Pattern-mixture Models Research Articles

Direct Estimation for Commonly Used Pattern-Mixture Models in Clinical Trials

Pattern-mixture models have received increasing attention as they are commonly used to assess treatment effects in primary or sensitivity analyses for clinical trials with nonignorable missing data. Pattern-mixture models have traditionally been implemented using multiple imputation, where the variance estimation may be a challenge because the Rubin’s approach of combining between- and within-imputation variance may not provide consistent variance estimation while bootstrap methods may be time-consuming. Direct likelihood-based approaches have been proposed in the literature and implemented for some pattern-mixture models, but the assumptions are sometimes restrictive and the theoretical framework is fragile. In this article, we propose an analytical framework for an efficient direct likelihood estimation method for commonly used pattern-mixture models corresponding to return-to-baseline, jump-to-reference, placebo washout, and retrieved dropout imputations. A parsimonious tipping point analysis is also discussed and implemented. Results from simulation studies demonstrate that the proposed methods provide consistent estimators. We further illustrate the utility of the proposed methods using data from a clinical trial evaluating a treatment for type 2 diabetes.

Possible Data-Generating Models of Longitudinal Continuous Outcomes and Intercurrent Events to Investigate Estimands

We aim to explore data-generating models to jointly simulate outcomes and intercurrent events for randomized clinical trials to enable the investigation of estimands. We develop four possible data-generating models for the joint distribution of longitudinal continuous clinical outcomes and intercurrent events under the scenario where they are observable: a selection model, a pattern-mixture mixed model, a shared-parameter model, and a joint model of longitudinally observed outcomes and a survival model for intercurrent events. We present a case study in a short-term depression trial with repeated measurements of continuous outcomes and two types of intercurrent events, and evaluate the potential and challenges of such data-generating models. Simulating randomized trials with outcomes and intercurrent events is a complex undertaking. We found that the four possible data-generating models can simulate different types of intercurrent events and associated longitudinal outcomes. They can be used to emulate envisaged patterns of intercurrent events and outcomes informed by prior available trial data or expectations. Model and parameter choice for a given application require further development. The four possible data-generating models could be used to investigate different estimands and their properties in-depth in the design stage. Thereby they are useful tools for the selection of estimands a priori.

Not all types of depressed patients who persist with their antidepressant treatment improve in side effect complaints: A comparison of treatment completers and dropouts in the STAR*D trial.

There is a "traditional belief" that antidepressant side effect complaints improve with medication persistence; however, support for this theory has remained inconclusive. We aimed to examine if side effect complaints improved over time by modeling the relationship between side effect complaints and time at dropout for patients receiving citalopram during the first level of acute treatment in the Sequenced Treatment Alternatives to Relieve Depression (STAR*D) trial. We categorized the 2833 patients into five patterns by week of dropout. We used pattern-mixture modeling to model change in side effect complaints (frequency, intensity, and burden) over the 12-week course of treatment, while accounting for attrition and depressive severity. Using post-hoc linear contrasts, we compared the attrition patterns with the completers' pattern for severity of side effect complaints at each respective last visit prior to dropout as well as averaged side effect complaints across the duration of treatment. We also reported frequencies and tolerability of side effects for nine organ/function systems over the course of treatment. Patients who dropped out early exhibited worsening side effect burden and patients who dropped out later showed improvements in side effect frequency and intensity. Treatment completers improved in all side effect complaints over the course of treatment. Early attrition patterns had more severe side effect complaints for both tests of post-hoc linear contrasts than later attrition patterns and completers. Side effect complaints from antidepressant treatment improve over time, but only for some types of patients. As a precaution for early dropout, clinicians should monitor patients who exhibit worsening and more severe side effect complaints-especially in the first 6 weeks of antidepressant treatment. In addition, clinicians may want to consider changing the type of treatment early on for these patients, rather than encouraging them to persist with their current medication.

