Ian Barnett

Study Objectives: While growing evidence suggests neighborhood safety and physical disorder may impact adolescent sleep, prior research has not accounted for mobility outside the home neighborhood. This study examined associations of crime and physical disorder surrounding adolescents’ homes, schools, and in their daily activity spaces with sleep patterns. Materials and Methods: Adolescents aged 15–18 years (n = 142) were followed for 14 days using smartphone global position systems (GPS) tracking and actigraphy. GPS locations recorded during adolescents’ waking hours were spatially linked to local data on crime and vacant lots to derive exposures surrounding adolescents’ homes, schools, and daily activity spaces. Mixed-effects linear and modified Poisson models estimated associations of crime and vacant lots with sleep duration, timing, efficiency, insufficient (<8 hours) and very short sleep (<6 hours). Fixed-effects models estimated associations of within-person differences in activity-space crime and vacant lot exposures with sleep patterns. Results: Higher crime exposure surrounding adolescents’ homes and schools was associated with shorter sleep duration, higher risk of insufficient and very short sleep, earlier offset timing (home only) and later onset timing (school only). Higher activity-space exposure to vacant lots was associated with shorter sleep duration, later onset, and higher risk of insufficient sleep. On days when adolescents’ exposure to vacant lots was higher relative to their own average, they had shorter sleep duration, earlier offset timing, and higher risk of insufficient and very short sleep, while such associations were not seen for neighborhood crime. Conclusions: Associations of crime and vacant lot exposures with actigraphy-assessed adolescent sleep patterns varied by setting.

Background: Cigarettes are a global public health concern, as cigarette smoking is the leading cause of death in the United States and throughout most high-income countries. Exposure to tobacco retail has been linked to adverse smoking outcomes, but research using naturalistic and causal approaches to quantify these effects in the real world remains relatively sparse. To address these gaps, this study used geolocation tracking, ecological momentary assessment, and neuroimaging to assess smoking outcomes in daily life and conducted a randomized controlled trial focused on the effects of exposure to tobacco retail. Objective: The GeoSmoking study aimed to evaluate (1) within-person associations between real-world tobacco retail exposure and cigarette craving and smoking, (2) causal effects of real-world tobacco retail exposure, and (3) neural cue reactivity as a mechanism for real-world tobacco retail effects. This paper describes the implemented protocol, including study status and flow reporting. Methods: In a 2-week baseline period, the study collected reports of craving and smoking multiple times per day using ecological momentary assessment, in addition to other measures. Simultaneously, geolocation tracking was used to quantify tobacco retail exposure through the creation of a tobacco retail database across 3 US states (Pennsylvania, New Jersey, and Delaware). A 4-week intervention period followed, in which participants were randomly assigned to make a purchase at either a nontobacco retail store 5 days per week (nontobacco retail condition) or a tobacco retail store 5 days per week (tobacco retail condition), or follow their normal routines (control condition). An optional functional magnetic resonance imaging (fMRI) session concluded the study. Individuals participated remotely, unless they opted into the fMRI session, which was completed at the University of Pennsylvania. Results: The GeoSmoking study was approved by the Institutional Review Board at the University of Pennsylvania. Data collection started on May 25, 2022, and ended on June 10, 2024. In total, 310 participants were enrolled, 282 participants completed the baseline phase, 244 participants completed the intervention phase, and 24 participants completed the optional fMRI scan. Conclusions: This study protocol was implemented successfully. Findings from planned analyses will quantify the strength of relationships between naturalistic exposure to tobacco retail, craving, smoking, and a range of other outcomes. These findings have significant implications for our understanding of health behaviors and outcomes, as well as policy.

Bayesian Cox semiparametric regression is an important problem in many clinical settings. The elliptical information geometry of Cox models is underutilized in Bayesian inference but can effectively bridge survival analysis and hierarchical Gaussian models. Survival models should be able to incorporate multilevel modeling such as case weights, frailties, and smoothing splines, in a straightforward manner similar to Gaussian models. To tackle these challenges, we propose the Cox-Pólya-Gamma algorithm for Bayesian multilevel Cox semiparametric regression and survival functions. Our novel computational procedure succinctly addresses the difficult problem of monotonicity-constrained modeling of the nonparametric baseline cumulative hazard along with multilevel regression. We develop two key strategies based on the elliptical geometry of Cox models that allows computation to be implemented in a few lines of code. First, we exploit an approximation between Cox models and negative binomial processes through the Poisson process to reduce Bayesian computation to iterative Gaussian sampling. Next, we appeal to sufficient dimension reduction to address the difficult computation of nonparametric baseline cumulative hazards, allowing for the collapse of the Markov transition within the Gibbs sampler based on beta sufficient statistics. We explore conditions for uniform ergodicity of the Cox-Pólya-Gamma algorithm. We provide software and demonstrate our multilevel modeling approach using open-source data and simulations.

