 1.9%) were reported. The rate of minor bleeding events was higher in the percutaneous group (11% vs. 3.1%, p <.001). The SC group experienced a higher rate of non-vascular-related access complications (minor 8% vs. 1.2%; major 2.2% vs. 1.2%; p < 0.001). (4) Conclusions: SC for TF-TAVI did not alter the mortality rate at 30 days and was associated with reduced minor vascular complication and bleeding. PC showed a lower rate of non-vascular-related access complications and a lower length of stay. The specific approach should be tailored to the patient's clinical characteristics.

Comparison of two propensity score-based methods for balancing covariates: the overlap weighting and fine stratification methods in real-world claims data

BACKGROUND

Two propensity score (PS) based balancing covariate methods, the overlap weighting method (OW) and the fine stratification method (FS), produce superb covariate balance. OW has been compared with various weighting methods while FS has been compared with the traditional stratification method and various matching methods. However, no study has yet compared OW and FS. In addition, OW has not yet been evaluated in large claims data with low prevalence exposure and with low frequency outcomes, a context in which optimal use of balancing methods is critical. In the study, we aimed to compare OW and FS using real-world data and simulations with low prevalence exposure and with low frequency outcomes.

METHODS

We used the Texas State Medicaid claims data on adult beneficiaries with diabetes in 2012 as an empirical example (N = 42,628). Based on its real-world research question, we estimated an average treatment effect of health center vs. non-health center attendance in the total population. We also performed simulations to evaluate their relative performance. To preserve associations between covariates, we used the plasmode approach to simulate outcomes and/or exposures with N = 4,000. We simulated both homogeneous and heterogeneous treatment effects with various outcome risks (1-30% or observed: 27.75%) and/or exposure prevalence (2.5-30% or observed:10.55%). We used a weighted generalized linear model to estimate the exposure effect and the cluster-robust standard error (SE) method to estimate its SE.

RESULTS

In the empirical example, we found that OW had smaller standardized mean differences in all covariates (range: OW: 0.0-0.02 vs. FS: 0.22-3.26) and Mahalanobis balance distance (MB) (< 0.001 vs. > 0.049) than FS. In simulations, OW also achieved smaller MB (homogeneity: <0.04 vs. > 0.04; heterogeneity: 0.0-0.11 vs. 0.07-0.29), relative bias (homogeneity: 4.04-56.20 vs. 20-61.63; heterogeneity: 7.85-57.6 vs. 15.0-60.4), square root of mean squared error (homogeneity: 0.332-1.308 vs. 0.385-1.365; heterogeneity: 0.263-0.526 vs 0.313-0.620), and coverage probability (homogeneity: 0.0-80.4% vs. 0.0-69.8%; heterogeneity: 0.0-97.6% vs. 0.0-92.8%), than FS, in most cases.

CONCLUSIONS

These findings suggest that OW can yield nearly perfect covariate balance and therefore enhance the accuracy of average treatment effect estimation in the total population.

Can propensity score matching replace randomized controlled trials?

Randomized controlled trials (RCTs) have long been recognized as the gold standard for establishing causal relationships in clinical research. Despite that, various limitations of RCTs prevent its widespread implementation, ranging from the ethicality of withholding potentially-lifesaving treatment from a group to relatively poor external validity due to stringent inclusion criteria, amongst others. However, with the introduction of propensity score matching (PSM) as a retrospective statistical tool, new frontiers in establishing causation in clinical research were opened up. PSM predicts treatment effects using observational data from existing sources such as registries or electronic health records, to create a matched sample of participants who received or did not receive the intervention based on their propensity scores, which takes into account characteristics such as age, gender and comorbidities. Given its retrospective nature and its use of observational data from existing sources, PSM circumvents the aforementioned ethical issues faced by RCTs. Majority of RCTs exclude elderly, pregnant women and young children; thus, evidence of therapy efficacy is rarely proven by robust clinical research for this population. On the other hand, by matching study patient characteristics to that of the population of interest, including the elderly, pregnant women and young children, PSM allows for generalization of results to the wider population and hence greatly increases the external validity. Instead of replacing RCTs with PSM, the synergistic integration of PSM into RCTs stands to provide better research outcomes with both methods complementing each other. For example, in an RCT investigating the impact of mannitol on outcomes among participants of the Intensive Blood Pressure Reduction in Acute Cerebral Hemorrhage Trial, the baseline characteristics of comorbidities and current medications between treatment and control arms were significantly different despite the randomization protocol. Therefore, PSM was incorporated in its analysis to create samples from the treatment and control arms that were matched in terms of these baseline characteristics, thus providing a fairer comparison for the impact of mannitol. This literature review reports the applications, advantages, and considerations of using PSM with RCTs, illustrating its utility in refining randomization, improving external validity, and accounting for non-compliance to protocol. Future research should consider integrating the use of PSM in RCTs to better generalize outcomes to target populations for clinical practice and thereby benefit a wider range of patients, while maintaining the robustness of randomization offered by RCTs.