Cystatin C vs creatinine eGFR in advanced CKD: an analysis of the STOP-ACEi trial

ABSTRACT Background In this secondary analysis of the STOP-ACEi trial, we explored the impact of discontinuing or continuing renin–angiotensin system inhibitor therapy in people with advanced chronic kidney disease on cystatin C estimated glomerular filtration rate (eGFR). Methods Cystatin C eGFRs were calculated at baseline, 12, 24 and 36 months using Chronic Kidney Disease Epidemiology Collaboration (CKD-EPI) cystatin 2012, European Kidney Function Consortium and CKD-EPI Combined 2021 equations. We excluded samples obtained after the initiation of kidney replacement therapy. Primary analysis used complete case analysis and mixed-effects linear regression model, adjusting for minimization variables, baseline value, time-point and treatment by time interaction. Sensitivity analysis was conducted using a pattern mixture model to account for missing data that was not at random. To model the longitudinal cystatin C data with time-to-event data, a joint model was utilized which incorporated the cystatin C measurements at various time points and accounted for the occurrence of kidney replacement therapy. Results The mean cystatin C eGFRs (CKD-EPI 2012) at baseline were 17.8 mg/L [standard deviation (SD 6.3)] and 17.9 mL/min/1.73 m2 (SD 6.3) in the STOP and CONTINUE arms, respectively. The estimated least squares mean difference at 12 months between STOP and CONTINUE arm was –1.46 [95% confidence interval (CI) –2.39 to –0.52, P = .002]. The estimated least squares mean difference at 24 months was –2.27 (95% CI –3.48 to –1.06, P < .001). The estimated least squares mean difference at 36 months was –1.72 (95% CI –3.48 to 0.03, P = .05). Conclusion Our results are consistent with the primary study's analysis and sensitivity analyses support these findings and provide additional insights. Our findings demonstrate the similarity of creatinine and cystatin eGFR results and therefore support the use of cystatin C as an alternative marker of eGFR in advanced CKD, particularly in those in whom creatinine is likely to be less accurate.

Trajectories of depressive symptoms and subjective well-being before and after the onset of the COVID-19 pandemic: Two six-year longitudinal studies

Previous research investigated the trajectories of mental health and well-being during and after the onset of the COVID-19 pandemic. However, less is known about the trajectories of mental health and well-being before, during, and two years after the onset of the pandemic. The aim of the current study was to investigate the trajectory of depression symptoms and subjective well-being (i.e., life satisfaction and positive and negative affect) trajectories over six time points (2017–2022), three before the pandemic and three after the onset of the pandemic. To increase the robustness of our overall conclusions and avoid reliance on data from only one country, we used data from two nationwide representative longitudinal surveys conducted in Germany (GESIS Panel study; N = 5184) and Switzerland (Swiss Household Panel study; N = 17,074). Using covariance pattern mixture models, the results revealed that a four-class model best fit the data. The Stable/resilient trajectory was the most common across outcomes (74.2%–90.1% of participants). Three additional trajectories of Chronic/Low, Upright U-shaped, and Inverted U-shaped emerged in the analysis of negative affect and depression symptoms, while distinct trajectory classes of Worsening, Improving/Stable, and Upright U-shaped also emerged for analyses of positive affect and life satisfaction shaped. In conclusion, there was no evidence of a long-term impact of the pandemic for the vast majority of participants (about 90%). For the remaining participants, the COVID-19 pandemic (along with its exceptional circumstances) was a turning point or a catalyst that reversed, accelerated, or flattened a pre-pandemic trend. These changes in trends were not only negative (e.g., greater depression symptoms), but also positive (e.g., less depression symptoms).

Protective Assets Reinforced With Integrated Care and Technology (PARITY): Protocol for a Randomized Controlled Trial.

