Variability in Causal Effects and Noncompliance in a Multisite Trial: A Bivariate Hierarchical Generalized Random Coefficients Model for a Binary Outcome.

Within each of 170 physicians, patients were randomized to access e-assist, an online program that aimed to increase colorectal cancer screening (CRCS), or control. Compliance was partial: 78.34 % $$ 78.34\% $$ of the experimental patients accessed e-assist while no controls were provided the access. Of interest are the average causal effect of assignment to treatment and the complier average causal effect as well as the variation of these causal effects across physicians. Each physician generates probabilities of screening for experimental compliers (experimental patients who accessed e-assist), control compliers (controls who would have accessed e-assist had they been assigned to e-assist), and never takers (patients who would have avoided e-assist no matter what). Estimating physician-specific probabilities jointly over physicians poses novel challenges. We address these challenges by maximum likelihood, factoring a "complete-data likelihood" uniquely into the conditional distribution of screening and partially observed compliance given random effects and the distribution of random effects. We marginalize this likelihood using adaptive Gauss-Hermite quadrature. The approach is doubly iterative in that the conditional distribution defies analytic evaluation. Because the small sample size per physician constrains estimability of multiple random effects, we reduce their dimensionality using a shared random effects model having a factor analytic structure. We assess estimators and recommend sample sizes to produce reasonably accurate and precise estimates by simulation, and analyze data from a trial of a CRCS intervention.

Racial disparities in cancer genetic counseling encounters: study protocol for investigating patient-genetic counselor communication in the naturalistic clinical setting using a convergent mixed methods design.

BACKGROUND: Despite decades of effort to reduce racial cancer disparities, Black people continue to die at higher rates from cancer than any other U.S. racial group. Because prevention is a key to the cost-effective and long-term control of cancer, the potential for cancer genetic counseling to play a central role in reducing racial cancer disparities is high. However, the benefits of genetic counseling are not equitable across race. Only 2% of genetic counselors self-identify as Black/African American, so most genetic counseling encounters with Black patients are racially discordant. Patients in racially discordant medical interactions tend to have poorer quality patient-provider communication and receive suboptimal clinical recommendations. One major factor that contributes to these healthcare disparities is racial bias. Drawing on findings from prior research, we hypothesize that genetic counselor providers' implicit racial prejudice will be associated negatively with the quality of patient-provider communication, while providers' explicit negative racial stereotypes will be associated negatively with the comprehensiveness of clinical discussions of cancer risk and genetic testing for Black (vs. White) patients. METHODS: Using a convergent mixed methods research design, we will collect data from at least 15 genetic counseling providers, from two different institutions, and their 220 patients (approximately equal number of Black and White patients per provider) whose appointments are for a hereditary cancer condition. The data sources will include two provider surveys, two patient surveys, video- and/or audio-recordings of genetic counseling encounters, and medical chart reviews. The recorded cancer genetic counseling in-person and telehealth encounters will be analyzed both qualitatively and quantitatively to assess the quality of patient-provider communication and the comprehensiveness of clinical discussion. Those data will be linked to pre- and post-encounter survey data and data from medical chart reviews to test our hypotheses. DISCUSSION: Findings from this multi-site study will highlight specific aspects of cancer genetic counseling encounters (patient-provider communication and clinical recommendations) that are directly associated with patient-centered outcomes (e.g., satisfaction, trust, genetic testing completion). Patient-provider communication and clinical recommendations are modifiable factors that can be integrated into current genetic counseling training curricula and thus can have immediate impact on genetic counseling training and practice.

An exploration of cultural competency training and genetic counselors' racial biases.

