G-computation of average treatment effects on the treated and the untreated

Harmful compared to what? The problem of gaming and ambiguous causal questions

BACKGROUND AND AIMS

There has been much concern regarding potential harmful effects of video game-play in the past 40 years, but limited progress in understanding its causal role. This paper discusses the basic requirements for identifying causal effects of video game-play and argues that most research to date has focused upon ambiguous causal questions.

METHODS

Video games and mental health are discussed from the perspective of causal inference with compound exposures; that is, exposures with multiple relevant variants that affect outcomes in different ways.

RESULTS

Not only does exposure to video games encompass multiple different factors, but also not playing video games is equally ambiguous. Estimating causal effects of a compound exposure introduces the additional challenge of exposure-version confounding.

CONCLUSIONS

Without a comparison of well-defined interventions, research investigating the effects of video game-play will be difficult to translate into actionable health interventions. Interventions that target games should be compared with other interventions aimed at improving the same outcomes.

Adoption of an Enhanced Recovery After Surgery Protocol Increases Cost of Cytoreductive Surgery and Hyperthermic Intraperitoneal Chemotherapy and Does not Improve Outcomes

BACKGROUND

Enhanced recovery after surgery (ERAS) protocols have been shown to reduce length of stay (LOS) and complications. The impact of ERAS protocols on the cost of cytoreductive surgery and hyperthermic intraperitoneal chemotherapy (CRS-HIPEC) has not been studied.

PATIENTS AND METHODS

We performed a retrospective cohort analysis of patients undergoing CRS-HIPEC from 2016-2022 at a single quaternary center. Propensity score matching was used to create pre-and post-ERAS cohorts. Cost, overall and serious complications, and intensive care unit (ICU) length of stay (LOS) between the two cohorts were compared using the Mann-Whitney U-test for continuous variables and χ2 test for categorical variables.

RESULTS

Our final matched cohort consisted of 100 patients, with 50 patients in both the pre- and post-ERAS groups. After adjusting for patient complexity and inflation, the median total cost [$75,932 ($67,166-102,645) versus $92,992 ($80,720-116,710), p = 0.02] and operating room cost [$26,817 ($23,378-33,121) versus $34,434 ($28,085-$41,379), p < 0.001] were significantly higher in the post-ERAS cohort. Overall morbidity (n = 22, 44% versus n = 17, 34%, p = 0.40) and ICU length of stay [2 days (IQR 1-3) versus 2 days (IQR 1-4), p = 0.70] were similar between the two cohorts. A total cost increase of $22,393 [SE $13,047, 95% CI (-$3178 to $47,965), p = 0.086] was estimated after implementation of ERAS, with operating room cost significantly contributing to this increase [$8419, SE $1628, 95% CI ($5228-11,609), p < 0.001].

CONCLUSIONS

CRS-HIPEC ERAS protocols were associated with higher total costs due to increased operating room costs at a single institution. There was no significant difference in ICU LOS and complications after the implementation of the ERAS protocol.

Simulating the impact of greenspace exposure on metabolic biomarkers in a diverse population living in San Diego, California: A g-computation application

Introduction

Growing evidence exists that greenspace exposure can reduce metabolic syndrome risk, a growing public health concern with well-documented inequities across population subgroups. We capitalize on the use of g-computation to simulate the influence of multiple possible interventions on residential greenspace on nine metabolic biomarkers and metabolic syndrome in adults (N = 555) from the 2014-2017 Community of Mine Study living in San Diego County, California.

Methods

Normalized difference vegetation index (NDVI) exposure from 2017 was averaged across a 400-m buffer around the participants' residential addresses. Participants' fasting plasma glucose, total cholesterol, high-density lipoprotein cholesterol, low-density lipoprotein cholesterol, and triglyceride concentrations, systolic and diastolic blood pressure, hemoglobin A1c (%), waist circumference, and metabolic syndrome were assessed as outcomes of interest. Using parametric g-computation, we calculated risk differences for participants being exposed to each decile of the participant NDVI distribution compared to minimum NDVI. Differential health impacts from NDVI exposure by sex, ethnicity, income, and age were examined.

Results

We found that a hypothetical increase in NDVI exposure led to a decrease in hemoglobin A1c (%), glucose, and high-density lipoprotein cholesterol concentrations, an increase in fasting total cholesterol, low-density lipoprotein cholesterol, and triglyceride concentrations, and minimal changes to systolic and diastolic blood pressure, waist circumference, and metabolic syndrome. The impact of NDVI changes was greater in women, Hispanic individuals, and those under 65 years old.

Conclusions

G-computation helps to simulate the potential health benefits of differential NDVI exposure and identifies which subpopulations can benefit most from targeted interventions aimed at minimizing health disparities.

