Martingale residual-based method to control for confounders measured only in a validation sample in time-to-event analysis

Determining the association of insurance coverage and survival outcomes in patients with olfactory neuroblastoma utilizing machine learning

KEY POINTS

We use machine learning to examine health insurance and mortality in olfactory neuroblastoma. Private insurance significantly improved survival even after adjusting for confounders. The regression model also found no statistical difference between Medicare and no insurance.

Clinical Analysis and Prognostic Prediction Model for Patients with Uterine Leiomyosarcoma at FIGO Stage I

Purpose

To reveal the clinical status and construct a predictive prognostic model for patients with uterine leiomyosarcoma (uLMS) at International Federation of Gynecology and Obstetrics (FIGO) stage I.

Patients and Methods

The medical records of patients with stage I uLMS during the study period were retrospectively reviewed. Multiple imputation, Martingale residuals and restricted cubic spline were used for data processing. Univariate and multivariate analyses were used to determine independent prognostic factors. The Schoenfeld individual test was used to verify the proportional hazards (PH) assumption. The predictive ability of the nomogram was validated internally.

Results

Ultimately, 102 patients were included. The median age at diagnosis was 51 years old. During the medium follow-up time of 68 months, 55 (53.9%) patients developed recurrence. The median recurrence interval was 32 months. The most common metastatic site was the lung (27 cases). Eventually, 38 (37.3%) patients died of uLMS. The 3-year and 5-year overall survival rates were 66.0% and 52.0%, respectively. Age at diagnosis >49 years, larger tumor size, MI>10/10HPF, presence of LVSI and Ki-67 labeling index (LI) >25% (P=0.0467, 0.0077, 0.0475, 0.0294, and 0.0427, respectively) were independent prognostic factors. The PH assumption remained inviolate. The concordance index was 0.847, the area under the time-dependent receiver operating characteristic curve surpassed 0.7, and the calibration curve showed gratifying consistency.

Conclusion

Age at diagnosis, tumor size, MI, LVSI, and Ki-67 LI were identified as independent prognostic factors for stage I uLMS. This prognostic nomogram would provide personalized assessment with superior predictive performance.

Genomic context and TP53 allele frequency define clinical outcomes in TP53-mutated myelodysplastic syndromes

TP53 mutations are associated with adverse outcomes and shorter response to hypomethylating agents (HMAs) in myelodysplastic syndrome (MDS). Limited data have evaluated the impact of the type, number, and patterns of TP53 mutations in response outcomes and prognosis of MDS. We evaluated the clinicopathologic characteristics, outcomes, and response to therapy of 261 patients with MDS and TP53 mutations. Median age was 68 years (range, 18-80 years). A total of 217 patients (83%) had a complex karyotype. TP53 mutations were detected at a median variant allele frequency (VAF) of 0.39 (range, 0.01-0.94). TP53 deletion was associated with lower overall response rate (ORR) (odds ratio, 0.3; P = .021), and lower TP53 VAF correlated with higher ORR to HMAs. Increase in TP53 VAF at the time of transformation was observed in 13 patients (61%), and previously undetectable mutations were observed in 15 patients (65%). TP53 VAF was associated with worse prognosis (hazard ratio, 1.02 per 1% VAF increase; 95% confidence interval, 1.01-1.03; P < .001). Integration of TP53 VAF and karyotypic complexity identified prognostic subgroups within TP53-mutant MDS. We developed a multivariable model for overall survival that included the revised International Prognostic Scoring System (IPSS-R) categories and TP53 VAF. Total score for each patient was calculated as follows: VAF TP53 + 13 × IPSS-R blast score + 16 × IPSS-R cytogenetic score + 28 × IPSS-R hemoglobin score + 46 × IPSS-R platelet score. Use of this model identified 4 prognostic subgroups with median survival times of not reached, 42.2, 21.9, and 9.2 months. These data suggest that outcomes of patients with TP53-mutated MDS are heterogeneous and that transformation may be driven not only by TP53 but also by other factors.

