Marginal modeling of cluster-period means and intraclass correlations in stepped wedge designs with binary outcomes.

Stepped wedge cluster randomized trials (SW-CRTs) with binary outcomes are increasingly used in prevention and implementation studies. Marginal models represent a flexible tool for analyzing SW-CRTs with population-averaged interpretations, but the joint estimation of the mean and intraclass correlation coefficients (ICCs) can be computationally intensive due to large cluster-period sizes. Motivated by the need for marginal inference in SW-CRTs, we propose a simple and efficient estimating equations approach to analyze cluster-period means. We show that the quasi-score for the marginal mean defined from individual-level observations can be reformulated as the quasi-score for the same marginal mean defined from the cluster-period means. An additional mapping of the individual-level ICCs into correlations for the cluster-period means further provides a rigorous justification for the cluster-period approach. The proposed approach addresses a long-recognized computational burden associated with estimating equations defined based on individual-level observations, and enables fast point and interval estimation of the intervention effect and correlations. We further propose matrix-adjusted estimating equations to improve the finite-sample inference for ICCs. By providing a valid approach to estimate ICCs within the class of generalized linear models for correlated binary outcomes, this article operationalizes key recommendations from the CONSORT extension to SW-CRTs, including the reporting of ICCs.

Ten-year cross-sectional and longitudinal assessment and factors associated with unfavourable self-rated oral health in older adults in the United States.

OBJECTIVE: To investigate the association of demographic and socio-economic characteristics with self-reported oral health (SROH) among older adults who participated in the Health and Retirement Study (HRS) in 2008, 2018, or both, and to describe temporal changes. METHODS: Data were from the University of Michigan's Health and Retirement Study (HRS), a nationally representative longitudinal survey of Americans aged 51 and older. Responses from participants who completed the Core HRS survey and Dental Module (DM) in 2008 (n = 1310), 2018 (n = 1330), and the "common group" at both timepoints (n = 559) were analysed. Using the common group, the outcome measure was 2018 self-rated oral health (Favourable vs Unfavourable). Potential explanatory variables included 2008 self-rated oral health (SROH), sociodemographic, and dental utilisation-related factors. Survey logistic regression analysis was used to identify factors that were associated with unfavourable 2018 SROH in 2018. RESULTS: Unfavourable SROH prevalence was 28.5% and 31.6% in 2008 and 2018, respectively. Among the common, longitudinal group, the unfavourable prevalence remained the same, 26.1% at both timepoints. A positive association was seen between 2018 unfavourable SROH and baseline variables of 2008 unfavourable SROH, male gender, less education, and lower levels of wealth. CONCLUSIONS: Over a quarter of participants reported unfavourable SROH. There was little change in SROH during this period. Sociodemographic factors influence the SROH of the older population. Policies and programs to promote and protect the oral health of older adults should be designed and implemented to reduce social inequalities and improve the SROH of disadvantaged older adults.

power swgee: GEE-based power calculations in stepped wedge cluster randomized trials.

Stepped wedge cluster randomized trials are increasingly being used to evaluate interventions in medical, public health, educational, and social science contexts. With the longitudinal and crossover nature of a SW-CRT, complex analysis techniques are often needed which makes appropriately powering SW-CRTs challenging. In this paper, we introduce a newly-developed SW-CRT power calculator, embedded within the power command in Stata. The power calculator assumes a marginal model (i.e., generalized estimating equations [GEE]) for the primary analysis of SW-CRTs, for which other currently available SW-CRT power calculators may not be suitable. The program accommodates complete cross-sectional and closed-cohort designs, and includes multilevel correlation structures appropriate for such designs. We discuss the methods and formulae underlying our SW-CRT calculator, and provide illustrative examples of the use of power swgee. We provide suggestions about the choice of parameters in power swgee, and conclude by discussing areas of future research which may improve the program.

Power considerations for generalized estimating equations analyses of four-level cluster randomized trials.

In this article, we develop methods for sample size and power calculations in four-level intervention studies when intervention assignment is carried out at any level, with a particular focus on cluster randomized trials (CRTs). CRTs involving four levels are becoming popular in healthcare research, where the effects are measured, for example, from evaluations (level 1) within participants (level 2) in divisions (level 3) that are nested in clusters (level 4). In such multilevel CRTs, we consider three types of intraclass correlations between different evaluations to account for such clustering: that of the same participant, that of different participants from the same division, and that of different participants from different divisions in the same cluster. Assuming arbitrary link and variance functions, with the proposed correlation structure as the true correlation structure, closed-form sample size formulas for randomization carried out at any level (including individually randomized trials within a four-level clustered structure) are derived based on the generalized estimating equations approach using the model-based variance and using the sandwich variance with an independence working correlation matrix. We demonstrate that empirical power corresponds well with that predicted by the proposed method for as few as eight clusters, when data are analyzed using the matrix-adjusted estimating equations for the correlation parameters with a bias-corrected sandwich variance estimator, under both balanced and unbalanced designs.

