A comparison of reweighting estimators of average treatment effects in real world populations.
Ontology highlight
ABSTRACT:
SUBMITTER: Lin CY
PROVIDER: S-EPMC8359356 | biostudies-literature |
REPOSITORIES: biostudies-literature
Similar Datasets
Project description:Propensity score matching has been a long-standing tradition for handling confounding in causal inference, however requiring stringent model assumptions. In this article, we propose novel double score matching (DSM) utilizing both the propensity score and prognostic score. To gain the protection of possible model misspecification, we posit multiple candidate models for each score. We show that the de-biasing DSM estimator achieves the multiple robustness property in that it is consistent if any one of the score models is correctly specified. We characterize the asymptotic distribution for the DSM estimator requiring only one correct model specification based on the martingale representations of the matching estimators and theory for local Normal experiments. We also provide a two-stage replication method for variance estimation and extend DSM for quantile estimation. Simulation demonstrates DSM outperforms single score matching and prevailing multiply robust weighting estimators in the presence of extreme propensity scores.
| S-EPMC9949738 | biostudies-literature
Project description:While randomized trials remain the best evidence for treatment effectiveness, lack of generalizability often remains an important concern. Additionally, when new treatments are compared against existing standards of care, the potentially small benefit of the new treatment may be difficult to detect in a trial without extremely large sample sizes and long follow-up times. Recent advances in "data fusion" provide a framework to combine results across studies that are applicable to a given population of interest and allow treatment comparisons that may not be feasible with traditional study designs. We propose a data fusion-based estimator that can be used to combine information from two studies: (1) a study comparing a new treatment to the standard of care in the local population of interest, and (2) a study comparing the standard of care to placebo in a separate, distal population. We provide conditions under which the parameter of interest can be identified from the two studies described and explore properties of the estimator through simulation. Finally, we apply the estimator to estimate the effect of triple- vs monotherapy for the treatment of HIV using data from two randomized trials. The proposed estimator can account for underlying population structures that induce differences in case mix, adherence, and outcome prevalence between the local and distal populations, and the estimator can also account for potentially informative loss to follow-up. Approaches like those detailed here are increasingly important to speed the approval and adoption of effective new therapies by leveraging multiple sources of information.
| S-EPMC8237350 | biostudies-literature
Project description:The experimental study of positions on policies and measures against various new types of threat is fast becoming a mainstream research practice. In this article we argue as follows: in security studies in particular, there is a risk that the experimental treatment is contaminated by subjects' previous experience of the real world ('contamination'), and this may substantially complicate the assessment of the size of the experimental treatment's causal effect. We discuss ways to decrease the risk of uncontrolled contamination. Using two experimental case studies we show two typical cases of contamination in security studies (one, where the contamination of all treatments was extremely high, and another, where the level of contamination was unknown and might have varied across the experimental groups) and consider what this implies for the substantive results of the experiments. An analysis of contamination should become a routine, especially when reporting security experiments.Supplementary informationThe online version contains supplementary material available at 10.1007/s11135-022-01354-4.
| S-EPMC8861618 | biostudies-literature
Project description:ObjectiveCase-control study designs are commonly used in retrospective analyses of real-world evidence (RWE). Due to the increasingly wide availability of RWE, it can be difficult to determine whether findings are robust or the result of testing multiple hypotheses.Materials and methodsWe investigate the potential effects of modifying cohort definitions in a case-control association study between depression and type 2 diabetes mellitus. We used a large (>75 million individuals) de-identified administrative claims database to observe the effects of minor changes to the requirements of glucose and hemoglobin A1c tests in the control group.ResultsWe found that small permutations to the criteria used to define the control population result in significant shifts in both the demographic structure of the identified cohort as well as the odds ratio of association. These differences remain present when testing against age- and sex-matched controls.DiscussionAnalyses of RWE need to be carefully designed to avoid issues of multiple testing. Minor changes to control cohorts can lead to significantly different results and have the potential to alter even prospective studies through selection bias.ConclusionWe believe this work offers strong support for the need for robust guidelines, best practices, and regulations around the use of observational RWE for clinical or regulatory decision-making.