Machine Learning and Health Science Research: Tutorial

Machine learning (ML) has seen impressive growth in health science research due to its capacity for handling complex data to perform a range of tasks, including unsupervised learning, supervised learning, and reinforcement learning. To aid health science researchers in understanding the strengths and limitations of ML and to facilitate its integration into their studies, we present here a guideline for integrating ML into an analysis through a structured framework, covering steps from framing a research question to study design and analysis techniques for specialized data types.

RKHS-based covariate balancing for survival causal effect estimation

Survival causal effect estimation based on right-censored data is of key interest in both survival analysis and causal inference. Propensity score weighting is one of the most popular methods in the literature. However, since it involves the inverse of propensity score estimates, its practical performance may be very unstable, especially when the covariate overlap is limited between treatment and control groups. To address this problem, a covariate balancing method is developed in this paper to estimate the counterfactual survival function. The proposed method is nonparametric and balances covariates in a reproducing kernel Hilbert space (RKHS) via weights that are counterparts of inverse propensity scores. The uniform rate of convergence for the proposed estimator is shown to be the same as that for the classical Kaplan-Meier estimator. The appealing practical performance of the proposed method is demonstrated by a simulation study as well as two real data applications to study the causal effect of smoking on survival time of stroke patients and that of endotoxin on survival time for female patients with lung cancer respectively.

Bilateral Femoral Cannulation Is Associated With Reduced Severe Limb Ischemia-Related Complications Compared With Unilateral Femoral Cannulation in Adult Peripheral Venoarterial Extracorporeal Membrane Oxygenation: Results From the Extracorporeal Life Support Registry

OBJECTIVES

Peripheral venoarterial extracorporeal membrane oxygenation (ECMO) with femoral access is obtained through unilateral or bilateral groin cannulation. Whether one cannulation strategy is associated with a lower risk for limb ischemia remains unknown. We aim to assess if one strategy is preferable.

DESIGN

A retrospective cohort study based on the Extracorporeal Life Support Organization registry.

SETTING

ECMO centers worldwide included in the Extracorporeal Life Support Organization registry.

PATIENTS

All adult patients (≥ 18 yr) who received peripheral venoarterial ECMO with femoral access and were included from 2014 to 2020.

INTERVENTIONS

Unilateral or bilateral femoral cannulation.

MEASUREMENTS AND MAIN RESULTS

The primary outcome was the occurrence of limb ischemia defined as a composite endpoint including the need for a distal perfusion cannula (DPC) after 6 hours from implantation, compartment syndrome/fasciotomy, amputation, revascularization, and thrombectomy. Secondary endpoints included bleeding at the peripheral cannulation site, need for vessel repair, vessel repair after decannulation, and in-hospital death. Propensity score matching was performed to account for confounders. Overall, 19,093 patients underwent peripheral venoarterial ECMO through unilateral ( n = 11,965) or bilateral ( n = 7,128) femoral cannulation. Limb ischemia requiring any intervention was not different between both groups (bilateral vs unilateral: odds ratio [OR], 0.92; 95% CI, 0.82-1.02). However, there was a lower rate of compartment syndrome/fasciotomy in the bilateral group (bilateral vs unilateral: OR, 0.80; 95% CI, 0.66-0.97). Bilateral cannulation was also associated with lower odds of cannulation site bleeding (bilateral vs unilateral: OR, 0.87; 95% CI, 0.76-0.99), vessel repair (bilateral vs unilateral: OR, 0.55; 95% CI, 0.38-0.79), and in-hospital mortality (bilateral vs unilateral: OR, 0.85; 95% CI, 0.81-0.91) compared with unilateral cannulation. These findings were unchanged after propensity matching.