Black women are significantly more likely to experience severe maternal morbidity and are 3 times as likely to die from pregnancy-related causes compared to White women. Using a strengths-based wellness approach within an integrated supportive care program provided by a community doula could offer pragmatic solutions for Black maternal disparities. The Protective Assets Reinforced with Integrated Care and Technology (PARITY) program consists of a wellness technology platform, including informational links to wellness content and reinforcing motivational SMS text messages, as well as community-based doula support delivered both in person and through the technology platform to improve Black maternal wellness. This pilot randomized controlled trial (RCT) and mixed methods evaluation aims to (1) determine the feasibility and acceptability of the PARITY intervention; (2) investigate the preliminary efficacy of the PARITY intervention on clinical outcomes (maternal blood pressure, gestational weight gain, and cesarean birth); and (3) investigate changes to wellness behavioral outcomes (nutrition, physical activity, sleep, and health care adherence) and empowered strengths (self-efficacy, social support, motivation, resilience, problem-solving, and self-regulation) in the intervention group compared to a control group. A 2-arm RCT and mixed methods evaluation will be conducted. Overall, 60 Black pregnant individuals will be randomized in a ratio of 1:1 to either the intervention or informational control group. Participants in the intervention group will receive access to the technology platform over a 12-week period that ends before birth. Intervention participants will be assigned a doula interventionist, who will meet with them 4 times during the intervention. All participants (intervention and control) will receive a referral for a birth doula at no cost, printed materials about having a healthy pregnancy, and community resources. Feasibility and acceptability will be assessed at the end of the program. Measures will be obtained at baseline (20-28 weeks), the 36th week of pregnancy, birth, and 6-12 weeks post partum. Summary statistics and distribution plots will be used to describe measured variables at each time point. A generalized linear mixed model with a shared random component will be used to analyze the effects of PARITY on clinical, wellness behavioral, and empowered strength outcomes, including baseline nutrition, physical activity, and sleep measures as covariates. For significant effects, post hoc contrasts will be adjusted using the Holm method to maintain comparison-wise error at or <.05. Missing data will be addressed using a pattern-mixture model. The National Institute of Nursing Research funded this pilot RCT. Recruitment, enrollment, and data collection are ongoing, and the estimated study completion date is October 2024. The expected results of this study will provide the feasibility and preliminary efficacy of the PARITY intervention, to be used in a larger trial with a 12-month PARITY program intervention. ClinicalTrials.gov NCT05802615; https://clinicaltrials.gov/study/NCT05802615. DERR1-10.2196/58580.

Interdisciplinary pain rehabilitation for patients with Ehlers-Danlos syndrome and hypermobility spectrum disorders.

Chronic pain is a common manifestation of Ehlers-Danlos syndrome and hypermobility spectrum disorders; thus it is often suggested that patients undergo generic interdisciplinary pain rehabilitation, despite there being little evidence to support this decision. The aim of this study is to examine the effectiveness of standard rehabilitation programmes for chronic pain on patients with Ehlers-Danlos syndrome and hypermobility spectrum disorders, compared with patients with other chronic pain disorders. Data, collected between 2008 and 2016, were extracted from a Swedish national registry. The patient data comprised of 406 cases with Ehlers-Danlos syndrome or hypermobility spectrum disorders, 784 cases with a whiplash-related diagnosis, 3713 cases with diagnoses relating to spinal pain, and 2880 cases of fibromyalgia. The differences between groups on key outcome measures from pre- to 1-year follow-up after interdisciplinary pain rehabilitation were analysed using linear mixed effects models. Sensitivity analysis in the form of pattern-mixture modelling was conducted to discern the impact of missing data. No significant differences were found in improvements from pre- to 1-year follow-up for patients with Ehlers-Danlos syndrome or hypermobility spectrum disorder compared with other diagnostic groups regarding measures of health-related quality of life, mental health, or fatigue. At follow-up, differences in pain interference (d = -0.34 (95% confidence interval [95% CI] -0.5 to -0.18)), average pain (d = 0.22 (95% CI 0.11-0.62)) and physical functioning (d = 2.19 (95% CI 1.61-2.77)) were detected for the group with spinal-related diagnoses in relation to those with EDS/HSD, largely due to pre-treatment group differences. Sensitivity analysis found little evidence for missing data influencing the results. This study suggests that patients with Ehlers-Danlos syndrome/hypermobility spectrum disorders may benefit from inclusion in an interdisciplinary pain rehabilitation programme.

Using proxy pattern-mixture models to explain bias in estimates of COVID-19 vaccine uptake from two large surveys

Abstract Recently, attention was drawn to the failure of two very large internet-based probability surveys to correctly estimate COVID-19 vaccine uptake in the U.S. in early 2021. Both the Delphi-Facebook COVID-19 Trends and Impact Survey (CTIS) and Census Household Pulse Survey (HPS) overestimated uptake substantially, by 17 and 14 percentage points in May 2021, respectively. These surveys had large numbers of respondents but very low response rates (&amp;lt;10%), thus, nonignorable nonresponse could have had substantial impact. Specifically, it is plausible that ‘anti-vaccine’ individuals were less likely to participate given the topic (impact of the pandemic on daily life). In this article, we use proxy pattern-mixture models (PPMMs) to estimate the proportion of adults (18 +) who received at least one dose of a COVID-19 vaccine, using data from the CTIS and HPS, under a nonignorable nonresponse assumption. Data from the American Community Survey provide the necessary population data for the PPMMs. We compare these estimates to the true benchmark uptake numbers and show that the PPMM could have detected the direction of the bias and provide meaningful bias bounds. We also use the PPMM to estimate vaccine hesitancy, a measure for which we do not have a benchmark truth, and compare to the direct survey estimates.

