Sampling weighting strategies in causal mediation analysis

BACKGROUND

Causal mediation analysis plays a crucial role in examining causal effects and causal mechanisms. Yet, limited work has taken into consideration the use of sampling weights in causal mediation analysis. In this study, we compared different strategies of incorporating sampling weights into causal mediation analysis.

METHODS

We conducted a simulation study to assess 4 different sampling weighting strategies-1) not using sampling weights, 2) incorporating sampling weights into mediation "cross-world" weights, 3) using sampling weights when estimating the outcome model, and 4) using sampling weights in both stages. We generated 8 simulated population scenarios comprising an exposure (A), an outcome (Y), a mediator (M), and six covariates (C), all of which were binary. The data were generated so that the true model of A given C and the true model of A given M and C were both logit models. We crossed these 8 population scenarios with 4 different sampling methods to obtain 32 total simulation conditions. For each simulation condition, we assessed the performance of 4 sampling weighting strategies when calculating sample-based estimates of the total, direct, and indirect effects. We also applied the four sampling weighting strategies to a case study using data from the National Survey on Drug Use and Health (NSDUH).

RESULTS

Using sampling weights in both stages (mediation weight estimation and outcome models) had the lowest bias under most simulation conditions examined. Using sampling weights in only one stage led to greater bias for multiple simulation conditions.

DISCUSSION

Using sampling weights in both stages is an effective approach to reduce bias in causal mediation analyses under a variety of conditions regarding the structure of the population data and sampling methods.

Comparative Effectiveness of Valoctocogene Roxaparvovec and Prophylactic Factor VIII Replacement in Severe Hemophilia A

INTRODUCTION

A prospective, non-interventional study (270-902) followed 294 adults with severe hemophilia A (SHA) receiving prophylactic factor VIII (FVIII). From these participants, 112 rolled over into a single-arm, multicenter, phase 3 trial (GENEr8-1; NCT03370913) that evaluated efficacy and safety of valoctocogene roxaparvovec, a gene therapy that provides endogenous FVIII in individuals with SHA. Participants from 270-902 who did not roll over provide an opportunity for a contemporaneous external control. Therefore, the comparative effectiveness of valoctocogene roxaparvovec vs FVIII prophylaxis was evaluated using propensity scoring (PS).

METHODS

This post hoc analysis compared 112 participants from GENEr8-1 (treated cohort) to 73 participants in 270-902 who did not enroll in GENEr8-1 (control cohort). The primary analysis used standardized mortality ratio weighting to re-weight baseline characteristics of the control cohort to better match the treated cohort. Mean annualized bleeding rates (ABR) for treated and all bleeds were compared between cohorts along with the proportion of participants with zero bleeds (treated and all bleeds). Sensitivity and scenario analyses were also conducted.

RESULTS

PS adjustments reduced differences in baseline characteristics between cohorts. Mean treated (4.40 vs 0.85; P < 0.001) and all (5.01 vs 1.54; P < 0.001) ABR were significantly lower, and the proportions of participants with zero treated bleeds (82.1% vs 32.9%; P < 0.001) and all bleeds (58.0% vs 28.5%; P < 0.001) were significantly higher in GENEr8-1.

CONCLUSIONS

PS-adjusted analyses were consistent with prior intra-individual comparisons. Compared with participants receiving prophylactic FVIII, the participants receiving valoctocogene roxaparvovec experienced lower ABR, and a higher proportion had zero bleeds.

TRAIL REGISTRATION

ClinicalTrials.gov identifier, NCT03370913.

Association of spontaneous abortion with bipolar disorder and major depression based on inverse probability treatment weighting of multigroup propensity scores: Evidence from the UK Biobank

BACKGROUND

Few studies have explored the association between the number of SAs and bipolar disorder and major depression (BDMD). This study aims to investigate the association between SA and BDMD, and the possible dose-response relationship between them.

METHODS

We conducted a cross-sectional study of 13,200 female UK Biobank participants. Participants were classified into BDMD and no-BDMD groups based on their BDMD status. The number of SAs was grouped into non-SA, occasional SA (OSA), and recurrent SA (RSA). Baseline characteristics of the three groups were balanced using inverse probability treatment weighting (IPTW) based on propensity scores. The three-knots restricted cubic spline regression model was utilized to assess the dose-response relationship between the number of SAs and BDMD.

RESULTS

The IPTW-adjusted multivariate logistic regression revealed that SA was an independent risk factor for BDMD, with adjusted OR of 1.12 (95 % CI: 1.07-1.19) and 1.32 (95 % CI: 1.25-1.40) in the OSA and RSA groups, respectively. The strength of this association amplified as the number of SAs (P for trend <0.001). There was a nonlinear relationship between the number of SAs and the risk of BDMD, with an approximately inverted L-shaped curve.