Research shows genetic counselors generally have pro-White implicit biases-both prejudice and stereotyping. Cultural competency training aims to foster equitable beliefs, behaviors, and attitudes in cross-cultural genetic counseling sessions, including those that are racially discordant (genetic counselors and patients are from different racial backgrounds). Therefore, cultural competency training has the potential to mitigate bias and reduce disparities. Here, we report the prevalence of cultural competency training among genetic counselors and associations between recency of training and counselors' racial biases. We conducted an online survey of genetic counselors and trainees in fall 2021. The survey assessed four types of bias (implicit/explicit prejudice and implicit/explicit stereotyping), time since last cultural competency training, time since last communication skills training, and frequency of clinic sessions with Black patients. Multiple linear regressions modeled associations between cultural competency training and different types of bias, adjusting for communication skills training, frequency of encounters with Black patients, and counselor race (White vs. non-White). Two hundred fifteen participants (107 genetic counselors and 108 trainees) responded, and 205 reported whether they had prior cultural competency training. Of these, 187 (91%) reported ever having cultural competency training, most (53%) of participants who had training had it within 6 months prior to survey completion. We found no clear pattern of associations between cultural competency training and racial biases (implicit or explicit) in adjusted analyses. Participants who had cultural competency training four or more years prior demonstrated less negative implicit stereotyping toward Black individuals compared with those having more recent training; but no statistically significant effect was found for participants who reported never having cultural competency training, compared with those having training more recently than 4 years prior. Overall, our findings do not support that cultural competency training is negatively associated with, or mitigates, Black/White racial prejudices and stereotypes against Black patients. These findings suggest more effective interventions are needed to reduce racial biases.

Predictors of smell recovery in a nationwide prospective cohort of patients with COVID-19.

OBJECTIVE: To determine which factors (demographic, symptoms, comorbidities, and treatments) are associated with recovery of smell in patients with COVID-19 associated olfactory loss. STUDY DESIGN: Prospective, longitudinal questionnaires. SETTING: National survey. METHODS: A longitudinal web-based nationwide survey of adults with COVID-19 associated smell and taste loss was launched April 10, 2020. After completing an initial entry survey, participants received detailed follow-up questionnaires 14 days, and 1, 3 and 6 months later. RESULTS: As of June 25, 2021, 798 participants met study inclusion criteria and completed 6-month questionnaires. Of demographic characteristics only age <40 years was positively associated with smell recovery (p < .003). Of symptoms, difficulty breathing was negatively associated with smell recovery (p < .004), and nasal congestion positively associated with smell recovery (p < .03). Of pre-existing comorbidities only previous head injury (p < .017) was negatively associated with smell recovery. None of the queried medications used to treat COVID were associated with better rates of smell recovery. CONCLUSIONS: Age <40 and presence of nasal congestion at time of COVID-19 infection were predictive of improved rates of smell recovery, while difficulty breathing at time of COVID-19 infection, and prior head trauma predicted worsened rates of recovery. Further study will be required to identify potential mechanisms for the other observed associations. Such information can be used by clinicians to counsel patients suffering COVID-19 associated smell loss as to prognosis for recovery.

Quality of life and safety impact of COVID-19 associated smell and taste disturbances.

OBJECTIVE: The association between COVID-19 and chemosensory loss has garnered substantial attention, however to date little is known about the real-life consequences of impairment in this unique patient population. The aim of this study is to evaluate the quality of life (QOL) and personal safety deficits experienced by patients with COVID-19 infection. STUDY DESIGN: Prospective, longitudinal questionnaires. SETTING: National survey. METHODS: A longitudinal web-based nationwide survey of adults with COVID-19 and/or a sudden change in smell and taste was launched April 10, 2020. Previously published questions on chemosensory-related QOL and safety events were asked at the 6-month follow-up survey. RESULTS: As of February 10, 2021, 480 eligible respondents took the 6-month questionnaire, of whom 322 were COVID-19 positive. Impact on QOL was substantial with 96% of subjects reporting at least one of the defined deficits, and over 75% reporting at least 3 of these. "Reduced enjoyment of food" was the most common complaint (87%), while 43% of subjects self-reported depression. The prevalence of safety-related issues was common in this population, with over 57% reporting at least one, and 36% reporting 2 or more events. Of the events asked, the inability to smell smoke that others could perceive was the most common at 45%. CONCLUSIONS: COVID-19 associated chemosensory losses have a real and substantial impact on both quality of life and safety, beyond mere inconvenience. The high prevalence of these issues despite a relatively short period of olfactory deficit should alert clinicians to the serious risks to an already vulnerable patient population.

Impact of adolescent obesity on middle-age health of women given data MAR.