Objective Concussion symptoms can be clustered into domains and understanding how multiple symptom domains present clinically may guide more accurate interventions. We investigate the associations between concurrent symptom domains and clinical recovery outcomes, as well as the role of sex in these relationships. Methods We analysed data from the Ivy League–Big Ten Epidemiology of Concussion Study and included sport-related concussions (SRC) across five academic years 2015–2016/2019–2020 with complete data (n=1160). We used symptoms from the Sport Concussion Assessment Tool 22-symptom evaluation, previously categorised into symptom domains. Symptom profiles characterise how athletes endorse concurrent symptom domains. Outcomes are time (in days) from SRC to symptom resolution, return to academics, and full play. Results Females more commonly endorsed headache, sensory, and affective symptom domains. Four classes/symptom profiles emerged: (1) ‘low’ on all domains, (2) ‘high’ on headache and sensory domains, (3) ‘high’ on vestibulo-ocular, cognitive, and sleep domains, and (4) ‘high’ on all domains. Time to symptom resolution, return to academics, and return to play were consistently shorter among class/symptom profile 1 compared with other classes/profiles. Compared with class/profile 1, the chance of having symptoms resolve was lower among classes/profiles 2, 3, and 4 (HR 0.74, 95% CI 0.63 to 0.88; HR 0.74, 95% CI 0.60 to 0.92; HR 0.50, 95% CI 0.43 to 0.57, respectively). Results were similar for return to academics and full play outcomes. Interactions with sex were not statistically significant. Conclusions Four symptom profiles characterised how concussion symptom domains co-occur. We found differences in recovery timelines among these groups, but not by sex. Findings inform and support targeted, symptom domain-specific interventions in concussion management.

Objective: Describe the factor structure of the 22-symptom Sport Concussion Assessment Tool (SCAT), using confirmatory factor analysis (CFA) for a priori hypothesized symptom domains. Study Design: Prospective observational study. Setting/Participants: Collegiate student-athletes with concussion. Independent Variables: Symptoms were collected via the SCAT symptom checklist. Outcome Measures: We created symptom domains based on previous literature, guided by clinical expertise. To determine which symptom grouping best represent the data, we used CFA and compared a single-domain model to 3- and 6-domains. We examined fit statistics to assess relative and absolute model fit. Motivated by differences in the prevalence of some individual symptoms by sex in our study, we also examined model invariance by sex to determine if symptoms were being measured as part of the same underlying construct(s). Results: Among 1160 concussions (male, n = 667; female, n = 493) between 2015 and 2020, all 3 symptom structures seemed to fit the data well, with 3- and 6-domains fitting better than 1-domain. The 6-domain structure fit the data best with the following domains: headache, vestibulo-ocular, sensory, cognitive, sleep, and affective. All 3 structures showed configural and metric invariance by sex. Conclusions: We demonstrate that the SCAT symptom structure is best represented through 6 specific factors; however, the 3-factor model also demonstrated good fit. Key differences between the 3- and 6-domain models may make 1 model more appropriate than the other depending on the research question being addressed. Symptom structures were configurally and metrically invariant by sex, meaning that symptom measures represent symptom domain factors in the same way across sex.

Background: Current methods of monitoring cognition in older adults are insufficient to address the growing burden of Alzheimer disease and related dementias (AD/ADRD). New approaches that are sensitive, scalable, objective, and reflective of meaningful functional outcomes are direly needed. Mobility trajectories and geospatial life space patterns reflect many aspects of cognitive and functional integrity and may be useful proxies of age-related cognitive decline. Objective: We investigated the feasibility, acceptability, and preliminary validity of a 1-month smartphone digital phenotyping protocol to infer everyday cognition, function, and mood in older adults from passively obtained GPS data. We also sought to clarify intrinsic and extrinsic factors associated with mobility phenotypes for consideration in future studies. Methods: Overall, 37 adults aged between 63 and 85 years with healthy cognition (n=31, 84%), mild cognitive impairment (n=5, 13%), and mild dementia (n=1, 3%) used an open-source smartphone app (mindLAMP) to unobtrusively capture GPS trajectories for 4 weeks. GPS data were processed into interpretable features across categories of activity, inactivity, routine, and location diversity. Monthly average and day-to-day intraindividual variability (IIV) metrics were calculated for each feature to test a priori hypotheses from a neuropsychological framework. Validation measures collected at baseline were compared against monthly GPS features to examine construct validity. Feasibility and acceptability outcomes included retention, comprehension of study procedures, technical difficulties, and satisfaction ratings at debriefing. Results: All (37/37, 100%) participants completed the 4-week monitoring period without major technical adverse events, 100% (37/37) reported satisfaction with the explanation of study procedures, and 97% (36/37) reported no feelings of discomfort. Participants’ scores on the comprehension of consent quiz were 97% on average and associated with education and race. Technical issues requiring troubleshooting were infrequent, though 41% (15/37) reported battery drain. Moderate to strong correlations (r≥0.3) were identified between GPS features and validators. Specifically, individuals with greater activity and more location diversity demonstrated better cognition, less functional impairment, less depression, more community participation, and more geospatial life space on objective and subjective validation measures. Contrary to predictions, greater IIV and less routine in mobility habits were also associated with positive outcomes. Many demographic and technology-related factors were not associated with GPS features; however, income, being a native English speaker, season of study participation, and occupational status were related to GPS features. Conclusions: Theoretically informed digital phenotypes of mobility are feasibly captured from older adults’ personal smartphones and relate to clinically meaningful measures including cognitive test performance, reported functional decline, mood, and community activity. Future studies should consider the impact of intrinsic and extrinsic factors when interpreting mobility phenotypes. Overall, smartphone digital phenotyping is a promising method to unobtrusively capture relevant risk and resilience factors in the context of aging and AD/ADRD and should continue to be investigated in large, diverse samples.