Comparing the performance of two-stage residual inclusion methods when using physician's prescribing preference as an instrumental variable: unmeasured confounding and noncollapsibility

Aim: The first objective is to compare the performance of two-stage residual inclusion (2SRI), two-stage least square (2SLS) with the multivariable generalized linear model (GLM) in terms of the reducing unmeasured confounding bias. The second objective is to demonstrate the ability of 2SRI and 2SPS in alleviating unmeasured confounding when noncollapsibility exists. Materials & methods: This study comprises a simulation study and an empirical example from a real-world UK population health dataset (Clinical Practice Research Datalink). The instrumental variable (IV) used is based on physicians' prescribing preferences (defined by prescribing history). Results: The percent bias of 2SRI in terms of treatment effect estimates to be lower than GLM and 2SPS and was less than 15% in most scenarios. Further, 2SRI was found to be robust to mild noncollapsibility with the percent bias less than 50%. As the level of unmeasured confounding increased, the ability to alleviate the noncollapsibility decreased. Strong IVs tended to be more robust to noncollapsibility than weak IVs. Conclusion: 2SRI tends to be less biased than GLM and 2SPS in terms of estimating treatment effect. It can be robust to noncollapsibility in the case of the mild unmeasured confounding effect.

Comparing Perioperative Complications of Off-Clamp versus On-Clamp Partial Nephrectomy for Renal Cancer Using a Novel Energy Balancing Weights Method

Partial nephrectomy (PN) is the primary surgical method for renal tumor treatment, typically involving clamping the renal artery during tumor removal, leading to warm ischemia and potential renal function impairment. Off-clamp approaches have been explored to mitigate organ damage, yet few results have emerged about the possible effects on hemoglobin loss. Most evidence comes from retrospective studies using propensity score matching, known to be sensitive to PS model misspecification. The energy balancing weights (EBW) method offers an alternative method to address bias by focusing on balancing all the characteristics of covariate distribution. We aimed to compare on- vs. off-clamp techniques in PN using EB-weighted retrospective patient data. Out of 333 consecutive PNs (275/58 on/off-clamp ratio), the EBW method achieved balanced variables, notably tumor anatomy and staging. No significant differences were observed in the operative endpoints between on- and off-clamp techniques, although off-clamp PNs showed slight reductions in hemoglobin loss and renal function decline, albeit with slightly higher perioperative blood loss. Our findings support previous evidence, indicating comparable surgical outcomes between standard and off-clamp procedures, with the EBW method proving effective in balancing baseline variables in observational studies comparing interventions.

Non-monetary motivations of the EU agri-environmental policy adoption. A causal forest approach

This paper investigates the non-monetary motivations of farmers' adoption of agri-environmental policies. Unlike the monetary (income) motivations, non-monetary drivers can not be directly observed but can be identified from observational data within appropriate quasi-experimental designs. A theoretical justification of farmers' choices is first formulated and a consequent natural experiment setting is derived. The latter admits heterogeneous, i.e. Individual, Treatment Effects (ITE) that, in turn, can be interpreted in terms of more targeted and tailored policy expenditure. A Causal Forest (CF) approach is adopted to estimate these ITEs for both the treated and not treated units. The approach is applied to two balanced panel samples of Italian Farm Accountancy Data Network (FADN) farms observed over the 2008-2018 period and concerns agri-environmental policies delivered through the Common Agricultural Policy (CAP). Results show how heterogeneous the farmers' response and the associated non-monetary motivations can be, thus indicating room for a more efficient policy design.

Model-based standardization using multiple imputation

BACKGROUND

When studying the association between treatment and a clinical outcome, a parametric multivariable model of the conditional outcome expectation is often used to adjust for covariates. The treatment coefficient of the outcome model targets a conditional treatment effect. Model-based standardization is typically applied to average the model predictions over the target covariate distribution, and generate a covariate-adjusted estimate of the marginal treatment effect.

METHODS

The standard approach to model-based standardization involves maximum-likelihood estimation and use of the non-parametric bootstrap. We introduce a novel, general-purpose, model-based standardization method based on multiple imputation that is easily applicable when the outcome model is a generalized linear model. We term our proposed approach multiple imputation marginalization (MIM). MIM consists of two main stages: the generation of synthetic datasets and their analysis. MIM accommodates a Bayesian statistical framework, which naturally allows for the principled propagation of uncertainty, integrates the analysis into a probabilistic framework, and allows for the incorporation of prior evidence.

RESULTS

We conduct a simulation study to benchmark the finite-sample performance of MIM in conjunction with a parametric outcome model. The simulations provide proof-of-principle in scenarios with binary outcomes, continuous-valued covariates, a logistic outcome model and the marginal log odds ratio as the target effect measure. When parametric modeling assumptions hold, MIM yields unbiased estimation in the target covariate distribution, valid coverage rates, and similar precision and efficiency than the standard approach to model-based standardization.

CONCLUSION

We demonstrate that multiple imputation can be used to marginalize over a target covariate distribution, providing appropriate inference with a correctly specified parametric outcome model and offering statistical performance comparable to that of the standard approach to model-based standardization.

[Relationship Between Hearing Loss and Cognitive Function in Elderly Chinese People: A Study Based on Propensity Score Matching]

Objective

To explore the relationship between hearing loss and cognitive function in the elderly population through propensity score matching method.