Impact of unequal cluster sizes for GEE analyses of stepped wedge cluster randomized trials with binary outcomes.

The stepped wedge (SW) design is a type of unidirectional crossover design where cluster units switch from control to intervention condition at different prespecified time points. While a convention in study planning is to assume the cluster-period sizes are identical, SW cluster randomized trials (SW-CRTs) involving repeated cross-sectional designs frequently have unequal cluster-period sizes, which can impact the efficiency of the treatment effect estimator. In this paper, we provide a comprehensive investigation of the efficiency impact of unequal cluster sizes for generalized estimating equation analyses of SW-CRTs, with a focus on binary outcomes as in the Washington State Expedited Partner Therapy trial. Several major distinctions between our work and existing work include the following: (i) we consider multilevel correlation structures in marginal models with binary outcomes; (ii) we study the implications of both the between-cluster and within-cluster imbalances in sizes; and (iii) we provide a comparison between the independence working correlation versus the true working correlation and detail the consequences of ignoring correlation estimation in SW-CRTs with unequal cluster sizes. We conclude that the working independence assumption can lead to substantial efficiency loss and a large sample size regardless of cluster-period size variability in SW-CRTs, and recommend accounting for correlations in the analysis. To improve study planning, we additionally provide a computationally efficient search algorithm to estimate the sample size in SW-CRTs accounting for unequal cluster-period sizes, and conclude by illustrating the proposed approach in the context of the Washington State study.

Lexicogrammatical skills in 2-year-old children with and without repaired cleft palate.

The purpose of the current research was to compare the lexical-grammatical skills of two-year-old children with and without repaired cleft palate (CP), accounting for the effect of variables such as vocabulary size at 18 months of age, maternal education level, and gender. Participants included 52 children with CP and 25 typically developing (TD) children. The CDI-WS was employed to measure vocabulary and grammatical skills. Significant differences were observed between the CP and TD groups with respect to the number of words, word forms (irregular nouns and verbs), word endings (overuse of plural (-s) and past tense (-ed) markers), the mean number of morphemes in their three longest utterances (M3L), and sentence complexity. In addition, compared to TD children, significantly smaller proportions of children with CP were observed to use words to talk about past and future events or use words to talk about an absent object. The difference between the CP and TD groups in terms of the size of vocabulary at 24 months of age remained statistically significant in the multivariable model. Among all predictors, the size of vocabulary at 18 months of age was identified as the most robust precursor of lexical and grammatical skills at 24 months of age. Gender was identified as a predictor of the M3L measure as an index for syntactic ability.

Is Fluorescence Technology a Promising Tool for Detecting Infected Dentin in Deep Carious Lesions?

The purpose of this study was to determine if the degree of fluorescence detected by fluorescence-aided caries excavation (FACE) correlates with dentin bacterial microbiome diversity, as assessed by 16S rRNA gene amplicon sequencing, and with traditional tactile dentin caries assessment. Unidentified human teeth were obtained from a dental facility. The included teeth had a carious lesion two-thirds into the dentin, verified by radiography, and were red-fluorescing (RF) using FACE technology (SIROInspect; Sirona, Bensheim, Germany). Two independent examiners performed visual/tactile assessment of the lesions. RF sites were sampled with a sterile spoon excavator and dentin characteristics were evaluated. Once RF dentin was removed, a second sample of pink-fluorescing (PF) dentin was obtained. After excavation with a sterile round bur to nonfluorescing (NF) dentin, a third sample was collected with a slow-speed round bur. The samples were processed at the UNC (University of North Carolina at Chapel Hill) Microbiome Core Facility. Out of 134 extracted teeth collected, 21 fit the inclusion criteria, yielding 61 dentin samples. RF samples had the highest number of observed operational taxonomic units (n = 154), followed by PF (n = 109) and NF (n = 100). RF carious dentin was primarily "soft," and NF dentin was assessed as "hard" 100% of the time by both examiners (rank correlation χ2: p < 0.001). However, approximately one-third of the tactile assessments of hard dentin still displayed some fluorescence, either pink or red. We concluded that the sampled fluorescing (RF and PF) and NF carious dentin layers displayed diverse bacterial taxa, and tactile assessments of soft, leathery, and hard corresponded with RF, PF, and NF.