CONCLUSIONS

This study showed no risk reduction for overall limb ischemia-related events requiring DPC after 6 hours when comparing bilateral to unilateral femoral cannulation in peripheral venoarterial ECMO. However, bilateral cannulation was associated with a reduced risk for compartment syndrome/fasciotomy, lower rates of bleeding and vessel repair during ECMO, and lower in-hospital mortality.

Immune tolerance via FCR001 cell therapy compared with maintenance immunosuppression for kidney transplantation: Real-world evidence analysis of safety and efficacy

While kidney transplantation (KTx) has traditionally required lifelong immunosuppression, an investigational stem cell therapy, FCR001, has been demonstrated to induce tolerance and eliminate the need for immunosuppression through the establishment of persistent mixed chimerism in a phase 2 clinical study. Real-world evidence (RWE) methods were employed to compare the safety and efficacy of non-myeloablative conditioning with FCR001 with standard of care [SOC] immunosuppression in a retrospective single-center analysis of outcomes among propensity score matched living-donor KTx receiving SOC (n = 144) or FCR001 (n = 36). Among the FCR001 recipients, 26 (72%) developed persistent chimerism allowing durable elimination of all immunosuppression. There was no significant difference in the composite primary endpoint (biopsy-proven acute rejection [BPAR], graft loss, or death) at 60 months (FCR001 27.8%, n = 10 and SOC 28.5%, n = 41; p = .9). FCR001 recipients demonstrated superior kidney function at 5 years (estimated glomerular filtration rate [eGFR] [mean ± standard deviation]: 64.1 ± 15.3) compared to SOC (51.7 ± 18.8; p = .02). At 5 years, FCR001 recipients experienced fewer complications including new-onset diabetes post-transplant, although two patients developed graft versus host disease. In conclusion, RWE demonstrated that KTx combined with non-myeloablative conditioning and FCR001 resulting in superior kidney function without increasing the risk of rejection, graft loss, or death among patients off immunosuppression.

Effect of Telemedicine Use on Medical Spending and Health Care Utilization: A Machine Learning Approach

Introduction

This study analyzes the effect of telemedicine use on healthcare utilization and medical spending for patients with chronic mental illness.

Methods

Using the IBM MarketScan Research database from 2009 to 2018, this study examined the timing of users' first telemedicine use and identified similar periods for non-users by using random forest and random forest proximity matching. A difference-in-differences approach, which tests whether there are differences in the study outcomes before and after the actual/predicted first use among the treated group (users) compared with the control group (non-users), was then used to assess the impact of telemedicine. Analyses were done in 2021.

Results

Comparing users with non-users after matching suggested that telemedicine use both increases the number of overall outpatient visits (0.461; 95% CI=0.280, 0.642; p<0.001) related to psychotherapy and evaluation and management services, and decreases the number of in-person visits (0.280; 95% CI= -0.446, -0.114; p=0.001) for patients with chronic mental health diagnoses. Total medical spending was not significantly affected. Additionally, no evidence was found of telemedicine use being associated with an increased probability of an emergency department visit or hospitalization.

Conclusions

The study findings suggest that telemedicine use is associated with an increase in outpatient care utilization for patients with chronic mental health diagnoses. No substantive changes in medical spending, the probability of an emergency department visit, or the probability of hospitalization were noted. Results provide insights into the effect of telemedicine use on spending and healthcare utilization for patients with chronic mental illness. These findings may inform research to guide future telemedicine policies and interventions.

High quality implementation of 4Rs + MTP increases classroom emotional support and reduces absenteeism