We analyze adolescent BMI and middle-age systolic blood pressure (SBP) repeatedly measured on women enrolled in the Fels Longitudinal Study (FLS) between 1929 and 2010 to address three questions: Do adolescent-specific growth rates in body mass index (BMI) and menarche affect middle-age SBP? Do they moderate the aging effect on middle-age SBP? Have the effects changed over historical time? To address the questions, we propose analyzing a growth curve model (GCM) that controls for age, birth-year cohort, and historical time. However, several complications in the data make the GCM analysis nonstandard. First, the person-specific adolescent BMI and middle-age SBP trajectories are unobservable. Second, missing data are substantial on BMI, SBP, and menarche. Finally, modeling the latent trajectories for BMI and SBP, repeatedly measured on two distinct sets of unbalanced time points, are computationally intensive. We adopt a bivariate GCM for BMI and SBP with correlated random coefficients. To efficiently handle missing values of BMI, SBP, and menarche assumed missing at random, we estimate their joint distribution by maximum likelihood via the EM algorithm where the correlated random coefficients and menarche are multivariate normal. The estimated distribution will be transformed to the desired GCM for SBP that includes the random coefficients of BMI and menarche as covariates. We demonstrate unbiased estimation by simulation. We find that adolescent growth rates in BMI and menarche are positively associated with, and moderate, the aging effect on SBP in middle age, controlling for age, cohort, and historical time, but the effect sizes are at most modest. The aging effect is significant on SBP, controlling for cohort and historical time, but not vice versa.

Statistical Power in Two-Level Hierarchical Linear Models with Arbitrary Number of Factor Levels.

As the US health care system undergoes unprecedented changes, the need for adequately powered studies to understand the multiple levels of main and interaction factors that influence patient and other care outcomes in hierarchical settings has taken center stage. We consider two-level models where n lower-level units are nested within each of J higher-level clusters (e.g. patients within practices and practices within networks) and where two factors may have arbitrary a and b factor levels, respectively. Both factors may represent a × b treatment combinations, or one of them may be a pretreatment covariate. Consideration of both factors at the same higher or lower hierarchical level, or one factor per hierarchical level yields a cluster (C), multisite (M) or split-plot randomized design (S). We express statistical power to detect main, interaction, or any treatment effects as a function of sample sizes (n, J), a and b factor levels, intraclass correlation ρ and effect sizes δ given each design d ∈ {C, M, S}. The power function given a, b, ρ, δ and d determines adequate sample sizes to achieve a minimum power requirement. Next, we compare the impact of the designs on power to facilitate selection of optimal design and sample sizes in a way that minimizes the total cost given budget and logistic constraints. Our approach enables accurate and conservative power computation with a priori knowledge of only three effect size differences regardless of how large a × b is, simplifying previously available computation methods for health services and other researches.

Quantifying Variation in Gait Features from Wearable Inertial Sensors Using Mixed Effects Models.

The emerging technology of wearable inertial sensors has shown its advantages in collecting continuous longitudinal gait data outside laboratories. This freedom also presents challenges in collecting high-fidelity gait data. In the free-living environment, without constant supervision from researchers, sensor-based gait features are susceptible to variation from confounding factors such as gait speed and mounting uncertainty, which are challenging to control or estimate. This paper is one of the first attempts in the field to tackle such challenges using statistical modeling. By accepting the uncertainties and variation associated with wearable sensor-based gait data, we shift our efforts from detecting and correcting those variations to modeling them statistically. From gait data collected on one healthy, non-elderly subject during 48 full-factorial trials, we identified four major sources of variation, and quantified their impact on one gait outcome-range per cycle-using a random effects model and a fixed effects model. The methodology developed in this paper lays the groundwork for a statistical framework to account for sources of variation in wearable gait data, thus facilitating informative statistical inference for free-living gait analysis.

Longitudinal latent variable models given incompletely observed biomarkers and covariates.

In this paper, we analyze a two-level latent variable model for longitudinal data from the National Growth and Health Study where surrogate outcomes or biomarkers and covariates are subject to missingness at any of the levels. A conventional method for efficient handling of missing data is to re-express the desired model as a joint distribution of variables, including the biomarkers, that are subject to missingness conditional on all of the covariates that are completely observed, and estimate the joint model by maximum likelihood, which is then transformed to the desired model. The joint model, however, identifies more parameters than desired, in general. We show that the over-identified joint model produces biased estimation of the latent variable model and describe how to impose constraints on the joint model so that it has a one-to-one correspondence with the desired model for unbiased estimation. The constrained joint model handles missing data efficiently under the assumption of ignorable missing data and is estimated by a modified application of the expectation-maximization algorithm. Copyright © 2016 John Wiley & Sons, Ltd.