Methods

We analyzed the data of 7605 participants aged 60 and above who were included in the 2018 China Health and Retirement Longitudinal Study (CHARLS). The non-substitutable 1∶1 nearest neighbor matching method without caliper value was used for propensity score matching and G-computation was used to estimate the average treatment effect (ATE) of hearing loss on all dimensions of cognitive function.

Results

Before matching, there were 3626 (47.68%) women, with 1409 (18.53%) of whom suffering from hearing loss and 3031 (39.86%) of whom suffering from cognitive impairment. After matching, 1409 subjects were included in the hearing loss group and 1409, in the normal hearing group, with both groups sharing similar distribution of basic demographic characteristics. The results for the average treatment effect of the population indicated that the cognitive function scores of the hearing loss group were lower than those of the normal hearing group, with the overall cognitive function being 0.593 points lower (95% confidence intervel [CI]: －0.916-－0.257, P<0.001), orientation being 0.183 points lower (95% CI: －0.302-－0.055, P=0.004), immediate memory being 0.150 points lower (95% CI: －0.218-－0.085, P<0.001), and language skills being 0.178 points lower (95% CI: －0.303-－0.058, P=0.006). The prevalence of cognitive impairment of the hearing loss group was 4.2% higher than that of the normal hearing group (95% CI: 0.007-0.077, P=0.020).

Conclusion

Hearing loss adversely affects the orientation, memory, and language skills of the elderly population and forms a potential risk factor for cognitive impairment in the elderly population.

Early special educational needs provision and its impact on unplanned hospital utilisation and school absences in children with isolated cleft lip and/or palate: a demonstration target trial emulation study protocol using ECHILD

Background

Special educational needs (SEN) provision is designed to help pupils with additional educational, behavioural or health needs; for example, pupils with cleft lip and/or palate may be offered SEN provision to improve their speech and language skills. Our aim is to contribute to the literature and assess the impact of SEN provision on health and educational outcomes for a well-defined population.

Methods

We will use the ECHILD database, which links educational and health records across England. Our target population consists of children identified within ECHILD to have a specific congenital anomaly: isolated cleft lip and/or palate. We will apply a trial emulation framework to reduce biases in design and analysis of observational data to investigate the causal impact of SEN provision (including none) by the start of compulsory education (Year One - age five year on entry) on the number of unplanned hospital utilisation and school absences by the end of primary education (Year Six - age ten/eleven). We will use propensity score-based estimators (inverse probability weighting (IPW) and IPW regression adjustment IPW) to compare categories of SEN provision in terms of these outcomes and to triangulate results obtained using complementary estimation methods (Naïve estimator, multivariable regression, parametric g-formula, and if possible, instrumental variables), targeting a variety of causal contrasts (average treatment effect/in the treated/in the not treated) of SEN provision.

Conclusions

This study will evaluate the impact of reasonable adjustments at the start of compulsory education on health and educational outcomes in the isolated cleft lip and palate population by triangulating complementary methods under a target-trial framework.

Target Trial Emulation and Bias Through Missing Eligibility Data: An Application to a Study of Palivizumab for the Prevention of Hospitalization Due to Infant Respiratory Illness

Target trial emulation (TTE) applies the principles of randomized controlled trials to the causal analysis of observational data sets. One challenge that is rarely considered in TTE is the sources of bias that may arise if the variables involved in the definition of eligibility for the trial are missing. We highlight patterns of bias that might arise when estimating the causal effect of a point exposure when restricting the target trial to individuals with complete eligibility data. Simulations consider realistic scenarios where the variables affecting eligibility modify the causal effect of the exposure and are missing at random or missing not at random. We discuss means to address these patterns of bias, namely: 1) controlling for the collider bias induced by the missing data on eligibility, and 2) imputing the missing values of the eligibility variables prior to selection into the target trial. Results are compared with the results when TTE is performed ignoring the impact of missing eligibility. A study of palivizumab, a monoclonal antibody recommended for the prevention of respiratory hospital admissions due to respiratory syncytial virus in high-risk infants, is used for illustration.

Dealing with confounding in observational studies: A scoping review of methods evaluated in simulation studies with single-point exposure

The aim of this article was to perform a scoping review of methods available for dealing with confounding when analyzing the effect of health care treatments with single-point exposure in observational data. We aim to provide an overview of methods and their performance assessed by simulation studies indexed in PubMed. We searched PubMed for simulation studies published until January 2021. Our search was restricted to studies evaluating binary treatments and binary and/or continuous outcomes. Information was extracted on the methods' assumptions, performance, and technical properties. Of 28,548 identified references, 127 studies were eligible for inclusion. Of them, 84 assessed 14 different methods (ie, groups of estimators that share assumptions and implementation) for dealing with measured confounding, and 43 assessed 10 different methods for dealing with unmeasured confounding. Results suggest that there are large differences in performance between methods and that the performance of a specific method is highly dependent on the estimator. Furthermore, the methods' assumptions regarding the specific data features also substantially influence the methods' performance. Finally, the methods result in different estimands (ie, target of inference), which can even vary within methods. In conclusion, when choosing a method to adjust for measured or unmeasured confounding it is important to choose the most appropriate estimand, while considering the population of interest, data structure, and whether the plausibility of the methods' required assumptions hold.