As digital phenotyping, the capture of active and passive data from consumer devices such as smartphones, becomes more common, the need to properly process the data and derive replicable features from it has become paramount. Cortex is an open-source data processing pipeline for digital phenotyping data, optimized for use with the mindLAMP apps, which is used by nearly 100 research teams across the world. Cortex is designed to help teams (1) assess digital phenotyping data quality in real time, (2) derive replicable clinical features from the data, and (3) enable easy-to-share data visualizations. Cortex offers many options to work with digital phenotyping data, although some common approaches are likely of value to all teams using it. This paper highlights the reasoning, code, and example steps necessary to fully work with digital phenotyping data in a streamlined manner. Covering how to work with the data, assess its quality, derive features, and visualize findings, this paper is designed to offer the reader the knowledge and skills to apply toward analyzing any digital phenotyping data set. More specifically, the paper will teach the reader the ins and outs of the Cortex Python package. This includes background information on its interaction with the mindLAMP platform, some basic commands to learn what data can be pulled and how, and more advanced use of the package mixed with basic Python with the goal of creating a correlation matrix. After the tutorial, different use cases of Cortex are discussed, along with limitations. Toward highlighting clinical applications, this paper also provides 3 easy ways to implement examples of Cortex use in real-world settings. By understanding how to work with digital phenotyping data and providing ready-to-deploy code with Cortex, the paper aims to show how the new field of digital phenotyping can be both accessible to all and rigorous in methodology.

Background: Growing interest has motivated recent studies to examine differences in recovery after sport-related concussion (SRC) by sex. However, heterogeneity in study design, participants, and recovery outcomes has led to mixed findings. Further work is needed to evaluate potential differences by sex and to investigate the role of related characteristics, such as sport contact-level, in recovery timelines. This study aimed to investigate whether concussion recovery trajectories differ by sex, considering a priori clinical and demographic covariates, and accounting for the sequence of recovery outcomes. Our secondary question was whether sport contact-level modifies the relationship between sex and time to outcomes. Using data from the Ivy League–Big Ten Epidemiology of Concussion Study, we included SRCs reported across five academic years; 2015–2020 (February 2020). We used Cox proportional hazards regressions to estimate associations between sex and time from injury to three outcomes: (1) symptom resolution, (2) return to academics, (3) return to full play, accounting for measured confounders. Results: Among 1160 SRCs (male, n = 667; female, n = 493) with complete data, median age overall was 20 years (25th-75th percentiles:19–21), and most occurred among athletes playing high-contact sports (78.0%). Males were slightly more likely to complete symptom resolution over time compared to females (HR = 1.18, 95%CI = 1.05–1.33), but results were attenuated in fully adjusted models (HR 1.13, 95%CI = 0.99–1.29). Similarly, the HR of full academic return for males compared to females was 1.22 (95%CI = 1.07–1.38), but was attenuated in fully adjusted models (HR = 1.11, 95%CI = 0.97–1.28). The HR of full return to play for males compared to females was 1.14 (95%CI = 1.02–1.28), and was attenuated after adjustment (HR = 1.06, 95%CI = 0.93–1.20) as well. The interaction between sex and playing a high/low-contact sport was not statistically significant across models, though differences were apparent. Conclusions: Among a cohort of collegiate athletes with SRC, recovery timelines appeared similar between male and female athletes, adjusting for measured confounders. Differences by sex, considering sport contact-level, were evident and may be important clinically and in future studies. This study used robust methods, accounting for nesting in the sequence of RTP outcomes. Results inform concussion management protocols and planned qualitative work to further elucidate how collegiate athletes experience concussion recovery. Key Points: Heterogeneity in study design, participants, and recovery outcomes has led to mixed findings in determining differences in recovery trajectories after concussion by sex. We found that having longer time to symptom resolution, and also the sequence of having academic return before symptoms resolve and longer time to academic return were confounders in the relationship between sex and RTP timelines. Time to sequential recovery outcomes appeared similar between male and female athletes, adjusting for observable confounders. Further differences by sex were evident when considering contact-level, and may be important to consider clinically and in future research. Results indicate that differences in concussion recovery trajectories by sex may be largely attributed to and driven by differences in sports with a men’s or women’s team only, such as football, and this should be explored further.