Two-phase analysis and study design for survival models with error-prone exposures

Increasingly, medical research is dependent on data collected for non-research purposes, such as electronic health records data. Health records data and other large databases can be prone to measurement error in key exposures, and unadjusted analyses of error-prone data can bias study results. Validating a subset of records is a cost-effective way of gaining information on the error structure, which in turn can be used to adjust analyses for this error and improve inference. We extend the mean score method for the two-phase analysis of discrete-time survival models, which uses the unvalidated covariates as auxiliary variables that act as surrogates for the unobserved true exposures. This method relies on a two-phase sampling design and an estimation approach that preserves the consistency of complete case regression parameter estimates in the validated subset, with increased precision leveraged from the auxiliary data. Furthermore, we develop optimal sampling strategies which minimize the variance of the mean score estimator for a target exposure under a fixed cost constraint. We consider the setting where an internal pilot is necessary for the optimal design so that the phase two sample is split into a pilot and an adaptive optimal sample. Through simulations and data example, we evaluate efficiency gains of the mean score estimator using the derived optimal validation design compared to balanced and simple random sampling for the phase two sample. We also empirically explore efficiency gains that the proposed discrete optimal design can provide for the Cox proportional hazards model in the setting of a continuous-time survival outcome.

Assessing the impact of unmeasured confounders for credible and reliable real-world evidence

PURPOSE

We review statistical methods for assessing the possible impact of bias due to unmeasured confounding in real world data analysis and provide detailed recommendations for choosing among the methods.

METHODS

By updating an earlier systematic review, we summarize modern statistical best practices for evaluating and correcting for potential bias due to unmeasured confounding in estimating causal treatment effect from non-interventional studies.

RESULTS

We suggest a hierarchical structure for assessing unmeasured confounding. First, for initial sensitivity analyses, we strongly recommend applying a recently developed method, the E-value, that is straightforward to apply and does not require prior knowledge or assumptions about the unmeasured confounder(s). When some such knowledge is available, the E-value could be supplemented by the rule-out or array method at this step. If these initial analyses suggest results may not be robust to unmeasured confounding, subsequent analyses could be conducted using more specialized statistical methods, which we categorize based on whether they require access to external data on the suspected unmeasured confounder(s), internal data, or no data. Other factors for choosing the subsequent sensitivity analysis methods are also introduced and discussed, including the types of unmeasured confounders and whether the subsequent sensitivity analysis is intended to provide a corrected causal treatment effect.

CONCLUSION

Various analytical methods have been proposed to address unmeasured confounding, but little research has discussed a structured approach to select appropriate methods in practice. In providing practical suggestions for choosing appropriate initial and, potentially, more specialized subsequent sensitivity analyses, we hope to facilitate the widespread reporting of such sensitivity analyses in non-interventional studies. The suggested approach also has the potential to inform pre-specification of sensitivity analyses before executing the analysis, and therefore increase the transparency and limit selective study reporting.

Imputing missing time-dependent covariate values for the discrete time Cox model

We describe a procedure for imputing missing values of time-dependent covariates in a discrete time Cox model using the chained equations method. The procedure multiply imputes the missing values for each time-period in a time-sequential manner, using covariates from the current and previous time-periods as well as the survival outcome. The form of the outcome variable used in the imputation model depends on the functional form of the time-dependent covariate(s) and differs from the case of Cox regression with only baseline covariates. This time-sequential approach provides an approximation to a fully conditional approach. We illustrate the procedure with data on diabetics, evaluating the association of their glucose control with the risk of selected cancers. Using simulations we show that the suggested estimator performed well (in terms of bias and coverage) for completely missing at random, missing at random and moderate non-missing-at-random patterns. However, for very strong non-missing-at-random patterns, the estimator was seriously biased and the coverage was too low. The procedure can be implemented using multiple imputation with the Fully conditional Specification (FCS) method (MI procedure in SAS with FCS statement or similar packages in other software, e.g. MICE in R). For use with event times on a continuous scale, the events would need to be grouped into time-intervals.