School-based social and emotional learning (SEL) programs are associated with improvements in children's SEL and academic outcomes, and the quality of classroom interactions. The magnitude of these effects increases at high levels of program implementation quality. This study aimed to (1) identify teachers' profiles of quality of implementation, (2) explore teachers and classroom characteristics contributing to their propensity to comply with high quality of implementation, and (3) examine the relations between school assignment to an SEL program, quality of classroom interactions, and child SEL and academic outcomes at different levels of teachers' compliance propensity. This study drew upon data from a cluster-randomized controlled trial evaluating the efficacy of 4Rs + MTP, a literacy-based SEL program, on third and fourth grade teachers (n = 330) and their students (n = 5,081) across 60 New York City public elementary schools. Latent profile analysis indicated that measures of teacher responsiveness and amount of exposure to implementation supports contributed to the differentiation of profiles of high and low quality of implementation. Random forest analysis showed that more experienced teachers with low levels of professional burnout had high propensity to comply with high quality of implementation. Multilevel moderated mediation analysis indicated that 4Rs + MTP teachers with high compliance propensity were associated with higher classroom emotional support and lower children's school absences than their counterparts in the control group. These findings may inform debates in policy research about the importance of providing the supports teachers need to implement SEL school programs with high quality.

Impact assessment of emergency care units on hospitalizations for respiratory system diseases in Brazil

Emergency Care Units (UPAs) are part of a national health policy implemented by the Brazilian Government. UPAs are fixed prehospital components of the Brazilian Unified Health System (SUS), whose purpose is to provide resolutive emergency care to patients suffering from acute clinical conditions, and to perform the first care in cases of surgical nature. According to the Ministry of Economy, 750 units are operational throughout the country since 2008, and 332 are under construction. Being a public policy in expansion, it is imperative to assess the impact of such units as part of SUS. However, we found few studies that assessed UPAs' impact, which have examined their specific impact on mortality rates. In our research, we aimed to evaluate the impact of UPAs on hospitalization rates for diseases of the respiratory system. To measure the impact, we used a strategy of Machine Learning through the Bayesian Additive Regression Trees (BART) algorithm. The results point to a decrease in the hospitalization rates by respiratory diseases due to Emergency Care Units. Therefore, these units generate a benefit for the Brazilian health system, being an important element for the care of patients with respiratory diseases.

Altered T-cell subset distribution in the viral reservoir in HIV-1-infected individuals with extremely low proviral DNA (LoViReTs)

BACKGROUND

HIV cure strategies aim to eliminate viral reservoirs that persist despite successful antiretroviral therapy (ART). We have previously described that 9% of HIV-infected individuals who receive ART harbor low levels of provirus (LoViReTs).

METHODS

We selected 22 LoViReTs matched with 22 controls ART suppressed for more than 3 years with fewer than 100 and more than 100 HIV-DNA copies/10⁶  CD4⁺ T cells, respectively. We measured HIV reservoirs in blood and host genetic factors. Fourteen LoViReTs underwent leukapheresis to analyze replication-competent virus, and HIV-DNA in CD4⁺ T-cell subpopulations. Additionally, we measured HIV-DNA in rectum and/or lymph node biopsies from nine of them.

RESULTS

We found that LoViReTs harbored not only lower levels of total HIV-DNA, but also significantly lower intact HIV-DNA, cell-associated HIV-RNA, and ultrasensitive viral load than controls. The proportion of intact versus total proviruses was similar in both groups. We found no differences in the percentage of host factors. In peripheral blood, 71% of LoViReTs had undetectable replication-competent virus. Minimum levels of total HIV-DNA were found in rectal and lymph node biopsies compared with HIV-infected individuals receiving ART. The main contributors to the reservoir were short-lived transitional memory and effector memory T cells (47% and 29%, respectively), indicating an altered distribution of the HIV reservoir in the peripheral T-cell subpopulations of LoViReTs.

CONCLUSION

In conclusion, LoViReTs are characterized by low levels of viral reservoir in peripheral blood and secondary lymphoid tissues, which might be explained by an altered distribution of the proviral HIV-DNA towards more short-lived memory T cells. LoViReTs can be considered exceptional candidates for future interventions aimed at curing HIV.

Impact of a Digitally Enhanced Diabetes Self-Management Program on Glycemia and Medical Costs

OBJECTIVE

To analyze economic savings and health impacts associated with a virtual digitally enhanced diabetes self-management education and support (DSMES) program.

RESEARCH DESIGN AND METHODS

Participants (n = 1,494) were nonpregnant adults with diagnosed type 2 diabetes and baseline body mass index (BMI) of 25 kg/m² (23 kg/m² if of Asian descent) or higher who enrolled in virtual DSMES between February 2019 and April 2020 for at least 4 months. Participants' changes in glycated hemoglobin (A1C) and body weight were calculated as the difference between program start and last recorded values between months 4 and 6. Outcomes for all participants were analyzed; subanalyses were done on 628 participants with starting A1C >7% (53 mmol/mol), who could benefit most from DSMES. Markov-based microsimulation approach was used to model the potential reductions in diabetes sequalae and medical expenditures if observed improvements in A1C and BMI were maintained.