| S-EPMC8206406 | biostudies-literature
Project description:BackgroundDriving is a complex, everyday task that impacts patient agency, safety, mobility, social connections, and quality of life. Digital tools can provide comprehensive real-world (RW) data on driver behavior in patients with Parkinson's disease (PD), providing critical data on disease status and treatment efficacy in the patient's own environment.ObjectiveThis pilot study examined the use of driving data as a RW digital biomarker of PD symptom severity and dopaminergic therapy effectiveness.MethodsNaturalistic driving data (3974 drives) were collected for 1 month from 30 idiopathic PD drivers treated with dopaminergic medications. Prescriptions data were used to calculate levodopa equivalent daily dose (LEDD). The association between LEDD and driver mobility (number of drives) was assessed across PD severity, measured by the Movement Disorder Society Unified Parkinson's Disease Rating Scale (MDS-UPDRS).ResultsPD drivers with worse motor symptoms based on self-report (Part II: P = 0.02) and clinical examination (Part III: P < 0.001) showed greater decrements in driver mobility. LEDD levels >400 mg/day were associated with higher driver mobility than those with worse PD symptoms (Part I: P = 0.02, Part II: P < 0.001, Part III: P < 0.001).ConclusionsResults suggest that comprehensive RW driving data on PD patients may index disease status and treatment effectiveness to improve patient symptoms, safety, mobility, and independence. Higher dopaminergic treatment may enhance safe driver mobility in PD patients with worse symptom severity.
| S-EPMC10525064 | biostudies-literature
Project description:In the presence of time-dependent confounding, there are several methods available to estimate treatment effects. With correctly specified models and appropriate structural assumptions, any of these methods could provide consistent effect estimates, but with real-world data, all models will be misspecified and it is difficult to know if assumptions are violated. In this paper, we investigate five methods: inverse probability weighting of marginal structural models, history-adjusted marginal structural models, sequential conditional mean models, g-computation formula, and g-estimation of structural nested models. This work is motivated by an investigation of the effects of treatments in cystic fibrosis using the UK Cystic Fibrosis Registry data focussing on two outcomes: lung function (continuous outcome) and annual number of days receiving intravenous antibiotics (count outcome). We identified five features of this data that may affect the performance of the methods: misspecification of the causal null, long-term treatment effects, effect modification by time-varying covariates, misspecification of the direction of causal pathways, and censoring. In simulation studies, under ideal settings, all five methods provide consistent estimates of the treatment effect with little difference between methods. However, all methods performed poorly under some settings, highlighting the importance of using appropriate methods based on the data available. Furthermore, with the count outcome, the issue of non-collapsibility makes comparison between methods delivering marginal and conditional effects difficult. In many situations, we would recommend using more than one of the available methods for analysis, as if the effect estimates are very different, this would indicate potential issues with the analyses.
| S-EPMC6001810 | biostudies-literature
Project description:Background:There is lack of real-world treatment pattern comparison data between ixekizumab and adalimumab which are approved for the treatment of moderate-to-severe plaque psoriasis. Objective:To compare real-world treatment patterns among psoriasis patients initiating ixekizumab or adalimumab in the United States. Methods:Psoriasis patients with ?1 claim for ixekizumab or adalimumab between March 1, 2016, and May 31, 2018, were identified (index date = date of first ixekizumab or adalimumab claim) from the IBM Watson Health MarketScan® databases. Patients were required to be continuously enrolled for ?12 months before the index date and followed for a minimum of 6 months until inpatient death, enrollment end, or study end, whichever occurred first. Treatment persistence, adherence, discontinuation, restart, and switching were analyzed. Inverse probability of treatment weighting and multivariable regression modeling were employed to address cohort imbalances and estimate the adjusted risk of non-persistence, discontinuation, and switching, and the odds of adherence. Results:A total of 646 ixekizumab and 3668 adalimumab users were included and followed for a mean of 14.0 and 16.5 months, respectively. Compared to adalimumab, ixekizumab was associated with 19% lower risk of non-persistence (hazard ratio [HR]=0.81, 95% confidence interval [CI]: 0.69-0.95), 26% lower risk of discontinuation (HR=0.74, 95% CI: 0.62-0.88), and 28% lower risk of switching (HR=0.72, 95% CI: 0.57-0.91). Ixekizumab users had higher odds of medication possession ratio ?80% (odds ratio [OR]=1.36, 95% CI: 1.10-1.69) but similar odds by proportion of days covered ?80% (OR=1.22, 95% CI: 0.98-1.53). Conclusion:Psoriasis patients treated with ixekizumab demonstrated longer persistency, higher adherence and were less likely to discontinue or switch treatment compared to adalimumab users. However, while patients achieving highly adherent threshold significantly differed by MPR ?80%, it did not by PDC ?80%; hence, further analysis using fixed-length follow-up is required.