Measuring the impact of long-haul low-cost carriers on lowering fares: A quasi-experimental design to assess the pre-COVID market

This paper aims to investigate the impact of the introduction of the long-haul low cost carrier in the North Atlantic market to present the competitive situation before the COVID-19. There are a number of challenges in estimating the incremental effect of LH LCC. Therefore, several strategies were taken. Firstly, a difference in differences estimation and propensity score matched methods were employed using six major routes in the North Atlantic market with IATA's ticket sale data from January 2015 to December 2019; a granulated data to present the characteristics of flight and economy class fares. The outcomes indicate that a 17.2-20.6% fare reduction in average on the routes where Norwegian operated during 2015 and 2019 after Norwegian's entry, compared to what it would have happened if they didn't operate. It implies the LH LCC entry lowered fares significantly, and the level of fare competition in the North Atlantic market before the COVID-19 was high. In addition, a certain level of viability as an LH LCC has been implicated. This output can be used for the airline's strategic implication and the policy proposition, particularly when LCC expands the longer routes after the COVID recovery. Frequent and specific (detailed) assessments by market and period are imperative.

Effects of Erectile Dysfunction Drugs Use on T-cells and Immune Markers on Men Who Have Sex with Men

Objective

Examine prospective relationships between erectile dysfunction (ED) drugs EDand CD4 and CD8 T-cells, and immune markers among men who have sex with men (MSM).

Methods

Data from Multicenter AIDS Cohort Study, an observational prospective cohort study, with semi-annual follow-ups conducted in four U.S. centers from 1998 onwards was used. Marginal structural models using g-computation was fitted to estimate the mean differences for the effects of self-reported ED drug use on CD4 and CD8 T-cell outcomes and immune biomarkers.

Results

Total of 1,391 men with HIV (MWH) and 307 men without HIV (MWOH) was included. Baseline mean CD4 cell count among MWH and MWOH was 499.9 cells/μL and 966.7 cells/μL, respectively. At baseline, 41.8% of MWH were virally suppressed. ED drug users reported a mean of 44.4 months of exposure to ED drugs. ED drug use was associated with increased CD4 cell outcomes among MWH but not MWOH. Mean differences in CD4 cell counts after 1 year of ED drug use was 57.6 cells/μL and increased to 117.7 after 10 years among MWH. CD8 counts were higher in ED drug users among MWH over 10 years than non-users; no consistent differences were found among MWOH. ED drug use appeared to reduce immune marker levels, such as IL-6 and increase markers, such as IL-10. We observed similar effects of ED drug use on biomarker levels among MWOH.

Conclusion

Long-term use of ED drugs do not adversely affect immune function among MWH or MWOH. Future studies on the relationships between different types of ED drugs and effects on T-cell subtypes are warranted.

Evaluating the robustness of targeted maximum likelihood estimators via realistic simulations in nutrition intervention trials

Several recently developed methods have the potential to harness machine learning in the pursuit of target quantities inspired by causal inference, including inverse weighting, doubly robust estimating equations and substitution estimators like targeted maximum likelihood estimation. There are even more recent augmentations of these procedures that can increase robustness, by adding a layer of cross-validation (cross-validated targeted maximum likelihood estimation and double machine learning, as applied to substitution and estimating equation approaches, respectively). While these methods have been evaluated individually on simulated and experimental data sets, a comprehensive analysis of their performance across real data based simulations have yet to be conducted. In this work, we benchmark multiple widely used methods for estimation of the average treatment effect using ten different nutrition intervention studies data. A nonparametric regression method, undersmoothed highly adaptive lasso, is used to generate the simulated distribution which preserves important features from the observed data and reproduces a set of true target parameters. For each simulated data, we apply the methods above to estimate the average treatment effects as well as their standard errors and resulting confidence intervals. Based on the analytic results, a general recommendation is put forth for use of the cross-validated variants of both substitution and estimating equation estimators. We conclude that the additional layer of cross-validation helps in avoiding unintentional over-fitting of nuisance parameter functionals and leads to more robust inferences.

Estimating the impact of cefuroxime versus cefazolin and amoxicillin/clavulanate use on future collateral resistance: a retrospective comparison

Background

Quantitative estimates of collateral resistance induced by antibiotic use are scarce.

Objectives

To estimate the effects of treatment with amoxicillin/clavulanate or cefazolin, compared with cefuroxime, on future resistance to ceftazidime among hospitalized patients.

Methods

A retrospective analysis of patients with positive bacterial cultures hospitalized in an Israeli hospital during 2016–19 was conducted. Patients were restricted to those treated with amoxicillin/clavulanate, cefazolin or cefuroxime and re-hospitalized with a positive bacterial culture during the following year. Matching was performed using exact, Mahalanobis and propensity score matching. Each patient in the amoxicillin/clavulanate and cefazolin groups was matched to a single patient from the cefuroxime group, yielding 185:185 and 298:298 matched patients. Logistic regression and the g-formula (standardization) were used to estimate the OR, risk difference (RD) and number needed to harm (NNH).