LIMITATIONS

The information of the SA and BDMD status relied on self-reported by volunteers, and the study sample was mostly of European descent.

CONCLUSIONS

Women who reported experiencing multiple SAs are more likely to have BDMD. Therefore, it is imperative to provide psychological care and interventions for women in the postpartum period.

Understanding the effects of Haemophilus influenzae colonization on bronchiectasis: a retrospective cohort study

BACKGROUND

Bacterial colonization is an essential aspect of bronchiectasis. Although Haemophilus influenzae is a frequent colonizer in some regions, its clinical impacts are poorly understood. This study aimed to elucidate the impact of H. influenzae colonization in patients with bronchiectasis.

METHODS

This retrospective study screened adult patients diagnosed with bronchiectasis at a tertiary referral center between April 1, 2003, and May 16, 2021, in South Korea. Propensity score matching was used to match patients with and without H. influenzae colonization. We assessed the severity of bronchiectasis as per the bronchiectasis severity index, the incidence of exacerbation, differences in lung function, and all-cause mortality.

RESULTS

Out of the 4,453 patients with bronchiectasis, 79 (1.8%) were colonized by H. influenzae. After 1:2 propensity score matching, 78 and 154 patients were selected from the H. influenzae colonizer and non-colonizer groups, respectively. Although there were no significant differences between the groups regarding baseline demographics, patients colonized with H. influenzae had a higher bronchiectasis severity index (median 6 [interquartile range 4-8] vs. 4 [2-7], p = 0.002), associated with extensive radiographic involvement (52.2% vs. 37.2%, p = 0.045) and mild exacerbation without hospitalization (adjusted incidence rate ratio 0.15; 95% confidence interval 0.12-0.24). Lung function and mortality rates did not reveal significant differences, regardless of H. influenzae colonization.

CONCLUSION

H. influenzae colonization in bronchiectasis was associated with more severe disease and greater incidence of mild exacerbation, but not lung function and mortality. Attention should be paid to patients with bronchiectasis with H. influenzae colonization.

Confounding-adjustment methods for the causal difference in medians

BACKGROUND

With continuous outcomes, the average causal effect is typically defined using a contrast of expected potential outcomes. However, in the presence of skewed outcome data, the expectation (population mean) may no longer be meaningful. In practice the typical approach is to continue defining the estimand this way or transform the outcome to obtain a more symmetric distribution, although neither approach may be entirely satisfactory. Alternatively the causal effect can be redefined as a contrast of median potential outcomes, yet discussion of confounding-adjustment methods to estimate the causal difference in medians is limited. In this study we described and compared confounding-adjustment methods to address this gap.

METHODS

The methods considered were multivariable quantile regression, an inverse probability weighted (IPW) estimator, weighted quantile regression (another form of IPW) and two little-known implementations of g-computation for this problem. Methods were evaluated within a simulation study under varying degrees of skewness in the outcome and applied to an empirical study using data from the Longitudinal Study of Australian Children.

RESULTS

Simulation results indicated the IPW estimator, weighted quantile regression and g-computation implementations minimised bias across all settings when the relevant models were correctly specified, with g-computation additionally minimising the variance. Multivariable quantile regression, which relies on a constant-effect assumption, consistently yielded biased results. Application to the empirical study illustrated the practical value of these methods.

CONCLUSION

The presented methods provide appealing avenues for estimating the causal difference in medians.

Semi-supervised mixture multi-source exchangeability model for leveraging real-world data in clinical trials

The traditional trial paradigm is often criticized as being slow, inefficient, and costly. Statistical approaches that leverage external trial data have emerged to make trials more efficient by augmenting the sample size. However, these approaches assume that external data are from previously conducted trials, leaving a rich source of untapped real-world data (RWD) that cannot yet be effectively leveraged. We propose a semi-supervised mixture (SS-MIX) multisource exchangeability model (MEM); a flexible, two-step Bayesian approach for incorporating RWD into randomized controlled trial analyses. The first step is a SS-MIX model on a modified propensity score and the second step is a MEM. The first step targets a representative subgroup of individuals from the trial population and the second step avoids borrowing when there are substantial differences in outcomes among the trial sample and the representative observational sample. When comparing the proposed approach to competing borrowing approaches in a simulation study, we find that our approach borrows efficiently when the trial and RWD are consistent, while mitigating bias when the trial and external data differ on either measured or unmeasured covariates. We illustrate the proposed approach with an application to a randomized controlled trial investigating intravenous hyperimmune immunoglobulin in hospitalized patients with influenza, while leveraging data from an external observational study to supplement a subgroup analysis by influenza subtype.