Quantifying bias due to missing data in quality of life surveys of advanced-stage cancer patients.

Many studies on cancer patients investigate the impact of treatment on health-related quality of life (QoL). Typically, QoL is measured longitudinally, at baseline and at predefined timepoints thereafter. The question is whether, at a given timepoint, patients who return their questionnaire (available cases, AC) have a different QoL than those who do not return their questionnaire (non-AC). We employed augmented inverse probability weighting (AIPW) to estimate the average QoL of non-AC in two studies on advanced-stage cancer patients. The AIPW estimator assumed data to be missing at random (MAR) and used machine learning (ML)-based methods to estimate answering probabilities of individuals at given timepoints as well as their reported QoL, as a function of auxiliary variables. These auxiliary variables were selected by medical oncologists based on domain expertise. We aggregated results both by timepoint and by time until death and compared AIPW estimates to the AC averages. Additionally, we used a pattern mixture model (PMM) to check sensitivity of our AIPW estimates against violation of the MAR assumption. Our study included 1927 patients with advanced pancreatic and 797 patients with advanced breast cancer. The AIPW estimate for average QoL of non-AC was below the average QoL of AC when aggregated by timepoint. The difference vanished when aggregated by time until death. PMM estimates were below AIPW estimates. Our results indicate that non-AC have a lower average QoL than AC. However, estimates for QoL of non-AC are subject to unverifiable assumptions about the missingness mechanism.

Addressing non-response data for standardized post-acute functional items

BackgroundThe post-acute patient standardized functional items (Section GG) include non-response options such as refuse, not attempt and not applicable. We examined non-response patterns and compared four methods to address non-response functional data in Section GG at nation-wide inpatient rehabilitation facilities (IRF).MethodsWe characterized non-response patterns using 100% Medicare 2018 data. We applied four methods to generate imputed values for each non-response functional item of each patient: Monte Carlo Markov Chains multiple imputations (MCMC), Fully Conditional Specification multiple imputations (FCS), Pattern-mixture model (PMM) multiple imputations and the Centers for Medicare and Medicaid Services (CMS) approach. We compared changes of Spearman correlations and weighted kappa between Section GG and the site-specific functional items across impairments before and after applying four methods.ResultsOne hundred fifty-nine thousand six hundred ninety-one Medicare fee-for-services beneficiaries admitted to IRFs with stroke, brain dysfunction, neurologic condition, orthopedic disorders, and debility. At discharge, 3.9% (self-care) and 61.6% (mobility) of IRF patients had at least one non-response answer in Section GG. Patients tended to have non-response data due to refused at discharge than at admission. Patients with non-response data tended to have worse function, especially in mobility; also improved less functionally compared to patients without non-response data. Overall, patients coded as ‘refused’ were more functionally independent in self-care and patients coded as ‘not applicable’ were more functionally independent in transfer and mobility, compared to other non-response answers. Four methods showed similar changes in correlations and agreements between Section GG and the site-specific functional items, but variations exist across impairments between multiple imputations and the CMS approach.ConclusionsThe different reasons for non-response answers are correlated with varied functional status. The high proportion of patients with non-response data for mobility items raised a concern of biased IRF quality reporting. Our findings have potential implications for improving patient care, outcomes, quality reporting, and payment across post-acute settings.