| S-EPMC7074803 | biostudies-literature
Project description:IntroductionHypoglycemia is the most common adverse effect of diabetes therapy, particularly insulin treatment. Hypoglycemia is associated with considerable clinical and economic burden, and may be under-reported. The aim of this study was to com pare the frequency of hypoglycemic events reported in real-world settings with those reported in clinical trials.MethodsWe conducted a structured literature review in PubMed to identify hypoglycemic event rates in patients with type 1 diabetes mellitus (T1DM) and insulin-treated type 2 diabetes mellitus (T2DM) from real-world data (RWD) and randomized controlled trials (RCTs). The search was restricted to English language, full-text publications from 2010 onwards, reporting on treatment of T1DM or T2DM with basal only, basal-bolus, or premix insulin.ResultsThe final dataset included 30 studies (11 RWD studies and 19 RCTs). Six studies (RWD, n = 2; RCT, n = 4) reported hypoglycemia event rates in people with T1DM. For all reported categories of hypoglycemia (severe, non-severe, and nocturnal), rates were consistently higher in RWD studies compared with RCTs. Twenty-five studies (RWD, n = 10; RCT, n = 15) reported hypoglycemia event rates in people with insulin-treated T2DM. For T2DM basal-oral therapy; the highest rates were observed in RWD studies, although there was an overlap with RCT rates. For basal-bolus therapy, there was considerable between-study variability but higher rates of severe and non-severe hypoglycemia were generally observed in RWD studies. For T2DM premix insulin, reported rates of hypoglycemia in RWD studies and RCTs were similar.ConclusionWe found that higher rates of hypoglycemia are observed in real-world settings compared with clinical trial settings, although there is a large degree of overlap. Due to the inherent constraints of RCTs, they are likely to underestimate the burden of hypoglycemia in clinical practice. Further, high-quality RWD are needed to determine a more accurate incidence of hypoglycemia in clinical practice.
| S-EPMC4801820 | biostudies-literature
Project description:Talaromycosis is a leading cause of AIDS-associated opportunistic infections and death in Southeast Asia. We have recently shown in the Itraconazole versus Amphotericin for Talaromycosis (IVAP) trial that induction therapy with amphotericin B reduced mortality over 24 weeks, but not during the first 2 weeks. Antifungal treatment effects in real-world settings have not been rigorously evaluated. Using data obtained from patient records at the Hospital for Tropical Diseases, Ho Chi Minh City, Vietnam from 2004 to 2009, we first developed a prognostic model using Bayesian logistic regression to identify predictors of death. Second, we developed a causal model using propensity score matching to assess the treatment effects of amphotericin B and itraconazole. Our prognostic model identified intravenous drug use (odds ratio [OR] = 2.01), higher respiratory rate (OR = 1.12), higher absolute lymphocyte count (OR = 1.62), a concurrent respiratory infection (OR = 1.67) or central nervous system infection (OR = 2.66) as independent predictors of death. Fever (OR = 0.56) was a protective factor. Our prognostic model exhibits good in-sample performance and out-of-sample validation, with a discrimination power of 0.85 and 0.91, respectively. Our causal model showed no significant difference in treatment outcomes between amphotericin B and itraconazole over the first 2 weeks (95% credible interval: 0.62, 2.50). Our prognostic model provides a simple tool based on routinely collected clinical data to predict individual patient outcome. Our causal model shows similar results to the IVAP trial at 2 weeks, demonstrating an agreement between real-world data and clinical trial data.
| S-EPMC8023982 | biostudies-literature
Project description:Introduction:Cholinesterase inhibitors (ChEIs) are widely used to treat mild to moderate Alzheimer's disease and related dementia. Clinical trials have focused on placebo comparisons, inadequately addressing within-class comparative safety. Methods:New users of ChEIs in British Columbia were categorized into five study cohorts: low-dose donepezil, high-dose donepezil, galantamine, rivastigmine patch, and oral rivastigmine. Comparative safety of ChEIs assessed hazard ratios using propensity score adjusted Cox regression. Results:Compared with low-dose donepezil, galantamine use was associated with a lower risk of mortality (adjusted hazard ratio: 0.84, 95% confidence interval: 0.60-1.18), cardiovascular serious adverse events (adjusted hazard ratio: 0.78, 95% confidence interval: 0.62-0.98), and entry into a residential care facility (adjusted hazard ratio: 0.72, 95% confidence interval: 0.59-0.89). Discussion:Given the absence of randomized trial data showing clinically meaningful benefit of ChEI therapy in Alzheimer's disease, our study suggests preferential use of galantamine may at least be associated with fewer adverse events than treatment with donepezil or rivastigmine.
| S-EPMC6944712 | biostudies-literature