Peak Bone Mass and Patterns of Change in Total Bone Mineral Density and Bone Mineral Contents From Childhood Into Young Adulthood.

The literature has not reached a consensus on the age when peak bone mass is achieved. This study examines growth patterns of total bone mineral content (TBMC) and total bone mineral density (TBMD), peak bone mass, effect of concurrent anthropometry measures, and physical activity on growth patterns in a sample of 312 white males and 343 females aged 8-30 yr. We analyzed data from participants enrolled in Fels Longitudinal Study. Descriptive analysis was used to ascertain characteristics of participants and growth patterns of TBMC and TBMD. Mixed effects models were applied to predict ages at attainment of peak TBMC and TBMD and assess the effects of height, weight, body mass index (BMI), and habitual physical activity on the attainment. Significant differences between sexes were observed for measures of TBMC and TBMD, and differences varied with age. For females, predicted median ages at peak TBMC and TBMD attainments are 21.96 yr (interquartile range [IQR]: 21.81-22.21) and 22.31 yr (IQR: 21.95-22.59), respectively. For males, predicted median ages are 23.34 yr (IQR: 24.34-26.19) and 26.86 yr (IQR: 25.14-27.98) respectively. For females, height, weight, and BMI, but not physical activity, had significant influences on attainment of TBMC and TBMD (p<0.01). For males, weight and BMI, but not height and physical activity, exerted significant influence on attainment of TBMC and TBMD (p<0.01), and also modified correlations between age and peak TBMC and TBMD. Our results suggest that (1) for both sexes, trajectories of TBMC and TBMD follow a curvilinear pattern between ages 8 and 30 yr; (2) predicted ages at peak TBMC and TBMD are from early to late 20s for both white males and females, with females reaching their peaks significantly earlier than males; and (3) concurrent height, weight, and BMI, but not habitual physical activity, exert significant effects on trajectories of TBMC and TBMD.

Confidence Bounds and Power for the Reliability of Observational Measures on the Quality of a Social Setting.

Social scientists are frequently interested in assessing the qualities of social settings such as classrooms, schools, neighborhoods, or day care centers. The most common procedure requires observers to rate social interactions within these settings on multiple items and then to combine the item responses to obtain a summary measure of setting quality. A key aspect of the quality of such a summary measure is its reliability. In this paper we derive a confidence interval for reliability, a test for the hypothesis that the reliability meets a minimum standard, and the power of this test against alternative hypotheses. Next, we consider the problem of using data from a preliminary field study of the measurement procedure to inform the design of a later study that will test substantive hypotheses about the correlates of setting quality. The preliminary study is typically called the "generalizability study" or "G study" while the later, substantive study is called the "decision study" or "D study." We show how to use data from the G study to estimate reliability, a confidence interval for the reliability, and the power of tests for the reliability of measurement produced under alternative designs for the D study. We conclude with a discussion of sample size requirements for G studies.

Opportunities and Challenges When Using the Electronic Health Record for Practice-Integrated Patient-Facing Interventions: The e-Assist Colon Health Randomized Trial.

BACKGROUND: Even after a physician recommendation, many people remain unscreened for colorectal cancer (CRC). The proliferation of electronic health records (EHRs) and tethered online portals may afford new opportunities to embed patient-facing interventions within clinic workflows and engage patients following a physician recommendation for care. We evaluated the effectiveness of a patient-facing intervention designed to complement physician office-based recommendations for CRC screening. DESIGN: Using a 2-arm pragmatic, randomized clinical trial, we evaluated the intervention's effect on CRC screening use as documented in the EHR (primary outcome) and the extent to which the intervention reached the target population. Trial participants were insured, aged 50 to 75 y, with a physician recommendation for CRC screening. Typical EHR functionalities, including patient registries, health maintenance flags, best practice alerts, and secure messaging, were used to support research-related activities and deliver the intervention to enrolled patients. RESULTS: A total of 1,825 adults consented to trial participation, of whom 78% completed a baseline survey and were exposed to the intervention. Most trial participants (>80%) indicated an intent to be screened on the baseline survey, and 65% were screened at follow-up, with no significant differences by study arm. One-third of eligible patients were sent a secure message. Among those, more than three-quarters accessed study material. CONCLUSIONS: By leveraging common EHR functionalities, we integrated a patient-facing intervention within clinic workflows. Despite practice integration, the intervention did not improve screening use, likely in part due to portal-based interventions not reaching those for whom the intervention may be most effective. IMPLICATIONS: Embedding patient-facing interventions within the EHR enabled practice integration but may minimize program effectiveness by missing important segments of the patient population. HIGHLIGHTS: Electronic health record tools can be used to facilitate practice-embedded pragmatic trial and patient-facing intervention processes, including patient identification, study arm allocation, and intervention delivery.The online portal-embedded intervention did not improve colorectal cancer (CRC) screening uptake following a physician recommendation, likely in part because portal users tend to be already highly engaged with healthcare.Relying on patient portals alone for CRC screening interventions may not alter screening use and could exacerbate well-known care disparities.