Results

Cefuroxime induced significantly higher resistance to ceftazidime than amoxicillin/clavulanate or cefazolin; the marginal OR was 1.76 (95% CI = 1.16–2.83) compared with amoxicillin/clavulanate and 1.98 (95% CI = 1.41–2.8) compared with cefazolin and the RD was 0.118 (95% CI = 0.031–0.215) compared with amoxicillin/clavulanate and 0.131 (95% CI = 0.058–0.197) compared with cefazolin. We also estimated the NNH; replacing amoxicillin/clavulanate or cefazolin with cefuroxime would yield ceftazidime resistance in 1 more patient for every 8.5 (95% CI = 4.66–32.14) or 7.6 (95% CI = 5.1–17.3) patients re-hospitalized in the following year, respectively.

Conclusions

Our results indicate that treatment with amoxicillin/clavulanate or cefazolin is preferable to cefuroxime, in terms of future collateral resistance. The results presented here are a first step towards quantitative estimations of the ecological damage caused by different antibiotics.

Using Metrics of a Mixture Effect and Nutrition from an Observational Study for Consideration towards Causal Inference

Environmental exposures to a myriad of chemicals are associated with adverse health effects in humans, while good nutrition is associated with improved health. Single chemical in vivo and in vitro studies demonstrate causal links between the chemicals and outcomes, but such studies do not represent human exposure to environmental mixtures. One way of summarizing the effect of the joint action of chemical mixtures is through an empirically weighted index using weighted quantile sum (WQS) regression. My Nutrition Index (MNI) is a metric of overall dietary nutrition based on guideline values, including for pregnant women. Our objective is to demonstrate the use of an index as a metric for more causally linking human exposure to health outcomes using observational data. We use both a WQS index of 26 endocrine-disrupting chemicals (EDCs) and MNI using data from the SELMA pregnancy cohort to conduct causal inference using g-computation with counterfactuals for assumed either reduced prenatal EDC exposures or improved prenatal nutrition. Reducing the EDC exposure using the WQS index as a metric or improving dietary nutrition using MNI as a metric, the counterfactuals in a causal inference with one SD change indicate significant improvement in cognitive function. Evaluation of such a strategy may support decision makers for risk management of EDCs and individual choices for improving dietary nutrition.

Mental health in people with Parkinson's disease during the COVID-19 pandemic: potential for targeted interventions?

The COVID-19 pandemic has introduced a myriad of challenges to the social life and care of people with Parkinson’s disease (PD), which could potentially worsen mental health problems. We used baseline data of the PRIME-NL study (N = 844) to examine whether the association between COVID-19 stressors and mental health is disproportionately large in specific subgroups of people with PD and to explore effects of hypothetical reductions in COVID-19 stressors on mental health and quality of life. The mean (SD) age of the study population was 70.3 (7.8) years and 321 (38.0%) were women. The linear regression effect estimate of the association of COVID-19 stressors with mental health was most pronounced in women, highly educated people, people with advanced PD and people prone to distancing or seeking social support. Smaller effect estimates were found in people scoring high on confrontive coping or planful problem solving. The parametric G-formula method was used to calculate the effects of hypothetical interventions on COVID-19 stressors. An intervention reducing stressors with 50% in people with above median MDS-UPDRS-II decreased the Beck Depression Inventory in this group from 14.7 to 10.6, the State-Trait Anxiety Inventory from 81.6 to 73.1 and the Parkinson’s Disease Quality of Life Questionnaire from 35.0 to 24.3. Insights from this cross-sectional study help to inform tailored care interventions to subgroups of people with PD most vulnerable to the impact of COVID-19 on mental health and quality of life.

Adjustment for Baseline Covariates to Increase Efficiency in RCTs with Binary Endpoint: A Comparison of Bayesian and Frequentist Approaches

BACKGROUND

In a randomized controlled trial (RCT) with binary outcome the estimate of the marginal treatment effect can be biased by prognostic baseline covariates adjustment. Methods that target the marginal odds ratio, allowing for improved precision and power, have been developed.

METHODS

The performance of different estimators for the treatment effect in the frequentist (targeted maximum likelihood estimator, inverse-probability-of-treatment weighting, parametric G-computation, and the semiparametric locally efficient estimator) and Bayesian (model averaging), adjustment for confounding, and generalized Bayesian causal effect estimation frameworks are assessed and compared in a simulation study under different scenarios. The use of these estimators is illustrated on an RCT in type II diabetes.

RESULTS

Model mis-specification does not increase the bias. The approaches that are not doubly robust have increased standard error (SE) under the scenario of mis-specification of the treatment model. The Bayesian estimators showed a higher type II error than frequentist estimators if noisy covariates are included in the treatment model.

CONCLUSIONS

Adjusting for prognostic baseline covariates in the analysis of RCTs can have more power than intention-to-treat based tests. However, for some classes of model, when the regression model is mis-specified, inflated type I error and potential bias on treatment effect estimate may arise.

