Faculty Bibliography

In causal inference, a variety of causal effect estimands have been studied, including the sample, uncensored, target, conditional, optimal subpopulation, and optimal weighted average treatment effects. Ad hoc methods have been developed for each estimand based on inverse probability weighting (IPW) and on outcome regression modeling, but these may be sensitive to model misspecification, practical violations of positivity, or both. The contribution of this article is twofold. First, we formulate the generalized average treatment effect (GATE) to unify these causal estimands as well as their IPW estimates. Second, we develop a method based on Kernel optimal matching (KOM) to optimally estimate GATE and to find the GATE most easily estimable by KOM, which we term the Kernel optimal weighted average treatment effect. KOM provides uniform control on the conditional mean squared error of a weighted estimator over a class of models while simultaneously controlling for precision. We study its theoretical properties and evaluate its comparative performance in a simulation study. We illustrate the use of KOM for GATE estimation in two case studies: comparing spine surgical interventions and studying the effect of peer support on people living with HIV.

Inverse probability of treatment weighting (IPTW), which has been used to estimate average treatment effects (ATE) using observational data, tenuously relies on the positivity assumption and the correct specification of the treatment assignment model, both of which are problematic assumptions in many observational studies. Various methods have been proposed to overcome these challenges, including truncation, covariate-balancing propensity scores, and stable balancing weights. Motivated by an observational study in spine surgery, in which positivity is violated and the true treatment assignment model is unknown, we present the use of optimal balancing by kernel optimal matching (KOM) to estimate ATE. By uniformly controlling the conditional mean squared error of a weighted estimator over a class of models, KOM simultaneously mitigates issues of possible misspecification of the treatment assignment model and is able to handle practical violations of the positivity assumption, as shown in our simulation study. Using data from a clinical registry, we apply KOM to compare two spine surgical interventions and demonstrate how the result matches the conclusions of clinical trials that IPTW estimates spuriously refute.

Marginal structural Cox models have been used to estimate the causal effect of a time-varying treatment on a survival outcome in the presence of time-dependent confounders. These methods rely on the positivity assumption, which states that the propensity scores are bounded away from zero and one. Practical violations of this assumption are common in longitudinal studies, resulting in extreme weights that may yield erroneous inferences. Truncation, which consists of replacing outlying weights with less extreme ones, is the most common approach to control for extreme weights to date. While truncation reduces the variability in the weights and the consequent sampling variability of the estimator, it can also introduce bias. Instead of truncated weights, we propose using optimal probability weights, defined as those that have a specified variance and the smallest Euclidean distance from the original, untruncated weights. The set of optimal weights is obtained by solving a constrained quadratic optimization problem. The proposed weights are evaluated in a simulation study and applied to the assessment of the effect of treatment on time to death among people in Sweden who live with human immunodeficiency virus and inject drugs.

Marginal structural models (MSMs) can be used to estimate the causal effect of a potentially time-varying treatment in the presence of time-dependent confounding via weighted regression. The standard approach of using inverse probability of treatment weighting (IPTW) can be sensitive to model misspecification and lead to high-variance estimates due to extreme weights. Various methods have been proposed to partially address this, including covariate balancing propensity score (CBPS) to mitigate treatment model misspecification, and truncation and stabilized-IPTW (sIPTW) to temper extreme weights. In this article, we present kernel optimal weighting (KOW), a convex-optimization-based approach that finds weights for fitting the MSMs that flexibly balance time-dependent confounders while simultaneously penalizing extreme weights, directly addressing the above limitations. We further extend KOW to control for informative censoring. We evaluate the performance of KOW in a simulation study, comparing it with IPTW, sIPTW, and CBPS. We demonstrate the use of KOW in studying the effect of treatment initiation on time-to-death among people living with human immunodeficiency virus and the effect of negative advertising on elections in the United States.

OBJECTIVE:To examine patient characteristics that impact serial observation adherence among vestibular schwannoma (VS) patients. STUDY DESIGN/METHODS:Retrospective chart review. SETTING/METHODS:Single tertiary care center. METHODS:We selected for VS patients from 201 to 2020 who elected for serial observation as initial management. Patients under 18, with previous management, bilateral or intralabyrinthine VS, and neurofibromatosis type 2 were excluded. Demographics, tumor characteristics, and follow-up status were extracted. Single and multiple logistic regression was used to identify patient characteristics impacting follow-up. RESULTS:We identified 507 VS patients who chose serial observation as initial management. Most were female (56.0%), white (73.0%), and married (72.8%). The mean age was 59.3 and most had private insurance (56.4%). Median Charlson Comorbidity Index was 2.00. Mean pure tone audiometry (PTA) average was 41.7 Hz. Average tumor size was 9.04 mm. Of 507 patients, 358 (70.6%) returned for at least one follow-up. On multiple logistic regression analysis, patients with private insurance (odds ratio [OR]: 0.39, confidence interval [CI]: 0.22-0.68; P = .001), racial minority background (OR: 0.54, CI: 0.35-0.83; P = .005), worse PTA averages (OR: 0.99, CI: 0.98-1.00; P = .044), and older age at diagnosis (OR: 0.97, CI: 0.95-1.00; P = .038) were less likely to follow-up. CONCLUSION/CONCLUSIONS:Private health insurance, racial minority background, worse PTA average, and older age were associated with decreased follow-up among adult VS patients electing serial observation. Patients with these characteristics may require additional support to ensure serial observation adherence.