Frameworks for estimating causal effects in observational settings: comparing confounder adjustment and instrumental variables

To estimate causal effects, analysts performing observational studies in health settings utilize several strategies to mitigate bias due to confounding by indication. There are two broad classes of approaches for these purposes: use of confounders and instrumental variables (IVs). Because such approaches are largely characterized by untestable assumptions, analysts must operate under an indefinite paradigm that these methods will work imperfectly. In this tutorial, we formalize a set of general principles and heuristics for estimating causal effects in the two approaches when the assumptions are potentially violated. This crucially requires reframing the process of observational studies as hypothesizing potential scenarios where the estimates from one approach are less inconsistent than the other. While most of our discussion of methodology centers around the linear setting, we touch upon complexities in non-linear settings and flexible procedures such as target minimum loss-based estimation and double machine learning. To demonstrate the application of our principles, we investigate the use of donepezil off-label for mild cognitive impairment. We compare and contrast results from confounder and IV methods, traditional and flexible, within our analysis and to a similar observational study and clinical trial.

Propensity scores as a novel method to guide sample allocation and minimize batch effects during the design of high throughput experiments

BACKGROUND

We developed a novel approach to minimize batch effects when assigning samples to batches. Our algorithm selects a batch allocation, among all possible ways of assigning samples to batches, that minimizes differences in average propensity score between batches. This strategy was compared to randomization and stratified randomization in a case-control study (30 per group) with a covariate (case vs control, represented as β1, set to be null) and two biologically relevant confounding variables (age, represented as β2, and hemoglobin A1c (HbA1c), represented as β3). Gene expression values were obtained from a publicly available dataset of expression data obtained from pancreas islet cells. Batch effects were simulated as twice the median biological variation across the gene expression dataset and were added to the publicly available dataset to simulate a batch effect condition. Bias was calculated as the absolute difference between observed betas under the batch allocation strategies and the true beta (no batch effects). Bias was also evaluated after adjustment for batch effects using ComBat as well as a linear regression model. In order to understand performance of our optimal allocation strategy under the alternative hypothesis, we also evaluated bias at a single gene associated with both age and HbA1c levels in the 'true' dataset (CAPN13 gene).

RESULTS

Pre-batch correction, under the null hypothesis (β1), maximum absolute bias and root mean square (RMS) of maximum absolute bias, were minimized using the optimal allocation strategy. Under the alternative hypothesis (β2 and β3 for the CAPN13 gene), maximum absolute bias and RMS of maximum absolute bias were also consistently lower using the optimal allocation strategy. ComBat and the regression batch adjustment methods performed well as the bias estimates moved towards the true values in all conditions under both the null and alternative hypotheses. Although the differences between methods were less pronounced following batch correction, estimates of bias (average and RMS) were consistently lower using the optimal allocation strategy under both the null and alternative hypotheses.

CONCLUSIONS

Our algorithm provides an extremely flexible and effective method for assigning samples to batches by exploiting knowledge of covariates prior to sample allocation.

Patient-reported health outcomes of SARS-CoV-2-tested patients presenting to emergency departments: a propensity score-matched prospective cohort study

OBJECTIVE

This study aimed to compare the long-term physical and mental health outcomes of matched severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2)-positive and SARS-CoV-2-negative patients controlling for seasonal effects.

STUDY DESIGN

This was a retrospective cohort study.

METHODS

This study enrolled patients presenting to emergency departments participating in the Canadian COVID-19 Emergency Department Rapid Response Network. We enrolled consecutive eligible consenting patients who presented between March 1, 2020, and July 14, 2021, and were tested for SARS-CoV-2. Research assistants randomly selected four site and date-matched SARS-CoV-2-negative controls for every SARS-CoV-2-positive patient and interviewed them at least 30 days after discharge. We used propensity scores to match patients by baseline characteristics and used linear regression to compare Veterans RAND 12-item physical health component score (PCS) and mental health component scores (MCS), with higher scores indicating better self-reported health.

RESULTS

We included 1170 SARS-CoV-2-positive patients and 3716 test-negative controls. The adjusted mean difference for PCS was 0.50 (95% confidence interval [CI]: -0.36, 1.36) and -1.01 (95% CI: -1.91, -0.11) for MCS. Severe disease was strongly associated with worse PCS (β = -7.4; 95% CI: -9.8, -5.1), whereas prior mental health illness was strongly associated with worse MCS (β = -5.4; 95% CI: -6.3, -4.5).

CONCLUSION

Physical health, assessed by PCS, was similar between matched SARS-CoV-2-positive and SARS-CoV-2-negative patients, whereas mental health, assessed by MCS, was worse during a time when the public experienced barriers to care. These results may inform the development and prioritization of support programs for patients.