RESULTS

DSMES participants with starting A1C >7% experienced average reductions of 0.9% A1C and 2.1 kg of body weight (-1.7% of BMI) within 6 months. If these improvements were maintained, simulated outcomes include reduced 5-year onset of ischemic heart disease by 9.2%, myocardial infarction by 10.6%, stroke by 12.1%, chronic kidney disease by 16.5%, and reduced onset of other sequelae. Simulated cumulative reduction in medical expenditures is $1160 after 1 year, $4150 after 3 years, $7790 after 5 years, and $18 020 after 10 years.

CONCLUSIONS

Participation in virtual DSMES improves A1C and body weight, with the potential to slow onset of diabetes sequelae and reduce medical expenditures.

Beating vs Arrested Heart Isolated Tricuspid Valve Surgery: Long-term Outcomes

BACKGROUND

Isolated tricuspid valve (TV) surgery is a rare procedure generally considered at high risk of perioperative mortality and poor long-term outcomes. Surgical treatment can be performed with either an arrested heart (AH) or beating heart (BH) technique. Aim of this study is to compare the outcomes of isolated tricuspid surgery with two different approaches.

METHODS

The SUR-TRI study is a multicenter international retrospective study enrolling adult patients who underwent isolated tricuspid valve procedures (n=406, 56±16 years; 56% female) at 13 international sites. AH and BH strategies were performed in 253 and 153 cases, respectively. Propensity score-matched analysis was used to compare groups.

RESULTS

After matching, 129 pairs were obtained and analyzed. The 30-day mortality rate was 6.2 vs 5.0% in the AH and BH groups, respectively (p=0.9). The rates of acute renal failure requiring replacement therapy (3 vs 10%, p=0.02) and stroke (0 vs 1.8%, p=0.08) were numerically higher in the AH group. The 6-year survival rate was 67±6 vs 78±5% in the AH and BH groups, respectively (p=0.18), while freedom from cardiac death was 75±5 vs 84±4% (p=0.21). The six-year composite cardiac endpoint of cardiac death and reoperation rate was 60±9 vs 86±5% (p=0.024) comparing AH-TV replacement and BH-TV repair groups.

CONCLUSIONS

Isolated tricuspid valve surgery performed with a beating heart strategy is a safe option and resulted in a trend of increased long-term survival and freedom from reoperation when compared with the standard arrested heart technique. Patients undergoing beating heart valve repair had the best long-term outcome.

Protein-protein interaction site prediction using random forest proximity distance

A front-end method based on random forest proximity distance (PD) is used to screen the test set to improve protein-protein interaction site (PPIS) prediction. The assessment of a distance metric is done under the assumption that a distance definition of higher quality leads to higher classification. On an independent test set, the numerical analysis based on statistical inference shows that the PD has the advantage over Mahalanobis and Cosine distance. Based on the fact that the proximity distance depends on the tree composition of the random forest model, an iterative method is designed to optimize the proximity distance, which adjusts the tree composition of the random forest model by adjusting the size of the training set. Two PD metrics, 75PD and 50PD, are obtained by the iterative method. On two independent test sets, compared with the PD produced by the original training set, the values of 75PD in Matthews correlation coefficient and F₁ score were higher, and the differences between them were statistically significant. All numerical experiments show that the closer the distance between the test data and the training data, the better the prediction results of the predictor. These indicate that the iterative method can optimize proximity distance definition and the distance information provided by PD can be used to indicate the reliability of prediction results.

Association between cardioplegia and postoperative atrial fibrillation in coronary surgery

OBJECTIVE

The aim of this multicenter study was to evaluated whether cold or warm cardioplegia are associated with postoperative atrial fibrillation (POAF) and the prognostic role of the latter on early stroke and neurological mortality.