Using Auxiliary Marginal Distributions in Imputations for Nonresponse while Accounting for Survey Weights, with Application to Estimating Voter Turnout

Abstract In many survey settings, population counts or percentages are available for some of the variables in the survey, for example, from censuses, administrative databases, or other high-quality surveys. We present a model-based approach to utilize such auxiliary marginal distributions in multiple imputation for unit and item nonresponse in complex surveys. In doing so, we ensure that the imputations produce design-based estimates that are plausible given the known margins. We introduce and utilize a hybrid missingness model comprising a pattern mixture model for unit nonresponse and selection models for item nonresponse. We also develop a computational strategy for estimating the parameters of and generating imputations with hybrid missingness models. We apply a hybrid missingness model to examine voter turnout by subgroups using the 2018 Current Population Survey for North Carolina. The hybrid missingness model also facilitates modeling measurement errors simultaneously with handling missing values. We illustrate this feature with the voter turnout application by examining how results change when we allow for overreporting, that is, individuals self-reporting that they voted when in fact they did not.

Simple Bayesian models for missing binary outcomes in randomized controlled trials.

Missing outcomes are commonly encountered in randomized controlled trials (RCT) involving human subjects and present a risk for substantial bias in the results of a complete case analysis. While response rates for RCTs are typically high there is no agreed upon universal threshold under which the amount of missing data is deemed to not be a threat to inference. We focus here on binary outcomes that are possibly missing not at random, that is, the value of the outcome influences its possibility of being observed. Salient information that can assist in addressing these missing outcomes in such situations is the anticipated response rate in each study arm; these can often be anticipated based on prior research in similar populations using similar designs and outcomes. Further, in some areas of human subjects research, we are often confident or we suspect that response rates among RCT participants with successful treatment outcomes will be at least as great as those among participants without successful treatment outcomes. In other settings we may suspect the opposite relationship. This direction of the differential response between those with successful and unsuccessful outcomes can further aid in addressing the missing outcomes. We present simple Bayesian pattern-mixture models that incorporate this information on response rates to analyze the relationship between a binary outcome and an intervention while addressing the missing outcomes. We assess the performance of this method in simulation studies and apply this method to the results of an RCT of a smoking abstinence intervention.

Dealing with censoring in a network meta-analysis of time-to-event data

Background The Health Technology Assessment agencies typically require an economic evaluation considering a lifetime horizon for interventions affecting survival. However, survival data are often censored and are typically analyzed assuming the censoring mechanism independent of the event process. This assumption may lead to biased results when the censoring mechanism is informative. Methods We propose a flexible approach to jointly model the participants experiencing an event and censored participants by incorporating the pattern-mixture (PM) model in the fractional polynomial (FP) model within the network meta-analysis (NMA) framework. We introduce the informative censoring hazard ratio parameter that quantifies the departure from the censored at random assumption. The FP-PM model is exemplified in an NMA of the overall survival from non-small cell lung carcinoma studies using Bayesian methods. Results The results on hazard ratio and survival from the FP-PM model are similar to those from the FP model. However, the posterior standard deviation of the hazard ratio is slightly greater when censored data are modeled because the uncertainty induced by censoring is naturally accounted for in the FP-PM model. The between-study standard deviation is almost identical in both models due to the low censoring rate across the studies. At the end of the corresponding studies, the informative censoring hazard ratio demonstrated a possible departure from the censored at random assumption for gefitinib and best supportive care. Conclusions The proposed method offers a comprehensive sensitivity analysis framework to examine the robustness of the NMA results to clinically plausible censoring scenarios.

Predictors and effects of toxicity experienced by older adults with cancer receiving systemic therapy in a randomized clinical trial of geriatric assessment

IntroductionOlder adults represent a large segment of the oncology population, however, they remain underrepresented in clinical research. Treatment of older adults is often extrapolated using data from younger and fitter patients, which may not be appropriate. Furthermore the implications of toxicity from treatment can be greater for this population. Predicting toxicity from treatment and its effect on quality of life and functional status for older adults therefore is important. Materials and MethodsWe analyzed data from a clinical trial of geriatric assessment and management for Canadian elders with cancer (5C study). We assessed whether the baseline Cancer and Aging Research Group (CARG) toxicity score, G8 score, and Eastern Cooperative Oncology Group (ECOG) performance predicted grade 3–5 toxicity using logistic regression and pattern mixture models. We also assessed the impact of toxicity on quality of life and functional decline. Patients were followed for six months. ResultsThree hundred sixteen patients were included. Mean age was 76 years old and 40% of patients were female. One hundred nineteen patients (38%) experienced at least one grade 3–5 toxicity. Neither the CARG toxicity score, G8, or ECOG were predictive of grade 3–5 toxicity. Patients who experienced grade 3–5 toxicity were more likely to have functional impairments over time (odds ratio 3.71, p = 0.03). However, they maintained their quality of life. DiscussionIn this secondary analysis of a randomized controlled trial of geriatric assessment and management we did not find any predictors of grade 3–5 toxicity. Patients who did experience toxicity were more likely to report functional decline over time. Older adults who do experience treatment related toxicity may benefit from increased supports.Clinical trial information: NCT0315467.