INTRODUCTION/UNASSIGNED:Excessive dialytic potassium (K) and acid removal are risk factors for arrhythmias; however, treatment-to-treatment dialysate modification is rarely performed. We conducted a multicenter, pilot randomized study to test the safety, feasibility, and efficacy of 4 point-of-care (POC) chemistry-guided protocols to adjust dialysate K and bicarbonate (HCO3) in outpatient hemodialysis (HD) clinics. METHODS/UNASSIGNED:Participants received implantable cardiac loop monitors and crossed over to four 4-week periods with adjustment of dialysate K or HCO3 at each treatment according to pre-HD POC values: (i) K-removal minimization, (ii) K-removal maximization, (iii) Acidosis avoidance, and (iv) Alkalosis avoidance. The primary end point was percentage of treatments adhering to the intervention algorithm. Secondary endpoints included pre-HD K and HCO variability, adverse events, and rates of clinically significant arrhythmias (CSAs). RESULTS/UNASSIGNED:Nineteen subjects were enrolled in the study. HD staff completed POC testing and correctly adjusted the dialysate in 604 of 708 (85%) of available HD treatments. There was 1 K ≤3, 29 HCO3 <20 and 2 HCO3 >32 mEq/l and no serious adverse events related to study interventions. Although there were no significant differences between POC results and conventional laboratory measures drawn concurrently, intertreatment K and HCO3 variability was high. There were 45 CSA events; most were transient atrial fibrillation (AF), with numerically fewer events during the alkalosis avoidance period (8) and K-removal maximization period (3) compared to other intervention periods (17). There were no significant differences in CSA duration among interventions. CONCLUSION/UNASSIGNED:Algorithm-guided K/HCO3 adjustment based on POC testing is feasible. The variability of intertreatment K and HCO3 suggests that a POC-laboratory-guided algorithm could markedly alter dialysate-serum chemistry gradients. Definitive end point-powered trials should be considered.

OBJECTIVE:Children with high-frequency severe-to-profound hearing loss and low-frequency residual hearing who do not derive significant benefit from hearing aids are now being considered for cochlear implantation. Previous research shows that hearing preservation is possible and may be desirable for the use of electroacoustic stimulation (EAS) in adults, but this topic remains underexplored in children. The goal of this study was to explore factors relating to hearing preservation, acceptance, and benefits of EAS for children. STUDY DESIGN:Retrospective review. SETTING:Tertiary academic medical center. PATIENTS:Forty children (48 ears) with preoperative low-frequency pure-tone averages of 75 dB HL or less at 250 and 500 Hz (n = 48). INTERVENTION:All patients underwent cochlear implantation with a standard-length electrode. MAIN OUTCOME MEASURE:Low-frequency audiometric thresholds, speech perception, and EAS usage were measured at initial stimulation, and 3 and 12 months postoperatively. Outcomes were compared between children with and without hearing preservation, and between EAS users and nonusers. RESULTS:Hearing was preserved at similar rates as adults but worse for children with an enlarged vestibular aqueduct. Fewer than half of children who qualified to use EAS chose to do so, citing a variety of audiologic and nonaudiologic reasons. No differences were detected in speech perception scores across the groups for words, sentences, or sentences in noise tests. CONCLUSIONS:Neither hearing preservation nor EAS use resulted in superior speech perception in children with preoperative residual hearing; rather, all children performed well after implantation.

Objective: To investigate the impact of adjuvant radiotherapy in isolated locally advanced oral cavity cancers (pT3N0M0) without adverse features. Methods: We selected all patients from the National Cancer Database (2004"“2019) who underwent surgical treatment where the final pathology was T3N0M0 with negative margins. Demographics, details of treatment, and outcomes were abstracted. The impact of radiotherapy on survival was assessed with univariable, multivariable, and propensity score-matched analyses. Results: We identified 571 patients in our survival cohort. Most were male (348, 60.9%), and median age was 65. Less than one-third (176, 30.8%) received adjuvant radiotherapy. The median length of follow-up was 29 months. Overall, adjuvant radiotherapy was associated with improved survival (87.2% vs. 77.7%, at 2 years, p < 0.01). On multivariable analysis controlling for age and comorbidities, this survival difference persisted (HR: 0.62, 95% CI: 0.43"“0.90, p = 0.01). In a propensity score-matched population of 278 patients matched on age and comorbidities, adjuvant radiotherapy was still associated with longer survival (87.4% vs. 78.5%, p = 0.014). Conclusion: In our study, adjuvant radiotherapy was associated with improved survival in completely excised locally advanced oral cavity tumors (T3N0M0). However, a significant proportion of patients do not receive adjuvant radiotherapy. These findings highlight the need for continued efforts to promote guideline-recommended care. Level of Evidence: Level 3 Laryngoscope, 2023.