Much work in the parimutuel betting literature has discussed estimating event outcome probabilities or developing optimal wagering strategies, particularly for horse race betting. Some betting pools, however, involve betting not just on a single event, but on a tuple of events. For example, pick six betting in horse racing, March Madness bracket challenges, and predicting a randomly drawn bitstring each involve making a series of individual forecasts. Although traditional optimal wagering strategies work well when the size of the tuple is very small (e.g., betting on the winner of a horse race), they are intractable for more general betting pools in higher dimensions (e.g., March Madness bracket challenges). Hence we pose the multi-brackets problem: supposing we wish to predict a tuple of events and that we know the true probabilities of each potential outcome of each event, what is the best way to tractably generate a set of n predicted tuples? The most general version of this problem is extremely difficult, so we begin with a simpler setting. In particular, we generate n independent predicted tuples according to a distribution having optimal entropy. This entropy-based approach is tractable, scalable, and performs well.

Background: Patients with advanced cancer undergoing chemotherapy experience significant symptoms and declines in functional status, which are associated with poor outcomes. Remote monitoring of patient-reported outcomes (PROs; symptoms) and step counts (functional status) may proactively identify patients at risk of hospitalization or death. Objective: The aim of this study is to evaluate the association of (1) longitudinal PROs with step counts and (2) PROs and step counts with hospitalization or death. Methods: The PROStep randomized trial enrolled 108 patients with advanced gastrointestinal or lung cancers undergoing cytotoxic chemotherapy at a large academic cancer center. Patients were randomized to weekly text-based monitoring of 8 PROs plus continuous step count monitoring via Fitbit (Google) versus usual care. This preplanned secondary analysis included 57 of 75 patients randomized to the intervention who had PRO and step count data. We analyzed the associations between PROs and mean daily step counts and the associations of PROs and step counts with the composite outcome of hospitalization or death using bootstrapped generalized linear models to account for longitudinal data. Results: Among 57 patients, the mean age was 57 (SD 10.9) years, 24 (42%) were female, 43 (75%) had advanced gastrointestinal cancer, 14 (25%) had advanced lung cancer, and 25 (44%) were hospitalized or died during follow-up. A 1-point weekly increase (on a 32-point scale) in aggregate PRO score was associated with 247 fewer mean daily steps (95% CI –277 to –213; P<.001). PROs most strongly associated with step count decline were patient-reported activity (daily step change –892), nausea score (–677), and constipation score (524). A 1-point weekly increase in aggregate PRO score was associated with 20% greater odds of hospitalization or death (adjusted odds ratio [aOR] 1.2, 95% CI 1.1-1.4; P=.01). PROs most strongly associated with hospitalization or death were pain (aOR 3.2, 95% CI 1.6-6.5; P<.001), decreased activity (aOR 3.2, 95% CI 1.4-7.1; P=.01), dyspnea (aOR 2.6, 95% CI 1.2-5.5; P=.02), and sadness (aOR 2.1, 95% CI 1.1-4.3; P=.03). A decrease in 1000 steps was associated with 16% greater odds of hospitalization or death (aOR 1.2, 95% CI 1.0-1.3; P=.03). Compared with baseline, mean daily step count decreased 7% (n=274 steps), 9% (n=351 steps), and 16% (n=667 steps) in the 3, 2, and 1 weeks before hospitalization or death, respectively. Conclusions: In this secondary analysis of a randomized trial among patients with advanced cancer, higher symptom burden and decreased step count were independently associated with and predictably worsened close to hospitalization or death. Future interventions should leverage longitudinal PRO and step count data to target interventions toward patients at risk for poor outcomes. Trial Registration: ClinicalTrials.gov NCT04616768; https://clinicaltrials.gov/study/NCT04616768 International Registered Report Identifier (IRRID): RR2-10.1136/bmjopen-2021-054675