Logistic regression frequently outperformed propensity score methods, especially for large datasets: a simulation study

OBJECTIVES

In observational studies, researchers must select a method to control for confounding. Options include propensity score (PS) methods and regression. It remains unclear how dataset characteristics (size, overlap in PSs, and exposure prevalence) influence the relative performance of the methods.

STUDY DESIGN AND SETTING

A simulation study to evaluate the role of dataset characteristics on the performance of PS methods, compared to logistic regression, for estimating a marginal odds ratio was conducted. Dataset size, overlap in PSs, and exposure prevalence were varied.

RESULTS

Regression showed poor coverage for small sample sizes, but with large sample sizes was relatively robust to imbalance in PSs and low exposure prevalence. PS methods displayed suboptimal coverage as overlap in PSs decreased, which was exacerbated at larger sample sizes. Power of matching methods was particularly affected by a lack of overlap, low exposure prevalence, and small sample size. The advantage of regression for large data size was reduced in sensitivity analysis with a complementary log-log outcome generation mechanism and unmeasured confounding, with superior bias and error but inferior coverage to matching methods.

CONCLUSION

Dataset characteristics influence performance of methods for confounder adjustment. In many scenarios, regression may be the preferable option.

CT Radiomics Features in Differentiation of Focal-Type Autoimmune Pancreatitis from Pancreatic Ductal Adenocarcinoma: A Propensity Score Analysis

PURPOSE

To evaluate the diagnostic performance of the radiomics score (rad-score) for differentiating focal-type autoimmune pancreatitis (fAIP) from pancreatic ductal adenocarcinoma (PDAC).

METHODS

This retrospective review included 42 consecutive patients with fAIP diagnosed according to the International Consensus Diagnostic Criteria between January 2011 and December 2018. Furthermore, 334 consecutive patients with PDAC confirmed by pathology were also reviewed during the same period. Patients with PDAC and fAIP were matched via propensity score matching (PSM). All patients underwent multidetector computed tomography (MDCT). For each patient, 1409 radiomics features of the portal phase were extracted and reduced using the least absolute shrinkage and selection operator (LASSO) logistic regression algorithm. The portal rad-score performance was assessed based on its discriminative ability.

RESULTS

After PSM, we matched 55 patients with PDAC to 42 patients with fAIP, based on clinical and CT characteristics (e.g., patient age, sex, body mass index, location, size, enhanced mode). A rad-score for discriminating fAIP from PDAC, which contained four CT derived radiomic features, was developed (area under the curve = 0.97). The sensitivity, specificity, and accuracy of the radiomics model were 95.24%, 92.73% and 0.94, respectively.

CONCLUSION

The portal rad-score can accurately and noninvasively differentiate fAIP from PDAC.

Volunteer Recovery Support for Adolescents: Using propensity score based methods to understand dosage effects within a randomized controlled trial

BACKGROUND

In a recently published randomized controlled trial (RCT) of Volunteer Recovery Support for Adolescents (VRSA), a secondary finding indicated that better adherence to planned VRSA telephone session frequency resulted in significantly higher remission rates relative to lower session adherence. However, interpretation of this dose-response relationship may have been confounded by participant characteristics such as baseline levels of substance use and mental health problems.

METHODS

The present study used statistical methods designed to approximate RCTs when comparing more than two nonequivalent groups that include an assessment of the potential impact of omitted variables. Classification and Regression Tree (CRT) analysis was used to establish the cut-point between high (H) and low (L) VRSA dosage groups. Because we were interested in generalizing to youth with poor attendance, the L-VRSA group served as the reference group. Balancing weights for H-VRSA and a services as usual (SAU) control group were calculated to ensure similarity of baseline pretreatment characteristics to the reference group, and sensitivity of findings to unobserved confounding variables was assessed.

RESULTS

Findings suggested that superior remission rates at the end of the intervention phase were the result of high adherence to planned VRSA session frequency. Recommendations to achieve high VRSA participation among a larger segment of youth and to test whether longer VRSA duration improves the stability of recovery outcomes are provided.

CONCLUSION

Few published dose-response studies have adequately controlled for selection confounds from both observed and unobserved confounding. As such, the present study aims to both assess the impact of different dosage levels of VRSA and provide a template for how to apply state-of-the-art statistical methods designed to approximate randomized controlled trials to such studies.

COVID-19 and the impact of cash transfers on health care use in Togo

Background

Cash transfer program during pandemics provide a social protection mechanism to improve the health of the most vulnerable households. This article analysis the impact of cash transfers on household demand for health care during Covid-19.