Asthma is not associated with the need for surgery in Crohn's disease when controlling for smoking status: a population-based cohort study

Purpose

Growing evidence suggests asthma and Crohn’s disease commonly cooccur. However, the impact of asthma on the prognosis of Crohn’s disease is unknown. The aim of our study was to assess the effect of asthma on the need for intestinal resection in patients with Crohn’s disease while adjusting for smoking status, imputed from a smaller, secondary data set.

Patients and methods

Using health administrative data from a universally funded healthcare plan in Alberta, Canada, we conducted a cohort study to assess the effect of asthma on the need for surgery in patients with Crohn’s disease diagnosed between 2002 and 2008 (N=2,113). Validated algorithms were used to identify incident cases of Crohn’s disease, cooccurring asthma, and intestinal resection. The association between asthma and intestinal resection was estimated using multivariable Cox proportional hazards regression. Smoking status was imputed using a novel method using martingale residuals, derived from a data set of 485 patients enrolled in the Alberta Inflammatory Bowel Disease Consortium (2007 to 2014) who completed environmental questionnaires. All analyses were adjusted for age, sex, rural/urban status, and mean neighborhood income quintile.

Results

Asthma did not increase the risk of surgery in the health administrative data when not adjusting for smoking status (HR 1.03, 95% CI 0.81 to 1.29). The association remained nonsignificant after imputing smoking status in the health administrative data (HR 1.03, 95% CI 0.81 to 1.29).

Conclusion

Although asthma is associated with an increased risk of Crohn’s disease, co-occurring asthma is not associated with the risk of surgery in these patients. This null association persisted after adjusting for smoking status. This study described a novel method to adjust for confounding (smoking status) in time-to-event analyses, even when the confounding variable is unmeasured in health administrative data.

Propensity Scores in Pharmacoepidemiology: Beyond the Horizon

Purpose of review

Propensity score methods have become commonplace in pharmacoepidemiology over the past decade. Their adoption has confronted formidable obstacles that arise from pharmacoepidemiology's reliance on large healthcare databases of considerable heterogeneity and complexity. These include identifying clinically meaningful samples, defining treatment comparisons, and measuring covariates in ways that respect sound epidemiologic study design. Additional complexities involve correctly modeling treatment decisions in the face of variation in healthcare practice, and dealing with missing information and unmeasured confounding. In this review, we examine the application of propensity score methods in pharmacoepidemiology with particular attention to these and other issues, with an eye towards standards of practice, recent methodological advances, and opportunities for future progress.

Recent findings

Propensity score methods have matured in ways that can advance comparative effectiveness and safety research in pharmacoepidemiology. These include natural extensions for categorical treatments, matching algorithms that can optimize sample size given design constraints, weighting estimators that asymptotically target matched and overlap samples, and the incorporation of machine learning to aid in covariate selection and model building.

Summary

These recent and encouraging advances should be further evaluated through simulation and empirical studies, but nonetheless represent a bright path ahead for the observational study of treatment benefits and harms.

Adjustment for time-dependent unmeasured confounders in marginal structural Cox models using validation sample data

Large databases used in observational studies of drug safety often lack information on important confounders. The resulting unmeasured confounding bias may be avoided by using additional confounder information, frequently available in smaller clinical “validation samples”. Yet, no existing method that uses such validation samples is able to deal with unmeasured time-varying variables acting as both confounders and possible mediators of the treatment effect. We propose and compare alternative methods which control for confounders measured only in a validation sample within marginal structural Cox models. Each method corrects the time-varying inverse probability of treatment weights for all subject-by-time observations using either regression calibration of the propensity score, or multiple imputation of unmeasured confounders. Two proposed methods rely on martingale residuals from a Cox model that includes only confounders fully measured in the large database, to correct inverse probability of treatment weight for imputed values of unmeasured confounders. Simulation demonstrates that martingale residual-based methods systematically reduce confounding bias over naïve methods, with multiple imputation including the martingale residual yielding, on average, the best overall accuracy. We apply martingale residual-based imputation to re-assess the potential risk of drug-induced hypoglycemia in diabetic patients, where an important laboratory test is repeatedly measured only in a small sub-cohort.