Bayesian pattern-mixture models for dropout and intermittently missing data in longitudinal data analysis.

Valid inference can be drawn from a random-effects model for repeated measures that are incomplete if whether the data are missing or not, known as missingness, is independent of the missing data. Data that are missing completely at random or missing at random are two data types for which missingness is ignorable. Given ignorable missingness, statistical inference can proceed without addressing the source of the missing data in the model. If the missingness is not ignorable, however, recommendations are to fit multiple models that represent different plausible explanations of the missing data. A popular choice in methods for evaluating nonignorable missingness is a random-effects pattern-mixture model that extends a random-effects model to include one or more between-subjects variables that represent fixed patterns of missing data. Generally straightforward to implement, a fixed pattern-mixture model is one among several options for assessing nonignorable missingness, and when it is used as the sole model to address nonignorable missingness, understanding the impact of missingness is greatly limited. This paper considers alternatives to a fixed pattern-mixture model for nonignorable missingness that are generally straightforward to fit and encourage researchers to give greater attention to the possible impact of nonignorable missingness in longitudinal data analysis. Patterns of both monotonic and non-monotonic (intermittently) missing data are addressed. Empirical longitudinal psychiatric data are used to illustrate the models. A small Monte Carlo data simulation study is presented to help illustrate the utility of such methods.

Exploration of model misspecification in latent class methods for longitudinal data: Correlation structure matters.

Modeling longitudinal trajectories and identifying latent classes of trajectories is of great interest in biomedical research, and software to identify latent classes of such is readily available for latent class trajectory analysis (LCTA), growth mixture modeling (GMM) and covariance pattern mixture models (CPMM). In biomedical applications, the level of within-person correlation is often non-negligible, which can impact the model choice and interpretation. LCTA does not incorporate this correlation. GMM does so through random effects, while CPMM specifies a model for within-class marginal covariance matrix. Previous work has investigated the impact of constraining covariance structures, both within and across classes, in GMMs-an approach often used to solve convergence problems. Using simulation, we focused specifically on how misspecification of the temporal correlation structure and strength, but correct variances, impacts class enumeration and parameter estimation under LCTA and CPMM. We found (1) even in the presence of weak correlation, LCTA often does not reproduce original classes, (2) CPMM performs well in class enumeration when the correct correlation structure is selected, and (3) regardless of misspecification of the correlation structure, both LCTA and CPMM give unbiased estimates of the class trajectory parameters when the within-individual correlation is weak and the number of classes is correctly specified. However, the bias increases markedly when the correlation is moderate for LCTA and when the incorrect correlation structure is used for CPMM. This work highlights the importance of correlation alone in obtaining appropriate model interpretations and provides insight into model choice.

A pattern mixture model with long short-term memory network for acute kidney injury prediction

Acute kidney disease is a serious complication characterized by poor short- and long-term outcomes in the intensive care unit. Impairment in renal function of the kidney significantly increases the mortality rate. Early detection of acute kidney disease could lead to preventive interventions, therefore deep learning systems can detect it before its symptoms and consequences appear. We developed a novel deep learning architecture like Stacked long short-term memory network with pattern mixture approach for kidney injury prediction. A total of 33,754 patients encountered were retrospectively analyzed from the MIMIC-III database. A selection and pattern mixture model was used for preprocessing the time-series data. We compared the proposed result with conventional algorithms like gradient boosted trees and long short-term memory model. Our model was trained on patient time-series data for different time windows and obtained the highest accuracy of 92.4% for 12 h and 92.6% for 24 h. A novel stacked long short-term memory model outperforms the machine learning model, revealing superior performance in predicting kidney injury 24 h before onset.