Methods

Using data from the survey conducted from 8th to 17th July 2020 covering all 44 districts in the 6 health regions of Togo under the direction of the United Nations Office for Project Services (UNOPS), we used propensity score matching and the ESR model. These models allow us to analysis the impact of cash transfers on health care use during Covid-19.

Results

Analysis of the results shows a positive impact of cash transfers on the use of health care services in Togo for beneficiary households. In addition, the health insurance promotes the use of health care among households’ socio-economic factors.

Conclusion

This cash transfer program is an effective approach to improving access to health care services for the most vulnerable households, particularly in the most disadvantaged settings. Thus, policy makers need to extend these cash transfers to a large part of the population during this Covid-19 health crisis as it has a positive impact on the demand for health care.

Iodized oil computed tomography versus ultrasound-guided radiofrequency ablation for early hepatocellular carcinoma

BACKGROUND AND PURPOSE

Radiofrequency ablation (RFA) is the standard of care for early stage hepatocellular carcinoma (HCC). However, the clinical outcomes of iodized oil computed tomography (IoCT) versus ultrasound (US)-guided RFA for HCC remain unclear.

METHODS

We retrospectively analyzed consecutive treatment-naïve patients who received curative RFA for HCC within Milan criteria from January 2016 to December 2018. Patients who underwent either IoCT-guided RFA (IoCT group) or US-guided RFA (US group) were included. Various clinical factors, including tumor location, were adjusted with a 1:1 propensity score matching. Subsequently, the cumulative incidence rates for recurrence and hazard ratios for survival were calculated.

RESULTS

We included 184 (37.9%) and 301 (62.1%) patients who received IoCT- and US-guided RFA, respectively. Before propensity score matching, IoCT guidance was significantly associated with multiple tumors, higher body mass index, lower albumin level, and tumors located at S8. After matching, the 1-, 2-, and 3-year local tumor progression rates of the IoCT group were significantly lower than those of the US group (4.4%, 6.9%, and 7.5% vs. 14.4%, 16.3%, and 16.3%, respectively, at p = 0.002, 0.009, and 0.016, respectively). In univariate analyses and multivariate analyses that adjusted for clinical and tumor location-related parameters, the IoCT group had better recurrence-free survival (hazard ratio = 0.581, 95% confidence interval 0.375-0.899) than those with US guidance but not overall survival.

CONCLUSION

IoCT-guided RFA had a lower local tumor progression rate and better recurrence-free survival than did US-guided RFA for HCC within the Milan criteria. CT-guide RFA is a safe and effective alternative to US-guided with similar overall survival. IoCT-guided RFA might have a better local tumor control than US-guided. IoCT-guided RFA may be more suitable for male patients, aged < 70 years, a single tumor measuring 2-5 cm, and a tumor located at the subdiaphragmatic/subcardiac region.

Methods for evaluating the benefit and harms of deprescribing in observational research using routinely collected data

Deprescribing is defined as "the planned and supervised process of dose reduction or stopping of medication that might be causing harm, or no longer be of benefit". Barriers to deprescribing include healthcare professional fear and lack of guidance. These may stem from limited available evidence on benefits and harms of deprescribing medications commonly used among older persons. Advances in pharmacoepidemiology and causal inference methods to evaluate comparative effectiveness and safety of prescribing medications have yet to be considered for deprescribing medication. This paper discusses select methods and how they can be applied to deprescribing research, using case studies of benzodiazepines and low-dose acetylsalicylic acid (aspirin). Target trial emulation involves the explicit application of design principles from randomised controlled trials to observational studies. Several design aspects, including defining eligibility criteria and time zero, require additional considerations for deprescribing studies. The active comparator new user design also presents challenges, including selection of an appropriate comparator. This paper discusses these aspects, and others, in relation to deprescribing studies. Furthermore, methods proposed to control for confounding, in particular, the prior event rate ratio and propensity scores, are discussed. Introduction of billing codes or mechanisms for accurately determining when deprescribing has occurred would enhance the ability to conduct research using routinely collected data. Although the approaches discussed in this paper may strengthen observational studies of deprescribing, their use may be best suited to certain scenarios or research questions, where randomised controlled trials may be less feasible.

The effects of competition and bundled payment on patient reported outcome measures after hip replacement surgery

BACKGROUND

Competition-promoting reforms and economic incentives are increasingly being introduced worldwide to improve the performance of healthcare delivery. This study considers such a reform which was initiated in 2009 for elective hip replacement surgery in Stockholm, Sweden. The reform involved patient choice of provider, free establishment of new providers and a bundled payment model. The study aimed to examine its effects on hip replacement surgery quality as captured by patient reported outcome measures (PROMs) of health gain (as indicated by the EQ-5D index and a visual analogue scale (VAS)), pain reduction (VAS) and patient satisfaction (VAS) one and six years after the surgery.