METHOD

This was a retrospective analysis of prospective collected data from 9 cardiac centers in Italy and the United States including patients undergoing surgery between 2010 and 2018. From the 9 institutional databases, 17,231 patients underwent isolated CABG on-pump, using either warm cardioplegia (n = 7730) or cold cardioplegia (n = 9501); among the latter group blood and crystalloid cardioplegia were used in 691 and 8810 patients, respectively. After matching, two pairs of 4162 patients (overall cohort 8324) were analyzed.

RESULTS

In matched population, the rate of POAF was 18% (1472 cases), 15% (608) in warm group versus 21% (864) in cold group (p < 0.001). Multivariable analysis confirmed that cold cardioplegia was associated with higher rate of POAF, along with age, hypercholesterolemia, LVEF, reoperation, preoperative IABP, previous stroke, cardiopulmonary and cross-clamp. Moreover, cold cardioplegia as well as POAF increased the rate of postoperative stroke as well as early mortality and neurological mortality Propensity-weighted cohort included 11,830 (70%) patients out of 17,231. After adjustment, both cold blood and cold crystalloid cardioplegia negatively influenced POAF, stroke and neurological mortality.

CONCLUSIONS

Warm cardioplegia may reduce the rate of POAF in CABG patients with respect to cold cardioplegia, either blood or crystalloid. This has a prognostic impact on postoperative stroke and neurological mortality.

Weighted nearest neighbours-based control group selection method for observational studies

Although in observational studies, propensity score matching is the most widely used balancing method, it has received much criticism. The main drawback of this method is that the individuals of the case and control groups are paired in the compressed one-dimensional space of propensity scores. In this paper, such a novel multivariate weighted k-nearest neighbours-based control group selection method is proposed which can eliminate this disadvantage of propensity score matching. The proposed method pairs the elements of the case and control groups in the original vector space of the covariates and the dissimilarities of the individuals are calculated as the weighted distances of the subjects. The weight factors are calculated from a logistic regression model fitted on the status of treatment assignment. The efficiency of the proposed method was evaluated by Monte Carlo simulations on different datasets. Experimental results show that the proposed Weighted Nearest Neighbours Control Group Selection with Error Minimization method is able to select a more balanced control group than the most widely applied greedy form of the propensity score matching method, especially for individuals characterized with few descriptive features.

Should the left gastric artery lymph node be considered as the predictive lymph node for extra-gastric lymph node metastases?

Background

To validate the prognostic impacts of the left gastric artery lymph node (No. 7 LN) metastasis and investigate whether the No. 7 LN metastasis should be considered as the predictive LN for extra-gastric LN metastases.

Methods

Between January 2003 and December 2011, a total of 1,586 patients who underwent R0 gastrectomy were retrospected. Patients with LN metastases were divided into three groups: (I) patients with only peri-gastric LN metastases (peri-gastric group); (II) patients with peri-gastric and only No. 7 LN metastases (No. 7 group); and (III) patients with other extra-gastric LN metastases (extra-gastric group). Propensity score matching (PSM) was adopted to accurately evaluate prognoses of all patients after surgery.

Results

Of 1,586 patients, 235 (14.82%) were pathologically identified to present with the No. 7 LN metastases. Patients with the No. 7 LN metastases presented the significantly lower survival rate both before and after adjustment by pTNM stage, compared to those without the No. 7 LN metastases. Patients in the No. 7 group were identified to present the significant lower survival rate than those in the peri-gastric group, and to present the similar median overall survival (OS) to those in the extra-gastric group. In addition, patients with extra-gastric LN except No. 7 LN metastases failed to show any superiority of survival outcomes, compared with those with extra-gastric LN metastases including the No. 7 LN metastasis.

Conclusions

The No. 7 LN metastases had the crucial survival implications. Nevertheless, the No. 7 LN failed to be considered as the predictive LN for the extra-gastric LN metastases in gastric cancer (GC).

Extremely low viral reservoir in treated chronically HIV-1-infected individuals

BACKGROUND

Small viral reservoirs are found predominantly in HIV-1 controllers and individuals treated during acute/early HIV-1 infection. However, other HIV⁺ individuals could naturally also harbour low viral reservoirs.