Gait Speed Trajectory During the Six-Minute Walk Test in Multiple Sclerosis: A Measure of Walking Endurance.

Background: The six-minute walk (6MW) test is a validated assessment method in Multiple Sclerosis (MS) research. While the total distance covered during six minutes (6MWTD ) is often used as the standard measurement of gait capacity (i.e., the maximum distance one can achieve), we hypothesize that endurance (i.e., ability to maintain speed over a prolonged time) can be inferred by the gait speed trajectory (GST) during the 6MW test (6MWGST ). Objective: To characterize group differences in 6MWGST between MS patients and healthy controls (HCs), and to assess information added by 6MWGST for discerning between MS patients and HCs. Methods: We performed a secondary data analysis on a cross-sectional cohort of 40 MS and 20 HC subjects with three repeated 6MW tests. We modeled 6MWGST using a linear mixed-effects model with time in minutes and replicated walks nested within individuals. We compared the discernibility of 6MWGST with that of conventional metrics using likelihood ratio tests and receiver operating characteristic (ROC) analysis on logistic regression models. Results: MS subjects showed a concave, quadratic GST during 6MW tests, slowing down more than the HC subjects, especially at the beginning of 6MW tests. Despite accelerating at the end of the 6MW, MS subjects were unable to attain or surpass their initial 6MW gait speeds. 6MWGST added useful information (p = 0.002) to the conventional metrics (e.g., 6MWTD ) for discerning between MS and HC subjects, and increased the area under the ROC curve from 0.83 to 0.93 (p = 0.037). Conclusions: The distinctive 6MWGST pattern of MS patients provided increased discernibility compared with currently used gait metrics. Both gait capacity measured by the 6MWTD , and gait endurance measured by parameters of 6MWGST , are significant functional indicators for the MS population.

Fibrosis-4 Predicts the Need for Mechanical Ventilation in a National Multiethnic Cohort of Corona Virus Disease 2019.

Simple tests of routine data are needed for those with severe acute respiratory syndrome coronavirus 2, which causes corona virus disease 2019 (COVID-19), to help identify those who may need mechanical ventilation (MV). In this study, we aimed to determine if fibrosis-4 (FIB-4) is associated with the need for MV in patients with COVID-19 and if there is an association to determine the optimal FIB-4 cutoff. This was a retrospective, national, multiethnic cohort study of adults seen in an ambulatory or emergency department setting who were diagnosed with COVID-19. We used the TriNetX platform for analysis. Measures included demographics, comorbid diseases, and routine laboratory tests. A total of 4,901 patients with COVID-19 were included. Patients had a mean age of 56, 48% were women, 42% were obese, 38% were white, 40% were black, 15% had cardiac disease, 39% had diabetes mellitus, 20% had liver disease, and 50% had respiratory disease. The need for MV was 6%. The optimal FIB-4 cutoff for the need for MV was 3.04 (area under the curve, 0.735), which had sensitivity, specificity, and positive and negative predictive values of 42%, 77%, 11%, and 95%, respectively, with 93% accuracy. When stratified by race, increased FIB-4 remained associated with the need for MV in both white and black patients. Conclusion: FIB-4 can be used by frontline providers to identify patients that may require MV.

Randomised trial to evaluate the effectiveness and impact of offering postvisit decision support and assistance in obtaining physician-recommended colorectal cancer screening: the e-assist: Colon Health study-a protocol study.