METHODS

Using patient-level data collected from multiple national registers, we applied a quasi-experimental research design. Data were collected for elective primary total hip replacements that were carried out between 2008 and 2012, and contain information on patient demography, the surgery and PROMs at baseline and at one- and six-years follow-up. In total, 36,627 observations were included in the analysis. First, entropy balancing was applied in order to reduce differences in observable characteristics between treatment groups. Second, difference-in-difference analyses were conducted to eliminate unobserved time-invariant differences between treatment groups and to estimate the causal treatment effects.

RESULTS

The entropy balancing was successful in creating balance in all covariates between treatment groups. No significant effects of the reform were found on any of the included PROMs at one- and six-years follow-up. The sensitivity analyses showed that the results were robust.

CONCLUSIONS

Competition and bundled payment had no effects on the quality of hip replacement surgery as captured by post-surgery PROMs of health gain, pain reduction and patient satisfaction. The study provides important insights to the limited knowledge on the effects of competition and economic incentives on PROMs.

Post-tuberculosis airway disease: A population-based cohort study of people immigrating to British Columbia, Canada, 1985-2015

BACKGROUND

Current epidemiological evidence of post-TB airway disease is largely cross-sectional and derived from high-TB-incidence settings. We present the first cohort study of post-TB airway disease in a low-TB-incidence setting.

AIMS

(1) analyze the risk of airway disease by respiratory TB, (2) assess potential unmeasured confounding between TB and airway disease, and (3) investigate TB effect measure modification.

METHODS

A population-based cohort study using healthcare claims data for immigrants to British Columbia (BC), Canada, 1985-2015. Airway disease included chronic airway obstruction, asthma, bronchitis, bronchiolitis, and emphysema. Respiratory TB was defined from TB registry data. Cox proportional hazards (PH) regressions were used to analyze time-to-airway disease by respiratory TB. Sensitivity analyses included varying definitions of TB and airway disease. Potential unmeasured confounding by smoking was evaluated by E-value and hybrid least absolute shrinkage and selection operator (LASSO)-high-dimensional propensity score (hdPS).

FINDINGS

In our cohort (N = 1 005 328; n_TB=1141) there were 116 840 incident cases of airway disease during our 30-year study period (10.43 per 1,000 person-years of follow-up), with cumulative incidence of 42·5% among respiratory TB patients compared with 11·6% among non-TB controls. The covariate-adjusted hazard ratio (aHR) for airway disease by respiratory TB was 2·08 (95% CI: 1·91-2·28) with E-value=3·58. The LASSO-hdPS analysis produced aHR=2·26 (95% CI: 2·07-2·47).

INTERPRETATION

A twofold higher risk of airway disease was observed among immigrants diagnosed with respiratory TB, compared with non-TB controls, in a low-TB-incidence setting. Unmeasured confounding is unlikely to explain this relationship. Models of post-TB care are needed.

FUNDING

Canadian Institutes of Health Research.

Estimating the causal effects of modifiable, non-genetic factors on Huntington disease progression using propensity score weighting

INTRODUCTION

Despite being genetically inherited, it is unclear how non-genetic factors (e.g., substance use, employment) might contribute to the progression and severity of Huntington's disease (HD).

METHODS

We used propensity score (PS) weighting in a large (n = 2914) longitudinal dataset (Enroll-HD) to examine the impact of education, employment status, and use of tobacco, alcohol, and recreational and therapeutic drugs on HD progression. Each factor was investigated in isolation while controlling for 19 other factors to ensure that groups were balanced at baseline on potential confounders using PS weights. Outcomes were compared several years later using doubly robust models.

RESULTS

Our results highlighted cases where modifiable (non-genetic) factors - namely light and moderate alcohol use and employment - would have been associated with HD progression in models that did not use PS weights to control for baseline imbalances. These associations did not hold once we applied PS weights to balance baseline groups. We also found potential evidence of a protective effect of substance use (primarily marijuana use), and that those who needed antidepressant treatment were likely to progress faster than non-users.

CONCLUSIONS

Our study is the first to examine the effect of non-genetic factors on HD using a novel application of PS weighting. We show that previously-reported associated factors - including light and moderate alcohol use - are reduced and no longer significantly linked to HD progression after PS weighting. This indicates the potential value of PS weighting in examining non-genetic factors contributing to HD as well as in addressing the known biases that occur with observational data.

Increasing antiviral treatment uptake improves survival in patients with HBV-related HCC

Background & Aims

Antiviral treatment is known to improve survival in patients with chronic hepatitis B (CHB)-related hepatocellular carcinoma (HCC). Yet, the treatment uptake in CHB patients remains low. We aimed to report the secular trend in antiviral treatment uptake from 2007-2017, and to compare the effect of different nucleos(t)ide analogue (NA) initiation times (before vs. after HCC diagnosis) on survival.