The impact of hypothetical interventions on adiposity in adolescence

In order to develop effective public health initiatives aimed at promoting healthy weight development, identifying the interventions/combination of interventions with the highest beneficial effect on body weight is vital. The study aimed to estimate the mean BMI at age 13 under hypothetical interventions targeting dietary behavior, physical activity and screen time at age 11. We used data from a school-based cohort study of 530 participants followed between the ages of 11 and 13. We used g-computation, a causal modeling method, to estimate the impact of single and combined hypothetical behavioral interventions at age 11 on BMI at age 13. Of the hypothetical interventions, the one with the largest population mean difference in BMI was the one combining all interventions (dietary behavior, physical activity and screen time interventions) and assuming 100% intervention adherence, with a population mean differences of - 0.28 (95% CI - 0.59, 0.07). Isolated behavioral interventions had a limited impact on BMI. This study demonstrated that a combination of healthy dietary behavior and physical activity promotion, as well as screen time reduction interventions at age 11 could have the highest beneficial effect on the reduction of BMI at age 13, although the change in BMI was small. The findings highlight the importance of a systems approach to obesity prevention focusing on multicomponent interventions.

Bayesian semi-parametric G-computation for causal inference in a cohort study with MNAR dropout and death

Causal inference with observational longitudinal data and time-varying exposures is often complicated by time-dependent confounding and attrition. The G-computation formula is one approach for estimating a causal effect in this setting. The parametric modeling approach typically used in practice relies on strong modeling assumptions for valid inference, and moreover depends on an assumption of missing at random, which is not appropriate when the missingness is missing not at random (MNAR) or due to death. In this work we develop a flexible Bayesian semi-parametric G-computation approach for assessing the causal effect on the subpopulation that would survive irrespective of exposure, in a setting with MNAR dropout. The approach is to specify models for the observed data using Bayesian additive regression trees, and then use assumptions with embedded sensitivity parameters to identify and estimate the causal effect. The proposed approach is motivated by a longitudinal cohort study on cognition, health, and aging, and we apply our approach to study the effect of becoming a widow on memory. We also compare our approach to several standard methods.

Assessing exposure effects on gene expression

In observational genomics data sets, there is often confounding of the effect of an exposure on gene expression. To adjust for confounding when estimating the exposure effect, a common approach involves including potential confounders as covariates with the exposure in a regression model of gene expression. However, when the exposure and confounders interact to influence gene expression, the fitted regression model does not necessarily estimate the overall effect of the exposure. Using inverse probability weighting (IPW) or the parametric g-formula in these instances is straightforward to apply and yields consistent effect estimates. IPW can readily be integrated into a genomics data analysis pipeline with upstream data processing and normalization, while the g-formula can be implemented by making simple alterations to the regression model. The regression, IPW, and g-formula approaches to exposure effect estimation are compared herein using simulations; advantages and disadvantages of each approach are explored. The methods are applied to a case study estimating the effect of current smoking on gene expression in adipose tissue.

Doubly robust estimator of risk in the presence of censoring dependent on time-varying covariates: application to a primary prevention trial for coronary events with pravastatin

Background

In the presence of dependent censoring even after stratification of baseline covariates, the Kaplan–Meier estimator provides an inconsistent estimate of risk. To account for dependent censoring, time-varying covariates can be used along with two statistical methods: the inverse probability of censoring weighted (IPCW) Kaplan–Meier estimator and the parametric g-formula estimator. The consistency of the IPCW Kaplan–Meier estimator depends on the correctness of the model specification of censoring hazard, whereas that of the parametric g-formula estimator depends on the correctness of the models for event hazard and time-varying covariates.

Methods

We combined the IPCW Kaplan–Meier estimator and the parametric g-formula estimator into a doubly robust estimator that can adjust for dependent censoring. The estimator is theoretically more robust to model misspecification than the IPCW Kaplan–Meier estimator and the parametric g-formula estimator. We conducted simulation studies with a time-varying covariate that affected both time-to-event and censoring under correct and incorrect models for censoring, event, and time-varying covariates. We applied our proposed estimator to a large clinical trial data with censoring before the end of follow-up.

Results

Simulation studies demonstrated that our proposed estimator is doubly robust, namely it is consistent if either the model for the IPCW Kaplan–Meier estimator or the models for the parametric g-formula estimator, but not necessarily both, is correctly specified. Simulation studies and data application demonstrated that our estimator can be more efficient than the IPCW Kaplan–Meier estimator.

Conclusions

The proposed estimator is useful for estimation of risk if censoring is affected by time-varying risk factors.