METHODS

We screened 451 HIV-1-infected treated-individuals with suppressed plasma viremia for at least 3 years and stored cryopreserved peripheral blood mononuclear cells (PBMCs). Total HIV-DNA was analysed in PBMCs with ddPCR. Individuals with <50 HIV-DNA copies/10⁶ PBMCs constitute the 'Low Viral Reservoir Treated' cohort (LoViReT). Longitudinal samples were obtained from 12 chronically treated LoViReT and compared to 13 controls (>50 HIV-DNA copies/10⁶ PBMCs) to analyse total HIV-DNA, T-cell and NK-cell populations, HIV-1 specific antibodies, and plasma inflammation markers.

FINDINGS

We found that 9.3% of the individuals screened had <50 HIV-DNA copies/10⁶ PBMCs. At least 66% initiated cART during the chronic phase of HIV-1 infection (cp-LoViReT). Cp-LoViReT harboured lower levels of HIV-DNA before cART and after treatment introduction the decays were greater compared to controls. They displayed a marked decline in quantity and avidity in HIV-specific antibodies after initiation of cART. Cp-LoViReT had fewer CD8⁺ T_TM and T_EMRA in the absence of cART, and higher CD8⁺ T_N after 18 months on therapy.

INTERPRETATION

Treated chronically HIV-1-infected LoViReT represent a new phenotype of individuals characterized by an intrinsically reduced viral reservoir, less impaired CD8⁺ T-cell compartment before cART, and low circulating HIV-1 antigens despite being treated in the chronic phase of infection. The identification of this unique group of individuals is of great interest for the design of future eradication studies.

FUNDING

MSD Spain.

Propensity score adjustment using machine learning classification algorithms to control selection bias in online surveys

Modern survey methods may be subject to non-observable bias, from various sources. Among online surveys, for example, selection bias is prevalent, due to the sampling mechanism commonly used, whereby participants self-select from a subgroup whose characteristics differ from those of the target population. Several techniques have been proposed to tackle this issue. One such is Propensity Score Adjustment (PSA), which is widely used and has been analysed in various studies. The usual method of estimating the propensity score is logistic regression, which requires a reference probability sample in addition to the online nonprobability sample. The predicted propensities can be used for reweighting using various estimators. However, in the online survey context, there are alternatives that might outperform logistic regression regarding propensity estimation. The aim of the present study is to determine the efficiency of some of these alternatives, involving Machine Learning (ML) classification algorithms. PSA is applied in two simulation scenarios, representing situations commonly found in online surveys, using logistic regression and ML models for propensity estimation. The results obtained show that ML algorithms remove selection bias more effectively than logistic regression when used for PSA, but that their efficacy depends largely on the selection mechanism employed and the dimensionality of the data.

Prognostic impact of D2-plus lymphadenectomy and optimal extent of lymphadenectomy in advanced gastric antral carcinoma: Propensity score matching analysis

Objective

To investigate the prognostic impact of D2-plus lymphadenectomy including the posterior (No. 8p, No. 12b/p, No. 13, and No. 14v), and para-aortic (No. 16a2, and No. 16b1) lymph nodes (LNs) in subtotal gastrectomy for advanced gastric antral carcinoma.

Methods

A total of 203 patients with advanced gastric cancer (GC) located in the antrum, who underwent R0 gastrectomy with D2 or D2-plus lymphadenectomy between January 2003 and December 2011 were enrolled. Propensity score matching was used to reduce the strength of the confounding factors to accurately evaluate prognoses. The therapeutic value index (TVI) was calculate to evaluate the survival benefit of dissecting each LN station.

Results

Of 102 patients with D2-plus lymphadenectomy, 21 (20.59%) were pathologically identified as having LN metastases beyond the extent of D2 lymphadenectomy. After matching, the overall survival (OS) was significantly better in the D2-plus than the D2 group (P=0.030). In the multivariate survival analysis, D2-plus lymphadenectomy (hazard ratio, 0.516; P=0.006) was confirmed to significantly improve the survival rate. In the logistic regression analysis, pN stage [odds ratio (OR), 2.533; 95% confidence interval (95% CI), 1.368−4.691; P=0.003] and extent of LNs metastasis (OR, 5.965; 95% CI, 1.335−26.650; P=0.019) were identified as independent risk factors for LN metastases beyond the extent of D2 lymphadenectomy. The TVI of patient with metastasis to LNs station was 7.1 (No. 8p), 5.7 (No. 12p), 5.1 (No. 13), and 7.1 (both No. 16a2 and No. 16b1), respectively.