Methods

A 3-month landmark analysis was used to compare overall survival in patients not receiving NA treatment (i.e. no NA), patients receiving NAs after their first HCC treatment (i.e. post-HCC NA), and patients receiving NAs ≤3 months before their first HCC treatment (i.e. pre-HCC NA). A propensity score-weighted Cox proportional hazards model was used to balance clinical characteristics between the 3 groups and to estimate hazard ratios (HRs).

Results

The uptake of antiviral treatment in HCC patients increased from 47.3% in 2007 to 98.3% in 2017. The pre-HCC NA group contributed mostly to the uptake rate, which increased from 72.7% to 96.0% in the past decade. In addition, 3,843 CHB patients (407 no NA; 2,932 pre-HCC NA; 504 post-HCC NA) with HCC, receiving at least 1 type of HCC treatment, were included in the analysis. Lack of NA treatment at the time of HCC diagnosis increased the risk of death (weighted HR 3.05; 95% CI 2.70-3.44; p <0.001). The impact of the timing of NA treatment was insignificant (weighted HR 0.90; 95% CI 0.78-1.04; p = 0.161).

Conclusions

The uptake of antiviral treatment in HCC patients increased over the past decade. NA treatment, regardless of whether it was initiated before or after HCC diagnosis, improved survival. It is never too late to initiate NA treatment, even after HCC diagnosis.

Lay summary

More and more patients who have hepatitis B-related liver cancer received antiviral treatment over the past decade. The timing of starting antiviral treatment, regardless of whether it was before or after liver cancer happens, does not really matter in terms of survival benefits.

Continuous versus intermittent short-acting β2-agonists nebulization as first-line therapy in hospitalized children with severe asthma exacerbation: a propensity score matching analysis

Background

Short-acting β2-agonist (SABA) nebulization is commonly prescribed for children hospitalized with severe asthma exacerbation. Either intermittent or continuous delivery has been considered safe and efficient. The comparative efficacy of these two modalities is inconclusive. We aimed to compare these two modalities as the first-line treatments.

Methods

An efficacy research with a retrospective cohort study design was conducted. Hospital records of children with severe asthma exacerbation admitted to Hat Yai Hospital between 2015 and 2017 were retrospectively collected. Children initially treated with continuous salbutamol 10 mg per hour or intermittent salbutamol 2.5 mg per dose over 1–4 h nebulization were matched one-to-one using the propensity score. Competing risk and risk difference regression was applied to evaluate the proportion of children who succeeded and failed the initial treatment. Restricted mean survival time regression was used to compare the length of stay (LOS) between the two groups.

Results

One-hundred and eighty-nine children were included. Of these children, 112 were matched for analysis (56 with continuous and 56 with intermittent nebulization). Children with continuous nebulization experienced a higher proportion of success in nebulization treatment (adjusted difference: 39.5, 95% CI 22.7, 56.3, p < 0.001), with a faster rate of success (adjusted SHR: 2.70, 95% CI 1.73, 4.22, p < 0.001). There was a tendency that LOS was also shorter (adjusted mean difference − 9.9 h, 95% CI -24.2, 4.4, p = 0.176).

Conclusion

Continuous SABA nebulization was more efficient than intermittent nebulization in the treatment of children with severe asthma exacerbation.

The Comparison of Latent Variable Propensity Score Models to Traditional Propensity Score Models under Conditions of Covariate Unreliability

Propensity score (PS) methods are implemented by researchers to balance the differences between participants in control and treatment groups that exist in observational studies using a set of baseline covariates. Propensity scores are most commonly calculated using baseline covariates in a logistic regression model to predict the binary grouping variable (control versus treatment). Low reliability associated with the covariates can adversely impact the calculation of treatment effects in propensity score models. The incorporation of latent variables when calculating propensity scores has been suggested to offset the negative impact of covariate unreliability. Simulation studies were conducted to compare the performance of latent variable methods with traditional propensity score methods when estimating the treatment effect under conditions of covariate unreliability. The results indicated that using factor scores or composite variables to compute propensity scores resulted in biased estimates and inflated Type I error rates as compared to using latent factors to compute propensity scores in certain conditions. This was largely dependent upon the number of infallible covariates also included in the PS model and the outcome analysis model analyzed. Implications of the findings are discussed.

Comparison of methods for handling covariate missingness in propensity score estimation with a binary exposure

Background

Causal effect estimation with observational data is subject to bias due to confounding, which is often controlled for using propensity scores. One unresolved issue in propensity score estimation is how to handle missing values in covariates.