G-computation, propensity score-based methods, and targeted maximum likelihood estimator for causal inference with different covariates sets: a comparative simulation study

Controlling for confounding bias is crucial in causal inference. Distinct methods are currently employed to mitigate the effects of confounding bias. Each requires the introduction of a set of covariates, which remains difficult to choose, especially regarding the different methods. We conduct a simulation study to compare the relative performance results obtained by using four different sets of covariates (those causing the outcome, those causing the treatment allocation, those causing both the outcome and the treatment allocation, and all the covariates) and four methods: g-computation, inverse probability of treatment weighting, full matching and targeted maximum likelihood estimator. Our simulations are in the context of a binary treatment, a binary outcome and baseline confounders. The simulations suggest that considering all the covariates causing the outcome led to the lowest bias and variance, particularly for g-computation. The consideration of all the covariates did not decrease the bias but significantly reduced the power. We apply these methods to two real-world examples that have clinical relevance, thereby illustrating the real-world importance of using these methods. We propose an R package RISCA to encourage the use of g-computation in causal inference.

Association between early antibiotic treatment and clinical outcomes in children hospitalized for asthma exacerbation

BACKGROUND

Professional society guidelines recommend against routine early antibiotic use in the treatment of asthma exacerbation without comorbid bacterial infection. However, high antibiotic prescribing rates have been reported in developed countries.

OBJECTIVE

We sought to assess the effectiveness of this strategy in the routine care of children.

METHODS

Using data on 48,743 children hospitalized for asthma exacerbation with no indication of bacterial infection during the period 2010 to 2018, we conducted a retrospective cohort study to compare clinical outcomes and resource utilization between children who received early antibiotic treatment and those who did not.

RESULTS

Overall, 19,866 children (41%) received early antibiotic treatment. According to the propensity score matching analysis, children with early antibiotic treatment had longer hospital stay (mean difference, 0.21 days; 95% CI, 0.18-0.28), higher hospitalization costs (mean difference, $83.5; 95% CI, 62.9-104.0), and higher risk of probiotic use (risk ratio, 2.01; 95% CI, 1.81-2.23) than children who did not receive early antibiotic therapy. Similar results were found from inverse probability of treatment weighting, g-computation, and instrumental variable methods and sensitivity analyses. The risks of mechanical ventilation and 30-day readmission were similar between the groups or slightly higher in the treated group, depending on the statistical models.

CONCLUSIONS

Antibiotic therapy may be associated with prolonged hospital stay, elevated hospitalization costs, and high risk of probiotic use without improving treatment failure and readmission. Our findings highlight the need for reducing inappropriate antibiotic use among children hospitalized for asthma.

Effects of prenatal bacterial infection on cognitive performance in early childhood

BACKGROUND

Previous epidemiologic studies have reported adverse neurodevelopmental sequelae following prenatal infectious exposure, yet long-term effects estimated from these observational studies are often subject to biases due to confounding and loss to follow-up.

OBJECTIVES

We demonstrate the joint use of inverse probability (IP) treatment and censoring weights when evaluating neurotoxic effects of prenatal bacterial infection.

METHODS

We applied IP weighting for both treatment and censoring to estimate the effects of maternal bacterial infection during pregnancy on mean intelligence quotient (IQ) scores measured at age 7 using the Wechsler Intelligence Scale for Children. Participants were members of a population-based pregnancy cohort recruited in the Boston and Providence sites of the Collaborative Perinatal Project between 1959 and 1966 (n = 11 984). We calculated average treatment effects (ATE) and average treatment effects on the treated (ATT) using IP weights estimated via generalized boosted models.

RESULTS

ATE- and ATT-weighted mean IQ scores were lowest among offspring exposed to multi-systemic bacterial infection during pregnancy and highest for those unexposed. The effects of prenatal bacterial infection were greater among male offspring, particularly on performance IQ scores. Offspring who were exposed to multi-systemic bacterial infection in the third trimester displayed the largest reduction in mean full-scale, verbal, and performance IQ scores at age 7 compared to those unexposed or exposed in earlier trimesters.

CONCLUSIONS

We find that prenatal bacterial infection is associated with cognitive impairments at age 7. Associations are strongest for more severe infections, that occur in the third trimester, and among males. Public health intervention targeting bacterial infection in pregnant women may help enhance the cognitive development of offspring.

Impact of a Pharmacist-Led Intervention on 30-Day Readmission and Assessment of Factors Predictive of Readmission in African American Men With Heart Failure

Heart failure (HF) is responsible for more 30-day readmissions than any other condition. Minorities, particularly African American males (AAM), are at much higher risk for readmission than the general population. In this study, demographic, social, and clinical data were collected from the electronic medical records of 132 AAM patients (control and intervention) admitted with a primary or secondary admission diagnosis of HF. Both groups received guideline-directed therapy for HF. Additionally the intervention group received a pharmacist-led intervention. Data collected from these patients were used to develop and validate a predictive model to evaluate the impact of the pharmacist-led intervention, and identify predictors of readmission in this population. After propensity score matching, the intervention was determined to have a significant impact on readmission, as a significantly smaller proportion of patients in the intervention group were readmitted as compared to the control group (11.5% vs. 42.9%; p = .03). A predictive model for 30-day readmission was developed using K-nearest neighbor (KNN) classification algorithm. The model was able to correctly classify about 71% patients with an AUROC of 0.70. Additionally, the model provided a set of key patient attributes predictive of readmission status. Among these predictive attributes was whether or not a patient received the intervention. A relative risk analysis identified that patients who received the intervention are less likely to be readmitted within 30 days. This study demonstrated the benefit of a pharmacist-led intervention for AAM with HF. Such interventions have the potential to improve quality of life for this patient population.