Impact of an Individually Tailored Light Mask on Sleep Parameters in Older Adults With Advanced Phase Sleep Disorder.

Objective: This study investigated whether light delivered through the eyelids of sleeping persons might create phase delay in older adults who are adversely affected by advanced sleep phase disorder. Participants: Thirty-two cognitively intact, community-dwelling participants aged ≥ 50 years (20 females, 12 males) with Pittsburgh Sleep Quality Index scores ≥ 5 (poor sleep) completed the study. Methods: This within-subjects, randomized, two-treatment crossover design study exposed participants to an active "blue" (λmax  =  480 nm) lighting intervention or a placebo "red" (λmax = 640 nm) control through closed eyelids during sleep for 8 weeks. Conditions were administered 1 hr after bedtime using custom-built light masks delivering a train of 2-s duration light pulses presented every 30 s for ≤ 2 hr (approximately 240 pulses/night). Dependent variables were subjective measures of sleep and depression (questionnaires) and objective measures of sleep (wrist actigraphy), analyzed using linear mixed models with treatment, period, and carryover as fixed effects. Results: The actigraphy analysis found no effect of the intervention or the control condition on sleep start time, total sleep time, number of sleep bouts, or sleep efficiency, either compared to baseline or to one another. Subjective responses of study participants, however, indicated statistically significant (p < 0.05) improvement in seven of eight reported measures of sleep quality with both the intervention and the control condition, but no difference between the two conditions. Conclusions: The participants reported improvement in sleep quality, but the intervention did not confer additional advantages after adjusting for period and carryover effects.

Fidelity and sustainability of Mouth Care Without a Battle and lessons for other innovations in care.

There are countless efficacious interventions that improve outcomes when conducted in controlled situations. Many fewer are effective when implemented in real-world situations, largely because they are not implemented with fidelity. Still fewer are sustained over time, for reasons including lack of institutional support and fit with existing values, among others. It is especially important to examine fidelity and sustainability when efficacious interventions are being implemented, because these interventions are the ones that hold the most promise. This project examined the fidelity and sustainability of Mouth Care Without a Battle (MCWB), an evidence-based program conducted in a two-year cluster randomized trial in 14 nursing homes. Results that triangulated two sources of data indicated that fidelity decreased after the first year; they provide guidance to promote fidelity and sustainability of this and other new care practices in nursing homes, including ongoing education, coaching, evaluation, feedback, and sufficient resources.

Power and sample size requirements for GEE analyses of cluster randomized crossover trials.

The cluster randomized crossover design has been proposed to improve efficiency over the traditional parallel cluster randomized design, which often involves a limited number of clusters. In recent years, the cluster randomized crossover design has been increasingly used to evaluate the effectiveness of health care policy or programs, and the interest often lies in quantifying the population-averaged intervention effect. In this paper, we consider the two-treatment two-period crossover design, and develop sample size procedures for continuous and binary outcomes corresponding to a population-averaged model estimated by generalized estimating equations, accounting for both within-period and interperiod correlations. In particular, we show that the required sample size depends on the correlation parameters through an eigenvalue of the within-cluster correlation matrix for continuous outcomes and through two distinct eigenvalues of the correlation matrix for binary outcomes. We demonstrate that the empirical power corresponds well with the predicted power by the proposed formulae for as few as eight clusters, when outcomes are analyzed using the matrix-adjusted estimating equations for the correlation parameters concurrently with a suitable bias-corrected sandwich variance estimator.

A population-averaged approach to diagnostic test meta-analysis.

The meta-analysis of diagnostic accuracy studies is often of interest in screening programs for many diseases. The typical summary statistics for studies chosen for a diagnostic accuracy meta-analysis are often two dimensional: sensitivities and specificities. The common statistical analysis approach for the meta-analysis of diagnostic studies is based on the bivariate generalized linear-mixed model (BGLMM), which has study-specific interpretations. In this article, we present a population-averaged (PA) model using generalized estimating equations (GEE) for making inference on mean specificity and sensitivity of a diagnostic test in the population represented by the meta-analytic studies. We also derive the marginalized counterparts of the regression parameters from the BGLMM. We illustrate the proposed PA approach through two dataset examples and compare performance of estimators of the marginal regression parameters from the PA model with those of the marginalized regression parameters from the BGLMM through Monte Carlo simulation studies. Overall, both marginalized BGLMM and GEE with sandwich standard errors maintained nominal 95% confidence interval coverage levels for mean specificity and mean sensitivity in meta-analysis of 25 of more studies even under misspecification of the covariance structure of the bivariate positive test counts for diseased and nondiseased subjects.