Method

Several approaches have been proposed for handling covariate missingness, including multiple imputation (MI), multiple imputation with missingness pattern (MIMP), and treatment mean imputation. However, there are other potentially useful approaches that have not been evaluated, including single imputation (SI) + prediction error (PE), SI + PE + parameter uncertainty (PU), and Generalized Boosted Modeling (GBM), which is a nonparametric approach for estimating propensity scores in which missing values are automatically handled in the estimation using a surrogate split method. To evaluate the performance of these approaches, a simulation study was conducted.

Results

Results suggested that SI + PE, SI + PE + PU, MI, and MIMP perform almost equally well and better than treatment mean imputation and GBM in terms of bias; however, MI and MIMP account for the additional uncertainty of imputing the missingness.

Conclusions

Applying GBM to the incomplete data and relying on the surrogate split approach resulted in substantial bias. Imputation prior to implementing GBM is recommended.

Examining Intervention Component Dosage Effects on Substance Use Initiation in the Strengthening Families Program: for Parents and Youth Ages 10-14

Family-based prevention programs increasingly are being disseminated and can be effective for an array of adolescent problem behaviors, including substance use initiation. Yet, we continue to have little understanding of how and why these programs work. Increased specificity in our understanding of what components drive program effects can facilitate refinement of programs, with potential for greater impact at a lower cost. Using attendance data, previously coded intervention components, and a previously developed propensity model to adjust for potential bias, this study evaluated content component-specific dosage effects of the Strengthening Families Program: for Parents and Youth Ages 10-14 on three substance use initiation outcomes by grade 12. Results indicated that greater dosages of program content related to (a) parental monitoring and behavior management strategies and (b) promoting positive family relationships had potent and robust effects on reduction of risk for initiating drunkenness and marijuana use and (c) self-regulation and stress management had potent and robust effects on reduction of risk for initiating cigarette and marijuana use. Results indicate potential critical components within SFP 10-14 and offer a path forward for continuing work in efforts to optimize this widely disseminated program.

COMBINING INFORMATION FROM MULTIPLE DATA SOURCES TO ASSESS POPULATION HEALTH

Information about an extensive set of health conditions on a well-defined sample of subjects is essential for assessing population health, gauging the impact of various policies, modeling costs, and studying health disparities. Unfortunately, there is no single data source that provides accurate information about health conditions. We combine information from several administrative and survey data sets to obtain model-based dummy variables for 107 health conditions (diseases, preventive measures, and screening for diseases) for elderly (age 65 and older) subjects in the Medicare Current Beneficiary Survey (MCBS) over the fourteen-year period, 1999-2012. The MCBS has prevalence of diseases assessed based on Medicare claims and provides detailed information on all health conditions but is prone to underestimation bias. The National Health and Nutrition Examination Survey (NHANES), on the other hand, collects self-reports and physical/laboratory measures only for a subset of the 107 health conditions. Neither source provides complete information, but we use them together to derive model-based corrected dummy variables in MCBS for the full range of existing health conditions using a missing data and measurement error model framework. We create multiply imputed dummy variables and use them to construct the prevalence rate and trend estimates. The broader goal, however, is to use these corrected or modeled dummy variables for a multitude of policy analysis, cost modeling, and analysis of other relationships either using them as predictors or as outcome variables.

A review of the use of propensity score diagnostics in papers published in high-ranking medical journals

Background

Propensity scores are widely used to deal with confounding bias in medical research. An incorrectly specified propensity score model may lead to residual confounding bias; therefore it is essential to use diagnostics to assess propensity scores in a propensity score analysis. The current use of propensity score diagnostics in the medical literature is unknown. The objectives of this study are to (1) assess the use of propensity score diagnostics in medical studies published in high-ranking journals, and (2) assess whether the use of propensity score diagnostics differs between studies (a) in different research areas and (b) using different propensity score methods.

Methods

A PubMed search identified studies published in high-impact journals between Jan 1st 2014 and Dec 31st 2016 using propensity scores to answer an applied medical question. From each study we extracted information regarding how propensity scores were assessed and which propensity score method was used. Research area was defined using the journal categories from the Journal Citations Report.

Results

A total of 894 papers were included in the review. Of these, 187 (20.9%) failed to report whether the propensity score had been assessed. Commonly reported diagnostics were p-values from hypothesis tests (36.6%) and the standardised mean difference (34.6%). Statistical tests provided marginally stronger evidence for a difference in diagnostic use between studies in different research areas (p = 0.033) than studies using different propensity score methods (p = 0.061).

Conclusions

The use of diagnostics in the propensity score medical literature is far from optimal, with different diagnostics preferred in different areas of medicine. The propensity score literature may improve with focused efforts to change practice in areas where suboptimal practice is most common.