INTRODUCTION: How to provide practice-integrated decision support to patients remains a challenge. We are testing the effectiveness of a practice-integrated programme targeting patients with a physician recommendation for colorectal cancer (CRC) screening. METHODS AND ANALYSIS: In partnership with healthcare teams, we developed 'e-assist: Colon Health', a patient-targeted, postvisit CRC screening decision support programme. The programme is housed within an electronic health record (EHR)-embedded patient portal. It leverages a physician screening recommendation as the cue to action and uses the portal to enrol and intervene with patients. Programme content complements patient-physician discussions by encouraging screening, addressing common questions and assisting with barrier removal. For evaluation, we are using a randomised trial in which patients are randomised to receive e-assist: Colon Health or one of two controls (usual care plus or usual care). Trial participants are average-risk, aged 50-75 years, due for CRC screening and received a physician order for stool testing or colonoscopy. Effectiveness will be evaluated by comparing screening use, as documented in the EHR, between trial enrollees in the e-assist: Colon Health and usual care plus (CRC screening information receipt) groups. Secondary outcomes include patient-perceived benefits of, barriers to and support for CRC screening and patient-reported CRC screening intent. The usual care group will be used to estimate screening use without intervention and programme impact at the population level. Differences in outcomes by study arm will be estimated with hierarchical logit models where patients are nested within physicians. ETHICS AND DISSEMINATION: All trial aspects have been approved by the Institutional Review Board of the health system in which the trial is being conducted. We will disseminate findings in diverse scientific venues and will target clinical and quality improvement audiences via other venues. The intervention could serve as a model for filling the gap between physician recommendations and patient action. TRIAL REGISTRATION NUMBER: NCT02798224; Pre-results.

Patient-Reported Needs Following a Referral for Colorectal Cancer Screening.

INTRODUCTION: Patient-physician communication about colorectal cancer screening can affect screening use, but discussions often lack information that patients need for informed decision making and seldom address personal preferences or barriers. To address this gap, a series of patient focus groups was conducted to guide the development of an online, interactive decision support program. This article presents findings on patient information needs and barriers to colorectal cancer screening after receiving a screening recommendation from a physician, and their perspectives on using electronic patient portals as platforms for health-related decision support. METHODS: Primary care patients with recent colonoscopy or stool testing orders were identified via the centralized data repository of a large Midwestern health system. Seven gender-stratified focus groups (N=45 participants) were convened between April and July 2016. Sessions were audio recorded, transcribed, coded, and analyzed for commonly expressed themes beginning in August 2016. RESULTS: Findings reveal a consistent need for simple and clear information on colorectal cancer screening. Participants desired step-by-step explanations of the colonoscopy procedure and information about bowel preparation options/alternatives. The desired level of additional information varied: some patients wanted to know about and act on test options, whereas others preferred following their physician-recommended testing path. Fears and concerns were prevalent, particularly about colonoscopy, and patients reported challenges getting these concerns and their informational needs addressed. Finally, they expressed consistent support for using the patient portal to gather additional information from their physician. CONCLUSIONS: Patient portals may offer an opportunity to build sustainable programs for decision support and assistance that are integrated with clinic workflows and processes.

Challenges and opportunities using online portals to recruit diverse patients to behavioral trials.

We describe the use of an online patient portal to recruit and enroll primary care patients in a randomized trial testing the effectiveness of a colorectal cancer (CRC) screening decision support program. We use multiple logistic regression to identify patient characteristics associated with trial recruitment, enrollment, and engagement. We found that compared to Whites, Blacks had lower odds of viewing the portal message (OR = 0.46, 95% CI = 0.37-0.57), opening the attached link containing the study material (OR = 0.75, 95% CI = 0.62-0.92), and consenting to participate in the trial (OR = 0.85, 95% CI = 0.67-0.93). We also found that compared to Whites, Asians had lower odds of viewing the portal message (OR = 0.53, 95% CI = 0.33-0.64), opening the attached link containing the study material (OR = 0.76, 95% CI = 0.54-0.97), consenting to participate in the trial (OR = 0.68, 95% CI = 0.53-0.95), and completing the trial's baseline questionnaire (OR = 0.59, 95% CI = 0.36-0.90). While portals offer an opportunity to mitigate human bias in trial invitations, because of racial disparities-not only in who has a portal account, but in how they interact with trial recruitment and enrollment material within the portal-using portals alone for trial recruitment may generate study samples that are not racially diverse.

Efficient analysis of Q-level nested hierarchical general linear models given ignorable missing data.