Sample size determination for GEE analyses of stepped wedge cluster randomized trials.

In stepped wedge cluster randomized trials, intact clusters of individuals switch from control to intervention from a randomly assigned period onwards. Such trials are becoming increasingly popular in health services research. When a closed cohort is recruited from each cluster for longitudinal follow-up, proper sample size calculation should account for three distinct types of intraclass correlations: the within-period, the inter-period, and the within-individual correlations. Setting the latter two correlation parameters to be equal accommodates cross-sectional designs. We propose sample size procedures for continuous and binary responses within the framework of generalized estimating equations that employ a block exchangeable within-cluster correlation structure defined from the distinct correlation types. For continuous responses, we show that the intraclass correlations affect power only through two eigenvalues of the correlation matrix. We demonstrate that analytical power agrees well with simulated power for as few as eight clusters, when data are analyzed using bias-corrected estimating equations for the correlation parameters concurrently with a bias-corrected sandwich variance estimator.

Patterns of periodontal disease progression based on linear mixed models of clinical attachment loss.

AIM: The goal of the present longitudinal cohort study was to examine patterns of periodontal disease progression at progressing sites and subjects defined based on linear mixed models (LMM) of clinical attachment loss (CAL). MATERIALS AND METHODS: A total of 113 periodontally healthy and 302 periodontitis subjects had their CAL calculated bimonthly for 12 months. LMMs were fitted for each site and the predicted CAL levels used to categorize their progression state. Participants were grouped based on the number of progressing sites into unchanged, transitional and active subjects. Patterns of periodontal disease progression were explored using descriptive statistics. RESULTS: Progression occurred primarily at molars (50% of progressing sites) and inter-proximal sites (72%), affected a higher proportion of deep than shallow sites (2.7% versus 0.7%), and pocketing was the main mode of progression (49%). We found a low level of agreement (47%) between the LMM and traditional approaches to determine progression such as change in CAL ≥3 mm. Fourteen per cent of subjects were classified as active and among those 93% had periodontitis. The annual mean rate of progression for the active subjects was 0.35 mm/year. CONCLUSION: Progressing sites and subjects defined based on LMMs presented patterns of disease progression similar to those previously reported in the literature.

Marginalized zero-inflated Poisson models with missing covariates.

Unlike zero-inflated Poisson regression, marginalized zero-inflated Poisson (MZIP) models for counts with excess zeros provide estimates with direct interpretations for the overall effects of covariates on the marginal mean. In the presence of missing covariates, MZIP and many other count data models are ordinarily fitted using complete case analysis methods due to lack of appropriate statistical methods and software. This article presents an estimation method for MZIP models with missing covariates. The method, which is applicable to other missing data problems, is illustrated and compared with complete case analysis by using simulations and dental data on the caries preventive effects of a school-based fluoride mouthrinse program.

Impact of Early Head Start in North Carolina on Dental Care Use Among Children Younger Than 3 Years.

OBJECTIVES: To examine the effects of North Carolina Early Head Start (EHS), an early education program for low-income children younger than 3 years and their families, on dental care use among children. METHODS: We performed a quasi-experimental study in which we interviewed 479 EHS and 699 non-EHS parent-child dyads at baseline (2010-2012) and at a 24-month follow-up (2012-2014). We estimated the effects of EHS participation on the probability of having a dental care visit after controlling for baseline dental care need and use and a propensity score covariate; we included random effects to account for EHS program clustering. RESULTS: The odds of having a dental care visit of any type (adjusted odds ratio [OR] = 2.5; 95% confidence interval [CI] = 1.74, 3.48) and having a preventive dental visit (adjusted OR = 2.6; 95% CI = 1.84, 3.63) were higher among EHS children than among non-EHS children. In addition, the adjusted mean number of dental care visits among EHS children was 1.3 times (95% CI = 1.17, 1.55) the mean number among non-EHS children. CONCLUSIONS: This study is the first, to our knowledge, to demonstrate that EHS participation increases dental care use among disadvantaged young children.

Alternating logistic regressions with improved finite sample properties.