Conclusions

D2-plus lymphadenectomy may improve the prognoses of some patients with advanced GC located in the antrum, especially for No. 8p, No. 12b, No. 13, and No. 16.

Novel ranking of protective and risk factors for adolescent adiposity in US females

OBJECTIVE

Rank the importance of potentially modifiable psychosocial, dietary and environmental risk and protective factors for female adolescent obesity in order to target and inform public health prevention efforts. Utilizing the largest dataset available that captures the onset of the adolescent obesity surge in the USA, the study provides a more robust understanding of paediatric obesity risk factors.

METHODS

Data were obtained from an observational, longitudinal study conducted between 1989 and 2001, the NHLBI Growth and Health Study. This study includes girls aged 9-19 years from three urban US locations, with Black and White girls generally represented equally. Data were analysed using multiple regression, random forest and propensity score matching to determine the strongest adiposity risk and protective factors during ages 9-12 predicting adiposity at age 19 with multiple methods to maximize the ability to identify possible public health interventions. Multiple linear regression and random forest analysis identified the strongest associations among 288 risk and protective factors selected from the study's literature review. For the 190 factors associated with follow-up adiposity from the data, propensity score matching was used to control for confounding of these factors.

RESULTS

Findings suggest that highest priority interventional targets across the domains surveyed are lowering specific nutrients; eating meals with others or during activities without skipping; parents fixing evening snacks; improving perceptions of non-extremes as the healthy weight; improving self-worth, physical activity and social competence; and limiting any negative impact of dieting relatives. Similar associations were observed for Black and White girls.

CONCLUSION

The clinical implications of these findings allow health practitioners to target behavioural change efforts and address social and environmental factors that have demonstrated higher prioritization value for early obesity interventional efforts for adolescents.

Visible and near infrared spectroscopy coupled to random forest to quantify some soil quality parameters

This study evaluates the use of visible and near infrared spectroscopy (Vis-NIRS) combined with multivariate regression based on random forest to quantify some quality soil parameters. The parameters analyzed were soil cation exchange capacity (CEC), sum of exchange bases (SB), organic matter (OM), clay and sand present in the soils of several regions of Brazil. Current methods for evaluating these parameters are laborious, timely and require various wet analytical methods that are not adequate for use in precision agriculture, where faster and automatic responses are required. The random forest regression models were statistically better than PLS regression models for CEC, OM, clay and sand, demonstrating resistance to overfitting, attenuating the effect of outlier samples and indicating the most important variables for the model. The methodology demonstrates the potential of the Vis-NIR as an alternative for determination of CEC, SB, OM, sand and clay, making possible to develop a fast and automatic analytical procedure.

Application of a Combined Model with Autoregressive Integrated Moving Average (ARIMA) and Generalized Regression Neural Network (GRNN) in Forecasting Hepatitis Incidence in Heng County, China

BACKGROUND

Hepatitis is a serious public health problem with increasing cases and property damage in Heng County. It is necessary to develop a model to predict the hepatitis epidemic that could be useful for preventing this disease.

METHODS

The autoregressive integrated moving average (ARIMA) model and the generalized regression neural network (GRNN) model were used to fit the incidence data from the Heng County CDC (Center for Disease Control and Prevention) from January 2005 to December 2012. Then, the ARIMA-GRNN hybrid model was developed. The incidence data from January 2013 to December 2013 were used to validate the models. Several parameters, including mean absolute error (MAE), root mean square error (RMSE), mean absolute percentage error (MAPE) and mean square error (MSE), were used to compare the performance among the three models.

RESULTS

The morbidity of hepatitis from Jan 2005 to Dec 2012 has seasonal variation and slightly rising trend. The ARIMA(0,1,2)(1,1,1)12 model was the most appropriate one with the residual test showing a white noise sequence. The smoothing factor of the basic GRNN model and the combined model was 1.8 and 0.07, respectively. The four parameters of the hybrid model were lower than those of the two single models in the validation. The parameters values of the GRNN model were the lowest in the fitting of the three models.

CONCLUSIONS

The hybrid ARIMA-GRNN model showed better hepatitis incidence forecasting in Heng County than the single ARIMA model and the basic GRNN model. It is a potential decision-supportive tool for controlling hepatitis in Heng County.