Shenlingcao oral liquid for patients with non-small cell lung cancer receiving adjuvant chemotherapy after radical resection: A multicenter randomized controlled trial

BackgroundLow quality of life (QoL) in patients with non-small cell lung cancer (NSCLC) receiving adjuvant chemotherapy after radical resection is a major global health issue. High-quality evidence for the effectiveness of Shenlingcao oral liquid (SOL) as a complementary treatment in this patients is lacking at present. PurposeTo determine whether complementary SOL treatment in NSCLC patients receiving adjuvant chemotherapy would yield greater improvements in QoL than chemotherapy alone. Study designWe conducted a multicenter, randomized controlled trial of stages IIA-IIIA NSCLC patients undergoing adjuvant chemotherapy in seven hospitals. MethodsUsing stratified blocks, participants were randomized in a 1:1 ratio to receive SOL combined with conventional chemotherapy or conventional chemotherapy alone. The primary outcome was the change in global QoL from baseline to the fourth chemotherapy cycle, and intention-to-treat analysis was applied with a mixed-effect model. Secondary outcomes were functional QoL, symptoms, and performance status scores at the 6-month follow-up. Missing data were handled with multiple imputation and a pattern-mixture model. ResultsAmong 516 randomized patients, 446 (86.43%) completed the study. After the fourth chemotherapy cycle, in comparison with the control group, patients receiving SOL showed a lower reduction in mean global QoL (-2.76 vs. -14.11; mean difference [MD], 11.34; 95% confidence interval [CI], 8.28 to 14.41), greater improvement in physical function (MD, 11.61; 95% CI, 8.57 to 14.65), role function (MD, 10.15; 95% CI, 5.75 to 14.54), and emotional function (MD, 4.71; 95% CI, 1.85 to 7.57), and greater improvements in lung cancer-related symptoms (e.g., fatigue, nausea/vomiting, and appetite loss) and performance status during the 6-month follow-up period (treatment main effect, p < 0.05). ConclusionSOL treatment for NSCLC patients receiving adjuvant chemotherapy can significantly improve QoL and performance status within 6 months after radical resection. Trial registrationClinicalTrials.gov identifier: NCT03712969.

THE LONGITUDINAL RELATIONSHIPS BETWEEN SOCIAL ISOLATION AND HEALTH OUTCOMES: THE ROLE OF PHYSICAL FRAILTY

Abstract Social isolation is a public health issue that is linked to poor health outcomes. However, the mechanisms underlying this association remain unclear. The main objective of this study was to explore whether changes in frailty moderated the relationship between changes in social isolation and changes in health outcomes over two years. We examined the mediating role of changes in frailty when the moderation hypothesis was not supported. A series of latent growth models (LGMs) were used to test our objectives using data from three waves of the FRéLE study among 1643 Canadian community-dwelling older adults aged 65 years and over. Missing data were handled by pattern mixture models with the assumption of missing not at random. We measured social isolation through social participation, social networks, and social support from different sources of social ties. We assessed frailty using the Fried frailty phenotype. Our moderation results revealed that high levels of changes in social participation, support from friends, nuclear, and extended family members, and social contacts with friends were associated with greater changes in cognitive and mental health among frail older adults with diminished physiological reserves compared to robust older adults. Additionally, changes in frailty mediated the effects of changes in social participation and social contacts and support from friends on changes in chronic conditions. This longitudinal study suggests that frailty moderated the relationships between social isolation and mental and cognitive health but not physical health. Overall, social support and strong friendship ties are key determinants of frail older adults’ health.

A Case Study of Nonresponse Bias Analysis in Educational Assessment Surveys

Nonresponse bias is a widely prevalent problem for data on education. We develop a ten-step exemplar to guide nonresponse bias analysis (NRBA) in cross-sectional studies and apply these steps to the Early Childhood Longitudinal Study, Kindergarten Class of 2010–2011. A key step is the construction of indices of nonresponse bias based on proxy pattern-mixture models for survey variables of interest. A novel feature is to characterize the strength of evidence about nonresponse bias contained in these indices, based on the strength of the relationship between the characteristics in the nonresponse adjustment and the key survey variables. Our NRBA improves the existing methods by incorporating both missing at random and missing not at random mechanisms, and all analyses can be done straightforwardly with standard statistical software.