Alternating logistic regressions is an estimating equations procedure used to model marginal means of correlated binary outcomes while simultaneously specifying a within-cluster association model for log odds ratios of outcome pairs. A recent generalization of alternating logistic regressions, known as orthogonalized residuals, is extended to incorporate finite sample adjustments in the estimation of the log odds ratio model parameters for when there is a moderately small number of clusters. Bias adjustments are made both in the sandwich variance estimators and in the estimating equations for the association parameters. The proposed methods are demonstrated in a repeated cross-sectional cluster trial to reduce underage drinking in the United States, and in an analysis of dental caries incidence in a cluster randomized trial of 30 aboriginal communities in the Northern Territory of Australia. A simulation study demonstrates improved performance with respect to bias and coverage of their estimators relative to those based on the uncorrected orthogonalized residuals procedure.

A new way to estimate disease prevalence from random partial-mouth samples.

AIM: Standard partial-mouth estimators of chronic periodontitis (CP) that define an individual's disease status solely in terms of selected sites underestimate prevalence. This study proposes an improved prevalence estimator based on randomly sampled sites and evaluates its accuracy in a well-characterized population cohort. METHODS: Importantly, this method does not require determination of disease status at the individual level. Instead, it uses a statistical distributional approach to derive a prevalence formula from randomly selected periodontal sites. The approach applies the conditional linear family of distributions for correlated binary data (i.e. the presence or absence of disease at sites within a mouth) with two simple working assumptions: (i) the probability of having disease is the same across all sites; and (ii) the correlation of disease status is the same for all pairs of sites within the mouth. RESULTS: Using oral examination data from 6793 participants in the Arteriolosclerosis Risk in Communities study, the new formula yields CP prevalence estimates that are much closer than standard partial mouth estimates to full mouth estimates. CONCLUSIONS: Resampling of the cohort shows that the proposed estimators give good precision and accuracy for as few as six tooth sites sampled per individual.

Enrollment in early head start and oral health-related quality of life.

PURPOSE: Dental problems in young children are widespread and can negatively impact quality of life. We examined the effect of enrollment in North Carolina Early Head Start (EHS)-a federally funded early education program for children under three years of age and their families-on oral health-related quality of life (OHRQoL). METHODS: In this quasi-experimental study, we interviewed 479 EHS and 699 Medicaid matched parent-child dyads at baseline (children's average age 10 months) and 24 months later. Parents reported OHRQoL using the Early Childhood Oral Health Impact Scale (ECOHIS), a 0-52 point scale with higher scores representing more negative impacts. We used a marginalized semicontinuous two-part model to estimate: (1) the effect of EHS on the probability of reporting any follow-up impacts (ECOHIS ≥ 1), and (2) the difference in overall mean ECOHIS follow-up scores. We controlled for baseline ECOHIS, language, and EHS and non-EHS group imbalances using a propensity score. RESULTS: At follow-up, negative OHRQoL impacts were more often reported by parents of non-EHS than EHS children (45 versus 37%, P < .01). In the adjusted model, EHS parents reported a lower odds of negative OHRQoL impacts (OR 0.70; 95% CI 0.52, 0.94). Mean adjusted ECOHIS scores were not significantly different (EHS: 1.59 ± 3.34 versus non-EHS: 2.11 ± 3.85, P > 0.05). CONCLUSIONS: This study is the first to demonstrate that families of young children enrolled in EHS report improved OHRQoL compared to their non-enrolled peers. These results highlight the potential effectiveness of improving the quality of life of low-resource families through early childhood education.

Randomization-based adjustment of multiple treatment hazard ratios for covariates with missing data.

Clinical trials are designed to compare treatment effects when applied to samples from the same population. Randomization is used so that the samples are not biased with respect to baseline covariates that may influence the efficacy of the treatment. We develop randomization-based covariance adjustment methodology to estimate the log hazard ratios and their confidence intervals of multiple treatments in a randomized clinical trial with time-to-event outcomes and missingness among the baseline covariates. The randomization-based covariance adjustment method is a computationally straight-forward method for handling missing baseline covariate values.

Sensitivity analysis for missing dichotomous outcome data in multi-visit randomized clinical trial with randomization-based covariance adjustment.

Dichotomous endpoints in clinical trials have only two possible outcomes, either directly or via categorization of an ordinal or continuous observation. It is common to have missing data for one or more visits during a multi-visit study. This paper presents a closed form method for sensitivity analysis of a randomized multi-visit clinical trial that possibly has missing not at random (MNAR) dichotomous data. Counts of missing data are redistributed to the favorable and unfavorable outcomes mathematically to address possibly informative missing data. Adjusted proportion estimates and their closed form covariance matrix estimates are provided. Treatment comparisons over time are addressed with Mantel-Haenszel adjustment for a stratification factor and/or randomization-based adjustment for baseline covariables. The application of such sensitivity analyses is illustrated with an example. An appendix outlines an extension of the methodology to ordinal endpoints.