This article extends single-level missing data methods to efficient estimation of a Q-level nested hierarchical general linear model given ignorable missing data with a general missing pattern at any of the Q levels. The key idea is to reexpress a desired hierarchical model as the joint distribution of all variables including the outcome that are subject to missingness, conditional on all of the covariates that are completely observed and to estimate the joint model under normal theory. The unconstrained joint model, however, identifies extraneous parameters that are not of interest in subsequent analysis of the hierarchical model and that rapidly multiply as the number of levels, the number of variables subject to missingness, and the number of random coefficients grow. Therefore, the joint model may be extremely high dimensional and difficult to estimate well unless constraints are imposed to avoid the proliferation of extraneous covariance components at each level. Furthermore, the over-identified hierarchical model may produce considerably biased inferences. The challenge is to represent the constraints within the framework of the Q-level model in a way that is uniform without regard to Q; in a way that facilitates efficient computation for any number of Q levels; and also in a way that produces unbiased and efficient analysis of the hierarchical model. Our approach yields Q-step recursive estimation and imputation procedures whose qth-step computation involves only level-q data given higher-level computation components. We illustrate the approach with a study of the growth in body mass index analyzing a national sample of elementary school children.

Systemic central venous oxygen saturation is associated with clot strength during traumatic hemorrhagic shock: A preclinical observational model.

BACKGROUND: Clot strength by Thrombelastography (TEG) is associated with mortality during trauma and has been linked to severity of tissue hypoperfusion. However, the optimal method for monitoring this important relationship remains undefined. We hypothesize that oxygen transport measurements will be associated with clot strength during traumatic shock, and test this hypothesis using a swine model of controlled traumatic shock. METHODS: N = 33 swine were subjected to femur fracture and hemorrhagic shock by controlled arterial bleeding to a predetermined level of oxygen debt measured by continuous indirect calorimetry. Hemodynamics, oxygen consumption, systemic central venous oxygenation (ScvO2), base excess, lactate, and clot maximal amplitude by TEG (TEG-MA) as clot strength were measured at baseline and again when oxygen debt = 80 ml/kg during shock. Oxygen transport and metabolic markers of tissue perfusion were then evaluated for significant associations with TEG-MA. Forward stepwise selection was then used to create regression models identifying the strongest associations between oxygen transport and TEG-MA independent of other known determinants of clot strength. RESULTS: Multiple markers of tissue perfusion, oxygen transport, and TEG-MA were all significantly altered during shock compared to baseline measurements (p < 0.05). However, only ScvO2 demonstrated a strong bivariate association with TEG-MA measured during shock (R = 0.7, p < 0.001). ScvO2 measured during shock was also selected by forward stepwise selection as an important covariate in linear regression models of TEG-MA after adjusting for the covariates fibrinogen, pH, platelet count, and hematocrit (Whole model R² = 0.99, p ≤ 0.032). CONCLUSIONS: Among multiple measurements of oxygen transport, only ScvO2 was found to retain a significant association with TEG-MA during shock after adjusting for multiple covariates. ScvO2 should be further studied for its utility as a clinical marker of both tissue hypoxia and clot formation during traumatic shock.

Just-identified versus overidentified two-level hierarchical linear models with missing data.

The development of model-based methods for incomplete data has been a seminal contribution to statistical practice. Under the assumption of ignorable missingness, one estimates the joint distribution of the complete data for thetainTheta from the incomplete or observed data y(obs). Many interesting models involve one-to-one transformations of theta. For example, with y(i) approximately N(mu, Sigma) for i= 1, ... , n and theta= (mu, Sigma), an ordinary least squares (OLS) regression model is a one-to-one transformation of theta. Inferences based on such a transformation are equivalent to inferences based on OLS using data multiply imputed from f(y(mis) | y(obs), theta) for missing y(mis). Thus, identification of theta from y(obs) is equivalent to identification of the regression model. In this article, we consider a model for two-level data with continuous outcomes where the observations within each cluster are dependent. The parameters of the hierarchical linear model (HLM) of interest, however, lie in a subspace of Theta in general. This identification of the joint distribution overidentifies the HLM. We show how to characterize the joint distribution so that its parameters are a one-to-one transformation of the parameters of the HLM. This leads to efficient estimation of the HLM from incomplete data using either the transformation method or the method of multiple imputation. The approach allows outcomes and covariates to be missing at either of the two levels, and the HLM of interest can involve the regression of any subset of variables on a disjoint subset of variables conceived as covariates.