A review of time scale fundamentals in the g-formula and insidious selection bias

PURPOSE OF REVIEW

We review recent examples of data analysis with the g-formula, a powerful tool for analyzing longitudinal data and survival analysis. Specifically, we focus on the common choices of time scale and review inferential issues that may arise.

RECENT FINDINGS

Researchers are increasingly engaged with questions that require time scales subject to left-truncation and right-censoring. The assumptions necessary for allowing right-censoring are well defined in the literature, whereas similar assumptions for left-truncation are not well defined. Policy and biologic considerations sometimes dictate that observational data must be analyzed on time scales that are subject to left-truncation, such as age.

SUMMARY

Further consideration of left-truncation is needed, especially when biologic or policy considerations dictate that age is the relevant time scale of interest. Methodologic development is needed to reduce potential for bias when left-truncation may occur.

Health providers pass knowledge and abilities acquired by training in obstetric emergencies to their peers: the average treatment on the treated effect of PRONTO on delivery attendance in Mexico

Background

A significant proportion of newborn and maternal deaths can be prevented through simple and cost-effective strategies. The main aim of this study was to evaluate the impact of the PRONTO obstetric-emergency management training for improving evidence-based birth attendance practices among providers attending the training at 12 hospitals in three states of Mexico from 2010 to 2012, and to estimate dissemination of the training within the hospitals.

Methods

The average treatment on the treated effect of the PRONTO intervention for the probability of performing certain practices during birth attendance was estimated in a sample of 310 health providers. Impact estimates were obtained by performing provider-level matching using a mixed Mahalanobis distance one-to-one nearest-neighbor and exact matching approach. A secondary analysis estimated the positive externalities caused by the intervention in the treated hospitals using the same analytical approach. Provider-level fixed effects regression models were used to estimate the rate of decay of the probability of performing the examined practices.

Results

Providers attending the PRONTO training showed significant increases in the probability of performing the complete active management of the third stage of labor, especially the first and third steps, and skin-to-skin-contact. There was a negative and significant effect on the probability of performing uterine sweeping. Providers who did not attend the training in treated hospitals also showed marked significant changes in the same practices, except for uterine sweeping. There was no evidence of a significant decay of the probability of performing the routine practices over time among the treated providers.

Conclusions

PRONTO is efficacious in changing trained providers’ behavior, but not on all practices, suggesting that some practices are deeply ingrained. The results also suggest that information on practices is effectively transmitted to peers within treated hospitals. Previous findings of the dilution of the effect of PRONTO on some practices seem to be more related to the rotation of personnel (mainly interns) rather than providers returning to their former habits.

Trial registration

NCT01477554. Registered on November 18, 2011; retrospectively registered.

Electronic supplementary material

The online version of this article (10.1186/s12884-018-1872-4) contains supplementary material, which is available to authorized users.

Intraoperative milrinone versus dobutamine in cardiac surgery patients: a retrospective cohort study on mortality

Background

Several choices of inotropic therapy are available and used in relation to cardiac surgery. Comparisons are necessary to select optimal therapy. In Denmark, dobutamine and milrinone are the two inotropic agents most commonly used to treat post-bypass low cardiac output syndrome. This study compares all-cause mortality with these drugs.

Methods

In a retrospective observational study we investigated 10,700 consecutive patients undergoing cardiac surgery from 1 April 2006 to 31 December 2013 at Aarhus and Aalborg University Hospitals in the Central and Northern Denmark Region. Prospectively entered data in the Western Danish Heart Registry on intraoperative use of inotropes were used to identify 952 patients treated with milrinone, 418 patients treated with dobutamine, and 82 patients receiving a combination of the two inotropes. All-cause mortality among patients receiving dobutamine was compared to all-cause mortality among milrinone receivers.

Multiple logistic regression analyses including preoperative and intraoperative variables along with g-formula analyses were used to model 30-day and 1-year mortality risks. Reported were standardized mortality risk differences between the treatment groups.

Results

Among patients receiving intraoperative dobutamine, 18 (4.3%) died within 30 days and 49 (11.7%) within 1 year. Corresponding 30-day and 1-year mortality for milrinone receivers were 81 (8.5%) and 170 (17.9%). Risk of death within 30 days and 1 year was increased for intraoperative milrinone compared to dobutamine with a standardized risk difference of 4.06% (confidence interval (CI) 1.23; 6.89, p = 0.005) and 4.77% (CI 0.39; 9.15, p = 0.033), respectively. Sensitivity analyses including adjustment for milrinone preference, hemodynamic instability prior to cardiopulmonary bypass, and separate analyses on hospital level all confirmed a sign toward increased mortality among milrinone receivers.

Conclusions

Intraoperative use of milrinone in cardiac surgery may be associated with an increase in all-cause mortality compared to use of dobutamine.

Electronic supplementary material

The online version of this article (10.1186/s13054-018-1969-1) contains supplementary material, which is available to authorized users.